From d0d633948ca1d5069d812c803134a8a5fefb189e Mon Sep 17 00:00:00 2001 From: Jonathan Gregory Date: Wed, 7 Aug 2024 16:16:22 +0100 Subject: [PATCH] GDT files --- GDT/0.html | 1890 ++++++++++++++++++++++++++ GDT/0.ps.gz | Bin 0 -> 103322 bytes GDT/1.1.html | 3222 ++++++++++++++++++++++++++++++++++++++++++++ GDT/1.1.ps.gz | Bin 0 -> 147763 bytes GDT/1.2.html | 3140 +++++++++++++++++++++++++++++++++++++++++++ GDT/1.2.ps.gz | Bin 0 -> 153638 bytes GDT/1.3.html | 3497 +++++++++++++++++++++++++++++++++++++++++++++++ GDT/1.3.pdf | Bin 0 -> 287970 bytes GDT/1.3.ps.gz | Bin 0 -> 153945 bytes GDT/1.4.html | 3573 +++++++++++++++++++++++++++++++++++++++++++++++++ GDT/1.4.pdf | Bin 0 -> 291067 bytes 11 files changed, 15322 insertions(+) create mode 100644 GDT/0.html create mode 100644 GDT/0.ps.gz create mode 100644 GDT/1.1.html create mode 100644 GDT/1.1.ps.gz create mode 100644 GDT/1.2.html create mode 100644 GDT/1.2.ps.gz create mode 100644 GDT/1.3.html create mode 100644 GDT/1.3.pdf create mode 100644 GDT/1.3.ps.gz create mode 100644 GDT/1.4.html create mode 100644 GDT/1.4.pdf diff --git a/GDT/0.html b/GDT/0.html new file mode 100644 index 000000000..1594741c9 --- /dev/null +++ b/GDT/0.html @@ -0,0 +1,1890 @@ + + + + +Proposed netCDF conventions for climate data + + + + + + +

+

Proposed netCDF conventions for climate data

+

Jonathan Gregory (1), Bob Drach (2) and Simon +Tett (1) +
+(1) Hadley Centre, UK Met Office; +(2) PCMDI, temporarily at the Hadley Centre

+

10th June 1997

+

+

1 Purposes

+This standard defines a set of conventions adopted in order to promote the +interchange and sharing of files created with the netCDF Application +Programmer Interface (API). This standard is based upon version 2.4 of +netCDF. Documentation of the netCDF API may be found in the ``NetCDF +Users' Guide'', Version 2.4, February 1996, available from URL +http://www.unidata.ucar.edu/packages/netcdf/ or via anonymous ftp at +ftp.unidata.ucar.edu. +

+This standard is intended for use with climate data, +and was designed with data generated by GCMs particularly in mind. +We recognise that there are limits to what a standard can practically cover; +we restrict ourselves to issues which we believe to be of common and +frequent concern in the design of climate metadata. +This standard is mostly additional to the conventions sponsored by COARDS +(ftp://ftp.unidata.ucar.edu/netcdf/Conventions/COARDS). +In the following standard, parts which are identical to +or paraphrases of the COARDS +standard are prefixed with COARDS, +and new material with NEW. +Material from the Unidata netCDF users' guide is marked UNIDATA. +All Unidata recommendations are supported here +unless noted to the contrary. +Comments indicate the places where there are differences +between the standards. +Comments and examples are not part of the standard, +and are given in italic. +

+COARDS: This standard also refers to the +udunits standard +supported by Unidata. +The udunits package is available via anonymous +ftp at ftp.unidata.ucar.edu. +NEW: See section 11 for details of how the package +is used by this convention to define units for physical quantities. +

+

2 Filename

+COARDS: NetCDF files should have the file +name extension .nc. +

+ +

3 Data types

+

+NEW: The netCDF data types +char, short, long, float, and +double +are all acceptable. +All numeric types are signed. +The byte data type, which is functionally identical to char, +is not recommended because its signedness may become ambiguous in +future versions of netCDF. +The COARDS convention deprecates char, rather than +byte. +

+NEW: NetCDF does not support a character +string type, +so these have to be represented as char arrays. +In this standard, we refer to them as type ``string''. +A string array must be implemented as a +two-dimensional character data variable, +serving as a vector of fixed-length strings, +the second dimension of its CDL declaration +(leading dimension in terms of Fortran) +being a constant StringMaxLength, +recorded as a dimension in the netCDF file. +

+ +

4 Attributes

+

+COARDS: This standard describes many +attributes +(some mandatory, others optional), +but a file may also contain non-standard attributes. +Such attributes do not represent a violation of this standard. +Application programs should ignore attributes that they do not recognise +or which are irrelevant for their purposes. +UNIDATA: Conventional attribute names +should be used wherever applicable. +Non-standard names should be as meaningful as possible. +Before introducing an attribute, consideration should be given to +whether the information would be better represented as a variable. +In general, if a proposed attribute +requires ancillary data to describe it, +is multidimensional, requires any of the defined netCDF +dimensions to index its values, +or requires a significant amount of storage, +a variable should be used instead. +NEW: When this standard defines string +attributes which make take +various prescribed values, the possible values are given in lower case. +However, applications programs should not be sensitive to case +in these attributes. +See Appendix A for a list of attributes described by this +standard. +

+ +

5 Global attributes

+

+COARDS: Although not mandatory, +the Unidata-standard attribute history is recommended to +record the evolution of the data contained within a netCDF +file. Applications which process netCDF data can append their +information to the history attribute. +

+COARDS: The Unidata-standard attribute +Conventions +is recommended to reference this standard. +

+NEW: Use of the string attributes +institution and +production is recommended. +The attribute +institution specifies who produced or supplied the data. +We prefer this name to ``center'' or ``centre'' because the two +possible spellings could cause confusion. +The attribute +production indicates how the data was produced. +If it was model-generated, production should name the model +and its version, as specifically as could be useful. +If it is observational, production should characterise it +e.g. "surface observation" or +"radiosonde". +A further string attribute comment is proposed, to contain +extra information about the file, +and additional attributes may be included as required. +The Hadley Centre, for example, will include an attribute +source to name the model integration. +

+NEW: The float attribute +appendices is recommended to record +the version number of the appendices to this standard used by the +application which generated the file (see section 12). +

+NEW: The calendar attribute (see +section 24) may +be recorded as a global attribute. +The global calendar +attribute is interpreted as a default for all time axes. +

+ +

6 Variable names

+

+COARDS: Variable names should begin with a +letter +and be composed of letters, +digits, and underscores. +NEW: Case is significant in netCDF names, +but it is recommended that +names should not be distinguished purely by case +i.e. if case is disregarded, no two names should be the same. +

+ +

7 Data variables

+

+NEW: The netCDF variables which contain the +physical data are referred to as ``data variables'', +also referred to as ``primary variables'' by Unidata. +Apart from the general naming rules for variables (above, +section 6), +the names of data variables are not standardised by these +conventions (since files may in general contain multiple +data variables of the same physical quantity). +

+ +

8 Axes and dimensionality of a data +variable

+

+NEW: A data variable may have any number of + dimensions, including zero, +and the dimensions must all have different names. +COARDS strongly recommends limiting the number to four, +but we wish to allow greater flexibility. +The dimensions of the variable define the axes of the quantity it contains. +Dimensions other than those of space and time may be included. +One example is a dimension in electromagnetic radiation wavelength +for a data variable of radiative flux. +Components of vector or tensor quantities could be contained +in a single data variable by giving the variable a dimension over components. +While there exist advantages for manipulating such a variable in memory, +we see no strong advantage in introducing this complexity into the netCDF +description, and do not recommend it. +Under certain circumstances, one may need more than one +dimension in a particular quantity +(see section 29 concerning multiple time + axes). +For instance, a data variable containing a two-dimensional probability +density function might correlate the temperature at two different +vertical levels, +and hence would have temperature on both axes. +

+COARDS: If any or all of the +dimensions of a data variable have the interpretations +of ``date or time'' (T), ``height or depth'' (Z), +``latitude'' (Y), or ``longitude'' (X) then those +dimensions should appear in the relative order T, then Z, +then Y, then X in the CDL definition corresponding to the +file. +N.B. In terms of Fortran, this means X is the first dimension +of the array. Non-spatiotemporal dimensions +should be placed to the left +of the spatiotemporal dimensions +i.e. as trailing dimensions in terms of Fortran. +

+NEW: Each dimension of a data variable must + have a coordinate variable +supplying the coordinates of the points along that axis +(section 9). +The coordinates of points within the data variable are +the simple ordered tuples formed by associating values from the +coordinate variables. +

+NEW: Dimensions may be of any size, +including unity. +When a single value of some physical quantity applies to all the values +in a data variable, +the recommended means of attaching this information to the variable +is by use of a singleton dimension (a dimension of size unity) +with a one-element coordinate variable. +For example, a data variable on a pressure level would use a +singleton pressure dimension to record the level. +Singleton dimensions also result from contractions, described +in section 23. +

+ +

9 Coordinate variables and +topology

+

+COARDS: A one-dimensional netCDF variable +associated with an axis of one or more data variables +is called a ``coordinate variable''. +A coordinate variable whose dimension name is identical to +its own name is referred to as a ``main coordinate variable'' +in this standard, +when it is necessary to distinguish it +from other types of coordinate variable +(sections 17 and 21). +Each axis of a data variable must have a main coordinate variable, +but other types are optional. +Apart from the general naming rules for variables (above, +section 6), +the names of coordinate variables are not standardised by these +conventions (since files may in general contain multiple +coordinate variables of the same orientation). +

+UNIDATA: The values in a main coordinate +variable must be +strictly monotonic +(all values are different and either increasing or decreasing). +NEW: If a main coordinate variable is +merely a ``dummy'', +not containing any physical +information, its values should be set equal to their indices (0, 1, 2, ...). +This is the netCDF default for an omitted coordinate variable. +This standard requires it to be made explicit. +

+ +

10 Coordinate systems

+

+NEW: A coordinate system for the Earth's +surface which is rectilinear but +based on a polar axis other than the normal geographical axis is referred +to as a ``rotated grid''. +To describe rotated grids, a two-element float attribute +north_pole is attached to the data variable, +specifying the (longitude,latitude) coordinates of the rotated north pole. +If the attribute is absent and relevant, +it is assumed to have the value (0.,90.) i.e. the geographical north pole. +

+NEW: In some systems, +the axes covering the Earth's surface +do not define a rectilinear grid. +We do not wish necessarily to exclude non-rectilinear systems. +For the moment, this standard is undefined for these systems, +and we invite comments from potential users on the appropriate definition. +The COARDS standard excludes non-rectilinear systems. +In principle, any coordinate system can be handled, +albeit clumsily, by replacing the relevant two or more axes by +by a single axis which indexes the points, +and providing ancillary coordinate variables to specify the coordinates, +point by point (see section 17). +

+ +

11 Units

+

+COARDS: The udunits package includes a +file udunits.dat, which lists collections of unit names. The names +given in the most recent version of this file +and their plural forms will be regarded as acceptable +unit names for this standard, with a few modifications which will be +listed in Appendix C to this standard. +COARDS lists some modifications within the standard, +but we would prefer to put in place a means to allow future +modifications to be made easily. +NEW: Users of this standard should not +define their own units, +because this would make their files less portable; +requests for new units should be directed to Unidata. +

+NEW: The udunits package also defines a +means for linear transformation +of units by a scale factor and an offset. +This convention is allowed when it is natural to express a unit in +such a form +e.g. density of sea-water in +kg m¯³ +in excess of 1000 +kg m¯³, +which can be specified to udunits as "kg m-3 @ 1000". +COARDS does not permit the use of this facility. +It can also be used to specify a scale factor and/or offset for a dimensionless +quantity e.g. sea-ice concentration in tenths (no unit). +It should not be used as a means of data compression, +for which an alternative is provided +(see section 34). +

+ +

12 Physical quantity of a +variable

+

+NEW: These conventions standardise three +attributes for specifying +the physical quantity of data and coordinate variables. +All of them are strings: +

  1. units, formatted as per the +recommendations in the Unidata udunits package, +with extensions for time +(see section 26). +This attribute is mandatory unless the quantity is dimensionless +(a pure number). +Dimensionless quantities will not generally have units; +a unit string specifying a linear transformation alone is permissible. +For instance, a unit of tenths is given as "0.1". +There are a few defined dimensionless units, +such as percent. +There is no need for a wide variety of dimensionless units for quantities +like sea-ice concentration, cloud fraction, probability and so on; +this descriptive information is the quantity rather +than the units.
  2. quantity, a short descriptive name. +The quantity attribute is optional for most variables, +but mandatory for coordinate variables +of longitude, latitude, vertical axes and time +(see sections 14, 15, 16 +and 24). +If the quantity attribute is present, +it must be one from a list which constitutes +Appendix E of this standard (see below). +The use of a standard list aims to clarify whether data from different +sources is comparable. +The list will include all quantities which are used as coordinate variables. +The list will also define the physical dimensions +of the quantity by stating an appropriate unit. +The units and the quantity of the data variable must be +compatible. +Appendix E will indicate the version number of the appendices at which +each +quantity was introduced. +In conjunction with the appendices attribute (section 5), +this enables the application to deduce the entire set of distinct quantities +known to the application which generated the file.
  3. long_name, a +long descriptive name. +It should not include the units. +
+The long_name is useful as the title of a plot of a data variable, +or the title of the axis of a coordinate variable. +In its absence, the quantity may be used, +or, failing that, the name of the variable itself. +In addition, other model-dependent attributes may included to define +the quantity of a variable. +The Hadley Centre model will give each data variable integer stash +and submodel attributes, for example. +

+Whether two physical quantities are different or the same is often +not a question with a well-defined answer. +Certainly if they are the same, they must have the same unit, but +various quantities with the same unit may have to be distinguished +e.g. temperature and soil temperature. +In practice, +the most specific description applicable should be used. +We intend to expand the list of quantities in Appendix E on an ongoing +basis +in response to requests by users of this standard, +since we cannot foresee all the possibilities, +and we will err on the side of expansion, rather than restriction, +when it is unclear whether a newly requested quantity is identical to +one already listed. +However, quantities which differ only with regard to +surfaces (sections 30), +linear unweighted contractions (section 23) +or other distinctions which can represented simply by attributes described +in later sections will not be regarded as +distinct physical quantities +e.g. net downward shortwave radiation at the surface, +and net downward shortwave radiation at the top of the atmosphere. +

+NEW: The optional modulo attribute + of a data variable, if present, +records a number which can be added or subtracted +without altering the validity or physical significance of +the quantity. +This is most likely to be useful for longitude coordinate axes, +with a modulo of 360. +

+NEW: We note that the Unidata-standard +FORTRAN_format attribute +may be useful for both coordinate and data variables. +

+ +

13 Topology of an axis

+

+NEW: An axis with ``circular topology'' is +one +which can be legitimately +transformed by shifting all the points one place along the axis, +moving the last point to the beginning, any number of times. +The main coordinate variable of an axis with circular topology is distinguished +by the presence of an attribute topology="circular". +A longitude axis which circles the whole globe is an example. +The value linear or the absence of this attribute +indicates an axis with ``linear topology''. +The topology is indicated only by the main coordinate variable, +but since it is the property of the axis it applies to any ancillary +coordinate variables (section 17) as well. +

+NEW: When a circular axis is rotated, the +main coordinate values must be +altered in order to remain monotonic. +Therefore the main coordinate variable +requires a modulo (section 12). +

+ +

14 Longitude dimension

+

+COARDS: Coordinate variables representing +longitudes must always explicitly +include the units attribute; there is no default value. +The units attribute will be a +string formatted as per the recommendations in the Unidata udunits +package. +The recommended unit of longitude is degrees_east (eastward +positive). Also acceptable are degree_east, degree_E, and +degrees_E. The unit degrees_west (westward positive) is not +recommended because it implies a negative conversion factor from +degrees_east. +

+NEW: A longitude coordinate variable must +have an attribute +quantity="longitude". +COARDS: Such a variable is also +identifiable from its units +string. The COARDS convention relies on the unit as the only way to +identify a longitude variable. This standard uses the quantity, but +requires the units to be specified as well for compatibility with +COARDS. +

+NEW: Longitude axes should have the +attribute modulo=360, +indicating that they may be interpreted modulo 360. +Thus, for example, +-180, 180, and 540 are all valid representations of the International Dateline +and 0 and 360 are both valid representations of the Prime Meridian. +COARDS assumes that longitudes may always be treated in this way. +Since we have introduced the modulo attribute, +we require that it should be specified to indicate this. +But the presence of a modulo +does not mean that the axis necessarily has circular topology +(section 9); +a longitude axis covering only part of the globe cannot have its points +rotated. +COARDS: Note +that the sequence of numerical longitude values stored in the netCDF +file must be monotonic in a non-modulo sense +for a main coordinate variable of longitude. +

+ +

15 Latitude dimension

+

+COARDS: Coordinate variables representing +latitudes must always explicitly +include the units attribute; there is no default value. +The units attribute will be a +string formatted as per the recommendations in the Unidata udunits +package. +The recommended unit of latitude is degrees_north. Also acceptable +are degree_north, degree_N, and degrees_N. +

+NEW: A latitude coordinate variable must +have an attribute +quantity="latitude". +COARDS: Such a variable is also +identifiable from its units +string. The COARDS convention relies on the unit as the only way to +identify a latitude variable. This standard uses the quantity, but +requires the units to be specified as well for compatibility with +COARDS. +

+ +

16 Vertical (height or depth) +dimension

+

+NEW: Whereas the two horizontal dimensions +are usually longitude +and latitude, whose direction is well defined, +a variety of quantities may be used for the vertical axis. +The quantity attribute is mandatory for the vertical axis. +The direction of positive, +whether up or down, may be useful for +applications displaying or processing the data. +For this reason Appendix E to this standard +will state the usual direction of positive +for all quantities which are commonly used as vertical axes. +This information should be taken only as guidance, +and may be overridden by the application. +

+The COARDS standard requires the units of +the vertical axis to be selected from a defined list, +in order that this axis can be recognised by its units. +It gives special status to units of pressure, +for which the direction of positive is defined, +and introduces a mandatory positive attribute +for vertical axes with other units. +

+We have adopted a different approach for a number of reasons. +Firstly, to require units for the vertical axis +means defining dimensionless units for any +dimensionless quantity one might wish to use for that axis. +This seems inconsistent with the treatment of a data variable which +happened to contain a dimensionless physical quantity, +for which the standard would not require that units be invented. +Secondly, the quantity attribute is more informative than the units, +and for most practical axes will clarify the direction of positive. +Thirdly, the direction of positive is mostly an issue for displaying the data, +and is to some extent a matter of personal preference. +If we introduce special treatment for the vertical axis, +we should also include information about how any other axis should be +displayed. +For instance, when latitude is shown on the horizontal axis of a plot, +is north on the left or the right? +This is the same kind of question, +but it strikes us as more a matter for a graphics application to consider +than one to be recorded in the data structure. +Fourthly, the vertical dimension is recognisable from the order of dimensions +(see section 8), +which allows any application expecting such a +dimension to know where to find it without any further indication. +The absence of the positive attribute means that vertical axes +with coordinates other than pressure will not necessarily be recognised +as such by COARDS software. +If this is a concern, the positive attribute could optionally be +included +with one of the allowed values up or down, +indicating the sense of the direction of positive. +

+For example, if an oceanographic netCDF file encodes the depth of the +surface as 0 and the depth of 1000 meters as 1000 then the axis would +use attributes as follows: units="meters", +quantity="depth". +If, on the other hand, the depth of 1000 meters were represented as +-1000 +then the value of the quantity attribute would have been height. +The COARDS positive attribute in these two cases would have values +down and up respectively. +

+ +

17 Ancillary coordinate +variables

+

+NEW: A single axis might require more than +one set of coordinate values, +for various purposes described in sections 18, +19 and 20. +The additional sets of values required to specify the axis should +be contained in extra one-dimensional variables +(formally two-dimensional for string-valued variables). +We refer to these as ``ancillary coordinate variables''. +Ancillary coordinate variables are recognised as being associated with their +axes by having the same dimension, +but distinguished from the main coordinate variable for the axis +by not sharing the name of the dimension. +

+ +

18 Component coordinate +variables

+

+NEW: A continuous physical coordinate may +require +more than one number to specify it at each point. +We refer to these as ``components''. +This standard requires that a main coordinate variable should nonetheless +be supplied which represents a combination of the components that can be +used to order the points on the axis. +As usual, the main coordinate variable must be monotonic. +

+NEW: The values of the components +are recorded in ancillary coordinate variables +referred to as ``component coordinate variables''. +Unlike the main coordinate variable, +the components do not need to be monotonic. +The names of the component coordinate variables are recorded as +a comma-separated list in a component string attribute +of the main coordinate variable. +

+An example is the hybrid vertical coordinate +eta = p/p0 + sigma, +used in the Hadley Centre GCM. +Atmospheric model levels are specified in terms of +(p,sigma) +pairs, +where p is pressure, +p0 +is a constant and +sigma +is fraction +of surface pressure (which is variable). +The eta +value is a linear combination of the two, +which cannot be uniquely decomposed back into +(p,sigma). +We would record +eta +as the coordinate variable eta(eta), say, +with p and +sigma +in ancillary coordinate variables +pressure(eta) and sigma(eta). +The eta variable would have a string attribute +component="pressure,sigma". +Information about how to relate the components to the combination +would reside in the application. +

+ +

19 Associated coordinate +variables

+

+NEW: A qualitatively different situation +arises where an axis has a number +of alternative ways of being labelled, +providing different kinds of information. +We refer to the alternative sets as ``associated coordinates'' +and store them in ancillary coordinate variables. +The main coordinate variable records the names of these variables +as a comma-separated list in an associate string attribute. +The main coordinate variable must be monotonic, as usual, +but associated coordinate variables need not be. +

+One example is a vertical axis where one wishes to store both the +physical coordinate and the ordinal model level number. +The latter could be recorded as an associated coordinate variable. +In this instance, if the physical coordinate was the +eta +of the +section 18, +it would also have component coordinate variables. +Another example is the value of a quantity along a one-dimensional +trajectory. +In such a case, we might have a coordinate variable containing +distance along the trajectory and +associated coordinate variables +giving the latitude and longitude of each point. +

+ +

20 Bundles

+

+NEW: If several data arrays containing the +same physical quantity +have one or more identical axes, +but are distinguished by the values of other singleton coordinate variables, +it may be convenient to store them in the same data variable. +The common axes of the separate arrays become axes of the combined variable. +One or more additional axes are introduced to ``bundle up'' the separate +arrays. +Such an axis does not correspond to a continuous physical coordinate. +It acts simply as an index of the bundled-up arrays. +For instance, +the Hadley Centre GCM can generate timeseries of the values of quantities +at individual points. +Typically, timeseries from many different points are produced of the same +quantity at the same sampling times. +It is natural to contain this information +in a data variable with two dimensions. +One dimension is the common time axis, specifying the sampling times, +which are the same for all the points sampled. +The other dimension is not a continuous physical coordinate; +it is simply being used to ``bundle up'' the timeseries, +the points being +irregularly scattered in a space of two or more dimensions. +This axis might have a meaningful coordinate variable, such as +a station number (in which case it must be monotonic), +but otherwise it may be no more than an index. +

+NEW: The singleton values of the separate +arrays are recorded in +associated coordinate variables for the bundling dimension. +They should not be interpreted as continuous coordinates. +In the timeseries example, +the latitude and longitude of the points would be recorded in associated +coordinate variables. +It might be desirable to have a string-valued coordinate variable as well, +to name the points. +For a set of temperature timeseries, +suppose that the dimension points specified the number of points, +and times the number of sampling times. +The data variable might be temperature(times,points) +and the time axis times(times). +The main coordinate variable for the points axis would be +points(points). +The associated coordinate variables would be +latitude(points), longitude(points) +and names(points,StringMaxLength). +This same form could be used for observed timeseries from stations. +

+Another application could be that of vertical +profiles at sets of points; +for example, a vertical temperature profile through the ocean, +or data from a radiosonde. +One dimension is the height or depth, with the physical vertical coordinate. +The other has a main coordinate variable containing just an index, +with associated coordinate variables recording the geographical location of +the points. +

+This section raises the question of how best to store a single timeseries, +or a single vertical profile. +Following the scheme of this section, +it could be contained in a two-dimensional data variable with the bundling +axis being of size unity. +The associated information such as latitude or longitude would then be +stored in singleton coordinate variables, +all associated with the same dimension. +Alternatively, these values could be recorded as +separate singleton dimensions (following section 8). +We have no recommendation for this. +Either scheme could be appropriate; +which is more natural perhaps depends on how the data was extracted from +the continuous axes. +

+ +

21 Boundary coordinate +variables

+

+NEW: Along a dimension, +the values might relate to points (at the coordinate values) +or to contiguous or non-contiguous cells. +The boundaries of the cells should be defined +as well as the cell coordinate values. +The convention is to define an additional two-dimensional +``boundary coordinate variable'' with +a left-hand dimension (trailing dimension in Fortran terms) of size +two. +The values for which this dimension has index 0 +(numbering from 0 i.e. in C notation) supply the boundaries +with the smaller coordinate values, +and those with index 1 the large values, +where ``smaller'' and ``larger'' refer simply to numerical comparison, +not to a physical direction. +The name of the boundary coordinate variable is recorded in a string attribute +bounds of the main coordinate variable. +We recommend that the it should be named +by the coordinate dimension with the prefix bounds_. +Thus, for instance, a coordinate variable lat(lat) +might have attribute bounds="bounds_lat" +pointing to a boundary coordinate variable +bounds_lat(2,lat) (declared as in CDL). +In C notation, +lat[0] gives the coordinate of the first point, +bounds_lat[0][0] its lower boundary, bounds_lat[1][0] +its upper boundary. +In Fortran notation, the declarations are lat(lat) +and bounds_lat(lat,2), +and the relevant elements are lat(1), bounds_lat(1,1), +bounds_lat(1,2). +Supplying upper and lower boundaries separately allows for the possibility +that the cells might not be contiguous; they might even overlap. +

+NEW: Boundary coordinate variables +are recommended if the main coordinate values +are not evenly spaced, +or if the dimension has a size of unity. +If the coordinates are evenly spaced, +and boundaries are not specified, +generic applications may assume that the main coordinates +lie at the centres of their cells. +Boundary coordinate variables may be supplied for ancillary +coordinate variables as well as main coordinate variables. +The upper and lower boundaries of ancillary coordinates are ordered +so as to correspond with those of the main coordinate. +

+ +

22 Point values versus average +values

+

+NEW: A subcell attribute may be +given +to the main coordinate variable for a dimension to +specify whether the data values represent +a point value (subcell="point") +or an average over the cell (subcell="cell") in that +dimension. +It is conceivable that the interpretation as point or average values might +differ for the various dimensions. +An important example of the need to distinguish arises for timeseries, +which might be instantaneous values (such as is usual with pressure +measured at synoptic stations) or values representing a time-mean +(such as the rainfall rate derived from the total accumulated over the +observation period). +

+ +

23 Contracted dimensions

+

+NEW: A contracted dimension is one in which +all data points share the entire contracted axis. +It indicates the relationship of the data variable which is +being described to another variable of higher dimensionality. +A contracted dimension will have a size of unity, a single representative +coordinate value and boundary coordinate values supplying the +range of the uncontracted axis. +If the uncontracted axis had the attribute subcell="point", +the boundaries of the contracted axis will be +the extreme point coordinate values of the uncontracted axis; +if the uncontracted axis had subcell="cell", +the boundaries will be the extreme boundary coordinate values. +

+NEW: The contraction attribute +of the coordinate variable specifies the operation which contracted +the axis; +permitted values are detailed in Appendix B and +include sum, mean, max, +min, sd, var. +Contractions are limited to common and +well-defined statistical operations +performed using the values of the data variable and possibly +the coordinate variable, +which may be needed to determine statistical weights. +If weights have been used, +the quantity used for weighting may be recorded in a weight attribute, +and should be one of the quantities listed in Appendix E +(section 12). +Note that some operations, such as taking a weighted sum or a variance, +imply that the quantity in the data variable does not have the same +physical unit as the quantity which had the uncontracted axis. +In this case, the units and quantity should be altered; +the contraction does not suffice. +This standard also states that two variables related by a +non-linear contraction, +such as the standard deviation, +necessarily have different quantity attributes, +even if they have the same units. +This will be reflected in Appendix E. +For instance, ``orographic height'' +and ``standard deviation of orographic height'' are +regarded as different quantities, +while ``maximum orographic height'' is treated as the same quantity +as ``orographic height''. +

+NEW: The optional min_interval and + max_interval attributes +specify the minimum and maximum distance between +two adjacent coordinates of the uncontracted axis. +Equality implies that the uncontracted coordinates were evenly spaced. +

+The most obvious example of a contracted dimension +is the time information belonging to an array +where time is not otherwise a dimension. +The beginning and ending times of the period over which a time-mean +(for example) has been +calculated form the lower and upper time boundaries of a contracted +dimension of time. +(See below, section 24, for more on time axes.) +Zonal and meridional means are further common examples of contracted axes, +whose boundary coordinate +variables would specify the ranges of longitude and latitude +over which the mean had been formed. +

+NEW: In some cases, several axes are +contracted simultaneously. +If the operation cannot be peformed on them separately, +probably because it involves a non-linear operation, +the contraction has to be recorded as affecting more than one axis at a time. +This is done by replacing all the affected dimensions with one new +singleton dimension +carrying the contraction attribute. +The information from the individual contracted axes is recorded on +singleton coordinate variables associated with the contracted axis +(see section 19). +For instance, the standard deviation involves a non-linear operation, +so an area-weighted standard deviation must be performed on both +horizontal axes at once. +On calculating the area-weighted standard deviation of an array +of temperature(lat,lon), +we might record the result in sd_temperature(area), +with area as a dimension of size unity, +having a dummy coordinate variable area(area) +with attributes contraction="sd" and +weight="area" +and associated coordinate variables lat(area) and lon(area) +with boundaries, units, etc. +

+ +

24 Time axes

+

+NEW: A ``time axis'' is one which +represents date and time, +which we will refer to hereafter just as ``time''. +Time is usually defined by a set of up to six numbers +(year, month, day, hour, minute, second) of various data types. +It would be possible to define it thus in a netCDF file, +using six ancillary coordinate variables. +However, it is more efficient and for many purposes more convenient +to encode a time into a single number, +giving the elapsed interval since a certain reference. +The units of the interval and the reference time are +matters of convention. +

+NEW: A ``calendar'' defines the set of +valid dates +(year-month-day combinations). +The standard calendar is the Gregorian +(the calendar of udunits), +but climate models do not always use this. +For instance, in the calendar of +the Hadley Centre GCM, all months have 30 days. +The elapsed interval between two dates will not necessarily be the same +in two different calendars, +because there may be different numbers of valid dates between them. +For example, +the interval between 1 February 1996 and 1 March 1996 is +29 days in the Gregorian calendar, but 30 days in the Hadley Centre +model calendar, +since 30 February is a valid date in the latter. +Therefore the encoding of a time into an elapsed interval will +depend on the calendar, +and it is necessary to know the calendar when converting. +

+NEW: This standard permits the use of the +standard calendar +(below, section 25) +and of other calendars (section 26). +In addition, we propose and recommend a new calendar +(section 27), +which has the advantage of being able to encode a date which is legal +in any of the other calendars, +and which can therefore be used in all cases. +

+COARDS: Time coordinate variables must +always explicitly include +the units attribute; there is no default value. +NEW: The quantity is also +mandatory, +having one of the values time or unitime +(see section 27 concerning the latter). +A time coordinate +variable will be identifiable by its units and quantity. +The quantity in conjunction with the calendar +attribute, when required, will indicate +the calendar in use and how it has been encoded. +The calendar attribute is described in the following sections. +If a time coordinate variable has no calendar attribute, +the global calendar attribute (section 5), +if present, applies to it. +NEW: The attribute time_format can + be included to specify +a format for printing the date and time, +according to the conventions of the +Unix (TM) date command. +

+ +

25 Gregorian calendar

+

+COARDS: Intervals between two times in the +standard +Gregorian calendar can be calculated by the udunits package. +This calendar is actually the mixed Gregorian/Julian calendar system, +as followed in England. +Dates prior to 1582-10-15 are assumed to use the Julian calendar. +Time coordinates in this calendar may have units of time formatted as per the +recommendations in the Unidata udunits package, +which specify a unit and a reference time +e.g. a unit of seconds since 1992-10-8 15:15:42.5 +indicates seconds since 8 October 1992 at 3 hours, 15 minutes and +42.5 seconds in the afternoon, in Universal Coordinated Time +(time zones can also be handled). +

+NEW: The file udunits.dat defines +second, minute, hour and day as units of time. +Units of months and years must not be used, +because both are of variable length. +Since udunits defines a +year as exactly 365 days and a month as exactly a twelfth of a year, +use of these units will not always give the expected results. +For example, 1 month since 1997-4-1 0:0:0 is treated by udunits as +30.4368 days since 1997-4-1 0:0:0, +not 1997-5-1 0:0:0. +Note, however, that this standard deals only with the encoding of time +in the netCDF file. +An application which processes such a file may have an interface +which handles intervals defined in months or years, if desired. +

+NEW: This standard recommends that +Gregorian +times be given in units of +days since 1-1-1 i.e. midnight on 1 January 1 AD, +in data type double. +(Note that udunits treats 0 AD as identical to 1 AD.) +This data type gives a precision of about 16 decimal digits, +which means that it can resolve tenths of a second +for years of up to around 10,000 AD. +The larger the year, the worse the absolute precision. +If this precision is not sufficient, use of a different reference time, +in order to make the year smaller, is recommended. +

+NEW: If there is no calendar +attribute applying to +a time coordinate variable in any standard unit of time, +the coordinates are assumed to be in the normal Gregorian +calendar, encoded following the udunits convention. +This can be made explicit by setting calendar to standard +or gregorian. +

+ +

26 Non-Gregorian calendars

+

+NEW: It is recommended that +times in other calendars should +be encoded in data type double with a units string +of days since 1-1-1, interpreted as +days since midnight on 1 January of year 1 +in the calendar concerned. +If a different base time is chosen, the udunits syntax +``days since date-and-time'' should be used. +The udunits package itself can process only +the standard calendar. +An extension will be required for other calendars. +Shorter units than days may be used, +but units of months and years are not allowed, as above. +

+NEW: Apart from the Gregorian, +calendars recognised by this standard are +julian for the Julian calendar (in which all years divisible +by four are leap years), +noleap for a calendar with 365 days in every year, +and 360 when each month has 30 days in every year. +If any other calendar is used, a suitable description should appear +in the calendar attribute, +but generic applications cannot be expected to be able to encode and +decode times in the calendar concerned. +

+ +

27 Unimonth calendar

+

+NEW: This standard proposes +a new means of representing times, +namely the ``unimonth'' calendar, +which assumes all months have 100 days. +A time coordinate variable in the unimonth calendar is indicated by +a quantity attribute of unitime. +This standard recommends that times should be given in +days since 1-1-1 +in data type double +in the unimonth calendar. +Dates can be converted into this unitime form and back +unambiguously, according to + +
days= ((year-1) x 12 + +(month-1)) x 100 + (day_in_month-1) +
+ (hour + (minute + +second / 60) / 60) / 24 +
+This formula defines the time 0 days +as midnight on 1 January in year 1. +We are not, of course, suggesting that the unimonth calendar should +be adopted as the calendar of a climate model, or in place of the Gregorian +calendar. +It is introduced only to provide a convenient means of encoding times. +

+NEW: The main reason for proposing this +convention is that +all dates (year, month, day) in any of the calendars recognised by this +standard (sections 25 and 26) +are valid in the unimonth calendar. +Hence any such dates can be converted +into and out of unitime. +Moreover, no knowledge of the specific calendar is required to do this. +This knowledge is still required for calculating some intervals of time +(see below, section 28), +so when the unimonth calendar is used to encode a time axis into unitime, +the calendar attribute should be used to record the original calendar. +

+ +

28 Intervals of time

+

+NEW: An interval of time is the difference +between two times. +The quantity is as for absolute time; +an interval is distinguished by the lack of a reference time in +its units string +e.g. plain seconds, +compared with seconds since 1992-10-8 15:15:42.5. +Both data variables and coordinate variables might contain intervals of time, +mostly probably expressing elapsed time since some event +e.g. days since the onset of the monsoon. +A further specialised application is in the representation of +climatological time using multiple time axes, +described below (section 29). +

+NEW: To be unambiguous, a time interval +must be expressed in units which have a well-defined value. +We identify two different kinds of time interval. +For intervals defined as numbers of days +or shorter units of time, +there is no problem, as these units are well defined by udunits. +We recommend the use of days with data type double, +but allow the use of shorter units of time instead. +However, months and years are not well defined +(as discussed above, section 24), +so must not be used as units of intervals of time. +We may nonetheless have need to express time intervals in terms of years +and months (and possibly a remainder of days or less), +so we need a means of encoding these unambiguously. +A climatological season, for example, is three months. +

+NEW: Any interval of time expressed as a +number of years, months, days, +hours, minutes and seconds can be unambiguously converted to an +interval in unitime (see above, section 24) +and back without knowledge of the calendar. +This standard recommends that intervals of the second kind above +(involving months or years) should be enocoded in data type double +in days of unitime, +according to the formula + +
days= (years x 12 + month) x 100 ++ days_in_month +
+ (hours + (minutes + +seconds / 60) / 60) / 24 +
+Shorter units than days can also be used. +Intervals of the first kind (not involving months or years) +less than 100 days in length +can also be represented in unitime. +

+An example may help to clarify these two kinds of interval further. +The interval between 1 February 1996 and 1 March 1996 is +29 days in the Gregorian calendar, but 30 days in the Hadley Centre +model calendar. +If we wish to record this interval as a number of days, +we can do so as 29 days or 30 days respectively, +of either ordinary time or unitime. +If, however, our intention is to record that it is one month, +we encode it as 100 days of unitime. +The interval between 1 February and 1 March is 100 days of unitime +(i.e. one month) in all years in all calendars. +The importance of the distinction is that it tells us how to add an +interval to an absolute time. +

+ +

29 Multiple time axes and +climatological time

+

+NEW: There is no bar on a data variable +having more than one dimension in +a particular quantity, so long as the dimensions have different names. A +particular need for this arises for multiple time dimensions, when there has +been a contraction over at least one of them. +This provides a means of describing ``climatological time'', +in which the data does not apply to a specific time or times. +

+NEW: This need must be described firstly +with reference to +the original uncontracted dimension +spanning a period of time which contains repetitions of a cycle or cycles +(most often the seasonal or diurnal cycles). +This dimension can be decomposed into an +absolute time, labelling the beginning of the cycle or cycles, +and offsets giving the phases within the cycles. +The absolute time is coded just as for a normal time axis +(section 24). +We recommend that +the offsets be specified in days, +as described above for intervals of time +(section 28), +with a quantity of time or unitime, +as appropriate. +Offsets in shorter units than days are also acceptable. +The first offset is relative to the absolute time, +the second offset is relative to the first offset, etc. +(In practice, there is not likely to be more than one offset.) +These time dimensions should be successive dimensions of the data variable, +in order to indicate their grouping, +and should appear in order, with the absolute time dimension first +in the sense of a CDL declaration +(last in Fortran terms). +Each offset time variable should have a string attribute wrt +(``with respect to''), +naming the next time dimension in the group. +The first of them might have subcell="point" or +subcell="cell", +and might have boundary coordinate values, +but all the others are taken as point values. +The actual time of a data value is obtained by adding up the +values of all its time coordinates. +Unless the offsets are all exact numbers of months and years, +this addition may require knowledge of the calendar, +which is an attribute of the absolute time axis. +The calendar attribute is not required for the offsets. +

+To take an example, consider a time dimension which has coordinates at +daily intervals from 1 June to 30 June in each of the five years 1990-1994. +This dimension has a size of +30 x 5 = 150. +The first coordinate is 1 +June 1990 and the last is 30 June 1994. This single time dimension can be +decomposed into two: an absolute dimension of size 5 and values at yearly +intervals from 1 June 1990 to 1 June 1994, and an offset of size 30 and values +of 0, 1, ..., 29 days. The data value for 13 June 1992 appears at indexes +2 and 12 (numbering from 0) in these two dimensions, respectively, and its time +coordinate is obtained by adding 12 days to 1 June 1992. A decomposition into +three might be appropriate if the original time coordinate had values at, say, +3-hourly intervals. The third dimension would then have a size of 8 and values +of 0, 3, 6, ..., 21 hours (expressed in days). Such a representation +would be appropriate for quantities measured instantaneously at 3-hourly +intervals. If, on the other hand, they were means over successive intervals of +3 hours, the third dimension would have subcell="cell", +lower time +boundaries of 0, 3, 6, ..., 21 hours and upper time boundaries of 3, 6, +9, ..., 24 hours. +

+NEW: Once a time dimension has been +decomposed in this way, +any of the resulting +time dimensions can be contracted, as for any other dimension. +That is the main reason for performing the decomposition, +and it is most useful when it is the leading dimensions which are contracted, +since a contracted trailing time dimension can be merged into a preceding +uncontracted time dimension with no loss of information. +After contraction, boundary coordinate variables should record the first and +last times of the contracted dimension, and +min_interval and max_interval the extreme spacings between +the +coordinates. +The combination of a contracted absolute time dimension and one or more +offset-time dimensions indicates climatological time. +COARDS recommends use of year 0 to indicate climatological time. +We do not favour this convention. +Firstly, it does not provide any way of recording which years were used +to make the climatology. +Secondly, udunits treats year 0 and year 1 as identical +(which is reasonable because year 0 does not exist--there is +no year between 1 AD and 1 BC). +

+In the example above, we might contract the first time axis. +It would then have a size of one, boundaries of 1 June 1990 and 1 June 1994, +and minimum and maximum intervals of one year. +The data variable would contain 5-year means for the individual days of June. +If we contracted the second dimension as well, +it would have a size of one also, boundaries of 0 and 29 days +(or 0 and 1 month, for time-means over the days +rather than instantaneous values), +minimum and maximum intervals of 1 day. +This would be a 5-year June mean. +The first contracted dimension tells us it is made from five years, +and the second that it uses one month from each year. +If we had retained the third dimension uncontracted, +we would have a set of values for times at 3-hourly intervals +within an average June day. +

+ +

30 Special surfaces

+

+NEW: A data variable may be defined as +existing on a surface, +the general description of a surface being a set of points which have +two or more independent dimensions and other dimensions which +are single-valued functions of the independent dimensions. +Existence on a surface is a property of the data variable, +and should be indicated by giving it a surface attribute +identifying the surface, +one from a list which constitutes Appendix D of this standard. +To provide a longer description of the surface than the surface +attribute itself, +a surface_name attribute may be included. +For example, a data variable of pressure at mean sea level would have an +attribute surface="sea level". +We intend to expand Appendix D in response to requests, +as for Appendix E. +Note that surfaces which can be defined by the numeric value of physical +quantities will not be included. +For example, the surface 1.5 m above the ground is not a special surface +in the sense of this section, +as a variable can be described as being on this surface +by giving it a singleton height coordinate with a value of 1.5 m. +

+NEW: If a plain description is insufficient + to characterise the surface, +then the values of the dependent coordinates which define it must be +supplied as additional data variables. +These variables should be named as a comma-separated list +in an attribute surface_coords. +If it were necessary to know the depth +of mean sea-level below ground, this could be done by specifying +surface_coords="mslheight" and including a data variable +mslheight(lat,lon) (CDL declaration), +lat and lon also being dimensions +of the pressure variable. +

+ +

31 Invalid values in a data +variable

+

+NEW: Invalid values are not permitted in a +coordinate variable, so this +section applies only to data variables. Invalid values are any which fall +outside the valid range or equal the fill value, as indicated by the +Unidata-standard attributes described here. An invalid value indicates bad data +i.e. a software problem, which is a different circumstance from unknown or +missing data (see section 32). +

+UNIDATA: The attribute valid_min +is a scalar specifying the +minimum valid value for a variable. The attribute valid_max specifies +the maximum valid value, while valid_range is a vector of two numbers +specifying the minimum and maximum valid values, in that order, equivalent to +specifying values for both valid_min and valid_max +attributes. Any of these attributes define the valid range. The attribute +valid_range must not be defined if either valid_min or +valid_max is defined. Generic applications should treat values outside the +valid range as invalid. The type of each valid_range, +valid_min +and valid_max attribute should match the type of its variable. +The +Unidata special treatment of byte type is not included here as we do +not +recommend use of that type (see section 3). +

+COARDS: A scalar attribute with the name +_FillValue and of the +same type as its variable is used as the fill value for the variable. The +netCDF package defines a default fill value for each type of variable, so it is +not necessary to define your own _FillValue attribute if the default +is +suitable. The purpose of the fill value is to save the applications programmer +the work of prefilling the data and also to eliminate the duplicate writes that +result from netCDF filling in undefined data with its default fill value, +only to be immediately overwritten by the programmer's preferred value. This +value is considered to be a special value that indicates undefined data, and is +returned when reading values that were not written. The _FillValue +should be outside the range specified by valid_range (if used) for a +variable. NEW: In cases where the data +variable is packed using the +scale_factor and add_offset attributes (section 34), the +_FillValue attribute applies the numbers as packed, so they must be +checked against it before unpacking. +

+NEW: If none of valid_min, +valid_max or valid_range +is defined then generic applications should define a valid range by using the +fill value (whether defined explicitly or by default); if the fill value is +positive then it defines a valid maximum, otherwise it defines a valid +minimum. For integer types, there should be a difference of 1 between the fill +value and this valid minimum or maximum. For floating point types, the valid +extreme should have a magnitude which is half the magnitude of the fill value. +We recommend a factor of two, rather than a difference of one bit, because +it is easier for applications programmers. There is no special treatment for +byte as we do not recommend that type (see section 3). +

+ +

32 Missing values in a data +variable

+

+NEW: Missing values are not permitted in a +coordinate variable, so this +section applies only to data variables. The missing_value attribute +indicates a value that is used for data that are unknown or ``missing''. This +attribute is not be treated in any special way by the netCDF API, unlike the +_FillValue attribute (section 31). +The missing_value should be outside the valid range +(section 31), so that generic applications +will treat it +appropriately. +The netCDF data type of +the missing_value attribute should match the netCDF data type of the +data variable that it describes. In cases where the data variable is packed via +the scale_factor and add_offset attributes (section 34), +the missing_value attribute +matches the type of and should be compared with the data +after unpacking. +This standard is unlike COARDS in giving +a particular interpretation to the +distinction between missing_value and _FillValue. +

+ +

33 Compression by gathering

+

+NEW: To save space in the netCDF file, +it may be desirable to eliminate points +from data arrays which are invariably missing. +Such a compression can operate over one or more adjacent axes, +and is accomplished with reference to a list of +the points to be stored. +The axes to be compressed might not be all the axes. +For example, we might wish to compress a three-dimensional +longitude-latitude-depth array of soil temperatures +by eliminating sea points at all depths. +In this case, only the longitude and latitude axes would be affected +by the compression. +The list is constructed by considering a mask array which has just the +axes to be compressed, +and mapping this array onto one dimension without reordering. +The list is the set of indices +in this one-dimensional mask of the required points. +In the compressed array, +the axes to be compressed are all replaced by a single axis, +whose dimension is the number of wanted points. +The wanted points appear along this dimension in the same order they +appear in the uncompressed array, with the unwanted points skipped over. +Compression and uncompression are executed by looping over the list. +

+NEW: The list is stored as the +coordinate variable for the compressed axis of the data array. +Thus, the list variable and its dimension have the same name. +The list variable has a string attribute compressed, +containing a comma-separated list of the dimensions which were affected +by the compression in the order of the CDL declaration of the +uncompressed array. +The list, the compressed arrays and the original dimensions +are written to the archived netCDF file. +The uncompressed arrays can be reconstituted +using this information. +In the example above, suppose the original data variable is +soilt(level,lat,lon). +We construct a list landpoint(landpoint) +containing the indices of land points +in a mask array landmask(lat,lon). +We compress the data variable to a new variable +landsoilt(level,landpoint) +and record the original dimensions in an attribute +compressed="lat,lon" of the landpoint list +variable. +

+ +

34 Compression using a scale and +offset

+

+COARDS: This standard +endorses the use of the optional Unidata-standard attributes +scale_factor and add_offset +for data and coordinate variables. +These attributes +can be used to provide simple number compression (packing), +to store low-resolution +floating-point data as small integers in a netCDF file. +After the data values of the variable have been read in, +they are to be multiplied by the scale_offset, +and have add_offset added to them. +If both scale_factor and add_offset attributes are present, +the +data are scaled before the offset is added. +When scaled data are +written, the application should first subtract the offset and then divide by +the scale factor. +NEW: This procedure is concerned only with +storage. +It does not affect the unit of the quantity. +For instance, a pressure variable with values in the range +900.0-1100.0 Pa +could be converted to short integers in the range +±20000 +by subtracting 1000 and dividing by 0.005 i.e. multiplying by 200. +The units of the compressed variable are still recorded as pascals. +

+COARDS: This standard is more restrictive +than the netCDF Users' +Guide with respect to the use of the scale_factor and +add_offset +attributes; ambiguities and precision problems related to data type conversions +are resolved by these restrictions. If the scale_factor and +add_offset attributes are of the same data type as the associated variable + no +restrictions apply; the unpacked data is assumed to be of the same data type as +the packed data. However, if the scale_factor and add_offset +attributes are of a different data type from the associated variable +(containing the packed data) then in files adhering to this standard the +associated variable may only be of type short or long. +We +exclude byte on grounds discussed in section 3. The +attributes scale_factor and add_offset (which must match in +data +type) must be of type float or double. The data type of the +attributes should match the intended type of the unpacked data. (It is not +advised to unpack a long into a float as there is a potential +precision loss.) Users should note that Unidata may provide a built-in +means of packing data in netCDF files in future. +

+

+

+

A Attributes

+

++ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
+Attribute Section(s) Description
add_offset 31 34 +Additive offset for packing data
+appendices 4 +Version number of these appendices
+associate 19 +Identifies variables containing +alternative sets of coordinates
+bounds 21 +Identifies a variable containing boundary +coordinate values
+calendar 4 24 25 +26 27 29 + +Calendar used for encoding time axes
+comment 4 +Additional information about the file
+component 18 +Identifies variables containing +components of a coordinate variable
+compressed 33 +Records the dimensions which have been compressed by gathering
+contraction 23 +Records how an axis was contracted to a point
+Conventions 4 +Identifies the netCDF standard
+_FillValue 31 +Indicator of invalid data
+FORTRAN_format 12 Format +for printing the +values of a variable
+history 4 +Evolution of the data in the file
+institution 4 Who +made or supplied the data
+long_name 12 Long +description +of a physical quantity
+max_interval 23 29 +The maximum separation between points on an axis +before contraction
+min_interval 23 29 +The minimum separation between points on an axis +before contraction
+modulo 12 +Arithmetic modulo of a coordinate +variable
+north_pole 10 +Geographical location of rotated +North Pole
+positive 16 +Direction of positive for a vertical axis
+production 4 How +the data was produced
+quantity 12 14 15 +16 24 27 +Short description of a physical quantity
+scale_factor 31 34 +Multiplicative factor for packing data
+subcell 22 29 + +Indicates whether data values are points +samples or cell averages
+surface 30 +Identifies a surface on which a quantity +is defined
+surface_coords 30 +Identifies dependent variables +defining a surface
+surface_name 30 +Description of a surface
+topology 13 +Topology of an axis (circular or not)
+time_format 24 Format for +printing a time and date
+units 12 14 15 24 +28 +Units of a physical quantity
+valid_max 31 Largest +valid value of a variable
+valid_min 31 Smallest + valid value of a variable
+valid_range 31 Smallest + and largest valid values +of a variable
+weight 23 +Weighting for points used when performing +a statistical contraction of an axis
+wrt 29 +Indicates the time axis to which a +time offset applies
+

+

+

+

B Contractions

+

++ + + + + + + + + + + +
+Contraction Units Quantity Description
max u N + Maximum
+min u N Minimum
+mean u N Mean
+sd u Y Standard deviation
+sum without weight u N Sum
+sum with weight +u x w + Y Weighted sum
+var +u² + Y Variance
+

+

Units: u means the units of the original quantity, +and w the units of the weighting quantity +

Quantity: ``N'' means +the quantity does not change, ``Y'' that it does change +

+

C Modifications to +udunits.dat

+None at present. +COARDS: The unit degrees is not +permitted, because +it creates ambiguities when attempting to +differentiate longitude and latitude coordinate variables. +This unit does not appear in the current version +of the file. +

+

D Surfaces

+This Appendix is not yet available. +It will contain entries such as: +

+

++ + + + + + + +
+Surface Description
surface Surface of the land or sea
+sea level Mean sea-level +
+TOA Top of the atmosphere +
+

+

+

Surface: Case, spaces and punctuation are not +significant in the name of the surface. +

Description: This is an explanation of the surface for +the purpose of this Appendix only. +Applications should not rely on this Appendix as a lookup table for +explanations, however. +If the surface needs a description, +it should be attached to the variable in a surface_name attribute. +

+

+

E Quantities

+This Appendix is not yet available. +As well as existing as part of this standard, +it will be made available +as a flat ASCII file for use by applications programs. +It will contain entries such as: +

+

++ + + + + + + + + + + + + + +
+Version Quantity Unit Positive Description
1.0 height +m U Height above the surface of the land or sea
+1.0 depth m D Depth below the surface of the +land or sea
+1.0 pressure Pa D Pressure
+1.0 longitude degree_east - Longitude
+1.0 latitude degree_north - Latitude
+1.0 time d - Time (in any calendar except +unimonth)
+1.0 unitime d - Time (in the unimonth calendar) +
+1.0 temperature K - Temperature
+1.0 soil temperature + K - Temperature of the +soil
+1.0 precipitation rate + +kg m¯² s¯¹ + - +Rate of precipitation in all phases together
+

+

+

Version: The version of the appendices at which this +quantity was introduced. +

Quantity: Case, spaces and punctuation are not +significant in the name of the quantity. +

Positive: For quantities commonly used as vertical axes, +``U'' means that values increase upwards along the axis, +``D'' means downwards. +No value is indicated for quantities not used in this way. +

Description: This is an explanation of the quantity for +the purpose of this Appendix only. +Applications should not rely on this Appendix as a lookup table for +explanations, however. +If the quantity needs a description, +it should be attached to the variable in a long_name attribute. +

+

+

+Jonathan Gregory | jmgregory@meto.gov.uk +
+ + diff --git a/GDT/0.ps.gz b/GDT/0.ps.gz new file mode 100644 index 0000000000000000000000000000000000000000..7b21117431c0670ef8a89267df65acbbd03184e8 GIT binary patch literal 103322 zcmV)rK$*WEiwFn<4xuvu128Uda{%o9{f{lTksu2G9)CqG76Z(~K3nABsl%*>_hN3b zk_!Vnz5UvahJnUlp_Z;}y|E=fNp^Sdh5z@>B;%80=E-xa?2+2{*t!)F{3e6Jh~P*5 zm;c*8{OV`#zx?)JKm6?Ua|i$BzohRzeE!3?-~HXgmw)>D58waKhsU2kPG3EwZ~yQg z-+ld?-~RC6pQg+IzWn=ie)z}Fzx?`(&wuwY|Hp@iUw!-45C8W0cORA<{^{!8fB)f|AO7L*H@{fE!L`~3GGe)#a+Qtp?3`1=2OaqfSA z_yP-l`0m5+zx~sPFEQ?~zWw8OUo2%m{L_bDKYaVgAO7&iwZop`&)wzcKKz|~dR#l; z`NMC%{^lRP`}T`pefZ(;u2&Df|M1)<%il?SYlU!AC{8-@ZGoY*;473HJj6;dnoyR`QfVvBG!^W z`}>az@y~0(XU`v(s(!gf`t#pEeD?V{MEzX|M~ghvw!->k3W3( z`8VJH?(-i$tWm!H{&(L$eD(S7zW?yBV1N348OsOnAL8lZ_kaA|O8RXfO+SBJXn?fe zEd%<8-~I9XC-<==t!4kWg&Dv8=Hr^<$A57TDw2Qrng8?0pZN#>XZZT74{Kp7Ke_+$ z#kb#l@%ayb{`#x6z<>Pk{kOmS;}2hd`_0nQKdqF1Zb;vJ_|3!j3wM8jQiv)+$6OMw zk!xGl8m~jMbi+p#=Vvt%{8l3U8Ef`W|G3si#npM&OfmI?fB5{%FW1!YPyhHYzyJJ) z@4o)`hcEv4-FFKYR_x<4-Vf`4->q}$;rl=S>%-^&`u&6Zxm$kz;deiLECEY_e}=EV z{qFPM{SITU6;|N?SXig#xQ^L>TN}Rg!yi6;tm9X>YZ+en?7uB@?T?@T?yCv;LfAht+s{nggD*5;DVVxI${KLo3zWVmtbt--2^DmdCEHz#w1vTtn*6#Z9tEE}rf3=MJl6$T1+E)vPt*(*2|F|;8 z8u8Z{?Ozsn?Yia9%EN0Pf=M)g{KL0@Sj#L7^Wm#STCVludM;7E`0daB?dvap`0d99 z`111~K3}84?>>C$NARcVVGub`VQgAns|618ZD=iYoq>|2~TCn`(Uh`(**5h_FWdVdas}>fjgnnW9 z0{;Cvf0lf{{<5;*cmKBjh0nhGqT;{#1M(>T{eRUzWlWa8ifcZ8|M{O*&h*RNTjt`g zN*`CwTgyE?EW-7}zZdvd!h;p96~m_f`e7YUo)x7Zz&&7xKVUyT@KBZRP@!d{OXD!= zvf%vi#phrA_QRJS{!6pZ(l(hM0ND=jWyC|Hnt_^gow>fB*TSR($v2 z`$fz6uy*Ck$Irg_Vrj+JNc@jgkgxnIowLMWo4Zb$m2Nv{4f)kWnL0~ql!{=OVI@)r z*WO&BuM=sNRZC$!8}-01RsH2hSclpFcgV z0{*lAj+jUy>@_GCQu#5b;!Q>QxT+pUY0kwJjTiYJYH(nbuIn%gQeCL`PXZ+ z{_t^arU2sO)v-|8dc-PzWsGe^l^K9})XC#rXBZKd+qf&+u{m z^N;`h&?;<8;op7x&w^Q z|KWF^|HuFK;r}A`I6tU3|Ml-b{Q95Pz;D0%bIH4^QK%gK8MP|pm}TMm`0+P|QUB>5 z%e-Fxzd(NN_Ej5tsEw*^EJsH-Tt1PF+cf_SM&m!2XRA zVV{4jIhHl~!|xWMT$o{9aiK^cr2cx%M56xpMN?jv;5BHS#HqwQ>Hg zrz6X!tY*5N#Nd>GQt74oL`jws^r~1h3VJc6NJNPwZ+Vh!P@owGBc=t+S1F){p+vf! z^4HJkK+|^nD0f}EK{2D|yHV>Yp<{?Z8F|gm%|fHuuc1V{On^!sHD_l`r6xqO60*)= znn|D*16~uyk zxU4gaPX2%kjn_-jQi5JymX%D4J0tt;B6baft-wZgY_)P-n^IFNXS49FMw>!SSg37#xqfkHOny!wc{wwH}9Fq{|pcYc```3K%TJ{V0qsVX|siq;?M_DVsjQkYD} z604RFs3fLvO^EWMDY~-Q*P5hFiZz9piELAdQAh=^gQiH6X&t6gnm`_{pJ?f6k~L4q z6t_VEMVh-vH&Ap&sf%<2&) zY%^`KwHRm(e+?A7?HZO02I6d*lmSvP?Iwvvhtxh*R7F)V4uz!5xzoVdSE?h&Ort*v z6`8H`rA5)}sIF+)rjlZ1586|-=b)=RMaEV~_#|2R!qS{^!R9d*Ee2&Sn7{^`bT$p9 z22S~}eW{z$S|@3C#$}zY9Y0Yb4H(@|Wfy;qn(+!o7dDGb&OrUctqy1f}j?2&B5misKp;znV1K2q;qo zJB8I0u<^yrPy`%+^yuPCTD6&9{sPF1l06K&MZy{kd&Nlc zkf4l%r+Eo?Y^t{KMaNR?bgwa_EX7-&uZ`lW?7>5N3{unhSkd|ieP^DEdT4n z$v^nhsh%$>W=|2vv8t={)_<|uZL1Nt^iH(e-c3JMDBMzNDKwu_oZdhC{O)mDR8QTF zlz$!&`~KE6kj~4an8y5Tw`6EbWc`6KUEson=CU%1#Axdnwm7clmdJYrisnW%8vm8X zK4Mx*b;J=(?L8S9v_vZU2yId{7!h^7TUy$n!O@7uHdBVkiB^Qj!{Zemg6%hnrM;YkrLf1pQb6~tY@=#P z?Mw!0t2OgpKyKubjg%}U?o-TMvypD2{53MV+O(s6bgi@PfQ(128Cus2KmD}x5i|q( zHjbhw|DUy*P1CMKIzsuB(O)4W79(>M257k zb;X~KCSlKy{Tu}A<)Fy31@z*>wiG*A%3qK`Da5ldXqK^{>L2OYXQhZSq%DgY+*7ip z%&K%=AkDBAq^Fp5zvLMrv)Fq<B|z3)D$B)`O3bTojw`&z4QraD%&QdE zGUh1j+`Ll!%++0hl>$oRL(hY1ColAgxFtSKQu+xu4JL^D|8 zO~X!5gA3%Uqn4>TA}hGHpk~RHc`1O2>o?-sY$p>)lGdV5=i@r2Z&)`f474>#Y3;E9 zh1Hol#LC4K6=*Rxz@{+8ngVa5Q2|Yn&tjIv2#c z_}oE**Cf_7c#GONhX+I7qIvKZwQ-+1`w;DKud|K%8b>toW>f-i6Dq%p2O575Nj!{!|5Gl>7zbFH|<^;X;;;fI~ zU(;gU0||3OoX50iXlPJ#1^KF_Yfc2Bbw0JU$yS<-|K=vqT5sDDaxPd}*)O(&q3-U* zMp~@(Za1+*&N(+UFG31N#9f-Wu?-D+V(i7)kQPh@VMN@f`M!ckYp-qf6&%yrUagCm z5FgzlhUY~LKUU+{tbpDK@u-&rkCqV2r&+1k5$#%9er6|Xf9-EF#dAFlrqGfZ>&{m& zDg*C+dYW3U-$_|L3+W)&>ayx!bPs&42XA`imEn0S8CgTSqj+o+x# z3Js=enj36Z7+7?Og)x?~nYBQm4MM4n&pbp$FlB1@YFCysGKrZel9N;9dvW+7u{u{y+25>?k6 zy1Xaiu|N{a5i608BM~g(E3xR0c$ozoHXArv*!6H`cs;Sp-MUtzf$4VuRvERdagcow zf{yE_5gdkaR-zab;MnhgOmQCeP%w-2AAu0=2Sa-RY*B0W_tg+{$v8ojTt%jsVjgi@m?_qrNjNGvBPLX1%p5 zan%pRFsCU&ez}Qgtl|e^88Ty>MDc9DsLdxYSZc8#*sQ##vFvt5LGvwGR@mggqW5EP zU&6koV9_nC>l=l6wic8mQwAqtQ#Vo;!(6TT%8X=NgNnK=h$zz>rlc9-B&U+hqbfHQ zPYCW#@+Mw#luf}xEdmcZBratA^ODKx$;i(rze3jbI?2Y;HehMU3l_C>HvOhBM|o=u zOsDwUv4B=PSoD4j-Xi*Zm^8%X!ib$}1=YX6{(B5li(ZhbMceswmmk6|mHKjXM`tYW0(Z>T)|J+-9Jb z9D~#`hpSK=K;k*j=aza%9w3aSuQ8CI%qM{|bLBqt6ZIS%g8aMz!G7=oC{(DNPGH4J zpoo(V&#kTU1GgY~Qj!~>2J?&Qb)ZjSSy|@i8M!jD20$tw`Hx!+_5wA=-fy1oKUR}Z zl&m+&Mp^u?ZQ@&4{MlOkW;FZ`#{2sQ1M!!!;o{4?sWTZio0DvS3bD^OBK48;p8+gqh=8?aP<;VW;MqCtX5|#fgq9Tuv#@@1Mi- zA@1ujxR3jK=A-SI8@aD%KDdwj2ovUm`?$|_%m?>zU(bASANTbb+{b-A^MOTGA-mky zV{*RH1nL5vJ#?gc+bOLOih;nrU9VFRR>k(Dv1~Cec-67>(Kstjk{Q9I&P~OPhDrIA z{c1Z(w(CvvR8TR6>%{aY4=)evdm&$}pQAas@?3~Rb-k`uGcRY+S9NI(C>Ft7)MLL= zLfM&Fko8(iED0-IJs7J+7GLVLn!=LoWnws^-ZjM)MiDz*Na1ea+3|tPDN)0g!jkhe znQKIRTo{qseU3K!5w2Y|k6|`dfDtxxMsTf5)-bAq6T+Yotcy3WZNX*(;KmkgHUMsD zL6`(aP7)Y2;`DhvdvtOG7Cr`Z+AMFuX$WgR2X6wVAsjYg$sb&`K3~p0zAV$mz4LT< z$hwI=nQFCK8Y3HBbhP@zb&FjE*aWVMW)bQpKCz8DDiuB{4C{|_C1y=jQ9!znSVm*i zQzI-G45-5>(5ZrHq-()6ykm@-Hp{iaSsmY!-K8GvVf5-6#!i?$wQEYQeEx+=0^IP;mt9G!qUN zTql3|oi3Mid`|C!e|Pcq#AK+-3MRw7Ph8K^QrJ=lRD;&L_K_+Omo&aApKDbvOPj(( z0)E9%0d=b=^7-OuDM%<$38ORK>mmzu60|HSe&SWH!-8xOq$2n%23nc2sueA1VT>h< zfkHhwq%5XYvLQz5Vpgc5EecxmEv9t_Z!s`!3Jo@?^s&L5!q{PrD6QQ})kTa}MxUQ2 z(G`4tnkZTEqdc+lU#gKsg$$XEd(2LjSFM1qqh?Th<-uZk50mS3s%=#7J#1<*$tajW zi$#rz_Ng&Bu&vdJ(~?hE)LH+|`%}?p?Ts4hwmP`$%NA@$Sn^wF-kH zYA)-wJ6bLbl-{-We$rTAt8x7|xB*6ra5OEvC=7)(x}i0;vJcNuaI1@`3Z7LaE|_O5 zJmp%o!WnraJTW=5lt(i-7(C9zleN}ij>bF>i^>eh=XDHb;Y^F;Ix`cm;@Xmk zTk%A>OI%hKoJI-N?~2zoW}>VG2Q|Ymg792X&^696aVb(b7y>Oo&dj#LIeLb*cyT#s zLTVxiJ8%mPJ}Dfq1v)}-7LPJ~nJPs1gcYi9mQ?QYVp6AqGfVq~OTVisRG8w-dd_M= zEmWZQUdW*?vT7IUGDk&qSB;`AJ5)PVFKl_FR2+RAi86(@G*t^-J6&sTj&5e> z5|nllqXD)?#WFHDl;W z_fbA&^q(lB*hsIb6=%73oXM$uBbOvdFM| z@`W94yy@58#=rab6UpdPXa5y5db7?xpH3-0zn_;Gvzlc^SN*W)s*m;465_7H^Q(&) zme9>y58h4H%EJH3td0}^EpaGr#J%s5$@&-Oh9G~*E;{xXjz+v8qpg}1#VXES)97#Q zj`c)4_Y3wBMWWK=L`$`iQnedS;oW(v4%DhNvK8*|alGqXtffb3nzLti+=2Mm!50XEq#xHqyyjZ~sQIFb99d+k^H; zoiU#OlEcZyCuci5t13#*k#*su{v(+j^u>=_+rIuS7HU9Cws_P9Zf5NzqzHWcL;Y1_9MB;{JYaam{Kbr_xi_;fB4UZ;%kyL>64z3nqW9k?U^(#1 zB9Fg%@XP-eaXgmp7CyaqA-;d|A-)5L0{u4+F+BLwV-@L+dPA9>)MIomSytg?=*mIO zILVGt-Q!;0cT##?8RGR-W%29zI{d59tSuxtJDO1(@_@KUvyLJBdH(v9B@Sp;#{?|V4I(O97}9D+uNWxP|L}@|GUEgtaettU zYfaN&BcdPDcE&`H+N}d+XmD^)uF~M(pj;gj7!e!Vp~(XsaXcs+H0eSZ5#5#kcO5Nc zW1;W#>)g|nV>J~eJxrPyi-<2On#kK+oXN9pT{*1m<)kbaqN?@k^qyMUTwLx+SFpS& z=$jAM#j*x*A(d9vs``Mx^g+>-jDZPC-Q?dnJ-%PkKZf+qd*ujq%?anKI#EUr=ZzZN z`W}_4el1HXRCn??dUb^hPvdTvSrA>bXlk5n3a(MMjiaNevLISpdyFeS3QYvVPgim1 zBg7V;kKnK!q1mi&)}oIc9B)JuHs%U4%j{VzDvgV+5_k@XO*v*hjdP!vw~SS+VG&grj_??L_RX$a^zn2~nq;CF8)%g)V|Y_h+hCLKoe$ zcA3c8PCRofWsJ)W|3aAU^^79e)h{;wVr_e%sBI;O?Vvz1JgVcUVj@yL<#L|1z3zJk zJ5{nwow_}zgo7&!El-VV)rmBz>3f_iQ2@b#4kFkH;YIECsXyB|Gz5FxT#a_CTeWzn zXSMoNhtES?F0!=xQ;S>uR@=1|XOTTM5!5!e?UH1Api_@pdtn2Q(nXVLAxYY3n4rgV zi*pYU&c{H9UF17_+Q6~1ebXhO!`(F;s0)(m0^fAOb-YggOahfoDJB%C(k0RDaEI^Snezmo&I1oNOBUfDiUL;lze`%83Q^6fK^jxvSyWD?y_btt*Ru0~~fE+@NZK zmEffow+n2(RZ9d@p@9ImnwJsa(0GsBbpd{MD8QeJ0I%+{dI@PA_E=nqGv0+b{x_tw zp3|ybR4+2D%}VmOcC8qI}jX=cnva_67GOR zgOmxe4~-t1k!j_Wi$#Q)%u|_^0?CvW$jJ>5zD;pDU(WHA-k+v-XZeV!oYoS^mInfM z1rD3A)q-~RkDx={2$neD%7f4Rz?FzcpZ*DObCQjDvwLsn*t}QY|9IWJuYckAY~QVx zvL8p48OhoWF-U#1*>;=r_RTD{X>fD8QKc#|LP^H(iiIm)=Mcv8|#jxm%1rqwZv9y?C?g5hL+I!MLIvp zZ1E@)?NEVN%Y!Qgb`@yL6Sh89;MMY)<3^hWUM&w=!bXAn z_i$?Lq%@Ir-DY3Dw6Bt~Y7%s63T;ZzWjFRZFgt3~xDTb9*G?%(3vArEB6M2)vsF4=NRFjb5V@coH}wTqJgMDxfZYrGIlZ_e#Ic7M>Mc#MD|NKjV$uY7N%g zxD~F4S1U{hTGzxgkLp~K#@a`9c=0Sk6>B0*^Z_(|q{Q6MR8BFi;%ntoWn-euO%!`F zvlgY33Rx#rPYtT4W12W9T}wwz3YIKLO(>&_r4}c>KEI9XN`2EYh`L5`8N`-DsaH4nd}gA)aIPeF^AM-X-e}$?lG3P+L1SKJ z$Tz}ZnlOf=(wHl!q9UVVP)HHfB=hGMM#f3pK7J>plnawW1dfWK@C};^*Ws%gd6nB! zZDgJNO2aNQl-d<1J4~vz)bcWqV)9^gC*`ZRc9mg_5tD%B7{A%|?g|W+Wwzjr$|)E& zE40Ph2G&;Sjg{Cc%lHm93bSQxG!je!ZU`UBQ=fy%6T5xu!B);kPUIwkqk>n_PwQnd zQ%Pr;1r7>c>%eCdCgI`C$Btg30BR*fpQzKUO zV1Xwm5%GrB`c{)HDlrfAQFnwLyB1MfaMX;g-7j0PUYNA|$2g$% zN~YnYjKn10-0^emzRmMWb~%)mu*WRtKB6>>8^{@)^U1&47%BS^z&3vtF^ASn50_Vu@Ve z>8tfGuP`EvH)zp>U?&uRAfv*Qf?Wp}ZiZPQTX4cU^5oF$k1$MaKsK=S+y?A%j)3$a zz4hG9msxsN!%-?tM-I_2o@M0pEsm&6<*3XF{f=hHbcvGwpj`86T`zt;;+S>2RqLkg zwt+lWRG*EHFVHpBRh&-sn@0u42^CfcWX6?!o!K)kHMRd3_V_`nsyFBAwn#X(>zkX& z0XaIVHCxytiR{J#am-(og<12&Ca$)cF`Y>Vnq*l+e@ca&)4ow-jnl#(NE%?C1RLFM z8p7&r)<$2vOCRkHuidO@tGl?3a=m4j?wca$$lkH`uE10-G?mzrVOB;CVCX7w1XdLy z*mxjYctaUw$I)p%uZOsd2Xf1EBHCHGccPX+r`AT+A0vxcXWq|K|Au)Cqu&xVk~yDE zI4I3OXw4rKqGcAkKoV!6pXSxizt)9;4jz;~iLBSMo>*)Onsk{Szpy=r-^c(-!DC|H zcvA+2MDnwYN1OtGcFp5A9;xBH>yX6nQkSL7x;m|ODB364=8wuz5zJm(V&Q;0SqdhV zBq>upPa+*QV9h>WOUtZZPY{jI!1cX@)WLNDW4_@dXZpoh#a3)MN}JZH##5Nb4{wKR zeT{~xXm-(fAoa9Yt5{fXh)_tL#T3qyE}g}2I|4j1li0=TWm@y1v4Zti3W9+r$5-e` z^=RSVv(;%Es3UQwxJeqJbxtocTKJ>HBRmr|EK@!1$>egr(EyTN!KB zUqF!XoSC5_KRdFZ{2q02FAysyE5)^=W+>8m;X}JRpR{lxXjczNc9mqC}sXDMhinC=aBv&roAU&eQ)^x>#&th&QKB z?*F|mBeWmZdP2nNRtD?;m1a~;aFHp`@$&vXpFONiguDCXOL~lcVUs+GD9U&Brju_i7kqF0ynF{3SHD(q+S} zDE;VH;F4wv@eBp7Fh2F0s6H!%M%C&#!UaF^CF8``hZEH!CcYue@66;$ZBv0PgktfE zsV5Kcpm>;gd2l2S3Oa%A5s`OF^ASyQ;%FQw)*qxcw0j@GWHLP9y*8UK2uKffeGw<# z2^E8B5_6328HK5@Yj=k0+V7sJrXB}UeIy`3@e0&;oS(Tyr_&;JU(#ifx|d)clrSDt zQy+_~izdlNXmEc8htLq>6+Ca?a-haWXz^t+H+XFEwnn+ZXN#}yfEJ&1jlvYN7RF(+ zpBIeI;<>?R2ZrF2+r`7QiQ^QpN=ApzjS|x0$=Mq5NeqK4KaT5heV~T+`E`K?PZkd? zjw&q^(1j^@)_V|+?`RKkx{0YW2rWKYJh~CCYD5MS&BX@ilN4hdcv4b?*y7IOv&Bii zK}#S!G=e-fcq8c5E>>0wiKEFz2RpFir)|7$8ZE2*Vog1#%XxZtc|>_cgCdiQDQdN$ zpX_3JvU-K#_*9%T~)6cXp!GeexYc;*My1X z$f5Ff*bdRY8p_a7TbTlFhlH-vb|^t8uoHHj5`_ZoScPu%b|{L$Nn;avm#BnV~>vi<@M6i~+lT#c)$Ka%Q@cEi#@Yvy#!L#y<@J7xKu9dL{uWu)4g>!`KlWi^T z>hLsIO|sj=Lx)QRp(Tv)*x|hf4c8**U;E^_#^G9o7FXXX^*3AGD_>d7pY{!Oerh7|!aC{nvIfekfEii)SsB1}=BxRU zCuiM(S{&6FmfJKlp51whr}TKfkmsq@x77P|o${QUdRWHh1M;Mc?C4;FL;k(zfGD7s zhIY|XKK-;sk?Ri~NllDyVN{EMo1h{_Iw{nX|Rqcsi(lzNO3^n z772win!8BS{XK=URTSt+Sdz5^w2!N{wV}W$lHT~JU1ScC)@_uA0?r(3Y1dWkBJanAXc0?u`jH05X)@iiJ9X`_pxZhc3=hsAOisaq?B!V+Cm@D+yE z-ysK~gu*kOa*mq9-pnRa=iG3eO!Kp*m|IlJ@7_u|bt;qf0z)y!a+T@bQj|41VH9g8 zugVCodGb3#sYRSmRm#En)GBk#kVhya7K1r z0t|Z6Hl_6{=s5&EiGHo`V4tGgtF6mHq}dS)X2kw9*Zb2hD&3k$=6GRNFXAYdkITwg zzT%alVaJ||lfJmIGKE!H!M|ABLexWEv(U+GYN<8vVl<2WfaZRq^lM8&l~1@HeLSSi zJzWoMcKBp)eNZQrW~dpGEa&iH=`DORbPlPX@A392m*f7l79aZ8j`-7BeAmCW zma@fup^dV6?8AFr|Vg8Tzw9MF>& zQ~~Q{AY4pqnvm!p;KQpK6+a?`#*<#}1Z1rYR0T_UtXQkDm8mfL)1HkAHu13bG#C-3 zHsnYY{uDJ!d*A4dqK4SfIz_#wfhL+md?Q+`-zpjmh~`wlkk+~g#N-pXbl9-u|#fxep!g$A!;jtmI-Ga>q4DE-V7Hn2}Xg}<S*sU#WN3Hy|YjB-Zrij)nn)LB|fDPLbD;parN|WRK(_VwP6J@Lamq_ge_tZMv<6W2NSf7RsoR6>C}%rfm$Lu(* zLVK9p&t}x@$@1dx6ZJajHbvu;s9Pcyr557-W>#%AMjkx$99G97?uG|Kr&=e~il`rw z#C3Xv0V>5V;2w8)c^5D0JhfI!^v-r_=+Fl!%-f?rI*z87KI$Ur&z)}4UxXfOpFtNe ztAkF4t}T|tJn2(b8HGs)XosbewPSiToE`gsipX(?Szmb>(4h&i7fSg3*m93V=(C}V z)u#fY(>vDu#q))3Pss#OQrv=S*Msbg%)lHN}bE z+=Pj=JQliI`Z*v>&JK>y?HC|EYRlNw&u5hY2B^dK>GfEtaxb*yuvHD+{j?GfHR28I1#EQ88*T#lGA+c3zZ(P{24QGnEm)zM@z)H_6uLAi?+lv=cKz5DMPvoX?jz4!H7t- z*>)F8AKisSgSjDQ+lQska?^#bw$F4S5YbAZ(!QRSnl4l{eH1F>)O1&E--yyLYUcpg%@@?zaQfw&dKzNV z^`-}u*vo)xOw$^Evx`JOpD+q>uu4x`qL^t z4&Y`x+?@BwvuRO}+ZMyME?MDy+?kC7PXxwiJMVMqaWzKOEvg7r^V!5}_bV@5COwti z^v3BymW<2;i}T$G7J_u!h{dPQ$Eo+GRQZl4r2Yms==j~tr}^RdmuEeFs_ti;K2D!h z2=nkXMCWHzS1g-PtCJ{`T&Z(dsG!}*LI>^6_Q-&2{LT&i3J#nwHyKgM*Kn)|Y*aOG z;Oz|<#=#QQx!Yv$_U21_S%`LN)Nb%bmr#q6!TH(>UG9js;PwWUx+tQ}9C3NE6lvJR zS)ORUmyL_E;Ls4}U3}-fSieC(!mpnfYd)>so(|nQFD5vH&=S_)wHv3S9xye!cH?xf z2ilWIZj84rN(L)t-_#Jc@v}q4r#62@?oF{8;thTqYT3ihF2A?fIw&JywA))Ca?Is6Jki{+O z_*5}CMUSHI;uZao9+R!# z_nEQO(hkqF4JIBy*6-+$IrXWeYcYE2cA};+>YRBxb&N%e&Dcjlxf%W9#e?`MM<^CTPFK&c&osB%1Nr6dgGN-}}U#?~)mb~SR=KDVdYeY5oM z^+~*nMasyaIM8tvm8@*SW`k7++iU_^RWH_rvGyBGc?`g!iua7TCBB?gc`PvNTx0QA zr$yGB98;Ebc)>%yDzj$G*R0pgUKzu=@(tF$xFyBqb6L!sr}P+#&)aAHn^ps&A@sOI zk%p(wCnx`Mk5>(HZ3lh&V8Slc+H}F4;I7C0j9N{`meFU=X2O%jLyPlr0+vwYkz11R zgq$R}Sx$VSt6j8iCG(As5spJN(2=1f?BKgPGPHyNen>}#oe0--WP`&X!j74+jc=aL zy58dE5X)_5!U)$V#gVYn1q0mn037TB`EB$7v;@_HeZ09!1Xj4yk^6K7XW8;{K40>p zX7x9ySRnU;*=2+W*fmbIDb#osI|(Xjf#K;)QH*SM@kKEO_2h_ z4h?$X-_fr+nnm2y7n?;M*3AYdzv5rj0@!>dvSy*ZyYu;SdBM#kGMn*Eh7(y;$dM$^ z4wa25FZS8$%w!=YuFuq)pU%D=C)Dx}%~`W;)<;ewYA{s3wHBsPuLnq8Xx&H!RdDj(W1Ik<~K^3w7S5 zM)Np;N57srgP^Iirs5O+>UaSda?+7&57@606E!r{na>_LOCC|JGpnV7C0lo=!#a(r z4yhA0t&nF+ZmCXF(})U8HX16ePa9ODuhTQv<~fYVX;~&x3K#!wis%hfRxoN1Bb?8j z$x0;KZ1k5TUuyo+x{m83f0BsS66>twuM4r{o=kq|d;r;wYnKiq1X-BKduXD`m7;$_Y8;`l8x2)MXM z?8mbBNqG4vOsYYHXp$X_tC}!k6`XBkP9v>IcaA=-;XIQ_NMq z)8}FqYu&$G9@9CT-aVaIvJlSgyl%DfP$pCLWuC*B-RylFfss9ser-HEqJ zXm-@^Bv!_9)WH>0vj7H$OwHDhIc;pnRP^QHFf$a8(8PgFORX0OPc(z{GVissL>1 z0(zZ2q6$E(7XYk4Q@`jhmX1^b7D0_F09(31SFL?jAS*}jmR?{~fq6?6fNfpiHdO$& zbpdyiDp309l&aX2Z`KZ_CY|U{&dK=VClQLHm&K-t_%IOh^=rfqQy?FZtOUm^VEqt# z9b+EInV>3rwb-REPF0^W9;P948{6YcTQ4ToDy8MRHm0{3ek zXk;JWZ!v5I^zeS`p%2&_-?AR0 z4J=(vpK3GT(Ng1~RY0Q}t_y;Cw>ZhS$3uRVLL0Te_<%}|uxImGENwZCealD8XQ^|Z z)FKj-OlPkIyyo`P`?HJBYOR%6tK$STGJ=hO*|9bBaib&hbyL>ri-B$oAG4O!W!^(A z*lYl({j4C{W7F!GLAdw>s^U}j@M@%3_UIzQrACCX^TEKlr;M;0!?SD_(Uk8+!8C+B z2?a((^{{@;ySJ7UOviCIO%bW68_KLWRT!%IAxv#Z4+HF=Xy&$#A506_NDya40 zfAfr9oiNG2s|8z*1r#oD)fBC^X8joYVpBr#fz|RC(8qI#Q}RkXaH53l_`2+o)sX{C ze|ll(cVN(t{f=cMS``~To-c-kiO+~dw!2t@i7YGcWK5WX{9)wgtzY1-IyQs#Gu#BB zxDkl$?2@OdA~$W{qkhnH==JB!1e!W%H1MW^y+Lzl1)-_o-4%rWF5IhD8O}tpv5zwQ zuQO{&3YysuFtY@tWpCZBBUMCuys1H_$H&VuA7#>wHH@e)VBe#9D46H$U8)EF_@jsl zsqit?qpfr^`nNc0X{DWtJO@Uqq^q@_Wm{wEZge#{l-m`)pcxRW{KxMj)McPLN$2{2 z*yKp^JCfdtlf7cwz@5Dy%0(jwoqT@criJ`G2`pX8FkV$t9Btkgkn57O^jO&?!Q$4x zk$eXQA(dH49R+)Wa|=(XNPtGL+v`dtFy&WQR0BtZ%O0p8WH`>^Vyv16-!_^VE{X@w z*-4Hq)fu?x_C@s|-98lmWN!X#G`H%fIke~9T9W_)CZnw=A{;o%dk!9?4RmwE2+I08ijuQ4| zBI-RNT1K*=)#1y^0)P=Q8k)9~x~}2!!A7?ofUWaQchyJlgjQ^;_Tp?ctoCTy=u+*c zjOL~fH)vjH+9=~WtF1HJT_`tOT5P+}!%QzODq_Zi^qMtjtFR#2nsev%@W}-4pC4U2 zJJqaoQ6}Sh791y`o_uA}0nR!-+0-OUYH}r6jI31CHD^2tfg<$MlfOb~bUl zg4nIf5M+9H*E1A&Gps;v3tA7h6T~}@k)Qe%C<`P;13|@=pdjE{CoGO3ZLP@E7p98L zOw|k_I4y9MH8cg2PDU3(!?MUWAmIZuLdkU5766Bi0|_(d!Ot>R&o*Er&G-mYZFXI6t8{_4EdSGReO z?j*?W>`|MMD%Zf`%-{eyquj$(rPE@F#Vk`{b%;**NXXg~xX*zu#wuyceqC7utKsT! z%SO)W)F~?iA;S<2B+HQSgK)J{`k5hQ*%3(gz+{yiK_KUFWOdwziezdkQz2+-s4J)- z@!~g7kUomG;{?J3P!89G=ekHw>n>O+J%yQ;hee13uRm8DS7`_GnXk=`+*M=8mFu8d zE#);Xg=7jO=4+s(5~s_;ix=9>qW!|DN}4LiSIunIwx*(&Wmj#obpqE1x7TTn&~te) z9UpjKpSTh*)Hf*C)em2zoJI9gkO%;IS8&!xgl~7&=|MSi08&eX%|W3({=HE-R6%fe zvQ1kpkll~drK(V;r@cd7bYZ1#WR$>lj-bisP>(N>>UCIN|M?+#1r^n;N|}k*eMPLc zCP8gqsGiCKAqXLh-p5qF9iq;DIf0x$pEFYz8J8(R=<`Dem$#YZ`UryJG|1e00wwN8 zXQ7Ok>d;VcJfKD6v@*qaW1KTbW8W`>${)6))=lI!oP^+nIf)J1ztjOV4D9S90(%0} z@UNUL)#cLX_JD4D*T1ZOU#mb@KHOM=%?^M&E6|l&j|fLRyGVOA^p{rKD(%p_>DZe| zXCv3iRm56-h1}7JY5W2Ky^Mt?b>#E4Rk36oF-MK_ha% zg4cI}t%@qn2^xKk)XHUHqUJ`^5Jrza5EAL$Me!o+!+7pDPy<$`fb7UPr^d34~U+`b^9bW?~ILlWSmw(8ihB>5^uKXaqraIMub0i516* zMo1t&U0k;SD&L3@5n^VxnwgOp@&Y9d1>k~11BrW2pvt2LqEk$~^-!%?;!2#ni5KV0 z_pYr~z5GrQU2C{sp`<|l0!xiQW7a4Qs^?>7i9k4rD8L9a-Fd1wjeM-b?_~VzKg0#l z66%zkSzOeaD5pa-o*c8t`katl#c}8inAC904URoyaOY^s%Kn}V9quD3d>-I@f}IqI z$m1+>ZVyKT#Zgb@sIH&6BP!jP#u}R|F-_ zvY6_*mU0(3B=m-sD}88M%+<`HCpZ{RWUx;Bb;V*i!^BG?D!H|`GT)uk-(t$l3EF?} zTAhF2TQPt|`xh^ba?I*kB+aO}k#Rlr@>Zf#lBF(o3B@9x5yW1(TRQ@hf2hqJ*N$eZ z9RbO?P3TV0;46DOM zvyHWxsvYr*ORIK7vu)Lmc;>BZ>)~0rc{{+HA<&Ykfyv-k&QkQ($t0p#E2|QU6tYYv zj$Y&%A4~JIVsm{LQq}nMbaurpkja89gIjIK%%!4eboIzXsZ`Dhb;4Y=;I(q7nv^?Y zFNKY&P{iqv0z{TlR?i$ak*bbAx}g9E<6Vsg2jg9h1|kq=jyF@h(n&KxHCj10`CKwh zWWN=sNKUyP`SXSS%E~f3d<{4x?Cm^(aQQ&i~?zvr)iOP&*&G>FJ)->loo9< zK2@J{Ha~uhHn}|fCHMG#N&Wcrd|9lAhxy^N{9Ewh!Jpj2mtXyYF6`BhvdrCo5uLQt9;23AE0358r^XWNW-oIZg=f{cP>DK3lL1tL}?O^6WU4sSC0>zEKKu*7yb>GC; zV7iZHy_Q?d&-$SopL~bOmAgqxyFDz5I-@YmC_BCPF&jn0U_R;IQ5bZAffth=tGggn z9eiM{B4ANsL}G#840ulh&YSG$T{{alB9rc0@{LAhQSqWeen zDv>+|q{j@mS!gliod)Rwi|I?HYUIsYU&X3=M5_2@ES5Gg^%hilDG)5GvG#fSeGHCA z-N)ca)bS_oZ7ZOn!34J0tT5ZB(&2zh5?94(IZ!oB%+*`0>e$5ONp5dB#~hY8PvI$M zn4)zuhh*AS#Eg{p->IDfL~8|uU-84Ls6} z9;=78H~J=+PK^zwyMN7-hu?WKZu+b&%lHv%5ZeGT6&gp1;Vd(b85Qe(4gnS(KZSWm zCDbrY{GdCw_(J4cXX-22TX5WfgUYB1mI3VPd_mlYfe`LJ4d*fJ6->q>s6O+Xuu9No zCEOEU;H_(VBtzA}M~Ew&37GD2&IpNDZE9 zU(*7kr^Su`!fO)$s4@!GbJ_Kgf2@7z=Xc&cK9m&-r2%fI&RbZOlU4DoGKyRjwOEoo zlg6sgmp+1BvHJEiFsjtrf=vp;6&o?%VAXm?ln$*U78`6*nEKGOrw^<~G8YJ~n3EHv z_W5PrM^acK<-A%QvkG(#mNmu?C!>!YKVjm*qP3rAg_Rc^1d9ze5mry`B^Br!Y*v`V z=ZZbeRl$JmNq&!3V}1v72BS+Y@(3F?Krk=oGMrN zBdvWJ)VsIJ2}qkEq~9q=r`Th-OrqrmW+A5ONVekaRNl6v0*GGDA%R_ zI~YWTW_zM=7c-xmn)LS)3wl2W_atmT`ew;i=cUB4 zCgQa6!oA>W6-fru=870EVA1*~-+@6NI4*HiH(hUtrcu=5oFpA-XqiSrcPf9lO*!&| z3|f)+Nn^5QP>ax*I&&opy3}pNSUVV74|d%4v1x$8ULavfH0lP6jNwVra`2N)gPFdY zIN^L|tOx?fSP+)Sj26Z363#{7r$ykOtH3YDMp^V}SD_42In+{50vE4ODlKCVa#15^ zVvxXFP%&AkoZOunXg%XXpn-6yG!Xmw^d(vna7tp!U2QqGr@j=I8RT!DIQ6AKklC%i zHK;y9=s71@Z?y+m$;YkZjdKh6F(eL0wN+1$oY*%IY9NX4sF0`3=kk8lT;sX?Tom3_ zWHYPGj-J^`tf!pLRWhDd*yIhTmi5@a6|Dxu` z4=PwJD{NL6(7j@kQukue20Deg8?g33^4%C5kGhY++hijjZhvJq_JiBU6I$OYJvfPo zTA8=8McF;Y^z6eko5gj8!|XKLVm!&ooYM=7_P{6y6rJI@6N9(OhW(GP%|@P8d1W@r zoPBK*;cc>sN`s(4Ici-0MP`}@7~N%_oqp&=e?XWqI-Q_(M&9AB$}nb)=PvjtLDgn? zHbkS%s+UM>nV4!}_cNe%SNVO$wa?)ad6c$3HBA1Es)E-jT6d&_7L%so>5(KYq?7hu z6B`<-n5n19hl=YZ{Yunk0ll1z_0%*v!;PI*bpd@{$G;Gf8^knif&_HaXj>&rV2Nd| zoAj`vLR+t!vgKcB=*+G(ptD>0%$IpOV2@T3RhryIDThZ@GL7Mi2G@5tC zoPDmuStb>|>tbn&`?-BYlPycOee{wwqB*9i`AcsEo7FF>`NTBhpe-6Lk0gGS0nB}j zwDZei7rZ>59>Zc6Ji$vjUh?(BWj$VUT922wG8Y`*J|-6o{jCRMGXJe~n7fGg{r!t% zG}n%$Ch^5;Z0dUPIX*tWgz#<=r05%Cx7%b!j;%Wq6LS-c!@{32OQUcxOO1KjmBxoZ zuB4OnHXoaxdY)|SZ&S|3U$_?G=wFZ$uhlFzO_IOKq1(guh3|^Ik|S$BJ1KRUI${@L zCIQGDJ2f##vM-UI&M%jo&I|vY-NJu-0inrV@}N=PtDkMjW-r=K8+8^OO1a(boFl2I zcaJpvg=-Ox{)NGG+xfdOJG_kc{&Z~1>1K7ZQll3>RG!Jc2@a%x@#Di|;lrmdSN>by zb6WWCm%pgzOLJN8)^%{b44Ux2dHv7JF~8aI6>;=66wM&lv;4U}G*F|bdJeZFDNX|a zbeX63E6cCmDRwQ)0_BY#dM8FnPD{!0`MobKxpq(z>^VjEd`YE3^1!A#o=Ol;6lI=> zCa$2-qnqjok9J=Z1OoQ?-I6nLW9zZHIcL4}Y*hMK&TvO}408qnw<F}PUgU%!9Yv3fB$VLr0vlkY~j9s_e z(X}$OA({A}2?x8%#~Oto95nS}MJCP&{60`n&jY0q#K%|j%tCk!k&K*KS`I(m2K&Ls zAvhS3myWIaLr)}AecU(LbE)mk>QCg(R*)W_2?tY;BmGmx8uv(f=$*WZAv_{{BfO;e z3waJs%kakMqNx{NRCfw;55)mQ2Q8e_LPv2lr^uW_s*@S05JUzuk4Wau&+_)vJgf~v zupfLJg15=$%FDsGKyZ7KF$A|K89^cQ4iG600!;%52fU;>`7p-|GGO)GWB&5}0($Gk z$yaLvdq9fC7s#pJ$c_MNZ+|7a<{P{0L$p zm&rTB@^NSOV0j%hT(Lo|Q9G-tUex;KO)YAJ6zZ!;l(C-CZ8W$V??5Ij)<*HdOz~nC z&yToI)t9)nLfijj!;Oq#ncGw!W#cE;P{nfAV4dFa9>~k8xfIfZYuv`Dr)`Yk3mY9z`((A7X)vUP10hG)T$IS zo{Dmc{nS~g&*>Vu0ET6qwwNoR+uGLl39 zM_#YxY!-8F$kZqrEhX}UI#fjVWZi7Tw-Iq*=#mZ{_oZh<>hvPZtJ75*f6v%mHFZ$) z#hf###`xxLtiq;L)nk_xH)&}T`7q_CH{BqJDAuMrFx)E>t7D1#EV1e~Ye;y5RpTq# z{#6JFq;(^7Q0A{%$j6QuDo`~2M37&gzC(gjcEY_@>Ms(0_?_Xb6$BBK zG`=Nw;~P*dPP^y<5^K6#4HFfkxlWpGQK)oS4|AAJ<5 z$tRe6uI33>bnxB$H4Ijlz1Dzj-5*q8y}r9&2?N7{o;U=E z7^{oAk>#=@iM9T!D@#di?U@sQ9UF|Fal9a2rlHSH3 zEs|Dy;?kOEXcU4~MyGWH~lmlZ>5HV#FIbF|1F3!WPd(16SsC%35JI!XIc&;5CLZ zye9FnjBa!u$Qigi?aWc>RBvFm3&c^QFXUI@uNMLu*+~t*<1MR0)!xi7Z9N9Dts4m;)1KhW zu+uQ;>rNUD22Kq74H)QyJhg*vgq61g4!cCIz;Bq5_T=2yG~F$rJ1lI%;{h?FrVu^M zLBWbGJZVMDN5o%rG=~{aGo)rzIoM+-%(4Q!#Q8ZvSpvlBn3dshWG9pCYQU^zZN3wR zg9(CiM%D3ZVAu-Sg7uqTdZK_vhA{X1J`7hQU7xOfjm%SIH#Yg^K3acD1LNfYH1ZmF zkdZoy*)$^L(-KKd(V)Exk$saF_Mt2uKYNhdC1{smb+s(?CZsn4l{lPW`PZ|A)GSchou5f}Qk*>p0oj2XoJGNkbvg)+xG{p1qQWbiVH7SZJG-viz3D2)^ zWEkq1i51D{hN|#tE6)|KA5N{>+9nJIDc{DaAH>;)H~w9RBPPrUkqQ^bUJsXV$xQ^W zC;1FqGjc0J-+nc43PX7P_^H6_KbTReCkBMIMOR=?9|kK<%GL4)ja0bk&evf5U-{Ys zW>gvLHSk{}qaX~Kgz@SL9pt+#Flg>9%*ly^Tev=pq)&BmrIS744??G?AT3cgFrx~tla?}y@_`f zg15;AcSK&FO?A?bf~n^=FBP^Bj7N$f8Bu;rJ;~4uJ2rOPVde==ko$i3g8TpNpb&YgQbA67jYnnc{ zVBJ@ff}Oy#a`&15<(<==8WhYoh^E>YP}Os)+Bd`Sn#3lfFbfj%P4L8UD6fm-$>0U4 zYQ~J3f`SuqsfJhIJEkovgyY{r3Ppo0vUobxJ=I8CN+HvRQXQL2+xY8R5^1W*QHJDb z2P@Iha1zW_L|5iwxr}Fds^>8A;L}N^5#;xva&(o24;_bM#;p)pO!{IVMf4&d)o~lS zu_edDys&p{h{EUm{DwZcLuJXUqr~1D)$=U-^?s4w8==$5_bnnUtQn$n;ZB_pH|NKT z;>b=wHevnXIUjD_trx=5sDOQ>TRVUgi_B61SvndR`ha?9 zD;XHf3(Lz_FeyZiRmi0ZIHIl}z zr=4Yme>O1nnOtu(T0;HIk@TPHKnwF=DILNIhdJ4N>;mFOVq3swja0Rt4J^&{29`e3 z9cDXHI@or+49H|h&3dP0XsHC$iUp3D$|6x>j4X5XOROS$lFJ_C!sEwvvqvKvC%MQ6 zGEMkmnQH|9%H)8{1nV>!9twUdM_Il?BZcK(XW(8GBTP#NuxOoVRA{)o5wo)rtpZm` zNegYxjo4&(Qqs}DN(kb^CAjw-wX_p>7C6pEs4+~sIB296p@r9$AnRe+3b+;i(PZRF zi(zA(N8~!piIuJc>x~8c&200t)ISb`V!zq0_ybsfW(E(i*Oq2^q(-0L#M3@A=Yv0c z?N8vG)inXtm0OoHx8EIcN);V*oKHa1UPZRdHdjcAAN<>O@y?_EDE zeV|bV_4!4tPsnoKST7tF2Dy3`vb~3qZ-QYmoO*g%c>Y;$<+ICV-u(+DqDalV|H5lC z_>|F)CZmJ=Lwx3+&o6LE_}t=A?$hV(rD%AAMt#cYr<_r|!Z#!Ci!X)woTxoS7d9ztuXgyrxew_iCUu?6-^IX z^<;b2EQ`Pq$V|z&cQ-jX#^AJ`eR?^~@%jC_ay~^Uo&l}C^BCj}SjP#nX?bNZ55sSf zkvRA3^hp@N8neUqsBw^mXQ*Gf(Gg#6p&qTYqhE| ztbA&)>ukFLK9W%|a9cZU4QM2H?bnkjsXh|5^Cg4iCLJ#yzx$aAzTp^uXD% z#5oG=y5lJAS3+BhZ8Nm=@}JP1*c+$$e2y>gpXpo{k5zAngGTyDLBU#s@8Qul%3bi0 zjJ&MlqD8lMI1s1fdX0vo?N#)S&@d3pMWV^46jmH#u{c`TZNluHAs)UL{!fz;?mLzt zsb{R9>J@b!s56BK*GOI{nAlv*S#y^+HpfX zwf_7u2#!X)2Ep58Bd4BTU9Z+IJ_LzU!d&M@_a#-ZBsrZI?PW6GBZy6fs6bi`+P@So zhb&zGd-eNx#3_}HW|WEQFVW4zbUS~4D!Xw%m5lUrKls<tMK&xqZ8WPi*RX7jbe zz=)2d6Vn=BU^d;weN$jk*6U@-tZFec%yw{=BXo{U=mO&gPesITX75LVrSDfS4{Pz+ z!ddBkyx3*Fx>YnCIV_vyP2uWSMYxCx)!CF&?Te&f=@+NrD1FD~Y*t=@T&*sF9gC;B z?AOLrbzmZxPhew!h9{l4f2l-Pfmv>qkxr<${|cn#qfFu+g8ks*&~?Z9L1_FDi#wo! zvVI@bUzxcR`s;KyE2bmwlJXgr!nHFqnfdZl`%oeDUzuC zUJOR!E8^V12;$!qyurUqV_lEk# z0c)C;KryDl-DOSp)+A`;HSkX~qu!eCjcL%xwl#fPG7uiowP-RW|g<9of;!@KHLd|Z*!i-Ei5qvNvnQBr** z>ap$AeBfjY2&348N^%$HbwEB&bNCtv;9PuR=ux zRgk)<6d2tyc!e4dMwIVAk_Lh26C%y9b)_UdgxaO=W9pO5o~Jozg!SS9JYRu9JC|Pr z{|GaRdP(86{}`;(E@sH0WWF$yl;PnN6Dm6>gfl^&AwJYZd3KO^0J+ct~B1ijx2qGf$|8N z9PMAT3QEf2>{vBGGcC}9PIM)*W9uBp>5411CIy&=b{Ks0v1#L_sf_6&U^ zMjdl=>Tw?&IjkxlUvs=rlpoNfJ3WOd&d(QhcEZz`&$STegbOSj>$|}7jjnL7snnhhRDm?R+Cr|oW!K-(r@{3&gFXjn=C3!=Q8+MB>mIElhV5sp~{;E$ZwKc;VKVA5caS{SwhWVG?FHg)+142a~!0rg{w4eSd` zGynB44O+&5Tj4Z-9m0tQkOq-p7z9jl0M~4AUc9W?GrT^bTF>_of}_tWo`yq~gJ&Zc z-=j#pQk4Ufr3HpB>YT45tBjwzO!=aT%#6?(6N;o0BQuCn$ZC4shyg+LSwwwCQO+)! zGmWC|muF(+RE|9t-~^V*U|+P+Cd4FXSM2LH63Kq=O_^)Y%8Yx`N0z0k^o}pW+_bD; z0BS#k282lGX-6X}B$O6sMC75U!!o-lk3Hj=piNjOUQD3p#RSsw0(>*tCiT?96h{o} z8G9ET?VnFvdxBdE?Ye7)jQX(}4}>tB9gCB&2~R8<>*y;)3pHK~3~Uzo0n5+Tzm?wsT53L_;3K=BV~S zwXXsn7?ERX<#l_8gMy=s%u}Oub4FQ77ZE7hRpYV|A`G5h&ZqbSb;3N%51-}#^WpLQ z;7_NAFTeT)M`QVDd_InKK7{PY8S0O&4rnoiKXF1xA@v8(oLVvy6?2kWaN{FS-dT!K zt~IhK9Q(E;TMh{tApN4MG<=Ag+}Iq%)r?JuNcV`Hl&Ik@o=qg;=!SS5z%nk4ijj$0Zr^!ZN!LDS;NG{*=8bj6*E(>pGZ+= zzd3g&6TBMx&q^U_}n*`0ZiDRa%m^Gr~ z;6^CdjMC}yz67f8?D6RoWr1C*eH3gihk67&>b-9iecaP7E=V(`FkUG zg#*;mKa$uZ?g|6$@}ksUVd6ZPNX_&C0=sEHS`Kyu0J!6meG8()Fr#UI14P3A^+L8n zKQiGziSWOzZ@htOAKfxOd#eCw;=UTOuTX%1u7~zD8B;DPwcVT=YfIaD7}&XI)j0!2 zCG0yv;KiXe(jv8@I?^H?L24B_sM?C?iwyteFEan+Ja!Aikw|}$ zj85;}OPt?d!sXrhQJkT|Gu(kPAt**Ru>$9*A0@M~s?$>Y9}nd@v_F<0(fwqwRLMtq4;YTgtT@<)486BEZ8Rd?2ALUa<|En|ltMbtxqoWg|pGIb`rMsr+bZqtyC__=B zDRy&aLuk2?sB6TbHHbNbUa)cyN2VO{THr=S`Zo6CV3rJZ-c0k88!SG91xmGe$>@B( zxcL6%<uz`8UvbdT=&Lzsi+0VyI1uXMAWjBcEU;;{ zPBI$iMib$)>K(>sg+ohd@VtXZjYCTyTxoVgT-lZd2jwEKoub_XEAEx&GaH>kl`ZFc z=B)@kEc>`TM=axNaFu^-%CmTN^yvazlb>($3r&8gc}b1h+1aFbP5M^dfzZT{Hoj}B z>znB8jFQ_yR^5t?TcXLn*UrGkk2b!OiEGpIwK<+H>fdt50l_v($}cU8MT6ROf4gyga)?)GO0OPao^x3eJ^mqrl$|s2p$?d zHaG#$5;kym4TqL+cFPJd=@p>kajTW%=wou}ZTuyP!(~V|$%p{2yhqGeGiRs9&FPse zA()f6FgYi7p|S)CW=+D#V!xHVi}+^`8a!YCL$f83$ zH>8-`hV+#BvZJvf5;Ua1uaHPSZ6EcNJi8_%-}+)*Kql6dw)@N5ww6S5P%&|8DEMea zOR*W@yC>Q^JU0}&)2}`GE004@(v(O#lGuWCLqXje3~6d2g^t3Ha)_8hlSWAL$b(R( zh7vjo>G~=oEz^?Dd5f^9ipdZD`zy>B<4w3lnKo_wlZjvYQhYg7DWlke0*&( z`FQSY&PVBVs@oG)S=C+V9Qj}`x0@n~t+h>28Yw7}5p2p^(^JS^R)bBaUk!ioO6&A* zpQ4Y9m`Fz{G7!^u!ZA|wkp#t4aEO5UX4074Dbr28l;Bdr*vLlm`O zdFV;^QEtd6>{8%Jq)PEaO4CQ0lx6&Hpa8^N#Z#iMQq%#B!Is}OO6)U=dyzINxl`CQ={Uvr8TFf~w^BAU8gE=~ zQ{cX2KaPz0?Hf%|qt@_aKhcfD=lpmHaglCl0Pta*Bb*C@?QHlvzZruoacS z9^Y9^M`9}x&`P0@M&plXh$tS$>!X=g<;0vF+en(T`*cbvXfuyZGM!Q9TI>Bi%E+99 z$*I0BVRFh^XHz4iD8i|fcLd{xpJjw8mZEr~v;$M$Ur{@1xP`BMad{%P{&G zYyB=L8nss9W%essGdb4`5(Z?Qb7Nbg3Q4S~J$1w^mNm?z5egDuFpXRjM@ZetAA-*( z=uUEU+Ei>a4YqO)4O$b6tCw^(lUmV8U@L=kO$4@kDSptC$*zQ6Q`l`e@i`wMQI6Ju zX`Eq(Qr!*wBuHkGQ(1-~Wgbu@qhxk3#+2DmyrzIAWg98qLV>+Vy_;*B0{11`&FGj@ zO;JLn=sq%}TPb?fq9$pIA5yZRIK@X%_NxKDQ$9Q&zVuru|3`4e&4CRWyU55X# z*fG!}^@j3){oO7)K2lt}9Ocnlyz3nzDIZap2$NU-j0& zUf}o_I!n_zZO{$^uZPKkIH-})eVSj#?z%?XfxBVNWQ4sy^9%h*=`h#~jD}OLhQVGS z|60}>gbjz^_bhld4E6%opICwWVO@=4e%@^3UKm6r_}d?USZWkD+b~_DRT^xyaX&1f zX*>uFxHykGr+$F72@{v!L`!-1+YiIBf!Z%$1+EPP@-&Y-U_XqFQP&NBkif2U z0xw2=SgdOHxC3I|3#)a?BY7mI&0ybX-V8$@JPbR49Q5NbS&ce*q?fAR^*3#XsVlKC zaE8nFVIM}d8ptEFf!R8X=glyC|2Es6*TC^zKv39I^3?*gzH5a!zi;w!W^b zjUd-2K( zocFZysNF444A5;ra2Rv8nW)IiEmL0)gL@OUVeRW_3+<9G=lDE_`2OAH@!>^1Du+MY zWm(rqtNHD(cSJW;G#Sw^Z6%UXXiUqj^dnRKOc2{Qs zt66p>1R4=DjS^i6L7CkIshZyuLf-6AqMJhCogkInXs|`giAEuf zVCxl%p&S|@U^;KuL%k@o<431Y27^8pDaz3@oU}vs+U8?5ww7f$+DH%D1_Q} z371oRKl>QK`cl1$6*^0S9pb4AARVH$%DAh1vl2*+g0=^g3R(72)KL&@S;Xx!yQxYg zt{WOg(#qJXi)_Z981pQySJO>31OB>t&TZ?4jtdo022j;#II(1j=ypW-KoKt$`kbF^ z9~k3>teLJH-WaHkdK(lasPWRsNubv65m1WcjUcN!NcF}c^g_VU$9yHkjDzxMv_U2$ z?}p4nU!5Pc^?FGA*-s7{`{?{;2y6%KhqP{V)wFF7`>4ZrVsy)47$qbP@-+_;t`v~GLS z?k5^a!s3V8k~V%*LU~;92K#g+XebM}LyCWog>KW?ZY-&5$goY-vlq2zjDDF;`*dwm zA+~y6zv$HkIjV%Mp56jbbULx6DF?|`p%mCjT zvJ*$+ra7ba1l#?hDTTHpv--_;ifEmzVH?roI;Ji`t+{5CjpTvWGMC@wbh*Sh2Z)d6 zI~aYTZCGn#*Pp16I|(JyIngF;W%R&8A^zyx2PW5T2g1Oj z)!B~Bys?wR=Gqye-Ujw6?$|$qQe^1IveoIrLKkLl3T?Ph451e}g)sE_IW2|`hUv>I zL!0(MSgslMfifN&VI!Ly_5!bj86&1W?6G6v?1#a%ghQXmAPnwJ*oKY0M^4STjj-H= z_1km@6SiTa;lC?WxHn-N)}8gaH&b|v+O)%eZ*6vm|7M=L?(;hkRu#4z=hRjhTuFFX zGPJ{gHK7}asY64zOUCiek-I}qJ4M(K{(Gy&*kRw_fx9$)bx@nn_jREy?pB=QRyVDJ|5$WwUa9BL0?AxGdb-g%hT1XaQ=!J{ zsFFuDrA8Y6k1? z%y1;oy(M(Fnn|2{j(?F3W$|5|WJtvQ`@O0&Ln&>-+`Rle;dVely1QB?SK0vj0PbvR zdDWDDzzi~}EVyM`j-5Mm5c}XVB=(m z7k-c~fP9|P@xuYfmebrLWZ)H^zx1ZwSRxBD$5XWSt78(pu0q8F?^yN;-!;!T_w6fN zAh=``#_kNBnHciGmmsUFb9u! zO9f=h1e_t`W@p}sh=i$QsTN;4lh^zE`za6W)nG_iJZ5DhS(m|=`zmKaPM&};rnDfu zM+By?Y6res)<^E zQHzRVkFtQOVX8!Rw@qgb^S`058pQ%!#` zxs^5&e=gPeh#xsh;Q#Om3P)+J*w64P&rM+1d4*%peA(AECE$PTI`Fjh1_Fq`*Olr0vTeElI}73N~UvtJZXnP1`>%Wl8WYmUq|*KR6cFZ6LQ) z%sz*Q5#U7m;lkv<#)i&1o)K`6mQqOuceDr3vvzq6SyqKD+weW_{5AdHO_=RfAZEyj z+}&&@B%XiOx7r%-!-l=AQ2atV!X42584i*}8nKc8jhJy~L35Pn>--lcvWjXyL?bRz zM!YNXbr*5y$t_UV?k6S}r^2~T->m?FCR>C3#o@iHF7u1S*dt4Yvc&0n<)F0=>Dl=f z@p7a`Plenyn3(=TQ;)1%^EB|%sb*p8Mq#M&?d7R}&U8>WavFj6W!FKwhASJ{1mhSM z(@av5=d$Fo^+JA?*tr0;jV>MDKd_%g>oAo*LygyZP>A`C=2RU-33lfkYM#RJg1dXVGDJ#`xc^vm*!$Lm8uvLbv5+GFE6`^bW3e z&jue*E+tj6)Ag{Ij#6?jM21j43j%EjBNr}y{m5;ahk(+=b#Gh- z2)j`XAo^zBQ?*MpyIzubSkB=&t}v(@=TtV&M7rsm#^~9b9wVMAqCRcVp9R)PIngC0lxfay^@O(Zu_u}fRQ4xpcx zCf$-N21?dTu17-z7&Y6DK&1(M-LEId1OF{=CcZSy_LD)}d%cuW*y3nrEv4t>PNe&B zthlI8-=THk9D5Lf>3v@Dqw{5RT`JQCd(%>4TZy88jOF?9>ZLP$i*JIL=A~49oRD?7 z0v|7tStaSVw92^pk+q3RPn_Sj`Y|nO;LCUAMlm<*@yklS)fM5GlIt-;4O5pEG|F=h zV>Y!H%qgCG>^^MN5`h}h_R?h0ExBOp*ERtcPp&_gnLmf!w^D49!Q~XlEivcH0||W2 zH1m#mvX~j9(XAx((~x+&Ua`m9(sslQoJ!6V$@aMQ9UER3TZzm32p`^ScDyl_0AXMD z+ICYGv(?z|4Bp#{f*2WLi+V9e>#^wKI->30t60_}N5v{TS5@7}I^{NVVEWHSJEwvw zr1o)+;))~dvr4Z*7P#x`)1Evz_&gB^4kPPxKrS9&PoL;kWc#lV3v2aN+PSNmwdiI1 zJsEr#k=3zx`CIDt)b5TFr$@O}+J&$_-<;QdV}mBK%Ds~Q7R3>Sm?6)~&ZP2$;c8Qe zO7&`Ei^qs`afm#*dwcyVLzB8aA+e9V2dM%+Nu=wDbbcKZkHR!QiAVccvx{l7}u{cX;~O8$y(5mS7lLX8KPS0-y??%nhEp5m>y zSai>xm?e|I6)(?y)^nWa_r)an2F&gG+KG2PVph3+PUQp&s#B;jYL4rUSD-D_1X2{p zPGjyWrbL-d>A9c(jX8O*Psxnl4P^hffvBE!nwIkDgA@XCtJFwMH(v4S>{Ny!NfD+d ziEqycon>_%PdnVqJ?TQnw#K?Y_l(n=JpJ^0p#4Yl>UlAsvl9TWWB!2)GmWF+gU|0qik$wna>?qHM2*@j1s(e z&ptPWI7n|-+ZrU6i;1oC+32wzo)HHptgDM@m!no>{!zda3X=U13D6-4G^OsrOifOCS4k_2gce4ZTdC>op45GIXD9m*f{F|? zdjDLtTa_i$8$>1iX9Sa%RJ406mj*6Vp#NV;?+v(0(c3PU2er$Uq;!@q=}U?DFRT|F z)~@49IPcn-_oLTL_XNv2teyU{A1&Bru3dHwMd#DFLsL81{DiVG-1-2e7+%U^zm9Mz zg?Gq7Gh=kz!C~0?F!e-2{3^hO#1-yq0^5784Nyk1iXQ7aigT&T2Jh+7N<6g#a%Edx z=v}aX*9r}@gJM!elVxA12A?Uj0teN1=V0Bw{mqMv4H@7&PMbE;fN z_fdUIzQk;~dY7V^dvhde7`hJE9$}N*!BUkNO7;d8X~z;S^rOrABnl*X{pP5Meko15 z`SIi4FB4gv*(GRf`E$@g)lSDl22yG}p?)m#EzT(In?xyTb?6cV!5 zPkU8jdYEBoOfr;V?{1gT95S!fnDUIYFdm@nyfQL6tS%{Z=N32gffEydL{uWqOA8qF zsuR0Y!s>9EVp5oDLS64aBWt{wQZLF#R`oMKudGOPIjAGEbx2Ial(l0f*LT-q%pg5Ey1^8=eBVu?HVz z!V694mbpy@%F3aS^uMP#4!I}Kbsv^YD|R<>pkB)a#%{Zn6y~i4C(@x+XX!mDR)c`5 z!K|6rcJ~|HZdtfY_QmqY&+a*?4t_DFKf_2D1s%9o^+0Zx)8tsK=_!H)GI2=bbJDh6 zdi<%y$+L61=bF8Dk)Fk0$AFW!PKH_4r2(*|eG4qruw4Z@fY z*ykrXKP%6z?eBeS0XeO(C^w(=`OXC2{{_^~V_1v()5#3n?iz*mFQeR_MqNkx%)lr;SJa>;W?NI^l-?7c{zp~!^%TJ zLNFuB2=f{6QW*KK0{BuwgEnZnin?fWlffP^ulx+PAADtRTJGV(I6Eia%6yh4bPz+w z5Oc%7h_UtCNfNUb80!c;w(O8tnr_7wPm7ze5~!^s*8W|F(jqn59!0D5v#Uje+H9Xv zs{$xq@unGBii>wiUfd=43xV-G=_Xuk@;fWtNAwOFw>qeHhZg4yk1LY0D)nag^ zecstGhULWr@t%yu?Ukz*k|3`e3~#k2jY3-@9mPuGel-JIH8$eM2`YjbUHF@k_YUO( zV0MgU+(suil@!LsFLpCkS_?;&)oW4JdVBlEVXHhFEd;)prdXTpKF({Z6oNcBRp+V3 zz76PyZ^=SP1{v0)b?7q?l~p9io>GGZg-WL_#CG|}4#o^XTd^?-nY1fdaph9wB+MbJ z)yN-qq1>T+*F%n$uRhWcmND<~QfdgG@A@4k<)dq`s+5oRpo;XiIG;;FOU7iOP1bo4>WV_xsORe$@NVxq)7N*%ZZ^?8Uh6a(=6oO{x{N5D!I>(=9v%|SVIPls<`R# z3;I9XtayQNafcxHDX6qpN_AVY0<@{SZl}5=UT-t}p7rNKp!F+~)uT3zQqh&7=-q<85=R4O9kKM-90gSVOeKSyIJ8 z%26oKZv3SJ8+m9oyirNSTMfHUwgfsY=yeU&{SJ|hT0_%~lPNZPxApXWmep1K`j_!K zLjaF2!*fNnz^cDs-Kt{Jh(SM`n9v>W7;bbP>iNG9^fQ12 zxij@L&Pe*)Y{u|XQ?mh$xMfe%OKE|3gM79|QXL=iJO^YYt3EBf5THLt$JoYlusF`G zgA>n6KI67H9E%Z?#(r&BI0&S`O=|(zfY@8t=0qP9{fhyD48kJV2`w0FnZ0y|bS&2> ziT+HfDQht!^9Khz)p>kbcgcz3^wS0u%q{yHa<4ZF{KM%lK7kK-#had&{!*C(NprAu z-eF1)Lj@^V(uWSOb%*U*m5h3LCHRQ1iR*g4x;RBj7#xtVCsnjC%6he%v=)_XEdz4= zZ9f)J6-KBJE6B)8xeDF{aM1GSEK!*If=~~x%qLV2Vxs|?=&Q_1I=1=Iu8Z~HQ0^IK zV(UG@=+${yFKZiiD$Rv!o*V}1su*u-bRewXI~busuGZ#?psnhBpJeffZFS8HBjEA^ z9Ra49K$t0@ryK%Rr_3o()qx%q>{yG(0>yB>+HG}<%c3Wv3*Nj9g#jq`pVzE%aD4yA43e;p z#3D6yw0;ipqnX01@0or^G*__?P)mpQBDJW?w41TAOIsVSd!qVT>;ZjG@VT$x>>^B> zh=;9xkLJgSV;TE_O-K*+TSY@_VS?Iz8*KaNk6xU z-c|(!TI#9PdRp^M;FQH0^rXuz>$@ox2ibMag>ZCpS|{){*kl2x2Bb`Bqzg;iYY zpD^xOZ;p)L;&h$mV6G~|Q5L2Qe2y-Og~3FwX35i=eL$mRi^8U4n9Hy~7rJ4Z_QDQ} zEd==N(WYY)bUNGbaFRPqemx(}W2sGu1u3zR$|cUpVJe{bp?;%)_fpF(ligWjP>qbS z$A)#6DTJ2b7KDszs}bS9W-xsaMxX@&C6LG9W{GO|N{gQ9 z%Q!CWN%vGUS`3=?+%!**lv0JQwtp*i>6b?_X|gIDKma8|-9)!l_cjs*bsQJEKx{FoXXHtwRbtr+T4 zuV64&yXb-Nz5~}A6~QSD-LY<~fc++khF|-Z9Bm70BHbItfkJ1`0mC5UzV{*Pp>5NV zkiK=CIKx1Qw-!thRNJ!2?ZgnNq1MU*8)3wskHz|n%*46#{>GI@rUYO^I!TwT5gM+9 ztsSS&vtZkOt3`K44N>f~Qh=Zp@>&O@b3l85m6v*CmtDdtSEzVEcqAX)47QWb9 zZ*HSlm=%6TmzFeZB@N%P9=5~k1(Mh!+qd~-@th3Rw~&RT!HojACqbVm9TPxi9RB+X ziI!Do9h`Zl+T6Sm<%^}I__W=Uh2GUyPBZj8>(-F_<;iG-HJsnYAtd=nWDD#0jpAy+ zaIBuMg2##jlgd8ROG3rNl6Mlii>-hnXhNECw3-&6f(LM^n@o((Y0lhP#tmIoO}B~H z5N#TM4VUtpL8@G{Q_N=3?cw?sSy8eC9XHeULtEF3Jo5}k;HCRzE76vD=VIv;4QMz; z91P9o{fUEueU5Zl~1}n;D*7eejsmy=KpwQVU z<=>5}t&|%-uE_s}^C~}Xfj+vBKBQbOirX3pq?Y;mg(i&CtkX-3SXm^HIR7z!ykU$K z&*Y-dA(kU0sy?}(`ONH-TEhP)?F(40^2bdy$`4aI=4Q@6v@(4 zncCVPm^*%ZZvFLJh*vR*&(84@Fs~<#7XI`r_`kamMhGyo$VC0{U$F?`%x{l5qv{qL z+lt&`srdpWiyzPapwXge2_OPtUnX zB^7<+k&(cg(p$d1U8o`MHI5dANhdkR@R z{^ddKE+=bQrspjT-C=_-h_|JW;I+p^(VeGA5$SGK=^ zI}<5;JSDgbYx8*cK1og8I;+P*~l=${$2dt4#$qTm#(S2Ek6KoT?6!sy9eGlt z{Hk1aI!Q=;n{t}X9svPd%*d53wt1vmDk#?-R-z!CSqmszXP(ECsmM>rlxvPvoS$E# zBfk&J0?PgTz+y9V(zT7V7VVvY0LxW{wB-eOMLGlaxQTnE9MlsMX=iUq=8PqKqY$X~ z_(l3#kKT`5k%B!bL-Rrjcn&K{!y0X->4<$N)LTuVcn>PHd+ygBIkHW(N!d1HOEgsI z-q5?hmIVuXYa@>qe{c~qGx1Vj1aaf45{!uoYB!4{#M&j>5-vGMUs&7$*3n=p9%(ex z2@do`X*}I!|H)R*;;net%Km4I`0m{5Q8$L^Ms6tDL>WR z>*Vy)iw8IJdl$3;EFR4F?8{hc+Tuqb^VK7#oE!e-PYU^1c0Zr0r_T@ji>j5KFBQe# zC)(5j<`Dr2%Q_N}zzFPXa$=bHaKbWqM32D#Y);rhdZl}MTJ+TLrHB2j*%Q4BT&e6G@;x1Pq#hKulda|bk@Ah zD34EBzlIL<+kyV>6y>w*5mmm6Kj_tCn#Q%o&zm`Xnx13TTn0bii%*nDya>gS6I$fw zseIsbck{_i`d&F!|0`z}B@Od_BsAt@9S4$B>4~UA#>1+8<20edDq%Y2vzRN+Dv;#_DHhE4^0JG~@b_c`UtW*=y-aJG2!C&DxNkL_4oV za@y2a&5>_v`lXtHJ<#8h(a@Ei?)d4&tq}ovr>@;(QzyvK9d;q2hxxK}3$-Q8>E0Z< z;K|+8CHd%R)<0y-%zk0)5?u9Are5^(d;CAyygx5jAzK`I9tuODC~0dv0yid8I#FAZ zU6x6+TS?zH%+v0!2I?K1odya1bLsy_rnh6Pp1H2~f1F?E&DS)O)sWdKoOzSK*NpA^ zWG1*hQ?YL-Uo0E3QU|4&{?PAUFcE7jxT@b3B)JMw>#npfn8mWO{gyaN;5m#uE=$!H z23s1jTxv3H=0>;-N_t#XUQbqO==iXCRu*EA8&m`Wz1{Plc-=(21*b5;HnGDd2OJu5 z?;6U-lP{tc*-M zP9|bss!1Hiszgw{J`U^g*IF`?&}~VC>EY`Ax%sWPp3iG(wEUtyO&8+o z+snGX!MM!B>)K4~I|mCI&u5@%osw2I<;*oz^ULgcSya}szW<+fY^ljR57^(bQ;{0# zl(cOF-km9toFd~^`jGlG+>J!x6Q%m1oDgSg3#`e6&y*cGNm5l~eA?eok98Gy;qa8S z<+!I_X~TQw_^k_SHG-=e0pvB#7vzAG!hGKMR<`r=_UR0|DdH;CT7KlExPfhzRA-Hh z^%d`Lyf*(Gx9h49881V;Fe>s_7^?jeTJkEzJ&rxIt_`Z|m8l6&(e;Va%w8QRUjh+7 z>a|Xeo^33B_G$w`#9aGG>NvAONo^qL=ZyN6GO^k6c|T(FM?a}BQfMTxInAtf>!goS z4d5ICp5x(`^->SZ!>~>ka*znyoM*Zz*L;M`eMY|&DMpY3G!EAWP(( z2)K`QT#g(|I4t`e9v+`x6I%re4jn|zURXxU;F^vv4C?4ZT*Z;B21$^9*>!465?(rC15+!zDc=)8QHFH- zBOnl-(i_^@Qx3ttn|@)|vR5o63Iy#-s(bZX`07PC6~jes4S32pttP6wGA;q!A#6-{ zyfRH{ogomBl!XU@_Cwk!M0iK@?`*vIvJrOf?L`M%`ZpdoP9hyxp_XHWh=qHQbV24) z_)V4Vk)E@2u-k?BSbFa~-ZI0F1Kd2>%j?MlDJl8J@L|${If~D@KTeL2M{iIACy3qik#+g3fP=Gc4SB^Zw#}QN%*Afp%Y!u6 zn&>^&Qy0Y5*#vedaaHowN4<9+aBWxS*_Rt@?(;PQ??it_BS=n%MYnp=h}E72A1iAS zx6;|wn30zKpLHSS07|Hrj?edt&GJ1h#Dk^yp6~ zXM`?rL@!K}GiCEeHI6a=^`__QgY|j_v|e7@U=y@ea~*jwiA9UWU-n`7unTCdNE&nM zp>XG;hbJ}BBpEc+Lt@$UD^KVx5Ayl+N(+2p{Yz}JSL->e)WuwUty zN|Ng-!s5SZ!|sDVKMMLUsn51l3^c!}`mtTiw}9Y@in1Y3#@A`xduV_kf@*a+p7 zPwRQ{8zHl%%QrYh?Qe2%uI!wdFU@v#Yj8}bqnEmm6;L5={rGc3p)<@Pvh4dzACQT+ zx+2kVgRj$Yv&nKWBedol;{dL$C0yZM`uo8YwjaL2Vr`m|_H8O;CAj}2Mq-3x*9Dxr zxht!`*`4Lo{3tDvmn$GoGYTMw51})y!O7=Jano7({MOpefczj1bJ*t2;TvYlNv}Hl zWT@`Wd!aBoerZtB$-BXp{2DlxJaANly#f~^vKFIKTIwV+Uk*xJTg9orp;4K|JvyHL zzd1Y!OUi!@o`hKgT)sJv)AQd4V3A8@qCN|0|8!3aQq{MO0eUat5jk&~D>1k=1Jxg; z(0%6O#v|gE=p`KeE?#B#uckthhbi(_f3j<3mMVR%Q5*34RF*1GGvq$enm!WBwdJ2f zN)b?2ipFh1oEjHOii)ek7Q1XypAD7nrd;>}8Z6Qa?c9nTSls^dqp6OA=AfLre@CZO%bc$UgNm%ASnv@E40T zkMnbvELVec9gW0$^Dae61wGQs770o%Oq6uX49eIxdaFM_*bfc>+=Z0xXPvt!P((dC zWh-pD?L&>^w7RJELNb96EmoDCbpY zve4-yv9=&P-L<>2&I1t0M<1{+^BXZsm?(q!dOlIVT!EeI z+l3PTauq-y6&+*L>Qz;ZiEQnK-C9fGzlN?O$5%+UN0$-gIbhAAP?$G_7z)AuD30vo za(gs0lI0|_XscXFr4Jgiy2$88!Lw04_$b~ApGnh(bw`k{DiR4(V5OFgd27Y|h&_`1 zIdd&$yUaeCY}l4+fZK{E}H4FhFkr_Z$vnV zz}JPy(D8u?)OzJ-NjT;!_I~2Cz1RpM<>-=2GywjL7E48}#i`g;FMF*yV$ios9|7lK zm#b+!HtwgWKl0K-GNX=koe?0bXbXvX!kX3Hym3)#h_>3dKS8sE zKK7iNG#E#fb_PCLE$6(~jVbYv51L=VK6K+9{jU1nrorua5Svr z-Em<5C9W*h2!)h$yc zamjG?Dd9|EpHpF-(i)4Tb=KJVb{QsssJ1at4y}2}uTAnTY^00Ca%>zK7-XtQv)Mkv z!!;4fP^PMD%dVXktHKYg<5otNbJh+=~%gP*y7Uliqw4NhB*{Ov`w|8)I#@XEr^tp3^vTDFPpHc zCv4SIVwYZIDnuuP*t?)ESg}@#;k`zs7$YiJPFi)`mNRpAd1s!f&MbKi{sfN^RcMTPzpgBoK7J zvFbaV$lc}|+;oTA{S;Wh6NZ-%*SFmg&gUumfIacKw47Gxa8G}sJtGu7q+qzsQ&HZN z(R@lXE1PE#L!~-oGB7MtJ?JD%Eu*&FWV6Uvx-=BjX9K!aS-mabrv_+h^!b5y7STs0 zq$_EwbkEzdX%iW_A=`@H!Gk;&vY%#wg3R(@)n&}z(J0ZM=$L!_ry!qCYbK1xa0NX|5?!Aah{kuIvq;fzjMqw zK7QFydg#iwOVR%#*b^7FQmNSos4RMN5AbG{R^byl!>r1zmGWO1pBxCK4_Kuv@l+T$ z&3&9zJFFLkCa{0uld^HxoqW7 z5k{5~XLs$X-Un8SU}==Df*PhBGXLC5vzTVjmM~pH;K8J5_*aQ@Mv#3_Sx&AJ0z?%4 zBmTI?9p;H!spL8qzKp@#xs)i8LN@xq){QD{En>)0zj+KhSIWn3B7Xaw3@6l5lhuHI zh2&K;=IZ+{&o-I36={faJmp$ZMRQZ)x)O$ERMPR}(MJxuW}-LUV{zP+vx!00cWr?V z6yTR&&JLos*u}5&C&(9^?WNS|RS;3-?xa7|(OIt23rB0?0r17kZP>*5#x@a{w^Sgt$`Pw1rqz}d*Dl8di`T`y zMZPsU4Ro0vh1_yvp_iKO((h(rQ3wyw67r3xGcKgv8E2Y;XT%u`eoTYdn>wVrgRH_% z`(@JoF<8Wc%$wS7C{cAt4hRrSo=4arDM3) zZ_xf)nt@h(+BwB@5F&JM{9o|sy7HOO{v(sXnAK%=^4+HK9Aad&+PZ_UJ+Lt0{blB1qpCj;Rf*xC>&O-tKdUKsN8n*E49P2mJ?J|vLrPbbQ)&7ViMaBy;ORg$D5wk91 z2fJ!Qe%DdQGStn8l!V(0hbfe1n)aciTd{vklx!YY=2HD1Ud&Ietg38x3^N}yd6(8G zJIoH;1QPY|Zt(4?4(U<_#AD7{cgr)%nNE8csnU6ePU^n6pO;1E|1O_e*&+=7W#n!lj??ETe&T@qDcx)Wan~ zW25XpSNTI>6ykfZ(o*3i+5H!iy*Zh`Bo~l@Hsb4pNMKLL$Pl53fci**N_)BtE<5o* z4}SvD)GNx?hiIBh)INdXUH>MX^y^I994z;;FG< zcqclHvr!r&dZ63z|MOr=IL)oyk(!=fFjZ?td){M;_4~%q%|C1scOB+q`vrgu9X2#YgEO?-m%uxkP*z#nY1ZKnS;Dd!`Iy0c**Amrq~^LT_aL>UAUt}Gc0mj?Qs+rSxRoI@B-3mRc`|-KpX^~ls;zAy7bS~iRpxT zlccT-fWqlte)>kv){M~gBvcDl96hf~wPpg3FNGXX>4Mj9Vr8Y`aEeTam3!9lG`|%= zDqbAcf6F=*W0?{jD}QeNVx;WoW%Y5B7LtpNgKu@Kg9LL`tl7H=;I-Y2rfF_2;KXSD z98QlXn$2tDo51B^nNs7LNXasmRr;AW?z50{5-7X(QY$`wL}d1afN2H z)gs7FrOW_A>EY*s1#|$GX;p_Wz$Q-fo4kA&c2h-$ZG{lAz08-SK1D=p zj;yFtgyHwC!9OM|XUSm^4EeSNIz!s-3s&Ua=gn(u-A(K|lw zm!1pM9~FZ|jEqG(&u#PDA%`4|bD|z!hA7{mEO&EwT;*c8T4FBWWfTXT+vaf7hfxlh z`V`8Y0}w(3zvI4liwu&|eaNUcqu4I3j0Uw-rMFG^pnRg=-zSDD%3A(kyLQOLMsADT zIPkmj)M)(&opB8W36!1p_M&lP2xG+&ncq5{oOwrWwreMwD2oKaM(LO*uLn5kd5i<0 z67nj4RXbfq)f5w|BFgg~`3h?qRI~Z)`r=ni>bK@3bq=Y=N>qG+GM*w`MyYbF?_|IF z&U&;5(SBh@VftRuaP5X&GI`{q0G+9KRu7Am-Iw7QF)P2HW%i@whdim5|AOKKhEuh3 z_UAyq{xH-p+iYjXyMH*Mn|G)F_)GqNfDQs~a8UCE`T=Y$|Du14ZXKhRB)ZPA7@t^c z+5#eo>;(!*S=RTm2C}y-lGgWF%*ofHP3R^C-ZcNVxkTPzQF&0liB=%+^8xpz7C zZ%29AVQIyuOzpp-75!KJjNY%S1@9CgEKXJ_!G-DpiiAEQ`4aWk!7+Z61>4>&9oy$+ zf6g%F{v3%Ihtpb0U9(}Gy;yysfj)luWpggA{#=q;1KxK&Ae}-_x@qRbdgUHvE(Cmz z=r>F?m+HIa`z`Q&Tc4|^$_prlc0t$DRW*xK?QDhE`YnBH>}l7y4sK)*Yq5|(r><`} zo!hZY<^81Ap4PlQpeVzYgIL9dm+)idFJlG`z-+NiJDWgD0z!K9h|YLpt$gAyp@s%t zI{Ye4NCh6MhMJ(-qK&SS9Yb1i?NnFU7sd;!=0qKKQoFhhbd;c!k7%$g5OL<;h{2WlP@k7s+Z4WA?oRx=CTFQa>QeBOx zE!SB?;=SI67K5`<^JV-Kaa%f+{B_Y1uzumh&~B%wlL~&&?oADr1Ybm6gI<8svs@q0UUW>d@h;wW*bvgo4rMki z3Of^o&N|mw-h-kSu^DhTZYBQutQQ84!ZmDZ2TZk~z=Kt`v>;7L;*5>f6U$gM4Q?dUU4;M&lxZa9RmmnTe_`$-*AF75dvLBy{kV#=zcQTp=Py}xNh zJZ?qgk2;Cv@P}ngFx2k;bag3=zj=(VBuuKSmM){9qrJA=%IiFxI40ADzMlPYtyt=M zcrXfNAXY4@Oce{xxeiWu=__hYg+AnS&ms<&>TnS>Ninie-|6~Bsk;=nvVI86Aj$=| z$LCWLO)jdGu78q;lPK}!8e^w*P(8`1+6a;d^5Q`w^UH#1i!-P3Y$v3^suH`rq-qk) z>0QOiV_vol#&Q}THG5<5Dj3xy7?;{9Mypwx(%U%K%d~6qbx{aatIFOAmXlLSKvg^^ zz>>91mGaR=qndzI*hD)Py(nAdY?4V;2PYnwoOIcNF{3=KLz&j3+l_bRc|e;CgV;d+ zU}+6rVMW2b<$@1Tn@GaO|1I0_gPvRw1+ zl4-mVxK;p+Z?|_6ke2e51p5_P7(?*4+YHBW_r4T`gwlfGx?l%iF`sUI8i>m`i~uo< zZkh}zyKz>5cei&wcn%H}?YM9p?XL5wdXVQvaDHltC!g+Dvo{uYUx{GS^dD!`EklSM zLA>5cr_;9#p7Uvz+is`Nv!-PK=$;&4KFf)u_`7i7%N3e;U`^0vYXRs z_v-$-Z1qvU?F zYog>XPAABlh-Skj1VIj2%E~hTED8@0koc)59rBb#oBzkp0fV$xCK&d%QT|O^u3}I6 zye9jw=I|>hW1fH7aE8ZkH%!8)+r)Emo&KwCdRF4oN+ZERL8a6XSIWGZ|1_W&c*X_@ z*EsdgA)8#XF+dE!?zwkLDL-?1eVSSykJnvZzrQ;tZwZPK&JPQEaA#6nQA#b6zX+Sr$2)zBs}GY*9Zj8{rH8U5lisn%2;waUj*;r+eJ6;?j~)}?&3Ss9irP(9ZHoVgQB)G$-wn+U8T zE#Hs~gpx>&fziAFBReA!%-zv+8mU%t=Z$|L_Uv4Xo_Dzr;RhEfM>4Ulga% z!eGqCKU2VfD;lDagmN2idH-g^uQnWe7(B6*i6sy@eCnxMj}N%gx@6UNHccMTysKcz z>&OKYWLprk$eqp4R3q?;Y%xnzZ#6K>J3GJ8@=wx-=#r?h9dnn77gALDnWrZ2P}xAv zA5`IEeq6X?SZ`?Iiaj!O_YbM&<1@6~Devwi9(5>LM0_xF9ipX!nI8G$Uh+;>$(_8} z%^u*@F`X3;<5=|;7G#gt|B1$as#11$jQsqB5VmtqV3EU`*vtcQAPZ`9SYIUz0PwhToFF zbWx(}B1xjd7FkwhTQs?RCV{egnzrgX6w~tlUq*by8K8o7ViQF;0dUS|@$8&h#dmmQ zKXfqCwd7H`ROzyP*ww%LP$0liC_YAQpSqG9zje$A_M&fUYU&2Kluy$i7Y_hgTuJ-Y zP$mD&z)SdgEE3e&?3+0kv>{E=kf>~>v&MdPc>mq8*hOAdx>F`jkG0q`!PZDHCPhq0 z0@r!^Y=s9hQeZw?p+8oKl_Utnl$E@{8hgGkJ zo}iTONhU}UoTf%ZOXAG@!kdX5kW91xvDml8csuk{+apK+yVIGc8>hkcmp5_2kt_(Ca?`m-d+ndUZT;+w{Pf^R))Pw~f)P z`0YI}21vB87x`OVyxslL5cAuAZ0UN_)&?!_fqLV@RJqarW?^TGG_5m}g2wVc4Y+6r_x~NO5F)mKxeP zd|{d>XHm9d1bkOvpK6{`Xq(nY{p7MywWVl=E}+^(2^30`!QU+Bk{6o&Io{~fD8wqoxR)c@W# zvmQ(V$-Pw#j0A#oYxb47s0|m2^%kdsm!u3hK0664HN0s&ROqjOcFuZ83WOa~&u>0k zXkMc-GzrV4T8&W|lRXTocPf4y?_39wH#tiyD9Mj|IGXP`nK0}TT)y~& zt`A>t_G9Qu=t)pGHzv+$AVgYpq}dbT>8!0*{=w#E=5(37f?Sn-y>%*^yL_GC4_HJ2h-3WVi+ii@v&JuHY6huQ z)JnG-^WdyVJA8ws0nQLPtR&{}IPeg;4|{nMf>vPO?{uaArMm zPUn$OAgJ1ZZ)0cwe^k9?R2)szHTn<&Bxr!(7J|FG2M_M)v18wO0SBuI{s|&N+K;Tex*Zpq0ewe*mmp@HI{N`lu3bEXn>$euD$w zs3eqMJF?ip!tIe}33PTU8}RyF>Eh@J5rl6s+@ilkRvFG8Ed*2W>y-1_6^l5h>0@u9 zv(t~KS4nekGa!YrEXXR!qqy*ITQ^+u3$^y_FPcJ$(=56pUP9Cz#WK-}+QlJ;(_oVK&O;-EIR6b4>DWgTH5>49gpjs$a^Fuj(j(<0X(!1dN z7OT+MmNDAW@6nQJdITCf;SLDXde${lCc1Vh9=HSIjAE`h)SnfTvt2)fX4iJEf3%HWbY~ev~&kI-70|N;ArinE826xgd6ODf& zmn%`(vQ?xt?B~bvB%uZ{Q4V@>0$`z&Ish9?OzW9QRa;vO@0}%4#Dn*PzS0A~>d{$| z>Ja083;rP$Nnyrq12Eqe6waTZ@MSzt!|t8a<;Pr*ZeBuZzxrDmCgD3$E9LXJY4Sc( zQ-}@}JC--!Q`ezlW=pLkO?0FZ*~Lvz;@r-exIEMh)g?O*hxor{7)20^@vYtR>_HT- zjLOANEoXf*;l8u#={#477v6I>i#>dk?Ul&1x;6r<%4H4U% z4Kse3me0jXCo;HTceKL)$-bz~y`6k!)HwX<)Z0ec*q6zffOergX_GSewe2|`MZXe>8mXqyWd@Y2lklWUXL`2pS9H-&oFefn&0`# z0iiz;pjhHpxZg1t`hS(eX;IGy{M32laQarf5$Y{9Wv8rOuX{XOF0lC;!vi?3EKV*s z)8I@hAdmlkEUkOhG?&3wVcjgf)9BZ;&b@EbU+>$^e5ssznqr|PQyeIj_EUY~38)># zG_~@#1{+*J`tln_^mkwAbM;c1v{F=mNog`*(GRYE!ZR7(Jh_GHy&~;rDXARoTGBj21ldntKakPqu7M&Cp_j9l2&q|{s-L- z^(lKh1v)?2&Y4>BwQNPmEw!|vi8sbA{Kk0Dajv0^J&{}?uHm6+Cw!QciT}!BD zY7>a%=@VF7uKtm-{&c1)@#+^o4#WN6s3)olQ#D4uc-EM#M(Bcg_~gvq&}YrD+e%wN z$hR{NX=e5!gf35|B)6{nTeiI)A>joMqW@$pj8~?-b~+B*_}#@<~44`cD?bhvCz=SH$vY@$Wn1 zwf;M%83l`2Plp3`x*DS2(&QjMwQNY60|bvB$W4+qW+|Kr6bz*i6d0ZddmK?wqZu-X z-&Kz9{zhWELUWnOgoK)3K97+h=5aX2AR&~UkNCe7`5El z6}Epp;VNuEi1{|t^{I_ZpT8i>#*k0ZY{I+VA`uk#NdQCC0EmE-)vrwO;&fU=@mIAWy1aa+Cz{ z5Cl;fWcxdGaT10zRw|?`O_d**QKs(_@HRo0iPC*zL=DL*Uj7nqqXgeg!~brBF*kQrc#wmOhX;; zZm}L|3rBC!+18pyL&B`2i7Hy;IYt)5ZWe!~mT!-w&_D}*sZb(bpqJC1l252gW@LNM z-I}6a&JG{3i+cKdj|II*Ura@>n4{`9C-W5W7Owkx1mJFxln zBV!$$4Z)uLFc>&=sB1g`i)&p(q=tGURnS3wqmBqk<_Xoagk%0Cb`LfTei5fCdz(W^ zf3HE4g>O!uUtTw=5`bG08)uSV^^TV}S)mzKT&2~d#jRJ8;;q~NvHTiu8H!&C;Zh;V zsFhu1l3lHp;dDqoRSGayyqr`4O3tIYF4K@|$EdAlmNO`tFx;hMESD@D*ReVwuuaK? zs>_1bNCt$nV4JPiG;NtTUI_#Xay;rU`8vXxw@m7AJ}656!$s;jD$&kAdN^Sjcw(|g zSBa1B#OG(HwlO$N%+9AGAyegHNuiQV zjK|Wz=JtZpUx8M86$>GRf-d7BBW2RPnEej{Oy7RGl${H!;VS&L7z;0=vbQPk5Z``| z%cYU7@FeDa5TCZSB}-W85)r3;D_+LUs@5g~G z>+yB9%gus>=+zJ3(s(+;ONzkRU9yX{RP}7aHw0Zt?V2~16@N6hol@kGT);3t9;mk?7i0~Pvm0z;e#H(pyA>tY-`zJ7!Er#EBw$CwX9lq z$%bax*JBuA|7~{Wl<$m88%%mJ8G$m{-2G3mr9Ik{;D5Og!L4PnjUU9)uvx+Zi-M)r z%H=ls;EM?N_71W*)uzt*RK_Ot!lwNO5?5OD2JU48%7+Jeb&AirJBH^ zB#4^VK}wS%=4eeKC9IZi$|=EfFnm{W877WD6UqI(%+>GjAvz0D6cs->hT_;h{hlfX zbQNh-{d?#c{j06LUv$tyrgj_#b9NldckbfCpa!y2dzLlG!-q?Y{#(YdW_y-i8d>yX zdS{qXpyBLl>}tVjeG#)3OiHvG{t`9Xs3C2|wDOnOOnpKQd=&gfi37{By*tN-pJGH~ zPSp<_Yr+jyg^_3UhpxXZGVo`mU!__XM-j(~@;R~-diwEJrcmMUO3pN7LurxeSB;!V z8m(gm+v}qzUq?a*)vMLdgt2A9wvvo>+Li0x$Mp0sq&U3)jPz;uB{G3g)K{e9vy)#s zmmzP*Np#B9g6=Ai(Fq!xQ8iHtb>RmLIyE=~GzV@CtfTWSIU{uI*Z{2VA+gByTATc;D%eQF?kXzf#mzdb)pCXu8cGi~u(eMal9bt#s_%RW!vz|D= zB3NNUu+io|sF^I2EOwU98=^LR;7jD452_~nN6Rs{P9ezbxxBGQ5ZII^>uvizOY_ zimpjC@VDI*Zyqczpg<0KGa@lO&1!E{N5XMDHP&&qM7p2OayEUZwkXR>RUS2OEWmTZ zvc}2#cU0zu95O6m_pv|^y~5Eo3XccyE*kY5aGI{LcUm|*c}SGQJL;LMO(N63M#{n{s)z> z*F3i6bsJyauW741Sc%nh;Ns9feMK+Ow0?$G-Z#XTqfmMfEqv;J4CXNFx>>B@6r-m= z>R|US>fj9)!&e>ucU7@Ag9IU~Lj2M5tuwFB`)jFlkVJ;EK11Ha@fAHJtk%v1M^@Kw z#^KjqWzX?Ymd^0WK;0#;Gu9%O*dcN6f!3ExlSQVfKIc6pQh};BLJh-@{Ny6uO8-44|$Qw^@(XBlG%=Xq3w)%x_f7KKCBk64W&&$m*;$3a9$RAMkyW@w{*rJf* z{QA-<$vy`>tmF%Z^aib|ogP-TI@DVpMFZLwKC<+ig9!@H8xhM;@LbfvLS|p}zT7h} z(V)T;O4ukR*jG_{|m678)U%Mjl|EgwmpQGD?GMw*S^1Y=Y2{xo$}dVL?{_JqR(#k z&lld0Qqa$!HJa-z%D-8Vy%dy>-^x1KTK#T50Rk|Rqg zSG3V%G+qiBZ1EcU#Bt-m8ZA9cHTI0#*QY<=1!K+6^#zJM{F5zF%F9ofmS0#sjyjWo z>t`O2`*)K3bWH#+%anJnMaYXu&3bg!mE6o?#WL-7Lizosss5qcwWm8JEO@#Mm#-1|8OC^76i%!D{)@ zPJ`6HH?FYutlpDbA9k`uthY@TCl4kFVksx5_(GRw75g~TY# z?(JuGtZFihI`Mc_xN{zD#%<${{1?aRfSPKir2o;)^mFbAqb>Z%2A$-wzJ7Q=Qjtv` znsAQ)sLpvH{{{@|kqOtq;4g8EJK@poGd^0d$W;178$3!A+Pb?UC0u&HLAxzX(**;U z<{dS<;wSUnv|xH{{GyEpCp%X72~G+_AHNIakxy~Hw+8qK;2&>ifc18lX?@9Kr>&1* z6VzVJHM0^UR&Dfiy^UWcw#*$hZ>;a%Db)_!e{A})|5h>g1vQUJDJguuZX@rRA0jWAVlLRQC?d|0VNjx&ra$-mq;rvuzcc^(HoiI_)h- z(8IR)*@CSRvW0S(pEM&sCH6I>n*mFT-H4pxU@z_xZsWGf?Llb>hUGb!HI(Q$Wv75j zx{)rydEI~9iw=HOPm0&u!`{(`p<`o#Gp{U4A%r6J_sBDdN;oY@Ahpb7R1rhTXa6rB zItGtHz#jb%H65&i(W_?bKeLbygq+=By3W`rt6t#*=;5BV%>O zwDuh6ks)3WjV!auWAz@7Sv3L%fzeS5zr3Zh?1h3%*8d5Nal9OhWtf|(;NCR&D?Ic^ zqzBiM5cWx{hnP*!TUCZlqAN0jFmNv35}IA~qZjZpto=Dd7}jTRK1teJ^)S+QvXI|X zl}GP2ejB6MR{cd~04j;07f}~!(;O@@9D2JlIK=edfWlw2Y7Lq zx|`EnI=bx!S4Loc_g7uJFaGTu>>NN<%0vx1a9Rpk>9RK_WJ?z64(U6(toAUE1>e6BmLU|9w4)Xx$*3vJiY`DEi?y)$&J#WFw7eJ=c<- zW%gS~bhzw?dKb~xgs<}tvqQSKSxHL{3fe##SodFsDyz?OUpbIPrk_ga>GRI`G3JOV zh&^WAr5;cZ*bPTc8S&jdcf`Y!;&~{-6Nw-!gTLMLH&>h4I5p*Z&wcCQOql;&LACm& z9j+r&lGUd3H2;urauR;AKxsHCVJ$U;l)3D4Wu6{|=F6qRMgv<9!*4&~Dm;|p%2wZ^ z`}Go~3SB;!iSV(dw+R|6wF*s%iu7N?95{V`< z-`_5QpXfjxm4Q66i0FZ()+ZCD^-?*fKKHn4(x;c>FW)|Y3N#G*FI1A%%li@lv|pQ8 zpCtRw++5nDqcm-yv@$Bt^Lc)qc&#O)Gx}coe(+g(_1^JOT>RB^)_(JR&SBJ`rt!ZE zooIPj^E7{#(XESWYsnT!)<&3k3?KxF?QScLpIC3j&|u-k$=a@G%E6<0Yr>iS9NSvJ za>a}9j2LHqki>yWkXztMM-g=K%NnfwOJ(c1nhhc?)U(5tcE2}){`}~m)YthKU-HTR zsJGSMrKLd~JiJTb?@~?Tu<3cd>!9_`<_ORQN!?m!@|$_XW^I|m19_wocim@!i_uyD zXR?5AW;p-v&;@wk`{-~W?-_OC<`*mk`qF1OFXK*4!GalfUv6uqg~vYw@_Ez|8I!ysEC#KHp< zLGMQLY`u(+bhudAx)C$Y}+>yNyIwr@5)9e=e+Q`o8bt{kwgz zE1_roGEDgTiTCrhUBqV4Py|N3X~ym{NPk)l8q>#B^m(heu*Ge|?xSPMY$OhY-|P#C_GO)?03TojB^n8B zJqPFyYA2@HUm@vvc;)IkR*r76!SQcfK8QAPwNhk~?;F$tFP%US-R_09uwTNvsb98m zYFi@f-tI$|3==O{^rjk=5^^{JyX6jn z*6QDXC`vljF-=VsuXM3PwmRY%8AfF2p68;NF4teWU94(1^t$~0=e6Y;D6n`vu6^vM zqtGjln)eqnbTmhDyp$3JbjX9)ZupCTF6V$rV@q2o9y4l+nO@DP&OfF+jRUz8|Cs$6 zn(VaYphYc#hx?P1 zGZV(@bjUt>Fh*_Jl9~2*D;$S8pWT!Kwu_R#-aq`mysn~HwIzP@vwV{CvuN`#lTU@6fE)!iT&h1jC{L}%<)d*OCL)A@9-@+6;cF_9 zic!GlGqE&qY&!tgv^lcAKwL{kk)T&V?x}3<5bi|5Evv7KyM2+JXd~uoyy!Mvi_GG^ zC)#wxnH0hoy9s<~EfQIo|8vAcT#(GA*@U{r*^)7L(&N~Fkp&oDalb$U-a9-!^=y-@ z7sz{!UE=y(Y$?tDwzoWEzW_E)ZKbMarCx6602C}vn&c+gdE5W6mv=kb&r=FlEd%qH zmb+cq&M!IchB}*6GVJx$ntnBDPxGk}Tu$8GME2RT>h^rqOOnG5KV1KsH(~!BAmkbN z)Q#hb`BZ!7mF&5CVqfx0vT<_Cs`(-sxSx5ZZ{K=oTOsMQQ_t#sICQR4+g{s!BLm_+ z@bf=G+5l}n7u#G=4>0-XgC(0J4DLNz&7XlIh6-v)^=Y{5=1kZPZs#0J5>wXXqlPy4 zuM_EBrwQG$pr2nQyRg}lPxz5rymrIOY~kekblHAeGQ%6=tF~F1I#n^4VFhjrwDjk9 zLOHv+Wzk;fbvsXVkrEdua4aqe8f&}M$ybth3Nlu>YA9|QU5*%g+%6vbnwQ3U_+0it zdNxPdkX@V(;E|EfB!of0KKQ7#I68*9)?xc))_bSKpo08cnW`!_w85{!w(6(gc$kA?_}NJe7_L}Y8%j%~kCq$m+_ovyr@ zSGIp??W^uiEf=xoAoSE;x2t$G-CdPY&6kbUf3tCA%TW+omm6+W94{{>x|=5|jf3!H z43+yPysx1IqgcPOa=9aiC3Mmf4e)i>D3g22%_+&2Ygcc5bNDSgg}3JS_^ZHTzH$vf zm0M0wX+GX9OM)25R=8~-TlbnnKoF(VLX?s^KtkSC7(5n2Zo3hbWyj*lU!<&Vw@dWe6`z zwfey_Idjf!R2X}-6Ju7=E)Twa$WJarAz`Gg3Es^zk$Sd)ICPuqAIcoie!^gW$$;%H z&6Bjb`czoqv&}ey;|WvNo)N-L7S!1QaJB|$uLV9_09Q|6ncl}CCitxFq+x1;EH1fW#Ev`0+E!rg-mbjrT8ovXU>tsiLKU$1xu00$AY zEM?*5ZuM-x9}%ttdXDp$UP)emc6Rn3_NX<#a&~`%=4{AW>~r^D<~~eZPGJPZ%!Zk+ zSxE2!*${*5J@yP_WOeMn2>>vAdgXfL^#37`x211n#GPK*?pq*t?zW0?uWoEGFNEin zd)d%ge-5(mNZ-I0N^jAW^HGhn(#g3un19X_KhDCaon7Q}aX|8xgqj_n$WrsYU0x!8 z1?v)+gAJutj<_uR^FBJ#H#VLu9B{*+Z~pFw9(E~)E33S=nY=~N7c5{W)5bS{F2V6N zw*R@kZU%NnbEW@Toz!^SBr0_94Sc`saqS1aU%qeRq~nfkJp}>cW7c*)RZpWgijv_0lJ&B6Ytrk<$dNgiN}rh@o!b@H5}mXsvJ0z&*O%mCiKMZi19tq z=JLvGS>FXP-Tx@G8VrdxIDSX3?u~xuSvKf;1il&)sBlHU@BUz|%8RD3f1dW7%;|i- z07(e7%|4MsP!fwV53a((_yWvZVlnc1lF8XUqz6l$E8HWUI`oEC%`J|p?AM^^XWg3v zNBuy(SCcm>?5X_aH~0B4?ck*lk0L~!phoK_rSa8Oy4F|Q&zT#jNyVhY(nr6|5xxuE zrI-p}S7+Jx+>TZI&E8*@a9rJkKnBu3etP6&eE=V${s-{QL)V#JD_AmH%=gz*WbTeT z(l&b%o7IkL#dTcPwSr+cbHe1+Y&;>8?Z6Bt;?q{PJv7p4IL ziNniBM8z7Lh5k1_oK7Z+gZp=|Xpyl&q(~rpH+uI;dNjr$;9S^ZT`*5{gWxGpCCHa>4Q|ka4C8 z)$v>kL?3W%78B&4FRyhp{;@V6FX6s+myw>}M~6d_nyZc5vej<&UJ~z{xC|TNPwoW= zH@+GFj?P&>F`dIp33pveodlsz3KKcc)>bdPZ!~vu9@q6XLHnsBiI3QOAKSgGtgC1p zH_h<7g-4&>&yq%g9)$Lkx+xjr*uu^z2 zuFwyiBC(Q}CEnC_*1nJX-6c5D)1;d%A2I?Z4=zV!~CKt?dJe{e|;IULuA2?^6 zmFEBp)4E|ZrH_ocDqfwANJ_a+#1X4$ z!a$EPkK@B2=36+2QHxS7zjm5q_jHj~6WL?#($a@-CGN8d`h;2HR_Dr?*&~~D;J0S& zdOmiaeGY`&HfvRq!Js6l+9?f7Imc$%5#;F0?50~O{BpHT*Y7;j#l!@+xZF%MYf;6b zf{Y6{+N1L~tP~jaf`K${M9Kzp>INO|AmxkwZ&uCtepUtq_}q{To5()*DT-w`qTEjP zCK{aP4Fi`+35QsMg88JKy0w2?HO-KwdHKpDo5Hn(w0EvMr<+`5Y`>O;lh*`<1^^FN$2CS)&qUM z+g{!xsyHyzy(fNMJN3Nl-}rmV%E%4E%3jXt3&b-uGqFZ{oVLcqb304 z{6|9jSu6y*_3;p}zq=6b_jcgZ*4ROb_k%Uz9D$rBdG?liO$uyIvE=JtU0)Nk8R z)^nG31#Sn_oI-C7J%rvoIV@j?PfxrYkFOpNdRDqnu_3oo8X$J^+r!MQ^W()}+x^b0 zm+tY&>9FC4f6$v5=yWNAK+@#Vd~aP^Y>7+9UIM=0=(;F0zO;K{CeS}`!eTus4>TDl z`RsTftaNF`pPJ%uYkB5i_gd_5+3Qi{4%vt*v6b9-+lLB0;b5LR=v?N;`#)CtHxgC( zhNmuWwjQ1hnCWn9{i|C@=1$pD6Jg-SN`mWk~U9xIEYw0QEU( zK7+n&XWiz#MFV`SQy2NPg8+D4rJ{xg6<19*G3Brt<uTySQSGMK!pe)Q9Qf(`1{t4oB?l#dp`R_3iaae5ttXsycPd(`<=EYgixTB$0|t$$|djf;4ZM~e$R_rM7gVq0K2`J|F@pI z|BHPlV*E4CGu4ujoJl9Wy8^kd4RWo+j%%t(xf+1)rU&nPlVJCK}f% zx02O~WfeF2dQSMgNWf!NOh2pVDp6d|lYbyX8p_4Av#4Dl^ti3HAJ7M(;d15P1(Y0(z3L8tG?2Rrwijj8Jx5S zm#^;n06Sh_^&?Z{6K%vA#>SgNBSIGK8}q$<*NFFEz*Z%_7^%^%#tzOg4o&NfUeHqj* z_x4~f3a$p4d2OSLycx#ZRpMk43T1$YxGa08xwoQxmH&Lv+{0eEHd$o=&SxPh3;ig! zwpz>csw*e`=?H3Cl-92SbBxOHHrEl3K_oK`w_IQ0Co!5Gul-Lo(Cyk;eUm;fg~$5J z1Gfn@CoIz~TVGX%(O_#a>5cGXIsn`PcaeYrOG7lD*~W(>0uG~26Qi9{!<4G7yk@%)KOLkfe?6;petHC5Nr-^njoVX-gi$}L7Ua)e7KXNXywt5v`-e2Noe|8msJQhC7d2GYZ z=b*4nw*X=9bLfTYEkT^z0cdk3Rw9(TjfbYKhsa~HWHnWx=}|K2rm7KN0?vk@fX#@d zLjS~U$=hRfrb9Dr)l}=Au05Ka&%4{8ZqcHayKYy~7~ht}LJb*9ot4OJf+DEG(JgU<$H(!tgs_J*ASDe_;XHCv)sHYi@o^$lfyZ0inH>K?V z3u>}mCEL~E!&k9UE2WxMk(@;yL`$?+YB(Go|C7~4;nWpIds$cVGPQyEP~m&KU(Y`5jQl#luc{LQI{bdRQ7&l*oS0)kwAG(A~e@f8XC^G9s4SwLHD@ZP*b z!@4Y&&e}g%DSI6Z)tx*cx>mM z1~E~ZLOu?b4sJ2>G!jbhC8WX2xE_{>0dM-lafFysAL~}<5s@>MBo{(ug#vURjCOSj z8DLupBXZ zb}5DMI2XFsQq=OD{r3%)zGI?7aLZhf?R{ym^@B&1-%F_227T+UM?PG{Y+>4s;>Z6X z)@L5T{q{8$s`v7;)y}7gD>Y7zS4Ra?SA<@^vln4kzVf~B(wpYuAsck8zug${8Dg7T zpG^{(4Hdp*fDVRZUML7{yaV)d$Qya3H{Cl4vERKmq!kWQL*orlZwA?u5imsCdKLJ* z2LEE|->h!V9+qY-3pD=gBKgBSaT{d%2M}p#ePgOqIczj&oxtnW#aS2nyTgw}`(&wv z#nj+M1LA6|6-!~4shyyyiO_!LQDv*DEmy=~F_^x3U%q0uV6u1gme6Xp(9_i+bGz2z zj;M_}eZCQhH+0l$5c1$YJgi|mys9~28UR**xxtki?I;*Jm`@EvaS-MX!skeQ1Y4`j zqLP{A@_jn`1ZIVd6TCjLd99*eH|dR+`!6Lwa~$Sj_seN0-0BckBX8aMQMr!tXsmV2~zFb^sE@G-V(2z56fhxn5KjL z=MxO}vb=u$iZb}L7!^iR@Tc3?hZirk6z*a^DlGO;-^Xq6aI%-=h?c$sf{6VubNat6 zk!**!>@aas^(o6Jh1Y4Jo2=At2MzR1GPWz*Uz|r1L*9QQAc#rRO4K5-)i+AA`#4@D z6Y$pHHw4JIJB<}F?Z^}2T!WAIMUazZ3RfDPoAlVTQ1D#U6tCRNKL-eR8v_=jy@?akp1mm7(uCbAL4?Q@r3dd$o= z4L4Cs6idg)v|!zjy$3bt$yRlzewOxKwVcpjDoj`wU?MHoGulhi4G{{EkbJixf;Zdx z^EeDPBssF9X}#}j6Pbuc!Qw*J9bw!YcCcRQqI-O5PzY~r!kIdN8S@mR9a(>Cy(V+; z#+^}8^;UCQ=cA$%XYOPkLeNGUro*zoOSp!Vb7lRFD*Xxnltm*R$%qNhA3*1d zfF}17 zz~*n}xH&Oe>f+dBl#fm?Qr=Pi|(IU!XPC*F7FqmI0?akmF6UonU2VjJ5g zgL=tD^1pGhVhZc4D$}c1C33lHkNL?#tt8>^8&f?E9lcBYa&CjhX?_E{t5k#c#R(Tf!;QX}@`Jsrs03eu{g+Ko zNsBY3q+a#EI0klbsC&e1RQ27fJa-lpapRUV$s;A_z3ewiGI)qgt?J_&EHf2e5BD>P$BVbyNSDqL-9$4?3aJz&Dl3 zmfwdFBJUL?#w_ed!Dh^zh==?@TV=F~2xE>}m;|0%@C+gm#cp#G;5cyS){l(K z!9k)Ry-wc_U~9hC>Ob7gLIPxym3P->C8_kWH52J9{W#7{QyL|OUiwPO1lBEI4_d_M zQsYlT)L9Z0h{u>gIC2mlcgi+^DvV6r0@>Wt8is^h&~jbjkhxr}Vfk89k%jvoa*8{N ztF9ohdx`@66xpcsk!wpY64K7|x!E)RV&7m#-8X7|OXw<#>Q=jnXfoX6bh<>#%2?60 zuLM5S$7%TobFFm=xAOKEda>XlQPB9_JUpK-^$ah(wAEQT4|K?)D_eF1P%6R@+5cX#;?)+r`+|WikZj5zf!$`ci z%+LFYA)e0obC$_nm(m;h?(Y-o=hWEztV^+*L~dRi|C3LVH!7aVSofUto*dH)p%-R9 z_qfIv1`X`-KQk?Msv0|l1+pWNT;{Cj&dmQ(t3p4ywSdJ6MlZkF`OG4+ECJD`^EvK- z?fmuQyZ|S9>AyWutx7elR>C#04h3G*dxn$#975)`qZ(7Zi0 zH-U8XDuqF*+_r}&Yt03H!(KL~OT=aeP0F!hn5_bJuG0>>FQ;=`xw*vq^#Ts0mQ-}L z@(I?@|ApR!rjACZB~)}J7i3)M<9aJ+eftSx=Gh1)j;)-Y6y39UO0i4$`(;MOJfOmI zhx4NS&W;G?0mJyU*Rs-S;#Q2ta=b=-g7f6j96~~_{U-XxYdjfYW9a#NNcghu<@&In zP3sei&6sZ=*edtGPh^!95td_RBrc|bC0A%VH`>ub&?(=P$!|51ngT)?z9;S1#0&f|fLXq+c({0As0Ixu3ix9VoUAhIa@@_3YbtGm)mkei>5V`Ge@s1*5LPzKx%>UCNn)BZM^x%X&mYxKIPrdgW$$KhuXjDwRA#BGheqI3z`iC zNI+M|l3u&OL8>>KOX+(;h5e9-QM*t(=w3qKj zq9eCi6NN1^1D-ux;@u2m0jI^sztG_;XOY|tA@8p~U#*~D{QjcgpWw}mq(matbm@X{ z-(jp_xUg$z-WX4?-hIHZvWvc)t+3;CB}#jB>vZR1bV`WV!s(C1S5G^RS8BRc1&L#0MniP5EDI${UL-MFR{xK#CqA1oK20)QBPUId~|uZE*n zlATI=fmiq5R{^@%*`Zw|&1BS3&WzP{)shZD7b6ve0KIeIYDh||W_vh@Cc%|-(J)_uP@NTn&h<8d<3|< zH5D;|u2;OyI8kIba8Lc0t$3w1iW>d7!JW$%^Zor3yM)u`HEsFuZjiJ61C^x(cvjE% z3*;(!YyD2Aq&@&g2>L<(-iD)^#v9Kb{``xs?-$OT>&qTXG!xiUwW>VTq z`oNZJiG&70^vjZI4w5**2RHk`FQkRn?u--Mi;rPu(wupVzU>Wn#vR^W?X8w)>i@D1 zanh9&2JjtHY?0gY zwDTjr+cpct5J(Rp^!C_BZCG`i1ii53otC_we$_CVfqYMtIe*|R06xYly+T7Vo0#z5 zGz(@qTtXK+M}Zq;hPOrf+=9zoN7T$u}b&|bi%D>`xgI-!B4 zgns(}@GxbHp=d;dhNMoC`d1gd;nG5S<4X>(Se32Z_IWJ{W&DtSSA(Iha{}*oM`$do ze!Xv&@8F&uLGSyS1N}So0Y+f=x}QFExRTd}xU^NNP1au10%MTh!gqi&TY5QVz~a+h z{*8K{oR>{6{pa1L3XsB8v@@bMA;G3G;37CYA3Mh$=gtZ;Z1rlL`1k$brMgnTmNycX zn099@;YVX?hyK^n{95LVVa=2Ah~Z=ke~n!hm8g8sL=p}8bTTs}8OAUB1uHXN!#q7H zyr({w=V`1wiI4S%*H`o5weOHPQ8X(kp8~IvEE;$)xj*aJAFxr(=3@)E5l#8WE$hCp zQfu8%S42%hUs-bOY;BntBxMfG)jhKgdu9~_0JX`mW&6kO9tvAz6te`11RLa)j&`7TO(2xD(0|zvc5uPmQOiohFO&b3BXhxMQfp>!7)1kEm?#w za+p`tSC%<3IEh1GTSZ)--;QTc@D!pwNVr>X=a2Odce|Qyd zphg@Wi|ZOk#+r}w>T{*8+$La#n=JvMaE|gKd6qHXeQMxz;%`xM4)>gF9|~r`fV(uL z_LM#hBmy5G$u!JfX!7|igbwM8oV@99#W$JOn#70X>R3#s5V3(m5Mt8ssMDov%1{qR zXyZs%La09@>t4O_ZDT%9QI%)e44h>F*oLKC246jAmdm5p%b0imF&^$qhfI(IbW3lc zixDWUXZW3h16~jpSS4r~B(8vaOqQ6zNfE&dW9j9Bt>L@;P{q+ zVA`2!Nj}!fGGp)xT*AzM*^**_t#qZc%)|x3Wq%CjNXUGwLke3kcATM86Ve$WHlc<3 z9b+3nzeF>>6g_>5mB>PX+>q@g&1FN!zv@>*2@6CZAO(xq4SH zt9~)ai`hilf?(p7{heNU2|je=bIIsQq-)2Q7nJ5rFL#}n7b9=B#Ft-j7BQdWegxRx zzX*yw&gX|jGe{3wWIM5wJ_rt7SrEz&XVN?0IKTp*-WUFk|Cj&&)bNlNGMOjp8vwb> zo%WaC&DwDQMR>D5W_p5PN`gF20s+{a?Emvh5kv0uoc&Rd8Vjs7S?WpdYarzXHxnmu%P_WrEDX-nRjSQ7eGpC&rDvEtv!mKDGmAg#hTV{MJDyDc{3? zC7U|{X>g6RzRIYkFyDf|zGfH97MW@^Yu@ydXfrTNKZ*zSJX~C9h%}8`gKfeR!eCMM z4-;mdwwl1FR%ymR_B>_WZ2Xql26Tjw#!&(nnG}4`3tq{1UyBixE;q_l=TeTMKduQx z%N{N0@fgbu#&QFqhk$CKSd8%+4ttY4VGD{1+Ck31OCDI`xW$3LuM(U3Fwp#n)WQ$C zvvtFD@LGOoM{^$ACcw}kHQx@^A9(FEg_?EPJ`HN|X$tkouLoH+cbrh4T?5_+trA;W z^hwU$g0v6xDSWDl#|QEKHYr;b$uuKLTxG;XTh)~hK|jmA6_E|6)T%%Sr5Hm&pjNUr z3@+XE^9eu$nY!{|4!0Z}HsLr>!zm3HCcwoRKtnj8Ugb-=#cPT3QtAXL84*QIQJOuP zL#N*l$x~QT3GEhK0hB_6O{ZB%ND*|_oq`5&COro7J3*AQzgo25kR*HHV3Dr6wd97M zKgL%8WmJ5=u4RIs8fBnxv!d4uE~7D{0;41NzlJl~(4@7thtbU}foe%x@fHPA@JpfF zdN;hzV&dSGjgVlb!VE0qpNbSD72YUy)?1s-lW(faICWmMyfuSU*_c4iWVK%qs!!W6 zi*1PdY{M)}aM{G8amg);`S&jk)K@ZGtK+N0>e|OcCHoA+IGK(Gvkb*2K@V@5kCkd0 ziS(bpCH=4D_s#A^!hEH`=>`8N+zRJ#m2)xS3FAT2xBX6on;6C^Iou*f_LC0H)~bZT zXAcOQbO(v)b(=`5r$k!pkxT>-__Vx;%I5Pq(`n^VOog|Kq~atSO0)%4+uF2|j~LD? z5Mmw8(M;Y{w^Q;sI4;pfg*iJmeQDzIEIX6&bHz)TleMh85oYehe>>;~2sQ0}Z9Kq) zR1+ug6(V1tNajd4T9S{3f(m7~0Z-w1X-NlDHiBy~8qzse}oqj=|%)ixLaGpNAvk=`d%r5zI&- zpRrD2_WOuMfpN;6QDD}}1HtOnDp|ltF%KlW3D(R8l;^!}KpCx?PjUeHo(Lo=JVwm? z5tb)r?N)tSRb-H@1N6;!%PwVLnQUj1>|C<=DTZKlj?5y2o~lYEAnuH?5Flt~uVqUB zkb=S{Vt7ABJiTFqU$-*7Q>GDaQ%o$ zBPIiNqw$D;S9FA| z6t~Uz($XJdg$C5f&Cfiej2<%q*S5%jGITstn-#APw3D-zm}BB9V7k_UJ>>neQaXc( zGqJ>kCbw1SCx}0VU6VTog_xkEF=uQxU;wRl#&c-GYt(&9T!AHMNh|o4;HIv`%ru-8 z^e3tK#)s>bZCuvH60C1G%HhrwS_*nn^g__?9?Qd;PxCgN|5Z+BBpBh4q7zEv(rn!sy?bp9j*l(%0a-{xeXO>@ ziBdusR|m1QAW$jDAs*7E|DyCd@6TE@-coz1_SXy;0n$Fh7uIcL3|rBW zdO%hYFEjRZ(Iive1xaACtRHpyDe5sVm~!T@qxDAaQhTd98FSfqGh^nUBm*tUs|YjZ zIt6Sjw|t!oq2G#D(ipIHGNyEzv{lBW?|KC`TFuSEHjGoK=r$PhJ>+|wFj zS;jcNP`RyHWSZf3k}OO5hO9?YIq={uV|G^JT;J9^hBiour91?*Ed3tP+MEb_5ZWh> z&CFrlo8Y-bE;zjO$_K_ipi$F+TMR}4xjYB4a3gyKN--3`Jg*;L2AXU1Oq4MteK^82~a#L-GL z@zOW2vM6L?Gu4-`{1${V!zvw-68mqvm+BLd9kyj+L;(b0NT_jYuN&v^3y7|0zEqct z?2mCn#?*Fv)f>Wo5c{ea(w7)M-Jci0 zAHgbD)bV%(_>L-;Vz+3=W4xp&;;JAXal!)<9b4=ya)~0U!cx|qC1Mm}R~Y(Cl+_v% zI&aJ=tC137Ql{vViQZ(rhcj~^(Hm?Ll-UW&b36@sv%&03PD4m$ATSZ>*u89DBIITy z@X-B{Qt{*b$25IxVkO|BxDCQ0VI}k)qUR8s7vE`r1_Ka3v+jI`UvLa3GNo1+(pEAw zj$YnBN8e(HuYKH#)J%F4=YPMz*UmhCA0u^O@D(iyTj>!QX2?@eLb1uCO z6+i<2dZvXoeilM5ld6Qs1KWU5jM$Qk%K*`Ye~Vro61i&DpQd*44qD+8h%}W3|IsqXg{o*l@P&j0zCU+(iBoBA4RUP zghk?RpblI?Yebwfz;#JzW3`5tIppl&z~i;A%wq(`zlsd)=Qwz=&jw83a8F6Xu*5YHQha?7P>Vnp!ygLkl`#SS z{doCr-swl*g?c&?7QJ61;{&F{arLQqHE)P+22|^t2wAcpA9FOU2d_eY;C|$@w@}@=~R4Os3%eK8E9V zoMw*oAEKCIFWLE#1TuZ|QdKbZpEgcB7DC{Tm6Ee>`K|1pE3f!F4q7xJu}BF=*I~AI z92qn&1x@Ogt}!{CX8dYm-BuP&WM>al?rF0=c9XJHeG^hxATK3$$N-?th6OZHvRmOZ zHq_GuQd6QQvn$_e-*1{oy`EO`)OVbP;TNNUd6ZuPEkb&23B_Fe;5h&jnWS# z^(UkydCPUiIq!R6`;P?YXVZkb8SNwm1BpCpj3#Fr6R~rlRUJ~SF=nx(904!1S^8`F zjeifD62Kmt)0~s`1oGxH5Z9YD%rNh@8R-w~xw6FqSUd+YAO!swUy3-ugAJvFjb>!l zXpu(^iAW;Bd^TlQ0eeSBGj3JbqTg}WH;wYt2@rnE5e?2RN+w$>6=*yrzPE0l(Bkjc zQUgE|Y0VlxsqX_#dS5^l4E>H__h6Q`VHK;$_z__h{;a%Q5G_X2SwFQI^TcEXj&yi! z5a~;Uj=c8~`CC@?zN#oFo8?1!@Gq>WRCMq)yrS^m49jBxCbMTM^*ths$k?RN5r-)w zj+YBtna}yJ`7?41DtBkUre&yvT0MOt%lR@r;9mN-AwC=e+pP`_{f-Pz^zZ~b($^R_ zffujRr~D|4!&Br*f2j_lt=+R!`oc(P(A1R9LS;(uvI(Nb%A1 z-)A$VAm}?x<~6jRYxDY^40krNZwzFuxdmf{K-1LkwC!p z_dYRYT8p<3E^jW&;>?B;^^N0TvBHzklb%UU4iyTItdb`ssBBp4(q<$ECD>g@$sNP0 zRwsB+Oq&DO3+jLGWrM7CHWbt?UNJO{celilt-A#@TFYH<2O9u*OgmAIG2KXNZmbxH zM_yel5=o*Aobl^Ym;|EL2;Cpo$f{zIvZ2wh1^ag2&i0~v=l6r}Sex#esF6+c6T8mv zgcxNS)=`#b6FIhqDmT(bLjut}mI|H~=q#|E4 z6$J<20%mP2KZ;GA94;DSI5E99IPXz}N~>T6@UP)852aoc7?|sG3{GhoDkVK#BX85l zK-5sG0v=hixY$T@9r6TWy>B1-0S+Dd!wz64IQ>b>`#bGT-LyAp6a8db2kGAhlE>Z| zV;G%fsb6)CBFL|^1R-aPBKa8UeU?!9n{1+qHHnI0$0Vx7B&wr5#eN@3La=zNpxLS+92&6#L#Aiv4Upb4IV9cl3IQ{4eYE4vM|xHwVR@ zV?<#NtXe;w1+|_$C~dV~d4S$j>s?2!?|qe8@6M?8Z)g%Nwcedk>w9mj^+Y7(61AR6 z2XJEBBVm`*sv!!CMq%*#KC{SfxLzX1L7$InxDc%;E1S%1eH~%3VJ1wNE_%ev^;DkhuI&rITsi+9(UJ1t;d2O?dN+KZVlUnx3{#S8y4r0E; zQg}GZ!NTl;N>ayyuF+03X~mBMFgPSl^X7gE!^m0=)teC6NuFCR{1jGGAvNGzpm2+S z*C?IF72_u=`2^%8#54OD_kpK6a(xCD3e;Z#Pz+r@gh_CE<{$+=v}D-l6%a{@xfPI( z-0u{DdSG1X^9Ijw>I{}@tl7sGd8|w|j!4D93MQKbOybk0l|(`U)gTN`cm>3cpi=z< zZfzmwI6D`BL|4L6x#C4NsLcqix+BD;y?K)XB#y~?ABlWqNx%$fRpezp56Gsa;}!Ce zbN|{a!-S?t7%{HMl52|H@#w;WlX{0kNLcMOw&yf;7b^RRUAPH2<2=IbcWCLde4qMr5r$Dktfyty2 zvDATxN)!WAb^b{l6Er|f+|vOEDuRdh0u*OOPf?(OLq}`DLomA!H^!2a^b&Fs_s}8c zB-2QWH!lP7OL-NUe}XMt%^)S@Cz3Ie13@vCw31H}81mbe*?uGyYNkS~Ln0az5s80B zz&HZkZ?u$y(7O=kTnaVR5hd0jSBcFNfM4QRGE5C$=X`;)X8A#sII>rQO!GYu2BN0_ ztRK-Lc~%FOay#h)VO%iNsR$N9skCHZ;eTa9p}z|Go45|jx4jiPr`;G>_ha) zppmTF1p0iUeeQ6HLmf~~AVgBmXVX1u8XhtAv{UYu@wf`yLlo@i3OT46Ae>CHU%~b3 zLJ@N8BtEGmH>VM*z>sJ3Bo`W`{HF@s2g-ir>oUs%Pf8q0D1A7Rkq6_N6B;5IHdrgB zjD~K4br~_B)^$X{zof`Q%$rO{Wz~T~lho0Xx(al1cGppX2T%*8SWjWdFbU-Av+H7H zB$Gz2S}}W6nT^0iq6G|V(Om^(j=`&}ovC zQBEd;Oy@%t_$0WxT!Jqy#iJ_3N)fTg8D5evwl?C+l2b7;*0hkcdMO5Pg_v-)`qdZ4 zJ&c6HV{3VDA{A9BWFL}?QV}1)0YQCYQb|tZJe(@kkwhml8SHB=RE0p?Y&;WAS&3^Q zMj@0oNaH>#tw<@Q_JVw-&hHO!A4AhlSZXM(74lb#@=#JDZITyeAUi6}l<}W3CcAU` zK#A{eIf{XRX%@&SJ>S=z(jI!4i`%s})gNGJI+K`VRiG8|$@2VpLkk+xdiMhD22>)7 z0RGHX^^|3k(mQ4esFSl0Fb8I1kYJ>wY|cfH&6Ad6841>y6zO11N|LVs+zZr4UGHA~ z6uagaG;`H*H2dd@`5TEn3T7sIDpm$j?YOQ&9rJ&@Bp?a<{~S!mL3ECE*~Lr+fA_}* ziKt8~qiGXb{jlnXB=?UMrz{PSu0)ZQVxI17E<6|4Toc$jR-tYeI0i*?P0;Umd$ndpU?U|PTDM5w{pPW}%`=1**Tai@ITVz<2*LII+k8dBv!3T^;iM0&-N7gvPC zvKp}prHj`g2%_sFh+>hUj13DrHVhVfFs+XG*gl*r#}*}vV&blx90`H$Z#mT{<5Di%BzvO#pPf z@vwOjxLX|q?7Px5x|J#DC7>a*^LZNUJmL#QqX zQ|uXdU<=ewQUe6LgJ{zTDZRr;HCZQb5pX@YvcOVA#d-&l_s&=8{urQAL}psm!sK2$ zcGs?UVI~j12HJX5m{`hh-tnj2DO`duWvQcEidR`4tRGk6#M>#aVpIr-B=jm~*q!N~ z_Kpf-prYlAsns`em(x>#3G=<00|vVcv|;k%HIL9p_tBbHYcfNIG{bDFwTxp({(FdP z(Z8{kYcXv_+h)C)oAst?w5?$WZeV7aY6u78X642kn>_;OAdff$R0a7#9o?#f15py90wv?}GDjcC70uUB-B#VRveA%fN0+K2NV$s^{ zA6OcMVm+d*kuf{qsxg8E9Y{4ntfyE`(B{%#0+Z7yT(K=5%O-H+C@&Pwg69)(jm%gj zC9@B_(_4jBDwouwn1Oy|u7Rt5J~ZHG_G(DpG`PflbA|A_aw!xS{tW4Cpx`!?Zc<@d z^d;W)u2mOY;d9{350zoaeA$cct z4+23&U@2C2$H-9501EM@qqFQ58*=BWubA8h5cZFUmau;^b{7+IA`zB`bvnZl!mOMb z8AbLRLazh2p`3?&j0l}nqfmm=6{GSytF4oU(cxfxN0}Twq5|JEOcET7_4WDWt#%dq zE7U?5DL;O}T0&sV|3Jus@Lg#|8BB~enHEdUNs65QaV}ao33rcXezc28p`$&XO`WEV;5AGzVUay_ za0nd*DWzO>N2cSWGf%0%tQ7xNKt`w5DPd^&2*wHyBy7Ae`2#FDZQOqFb7=z>?J;45&6ewL=)wk^AQeLFfRI$!LTHn(IRebs$JyT^Z;MR?9v}+SXqP4c*Pnp z`Hw;BDCWUJN{hTXVFTCBxf5;=sHtm9FE0-0oGe8WY=792o!e5%z#7*Ig20Wb^@*x9 z?8mNnLUU*udX0sX#HVcT(KIHnq+~dF-Gx(U=*GopB$4lPF|!6yZSaU{F%;6UW=;z# zuM_CaHLUl*PE6t2V`X8o2v*FUE!CjGm!1i}*hu57C-_^g_)C+L_W$DNg*ZrvVN!I?}hk1N-@DODS%D7RDX=c7=wgxs&X9cQHf zB+nvw>1Sid6vTL=jo7c$Nu5BD*#d#?b)jt|i^MrFg#}v^?7D}4PFHVj=@VFA3)Gpu zc~i264T^S(p<%|6l3SBF4y%A*C?DC!yxUWz+1$(qLTgEItUOdgV+P;?L!D}?EnHA$ zS`PbJ%V9t7&E32=OC9z`G0Cl^wADhSC}K_CxwuxuZHSHsrL@~oc0w~oPu6kIuH*a- z%M4$QI@!8kH}EG)1q;~Ntl$NJ1b@h;J~5HI<+QZnG_Tzz&i;HT>c;n{?UFqWX{m{+ zmMdKGqB>NJYV{M{C_hiR-^+1cYi@U}=xT5iDeZEH{WD3)sFise0|eim`tFkw+B_>D zOVTs}xOQF3xmvYIVN%EZT#XHhY&laf{!3@n;VL-piK75XCZ#xE)x}|}b2Ncun@hSJ zxsdsht6xO+1jmC`ooC<@uh`GV@AAYt8+)aL&qV$BdXB>(73YgH_qhh81nX0G3;X~4 z2)cI1C3}^QLXJ&K={~&5MnstL9dZNHC?5MZ!!wTvaV|YzF_P2p`CZV`LI<4b!8KPK z3K=(S32F2M_^C9zkdQM;B!U}H)vSGjGp)0?=`CSyGrJ(J= zBdzv`C>^Kg-PsYJkK?D0|{!lENxCI2T%@l zl8U@zE-Y3)q*x;t)?k<KsPNEX&1c=H<-{%i)ZY8M!%py4&G8jijL zl==kec1iu7nUznjLDeY_qL&_+i~)uAZ03Mu&|^DU&Qw-HHI1EOAY^k69n(>@3#Y67 zNl?K`!ns&)}_!Pulbmel%P>zBTX{39eHq#l`TV7=%djdozVt7v5_hD`UINoR$A?Gi!-129U2 z5*sg}Fs95oMaYnbpquQh?aN#4oZ0(^VPDKsBX6u3PB7)M%Pf@!U0qfO#oWhCWOzk2 zWhW*Xzh_fW1Qx}zG7Rvb?wXs{aN#*xM}{2XoH-Ihf&dOUztRdY;Vz5%4g*!ly`_7b0;R%$|%XbAH~#YstQRsO)FI{X3A=Sfp6CejpM?ne2-^*qzXwj?GhmBgW3ii?~z~;BVuefxp4ecd;TlZZRzv4ZuKS zF(J$lMeOppld@5$Vxx z0aoQW6Ee6UQOs>0HOj+m!woxa^@gjTuq}UAe*SlFLH?~@BSZXe(z%IKTuH&Yk&LD( z^o|CPl59%9?k0s~;SZ8BabCrL@Vn_#TIC^;n_94qp!jeN2>X#lv8UkrHBP6;bIA)< zzQJ*r#PTFZxjbq}p`D#*b8&WB*_hpl$ib+N2NfPx+I+{mm`|x7|Je{MW>aQDOC7md z#^Id&!3oxChE~1TK|0lIy~O#!EsU~M#O5`}7QY=}t#cju#CHd8!B$cb9Y%PI`O=Ms z!yy;$Qu|GCs^J8BKCT1FbU4&nE>N8Dt-&ZRo=P@Dz6gpqM~iAnbirxYef&MDrsl(- zU9@bQj+YLr?vb%Ph)h9sx-w9`bG!UapY9~_z}4rS*7$a;yEDp1l&Dcx@tp>2UWWnO zm^oK6c$e^$rIs>Rh5@>tcQ%StBqahYi++h%ox$*!axT~a6a$T#$>C@)L^kRmtR*W& z?4~xzx-T^(SL!1JA=@nLZkds-Y_ikhR5M{A!0{IF>R_VNA@#caVqHC%tN}KxY}b{@ z?kHe@Am0-CF;em3blr%g7p03U`gwEaJ))CC=97-TY_3`){x_4@YRZlE*?gmJgvt2dg|v>7m+UY;6VNru@9pz4q+#!-Y`JHITK z{4%kB{wfx~Xx~UwX(uT~KB<@x&wJt;=|3}S>utkvoeZYEY@TFaBx4Z@70Ml4uJAi{ zqG;ZmR5*x&0o4N%FBaYt@?+QK^Fi|Zs*v!#j1)tS!Z6m%e>Nm&Ys?&ff$2_2c6B=( ztx{qV{{r!#ITOTt6Zi@`Wi#epKedsej3|hl|L45Tot9@B6xp2!p8NW>@Qc8x#w@f7 zaz@N3*P6oy2d+nMs7uGu2sjvQdc;O*GVZ!y*?Ut=q>Y74odRB2#G7}5f)=W^U`yBH z?&d6ACbvs}^DC;_Y!0Gwgd-kR957Ig`i28TIUb<1WE~7+PGSfYK?=fD_8`&$gknvx z3kzlydMo*ZoJ|&8aHbqO`{yJ9Ay8{5rp3f7+@*{N;at1Fk3D{6@8{c4+jl(qp_E6B z%;WcKP7O2Jpj;7F^TnmuA4EpHNdT$AP(JRt4@06!qkK_6*cfO?G8wHH`&{2nNcrP?!C*`2K(z}Z2eYBT}$naOGf3lgt`)~XSj%(#N~b$cVl zY;`0Y+}}$R-T?><+N}F~Jh*>{?aL_w;un?$n2=YJ7($jh*Vb(gp{5sm!|(($+mkp{ z;8wb3BC!NGou+lz6?>kb-vEf1&WtSw113U*4xpV2;60=dKhS3mQj%g7>X^O(_p$hA zJ{3daP|-6KhzQcbsFVZ=esgm)oCGoCN@KtRheNRn&$r*N?B;Wp`GoS1AtFArA3}Qd z+$+3B&dp#-MMsp1N~LJIk11X^3kg(+0tM`edq}F#Uou9l%Yg^i9|ZUK$AHrZ+0j!4 z?_HRRCN}yv@3ir*7>}e`f9d(7`-NWh4#BSBkio{r0)a|C?TLS|jHZLGqqjijqiLt@ zVF>d)mrn{WJ$jX2albyY5L~LV&*fjzq)=g!y7}SXQ4pe+O<*Nd{7Gx2T<8GGHX2tit%W@>Xl976q{LF}oGUWFF3=ufvg9w*$5(*s z9$8U9hB>8N@0pn@09G-!S5gYXj9NDIQU}k)QTsnKytsDq;;J9qop^7r{a*2c`GLYm zJBJ9YkvMux=2k@r^?StlCX#QbcC1973<`f1Tw!6kI*THXy&8L+7H^Ehg-5P(bto#W z%~EC};CHJ}z`#%Wk0$_j*Ot25wj#4`xNOdFTf48b+~KlqM|Nc%Tc-`r{Pv^=T>3yB z9ggEd=|Jf;gAs@XiDzMrZ323S5Hrw!_KB$G(@tYIOhcCq*|OSzIPhR|Ue&37hy^Dh z`m8^;yrqCd{7FtELzVG-!3TTEWmirMRj7rjes>nKDK$6_)fU4D$oY)T5Xlj7R{?7-P{LR_gYq;Zr6Yk_*p?@zCVf%3a$3M@cTNe6%+)=?I7hzH z@{npM?wpxn;%q&IX6G}cj)pCbrn6el8gZ5oJFQ2mOYq)OX00D zNG1U?NyX-TJ{n^xcR+X>EY~>=>{_Bcj=_Mr*u8%oY+z#|!50x4$@Ry68gmllh8VF< zT;XLye>;K&)-q&sv;CNety7lp15|Ry<+5SpHEnv<7AI;$XEzI;l>zX(uPi)OqY}*OKwW)zCLBm*<(1kou){|ojtzamOxt^L4?5pAAqA&u#d52h!RFENU zh@oFXwpI#0D8M8!-nEgzp^aiQcivC`6dG&1z99lsq1m_5c?hrwU!t+yE~-ZR9O}rM zqa{$R*_gc)R~o<@vs%D z;T*34L6^u4EIOgA*#^=mTPV1;AX5lt_~2JgR+)92SD`#$CpydZL}48u2Ewdr{x6Lv zoMe3k_}zV=lR9YfffDX&(erLHADndoW209i|23-Wy>5wOXn|u$;(30Z2-xS7?r0%Z z>yo#GiZZ2i|Glk{L88Pqy-rm+zlm36;NMn)+aBW*nOn{eW(-Nhdh-NFKE&YW|~b6E}jBuSqM+c%?^@MO4TAPVVTEzqgg`R zO{3sqQkE!N@^lBWwiYG>$Tv%Y%=HQ&ikh}lkZ2yN;64!#7?g(eYeP340*Whn_JU*O zY|T&?GM0MV@Pb%3&&qc#Fjj>ir1CarH;J_QW4~z;Y$VJp6!pgmw+&^kvm);n$jAdw zbM5z03i6}$Hp)eNWe0-fe|=#0TiW#|8EC18XYl7AT~zV}vTKw)|GF#X+Fiyt*f<{( zXw)79A^UCAP5-Wv^Prdq-NAZm)^9p>g+KPAOw2 zqP8lTeBHfQk=5&R8P8usV|4K}L$1qc9u!TsdQ;6dfu5iJEEnn!eSjUru13(HdTMb= zyn)=8%9u9#N(@z?_-^-@fhXV5c8sl)o19rs0#CuXoI7%P4A`V@>-4Twk};e4wf$eK)KcP*^VAmG7@O+hOz1>;xm)y_!Ig=| zwzNLSxAlC1p<44L;*h0;tFVqpeeNJha7?!+i67*ts--Z?Q*-3R#dO{?$wZ(i#IbBP zxCI_1OL|}TzgGZYq)laQ^~scXR6x&D}>!}I9^ZHmOe|{?Wp(YEs z-(&yS#iFeyVhZ0eG@5BYL+Xcb>!X@A?_-2z#Yg$!D9&b72~WmYFf)8bYEzR+5k*R! zs=GzVnBgDK9{`2&m9L{)f@_J0E23Qz3Etgy(O*uBi;)O53L;%UIl)b)^x&dIuvI*0yHc|&;fmBIpQBw2fV}zN;Y!zf9KW2xt9pO$58>wN^u)t@spEy|cq1TzbA>d=+L$D1DQ-2# zzg`~UB8~y;TG@=46k@XNIBHwStqgR(h?=klkr4?29<)W>8Y*%6iSq4wRt~Tdk`JeI zJVJHbruOdiT%o;gJYOo!{C5G)%!i;@c;W~@1J(luO2!azy zK&Xqya-mFKZ5L=_4O|qI6cT|C_LkXARPcxL(-1(9m#i7r{;DVmFrTASz80Un5X9v~ z%BPija%N6Ip@rSUCpe($)QeQZ9TIVfSQ3hEYT-MQYEU?+8U&&W z&wH8xtEpe5Rw^^}6O5gcmm@h565b{-B~T?P9;H@5bX`$2bqg?VNOPzmmSQd>Eu0b9 z`3!4d%eNw>S3gw>I#G6j=tAO*Ri;8uf!p_kf#QozyNDe)i+KKWvxp9$k#J@{1zUw9 zv6MwLBr*m@aU{XikN~kxW2%jouhYV@m*!cDBg9Y=4~M@X{0#S8JtXF;YHxhwuF@hd zg8rq?<|C%z%O~zt%Ai=x9>#SWmlB|H#Arqz_-~LcqZqV!LSbTDT_tqyjNnXP8#(8b zv{BqDQzN^W0wtq~Ax>&<0k}RsYoh25gyCuAI^_u)v=FG4SSu!HG@W!+n{%ott8JYy z07T5}kn7IN8A#)062p?Ux>C5CadzSXc9Nv45l9v zpC6?$rJNcnhK1(tYh|?5Ha@U)-jGY#i>FP9Q=&$YT0r6hO-=72C(=LS@+1+l3>qX- z=qoW5>rg=QX|T2dYw?uf_0Rj=L-~N5cqSbvL1gYEbyo3uyA663xTUNHxO+ya`XZ;2 zmeR(n;510}L%3oBt&WDaThx#K6oFd7bs7SDfs`O%&#R=ZLX9m&s1avZF#@w6w6>7s`E}k2TPR$P~f#wvlU8bHYdw2RgfphYYw2bYhE&c*rNE4<-(KmHPjXtcbqNGmiB3Srcblh5VPt8 z-h+K8VPz_Zc2XJ%IJ5zXRIxsUCLW^hfQh`F%>sz+Zh}2>ppw#@5xzBEyR7i4Gyrfw zkH230mmc)#q!C^Qml3?g&LvvJj0Z>c%2F~KHCMf(79j3Gynqxs;#?6a|F-|nl!BD=V zHTyuEH6oH2lvi!a=L1KVC$Rez4-%pIQ4zmQ+ ze&NOJl#tMLc%K+sG_(6wge_}F*zz91L30YVwIXa>rcH41XvK3y70%@ii3P^aD zGv6&s2aE7D7r;S++er9_quM4U%`ML+=SYT}fRMh-t&GKDd2}sRR^Fk?f;IWuo8M7i zQYs=p%*XX%AZ!@~BcUD?L)M>`C78P=v0OmG<~whP)tTkWwTjzbvX9nFwvy=%5n?TQ z7QNVwh&c`P2ya}2q=x8u*Z5k?$z6rn#KCB7GkV`1Vd54L?+goX=TzY38^eZ;q_Qpr zOxo!-v*liMkd0|pt+pP$3RhtwgC8?0zQ-4}vq`_t1|e0ddf1-=Gj=wgPX=1Oo6>!L zr)nrMCei|sJ0OSiIrX@FmCOU>Vd9Y6Z!N_VZxm=UK{3YIO0rUNy-D%X?4ezfi|3 z(uj`|Hm1e#DPp&h<&R_%dav$_wH`HgW8|C|O!9};G zyfU&q&=8&c9;Kki8P2#UbhzKZn7{r*s;uY=xP%37(Wooot3l)v}OMF7Uc9 z8TZgbQcb1ORcd0Z>Yoy;`(a@IQ1gbAMMQ~GOz)p_KHB}qkG>hg4 zx0<3zn#;EqS4;htmI)Lpvx$b}ooDYOd9 za^=?~>uX-y+*-683peZ#68#_K%d7}@>dSBxtPORc2O+TZjC_nlwWQri83eYm9ywhr z#(^7+-dD*WN_co0D!AsYQLZt{jO*V%p0-*|Q-WgbcTm% zX#yT?q--BIZncy2D0aPSEkCikeq3=of~`Xoe~6!*Bjqh<&=nxe(*e;Ry zG91q&@ntN&pJ?ka<3l{;mN+DjWJ4! z9!DWpuiPMog~`x3#>tKkoAxaBng6e6)VBL54Xw$_E}Wz-D|=X~Q?i@}d1|>ho&s4f zCv_-7;B!RfqlQyQDgM=eK5ZKIcY_3tX^?bBz#IH}I95+rT0qZKv(-ph;3sYRx7&Jy zd4#9#2QDMcghS0vD)O!R*?5+}&xY@_@nB^$Lm5_rBsldwOQv%JN#^I9dC>r(=00M>P;GwhqW!RD47TA-Bv-2BRuLnx%dYR-Ji2(agG`GOLvmm6g=9fc#cED5Ri8#JI6CZyDS$F)ar^BIjDQ<6sLt(Hkj70R}YKLrh`o zf1mmP{KMk7i`0#fuk^`*kUJNot?o{w5z->U3g2?AW%0hDIfj!;zlm4bazEsz-R?B* zeHE5)O_O0L&pvsZ8F987z1g_NYbD_OnWOmk8qWWZ-4=~nsFtS6J?mU4X)Opx*zY0B zo>U&vRf)pvOokp*j_C9h;2JNS>m}i8NQa2tEA`)AA=VPyx)sJ!%;#agn2cdyvj*P?BS|SmZP3AiWTeK2psVOP zvRQE<_mu!-gXf)XOc$8U{2hiQD@vBck(3w|A;D6WsfH{Jj2uT}_Xvp*b~0I%z8}3O@5}Jl{{MxldCNe7W8rsr)dR5AOYKQ1M5M*= z+OL?I7clUuR4PYY{754ok-(rbfiqPZy#sfS{hdp_lj%%n0?H%U^A#_Ib;oi4$T+i6 zC^fT+CpiUss&d1nO@UfYV%i<22@{YJl@IOlJREtNiI~)_7pU-utuW8UjBkoV1R4H_ zDq4p=`9k1*Q{*hFUxxi1D_58!33A@JFIOdIQJBX}eiz9X#{5`Mr(=bP*(pu52|<>) zryE9m>z1Y&B1|n)7Q;f|3QWV1txl2|AI!*vlzFYnNqA78cN_M_x~@bnBX?@L)k31F>e)Kn2cKRpW%@>ac)=UfBNvlmHHJs zu`Mm_ejHnLta3WY zYt4SPZPbxdicfE9nPiWu+y)%^3VuB& zm4&@560i{cUd6)!LSV~o4TIE*la2jd zA#CL|Y?ZgeR)zmRNB_**VJkWq3Mo}w3FM&Vz&C+1C-&HqH5snfEGqR=`B8@)@+cm3 zLW6ZBiL1`icAQBq3kbKBbwbtA3FY2!0#uN6uVPT{akZU$dAq|5?A2yWe=o%T+ZGQ$raL(9t zXCsh8EZPV-D=o(TR4x(LQTJgm7)c#J9vM}P=vL~u7<1|}`OyJ}(eZi|CmgQUp>yKC zBGR`^EbWO{+P~jz-(Rq{NqnJZQCKSQa{VY=KUg0rBa8?Kmgho{*L8`nT$hG08D|Ln zt_U7FE?8tFaUn*Z(hS>_j>gnK=@_4#97OsAt``D0Wj#)I*6(w7o}QMt)r#t zbBZLFFUnXQNX{52in>kxTQVM^;9H6D4{?psb_F7twkERyuiNs1vhT`62&zl2I;8gQ zPsWrH6cTep>JL1o7d%NKb%Ft&RDhiG)?l;AtcbpLi3ZwpB}i9B*1s~&t%HV}>TPpqk@q2z z5LIe47QXUFbt`dAANpt{2C3C=lX%eX&$B6n1N}gDMDwX(xX>c_>{nvNauRM;J{uF1 zT(+=zwmbY%X+uJ}9$~pP?GWif7NMC^$NqdVEnZpCP-s<1>?tNgb~?jpzw*Cpxavs~ z)7Jq3T?`Y8Gktw1#3Nt1X1|`_`bDncIWj@p!7`1n>T&~H!l^kjS-TVe{Ij^UgWxSW zaSi>C*_p{Hqa}A*E_JSOghj0sSL-&zmB7>=wSLShF<&JBtQ+NGXb~2<*H8tw0#|$i zRwW!T3sa=e1VXIz^j(;8IrK*gI`NVB=WX-ER3U4Ya>l)!&k6nSLTq_x?3%n~*Tf(# z>)RZXRaKb^Cz8OofWxE|7uNy~LXF5uPuwyR4ewBkL@@uVl0cUT+ocj!=fHBJWyPuC z=pdJ~;V^&z#cR<2AjI?M!fPEb%xGxMU-HA}Q%1xFZ%GPXVg=HLkd+l9C5OP$lJ3R#VNZ5xwQ5=NAgpyp=NS6U?T@auuqH8uGteoI488fk3Xu6EM6sFu{> zJ{hbLQUvT^EzP+F%4Bh!k3~m}s!;-h2sjPiLPs_Qb(kv0*;H#eRUYKLr8`QlAX%R< zPU9J<2VxG2;9^>FpV4qmSpJy?aQw*3#Jc_sf;_})MIC=7NsA(+}Zj$h`eC zrF!xtdjDlGV`?PUgWxUA1!(gxgH>!YoY1Uc{tF*}V&d+5cMCr>Sg(^Ie z`H27gHSL{IHWgtTS2$*qGKNZ;+vwX=MsA*U`umKrBxQ2p4Z?|lM_Sn%dH zPf$5^aH5)T2KG~_V1?QylW|hrE;qnQQz3p|l?N0LU@nbt`YC{;>6pMy2;UpVpNK5x zql(doA1?vc{d`SPM~;sy;!zjNtWs5|u^6tXx~{F}3T_gkx~?oXEQI>0;rUpS02*Ml zLPVkin}ok&Gy`Qnefwh~)>qaXAu@q~gC~dc0k+Z<_4(ETQfejssw43qoe}vQaUmbx zI?`cDv}Qa+X}I6*_En@Jyka+w9|}R-ZK`D*$h5?sGW0(SYl~KD=8MI|>|m(p8=%2O zIFxCMS|4c+%pgXT_+AqDxi|LLqa?~MOxDthQ!JaazUxWYimbFSw5Umudg}n;sQUSg zvCxMdgUPC>!Hn~9TZ36!s{qRd$TIOwvTh1Zl%ub<4qH5x!;C0AHr1+IOif?~0|uNW zUPK~`H9e5p$ehrpQizyQo4Vn;&T+{lX!@epXm6MabL6gv`TQ>fy|OzP?Qdm@R54aaR{J->ktR~ z;9JQBHZ(aObcrfv{b|CP2sTV8828{2Y+RebP~$k2XfeCZt_Jr}Hr1Zygdo?1C`mSN zo;R_`#xcNSIM4qv1@NtcD0Q^)YJ-Pc_4B#l6nb1HqgxP{ZHb*$4-`jOv^{nwBvIN>aRc1FofMCTLv89>dUOEd z(vh;~s?s)ss$mof^vFR6v@uZwYzX=VKO@t59jfsBN_3)7B1lVXm|@MpCdsZUMX2zO zi$lxkBt^5U_8xHg?5A+-Bop6UH`v+OfGj)jIpwK}PE+B7(83xVeH&}^O^0WF2I^HK zwW%tXs>XFw1fd!Shs&bv$X&B*RQ$LvaKr$WY-8{D;?L!}8UDOkoDQsUq`qCN)Htuz zj2P2@*fHX!&pGFzWHYo#QC$qUM^wIUn#UoGN-fAA7?W|vuqA~F-bvAn?QTnS+#7Pw z3nX_Cnfizdkezy!Ze6lkX>1o~TaZgIu;$81x>&<_DyZ4`&M8NOFIrZ^VNWV!l)`CM zkpZ*BK7iy7fgFab_o)YP8`F4JA>mhgK#HCQzcG9kC$g3u;??3*%54Jm?PR2s*IMr) zN^jopbyA)35Qw3HQWKIAMj>f=)P`MK22&8rq5F{$b|jRj=s+|+uoZ%e#b=bCWF63p zgV_DZa8S^}83E+DgoX`n9j*f2h^HzhAOlY#h`yoY`V)k+!P~#|dXo zcPu7ecqQ{gv^$xL(2b6>Kwa_J*?6I9h3>ycY`>rd1s@yNE!sf>B}UL$Lp2m$?<6g) zEMO*{W`)Hu657II}Ko{iL)mJW0M#MvjfYGw+HNL< zfNCf-3Y(YdISy-`fOh3pp_K_JuazDAP!D^1dWuMNLibN4t#dRM$2y%|FI6mbOgwn_ zZKt{zxUBsC^kaz#5_&(Hum(V4l|iy+ehf@E&tlLoLVaD4?_Tq0B=)Pka%6GuH35VT zT`p5H3v@MD#w$NoGDGX#i??|xMVCDFSnx-}9qy2~)d*47`Vut{Gc6;fs5!EGkMjjr zEabT8ch&JbE9=QoMjSw2;QK~g`G5l#b54Ot{hH%Ynt5h-#ko}T>4j8t<-cCx*JK;a z*xQ4c616B!O(7u=+Zt*xmqWN`d@0r|6O>^9C~74MuS!{fp1%e@szWgPp!9K-jWesf zQ{OeMk|K!|QXWr}sM%$u32aSJAS4HHl#juITgi7t@ds0;)C9&dUP5tT0Im%8%A|}3 zZ`#VmOo8!iEP?j9I1_)T&WCYGXRS%iEk}lU&B(!?;EQ99H<_&Hv|6qHID)$t z9WGS|X_t%A^3DTEa&4H9#mV8J9CJz04;&X1iz%s``e_4zz-XRRwd5ADv}Dr4vc(aJ zQp2kloN6q++%jQ1{@7tV1~5RU{{uJOG@@eDq*3sa1c9aK4MhP)lKaQv$-03ml@qZ! z$YK=qittPX7EjF(Ee9Q1zg1f8s62AUcZn08B?1?{$R=C}izp>3QFy6igzlo*SO&PV z1x|#-Qg&tC-?vr@v|ixQuGpQ0>5AFnN=1&TX`J_QWSJcX&@wc zau~@JV`k&g&^o~*8v}ahm#0{dYMe&~q{Q78w4#`fV4cXWVLL|LkLvMXbog@#zMUgn zX83Cl&m~jBQsTfsYcVu1UI(khDQ~mv`NiI3=2#jNOdcD}Vv~UCs=s($Y$~-D@W+AK z)b5#umuQDZQyig3U0j1Yv2SS}=j;fh4RYEu;VHk2l~(}m1MH*4(OW^Oi)aqQ3OxxX zXBkYt9n(=*xs}98M(sGs*#DoSe~vnJJz$%_9kMCeQ@V+iOE`XTRmpglQ8uq-z+^3P zAGNliA660phYAGpeIatviGN;~oGd~M1o1}#aND<1SR~ZlM&L$!8lqH;3ymP#N8vFb zF}jlaUn{5B=#|O|32xc6Xfl7L1VQi+5$uoL%h9bDtCFeZ7P}bJX4A5XA&cEO3mAKw zZkm=!tK}V{aqOaFs};Qxn0M?CR7njJVh3^-SSPSV_Gw@{Xq4@sXHiNHZb&dilyC_? zQiN}h4iZEh`w|7xJAGu&N>g0xyz1mu< zFz6ay;Nnq|TgFE^&rVn|PCdjc4U~?B%ahy+2oa?(T_rDNAfICNwIOwQhWu82Kf$1!!uN}LQ=%Xd8MWxcw1fgCZ*0yC#*|(`7;XY z$kHCm>kcd(g$)RQ+-Op8a*7AzRw;;Z*d&GAwRTEON980A=|kXiz*skQz25Im&wlrV zN1t`pGoVc|(eN@FQCYZXjAK5bsHCat6$R(oSmOpn=?&q7A_KVY*3A$9jx&pFS`AaN zV}b&p2+*KcoeIVn1@`1h4aMt7A#_`6Msrw6A;r3#{RtqR78Oqtv`Zk71>YYd>e#w3 z`y=n#d#^f33_;3t`hAax{yW<873o??vv8rnhZq{$>a#)z6UJW?qFeK93DzeFU$gtf z@kfxnP&_fIZ}w>FZiYHtN-AWtInGs%yozk^iwo~>zNPS#0r%E|(_xr6K*!2!?T1S0 zdvFTUmXBo!>WgJIQAD=UTuEGn?}0RvBpOfrNjNMyi)`H~ z&F0%1D2Ev2h_qWyx<8kybP(+IpSN%@6GQjgAxwNE?Z_M}vsEW&dTo5= zp~lBCs5r_AH3*rYF7Uxnu(9J_4pr2HKc;T;F+l~gCE^SMsEN`dQWN2v(>gc`y5Qu> zi;OGE3d1rx!iJ))l4t_4`UZR0eikvpSZlw&nE%Q~KUN{f-V5P|WU18u3n2?j;;Xej z*`PU(SWv=ilwK{l98Cl!Fwqck`o{v7%x68ItcGoXt{2020TM#1C6Je48nlq$3;l$w zJ?j==mhAcPzx#iT!jay{FIMf<>esap6f4h=|HMqVP<(UpnRR*7#W4$v-;hIaJXF|6 zR;Y&sTQr{87YKrth@`e-Kn+)KqZ(g0A;b+4hiBdA{sl@R!MieHUd_n>Ei2W)M7MmH z8iH-4{Tt2+AY+(2P4rMdos>v>eA)e){b_iMd$5V2_FE7}7}}K|vryQBU8z!iwdMlc zeg;L`-20Uy=e#bSRkCWpZsD)RQ5dRQC8Nz@Mh5{9e2w)PJtidvDdjN9(W*r~xaeG{ zOpC=fB%L-RS>Pv`T5*m4BpeY~83Ui9DW=V9O^S3u`qp9}WJNODIpbHP_slYWP?jUs zZK}^`OCAY{_^3kUXFNoq%S)=U&y6UTJkdvRF?S|ObKs@ z%s-D`QpzG)w(!HtnG;||bk>9aAe=nDKaG02bvn&#=3d#IfzJ%FgL7+@rTu7T)@mBw zMN2vri$g&VvNY&LY;1UeMh9k$2?p0erDFXOQ#n1z)iTU7zXF-*D=nnrJw>XJwGx)O zJ=!_junAo9p?F8!Ua)LlM58;d;a~!hE1959IM-zl&cZ$j&4Jd9m0eJSCN%rByj`-q zsU0~@0|(Z~D*?cmED{wCV<(7SF8hH=1pgc@y`QK+vYamKGewqJA%=4D67ouvycUl= z2l!f;nBIeIr&uy;AO%ZE9@%oxISUht3@1gX0dkOnAbzVskcD|!CG*0CWToVo76%!` zEHa{s5G>FuyY6Eo2EM}tQ{I!>Kn!6Hix>uy@H0i(w`TtzRS6+-USXd($k(8FH2Xj( zRkThx9Z)Qc)p-TDMjcl!8MIVAvQrc&GDPd52{l6|w6MFPal~tA$+m6rUOmgyl#o(2O}{O?``VPWJwTrs7)oob~syQV^ZCXS`vq zQg3k}n+d+_X)i$3qu~ON0a4AcjU=^31jb{G;kuRBp13E)jZ>)-7yx_JrPTyAI4c2{TAe5Y@xl@bv;(08*qU71 zBJFKaaV;7WbZs`sy48OGF9Hg8$==h%f7_sdaRXms5x|^}i7K1AI-jb2ZOE*d9-eB} zqZkOWSX4`Powt3(O=tSl@r67AF9bHbg8_?$sx;Sw3DO2p@Bk>^b|uJA<-^*Xk6$Q2 zeBjd~I5|?FyU|AZ9Dq()1|qVjgmRSGv|Shr_rKo#>4+r3g;qn z&$jBS{6qoqkSJtxta$Y!GJMw{S|bWH#K>}|S_L;k2(6|dGYATRByv$}x1>D2l)?#7 zosYEx)J~HxrOkLe9Gi;#Lp-bT_sKxthdfFoHDN;w_$&+$E6eq(`0;;ADf@(3@`^K^ z@-}BgT9i}HJ2PE49EYO0*@g?~L<6H8#v=wqY288gR~z0WiWmh(xQBGm%D}8+8$ue% zur)=aFh%1b3e8uao5S({dt4jUXQcin1en`hy9~qm@Pq8F6$5U*zfr{eI3r$$;O*u! zB^IYFRf+gSxs$XLUYKBn6n?W9wO(JWGG-;@F;>9wl4C0Qhm)XguUOrdx(O>Tt&)$K zMe9Fg8NqV&lHp2J8<$jA40m)5StapbrG_l4PnARYQmO%~l3Z)TBG+UDaATr7))JJD z>xa>_vHWE)Q^i@l3T8zs*Mwx!_`R7&Jl-j+xI!S31+uBRHgw|=@?1l9EjN$4DU^gi z3OO)l?Lu`$=;>(?RQFpdVL5?`utowQx#DMmp5Y!=JcaA%1FGYY5os2ZT`5jtvAAFuY3VAm6v?|^hc&m7+)iC{ReQ1V!u5^4=Is}AuyskL9G;~{VUuMA zWQwf*I4l;(LuW{gyWi;*(rQK54 z@W3*tDqDGkdYzc(!hpS6pDZvs{2}srrm!bTeKDZDS{w~3;@Qh$U}j8i7JUgtTDg@7 zGB{p(ei*OV-SR@Z?n?+DiQrqC%33Zu7ucFm<&rsQ)v~VpHeF3(@hh$!$dTLflb#_U zH=^dsV!*-z{}yAUiXma7e6N3c9vnF(j9qZv{wH3?F87u~Og``;Q_+{AoPdN0+>sF$ z5i?G>+9r-Xwq}r@&7?8-axCfsc9(J!!tc+qzPM6xaSV0guzpD z$)yaJNxo;)OWBzyWk*8QEE|h;vcK7)3^ZKkmLefETube^1@%fe{8Jm_ zOgk);inJ;Jnp+y(iaAz=b4%;-V&TbnX;@|usAQ@=!^zgp8sL)q)D9_|LeDLD{@Svx zL<;hNDrvZcsuFIobA{VP{=YUEi^+ns#PMOhG>{2PCYGZ9cx!gkz)CZ@dEP!bP}N>G z8a_PBD57TD5A5!vK;1k9!Z;g7vLdv8o+Qq-5o7~S$R^ZX1#Y#@iq%D}9bNNDVt9|j zFYN?vVcTMwvMWhZQJ)k)^W2-v3x^xkiky49kZ^q6vI(+3)s|lcI{&j}w&s8{az&gj% z?LZ)}5Y8ohMuZ-2OG)O_u0UbUKLlXKd&KlW(L$P&<_(53t8-d&u2`gNUE0IgY$1@K z#++7#%^psuvjR#N%$q_afQQw*vnj|PI#oU=0?#=J=WNFT3EcKNDEj#CC#hI39t73UY3gCn1NE9NylCyGgd>*PA5BVw`hLvjcNXG8Xw!JG^hpK(lp zd&wA&Qc_L1qSpisc?nosKMQ_+Imvsq!hdXq|9h7ENmza3Uju8(Zu%<(%8F??qO-X5 zn7w2uuW)hmvB7|Cb9jz>a3ShMAuJwYh`inB%@TYj8Kd94dpc}*XT1vrC^(+Xf{>M) z5}{bX92`3aKzEjbF#_V^BmB7fCR?hSc9H{rnNV*Sq7bUxl%Z}o8J1f4YQyz;v|1UL zLuUSb!DuGpW!y{rsH=z$raU*8lO{wdhd7>6X@`0_6HZ9;RzxON1ZZ0VXG^nWTOwSb z9XbyA)n%en&JEa(GKcc{8T!y4IYm$QK2agf-9w79V)b@kg##g!-aBfXfEE5arNv|3 zhm7IMpt?wPXrZMEOe@O4hpk~awnIT_O>>k#p{~<0BtDkt=YJ@dap55MiSqysyVa4| zLY^>mO{9@X3;-tP8FU%XyGoC|0gmPC z9$AzuCt{r*Tip)6CWy za`~B?oj~so;pB#gj9a-#7Ir>OxBM07ZulAAw7yx4=2`gW+`Y(lpePR8n;5e~VSBLs z7ss8@2viBtiXtjqta9EO-l{viOvdri_9Y#0EbbQam9aO2m1f+Jrm#f?Vhz=h;EdrZ zS-#k@T$Q~w>&B@dj!~XcEqb%IW0vP$&f}9bPmY&Yt-;4Sxph7q##?1d>%_`b#O4Pr zx;dI>Qp`~hI-Hnn#6~108E6@WWQ|T#tc83W1@gv{cBYjspZMj0&q^yrCZc$^sCQLD zS+cf-H*P$H<=U9U*JQvZ*{p|qbPz}fa}1bXr8{64@0`nUMXIpyO#%J>Ryd8hZS{F)*I~F1N$agx-$*u%$yQsDAi&{+>{CDB^JpPgak^drkiBg z3oktI^<>zUAtzl&1vk{0AzbN_OUAGmd!y05mB!G|Q9f<+LyKQQgZWS8Q4de?=f-ZV zrLH=I^4ijEEZMgR_RvUZ&9rl7CJCZ6D)~gbURbDPl~7+EUP&o;<6=Hyu<`LVsH4!a zj4_*9)*KH^;)8lhBi<8gsn_RY+w9Nh%A8^&!L8nIzL*UnM!yuMw`(Q zuoT!$G6E-lBPH>ussBGm|D3cl0;SDuIx*Yquur+1ybRVV7CW=bFI#3i4(|xl4=t^g zEuig~3e*T}$|hh|d{LGBp3P@7v{O8i(~)b<>re{8*zF6!)aWcfhVcQezj1+D-kt); zPd@FnVzsRYYW3VWr@1EGv(0))A^^~F6Tz#=gS>ed14Juk4#6n;RKN%KW}v2qI1u)s?O}?JTRN4KW3I zn>ie5=lGz{fQ)qKIg?r$_BF1pisk1`YlV`G&J z9NY5|5zocN_to*V$GFCN$%@cRkLB=*q2pXZfUouvjOEQGV`U<>&%L?7*+1Y#qF13? zTc?m{oE&tq>WA8fI@Wyr`lZ*(bY07&tU8J_wNW$`(VHEz3ptz3Z%qI?NP6B60)uWh z7I;0FO@rW8xHC!vps_Oo#djO27|f>3?pV$V!8sd=LFd??;7N)7@= zQbre5XIPa`@&h}$R@o-DiDx&TjqT05)e89zsaiW&C7H;@h@<@UeXshAeb4w6)2uyV zkExhj7*~QMCGvcAGq_kivpokKUh?u^;WC$+m!=#rN&G4qtg@z@dPvg12Y@>EltBC} zWU_udB_|5mh@I%KhR?SQjnS$`LD!`;AUtYRx^uc;HYp_I(I|AGgnMe3c=>Mc6s^^D z6lk^^$b!ZnMzOXs=Qf z9K-!UbnjQj!$)F~MCaebZJTFgl$YB2`YI-h;*8Jv2xFdN7$@ z5pfA`B2I4-&Vk#bsf4eMxhq+n)21*UgE|(_u@qB-yNk3ERNtbkOXLPt;s!5VFwxit zWI_pK1Lo&#K-g|j#DfvY{(0a~#C)rT59WgTmb-^@HruFevswO5vDv8A1t-7fMKKY- zeXa~i;>jF~`C#4k?m;PtRtD`eH@4KU9TprH?90^G!WK^*++53}5`%S>dg<%nrFUNd z?Tf$cpk)9Ct8gr@&Cx0mLa48z?4JcLfxo1uwC?VK(fQm7I;@5h4mOHjB$4zEP$ejk zYZh0DK{hj*rlIy1mUxaFGJ!*((_Kp}z__3VE`NV7&3l&=Kqy?|4^s@W)ePV|iY(S4 z7slwwl{4u%97cpfO@9_z&d-<~AO85`6?A6d$OsxyR?MdwZJdtIoAm~ma&P9a$Y{}G z6)LMpIw{aqZ6x!P$aad{xmTkc$B3#b$9kA1qL<>GdeFC9LQ0-QgfK0a3XeH1^(Uju z=VlE5MK+d5&&#l)wP3!3_uyP{-RS#BM#AL zxg%1DhAhmu)ZeB%o4hWN|VT1 zb=|L2c|U`yJv4hsJ7~F*mM#UnHX)=rVpuAZ(v|ytAa*iQW=5tmM|!dkg~(FI+#i`_ zGyHV2iHC5=#IcgJOs{hCInPu>hPaQhn-$ZBN}LgpTY%up<)a@UMeu0B2*F&yG+xlx zHd-L1Ltv+#N>Vlj5^9v+WQ@cgC*Cc`p+dh8b|^YOwb1~K$mc5f4+FAm!~k|yDW9}F z2qV4_|JG6@tox+T1-{OZTC>7eD3gZG41r|gO#xT;NJ?ZSZ>F+Uh#G>9mctUeLEUYW zmh`GydH(MMJVa~f5T+HD!+YY`!hbNVhZc%z_pc!_hvM-KPQXg(LMw)@Fr4Hd!9Z`# zJcu@->y~|I1`ItavX#*gNt}6xhXPNs6hu!h1n`*P)ZR?V!N_nsWJ%B7UA zDIX&vO%s4&t~r|_(c=hRV5a}*_a*|`kRMpryrZ1-eW|qt7@B$1BM+)|CG}|nFvou| zS+JDkMT7BZsnIy#7P)}f@H1Oq=p$&|WIfH0ilbnAtGVj&z9XDY+>{Jd9Wqhui6RqwXxTsL>Niux#2+ z-kSOUbM((?J6Lwwii;?88H3O|2M!+_ZYG$7;$bCYXLSKX?GpPSP_aG6q551cPCO5~ zgEz2jv7aP|%N&U638&p46}9JLS$J{*f@Nw z;)nvipkb+Y)HinzfwXF+tbKJGJ7T&nPHO=6BPA`I8o9W=+GyOBqH%+YK_0I6*x#Ut zjJQh!J=S1QE$7I>mMRV^v0Yn+&AyQhwu%C84fN=MojA*-Yff&*LQgmMsZC(1nri2c~uh%LqVdN9{e88q^wH+9k| zTiw!ec@xS7de1Q<8WQZ#5ob0g^+Z$pi5X|dPzpALU2?^O=Tr|V-I>G8))1t`q=`a#UkTcP=c>G zqsb~gVq=Grww=U$FsciokB8st#g6w%Ki?W>9-G16yy&6wuk&a~GQ_s!s+9r-`I#wh zEq{sdAVANq@F0bnR0{p(oeaji!AHlH0ECv;G#fKWu<}|}CVd7{8fYR{~qlVO+_x^rjV1e4` zOkC=HOk%R}DW<@Z0RZg%Nr*~Axb8s3yV-3_>StRZ1Le;2JchHh>Llfbak;CcDV#25XVbU*+igg&2;iSIV@3a91;}(4M8Kt_l2r9@&vg*$F0niKJ3O zQuhB!acC~B&MygI51ukn0{x!JdXny;8J0OM)+hL*z%v zh}FUn5)X4h3k>QU5#OW=UZ?nnvxWYQqlil)BdJ0`WQkX|1m!d zX;j3TjSP0u1T)_pHk3d^(!zL?Q+;Spbz#$mj}|C6vZWS^H<7G%osV!r{Q2Og z>p%IQe|zHiTcAXLkLqGDDFZwD!d4wQrHD=&-0=$Dc*3!iG`bV=7FJO{X9TCCd)=GS zo$p8@X?;h|g|39#fk^G3S89v|8Lpab#AWI?5?kdW^?aBn$oWtylIm8nM%YMI1)SGv z?^pqaMG;0CJP%{qx&+43aI6iASp7&tB4%{Rp6deO{2?= ze6I`14mRE40J+&If$Wb+ty6#+cUTpkkrBVM@P%@iM3x{u9h)SAun@@(b zp58lkZ$u?IZX>neDWOh^`X@;O4?Fs=_$= zT80HasSBwf0&}!2T}UBgfuxT z%}4`k#%i-GTVrj@XnW;L6DlZSk7@`DECy8u?UFTRZOs4~I+_;Hp$>(gw@VfcRx+DY zIYh{w}-Y7@q_OiejeFcu#JmxU}8JS56M-1{{%UytS}_pmxqgpd`(>kdQPtnK8V`~P$F z&spl^f9!t{H^+y;rXN=Y?UmkS>P4tm)8Q;2q~f$U>ScN6|D5H1XJi~E_&@`!E}7>Z zQP}ufz1}eLtMR4(6JtN_`zR3xHdegqFC@$S?>`s%ke777~v7jNtP>%f|bL0GRfulmk%-=^U;Aq5c z%jViRdI90oXq0|b&MC=Gh7?XuK>;(i55aPC^!8sFhZyBrvZzm|x(P*L*vqDsBdRSIo|6!^vXsWQ>Zi%aJs$Du&N44sN^!^n$}6J4-$Alq zG-)kvd6ZZ8he!{TkXR42dfe#<*6now^;fzC6ETJMF5T zg;wz|enGf%3!+U$b9OZ-i}|+V^D`K+B`=7q7?S~)I!rNnt+l4ZEv+RWKD=$zw~;pn z%1#sxotwa`7Q{(WgwFBfyxRW+;_gHA!LO{}T>$cq5%LM?pL!J$*OE2v)ucHSnWG_^ z!&W&1`O^8~Mo@32<4JO&OuY&&ecFb~r7R+GukV!CcX?ap^?PdT`1yiZ%d#ntQ$Cz!(waZ~>E}Ov_`M(0(P+%qE%_dAd1)!0 z*OieCv)&8kFJ-y5RNl`m4H+f$JoGD$&p+>Z8Vw6#T0bpRn_Qhq$kR^IryO@GFSJwn zP&(ypEV~WtGgKG2X50}*UUb?l!~V*gO_QVwt^p~R*a_FiqS|~!N#D==W@|1WRk*;rqQE8 z9~;Yhq*<0x<5%-6|N9R5LWe1msqCKDh+c!zUOm1pxO}zI{61>*)Nd4XQsm=_G#w=a zSd*k{XNwF->+L)8A=0%Wmsat-<(pKO4A_f=v_%jRY?Ux3%BE&Lrg)nrSkW zrNU}hz{NO&TlZTvpmi9yO$BQG3}xdn@H2PyZv!ZM-#XL zCTTeLdAHHsg=ag~PBsi}=XZ14e_K(G`SV(8LJ&R(9<4HNw5Gq`9mQY!d$hr=`hSEd zi!JAO7N{eXNz3z@*gemjYFUY7l3RiTOFRx@fblQfv>%#%wTR7F+oVkw>j^Vtf*p!S zbfe-`p-q#o4N2?E95-z=)TZtIgzG9n!v_+G9vq zR_njTW?&fV5-GMKe>&^Vu9%^g*)uzNxITM@w5XtUQzSI5WZxFnNJd_l<~bP+hg0pA za#zB$z;}dh6=D0!FXt^^A^Tbx^6OBaQ#^!VY-}wW@xCo9u?j|6K9%6}wo)b5$(O19 zqoG)jqkJ@Lx1ZERTqgnNKD7uq(K~uxCl_j>_EG#N{&#oLMBz)flFF^P zAB>Ckv7qu_edP7@TXbkaK0qG|~i zrpRfEHu6){#>q%*%F$#fHsx{mOcjDu4#_Q$Lr}zn@g=|T$2D~aD;ZrWN03}Yy9%F zMuaNOM{$Cc9e0m{rhA{XEQMqN3*dQV;s$vmhtc7(d?{T_f7PBD-S*ixt*9D zUda7?JZP}slIk0FqzXiN1u=*XhuXIJRN58^O$uCfc1*d57Bw4>cJ%dNKRS7}=JYRCXhDCbb1e5oI{w(Ll`)|$9G#*!z#y5JP)1&Tk01#QNkUw*C&yjnLhCv6sm zHaYCq*v0fQUhQKNUe=Xd6wNR|NOg9}u!gAgxRVf@h zEd!rWT2pYL&bILg6&((gJ$X3BnrAKBU>(HNI#C;?ZPd+aK4Z0U?Eh0bxDUPazmkIB zw>$SPs1EX6;#H{*G1P55tdV2O<`)jMQk0AF2kzPg$!nUwuPx@5(C>6ph~UH+QORC> z8z9Pw!NKL^S84}{z7IH9VlCDIqtA_V3J z;hPy-wTu2q+a+P0TsGhD=_I1o@b%FUs%tdwy#N5{2EWp`IsQUl4dEb`QTVFFiaKAF zs7>dqd|kd8ZEE?de9M28$zAEIAcikKlhV*M7iiv^{FZCtjl+O3#RizS`Z>a2R}66 z9M$}~DiKm>8$TwxWNe$`>)oN?>(VwFdTs|UJ8B2Vb*LQ_2MX;#P#dtWZX+B&LjqMA zk>{o~{F<)4=%DevSt0J1FlZ9-8nR5~*rABkN)U60xKyN-w1GLcLk$j*r%wnv@rK3E zZpL~jwJjUZM!h-XJ*s{lHn2pR=GjnRgx1OC;T)|mem0Qi~T8r|vG9WCIaIowmM0=D*hin_EB}xO1rK4b~@IL(n z+*Rh>qHxrRTvM%=553@*42sKiwjGSmg<@qc zpACge^yg5#DJ}8@OP+_F*h(u<3YZ8`voXA)KT19X=vHX_lg~NICP6H!^H~fhHUi`=wPHwWrFMc8h%x1|(0yK3j=O3=KAnW$@tI;crRAP{ zhG_KD&rmqIWunuU9B%zfbY5vglIq(VNdy9|d{vUTLo4`M{v;~{{IQ^Vpe_@w1}^TQ zz0-TewSmJ{c&ktcgl1+F`s(j?`W2?W=dT1+mKYB&8&WVrD;B?js%;B`D|K@2ib41fU^ksD1TXt18u!6#en^Oj1iw@9eNJojP1m zB3dz@l7nj90^ME@rr5bslBnJ9ZB;FWbU7){76?{4d^%o|GIMf7^hMezgNBtVV{jpY zJ!Y2HwBw>Xblny0;%v&}8t{t=^sS!{TTIY7g^C^hbJ#X(Zbc~K4BqsmkQcZEtze{N zk#r8gqR9Y85sd2~7<)!Aj_H*7Ak-lkc%+9RFc1nlt{9E*T<#6oN`m9PqHJ!AID{*L zb|FJ1dn?vskx8+*gR8?3C;c^h>yjsoB7a}&aXv`}x)RAi(#0d*H5939Ir z8;uLOVM6LN6hsnRL{h}p$=~|Ifvd>;Np!UNRPH!yQ%)yC{n5sd6)$c zh;RFp!ifFsdMGWXpZ&=(n=^@$)Ekr5@{iL@JW@~ETE0H^3)}aCE>Z*iMRK}K2w3@V zs%}al6yT%+PA3#>3l$Wpn#9;!7qiJ&^4f8dbdD_UD}SDf`eqcda=u%{`D?qyzr=nq z(G6;^6!?+c9+O$CU-A@p><8_bTrDO;M?+69TJj?2^vt%%XRU|&;~yQ4-K)UqFZ4!U zPuPa#)0vopw35ouIevJYr=7!9=Q~F$FTRhP z#*a|b3XXrx$aDyea%dF)6dD!vQEe69FrJ8Z!YVYau_;h(2rIKRrKVNoA#4jg8gw9f zLI1z&hLG15vxz`3^rPH#{uPyp3SDDulE;zV}#_mCziQu37pV;@CT?{ z^Vx`*(6d=SOFQI$vhzWPNbHl7pC^+^yHETBS^Kt6xav%wPz~wcu-n+Sf?i^yotlqk zI4FE)O(*y}89}cQ-U~-UKJ@F$b_F+>u8=85tZ8Pq@-1#O5%|;puHBJm#}K8VJ91UJ z!?~i>ld`dyT`MVExZ^s!*-S}7! zGLU5t{2HdtOyT^XyoMo-6MRP5EX(=J%tY!hjE9|j$-k88*4Jibqe9wDPsZWlpLrtx z6aco&%#(IuCHv1SGB{82Q4TEY8kC!lUX`sm#!t~48asY9rFpY+`+?Myc~Db+jijvL z+880a60G&aTudSQs9;{%AiL;-Z9LmLLQFUu5^*N5{H1Pr*TF;tC&oC--!{&a+mMyr z)HZXV=zh~UJLgrqhVOw1BOh8y*sp~(zT_iI zp&yP<3opy>cdVBu5k#>{Yg0r}F-@(mV=5>-3yM7f?F$_y-Vlk1b-2lyTAVSAi7Plw zTIPy>+th}9b4uCfbjnU*?BnoM$A~3p*tCfDvHL~#aq?#O%k5)7mLo8Zbj{<5n8zV! zxYK>Lk^zRP-m;Qo@o;M;H#Tv*Z6%M{N=E4b^0Nq;i{x4Q + + + +GDT 1.1 netCDF conventions + + + + + + +

+

GDT netCDF conventions for climate data

+

Version 1.1

+

Jonathan Gregory (1), Bob Drach (2) and Simon +Tett (1) +
+(1) Hadley Centre, UK Met Office; +(2) PCMDI, LLNL

+

12th June 1998

+

+

1 Purposes

+This standard defines a set of conventions adopted in order to promote the +interchange and sharing of files created with the netCDF Application +Programmer Interface (API). This standard is based upon version 2.4 of +netCDF. Documentation of the netCDF API may be found in the ``NetCDF +Users' Guide'', Version 2.4, February 1996, available from +http://www.unidata.uca +r.edu/packages/netcdf/ +or via anonymous ftp at +ftp.unidata.ucar.edu. +

+This standard is intended for use with climate data, +and was designed with data generated by GCMs particularly in mind. +We recognise that there are limits to what a standard can practically cover; +we restrict ourselves to issues which we believe to be of common and +frequent concern in the design of climate metadata. +Although this is specifically a netCDF standard, +we feel that most of the ideas are of wider application. +Our main purpose is to propose a clear, adequate and flexible definition +of the metadata needed for climate data. +The metadata objects could be contained in file formats other than netCDF. +Interconversion of the metadata between files of different formats +will be facilitated if they are based on similar ideas. +

+This standard is mostly additional to the conventions sponsored by COARDS +(ftp://ftp.unidata.ucar.edu/pub/netcdf/Conventions/COARDS). +In the following standard, parts which are identical to +or paraphrases of the COARDS +standard are prefixed with COARDS. +Material introduced by us in our original proposal of 10th June 1997 +is prefixed with OGDT +and material newly introduced in this version with NEW. +Material from the Unidata netCDF users' guide is marked UNIDATA. +All Unidata recommendations are supported here +unless noted to the contrary. +Comments indicate the places where there are differences +between the standards. +Comments given in emphasised type +and CDL examples given in slanted typewriter type +are not part of the standard. +Note that examples typically show only the details relevant to the +point under discussion, +and hence may be incomplete with respect to the provisions +of the complete standard. +

+NEW: Successful transmission of data +depends upon the receiver of the file having +software that will correctly it. For this reason, a +strategy that is as conservative as possible in the use of attributes +and encoding techniques will best promote portability of data. +

+COARDS: This standard also refers to the +udunits standard +supported by Unidata. +The udunits package is available via anonymous +ftp at ftp.unidata.ucar.edu. +OGDT: See section 11 +for details of how the package +is used by this convention to define units for physical quantities. +

+Useful comments from John Sheldon, Jan Polcher, Harvey Davies, +John Caron and other contributors to the netCDF news group +have influenced the development of this standard. +We have made some changes to gain greater compatibility with +the NCAR CSM netCDF standard. +

+

2 Filename

+COARDS: NetCDF files should have the file +name extension .nc. +

+ +

3 Data types

+

+OGDT: The netCDF data types +char, short, long, float, and +double +are all acceptable. +All numeric types are signed. +The byte data type, which is functionally identical to char, +is not recommended because its signedness is ambiguous in netCDF. +The COARDS convention deprecates char, rather than +byte. +

+OGDT: NetCDF does not support a character +string type, +so these have to be represented as char arrays. +In this standard, we refer to them as type ``string''. +A string array must be implemented as a +two-dimensional character data variable, +serving as a vector of fixed-length strings, +the second dimension of its CDL declaration +(leading dimension in terms of Fortran) +being a constant StringMaxLength, +recorded as a dimension in the netCDF file. +

+ +

4 Attributes

+

+COARDS: This standard describes many +attributes +(some mandatory, others optional), +but a file may also contain non-standard attributes. +Such attributes do not represent a violation of this standard. +Application programs should ignore attributes that they do not recognise +or which are irrelevant for their purposes. +UNIDATA: Conventional attribute names +should be used wherever applicable. +Non-standard names should be as meaningful as possible. +Before introducing an attribute, consideration should be given to +whether the information would be better represented as a variable. +In general, if a proposed attribute +requires ancillary data to describe it, +is multidimensional, requires any of the defined netCDF +dimensions to index its values, +or requires a significant amount of storage, +a variable should be used instead. +OGDT: When this standard defines string +attributes which make take +various prescribed values, the possible values are given in lower case. +However, applications programs should not be sensitive to case +in these attributes. +NEW: Several string attributes are +defined by this standard to +contain ``blank-separated lists''. +Consecutive words in such a list are separated by one or more +adjacent spaces. The list may begin and end with any number of spaces. +See Appendix A for a list of attributes described by this +standard. +

+ +

5 Global attributes

+

+COARDS: Although not mandatory, +the Unidata-standard attribute history is recommended to +record the evolution of the data contained within a netCDF +file. Applications which process netCDF data can append their +information to the history attribute. +

+NEW: The Unidata-standard attribute +Conventions +is recommended to reference this standard, containing the +string "GDT 1.1". +This standard is registered with Unidata under the name ``GDT'' +in the directory ftp://ftp.unidata.ucar.edu/pub/netcdf/Conventions +and is available from +http://www-pcmd +i.llnl.gov/drach/GDT_convention.html and +http:/ +/www.met-office.gov.uk/sec5/CR_div/GDT_convention.html. +

+OGDT: Use of the string attributes +institution and +production is recommended. +The attribute +institution specifies who produced or supplied the data. +We prefer this name to ``center'' or ``centre'' because the two +possible spellings could cause confusion. +The attribute +production indicates how the data was produced. +If it was model-generated, production should name the model +and its version, as specifically as could be useful. +If it is observational, production should characterise it +e.g. "surface observation" or +"radiosonde". +A further string attribute comment is proposed, to contain +extra information about the file, +and additional attributes may be included as required. +The Hadley Centre, for example, will include an attribute +to name the model integration. +

+OGDT: The float attribute +appendices is recommended to record +the version number of the appendices to this standard used by the +application which generated the file (see section 12). +This information could perhaps be recorded in the Conventions +attribute, but having a separate attribute +for it will allow applications to extract the information +without having to parse a string. +

+OGDT: The calendar attribute +(see section 24) may +be recorded as a global attribute. +The global calendar +attribute is interpreted as a default for all time axes. +

+ +

6 Variable names

+

+COARDS: Variable names should begin with +a letter +and be composed of letters, +digits, and underscores. +OGDT: Case is significant in netCDF +names, but it is recommended that +names should not be distinguished purely by case +i.e. if case is disregarded, no two names should be the same. +NEW: It is also recommended that variable +names should be obviously +meaningful, if possible, as this renders the file more effectively +self-describing. +However, nothing in this convention relies on the use +of particular names for variables. +

+ +

7 Data variables

+

+OGDT: The netCDF variables which contain +the +physical data are referred to as ``data variables'', +also referred to as ``primary variables'' by Unidata. +Apart from the general naming rules for variables (above, +section 6), +the names of data variables are not standardised by these +conventions (since files may in general contain multiple +data variables of the same physical quantity). +

+ +

8 Coordinate variables

+

+OGDT: A one-dimensional netCDF variable +associated with a dimension of one or more data variables +is called a ``coordinate variable''. +A coordinate variable whose dimension name is identical to +its own name is referred to as a ``main coordinate variable'' +in this standard, +when it is necessary to distinguish it +from other types of coordinate variable +(sections 17, 18, 19, +20 and 21). +Apart from the general naming rules for variables (above, +section 6), +the names of coordinate variables are not standardised by these +conventions (since files may in general contain multiple +coordinate variables of the same orientation). +UNIDATA: The values in a main coordinate +variable must be +strictly monotonic +(all values are different and either increasing or decreasing) +because this assumption is frequently made by software. +

+ +

9 Axes and dimensionality of a data +variable

+

+OGDT: A data variable may have any number +of dimensions, including zero, +and the dimensions must all have different names. +COARDS strongly recommends limiting the number to four, +but we wish to allow greater flexibility. +The dimensions of the variable define the axes of the quantity it contains. +Dimensions other than those of space and time may be included. +Several examples can be found in this document. +Components of vector or tensor quantities could be contained +in a single data variable by giving the variable a dimension over components. +While there exist advantages for manipulating such a variable in memory, +we see no strong advantage in introducing this complexity into the netCDF +description, and do not recommend it. +Under certain circumstances, one may need more than one +dimension in a particular quantity +(see section 30 concerning multiple time axes). +For instance, a data variable containing a two-dimensional probability +density function might correlate the temperature at two different +vertical levels, +and hence would have temperature on both axes. +

+COARDS: If any or all of the +dimensions of a data variable have the interpretations +of ``date or time'' (T), ``height or depth'' (Z), +``latitude'' (Y), or ``longitude'' (X) then those +dimensions should appear in the relative order T, then Z, +then Y, then X in the CDL definition corresponding to the +file. +In terms of Fortran, this means X is the first dimension +of the array. Non-spatiotemporal dimensions +should be placed to the left +of the spatiotemporal dimensions +i.e. as trailing dimensions in terms of Fortran. +

+The reason for this convention is that these kinds of axes may +have special meanings to particular applications. +For instance, an application might want to plot a longitude-latitude map, +or integrate vertically, or extract a timeseries. +In the COARDS standard, the indication given by the order of dimensions +and information in the attributes of the coordinate variables +have to be used together to identify the required axes. +For compatibility with COARDS, we uphold all these conventions, +but we also introduce a new attribute to make the identification +straightforward and unambiguous, as follows. +

+NEW: If the last four dimensions +do not have the interpretations TZYX +(in CDL order, omitting from the left if there are fewer than four +dimensions), +an axis attribute should be attached to the data variable. +In other cases it is optional, but recommended. +This attribute is a char array of size equal to the dimensionality +of the +data variable, +having one element for each dimension (in CDL order), +indicating the interpretation of that dimension. +The permitted characters are T Z Y and X, +with the meanings given above, +and - as a placeholder for a dimension which has none of these +meanings. +Each permitted letter may appear no more than once in the array. +If a data variable has more than one dimension which could be given a certain +interpretation, +the axis attribute will therefore clarify which should be chosen. +Note that if there are multiple time axes (section 30), +of which only one is not collapsed, +this single ``climatological time'' axis would normally be the +designated T-axis. +If the axis attribute is included, +the dimensions may be put in any order, +but this should be avoided if possible +because applications not able to use this attribute +may not process the data correctly. +

+

Axes for an ordinary time-mean longitude-latitude-height +variable: +

  dimensions:
+    lat=18;
+    lon=36;
+    pressure=15;
+    con_time=1;
+  variables:
+    float xwind(con_time,pressure,lat,lon); // order T Z Y X
+      xwind:axis="TZYX";
+    float lon(lon);
+    float lat(lat);
+    float pressure(pressure);
+    float con_time(con_time);
+See section 11 for the description of the units and +quantities +of the variables and sections 14, 15, 16, +25 +for details of longitude, latitude, vertical and time axes. + +

+OGDT: The coordinates of points within +the data variable are +the simple ordered tuples formed by associating values from the +coordinate variables (section 8). +NEW: If a particular axis does not have a +coordinate variable, +the coordinate values are assumed to be equal to their indices along +the axis, numbering from 0. +

+OGDT: Dimensions may be of any size, +including unity. +When a single value of some physical quantity applies to all the values +in a data variable, +the recommended means of attaching this information to the variable +is by use of a singleton dimension (a dimension of size unity) +with a one-element coordinate variable. +The advantage of this method is that all the attributes of a coordinate +variable (quantity, components, boundaries, etc.) can be used to describe +the single-valued quantity. +Singleton dimensions also result from contractions, described +in section 23. +

+

Longitude-latitude field of temperature +on a pressure level: This would use a +singleton pressure dimension to record the level, thus: +

  dimensions:
+    lon=96;
+    lat=72;
+    pressure=1; // single-valued coordinate variable
+  variables:
+    float temperature(pressure,lat,lon); // axes in order Z Y X
+      temperature:axis="ZYX";
+    float pressure(pressure);
+      pressure:long_name="pressure";
+      pressure:units="kPa";
+  data:
+    pressure=50.0; // Pressure level of 50 kPa = 500 mbar
+The units and long_name attributes +are described in section 12. + +

+

Surface air temperature: Surface meteorological +measurements are made at +a certain defined height e.g. 1.5 m, which can be shown thus: +

  variables:
+    float temperature(height,lat,lon);
+      temperature:axis="ZYX";
+      temperature:long_name="atmospheric temperature";
+      temperature:units="K";
+    float height(height);
+      height:long_name="height above the surface";
+      height:units="m";
+  data:
+    height=1.5;
+The explicit height should not be given if the surface of measurement +is included in the name of the quantity e.g. screen height. + +

+ +

10 Coordinate systems

+

+OGDT: A coordinate system for the Earth's +surface which is rectilinear but +based on a polar axis other than the normal geographical axis is referred +to as a ``rotated grid''. +To describe rotated grids, a two-element float attribute +north_pole is attached to the data variable, +specifying the (longitude,latitude) coordinates of the rotated north pole. +If the attribute is absent and relevant, +it is assumed to have the value (0.,90.) i.e. the geographical north pole. +

+OGDT: In some systems, +the axes covering the Earth's surface +do not define a rectilinear grid. +We do not wish necessarily to exclude non-rectilinear systems. +For the moment, this standard is undefined for these systems, +and we invite comments from potential users on the appropriate definition. +The COARDS standard excludes non-rectilinear systems. +In principle, any coordinate system can be handled, +albeit clumsily, by replacing the relevant two or more axes by +by a single axis which indexes the points, +and providing associated coordinate variables to specify the coordinates, +point by point (see section 19). +

+ +

11 Units

+

+COARDS: The udunits package includes a +file udunits.dat, which lists collections of unit names. The names +given in the most recent version of this file +and their plural forms will be regarded as acceptable +unit names for this standard, with a few modifications which will be +listed in Appendix C to this standard. +COARDS lists some modifications within the standard, +but we would prefer to put in place a means to allow future +modifications to be made easily. +OGDT: Users of this standard should not +define their own units, +because this would make their files less portable; +requests for new units should be directed to Unidata. +

+OGDT: The udunits package also defines a +means for linear transformation +of units by a scale factor and an offset. +This convention is allowed when it is natural to express a unit in +such a form +e.g. density of sea-water in +kg m¯³ +in excess of 1000 +kg m¯³, +which can be specified to udunits as "kg m-3 @ 1000". +COARDS does not permit the use of this facility. +This facility should not be used as a means of data compression, +for which an alternative is provided +(see section 34). +

+ +

12 Physical quantity of a +variable

+

+NEW: These conventions standardise two +attributes for specifying +the physical quantity of data and coordinate variables. +Both of them are strings and defined by Unidata conventions. +

+OGDT: The units attribute is +formatted as per the +recommendations in the Unidata udunits package +(see section 11), +with extensions for time +(see section 27). +Case is significant in the units. +This attribute is mandatory unless the quantity is dimensionless +(a pure number). +There are a few defined dimensionless units, +such as percent, but +there is no need for a wide variety of dimensionless units for quantities +like sea-ice concentration, cloud fraction, probability and so on; +this descriptive information is the long_name rather +than the units. +A scale factor and/or offset may be specified +quantity e.g. sea-ice concentration in tenths +may be given as units="0.1f". +A dimensionless quantity with no scaling of offset +may have units="1.0f" or +units="unity". +

+NEW: The long_name is an +attribute +containing a descriptive name, +which should not specify the units. +Case is not significant in the long_name. +This attribute is optional for most variables, +but mandatory for coordinate variables +of longitude, latitude, vertical axes and time +(see sections 14, 15, 16 +and 25). +If the long_name attribute is present, +it must be a quantity chosen from the list in +Appendix D of this standard, +optionally with additional information enclosed in parentheses () +in the case that sufficient detail for local use cannot be given by +a standardised description. +The purpose of the list is to ensure that comparable data +from different sources has the same description. +Each quantity in Appendix D will be labelled with the version of the +appendices at which it was introduced, +enabling a generic application to deduce the complete set of +quantities which was available to the application which generated the file. +

+

Quantity attributes: +

    float tempt(pressure,lat,lon);
+      tempt:long_name="atmospheric potential temperature "
+        "(after timestep)";
+      tempt:units="K";
+"atmospheric potential temperature" is the standard +quantity, and "after timestep" is additional information, +which a generic application can disregard. + +

+Whether two physical quantities are different or the same is often +not a question with a well-defined answer. +Certainly if they are the same, they must have the same unit, but +various quantities with the same unit may have to be distinguished +e.g. atmospheric potential temperature and soil temperature. +In practice, +the most specific description applicable should be used. +We intend to expand Appendix D on an ongoing basis +in response to requests by users of this standard, +since we cannot foresee all the possibilities, +and we will err on the side of expansion, rather than restriction, +when it is unclear whether a new quantity is needed. +

+NEW: The subgrid attribute (see +section 22) +can be regarded as a modifier of the long_name; +it applies only to data variables, not coordinate variables. +The long_name and subgrid attributes together define +the physical dimensions of the quantity +(through information given in Appendices B and D), +and the units must be consistent with this. +

+OGDT: The optional modulo +attribute of a variable, if present, +records a number which can be added or subtracted +without altering the validity or physical significance of +the quantity. +This is most likely to be useful for longitude coordinate axes +(section 14), +with a modulo of 360, +and climatological +axes of seasonal or diurnal phase (sections 27 +and 30). +

+OGDT: We note that the Unidata-standard +FORTRAN_format attribute +may be useful for both coordinate and data variables. +

+OGDT: In addition, +other model-dependent attributes may included to define +the quantity of a variable. +The Hadley Centre model will give each data variable integer +stash +and submodel attributes, for example, which are codes identifying +GCM diagnostic output quantities. +

+ +

13 Topology of an +axis

+

+OGDT: An axis with ``circular topology'' +is one +which can be legitimately +transformed by shifting all the points one place along the axis, +moving the last point to the beginning, any number of times. +The main coordinate variable of an axis with circular topology is +distinguished +by the presence of an attribute topology="circular". +A longitude axis which circles the whole globe is an example. +The value linear or the absence of this attribute +indicates an axis with ``linear topology''. +The topology is indicated only by the main coordinate variable, +but since it is the property of the axis it applies to any ancillary +coordinate variables (section 17) as well. +

+OGDT: When a circular axis is rotated, +the main coordinate values must be +altered in order to remain monotonic. +Therefore the main coordinate variable of a circular axis +requires a modulo (section 12). +

+Note that the topology and modulo attributes convey +different information. +For instance, +a longitude coordinate variable limited to values in the eastern +hemisphere between the Greenwich meridian and the date-line (e.g. 0E, 25E, +120E, 130E, 180E) does not have circular topology. (This might be from a model +of a limited area of the world.) When +making a contour map of a field with such +a longitude axis, one can interpolate anywhere within +the eastern hemisphere to draw the +contours, but it is not legitimate to interpolate over the western hemisphere +and draw the rest of the world, which is simply missing. The +implication of circular topology would be that one could put any longitude +at all on the left-hand side of the map. However, +this coordinate variable does have a modulo (of 360, +as required in section 14), and the points +can be labelled +in any way which is equivalent under the modulo to the coordinates +in the file. +The coordinate values 0,25,120,130,180 are thus equivalent to +-360,-335,-240,-230,-180.} + +

+ +

14 Longitude dimension

+

+COARDS: Coordinate variables representing +longitudes must always explicitly +include the units attribute; there is no default value. +The units attribute will be a +string formatted as per the recommendations in the Unidata udunits +package. +The recommended unit of longitude is degrees_east (eastward +positive). Also acceptable are degree_east, degree_E, and +degrees_E. The unit degrees_west (westward positive) is not +recommended because it implies a negative conversion factor from +degrees_east. +OGDT: A longitude coordinate variable +must have an attribute +long_name="longitude". +COARDS: Such a variable is identifiable +from its long_name +and its units. +The COARDS convention relies on the unit as the only way to +identify a longitude variable. This standard uses the long_name, but +requires the units to be specified as well for compatibility with +COARDS. +

+OGDT: Longitude axes should have the +attribute modulo=360, +indicating that they may be interpreted modulo 360. +Thus, for example, +-180, 180, and 540 are all valid representations of the International Dateline +and 0 and 360 are both valid representations of the Prime Meridian. +COARDS assumes that longitudes may always be treated in this way. +Since we have introduced the modulo attribute, +we require that it should be specified to indicate this. +A global longitude axis should have the attribute +topology="circular". +Note that the presence of a modulo attribute +does not mean that the axis necessarily has circular topology +(section 13); +a longitude axis covering only part of the globe cannot have its points +rotated. +COARDS: The sequence of numerical +longitude values stored in the netCDF +file must be monotonic in a non-modulo sense +for a main coordinate variable of longitude. +

+

Global longitude axis: +

    float lon(lon);
+      lon:long_name="longitude";
+      lon:modulo=360.0f;
+      lon:topology="circular";
+      lon:units="degrees_east";
+ +

+ +

15 Latitude dimension

+

+COARDS: Coordinate variables representing +latitudes must always explicitly +include the units attribute; there is no default value. +The units attribute will be a +string formatted as per the recommendations in the Unidata udunits +package. +The recommended unit of latitude is degrees_north. Also acceptable +are degree_north, degree_N, and degrees_N. +OGDT: A latitude coordinate variable must +have an attribute +long_name="latitude". +COARDS: Such a variable is identifiable +from its long_name +and its units. +The COARDS convention relies on the unit as the only way to +identify a latitude variable. This standard uses the long_name, but +requires the units to be specified as well for compatibility with +COARDS. +

+

Latitude axis: +

    float lat(lat);
+      lat:long_name="latitude";
+      lat:units="degrees_north";
+ +

+ +

16 Vertical (height or depth) +dimension

+

+OGDT: Whereas the two horizontal +dimensions are usually longitude +and latitude, whose direction is well defined, +a variety of quantities may be used for the vertical axis, +if there is one. +NEW: The axis to be regarded as the +vertical axis +must have both a long_name attribute (section 12) +and a positive attribute, +with one of the allowed values up or down, +to indicate the sense of the direction of positive +since this information may be useful for +applications which display the data. +

+

Vertical pressure axis: +

  dimensions:
+    pressure=15;
+  variables:
+    float pressure(pressure);
+      pressure:long_name="pressure";
+      pressure:positive="down";
+      pressure:units="hPa";
+  data:
+    pressure=850, 700, 500, 300, 200, 150, 100, 50, 30, 20, 10;
+ +

+The COARDS standard requires the units of +the vertical axis to be selected from a defined list, +in order that this axis can be recognised by its units. +It gives special status to units of pressure, +for which the direction of positive is defined, +and makes the positive attribute mandatory +for vertical axes with other units. +

+We have adopted a different approach for a number of reasons. +Firstly, to require units for the vertical axis +means defining dimensionless units for any +dimensionless quantity one might wish to use for the coordinate variable. +This is inconsistent with the treatment of a data variable; +the standard does not require that dimensionless units be invented +for dimensionless physical quantities in data variables. +Secondly, the long_name attribute is more informative than the units. +Thirdly, the vertical dimension of a data variable can be identified +from the axis attribute or the +the order of dimensions +(see section 9), +allowing any application which expects such a +dimension to find it without any further help. +

+We are not entirely convinced that +the direction of positive should really be recorded as part of the +data structure. +It is mostly an issue for displaying the data, +and is to some extent a matter of personal preference. +If such special treatment is given to the vertical axis, +why is it not also recorded for other axes? +For instance, when latitude is shown on the horizontal axis of a plot, +is north on the left or the right? +This is the same kind of question, +but it strikes us as more a matter for a graphics application to consider. +Nonetheless, we have required the positive attribute for +compatibility with COARDS. +

+For example, if an oceanographic netCDF file encodes the depth of the +surface as 0 and the depth of 1000 m as 1000 then the axis would +use attributes as follows: units="m", +long_name="depth below the surface", +positive="down". +If, on the other hand, the depth of 1000 m were represented as +-1000, +we would have long_name="height above the surface", +positive="up". +

+ +

17 Ancillary +variables

+

+NEW: ``Ancillary'' variables supply +additional sets of values belonging +somehow with other variables (usually coordinate variables), +referred to as ``main'' variables. +Ancillary variables +are needed for various purposes described in sections 18, +19 and 20. +These sections also define attributes of the main variables which enable +applications to retrieve the ancillary variables. +The dimensions of an ancillary variable must all be dimensions of its +main variable. +

+ +

18 Component variables

+

+OGDT: A continuous physical variable may +require +more than one number to specify it at each point. +We refer to these as ``components''. +The values of the components +are recorded in ancillary variables +referred to as ``component variables''. +The names of the component variables are recorded as +a blank-separated list in a component string attribute +of the main variable. +The dimensions of a component variable must be identical with those +of its main variable. +OGDT restricted components to coordinate variables, +but the concept has here been generalised. +

+OGDT: When a coordinate variable has +components, +this standard requires that a main coordinate variable should nonetheless +be supplied which represents a combination of the components that can be +used to order the points on the axis. +As usual, this main coordinate variable must be monotonic, +but the components do not need to be monotonic. +

+

Hybrid vertical coordinate: +A vertical coordinate +eta = p/p0 + sigma +is used in some atmospheric GCMs. +Atmospheric model levels are specified in terms of +(p,sigma) +pairs, +where p is pressure, +p0 +is a constant and +sigma +is fraction +of surface pressure (which is variable). +The eta +value is a linear combination of the two, +which cannot be uniquely decomposed back into +(p,sigma). +We would record this coordinate variable thus: +

    float eta(eta); // main coordinate variable
+      eta:component="pressure sigma";
+    float pressure(eta);
+    float sigma(eta);
+A generic application would treat the component and main coordinates +as independent information. +The extra knowledge required to relate them would reside in any +specific application which needed it. + +

+ +

19 Associated variables

+

+NEW: An axis of a data variable, or two +or more axes in combination, +may have alternative sets of coordinate values. +These alternative sets are recorded in ancillary variables +of the data variable, and +referred to as ``associated'' variables. +The data variable records the names of associated variables +as a blank-separated list in an associate string attribute. +This attribute may alternatively and equivalently be named +coordinates. +This possibility is included for compatibility with the CSM standard. +However, in this standard it is deprecated because of possible confusion +with the normal definition of ``coordinate variable'' (section +8) and because the use of associated variables is +wider +than just coordinate variables in the usual sense. +An associated variable may be named by +more than one data variable in this way. +If an associated variable itself has an associate attribute, +variables named by this attribute are also regarded as being +associated with the data variable. +An associated variable must have dimensions which are all dimensions +of any data variable with which it is associated; +the associated variable can be regarded as a function +of the indices along these axes. +The values of an associated variable do not have to be monotonic. +A generic application is not required to make any +use of associated variables. +Associated variables are not indicated in the axis attribute of the +data variable (section 9). +

+

Vertical axis: Many associated variables +will be one-dimensional, +giving alternative sets of values for a single axis. +One example is a vertical axis where one wishes to store both the +physical coordinate and the ordinal model level number: +

  dimensions:
+    lat=90;
+    sigma=19;
+  variables:
+    float xwind(sigma,lat); // 2D data variable
+      xwind:associate="model_level";
+      xwind:axis="ZY";
+    float lat(lat);
+      lat:long_name="latitude";
+    float sigma(sigma); // physical height coordinate
+      sigma:long_name="sigma";
+      sigma:positive="down";
+    int model_level(sigma); // model level number at each height
+      model_level:long_name="model level number";
+      model_level:positive="up";
+ +

+

Surface height: Pressure is given at the surface, +but we also wish to define the height of the surface itself, +in case some subsequent application needs to convert to sea-level +pressure, for example. +

    float pressure(lat,lon);
+      pressure:associate="z";
+      pressure:long_name="pressure at the surface";
+      pressure:units="Pa";
+    float z(lat,lon);
+      z:long_name="height of the surface above sea level";
+      z:units="m";
+ +

+

Trajectory: The +value of a quantity along a one-dimensional +trajectory. +In such a case, we might have a coordinate variable containing +time of travel and +associated coordinate variables +giving the latitude and longitude of each point: +

  dimensions:
+    day=10; // 10 sample times along a trajectory
+  variables:
+    float hice(day); // sea-ice thickness measured as the floe drifts
+      hice:associate="lon lat";
+      hice:axis="T";
+    float day(day); // number of days since the beginning of the journey
+    float lon(day); // longitude at each time
+    float lat(day); // latitude at each time
+The main coordinate variable (day) must be +monotonic, but the associated coordinates are not necessarily. +An important application +rather similar to this one +is described in section 20. +Note that lat and lon cannot be +indicated as the X and Y coordinates in the axis +attribute (section 9). +This is reasonable because, even though they have the interpretations +of latitude and longitude, they are not independent dimensions in +the sense normally expected by an application which might look for +such axes. + +

+An alternative convention would be to associate lon and +lat to the coordinate variable day instead of to the +data variable. This would have the effect that all data variables +that share the day axis would +automatically share the associated coordinates, +whereas in the current convention the association has to be repeated +on each data variable. +We choose to follow the current convention for two reasons. +Firstly, it allows the possibility that the same main coordinate +variable might occur in combination with various different sets of +associated coordinate variables. +In this example, there might be more than one trajectory, with the +same day coordinates but different longitude-latitude positions; +it is therefore more flexible to allow different associations for +each data variable. +Secondly, association on the data variable, +rather than the coordinate variable, +is the only possibility for multidimensional coordinate variables, +such as we discuss in the next example. +

+

Transformed coordinates: Associated variables +of more than one dimension can be used to +describe alternative coordinate systems. +For example, vertical profiles of atmospheric humidity might +be available on a regular longitude-latitude grid, +but we might also wish to give the national grid coordinates of +each point. +The national grid x- and y-coordinates are each functions of +both latitude and longitude; +the x-coordinate does not correspond specifically to longitude, +nor the y- to latitude. +The appropriate representation is therefore: +

  dimensions:
+    lon=10;
+    lat=20;
+    pressure=15;
+  variables:
+    float humidity(pressure,lat,lon);
+      humidity:associate="x y";
+    float pressure(pressure);
+    float lon(lon); // 1D main coordinate variable
+      lon:long_name="longitude";
+    float lat(lat);
+      lat:long_name="latitude";
+    float x(lat,lon); // 2D associated coordinate variable
+      x:long_name="UK national grid eastings";
+    float y(lat,lon);
+      y:long_name="UK national grid northings";
+This tells us that humidity[*][10][5] is the vertical profile +of humidity at the point with latitude lat[10] and +longitude lon[5], +which is at national grid x-coordinate x[10][5] and y-coordinate +y[10][5]. + +

+

No main coordinate variables: +A related situation is when the 2D grid is staggered or transformed +geometrically in some way (other than rotation--see section 10), +so that it is not possible or easy to give +one-dimensional coordinate variables for the axes. +In this case, there would be no main coordinate variables, +and the default of plain indexes would apply. +The physical coordinates are functions of the 2D gridpoint indices, +and would be given in associated variables just as above: +

  dimensions:
+    x=90;
+    y=45;
+  variables:
+    float orog(y,x); // 2D variable on a horizontal grid
+      orog:associate="lon lat";
+      orog:axis="--";
+      orog:long_name="height of the surface above sea-level";
+      orog:units="m";
+    float lon(y,x); // 2D coordinate variable on the same grid
+    float lat(y,x);
+The lat and lon variables are not indicated as +the X and Y coordinates in the axis variable +(section 9). +An application which referred to latitude and longitude coordinates +would not generally expect them to be two-dimensional. +If it can handle this situation, it should identify these axes by +their long_name and units. + +

+

3D associated coordinates: +These could be used if one wished to describe +a field with alternative 3D coordinate systems, +for instance on both a regular Cartesian grid, +and in cylindrical or spherical coordinates. +The values of the alternative coordinates would be given on the Cartesian +grid. In the spherical case, for instance +

    float temperature(z,y,x); // 3D variable on a Cartesian grid
+      temperature:associate="radius theta phi";
+    float radius(z,y,x);
+    float theta(z,y,x);
+    float phi(z,y,x);
+ +

+NEW: A particular technical application +of a one-dimensional +associated coordinate is to deal with the limitation of netCDF to a +single unlimited dimension. +If several data variables have unlimited axes of different lengths or +physical significance, +they can all share a nominal unlimited dimension, +and each have associated variables specifying the meaning of the axis. +

+

More than one unlimited axis: +Consider a file which contains data variables with unlimited +axes measuring elapsed time with different sampling frequency, +and hence of different lengths. +

  dimensions:
+    time_counter=UNLIMITED;
+  variables:
+    float sw(time_counter); // sampled every 3 hours
+      sw:associate="time_3h";
+      sw:axis="T";
+      sw:long_name="vertical component of "
+        "shortwave radiative flux density";
+      sw:units="W m-2";
+    float latent(time_counter); // sampled every 30 minutes
+      latent:associate="time_30min";
+      latent:axis="T";
+      latent:long_name="latent heat flux density";
+      latent:units="W m-2";
+    float time_3h(time_counter)";
+      time_3h:long_name="elapsed time";
+      time_3h:units="h";
+    float time_30min(time_counter);
+      time_30min:long_name="elapsed time";
+      time_30min:units="min";
+ +

+ +

20 Bundles

+

+OGDT: If several data arrays containing +the same physical quantity +have one or more identical axes, +but are distinguished by the values of other singleton coordinate variables, +it may be convenient to store them in the same data variable. +The common axes of the separate arrays become axes of the combined variable. +One or more additional axes are introduced to ``bundle up'' the separate +arrays. +Such an axis does not correspond to a continuous physical coordinate. +It acts simply as an index of the bundled-up arrays. +

+OGDT: The singleton values of the +separate arrays are recorded in +associated coordinate variables for the bundling dimension. +They should not be interpreted as continuous coordinates. +

+

Timeseries: +The Hadley Centre GCM +can generate timeseries of the values of quantities +at individual points. +Typically, timeseries from many different points are produced of the same +quantity at the same sampling times. +It is natural to contain this information +in a data variable with two dimensions. +One dimension is the common time axis, specifying the sampling times, +which are the same for all the points sampled. +The other dimension is not a continuous physical coordinate; +it is simply being used to ``bundle up'' the timeseries, +the points being +irregularly scattered in a space of two or more dimensions. +Thus: +

  dimensions:
+    points=15; // measurement locations
+    times=20; // sampling times
+  variables:
+    float snowdepth(times,points);
+      snowdepth:associate="lon lat sitename";
+      snowdepth:axis="T-";
+    float lon(points); // longitude of sites
+    float lat(points); // latitude of sites
+    char sitename(points,StringMaxLength); // string array of sitenames
+    double times(times); // times of measurement
+See section 24 concerning the time coordinates. +This same form could be used for observed timeseries from stations. +The bundling axis (points) is simply an index. + +

+

Vertical profiles: +A similar application is that of vertical +profiles at sets of points; +for example, scattered vertical temperature profiles through the ocean, +or data from various radiosonde stations. +

  dimensions:
+    station=10; // measurement locations
+    pressure=11; // pressure levels
+  variables:
+    float humidity(pressure,station);
+      humidity:associate="lon lat";
+      humidity:axis="Z-";
+    int station(station); // station numbers
+    float lon(station); // longitude of stations
+    float lat(station); // latitude of stations
+    float pressure(pressure)
+ +

+This section raises the question of how best to store a single timeseries, +or a single vertical profile. +Following the scheme of this section, +it could be contained in a two-dimensional data variable with the bundling +axis being of size unity. +The associated information such as latitude or longitude would then be +stored in singleton coordinate variables, +all associated with the same dimension. +Alternatively, these values could be recorded as +separate singleton dimensions (following section 9). +We have no recommendation for this. +Either scheme could be appropriate; +which is more natural perhaps depends on how the data was extracted from +the continuous axes. +

+ +

21 Boundary variables

+

+OGDT: Along a dimension, +the values might relate to points (at the coordinate values) +or to contiguous or non-contiguous cells. +The boundaries of the cells should be defined +as well as the point coordinate values. +The convention is to define an additional two-dimensional +``boundary variable'' with +a left-hand dimension (trailing dimension in Fortran terms) of size +two. +The values for which this dimension has index 0 +(numbering from 0 i.e. in C notation) supply the boundaries +with the smaller coordinate values, +and those with index 1 the large values, +where ``smaller'' and ``larger'' refer simply to numerical comparison, +not to a physical direction. +Supplying upper and lower boundaries separately allows for the possibility +that the cells might not be contiguous; they might even overlap. +If a lower boundary value is equal to the valid_min for the +coordinate +variable (section 31), the cell has no lower boundary. +If an upper boundary value is equal to the valid_max, +the cell has no upper boundary. +The name of the boundary variable is recorded in a string attribute +bounds of the main coordinate variable. +We recommend that the it should be named +by the coordinate dimension with the prefix bounds_. +

+

Boundaries for a one-dimensional latitude coordinate +variable: +

    float lat(lat);
+      lat:bounds="bounds_lat";
+    float bounds_lat(2,lat);
+In C notation, +lat[0] gives the coordinate of the first point, +bounds_lat[0][0] its lower boundary, bounds_lat[1][0] +its upper boundary. +In Fortran notation, the declarations are lat(lat) +and bounds_lat(lat,2), +and the relevant elements are lat(1), bounds_lat(1,1), +bounds_lat(1,2). + +

+

Albedo as a function of wavelength and snow cover: +Characteristic values of albedo are given for various wavelength +bands, dependent also on snowdepth. +

  dimensions:
+    lambda=4; // number of shortwave frequency bands
+    snowdepth=10; // number of snowdepth categories
+  variables:
+    float albedo(lambda,snowdepth); // no units for albedo
+      albedo:axis="--";
+      albedo:long_name="surface albedo";
+    float lambda(lambda);
+      lambda:bounds="bounds_lambda";
+      lambda:long_name="wavelength";
+      lambda:units="nm";
+    float bounds_lambda(2,lambda);
+    float snowdepth(snowdepth);
+      snowdepth:bounds="bounds_snowdepth";
+      snowdepth:long_name="mass per unit area of lying snow";
+      snowdepth:units="kg m-2";
+      snowdepth:valid_max=1e9;
+    float bounds_snowdepth(2,snowdepth);
+  data:
+    lambda=250, 385, 570, 795;
+    bounds_lambda=175, 320, 450, 690,
+                  320, 450, 690, 900;
+    snowdepth=0.05, 0.15, 0.35, 0.75, 1.25, 1.75, ..., 450.0, 1000.0;
+    bounds_snowdepth=0.0, 0.1, 0.2, 0.5, 1.0, 1.5, ..., 400.0, 500.0,
+                     0.1, 0.2, 0.5, 1.0, 1.5, 2.0, ..., 500.0, 1e9;
+A first index of 0, for instance, gives albedo values for +the wavelength range 175-320 nm. +The deepest snowdepth class has no upper bound; any value above 500 +falls into this class. + +

+

Probability density function of precipitation amounts: +

  dimensions:
+    ppn=10;
+  variables:
+    float pdf(ppn,lat,lon);
+      pdf:axis="-YX";
+      pdf:long_name="probability density of "
+        "depth of water-equivalent precipitation";
+      pdf:units="mm-1";
+    float ppn(ppn);
+      ppn:units="mm";
+      ppn:long_name="depth of water-equivalent precipitation";
+      ppn:bounds="bounds_ppn";
+    float bounds_ppn(2,ppn);
+  data:
+    bounds_ppn=0.0, 0.1, 0.2, 0.5, ..., 0.1, 0.2, 0.35, 1.0, ...;
+pdf[3][10][12] gives the probability density of precipitation +amounts between 0.5 and 1.0 mm falling at the location +lat[10] lon[12]. + +

+OGDT: Boundary variables +are recommended if the main coordinate values +are not evenly spaced, +or if the dimension has a size of unity. +If the coordinates are evenly spaced, +and boundaries are not specified, +generic applications may assume that the main coordinates +lie at the centres of their cells. +Boundary variables may be supplied for ancillary +variables as well as main coordinate variables. +The upper and lower boundaries of ancillary coordinates are ordered +so as to correspond with those of the main coordinate. +

+

Boundary values for a hybrid vertical coordinate: +The atmospheric column is here divided into three cells in the vertical; +from the surface to +sigma=0.7; +from there to 20 kPa, and finally to the top of the atmosphere, +using the hybrid vertical coordinate introduced in an example +in section 18. +

  dimensions:
+    eta=3;
+  variables:
+    float(eta);
+      eta:long_name="pressure-sigma hybrid";
+      eta:component="pressure sigma";
+      eta:bounds="bounds_eta";
+      eta:positive="down";
+    float bounds_eta(2,eta);
+    float pressure(eta);
+      pressure:units="kPa";
+      pressure:long_name="pressure";
+      pressure:bounds="bounds_pressure";
+    float bounds_pressure(2,eta);
+    float sigma(eta);
+      sigma:long_name="sigma";
+      sigma:bounds="bounds_sigma";
+    float bounds_sigma(2,eta);
+  data:
+    eta=0.75, 0.45, 0.1;
+    bounds_eta=1.0, 0.7, 0.2,
+               0.7, 0.2, 0.0;
+    pressure=0.0, 10.0, 10.0;  // does not need to be monotonic
+    bounds_pressure=0.0, 0.0, 20.0;
+                    0.0, 20.0, 0.0;
+    sigma=0.75, 0.35, 0.0;
+    bounds_sigma=1.0, 0.7, 0.0;
+                 0.7, 0.0, 0.0;
+ +

+NEW: Boundary variables may be given for +associated multidimensional +coordinate variables (section 19). +Each dimension of the main variable requires an extra dimension of size 2 +in the boundary variable. +These extra dimensions are placed on the left +(right in Fortran terms) of the coordinate dimensions, +and in the same order as the coordinate dimensions. +

+

Boundaries for a two-dimensional latitude coordinate +variable: +

    float lat(y,x);
+      lat:bounds="bounds_lat";
+    float bounds_lat(2,2,y,x);
+so bounds[0][0][4][5] contains the latitude of +the lower left (smaller x and y) +corner of gridbox [4][5], bounds[0][1][4][5] the lower +right corner, +bounds[1][0][4][5] upper left and bounds[1][1][4][5] upper +right. +In Fortran, the indices of the boxes would be (6,5,1,1), +(6,5,2,1), +(6,5,1,2), (6,5,2,2) respectively. + +

+ +

22 Representation of subgrid +variation

+

+NEW: Since a data variable usually +represents a physical quantity which varies +continuously along the axes, +in reality there will generally be variation of the quantity between +adjacent gridpoints. +The data variable can give only one value for each cell, +despite this subgrid variation. +For many purposes, this can be taken as a ``representative'' value, +and it is not necessary to define precisely how it relates to +the subgrid variation. +

+NEW: To be explicit about how each data +value reflects subgrid variation +along a particular axis, +use the subgrid attribute of the data variable. +The most important application of this attribute is to +contracted or collapsed axes, described in section 23. +This is a string attribute comprising a list of blank-separated words. +In this list, "name: method" +indicates that subgrid variation along the axis with the dimension whose +``name'' is given +is represented by the specified ``method''. +The method, which may be several words, should be one of the +permitted values detailed in Appendix B, which include +mean, maximum, minimum, mid-range, +standard deviation, variance, mode, +median, cell, point. +Case and punctuation are not significant in the method. +Like Appendix D, Appendix B will be expanded on request by +users of this standard. +Some methods imply a change of units of the data variable, +and this also is specified by Appendix B. +In the above list, this is true for variance. +The method point indicates that the data values apply exactly +at the coordinate values, and do not at all represent the variation between +adjacent gridpoints along the axis concerned. +The method cell indicates that each value +should be regarded as a property of the whole cell +along the axis concerned e.g. a sum or integral. +The method can be differently specified for the various dimensions. +It must be remembered that the method applies only to the axis indicated. +If a precipitation value in a longitude-latitude gridbox is given the +method maximum for these axes, for instance, it means that it is +the maximum within these spatial cells, and does not imply that +it is also the maximum in time. +

+NEW: The absence of any specification +means that generic applications may +regard the data values as representative in whatever way suits their needs. +For quantities calculated at gridpoints by numerical models, +this kind of vagueness is unavoidable. +If a model provides a longitude-latitude field of temperature at gridpoints, +an application used to draw a contour plot of the field will generally +assume that the temperatures apply at points, and will use some interpolation +scheme to compute values between them. +An application which calculates the mean of the field, however, will +probably assume that the temperatures are gridbox means, +and average them by weighting each with its area. +Both of these approaches are valid. +The finite-difference scheme by definition does not have any information +about subgrid variation, +and may itself treat the values in both ways; +it might calculate gradients between them, regarding them as points, +or enforce conservation properties, regarding them as means. +It would be unusual to regard the values as extrema, however, unless +this was explicitly indicated. +

+

Subgrid time variation in timeseries: +Consider 12-hourly timeseries of pressure, temperature +and precipitation from a number of stations, +where pressure is measured instantaneously, temperature extremes over +the preceding period are recorded by maximum and minimum thermometers, +and precipitation is accumulated in a rain gauge. +For a period of 48 hours from 6 a.m. on 19th April 1998, +the data is structured as follows: +

  dimensions:
+    instanttime=5; // 5 instantaneous measurements at 12-hour intervals
+    periodtime=4; // 4 intervening 12-hour periods
+    station=10;
+  variables:
+    float pressure(station,instanttime);
+      pressure:axis="-T";
+      pressure:long_name="pressure";
+      pressure:subgrid="instanttime: point";
+      pressure:units="kPa";
+    float maxtemp(station,periodtime);
+      maxtemp:axis="-T";
+      maxtemp:long_name="temperature";
+      maxtemp:subgrid="periodtime: maximum";
+      maxtemp:units="K";
+    float ppn(station,periodtime);
+      ppn:axis="-T";
+      ppn:long_name="depth of water-equivalent precipitation";
+      ppn:subgrid="periodtime: cell";
+      ppn:units="mm";
+    double instanttime(instanttime);
+      instanttime:long_name="time";
+      instanttime:units="h since 1998-19-4 6:0:0";
+    double periodtime(periodtime);
+      periodtime:bounds="bounds_periodtime";
+      periodtime:long_name="time";
+      periodtime:units="h since 1998-19-4 6:0:0";
+    double bounds_periodtime(2,periodtime);
+  data:
+    instanttime=0., 12., 24., 36., 48.;
+    periodtime=6., 18., 30., 42.;
+    bounds_periodtime= 0., 12., 24., 36.,
+                      12., 24., 36., 48.;
+It is not appropriate give a subgrid method for the station axis, +since this is a bundling axis (section 20) and not +a continuous physical coordinate. +The instantaneous and period measurements have to have different time +axes both because of their different dimension, +and because they do not coincide. +If the pressure measurements were made at times half-way between the +others (noon and midnight), the time axes could be shared. +Since the precipitation is given as an amount, it is a sum over +the interval of time by definition. +It could instead have been expressed as a rate in mm h-1, for +instance, in which case its subgrid method would be mean +rather than cell. + +

+

Thickness (geopotential difference): The +``thickness'' is the difference +in geopotential height between two pressure surfaces in the +atmosphere. This quantity is by definition one which relates +to the whole extent of its cell in the vertical dimension. +

  variables:
+    float thickness(pressure,lat,lon);
+      thickness:long_name="thickness";
+      thickness:subgrid="pressure: cell";
+      thickness:units="m2 s-2";
+    float pressure(pressure);
+      pressure:bounds="bounds_pressure";
+      pressure:long_name="pressure";
+      pressure:units="hPa";
+    float bounds_pressure(2,pressure);
+Here, bounds_pressure[0][0] and bounds_pressure[1][0] +will be the upper and lower pressure bounds of the thickness field +thickness[0][*][*]. + +

+NEW: If more than one subgrid method is +to be indicated, +they should be arranged in the order of the dimensions of the data variable, +unless their order is significant and different from this. +The right-most operation is assumed to have been applied first. +Suppose a quantity varies in both longitude and time +(dimensions lon and time) within each +gridbox. +Values which represent the time-average of the zonal maximum +are labelled subgrid="time: mean lon: maximum", +i.e. find the largest value at each instant of time over all longitudes, +then average these maxima over time; +values of the zonal maximum of time-averages are +labelled subgrid="lon: maximum time: mean". +

+NEW: If a data value is representative of +variation over a +combination of axes, +a single method should be prefixed by the names of all the dimensions +involved, +which should appear in the order of the dimensions of the data variable, +though this order is not significant. +Dimensions should be grouped in this way only if there is an essential +difference +from treating them individually. +For instance, the subgrid standard deviation of +topographic height within a longitude-latitude gridbox +would have subgrid="lat: lon: standard deviation". +This is not the same as +subgrid="lat: standard deviation lon: standard deviation", +which would mean finding the standard deviation along each parallel of +latitude within the zonal extent of the gridbox, +and then the standard deviation of these values over latitude. +

+NEW: To indicate more precisely how the +subgrid method was +applied, +extra information may be included in parentheses () after the +identification of the method. +This information is not standardised and may be ignored by a generic +application. +A mean over latitude, for instance, may be area-weighted. +This could be indicated as "lat: mean +(area-weighted)". +

+NEW: The subgrid attribute +cannot be used +to show how a value reflects variation over a coordinate which does +not have a dimension in the data variable. +This should be done in the long_name instead. +It is generally more informative and precise to introduce a singleton +dimension +specifically for this purpose, however. +For example, we could describe a quantity in its long_name as +being simply a temporal variance, +but it would be more informative to record it as a subgrid +method, by giving the variable +a singleton time dimension, which could also +be used to should the range of times +it covers and the time-interval of the data from which the variance was +calculated. See also section 23. +

+ +

23 Contracted +dimensions

+

+NEW: A contracted axis is one which is +formed by aggregating +the values of an axis with a larger dimension into a smaller number of groups. +In the commonest case, the dimension is collapsed completely to +a singleton dimension (i.e. a size of unity, section 9), +where all data points share the entire collapsed axis. +OGDT: The collapsed dimension +indicates the relationship of the data variable which is +being described to another variable of higher dimensionality. +NEW: The boundaries of the cells along a +contracted axis +will be the outside boundaries of the groups of cells along the +uncontracted axis, +or the outside coordinates if boundaries were not given. +The main coordinate values of a contracted axis will be values representative +of the coordinate ranges spanned by the groups. +OGDT: A collapsed dimension has a single +representative main +coordinate value and boundary coordinate values supplying the +complete range of the uncollapsed axis. +NEW: These boundaries will be +the extreme boundary coordinate values of the uncollapsed axis, +or the extreme main coordinate values if boundaries were not supplied. +A very important application of collapsed axes is to indicate +climatological time. +This is discussed in section 30. +

+NEW: The subgrid attribute +(section 22) of the data variable +with contracted axes +can be used to indicate how the data values +of the variable with uncontracted axes +were aggregated to reduce the dimensions. +The new subgrid information will be prefixed to the existing +attribute, +if any, +indicating the name of the newly contracted dimension. +Any existing references to the uncontracted dimension in the +subgrid attribute should be modified to refer to the contracted +dimension, since the uncontracted dimension will no longer be a dimension +of the data variable. +

+As explained in section 22, this attribute will +indicate +that the data value is the mean, maximum, minimum, etc. +The allowed subgrid ``methods'' +are listed in Appendix B, which will be expanded as need arises. +As foreseen at the moment, the idea is limited to operations which give +a single value representative of each contracted group of values, +without reference to any external constants. +For example, the number which exceeds 20% of the values in the group, +or equivalently the 20th percentile, is a single number +representing the group, but the +procedure of finding it is not treated as a subgrid method +because it requires the constant 0.2 to define it. +Instead, the relationship of this new variable to the old should be shown +by changing its long_name to indicate that it is a percentile value, +and giving it a new singleton percentage axis with value 20, +or cumulative probability with value 0.2. +This kind of transformation is analogous to reducing a variable on three +spatial dimensions (say) to two by extracting its values on a specified +surface. +The contraction or collapse is a special case, because, +in general, the percentile axis need not have a size of unity; +it might be a new multi-valued axis (in cumulative probability) +replacing the old one (in some spatial dimension, for instance). +This is like regridding a vertical axis of height onto pressure. +Having said all this, however, we note that median is in fact a named +instance of this operation--extraction of the 50th percentile--but we +allow it on the grounds that it is a common method +for choosing a single representative value. +

+Singleton axes are not necessarily the result of collapsing an axis. +In section 9, we recommend singleton axes +as the means +of attaching characteristic single physical values to a data variable, +for instance the height or pressure of the surface on which a variable +is supplied. +If no subgrid method is specified, the application knows only +that the single value characterises the data in some way. +All information in the subgrid attribute is entirely optional. +For instance, a time-mean quantity should generally have a singleton time +dimension +to indicate the range of times to which it applies, but +it is not mandatory to indicate in the subgrid attribute that it is +a mean over time. +

+NEW: On the coordinate variable of a +contracted axis, +the optional interval attribute +specifies the typical spacing between +two adjacent coordinates of the uncontracted axis, where +``typical'' is not well defined. +Further information may be given by the +the optional spacing attribute, +which may have value uniform, indicating that the coordinates were +evenly spaced with the interval specified (if any) and the cells +contiguous, +or variable, if they were not evenly spaced but still contiguous, +or disjoint, which means there may have been gaps between them. +The coordinates of the uncontracted axis may be explicitly recorded +in separate variables; +if so, the main uncontracted coordinate variable +should be named by the attribute expand of the +main contracted coordinate variable. +

+

Area-averaging a longitude-latitude field to one of +lower resolution: The original resolution was 1 degree, and the field +has been averaged into 10-degree boxes. +

  dimensions:
+    con_lat=18; // contracted dimension
+    con_lon=36;
+    lat=180; // original uncontracted dimension
+    lon=360;
+  variables:
+    float sst(con_lat,con_lon);
+      sst:long_name="sea surface temperature";
+      sst:subgrid="con_lat: mean con_lon: mean";
+      sst:units="degC";
+    float con_lat(con_lat); // contracted latitude axis
+      con_lat:bounds="bounds_con_lat";
+      con_lat:expand="lat";
+      con_lat:interval=1.0f; // original resolution in latitude
+      con_lat:long_name="latitude";
+      con_lat:units="degree_north";
+    float bounds_con_lat(2,con_lat);
+    float lat(lat); // original uncontracted latitude axis
+      lat:bounds="bounds_lat";
+    float bounds_lat(2,lat);
+  data:
+    con_lat=-85, -75, -65, ...;
+    bounds_con_lat=-90, -80, -70, ..., 80, -80, -70, -60, ..., 90;
+    lat=-89.5, -88.5, -87.5, ...;
+    bounds_lat=-90, -89, -88, ..., 89, -89, -88, -87, ..., 90;
+Instead of an area-average, the contracted field might instead have +represented the subgrid spatial variation of SST. +In that case, subgrid="con_lat: con_lon: standard +deviation". + +

+

Mean over time and longitude: Here, the time-mean +zonal-mean humidity is given as a function of latitude and height. +The means have been formed over the complete time and longitude +intervals of the original data, so these dimensions are collapsed. +

  dimensions:
+    con_lon=1; // collapsed longitude dimension
+    con_time=1; // collapsed time dimension
+    lon=72;
+    sigma=6;
+  variables:
+    float humidity(con_time,sigma,lat,con_lon);
+      humidity:long_name="specific humidity";
+      humidity:subgrid="con_time: mean con_lon: mean";
+    double con_time(con_time);
+      con_time:bounds="bounds_con_time";
+      con_time:interval=0.125; // original data was at intervals of 3 h
+      con_time:units="days as %Y%m%d.%f";
+    float bounds_con_time(2,con_time);
+    float con_lon(con_lon);
+      con_lon:bounds="bounds_con_lon";
+      con_lon:long_name="longitude";
+      con_lon:modulo=360f;
+      con_lon:topology="circular";
+      con_lon:units="degree_east";
+    float bounds_con_lon(2,con_lon);
+    float sigma(sigma);
+      sigma:bounds="bounds_sigma";
+      sigma:long_name="sigma";
+    float bounds_sigma(2,sigma);
+  data:
+    con_time=19960901.0;
+    bounds_con_time=19960301.0, 19970301.0;
+    con_lon=180;
+    bounds_con_lon=0, 360;
+    sigma=0.99, 0.96, 0.92, 0.8, 0.5, 0.1;
+    bounds_sigma=1.00, 0.98, 0.94, 0.86, 0.65, 0.30, // decreasing, like
+                 0.98, 0.94, 0.86, 0.65, 0.30, 0.05; // main coordinate
+This is a mean over the complete +range of longitude from 1 March 1996 to 1 March 1997 +(see section 27 concerning the time coordinate). +The longitude axis indicates circular topology because +this was the case before it was collapsed; after collapse, the topology +is not really meaningful. +If the humidity was subsequently meaned over the depth of the atmosphere +as well, +subgrid would be prefixed with con_sigma: mean, +and con_sigma would have bounds 1.00 and 0.05. + +

+NEW: If the same axis is contracted +repeatedly, +the methods may all be +recorded in the subgrid attribute of the data variable, +but only the most recent interval +and spacing will be shown on the contracted coordinate variable. +But if the axis before contraction is retained in the file (identified by +an expand attribute), and was itself the result of a contraction, +it can record the previous interval and spacing. +

+NEW: Repeated operations of some methods +can be regarded as equivalent to a +single operation. +For instance, meaning longitude cells of 1 degree width to 5 degrees, +and then from 5 to 45 degrees, gives the same result as meaning in one step +from 1 degree to 45 degrees (apart from complications with missing data). +Similarly, meaning a time axis from days into months, then into seasons, +and finally into years could be represented as a single operation of meaning +from days to years. +In that case, the subgrid, +interval and spacing attributes need not be modified +for successive operations. +The choice of whether to take this approach +is left to the application. +

+ +

24 Time variables and +intervals

+

+OGDT: A ``time variable'' is one which +represents date and time, +which we will refer to hereafter just as ``time''. +An ``interval of time'' is the difference between two times. +

+OGDT: It would be possible to describe +time in terms of six components (year, month, day, hour, minute, second) +in a netCDF file, +using six component variables of various data types. +However, it is more efficient and for many purposes more convenient +to represent a time as a single number, +giving the elapsed interval since a certain reference time, +which may be either implicit or explicit. +NEW: We refer to conversion from the +components of a time into a single +number as ``encoding'', and the reverse as ``decoding''. +Encoding and decoding are complicated because +year and month are units with lengths that depend on the date +and the calendar in use, +so special provisions are needed for time axes. +

+OGDT: A ``calendar'' defines the set of +valid dates +(year-month-day combinations). +The standard calendar is the Gregorian +(the calendar of udunits), +but climate models do not always use this. +For instance, in the calendar of +the Hadley Centre GCM, all months have 30 days. +The elapsed interval in units of fixed length +(days, hours, minutes, seconds) +between two times will not necessarily be the same +in two different calendars, +because there may be different numbers of valid dates between them. +For example, +the interval between 1 February 1996 and 1 March 1996 is one month, +and equals +29 days in the standard calendar, but 30 days in the Hadley Centre +model calendar, +since 30 February is a valid date in the latter. +Therefore the encoding of a time into an elapsed interval will +depend on the calendar, +and it is necessary to know the calendar when converting. +This standard permits the use of the standard calendar +(below, section 28) +and of other calendars (section 29). +The calendar attribute, +described in the following sections, +indicates the calendar in use. +If a time coordinate variable has no calendar attribute, +the global calendar attribute (section 5), +if present, applies to it. +

+NEW: This standard permits two different +methods, +distinguished by their units, +of encoding a time into a number. +These methods, referred to as ``relative time'' +and ``absolute time'', +are described in the following sections +(26 and 27). +Relative time is a more familiar method, +but absolute time offers important advantages. +

+OGDT: Time variables may have +an attribute time_format, to specify +a format for printing the date and time, +according to the conventions of the +Unix (TM) date command. +

+ +

25 Time dimension

+

+COARDS: Time coordinate variables must +always explicitly include +the units attribute; there is no default value. +OGDT: The long_name attribute is +also mandatory. +A time coordinate +variable will be identifiable by its units and long_name. +

+ +

26 Relative time

+

+NEW: A time encoded as a relative time +gives the elapsed interval since +a specified reference time; +units takes the form "time-unit since +reference-time", +as per the recommendations of +the Unidata udunits package (but see below concerning +the time-unit) +e.g. a unit of seconds since 1992-10-8 15:15:42.5 +indicates seconds since 8 October 1992 at 3 hours, 15 minutes and +42.5 seconds in the afternoon, in Universal Coordinated Time +(time zones can also be handled). +In order to decode the values on a relative time axis, +the application will in general need to know the calendar; +the encoded time values are meaningless without this knowledge. +Furthermore, a given date may result in different time values +when encoded in two different calendars +with the same units. +For instance, 1996-2-1 15:00:00 is 62.625 days since +1995-12-1 0:0:0 in the standard calendar, +and 60.625 days since 1995-12-1 0:0:0 in the 360-day calendar. +

+OGDT: The file udunits.dat +defines +second, minute, hour and day as units of time. +Units of months and years are disallowed by Appendix C +of this standard, +because they are not well-defined; +since udunits defines a year as a ``tropical year'' +of 31556925.97 s (674.03 s less than 365 days) +and a month as exactly a twelfth of a year, +use of these units will probably not give the expected results. +For example, 1 month since 1995-4-1 0:0:0 +is treated by udunits as +30.4368 days since 1995-4-1 0:0:0, which is approximately +1995-5-1 10:29, +not 1995-5-1 0:0:0. +Also, 1 year since 1995-4-1 0:0:0 is about 1996-3-31 5:49, +not 1996-4-1 0:0:0. +NEW: The udunits unit +common_year (exactly 365 days) +is permitted, but not recommended. +

+

A relative time axis for instantaneous measurements of a +quantity: Measurements are made +at noon on 2nd-5th June 1996. +

  dimensions:
+    time=4;
+  variables:
+    double time(time);
+      time:long_name="time";
+      time:units="days since 1996-1-1 0:0:0";
+  data:
+    time=1.5, 2.5, 3.5, 4.5;
+ +

+

A relative time axis for monthly means: +Means are calculated for February, March +and April of 1990. +

  dimensions:
+    time=3;
+  variables:
+    double time(time);
+      time:bounds="bounds_time";
+      time:long_name="time";
+      time:units="days since 1990-1-1 0:0:0";
+    double bounds_time(2,time);
+  data:
+    time=45.0, 74.5, 105.0;
+    bounds_time=31.0, 59.0, 90.0,
+                59.0, 90.0, 120.0;
+In this example, the main time coordinates are merely representative +values, being the mid-points of their respective months. +Decoded, they are 1990-2-15 0:0:0, 1990-3-16 12:0:0 +and 1990-4-16 0:0:0. + +

+ +

27 Absolute time

+

+NEW: This method of encoding time refers +to the separate components +of time, rather than to a single unit of fixed length. +It offers two advantages. +Firstly, the encoded times are meaningful and can be decoded +into components of time without knowledge of the +calendar, although to calculate intervals between them +this knowledge is still required. +Secondly, ``partial'' times can be encoded, +which omit the year, or the ``seasonal phase'' (time of year, +time within the seasonal cycle), or the ``diurnal phase'' +(time of day, time within the diurnal cycle). +By contrast, relative times can only be ``complete'' times, +which include information about all three of these. +

+NEW: The units attribute of +absolute time takes the form +"time-unit as time-string", +The possibilities with the recommended data types +and their meanings are as follows: +

++ + + + + + + + + + + + + + + + + + + +
+Format Data type Interpretation
second as %S.%f float Diurnal phase
+minute as %M.%f +float Diurnal phase +
+hour as %H.%f +float Diurnal phase +
+day as %Y%m%d.%f +double Time
+day as %Y%m%d +int Year and seasonal +phase
+day as %m%d.%f +double Seasonal phase +and diurnal phase
+day as %m%d int + Seasonal phase
+day as .%f +float Diurnal phase +
+calendar_month as %Y%m.%f double + Year and seasonal phase
+calendar_month as %Y%m +int Year and seasonal +phase
+calendar_month as %m.%f +float Seasonal phase +
+calendar_month as %m +int Seasonal phase +
+calendar_year as %Y.%f +double Year and seasonal +phase
+calendar_year as %Y +int Year
+calendar_year as .%f +float Seasonal phase +
+

+Standard abbreviations and plural forms of the unit names are acceptable, +as usual. +The time-units calendar_year and calendar_month +are units of time defined by this standard (Appendix C). +

+NEW: The time-string +codes show how the year, month, day within month +and time within day are encoded into a single number, +after the fashion of the Unix (TM) date +and printf commands, thus: +

++ + + + + + + + + + + + +
+Format letter Interpretation +
%Y Year (including century)
+%m Two-digit month +(01=January)
+%d Two-digit day within +month
+%H Hours since midnight +
+%M Minutes since +midnight
+%S Seconds since +midnight
+%f Floating-point +fraction of the specified time-unit
+. Position of decimal +point
+

+Since an encoded time is an ordinary number, leading zeroes in the +integer part may be omitted. +

+In absolute time, +3 p.m. on 5th April 1998 is encoded with value 19980405.625 +and units="day as %Y%m%d.%f". +The advantage of this method of encoding a complete time is that it +can be done without knowledge of the calendar, +whereas if we encoded in relative time units of +days since 1900-1-1, the value would be 35888.625 in the standard +calendar, and 35374.625 in the 360-day calendar. +We also know, without reference to the calendar, that the value +19980605.625, with the same units, is a time exactly two calendar months +later, and 19970405.625 is exactly one calendar year earlier. +But to calculate these intervals in other time units--days, hours, etc.--we +still need to know the calendar. +

+NEW: The only complete form of absolute +time +is "day as %Y%m%d.%f". +Note in particular that the forms "calendar_month as +%Y%m.%f" +and "calendar_year as %Y.%f" are partial times which imply +no information about the diurnal phase. +This is a very important point. +For instance, 1998.25 calendar_year as %Y.%f means no more than +``a quarter of the way through 1998 as regards the seasonal cycle''. +This meaning is the same in the standard and 360-day calendars. +Because this representation carries no information about diurnal phase, +it is not permitted to decode it to 1998-4-2 3:0:0 +(i.e. 91.25 days from the start of the year) +in the standard calendar or +1998-4-1 0:0:0 in the 360-day calendar. +Similarly, 199804.3 calendar_month as %Y%m.%f means +``30% of the way through April 1998 as regards the seasonal cycle''. +Examples below show the use of such partial times. +Note also that the only form of partial time which is composed of the +seasonal and diurnal phases is "day as %m%d.%f"; +there is no method of encoding the seasonal phase as a fraction +of a calendar year or month in combination with the diurnal phase. +Should this be required, the application could construct it as a +two-component time variable. +This exclusion seems reasonable because data which resolves both +the seasonal and diurnal cycles must belong to a known calendar +(it will exhibit a certain number of days in a year, for instance) +and so its seasonal cycle can be labelled by month and day. +The calendar-independent representations of the seasonal cycle, +shown in examples below, +are more useful when portions of the seasonal cycle have been averaged, +in which case the seasonal and diurnal cycles, if both present, +will be on separate axes. +The forms "calendar_month as %Y%m" and +"calendar_month as %m" +are shorthands, respectively, for +"calendar_month as %Y%m.%f" and +"calendar_month as %m.%f", +with the fraction assumed to be exactly 0.5 +(i.e. the middle of the month) in a main variable, +and exactly 0.0 (i.e. the beginning of a month) +in a boundary coordinate variable. +

+

An absolute +time axis for instantaneous measurements of a quantity: +Measurements are made at noon on 2nd-5th June 1996. +

  dimensions:
+    time=4;
+  variables:
+    double time(time);
+      time:long_name="time";
+      time:units="days as %Y%m%d.%f";
+  data:
+    time=19960602.5, 19960603.5, 19960604.5, 19960605.5;
+ +

+

An absolute time axis for monthly means, encoded in days: +

  dimensions:
+    time=3;
+  variables:
+    double time(time);
+      time:bounds="bounds_time";
+      time:long_name="time";
+      time:units="days as %Y%m%d.%f";
+    double bounds_time(2,time);
+  data:
+    time=19900215.0, 19900316.5, 19900416.0;
+    bounds_time=19900201.0, 19900301.0, 19900401.0,
+                19900301.0, 19900401.0, 19900501.0;
+As in the relative time version of this example, +the main time coordinates are the mid-points of their respective months. +Although they are encoded straightforwardly, +their values depend on the calendar. +If one was comparing means over these months from data sources which +used different calendars, that might be inconvenient, +and could be avoided as in the next example. + +

+

An absolute time axis for monthly means, encoded in months: +

  dimensions:
+    time=3;
+  variables:
+    double time(time);
+      time:bounds="bounds_time";
+      time:long_name="year and seasonal phase";
+      time:units="calendar_months as %Y%m.%f";
+    double bounds_time(2,time);
+  data:
+    time=199002.5, 199003.5, 199004.5;
+    bounds_time=199002.0, 199003.0, 199004.0,
+                199003.0, 199004.0, 199005.0;
+This method shows directly that the main coordinates are half-way +through their months. +The units could alternatively be given as +calendar_months as %Y%m, which is a shorthand. +The coordinates would then be +
  data:
+    time=199002, 199003, 199004;
+    bounds_time=199002, 199003, 199004,
+                199003, 199004, 199005;
+These have the same interpretation as the above, by convention. + +

+

A partial time defining just the year: +An axis of this kind could be used to record +the number of occurrences of a particular kind of event: +

  dimensions:
+    year=3;
+  variables:
+    int year(year);
+      year:long_name="year";
+      year:units="calendar_year as %Y";
+    int count(year);
+  data:
+    year=1991,1992,1993,1994,1995;
+    count=0,2,1,0,1;
+ +

+

Year and seasonal phase defined in calendar years: +By contrast to the last example, if it was appropriate to indicate that +each count applied to the whole of the continuous period of time of its +respective year, this could be done thus: +

  variables:
+    double year(year);
+      year:bounds="bounds_year";
+      year:long_name="year and seasonal phase";
+      year:units="calendar_year as %Y.%f";
+    int count(year);
+  data:
+    year=1991.5, 1992.5, 1993.5, 1994.5, 1995.5;
+    bounds_year=1991.0, 1992.0, 1993.0, 1994.0, 1995.0,
+                1992.0, 1993.0, 1994.0, 1995.0, 1996.0;
+    count=0,2,1,0,1;
+The use of floating-point years allows us conveniently to represent +exactly the beginning and ending of a year and a point half-way through. +In the standard calendar, of course, the interval from 1992.0 to 1993.0 +is longer in relative time than all the other years. +But for some purposes, it might be more useful to record that each interval +is a calendar year. +This could be especially helpful when comparing data from different +calendars. + +

+

Seasonal phase as a function of year: +Here we show the date within the year of a particular event, +such as the highest daily +maximum temperature, or the onset of the monsoon, as a partial time +within its year. +

  dimensions:
+    year=5;
+  variables:
+    int year(year);
+      year:long_name="year";
+      year:units="calendar_year as %Y";
+    int date(year);
+      date:long_name="seasonal phase";
+      date:units="day as %m%d";
+  data:
+    year=2011, 2013, 2027, 2028, 2051;
+    date=629, 627, 626, 703, 710;
+The event concerned occurred on 29th June 2011, 27th June 2013, +26th June 2027, 3rd July 2028 and 10th July 2051. +Clearly the date variable could have been encoded as a complete time, +perhaps in relative time units, +but this would have included redundant year information. + +

+NEW: A time variable which indicates +seasonal phase but not year has a +modulo of one year. If it spans the entire seasonal cycle, it also has +circular topology. Similarly, a time variable indicating diurnal phase +but not seasonal phase has a modulo of one day, and has circular +topology if it spans the entire diurnal cycle. +These kinds of time coordinate are particular useful for +representing climatological time, in conjunction with other contracted +time axes. See section 30. +

+

Average seasonal cycle expressed in months: +Data for solar radiation as 3-monthly averages. +

  dimensions:
+    time=4;
+    lat=72;
+    lon=96;
+  variables:
+    float sol(time,lat,lon);
+      sol:long_name="vertical component of "
+        "solar radiative flux density";
+      sol:units="W m-2";
+    float time(time); 
+      time:bounds="bounds_time";
+      time:long_name="seasonal phase";
+      time:modulo=12.0f;
+      time:topology="circular";
+      time:units="calendar_month as %m.%f";
+    float bounds_time(2,time);
+  data:
+    time=10.5, 13.5, 16.5, 19.5;
+    bounds_time= 9.0, 12.0, 15.0, 18.0,
+                12.0, 15.0, 18.0, 21.0;
+The first time point applies from the beginning of month 9 to the +beginning of month 12, i.e. September to November inclusive. +A representative main coordinate is given of half-way through October. +The second point runs to the beginning of month 15, which is equivalent +to 3 i.e. March under modulo 12, and thus covers December to February. +The use of the modulo allows the main coordinate to be specified as +monotonic, as is generally required. Because the axis is also circular, +it would be permissible to rotate the values in order to begin with a +different season. + +

+

Average seasonal cycle expressed in years: The time +coordinate above could equally well be given in calendar years, thus: +

  dimensions:
+    double time(time); 
+      time:bounds="bounds_time";
+      time:long_name="seasonal phase";
+      time:modulo=1.0;
+      time:topology="circular";
+      time:units="calendar_year as .%f";
+    float bounds_time(2,time);
+  data:
+    time=0.7917, 1.0417, 1.2917, 1.5417;
+    bounds_time= 0.6667, 0.9167, 1.1667, 1.4167,
+                 0.9167, 1.1667, 1.4167, 1.6667;
+Here, the periods have been constructed as exactly quarters of a +year, beginning two-thirds of the way through the year. +In the 360-day calendar, this is identical to the last example, of +periods of three months starting at the beginning of September, +but in the standard calendar it is slightly different, +since a quarter of a year is not exactly three calendar months. + +

+ +

28 Gregorian calendar

+

+NEW: This standard recommends that +Gregorian times be given in +units of +days as %Y%m%d.%f with data type double +(section 27), +unless compatibility is essential with applications that cannot process +absolute times. +COARDS: In that case, Gregorian times may +have units of time +formatted as per the recommendations of the Unidata udunits package, +which specify a unit and a reference time, i.e. a relative time +(section 26). +NEW: The recommended unit is +days, with data type double. +

+COARDS: Intervals between two times in +the standard +Gregorian calendar can be calculated by the Unidata udunits package. +Udunits implements the mixed Gregorian/Julian calendar system, +as followed in England, in which +dates prior to 1582-10-15 are assumed to use the Julian calendar. +Other software cannot be relied upon to handle the change of calendar +in the same way, so for robustness it is recommended that the reference +date be later than 1582. +If earlier dates must be used, it should be noted that +udunits treats 0 AD as identical to 1 AD. +

+Data type double gives a precision of about +16 decimal digits, +which means that it can resolve tenths of a second +for years of up to O(1 million) in relative times. +The precision of absolute times is an order of magnitude worse, +since a year looks like 10000 days, rather than 365. +The larger the year, the worse the absolute precision. +If very large years are needed and the precision is not sufficient, +the reference year will have to be modified to keep the interval +small enough. +

+OGDT: If there is no calendar +attribute applying to +a time variable, +the values are assumed to be in the normal Gregorian calendar. +This can be made explicit by setting calendar to standard +or gregorian. +

+ +

29 Non-Gregorian calendars

+

+NEW: It is recommended that +times in other calendars should be encoded in units of +days as %Y%m%d.%f with data type double +(section 27). +Relative times are permitted, the recommended units being +days since 1-1-1 (midnight on 1 January of year 1), +with data type double. +Since the Unidata udunits package can process only +the standard calendar, +an extension will be required to process relative times +for other calendars. +

+OGDT: Apart from the Gregorian, +calendars recognised by this standard are +julian for the Julian calendar (in which all years divisible +by four are leap years), +noleap for a calendar with 365 days in every year, +and 360 when each month has 30 days in every year. +If any other calendar is used, a suitable description should appear +in the calendar attribute, +but generic applications cannot be expected to be able to encode and +decode relative times or calculate intervals in the calendar concerned. +

+ +

30 Multiple time axes and +climatological time

+

+OGDT: There is no bar on a data variable +having more than one +dimension in +a particular quantity, so long as the dimensions have different names. +A particular use of this is to decompose time +into multiple partial time dimensions (section 27), +of which one or may be collapsed (section 23). +NEW: This gives a method of indicating +disjoint intervals of time +belonging to corresponding parts of the seasonal or diurnal cycles. +When a variable has two or three time axes, +the first interval of time which they cover is assumed to begin +at the earliest boundary values of all the axes. +If there is an uncollapsed axis in combination with collapsed axes, +it is a ``climatological time'' axis. +There may be more than one--see below for an example. +

+COARDS recommends use of year 0 to indicate climatological time. +We do not favour this convention. +Firstly, it does not provide any way of recording which years were used +to make the climatology. +Secondly, udunits treats year 0 and year 1 as identical +(which is reasonable because year 0 does not exist--there is +no year between 1 AD and 1 BC). +

+

A mean of a corresponding months in a number of years: +A longitude-latitude precipitation field with time axes +to indicate the mean over the months of January in 1961 to 1990 inclusive: +

  dimensions:
+    con_year=1;
+    year=30;
+    month=1;
+  variables:
+    float precipitation(con_year,month,lat,lon);
+      precipitation:axis="-TYX";
+      precipitation:subgrid="con_year: mean month: mean";
+    int con_year(con_year);
+      con_year:bounds="bounds_con_year";
+      con_year:expand="year";
+      con_year:interval=1;
+      con_year:long_name="year";
+      con_year:units="calendar_year as %Y";
+    int bounds_con_year(2,con_year);
+    int year(year);
+    float month(month);
+      month:bounds="bounds_month";
+      month:long_name="seasonal phase";
+      month:units="calendar_month as %m.%f";
+    float bounds_month(2,month);
+  data:
+    con_year=1975;
+    bounds_con_year=1961, 1990;
+    year=1961, 1962, 1963, ..., 1990;
+    month=1.5;
+    bounds_month=1.0, 2.0;
+The representative year is not likely to be particularly useful in +this case; the important information is the boundaries, which indicate +the range of years used to form the climatological mean. +These years are also given explicitly, and optionally, for reference. + +

+

Climatological seasonal means for several decades: +This is an extension of the previous case, +and of the +example of an average seasonal cycle in section 27. +Here, the axes are set up +to indicate climatological means for two of the seasons +in three successive decades. +

  dimensions:
+    decade=3;
+    season=2;
+  variables:
+    float precipitation(decade,season,lat,lon);
+      precipitation:axis="-TYX";
+      precipitation:subgrid="decade: mean season: mean";
+    int decade(decade);
+      decade:bounds="bounds_decade";
+      decade:interval=1;
+      decade:units="calendar_year as %Y";
+    int bounds_decade(2,decade);
+    int season(season);
+      season:bounds="bounds_season";
+      season:calendar="standard";
+      season:modulo=1200;
+      season:units="day as %m%d";
+    int bounds_season(2,season);
+  data:
+    decade=1966, 1976, 1986;
+    bounds_decade=1961, 1971, 1981, 1970, 1980, 1990;
+    season=115, 415;
+    bounds_season=1, 301,
+                  228, 531;
+Here, precipitation[0][0][*][*] is the data for December-February +(i.e. 1 December to 28 February inclusive) +of the decade 1960-1970 (first December in 1960, last February in 1970), +while [2][1][*][*] is March-May 1981-1990. +The choice has been made to give the seasonal phase in months and days, +rather than months alone; +hence the modulo is 1200 rather than 12. +Under modulo 1200, midnight on 1 December can be expressed equivalently +as 1 or 1201. +If 1201 were specified, it would mean that the first interval of time +began on 1 December 1961 (rather than 1960), +taking the combination of the lower boundaries of both time axes; +the value 1 is a year earlier. +The drawback of this %m%d scheme +is that it is awkward or impossible to give accurate +representative dates for the middle of the periods, especially since +February has variable length. +The absolute time format %m.%f for seasonal phase +is better from this point of view. +The season axis is not shown as having circular topology because no +information is implied about the other two seasons. + +

+

Average early June maximum temperatures for several years: +In this example, the dimensions indicate that +maximum daily temperatures (between 9 a.m. on the day +of record and 9 a.m. of the previous day) were recorded for 1-10 June, +and an average maximum found for these ten days in +each of the years 1980-1984. +

  dimensions:
+    year=5;
+    con_season=1;
+    con_day=1;
+  variables:
+    float temperature(year,con_season,con_day);
+      temperature:axis="T--";
+      temperature:subgrid="con_season: mean con_day: maximum";
+    int year(year);
+      year:long_name="year";
+      year:units="calendar_year as %Y";
+    int con_season(con_season);
+      con_season:bounds="bounds_con_season";
+      con_season:interval=1;
+      con_season:long_name="seasonal phase";
+      con_season:units="day as %m%d";
+    int bounds_con_season(2,con_season);
+    float con_day(con_day);
+      con_day:bounds="bounds_con_day";
+      con_day:long_name="diurnal phase";
+      con_day:units="hour as %H.%f";
+    float bounds_con_day(2,con_day);
+  data:
+    year=1980, 1981, 1982, 1983, 1984;
+    con_season=605;
+    bounds_con_season=601, 610;
+    con_day=-3.0;
+    bounds_con_day=-15.0, 9.0;
+The diurnal phase of -15 h means 15 hours before the beginning +of the day in question, i.e. 9 a.m. on the previous day. +No bounds are given for the year, because it is a discrete quantity, +and there is no further information which could be added. +But if the five years were averaged together, this would collapse the year +axis, and the extreme years of 1980 and 1984 would be recorded as +the boundaries of the collapsed axis. +If, say, 1981 were not used in forming the average, +the collapsed axis would have attribute spacing="disjoint". + +

+

Daily values as an average of subdaily values: +Instantaneous pressure measurements are made at intervals of 3 hours +(first measurement at midnight) +throughout the days 6 May to 9 June 1937, +and daily means formed from midnight to midnight. +

  dimensions:
+    con_subday=1;
+    day=35;
+  variables:
+    float pressure(day,con_subday);
+      pressure:axis="T-";
+      pressure:subgrid="con_subday: mean con_subday: point";
+    float con_subday(con_subday);
+      con_subday:bounds="bounds_con_subday";
+      con_subday:interval=0.125f;
+      con_subday:long_name="diurnal phase";
+      con_subday:spacing="uniform";
+      con_subday:units="days as .%f";
+    float bounds_con_subday(2,con_subday);
+    int day(day);
+      day:long_name="year and seasonal phase";
+      day:units="days as %Y%m%d";
+  data:
+    con_subday=0.5;
+    bounds_con_subday=0.0, 0.875;
+    day=19370506, 19370507, ..., 19370608, 19370609;
+Note that the con_subday axis is shown with two +subgrid methods, referring to subgrid variation before +and after its collapse. +The only point here in having separate axes for day and diurnal phase +is to show when the first and last instantaneous measurements were made +in each day. +If this is not important to record, the two axes could be merged together +thus: +
  dimensions:
+    day=35;
+  variables:
+    float pressure(day);
+      pressure:subgrid="day: mean day: point";
+    float day(day);
+      con_subday:bounds="bounds_day";
+      con_subday:interval=0.125f;
+      con_subday:long_name="time";
+      con_subday:spacing="uniform";
+      con_subday:units="days as %Y%m%d.%f";
+    float bounds_day(2,day);
+  data:
+    day=19370506.5, 19370507.5, ..., 19370608.5, 19370609.5;
+    bounds_day=19370506.0, 19370507.0, ..., 19370608.0, 19370609.0,
+               19370507.0, 19370508.0, ..., 19370609.0, 19370610.0;
+If the 35 days were then averaged together, +the date axis would collapse with bounds of 19370506.0 and 19370610.0. +The subgrid attribute would not need modification since +it is already shown as a mean over the day axis. + +

+

Average diurnal cycle: +The following axes are appropriate for +the average diurnal cycle of +precipitation rate in July 1970-1979 as a function of latitude: +

  dimensions:
+    con_year=1;
+    con_month=1;
+    hour=8;
+    lat=45;
+    con_lon=1;
+  variables:
+    float ppnrate(con_year,con_month,hour,lat,con_lon);
+      ppnrate:axis="--TYX";
+      ppnrate:subgrid="con_year: mean con_month: mean "
+        "con_lon: mean";
+      ppnrate:units="kg m-2 s-1";
+    int con_year(con_year);
+      con_year:bounds="bounds_con_year";
+      con_year:interval=1;
+      con_year:units="calendar_year as %Y";
+    int bounds_con_year(2,con_year);
+    float con_month(con_month);
+      con_month:bounds="bounds_con_month";
+      con_month:units="calendar_month as %m.%f";
+    float bounds_con_month(2,con_month);
+    float hour(hour);
+      hour:bounds="bounds_hour";
+      hour:modulo=1.0f;
+      hour:topology="circular";
+      hour:units="hour as %H.%f";
+    float bounds_hour(2,bounds_hour);
+  data:
+    con_year=1975;
+    bounds_con_year=1970, 1979;
+    con_month=7.5;
+    bounds_con_month=7.0, 8.0;
+    hour=1.5, 4.5, 7.5, 10.5, 13.5, 16.5, 19.5, 22.5;
+    bounds_hour=0.0, 3.0, 6.0, 9.0, 12.0, 15.0, 18.0, 21.0,
+                3.0, 6.0, 9.0, 12.0, 15.0, 18.0, 21.0, 24.0;
+ +

+ +

31 Invalid values in a data +variable

+

+OGDT: Invalid values are any which fall +outside the valid range or equal the fill value, as indicated by the +Unidata-standard attributes described here. An invalid value indicates bad +data +i.e. a software problem, which is a different circumstance from unknown or +missing data (see section 32). +Invalid values are not permitted in a coordinate variable, +but the attributes which define the valid range may be used in boundary +variables (section 21) to indicate unbounded cells. +

+UNIDATA: The attribute valid_min +is a scalar specifying the +minimum valid value for a variable. The attribute valid_max +specifies +the maximum valid value, while valid_range is a vector of two numbers +specifying the minimum and maximum valid values, in that order, equivalent to +specifying values for both valid_min and valid_max +attributes. Any of these attributes define the valid range. The attribute + +valid_range must not be defined if either valid_min or +valid_max is defined. Generic applications should treat values outside +the +valid range as invalid. The type of each valid_range, +valid_min +and valid_max attribute should match the type of its variable. +The +Unidata special treatment of byte type is not included here as we do +not +recommend use of that type (see section 3). +

+COARDS: A scalar attribute with the name +_FillValue and of the +same type as its variable is used as the fill value for the variable. The +netCDF package defines a default fill value for each type of variable, so it +is +not necessary to define your own _FillValue attribute if the default +is +suitable. The purpose of the fill value is to save the applications +programmer +the work of prefilling the data and also to eliminate the duplicate writes +that +result from netCDF filling in undefined data with its default fill value, +only to be immediately overwritten by the programmer's preferred value. This +value is considered to be a special value that indicates undefined data, and +is +returned when reading values that were not written. The _FillValue +should be outside the range specified by valid_range (if used) for a +variable. +OGDT: In cases where the data variable is +packed using the +scale_factor and add_offset attributes (section 34), the +_FillValue attribute applies the numbers as packed, so they must be +checked against it before unpacking. +

+OGDT: If none of valid_min, +valid_max or valid_range +is defined then generic applications should define a valid range by using the +fill value (whether defined explicitly or by default); if the fill value is +positive then it defines a valid maximum, otherwise it defines a valid +minimum. For integer types, there should be a difference of 1 between the fill +value and this valid minimum or maximum. For floating point types, the valid +extreme should have a magnitude which is half the magnitude of the fill value. +We recommend a factor of two, rather than a difference of one bit, because +it is easier for applications programmers. There is no special treatment for +byte as we do not recommend that type (see section 3). +

+ +

32 Missing values in a data +variable

+

+OGDT: Missing values are not permitted in +a coordinate variable, so this +section applies only to data variables. The missing_value attribute +indicates a value that is used for data that are unknown or ``missing''. This +attribute is not be treated in any special way by the netCDF API, unlike the +_FillValue attribute (section 31). +The missing_value should be outside the valid range +(section 31), so that generic applications will treat +it +appropriately. +The netCDF data type of +the missing_value attribute should match the netCDF data type of the +data variable that it describes. In cases where the data variable is packed +via +the scale_factor and add_offset attributes (section 34), +the missing_value attribute +matches the type of and should be compared with the data +after unpacking. +This standard is unlike COARDS in giving +a particular interpretation to the +distinction between missing_value and _FillValue. +

+ +

33 Compression by +gathering

+

+OGDT: To save space in the netCDF file, +it may be desirable to eliminate points +from data arrays which are invariably missing. +Such a compression can operate over one or more adjacent axes, +and is accomplished with reference to a list of +the points to be stored. +The list is constructed by considering a mask array which has just the +axes to be compressed, +and mapping this array onto one dimension without reordering. +The list is the set of indices +in this one-dimensional mask of the required points. +In the compressed array, +the axes to be compressed are all replaced by a single axis, +whose dimension is the number of wanted points. +The wanted points appear along this dimension in the same order they +appear in the uncompressed array, with the unwanted points skipped over. +Compression and uncompression are executed by looping over the list. +

+OGDT: The list is stored as the +coordinate variable for the compressed axis of the data array. +Thus, the list variable and its dimension have the same name. +The list variable has a string attribute compress, +containing a blank-separated list of the dimensions which were affected +by the compression in the order of the CDL declaration of the +uncompressed array. +The presence of this attribute identifies the list variable as such. +The list, the original dimensions and coordinate variables, +any ancillary and boundary coordinate variables, +and the compressed data variables with all the attributes +of the uncompressed variables +are written to the archived netCDF file. +The uncompressed data variables can be reconstituted +exactly as they were +using this information, +except that their original variable names are not known. +

+

Horizontal compression of a three-dimensional array: +We eliminate sea points at all depths +in a longitude-latitude-depth array of soil temperatures. +In this case, only the longitude and latitude axes would be affected +by the compression. +We construct a list landpoint(landpoint) +containing the indices of land points. +

  dimensions:
+    lat=73;
+    lon=96;
+    landpoint=2381;
+    depth=4;
+  variables:
+    long landpoint(landpoint);
+      landpoint:compress="lat lon";
+    float landsoilt(depth,landpoint);
+      landsoilt:axis="Z-";
+      landsoilt:long_name="soil temperature";
+      landsoilt:units="K";
+    float depth(depth);
+    float lat(lat);
+    float lon(lon);
+  data:
+    landpoint=363, 364, 365, ...;
+Since landpoint[0]=363, for instance, +we know that landsoilt[*][0] +maps on to point 363 of the original data with dimensions (lat,lon). +This corresponds to indices [3][75]. + +

+

Compression of a three-dimensional field: +We compress a longitude-latitude-depth field of ocean +salinity by eliminating points below the sea-floor. +In this case, all three dimensions are affected by the compression, +since there are successively fewer active ocean points at increasing +depths. +

  variables:
+    float salinity(oceanpoint);
+      salinity:axis="-";
+    long oceanpoint(oceanpoint);
+      oceanpoint:compress="depth lat lon";
+    float depth(depth);
+    float lat(lat);
+    float lon(lon);
+This information implies that the salinity field should be +uncompressed to an array with dimensions (depth,lat,lon). + +

+ +

34 Compression using a scale and +offset

+

+COARDS: This standard +endorses the use of the optional Unidata-standard attributes +scale_factor and add_offset +for data and coordinate variables. +These attributes +can be used to provide simple number compression (packing), +to store low-resolution +floating-point data as small integers in a netCDF file. +After the data values of the variable have been read in, +they are to be multiplied by the scale_offset, +and have add_offset added to them. +If both scale_factor and add_offset attributes are present, +the +data are scaled before the offset is added. +When scaled data are +written, the application should first subtract the offset and then divide by +the scale factor. +OGDT: This procedure is concerned only +with storage. +It does not affect the unit of the quantity. +For instance, a pressure variable with values in the range +900.0-1100.0 Pa +could be converted to short integers in the range +±20000 +by subtracting 1000 and dividing by 0.005 i.e. multiplying by 200. +The units of the compressed variable are still recorded as pascals. +

+COARDS: This standard is more restrictive +than the netCDF Users' +Guide with respect to the use of the scale_factor and +add_offset +attributes; ambiguities and precision problems related to data type +conversions +are resolved by these restrictions. If the scale_factor and +add_offset attributes are of the same data type as the associated +variable no +restrictions apply; the unpacked data is assumed to be of the same data type +as +the packed data. However, if the scale_factor and add_offset +attributes are of a different data type from the variable +(containing the packed data) then in files adhering to this standard the +variable may only be of type short or long. We +exclude byte on grounds discussed in section 3. The +attributes scale_factor and add_offset (which must match in +data +type) must be of type float or double. The data type of the +attributes should match the intended type of the unpacked data. (It is not +advised to unpack a long into a float as there is a +potential +precision loss.) Users should note that Unidata may provide a built-in +means of packing data in netCDF files in future. +

+

+

+

A Attributes

+

++ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
+Attribute Section(s) Description
add_offset 31 34 +Additive offset for packing data
+appendices 5 Version number of +these appendices
+associate 19 20 Identifies variables + containing alternative sets of coordinates
+axis 9 16 19 + Identifies special spatiotemporal dimensions
+bounds 21 23 30 + Identifies a variable containing +boundary values
+calendar 5 24 28 + 29 +Calendar used for encoding time axes
+comment 5 Additional +information about the file
+component 18 Identifies +variables containing + components of a variable
+compress 33 + Records the dimensions which have been compressed by gathering
+Conventions 5 Identifies the +netCDF standard
+coordinates 19 Synonym for +associate
+expand 23 30 Records coordinates + of an axis before contraction
+_FillValue 31 Indicator of +invalid data
+FORTRAN_format 12 Format for +printing the + values of a variable
+history 5 Evolution of the +data in the file
+institution 5 Who made or +supplied the data
+interval 23 30 + The typical separation between points on an axis before contraction +
+long_name 12 14-16 + 25 Long +description of a physical quantity
+modulo 12 14 27 Arithmetic + modulo of a variable
+north_pole 10 Geographical +location of rotated + North Pole
+positive 16 + Direction of positive for a vertical axis
+production 5 How the data was +produced
+scale_factor 31 34 + Multiplicative factor for packing data
+spacing 23 30 Indicates the spacing + of points along an axis before contraction
+subgrid 22 23 30 + Records how the data values represent subgrid variation
+topology 13 14 27 Topology + of an axis (circular or not)
+time_format 24 Format for printing +a time and date
+units 12 14 15 + 25-29 + Units of a physical quantity
+valid_max 31 21 Largest valid value of a + variable
+valid_min 31 21 Smallest valid value of a + variable
+valid_range 31 Smallest and +largest valid values + of a variable
+

+

+

+

B Methods of representing subgrid +variation

+

++ + + + + + + + + + + + + + +
+Method Units Description
cell u Value is a property of the whole cell (e.g. an integral) +
+maximum u + Maximum
+median u Median
+mid-range u + Average of maximum and minimum +
+minimum u + Minimum
+mean u Mean (average)
+mode u Mode (most common)
+point u Value applies at gridpoint
+standard deviation +u Standard deviation +
+variance +u² + Variance
+

+

Units: u means the units of the quantity whose +subgrid variation is represented by this method. +

+

C Modifications to +udunits.dat

+NEW: The unit unity is defined +as a dimensionless constant +equal to one. +

+COARDS: The unit degrees is not +permitted, because +it creates ambiguities when attempting to +differentiate longitude and latitude coordinate variables. +This unit does not appear in the current version +of the file. +

+NEW: The units calendar_month +and calendar_year +are units of time, but cannot be converted into each other +or any other units of time, +except that multiples of 12 calendar months equal integral numbers +of calendar years. The units year and month are not allowed, +because they can cause confusion. +

+

+

D Long names for +quantities

+This Appendix is not yet available. +As well as existing as part of this standard, +it will be made available +as a flat ASCII file for use by applications programs. +It will contain entries such as: +

+

++ + + + + + + + + + + + + +
+Version long_name + units
1.0 depth below the surface m
+1.0 height above the +surface m
+1.0 latitude degree_north
+1.0 longitude degree_east
+1.0 pressure Pa
+1.0 soil temperature + K
+1.0 specific humidity + unity
+1.0 temperature K
+1.0 time s
+

+

+

Version: The version of the appendices at which this +quantity was introduced.
+long_name: Case, spaces and punctuation are not +significant in the long_name. +

+

+

+

+Jonathan Gregory | jmgregory@meto.gov.uk +
+
+Robert Drach | drach@llnl.gov +
+

+LLNL Disclaimers +

+UCRL-MI-127703 + + diff --git a/GDT/1.1.ps.gz b/GDT/1.1.ps.gz new file mode 100644 index 0000000000000000000000000000000000000000..5cf6318bfc3ca53b69c739636a07a854c7f6ad10 GIT binary patch literal 147763 zcmV)pK%2iGiwFqAM1eH`12HZ!E^u=I?EQO>Eje;02>x%MA{L7VX0TIA`tgXMhoHUa z8y?PqfNiZtB?tuwjX|pNW>rp?GHWuk^w>mw_jM=j@iUU{$Ipo?snp0*l{d_cbUJA? z8qJIk|1bZKzy9SfK7RGxKYaSd<o0%v>4!gmcpMr2|9=0~r}bSQ zfB5yMZ@*e;uiyRQ+poU)_BXHJ{mY*}g!O+=I{P-OZtKS%iG`m0kcrx1YXT>+$f5@94k$zv_yw4|@u~y*-^y8=R*Lr;Q`)~f=&msJ855MM$JbeG@AHVyjPhWA~Uw-$8?|;44<>7BX{p#Vn zKm7RnKWvPAnf@$1|JiW)^Wiy6TjM|d=9_Q-`up#G{mV~3{`vLl(hr}$|EEviuYK_2 z!#`|~uI&2sFW>xF$n?XHU;g-qA1sxA{QeK09)9@rV=3exzx(P_VIIG+slboxn|}ZO zcl59?(z%TsvB<$d+(>jxp0H-GV0Kdr<+ZV6w!{U}^+rQcB19`FR zQYb(DFg$$u)mPio^tXTW_y73ikKcduFAu-|!}s5>QdqE`*8YCj{`-C#TMs|{;U6Bp z{D&VN!k>lpH$43A$DeY-`ocfb*WZ2r)89lm%sb^)2FX~0wMgu$R+bS2sUZ>ufP88_uqW_;ioVDp9OgM)mBLQ zZ(siS_e1WYU*tdkeSK+qef|5u-L|6NZW6!#{mZ5N{q>LE7vQU}fAQm&-~8?;+OkYV zX8-D8g;`kBci;Z_xBvS4PY>gkd&B-)Vcgf>{`mdB2<>%MLJ#XJAO30`OW&_b>EVli z_;T%>Up;*J{r6w~>niBqtk13L;a4m8hi`uO)%@EpzyE3bCu5(!-}tnA`|IyEh4!#k z@~`OAHZK0~`=7q}`n&J8q4@Rsnm=B+ytVo3KmWA7|99Vg`PJH#wZ@yKkcR#J#;&iv zUYqs9*K5D8rI-3{yjn4Cb;&nn@^JQx1YB8 z0&_la&c`a}zpP5`CmJr(LnrennYX|Ez`t$}bn;%dJhbH@cVz>{?ZrU|GMD6`8~@h z|NY;~pWG+wUyz!ge)#gAHpv{;vA2%JUuGT`$=lw0d03U}r+>-tFHHsuS}Mj({nf+P zpE@csA1FL9!ylNB5873k9gv!PIya89t`p9Wzy9*qzy0*pr+-=}CCg()Yy;bd?Jxdp z&;q&eT+W|m*8lgPgz10G|Ni5btGV(0ryo}9}f?DfQd-xT>Kug07; z|JK}X&}?+ioGs;-4|(XUPqS7eCkz`=7+koyX5R+VrmNPM>1f1`U#t3;KhgT?LU`K< z|M1=KfBM4@JUMUJFMj*ocmLzze|-DhKY#m^pj(Mw{PHg!a)tN{`AvI{^iz#MN4lC28~CswU;XWt z_}%w^%w-odijAW`vQ@2CJC+NDKn|izW!!a*uMc4_T^8d#5^Z|`rWFOvoLHkE@chG)L(6d zSl0ixTFToLyd`bJSZ3&d$aKMe{nIZVAO71PX)F7uUyLibLZ5#6#ic-BZ^mwhezS^4 z27db!Q@;6$GyPvb@pJ#sG!ISa6$@nrnO_yo)C7;GAysQpoRxQu`-~ZA&s{bn$ZMNEhU#|A!FswG@*MAY> z5{8tfF{L@Cv~2&Sou8!BmU(+9KfM0$d7RSY#}ALs4{v|5+23p4CBSdUk0U8&3av%> zJeX^g_f8FE_T{Vvr6eS)@6=FcqfDhZplp8#YQUt?$`+`lM15N<&(rqYB-Ng!iL*3z zswn!aShp6`kb0DOl|pASHI%XLxx#EDH652~N~|e^_L`D35AxqrM!~d>vEbM=%`cSv zu254ZM;Vj!nv$Y+4N!a^gB%#8{H2zJl=(D;jx!mr_9=-A8j>|c8Co?hrivprmPh6q zMIBq57_FOzA<+++@Q2Z)N^wR{v^AROG@l!A$X0IX37?zPlFd68a0Ano*MaJWz!%YV5 zoGGMrDWg9bQ&>eL7-d^)U`F7O zRCiAI`$_B?Mtg~s=~!FkSesl^uW8@$`;RYZFi{CxR}rdZ5tV|))7bVLdDO7MVUl4V zqw`shF*=|17^Cx9k1;x*^%$eiQ;bfu?=6O1z4sUM`MV#?5iy!YAy$~vrGSw~H|(ew z|HAQc)qjs~)2jby2u}~r(w?D;L9C+f$F@BfNJR^a%b0VOH;nQorD#or@`e-x#6lK*+P$f3eVVWK`o`N(PJD}%kXu{>Sz#qj0brcUZWJk z!#Mv&Z(MGex_)GqswPHYmM%rQyt_fkz~W4SUZ4wyA$E8A7lyzEYU zvh_S-m8YoK76_lLs3D8A7M!uAk42AB9t$q9!X_-H!l2_Iem&&6fz>)$vlCD2>g@z! zePIR!=ZZA!VbhAeDz+8;Vj06$E@@iGoYX#!3&zFbjI1A@ABXsyLv+LNBr&=wRz5i2 zr8a|lI9j6uQe zFV%#5|Lzx*wP3u0TU3d1-3KFx>0&cZs7(Cg(x4NNhX^x8$Ad9RhK81y63#p>yDX)E z8XJ9nb2EUz9WXzI_%?6W%wPTmi5q2W7=DRFEf|f8iQ!>M8--d@OKA&gDWtE(HH33D zNf$??@t}2Vl77c}%gl+@h9F@IGH4-GXd$JRHdrDA8_Qopq6#L|dt87fMvl^~+7zMj z8nx=skZ~F$1qMq5bO1jpDJ>vv_La73k7A%IY(e_Kn_HzVe$bX``yFzW>+Ne@>h1SAlxp{A+z?Q;8_}zV zT<=)XwrbZl;wT~cy53a<&(KKgSk<~ui?u#eTD4(eY$|w`MzX!FsdqyQZ3TgbHLBTK z(VoJphL=?hKYjSizu;=g+eQM8fkq!evBt8uoNe8_JThyYM~T3&3lr3TD~mFP0qlFs zrMOw!8_OCBuiPxf+7~23f`x<-P6`ht{_S7_L}!YVFfcqBuj_;!9QWag5D|Fn!U@-) z4IXk#mk0l_K&))Br{ioZ*s|+3h`z}J%`v|rS+q!I`5-)Wiy}GuWxk~N_VH~TKMY~W zA^pgbpmmS3fHsTYy!Wr>q*)#wTJ3ww&+XqWyCwGj<-s}h{-wQyTmSv{D<~cQhUxKn zwZxt-fj-DSjtIMKixQAPX~9Gpq@;_o6Js%DOMhn62KTrv9tDYzW1;yoCBo?_M!@_CAph1_=+GfTkg{$h$aO(QSk3wkRy#>-orKdwf? zDv!93ZzJDnAg7zeN73NmCak*gN`{13^Hb1SXq1QDKMveN`qSFFDDI+9UN%^sXr@blLPi6ar>BX}3E=xA6!|fkp%~&Frt5zI1BKx`w zWd>fgp+H=Yhl*|+O7Ev)G%e-z1sWC%73@IfGt%54-UHF+DW>%PXDCM2ymYoY16?)^ z^!9PSEUN-~%(hG}Zqx}d+L@+F0}fNMQN}V|7E#@x66K$^<-k(2i7oIAl4XPr6elt1 zE2zX9G6?c2&bKN<(~G0(M6=yDF`^_xN}BG|6D&4DJi9PjUI|T*rp=md3OSbM|4z*vk}bV4bst<@o|-r*JYE+yy*7php1<| z155ra+1VzJLwWO-kOx-|28CgxjB|;b53BN-V}7u}#o*p_TYM`H+cd6^y-f-yZo(8f z#;_T|)3XR{1usFPs?(eYYI)fl265R-#!$A_E0PB=t4o0ko0dMTqkb!>AgVBWaH&f| zm@SD(VM1gq1|Gt73w9eoH@0B60d%?rV-pu05hj-qx^Ucs7k~+nv3k%bgqhc~XoU)n*Q$#x8yyWN9yWTqaJ|0U=g2OV2a?c3<= zi?6+n3_N(m$?667fpyH*-s4AEgkXkHCZ)Ki6t0hc`7KSeTNWo@)#Kg*D8awl-u~{V0 z+qI*$N&kwcc*4w8yY|4!rH;|N)qa=|8`W#v{vAUHBSxcZ;WJ>Yq&XIc*d{*UbG)Tu z`;}GR6&n;AUs9=BmW;GMF7D!{#N+1M7bb1Y)?gDmnYkK64vM~x(}!*E@7u!`mH2AA z^EC%0)MH|QO9M>eWF`L#^#f}y9^>@#@$v0LfLohz43;V_H&EzIrk@M?pQWH!2^meb z*rj$gd~P0_nxg;XoAz@-|7HcX-Cv6cos*i!%j0D=podk3ZC7Qs8K|P%CXMj>kH2l_ zPpzPznf>or&}U=zZ-{D{SMfz1($Odn=AlJXyI*q*a8=0A_*(!@2NorC3CB7p- zM6CMjO!7Ls>|)=9`6VKv9z{XNon$7D;=4)|B|E#N@!Q{QSC1uvn*Ezl{^rLQUQLh- zlNLD$EMvm{u1ueq=}%>+lZF#?yQ45Y2UBv`7Pz0`&Ck%k#q-V%5c~}@@@htZ$}{hP zIU0??F$oJD7!E*h4l<;w7onGn9K|G8MR<9s|H|fp_3tykAWGB3=Ned2RdR)?9Hjw? zFKet{=d0mY#nV?~9=?1GWBm9sjB)c9afBRmM5&ml-$-v5W z4zN0G`M0+7A1FSbZ5*DdCHqFU}EF#LQ~{RzODSc%X8b61R)u|=sK2!YHxC` zpuAHC;o;!@AQupw`cUeD|LhVBc1HTCfZf&^z|rCs{crPE#l zbmkr%$pd|+-q)?XyAPa|3fdQ!+oZzINcA5Ws?LdfblExEdEhoW?TE##y|33zoqk|X zDrmR&Z2$T1Cr|J6!De=E+xqYjN_kjTw0SDAT5eO6 zOh@?AY4gO#;z=x3!XABMI?fW?)I|BZf{=W|Py&;%-9zZ*$kBj(x$RZomJ>F1xSN6F zR-3X9(M|d>pN`mOF3-8LDUr$)k9|2L+n-qvn~5Jpaxu8d#%b3xPcKg&r_22G`1)ZO z;HFMCEBw+%5>j;$&Qj;`y*B&))veewEO77w+G6EQkT;kl*xRCR&{=d5S$06x`~JF= zS>H9=U#dux&t#UADlA=O^%{m+(pgGu3#xR;T@>>3d0C}sChIGdd8eSDoE?#~EoFu_ zQ%eF$JVy~8jh+NWrc;y!3PO^-ci0iWiXtvzO)|=eOlv8^v`Gr8NJfe0D1i$afHb;c zkg|SjC{>qJq_mSdh({EF~40eK2-q~J5XdP>?TOGO!L z3Z>E4F>P1iRMYhA*V4C%scP!j9<+9js#d@gLiO?x9w*|dXM~MqLEi@V2iUA&s#abY z?WAqGlJ9D^GG=Y~Xg-^S<_79LK;36eQ_Ky7qt;ZM?y&=PWu>&`rMA4mX<8_{Em@%0 zeFsD2UeZ8A{o)-6ZZ7Oq;bdhpL^T)s{5Xcxs_AsZ%@BQ_VzO^+@_Pi0T}YY-8>)xY zLcNnFcp^nd&S+!JUIXzID-rJR&~jj5O$@e6e142!{&-0r9`!>RA!=^KJBUE;5%o3exSVxB1kE{{_4z2c1dhU%6d=iO_sz@9OknzB8hOO zjS~q+`DI$Qz+2V=`Kas?ghP5e#FR~nnf+e-ms*niTOQzt$5hs9*pLkRFYkXVJHCsq z;+?9NU-DQYlP7IuQZh;LW&7-Kxm;qLmTmFuan*2gj56;0P5PIRurz;V;b@q{#~5Ee zUS7G8+Z})z_^D3zJO=`@SvyV}#8k|~ebb(FIFTEz- zB>|rYIiY(MV zf2%IUc>MVI_(<`29>&)XPaKo$=|oj=I|?xWOYEJo*a=JLfcp8&@0JZAtP^#3ap%Pn zsY|k0+5U@7nruk&P~QH1zRdA4ugucZdR237uTmqMu~)zRx7r4z<$3#ykELP*ys zL*K#nl(C`24y7q5HJKX9&``$qZnd#iKt`xGO@d+wABO{KSep-CnciGjUk*Wvebwf)?s})J6*)grDld+~9K-3UJC!;ZGQmCR) zMXcD>rP;VOp@MnVhzV)6uJtT5E|6&1mWQL`wK$xpG;yjRS2mP~qN&S7@Y8X?Ke)uy z<+D~cG8|h{c>G9@kE_ldp9WmVkd7M7x>KvkBz+LcA9$$2o9BS>MtQ5BW%G_wTOca? z1-SRI()xhCcJ_%~k)VhYV~jD5n+gjR&1n+Y3IkJ6qum2&CNpou58K{hI}gO2GZXKC zhVLjes;p{t`oJ8I;`()IOt7%AHyoo(hD%LLo+iNvygemIW#&6*vh$$H=v#7~=6t@` z$>>|WPouLLJ!@JToz2*%nJ_@pG#X^=8oIPrlK7COC+bB-juskP^fV5w>wE^WpLRgO z=eU$6R<8S;!r06kW>&lTzt*Xzh(| zrkwSj#m0Prt~Dt3X>)ChNQiBpX3VYT1OnpN$+$zqxLEO(bVLq>6CnW& zT0B)Wc?6APOTR9L4UKvOQ|o0ktAnx2BEHdXSsih#_C{^bIl;G)CnkKIc391is7lEc z-yOy=lB^4

3ZQCUM(*Ls^IF)H{A|#r)yA<(OOIAfC5}ctVz@Hr`v4FdBRyg-Qsjd|4J}b;cw4Jn z5M-{30ZFs@5!P>zolbf2vrft72*4a|^lN8=HWbCaS=T}`O$u8va*@;wtkI7(HJD#M z#Gwi@&j3mH^SceYmF^0fQLMFfHPol;UcpvhAJG$b_5ua935qond7p0fdLuAb=v54n z_7d{XM+*<>q-dz1BYLL~XfHvqvwfiYXph%t`=DMqIL&MOK|wxY+IgTdXDT}FB?!HS z_ULj-OmmE<3}USx0p_{plVTzlP83*j~{b@j#v<-MdHCuJGe&ty?-tjEAQcT6eyON)(A-h z)LncqoGXN(tGsj|L2B{UJzmGKlE(+}`=rUE#r7oLBslxr3*w`dlW|I(5bX~RW<*&P zRd}_n$`!D^_pR7AkIx_Cvr{i1d6{7GbuZVTNb0nqn+A&+dk>CUKY2X{4xMx46Me-a z6yl>CK5dhuh>7_2byiDf(L#H%(tJoRSwb; z?w9vzL!!#2DX}gdeQXnS(#FN74cpu=iE$29M;)cjnpW%)T6&2?!sVP$bQGl3ofzc9 zljG7PPEvPHn|o>b+Gf^A(ixgqSwh#+0;f^A3$Tgi=R`S~<2s5d^0_NXBTUSp`nnT2 zEfecbgft3uN0r1}(WVninp(QpPgEuPPAsgHPZC!hmD6%XM`vht5`9N)q&HdknCmp^ zCPsNsXQtg17>tYKBrJ*bk_-mft4=KJ64S4n*rv%8sB@(zktNPe;-C-Au5hjyrxn8h zyzHpcj!0HLMQK`U)A9tFmVRPRo9ncw^Ikp%7@?pLo_T71es$YLgljg{L4gC#qy(8< z!ZCdTx=yq(%?^80Y#oJCy2>o=CGfk&PrOPOpGM}Ixq6_bmnpUs0Cz;tUII_8Vhm2# z=-Xod3`<8C+1sN}xpEe3c#x`kUQeMzryw7K%?U2zXNIlkb)iqb4=5*$ zZ9)+De6IUoQ8j^dl5=e3IoCPC+VqInaL$o!09~ri2}WNUdTs*JVb4$+>dxV+zSMo7bm#+#jBRy&Yj`aqyrqYluG{lq zD3%9iX^Nwddvq;mtRI+w97m6#OTr@Lvb{{WuDPrv{>-?&A-$%Vebe*~_yM5O5?k~%pZ4Z~4OP~GBaxLMo` z6bxm=&a|I^rWRz6s9=W%0hs}5)FT2doS@|f)SE_divbLV3%+c!2uhn&;$cgffpAs~~|VH9@j&V619^ z{?Z|h9C7YstZ8+KkZIJ-IGvzXF+}PSl~z}ru#m3s(9w=qs5RQ*VMX(c zj#^quutNuk*J*x@k!$q$k;8GzhbiE(^H}#lOjQ*jW>)Rcan}0mdWl^tD4mxdELLNl zRqFwWFQtnXR#I|FWbX<#7mlbAE0C4awI!3&41SLM_R)BO_-h)XijanVXqTQ0N{O z`F<^JqKdRMnJP-GmJWsD3hGKxaFjvCabgRXrI#W6uT;}%R{RLaSV%1ghipWB$zEJ`v zzvSnbms+X(ht~nB_+At<+bqrr@xgF0G^HXFl%ET#31V>y&VQC`I5~nXhF0OqeVivv zyK2C7SS>BJWYlt*+|3t~!MjjSC|!gT@C7YDd}63@AI^jl;sWJdQck$WbC5XZ`DxoDavT8gauH!}VMb5w# zf?kcLxENh-oPf8?+4B5oCCh>3D$Y|y6+oq=a!s>5_<)P$;P13MC+g z7Ktb&$fRa(OA&#pjt!UF35ztDlvO(@9A#2U07|~T)gP@HQcIcaCT@Ow(vSU$r+~*y ztBI;hS@tLg2eG@9d5?nUx_qWSW!j_ILU^62PYHVzJY?)M-Nquk&*5DxLcj4{EW#}w z?_v?0qj(>S+H(~0^OQEc36zNoN)@R`X$yLMd>o(Ce0kh%AJy}U(qQ&mw21Qjd`K(r zp_XX2D7Fq(`C#%jF4#~r3(g*%8`{vNP0h}JvlxLEk-q{i6^-D(DY+(o- z+0G>b?y6^<)q1p<8X_DIg?MI<7LZmHBz!gA)N+qDJDM2|)j31IPARltza)-jOQX27 z#M<1f7W71z+_&i^Lg%kZEHr=Ze~}0GUZ!~`tq}c-g*H_*cN~prl&Z>t zh=Z=^Hs!ygqfJUPry)36*-m2~g2hu|sYxA#siEn3#IU*xrY2t=(5hpftmUXB!k||g z^z@j#Mq?W7YBW?L(v}{L-{GeIMth}DDU3pE6jibnb!Wul0t|_;Z*265e<#iWsoCOEgaWlTpE|O)BlzPPKx1*G54Y{YO z8rtz98F^w|UXRvaB%^M|h7H$Q*qkzU85`E@v#>r#QOk&0v)5?UA$AIUoF5;bUyI6p zk!3FTr8;t+LYesF+Z<)ORBgptS29kfM8`bS!jduvAo;o3FbhWVbCabpQI`3lxa`Z` zpP~LUyFa9Xe5H==kHz}APE#MVC9J3wp0I`~T)~ny?OL1ipY2=F*0i()Xkm6i$fI)? z8x$K;`SzBzx7>@oNYlmyy*8CAk{8yfap9O%6l^PJ+;;O#8jD$(Eov-F4TLCog=Z1D z2{I9k<7J6Yt3mMk;U!YJ7TtLv7SY>>F~7vY{IX;PPdzwxVDYX3Bq*GA;iZB7nrs8l zJ-FKdKBYVfm&W|%FpVxMYcTqp`DmjqcjIh_i%0Er3BM(a$C=noFvt+53C7z5qilkq z*1%H>)8vL&1&0GTu}9A^O*W8H=)i-A7Yh(Oc%~3gS)i(D;mGWo0G^qFWJt+6gBqND zd|AJ_p;gIQ3p=>iz|J)JcfN#1qQm9o@i{)tkE1R>4RAFbdMeSZ8Z?Ds)*KSgP{*>g zoj}#8DqjRH!Vm?X1Qxy};}wDAC6K&WEJ=oEP~%g7h{%}5g_y%Zt`cN$(1P{{`SE~9 zn9gL{)@98rP%#LirE@_eAUt?Vo}_CEqQZ?k8-vAQV*7wt#FF)RL`|yH%M0sgT)VX8 z3cKk-lj0T3`yv=tMV;8e<~Z9%0^aj1EhOm=EnN-jwaPFy@a0gUR(G z7_8jO1r^p&VvccetTx9w)>xUL%)gb%<2j+fH!g(VMZs3HBjtPO?orBbRs&My+8G?!l_q z(>oaDO^CY~g+nB~i&6QnMZ=)p$C$P$pB-1jZ_v3^%6B7lsxc5z_^x*H-_%Z7CpZj8 zFC2UGYG$2G;GB}PP@8p`OTHE)(2h zDSdVjCrZX}FlB6pggh4XV`jrC<0>?N%v2(eGFR=-(8Ku8h?v@a2e#oU(~u5q!{$%h zH2Z^1JFPDu1|a($Ar1m9!lVGa-X6R>4)G&c zEQE1!@2#dQn17B?8yrvc$L=Y46tS>}bvxHKE?u9K`xsZ7JbJ2y=-W=B&PbFvI1jJZ zbtM`lEGe;Fc`XaRM0~geOO&`*H)|w47#pE#y0y>O3TDGfm!yYER79Dz| zgObxhwIU{HQ^zgOH_xbgP0I+4$w%t7+N5YKiJ(^py|2bFsc8A+W5Kph66-?9hQ=iT zrvCJX4TggI0>r)^Tr-4^rpZs6TYdVV*Vw+qTzlvw?D!oOk`4Qdcd);Xqcj@p94q}~&@V5tCfH^H;+V^5e<%6NU;Yg8w**z}Z~n$-5x*A(v&%cK{!qizPr#|ngv{NifEqsBB1Tt!pSvKK zr$%Tb%-#c-I3vUWvH&g;7BPpyM?g%PBEkzu*UWuDh?H`A@CeZ+z9LHd!?{?!l?jHY z9vmCE<61sXlxOM}DW`|Q!vVUcV0=x$i#L<@PZXu)W3laQa~g|o2)PUq1qu`va35my_E09ETJWoHE) ztXYtH`#62s8`T!ZHj6Ek9_r*2HRhAo-+7H*ZVQ~Sp)kbqT9qF zLhTsmVuj>I_Ji!C&;Ju2>}*!m?(J21oDUv>r>o5|Tus>ld5 znsg2P9ooM${xvh0l(1*C^()Z)r8_Ioyr1r@K=Xc?LvF1=zOyiEJKd=XEQX&B1y2og*QXP@NpHQX?`4*EJS+8R^SknMT8bB%*Jk+ zig=I7i1(arq^XFeevmmebduUMYRT}?MXJ;`-KZ32jO z6`knQ2@`w(!=~hhz9?N;Kzj)|DSO`7e2jFe5}>?spYCrJm8e)uviaM0j`({uZ~F9d z!mK1zypj3jh;DsAHni)`k?E%8b(v{Aoov#nc)iib9$h9PPbaWBIX)rkKqjYP+qyz< zVj)o!G5gt*TA1z#rZT7^Xv>|Bm@3sU zRVraB$FN*Tu2uo{(Dy^$MmoVrtA=Df;_8v+g@qdFsimeFVU;jQ(`-cfpko&Yay4LN zgm4RR=Yv7|vT)beJ%%E#jGA7HKUJ8O6&d)qyZt?`J#$zS&bC4dw`q z;)aUV3imWFDF0$nzZJ269AI0Fn8E|z0|7?!P)uZb@$C9_6b=-<6Lbz4)LsomL%`tI z@mpeZ0;|l5<*hpy(A*h3j9zR5Z+W9A5k@9RXy9$sB@e5tCJC+@QWyg8q}GLpJ8n3+ zmJJ?;P_YGqa&Jc%m5~c_s)Ys5qysyGqXS}p`93Ds`79Nt4=zc!xM$&!Nec#Bgid1x zDLzDYtXPPLO*zh`dU0Vr%X*4cdX5)R7PcZ(TMoIJsH~bzgAIZ~?}9IZ5m|02mF`fI)hafyO@q!dVzVOxW-SDaj5@hIpc>=OpQEN7RUg#4 zjbxQLYntu=XAzV!2^t^^p3W&Ok}QWL=p>0Rrl7{GGbNC69MOTD5LZGv6;?-d+VbU_a%^iw0$~oYGF{cJ1ywL zvv3Px9<$JbfWsE|ydc|c-z;nRXlJnUZQ;@om6Z|Tt&jzbH5BgwucI71Jon&n55|wZ zjR41eIP_o$u0~J^2&Wx5iq5z!>#%qt-4-z_ z)qms?-qC-cXi=4Vcmn=JV48qJg^+=Owo+Zdl zPa*KP9Tpb6d|Q&})k*bc3q>{BDkzF|M1tq%f*4XdV<&OGI{5^q{C;Hj0~G5>rGBi0m1Zx7f{66LVQ-Ayt)Jzi9WF z4_0Tkp2B5X`*&LVS6rb{nJ3K@17!{nT+_d!{hfM$qx}u~7sbf*^3wY|*xy;T`oap` zeW5ESNQNsFXx>jJD{z$obdCep7%(m*(q1YM*~^o^;`ODn(`R5?f7HJ$4GXV^v9Zak zAH}bZNo{By2rL~NdI^t1jIZhP%wyk~g`jV2_F+z`y~{$%ej9K1r3gYk8z-{DhX`Pr z6)$6Rs&EqDN19=NUyaYD{ISKSz?)8hz%5G>P7()V=BkI!c22=6$vIU_pv#o^!(jj5b&GMrd=Siy|xIz9p(CAhQINLV0iudK27dKs|* z1RupD2|ChPtpyv3Lwd8Dg^HS1laD?5s7z7ki#;rv3R$UgRCFv+(I8T6y&<3LD)pX4 z)QQdWiJgZ#6jt;}uifa0@Oes!hc_gn)WUv5Ay0}hE>b%>>M>G` z0lAdWbTC&)jLmk`v|~p{kREw1j;44S3EWT{P8L;reCBxlEc^OJ&*@9FCyI?bfMvrm z*D{*M&H0?lxza_pning&$Z6S0E1y*?LqpY9mK0T@Hb>^UNQM_rWl2`|63hIk9=Bpc zd3)JXz&3W?-UeZlNjcKS&kUOyjJk#9P$~2vIeQs{=!bzE9IwWmXkp|z9K!z^m2le z`0^pkGEz3l{P=NIkW;JfD+Ae?$KE)uOk-~td$ZUX#oi>gwSer+dUMzt!`>7I%K@>R z!!&l8A8HdgRTx^o&ekPw`g+56fELy^ueWgJ@T^N5u%~pWcuTj%$a=`P$t>7CEILda z8cPH$DW*y_m_A($iyH44K!XMG<(*2^;lE{jGi5OHmR5_qg28N(D<4B--OsI;=hZ{w|jK$>_6*d50m{F$P20+Hg)s@|> zGC9CG@SOm2(B{Zqk312WTWp_8lq*vc&tOH=@;8iRnOZxFdr6;_stKH5%QBncxx^p} zY0^$nCO{*hh1PcWiAOv=0C{dMk}&_(BDo52vLcI$SvYbuOJpH%)XyGMftsX99WTPL z%wGBMnDDX4k-~#ss9)k@QJ>}Aa0|ay#7H}fE31~_Y>de!{YQ?GCHV1iPOBp(dTiCH zkKr*szvR^^S{}YgzkP_q!*CfNzWVwvcuQ1i`@lg!_FGF-4e%TK-PvS`F?>gNtu9DLnF4k|G?kZ6(8{MN}k3Wi~DiTzKo?p~zz> zaQPm2M#@2xMPUhz$0TE7wBWXbMORAnbqa$942@5tjV&Hj{h=NcoM>|6eoqG+lu3xR zW$x>hx_n&+_7%OoOgYlB-*{I}oD|leWtp`|3yorv6iqV)5hjG+*kCE9oWc3NASqOi zqjd$VBTT^w6V;NnG%yI(QZ_3WxR}2@{VZe=tuk^Lo<$3niBmtc6N+>N9-o)=_?CCM zKhlRW>UZ{rMLNRulgO0?yWthmZF-%aW)~5Wp0w2n*YkBU?c|eLYSh1zdrq(N7mG^` zvE%3W%CW!I$vAz%&3~n!x%zu|K}}Wt-@1v?^g;hpUG#1V@!Ur@f-e-~73o zOL%;H{!BmDR08C!#Hntd!K2(tEb^z>xIo z1h^+_)_jvBns>A4qg5h7s<}!iZz}30{+B_NvmS}Zt1IyO{&Xu$h*+WaDHAr`*_7U3 z4yuK+U|pO%UbZXeZsmr7Z8jb9zP7;j{3`_T#H};;s@Uwh%f|VE5xl@L;Yd~9jpCXy znRMEzUuKoy4Y6BzNp{m_BcE?tD3L)5j7;SebD7_V%#O*k2wB&{=)g}m!q!cL?$x`z zTwdez>&Sw^n-#cmL2PMl7XDkT>tYWl{a0f&vsPXXITS;q@DN^8vQHY`w&gP!8sPng zVH{p#nC4B(Kdx4W2m(h3#1#bZFHO6!dW*>%ddYK2vM)^v<&4h}q}m?x+K)KLi=WGON$%j7~X>Bdv4zE6(9KH~7g7tPz{tWDc_n zhQ&Ksv;ieRW-hO~zv8>1A}7)3rNNk79^+W(tVJNqlT!{3$?P-~H<#->8%OeN(}$G{ zDeq~Ty9qiIo`Po!mD--V+u^N1LL^;*%L(E>tYT!&d4MlqO6aR$`D1Mrt)#2t(0Q5%swPRG2J+*m^S2Ie?EIx0Td>;z3cBeq!4}TFpX9r2-&ubG8F0toam_~p{6h9d{T$%s zISaA*7{vV=WSvuMuiyQhwK*2zmC*BI?(C436L8wgg9Sl=2XvSG1_w88L>mW17vYB3 zD%IuS3<%wTQ|Rc1GG~X|*h-!Rc`X0+ECkN*K{3~}(CK25WsyrA%2V09ok@^UB@SJV zeYd2B^}D9}@Uq2QSNZ>lIuO=F2hgXo)+K692F9-ZC#~`zOYD>OUYSBhkbhmeo3&dT zcyzFh7v8MrM0L-B^$~jo18Zz9$2w_r5{Z3^hLS<5A0y}vQx|4GDZ}RaVMz*BqvTq% z=hyGWR-kFF%_m_AzHjiDprB~&vl^x)*yN1x493cAdM$$GQoL#I*rc5wd6o>aUNuPE zAj_33k}}(rj~H(W3E1nlhuNluWfe$27HVyiy0G;vPxyAL8$?*mDo+S6(MH>>@K`?7 zK*DJZ&>3tO%gY)mKMNlP4zl>qp)%{G0eS}Q9D&O=HnVf|3&N#0^jU#<+paz0* zZi3zCQQdG{dN5ROu(4n;?IqyDm$!8)cro#Jn@P%#*p4Z`c?ZTkp>HP;B4^>A>ZXQ4 zCtzolpvQj@sa8{?oB|RZrHq#rBd>F=!8jDk7D3r<&P)qn6b98eWj$n+Np@17q_EQ) z*T*oO$>d>C`6bK_1=*)>5uQ}dBV2q>Qy^y65+d3+va6)zpUhb_wgBs*>_h`zi_CIW zgze5vESa>qHep|uEdMj8|G{f{)6Q0yy*rcTlNnqS6%Co4N81+^WCvdgP8Xg%|+t?-^^pGJ`Xj38Y=w)Ge*X7q=*q9Y2?>p?lvn%NG zHpDo6d|8?FJdo@uhtRHUx0f2sM)tE+fw?L&ZOTJ3-8mWh9y8 z#!W5rQ~bJOm=#!$88IwXVsE#cUF6woNCIcEocw;tNh;vrQQWWX=Iu>NWyXyN%Y-HH z)WS)2K}1@DMv&r-a6*`7KnxxnJSjXnI0R{Z-ZGy!-py|1z?Pt}OU_LMS=Hd%|8-I- zp$jS{%Ql%}DZ@BdGk1nCI?pv2wOMSJu^6spz+1(u>XSz{ehHvs?XMxH`jeYeB;T z1&t8XQr-y+ZVB!Z@i6m9bXDMaR4w#HH6>fE=IE-GAa8*?o%GIb5jzaT%llJ*r7z)? zce*G)1{)6Lz!5s>Cf<|7Xm;r0FjpyD097+ob?8EX+DR32X1ycpgM|)*lOgq=xy8^U z(yo);Mh@M>mKM|bv0(lPPr4NoE`X12ZpD2V8K*72VWay9iwT%0`|*6DZVE%zgdU@e zrMP}IIu7H8IK#j{5=^65GAU*<$0pzp2XPP!Dk%xSX?u>R zyBO7pgBz!Efez*oHh7GV5?CW9Z0T95gW*cbJIc|4d5c;4)7Xd#V*+FPpZxOfMIDL!#(b{&wS_!rhb93xUZ_`XBF(E zLQftQht0D3VzNESlgty@B%|z4eiKa#_s&b+7j@pe5~>9vNt4@i`}IEmhHpzice$t5K};Wf&&55Sia5# z8Co2s-!*AOeEt9S0YI^1jG*|ivGo|Y<*CiGl9g}fVJJI^=L39kI4=tm5NvN`850xW z7?SGZbS;ChM*5(_ogw20H{s9+R1vJD9bA^%A}l=TdMCC%d2Z2oM2+JEy#hBUieNc_ zhqQTbGM6qoPo^H0PhZTM73R`xPs%DiGv(D|ONHs36e~3+DwdbUl{XcZ8jKXp$9GsR zqsNwl%{$m+Vx%ZwA~MNi`GYIE@xU-*basg~2D>K_%U5(WqsXHfTMjX0qQ|!N=EeBd zVCfK3mSz*pc)GpS7_U(ylZ;dswE(P?gQv(Lj)69_981ylY`}sX>?RlZ!u!$N?r z5hELq#d!nr7L0CPecQv_&cI&e{wzj2iSVBe;`qZr#Rpn`Po7rB;{Z z>VnSNZmYB)FVeN|384^+ZX}3p6j%$=x2O8bDKRP$Zg z7nIo)BjKG~w3BDF#R3iGLmsT#rp(Y%lg({<&6};Jk>37dXOr@bYt0D9E=;^Du?YWl zc-$OM?!Zy*UYfJdHyKi{V37Tr&aWv!SfKRx_q9))2 zGi4FfGY<1uc@yo1qEqFF;H zY$J~zVRF{pV{ES{7hH?!EG8zlVRV&?*D$&4`$nGbV(#ev37+m^w#6LXCBu@(kT|#p z$TQRLmBafWY<$U~v=FqoXbV?KzbaKQoYV2hh01+^te|P5&Fj#h{M|vq!Cxn+S zjEm_4Q)#Rvj4cy}VlRX)EQ&IRpVdJ0FoxO?_685bR3)f_PsfLK6&8UaT<=1dBQ6Y& z4h{<2#6pNfyWOdRO_F2k!xX^?jotBdIKaT#TLY3 z2b%<2Z5(=2bTfHg44XELSvwzKZySqW^#VCY``Ch55qkBmX2!^)iut`$yO{h_x_=2l zjxn#m0D~{^h?4_SB}6Wl3616iX_#h`M4#0c49KcJ#^I7O+-wzPM-M9 z_AWR31;kAGZ0ZxRgnqy}&w6X#+wctxoYK6kEVc8(^m20GeS=Dchpc zVr4?KCrMfM7uSvT<}X;-2@-4&qpnzba`N6|l9R;m!D7Zoh;0jP%aJVG3CvzxCbYrr`&&|csjv3*g zQkYK(+o6){2tR;WWjaZcV|2_2XM-hbG_n4SI7+A-PdZ0zDW>LO-m|p*qcCyY6%d)>M=e*i* zi=~`N3z@G$cJGxj5;lsNwDzE}3r!8w{U}@PLg5gyb2F&}EvFzo!QX50%`J92L$5%G zEv74tKiy*5Y4Xi2)>VUZoxI=TNJRS|?ev5H zb_s{Ue^V!q)#PB42Avxj@hy*xm$Cb?xm*xVLw9~`ZjNu;hjU+^c%wl|(DLrJMz|Zq znpUc!w?Cz%m;Mgm1j%iSv%<2`XHi%@Y_myera^AafwshWovi80afFve~yRztVC*j(s zOk!Y}o$ksD!f5Vxk<2VIY$}t<&zKON zl`TUB4-TFgrd#iM#T_D(>eXB$Lkx}%?#CjwirMlf0W!nzF&Mvt#IYatO@oZcje#WL z1vMEKX1FTHU@$8l?X&V3Rhg~h&LUkYbt#0u&aD0ax8Z5oq#h?Ct! zWIHP1W*UygZSu55#0I?NWl~;ipYR})p0(h(%v_i0vQT{fi1+)u84{%eqcaF|{MNPz zUX4^$fvO5_t(RNVEokOVcd~I7|1yTE0}`8 z3!VcG!P_u54r2Aw;sim}8Q?|LJgyLkxmwDbi|wdGG{~-C?Y{(yz}xyTyDifFM%NPa z2fmj$Rl0NZx>JNHHrHfmc7<_kPq9oTdoXyhm?0h{E>^i|h*K=Trf8i)MHhovrh-A$ zknF;Xe>n7PR0=x@@^Wvy?!#C1nx4O@;L2Xp9*g?$@m|v#E^|2EYkHMMX19ZLZ5G=x zj7I;seI+z+TPFDM1MnoWL26FuOR5|RH|v(KgQ~%(({M3O*@sxl@aEb~rNL3=3KJve zZwq_Tb5QK&4VinJE)l~WXX4K|Eo?IAcLws_GQ}2kmWt+~!Ouggmqr z?-GYNR&}dFfZrL6x`_*7s905qygV*2&uXgI7yfXScr@Tu%3u3bHEuZE3U4r6+Er!B zt%;6D=x%syg2w9|%JzEq%|du1uAB$CPhDim!bcL0D1i!6AKfuSIVIxW5e78Z{Us?@ z-|6u&rjM`n0dswYVQgc;R#T^J>ir!d;G;dPqiAc)7vn_?;$cUL4c0D;>XyNLi5NIQ zL)ihLgK-8PcTbne056w*nldO>tzwJACWi%7Fz#+dk&!;;HB+cD(%OiO&A=KowXR^~ z(0Y%zMLD!9_b}Q`L~F+fg}O?cjuQ7&xM0XALKF@dh)uId+@tjMz=Or z@hv9zaXtOG1x7Md9!JnDFu3cC0D);>gU1~&e9x~cNY#;9xOUR42L;JjpH{hNVMc3o z7?e4?6h%GaJxfu#3*sFMGr7f$0SVy?6UWR@g{BTjl!*aR0^qC2Q_wLMVs?Zcg|0=GB~21`NSQ8?{17%T;O zN8zkr1+aOK_xM!++ZItvxQ9jNJ-(Ag?Vf`h*rF|S)=_|j{0DpvJ_TTNrcQ&C)`6Pu zz$SxbGC-prZ9WuULea0|8bZFzPfu~NPdDNI+~U@`!C#q$dzjy?EY>!8^bA&+bBBJ& zSqy}rSXq-x)Tx^>J2XA01<+u`@wSw#qpcY3b~PCT2;7LiL9mbE6lNq;r+AxH=-;hp z7=cJf`wsn9y5FBZqwH^XFJN!ZeelZAVVkIpPQDyA!}gv**EtHdH2#)ZIq0_81%4xrd96T>c=1fi_EDl@T5yz`O?f-_Pp9-T$qWe=$gKPu~15g0WcGAc^f=SU7|*V`$QQi4vMGL)*_AM02np7>?hM-fS7dnc)IP6O3g#ZmMF^KerQ=u0hUO9@yTnWAGhgXiU zPUz)Sc;!s!g^pq`w}f8kDE6`+db!LH>D(%*Q;6>D8bqpOz78r;DrqH=#YIkGO&Z;m zZArGTsO)Q*6p}vVcN!EETr6Oa?E^qaY96}~wL~3g%K& zI&e`d=bcY5ydwMD8>`vqFGa*;@AAvYCCeO=XqCmLezt9CQBmP!7O(}RYk|=_yOC(S zacFNR=?udIzszX|kv&ZPgnW_}Venu3Fl<~*se)y$PcZK<3gaz!W$kGMR*W#rbhrB= z<2K6Hfzl^4=gj^tNuMQuta8>)Q}$cLAdfB6BAJ~QgSc;Nhr-lNxFZBfy@Z_*q|+oe z3fSde09l$$^6+JTyu=r8LW>5>n>~q*cP@$t4r7xHzJevhUIoif#Bvk4j>9a%hM5wR zz2DMLWqfI!gA0gb6tBi+hfNMEe`OmCs;GJ~_Xw-PeP>ki6^stE-f%Z49VH$pXxBm~ zk22omDKlcuVY=NEi^(?h4ANJ;`@mq5TTEcThkQv9!dDp7(#`88I~o{pK8Z7qMw#O) zI3ldpwdu5_6ja($G?DfD08$70rh|)cT4-0BHbKg4k7IDo?Xw4|fU}&^TIn!x*}2-| zE<6KFbCZw|p4B4a@B-VolW;%qc5~wBCiK$+aIWYA{kIrj;yfr<#)+)KdGq8880g9% z{NTz?=uq|=-xf&2F~~At`(N#=3&)W(yJ2$vszGIMYB zAPEy$z^DsFfTZn7ApXQpnJY>jah&RSc4>U_Ze9rB<>@}^%QQ|gzCOz}cE!#}G7psi zQYo_>SU@jufI#r%Jk0hNAp*ga^Du9yCPc2Y zq60HaiTyE6;8jcEyJQv3iM##Wc7Fn7kRd7hXp*%`)xMW`ouytre$=ZDi!NKu#}Wz} zZ1i)C2w}{F4^&y&rIZmcyKJ=V?49A2gUGDD-7yL`bR$d27P_3xqy!=}2J)r)$KZ{*}{*3>cq@{?>8#U92vNnYZMadVAf7KN%Y@r!iYLQJ?? zV=yln2HSG1<63NV;*QtI${Wa|g27nuSVGyvnTz~{)xluYcx){rnh+CCV`WmC0DHl~ zpci%aFnV?9L1e^_p&Emb7+U@X3;S3dH^9MPjrOx+Ul`io)~_!dYuyI#OM265E+{!w zmliB@Qv)#f``t@Ux5+6ZON7QprxQoaSciz79xoIZ9kqUC$V`X%3YLi0wb_xU8@o?L}IhV9##Nq6PHlSNtq>zxIHWnU!-p+K0OSV zr-!e;w#ctYiswa9uP)FuZ9&`R=p{aDcw(#rfe;n#e1dQKCngKf^pwm^MIV({HP=^v_yV<+S2#ji(`3sNfL z4lVX*)-=oE)UI0Z=V!hx zh~ZpEK?XPn7;f(13xI)7)y@ai<`y%YWL< zm~I<9erNf~j`GG4_u5%r4_D-$!f?RXrKK4chIz{{a!|p{v$1SEGV^xZzq7ax!4_Ik z)L@BaTDguU0I}S&jI{L)SrVawc_zw}5zO>4&q3}bJXz(B?E+)61LfBTqnJk^N00N< z$MyNAasDt2^1V53cx5G{#w)`HFICTP#|g5$%<>mALbI1Jh`F7m_7H1dm{~hGgo6=R zo|=S!O7DNIIJ}K4Bof6j#Evj7B#DRX&3irx8Xg0vy`#~Lz5n-Zib=3 zg^^=}b1rg(vC1%3QUfkF-?W(E5-rd>Q7T+m5^Ee;wpqDU5hQe*|E~6^;H>nyLi!5bRGUGa538Zy9TI|n-`XD0;Su+` zR9i)44agb({;NdWGZ0(Nu|7}^Ax|iRmoS*^mgVPLTq+?>y~DSVwi{~Wn}x}|4R9SN zZ;aJ8S8ZGE;AUsl_Edu}T^u+-)-cSVEO59f>>8iM$u|P93|Rm!yCq;Wmuf-}hcjf8 zYay6SR1*+SHLe_}@kK({jzvR*Pdj)r9GN{?Zl7u;4rcXgDl-e2&I4cg{^$1SVvG{tof>Na>BH zmMCzvvu;(`5#LWeInVc7RnYMTHaPFcieJ zmQu$u_+f4-OAS2zSrcTloL@B)CYQ*0bW^Nz0sAD3bs#>)lZ<8 znPIMVg9;|AqZq1$%8<}UTe=B4?Im>R!!CWQ3K}87)5BFdbrNdlq#HQcei7)dAY326 zPIqxHKEB&YIH31BO}BM437uXBP4DR?Qrqs0tAh=`Gsixipyj@oJM=2};5r#U7vXz3 zE+^@u6x1dFNAjbduB>yNNqY%<^r4~iF>vSvX9Hp9?9s84(Z_ZPAbqa)JM@_g61wmG zsD6;|{mtF!(aHQE-}@c<(}U`6eRP0-TSjRT)y|C(OoA~oW%q4 z3FAC^x+E_WJn?k+7J}d0&GP`=m^%qKB=)Hs@6i48c$g*KhaTT2AZ}1#W9=AztEDy@Rl2!wB#_*b_y&d zQODaq7NmUTqS_zzM)%4Q!x3?l+4kH zdOAYMrCt#o2qrOj)d)V)~X|Q_oqNuRX4m(NSczz$kG-t7Z_nGjpSu;~+~L zW$;QR<#*xKA6PL(D3*$Y2es93 zEpPZy_l|!P!03&Uc&Wvz2$5wfxaLFMk!?RD5>+B*9p0Z%4OLILX&1(6nm`u_m zGP|YV#)bMTNXZ@qj0=%upbSRQn5?wU1z`!sg>*bnOqZzcN$^H(doC!`t!PFTC>ba^ zQnk`zLuAtE9K)m)5|K(;p#&6-T9kq!D_xV;IC))CCK^}H_UdSD%mrzn8Eiw&|Ev_L z?BmnZ5XbrBQ}{4$pRJK}a~;jyHPM_7vW@05+86+{R*Q2_>^wlVTJW{UQQLVLJW9vU z+K4(VC27;Z$?@L$>5_e%$G>F%9i6vP|8+g6*EoadWMvMkH7?fuCpohZQPrlNJe`8* zo`+)xq8_me*=XrP)Dj`$WK)cr?;0q=zBzrhK`_DC>}Dm3I=DPOKgJ8qA1~8~$DoVe z&rL=g@cZwhTbm$I%|>#>J|7Pu8zN(ew|782HF6h32gx^$8>)zRjQgIdy?NYQRjF6& z!)mpa$u;nc;MZT=ecKdtA>>fKotj(dd@08>fJMU3&S{+op!;u~UTsK9;#$yUT-89z zTA;YLVPucCoWNY%f7#QW=TgMLj_RjZ}H>wGAtkH zadR?b6E-a|)$;FJ;!~Vn;$_?_+|;-HM+GPsV`N@@3?JijehDrv1g9h`&M~K}phAP! zQI=+%M+$NTc%ctwv^!wj`{?8gufC6J!p;lZxKs5$I{Ct@@2hs??7r~o`{?8g_i%lm zhKr6XaJ_+46LwyB^?h{mg;(E4HDTw4rTl{iSLWL83$MNpYjpI&tM9`Q#?cE8aPb)J z1=q&N;rr<9puF}zICk^C2GXq}v^OSgSNHJ?ufC6N(>nX_6KCz+V}xo#^?*mi__55S z<3*>K5iAiriM+)^CB~NQ)ezdu)3)v%3+9FvT7o8Nc-3GFiQvBy7@5KBE8!M^ad!^F zxJ$=^ni}N8^Qv|SsRWRAVdWCp`b`eEbzrMRJb(!iT?Sju!CJZ5mgO?C2Vbr96!eRA zxdnC$qO#Y)3$PmDZ1%N%FgSSX!Vp&zy{4;qK?Dc?O$eKC$&T__%V8}d@w*%cEMh%i zXm>3`2Omg2jtB~{=yLh^_!yt&0S1zptX=n3Z8qvz-u2eoTn`(6NfWig$!^57gW-Cj zP9-eZ;`8nWE>scEUe?lHMH7%mutU+H08`pudA}g6!>*e3NJCW%g}409#h%4ArkH0rTWaBtIhU2N z;$W9+BqCp%yp$K{Uc55XCGNjcKwwx7tQQexE9>&Ms{F|fB=@-qTVNg+<8lI)Ea0l= zb~6(}K1x#yqsQ4`TsDdrqfcCsEP7!Ucqk${@Hl#|U>Q@-dc1J$ujd_D$H)q+$u*We zsl29^mr`ZntGw7IXfL=5I{(20ly>vOUgZR86E|0({aY=i8tklOG^z0ckZ*FeI0sKX znA%}3OZG(?rz4n;TbF{02($h-BaCe=GyXWB#;U1-+3t&0j=-h+ib-Dd_GFD5!NSf1 zROi_^`EG*NO3U*&ik0akvHEHseyFrtU(@wq+|kXo30Gkh0JlM?FRBL>?NqMbKfPO_ zP9+>{L+#v)L#4UV3*CL$UL(6@(*D%2SLWnku=}hiRHNez%g+EOyH_}SsbaF^uQ@MT zj>20Vbm!}2LY#sVXo2%`nBH4I8B@#Hoix90F->|)aw^AoO7VQhmas&IM%g#B!Bfzj zY36mU2y=?*IdWFPIdSYt2S*6#W&Ckg$ZMM0D=Q}=T&xoZREko>l9}52t?b-h1y}G%r1-v6d?LSE);_>PE5;nnD7L{VApW3u$}o@ zTQjNE@iyzwE2gpYi+N6f5DO~(T+t330It2qb8zR=eYBTQs^wst!XiXzVu@=IkH0DC zX1f%Q;8Ho9Ml&qOf^Z(4{fiY23WMn=pX_GcKPwwySD?mWEI3uR6ebB%z?5pAc*uo- z9)Qrqyw0!Q;j28p1p8;q*#R(h2^@9k;4mCQe zS#8fb+H;f)_WE(}LyEPS$@8OKze(i`jcT2jS>~;jUMGz{rR>HqyaX9weCJJej0d?T zK35`50EY(7;=FWW@(D$rip(QC9l?p?eJ20zj_C#-+A#t!xdF2nXAzE8Y2CeyaAvk} zEf@zEg=Jr@x!LTUKpa8z+5q!&&^PV}LY52p31xcZV6q6Xvsl+X2?&$R;#k0+E4&K)pigwq_HTi|y=WpO*mNL1C zbLBaU36z(@jXyelNyP(~!Lt;4=`XZ zu-ljIPiuRY(-c6{?!36?;Ch(VRvm%7I-fQbia24P`Q&XnhN-PJdeiFgE$cYC90X1X zr!;-fc3=nQzyNcj-^9-I&SakY0RM7arGewzhq==LaqtM1|G4yFP+79Y*aB7D_JLv3 zVtn93>{_$h%7ksUtll59UG-;EwsCH7r}!W!5)<5uhvq}FT*G2d$4N6~$Bf~8;i1Hh z^POvOreTy%cW9Gj4Y25Zo~SHNQ7#O}LtJ_WXx0H+U@=K%8?b3Ip3S>^|o_ zhrrua57#}Q!afC^XHrm@?LM^N5{mh3HDh?voUwxSgtCif_00{r2GQA=0+UY|lgz0K z&qr`nnAE8-gXtKSk9$;b{#1k}$k)>fY)sRyOspB-3?5&T3KT`IWIhBAAF7#5L(XS3 zI(ap|XC_}ay{VnU2DIl5( zjwcmM=x(0e*|Pf!GM{RPC;ngl=EV3`RAEtq<=^@HcgttoAi3CU^l+9ewcKA2%-)aZWmlQW1bpVK4wWynokqW z`* zjfN`2I+;5>)_7H;{P{qv(c&`~YzLL7o6{oY3NE{45^l12jOGJwDd>>BI9At0O(ff!?!irK%_@je~xz63z!C#~c$z{GzV- zwq~%3v#tAk?CP#&ZqPdESdT{0-lUeaONkxIR2O8|K{lCE2k`vBrWt9eDB7UGD07uL z!gq>mh5#RfFA--o@uDX49jZa-{6u-}>|Q0UPaN)Hk{|0v!LZk4PvdK0OQzNYBQ;WB zs8|&1)(4wJ9BGF#8l{&+ySZyh>`QlBdA8X&VX(7*(ViXgLE- z)R#`W>13|nD!I{+nteH{C1rz;Rkn9i*Hlx$OK^g=7W6o*-tP91%BS`NyJLXo0_yZq zrjF3fFBD3uPC0Gg&_b@rMQ|$F3twprBi4^EUB~i#ss=5L5V>U68=k0jr+@<* zmm~|Gts@M#1?(2VCx_#ARTtlBNKhP&u?C;Lxmj)#@dSKfg9n{}5D!+P;48HK#s*&w zajFTi!CQ;Ehtpn$ss&eZI?9l*Upc@d;Kd8A3aIf)2zX1R69QdR&4k%!;QbYbPacmb z#S~olMPCTGT%@7nqMEph!vyC;UpHX?!x#;wgsy^k(~3RpYM{U zIbnPZk5jzNPmj+McdG3zP;y4m4}SQ1c)`}~@dqXtH8U(27}qIwg)!NzQBMsmB{Y+# zQX;h^usv%^+@&OHN#?C|C{v#@-~j<4Ew@nQ)5-9?7bVR{l+dR{_?p44nSF+m@MaM= zQw39jCn#XkN_8DWTFMU7M!fR>JVVf>Smd3x17Xca%U38Hfeha@PmPO04!Ngd|?53`hwf3=Ji; z6!;SByI3?(pGA6;>by^JA%JqPVeN(iMKyX$@xhIJ^aZ`Q@|HSqVJDQTSf5hVUUQ~R zqtk|^ri>K@_GQ69YmJB2z^LLkaL}aniZ#Gjd&Rof!_*G7q#mWb;_<>-?$T&Ewbt>e zJ&J0%h&r|W#ibsVivx;sxTa7`vV#Z8{5n|Fq0m95*r7-p8w#n;fIZzzNgYb)QX&?l zqk2$Dxn&5GC*7yz;M%>O;=`Rl(E1&{M=eQyEJ6V(0c--L_=`LmRd}P8^Jdl=7VS{T zk;aBaW6dJ>0}AbCs$FK(X&p)(r`2Ecl?1mm(^BdbEM-@n=Y%L9YL+d}vky=YZ{D>S zQkuNsr;60o5-o2ZK9kRM=}_bIf#4~?`vr%N)T#KhQ@wI#7X*FKEkDyTY)!AT^_Ukzsp`yYD0dsuRT)j|FqlVa+<;OG( z)3ysk=<0*cIG_#q>2#@>+bGb4Cw6G0eUGf^(W+Y2w0fScdDzhQ*hM{JsMvLeg>;37 zS6Qfb|I#5I)A}r|R%o$eJ=#7CtM$-Z#n387?-N6eiN8w>ja{`da-$d;O{V@k#n5Q= zV&9EoXf(HPxzDb1EF`f4-AzjOYskk}wgxBGx3PJmg*Qwfatw6Swui$Nm|WZUjy?J# zkO*tXk`ssHBe;h!-PnTN2GETy*oEoF7VI{FZfwDB1L(#U>^6XIY{6~==z0sr6I}qh ze{UERxIB`#yQhAi*O-tX??&WHEWO`X3#WSz#oL7-dlCPd%U!g>+Ai=Am9^%C!CA@8N% zZN;e;4#D->=wbsN;|nijE!aZ7|CXy=VKZYE%>wg!##TfZlSnN#ir2>n zP}Q_WOd@z3Ia~l61UX@uq{ur-x0PxF!B=j`1XO~^MmcDMzaK0w&$yK0(9_O zebCN8`qaxa(q016-;5q%(E#ZS?IpN?rApy;sgThppgXf~0$tmuN9XY&9Se%4ix)ic za-BZRBQ8BB+n+Sy;9nel>d@oHY)sJoNzP3)`r_y_($Nsr1jypB?Q#wmGzL0hCrAQb zjC+Q@?R~{)h9p!6VWLcrnLfAld@zRk3C?~Ndh`l-5qbt_+DT~XWuGbb0n!OOA;>kW z6|aL_t-4D`15z(ELW?DUL6PZ8L+86!sY%egSFI2F^ib&hm-Z4G`c_2M2OB+CXvBgR zm*9B4RlVNb>F5zIZRe?B@e_?s5(}&6ut%pVA<5*TtlOg(jn9f7j#2To>^}Tm(UI3$ z0@H0dG{3U>L^Q{dBhEj8=MsJm$5-Z^go++K{pn+whMpoDmZWnKe1(0FHdi#5Uf_}v zBSP2cBlsD1rP&spr=Lp?rwuCl2T54#rJTkixB1EOlBD($qU$H}D7t4RFi%)_ZZIW^CzaoY-GUDRd zrP>EsD70ED(RD0dlr`VF^z(D#Y(K$mD}8*y;UsC)7o3b!>uiEb4yEyH2qMpGegZZ zozBn#?qC@mk!b?OhBX};d7>86eqbGSlC9&}H|W6gR*}0#lae9+O?JAdT0eJ307|u+ z`?K&mQNBNE!t@a8*65X8&h4=W^-^L>n7a>Z!7`00&li5&Z-K6F>Q9Tq-E5~?q|}58 z!MZ#fP8IIfKPGP@B{ny4w@@8|tz*Zr!NDiXU#)jeS`QD14_zF#9XUTdC3hD6I==GA z9e>)?#|`+XxCV|F*P<4`#sTt#4L-)*upSJD#m9Q#tHz;|VkIPH5~&Z4;_-%J$QthL zR6H}s>+_aP%QIZhMl4uffI}?7tD}esuAs@@pG%lAdx6}o#Q32ad8gYNh}ZovvsABb znxeZKq6sv1j#y4o<)THbSYeJT74{od>m;i=Fi-sl+pHxC*YLS2SPPOj&sXqi50?|P zK102FY^%gw+(M9g_C@DkwG{Bp^@1=g2FtmHJr>Lkw___9>MD2%9Qzl;K@oIwEaYVa zz@!8s(A9{u-`@T(!$D;`7Vz)-elb&U99y4VhU@#orWVHcDqmR0U=PUY}ia1@DbNCagj{%CU(|T`7Ew!Heed% z!+!y%y@Uh2KFO|PG*&g)!&5b6ovNUk0Ju5q6>hr2aI%Db7EF$ATl;Wwv!%ZJKm{C8 zp8`&vk^3|#p2HcaJ5UCKtjtxq#WEykbm}$_INarfEW5bQ%B5*QsOR%84r<)t)admc zyx#OPR((e`p~m-TIxn466KdRv^MXR)aT0>fFvTr*I9O}j-+T`=q;J6(jY4Jn8+6Im zO9^#LTKpY)qgTWf_#=Ri&sq7z57R}zi_YzWSS44PF}QnF1&baN@tiA+n#9z{rVfUIIDcDa zJqOn7nK+yH?c9-^k&2)7C>E@97Qj~?AhMa?;m(L#4Q^ZH)Ahyk*4Bfjl1c?@Z!)85|@;D zYi3C@+)W?D>^|rrIMoEixr;)rMd{Z! zw%?c+WHX+}L6__Oul!RnV*F?%<@ZeoxLXgE7FD?St<^~JIIJ6V?sxTg@c80zstI^t zv5Uj%CW@5>>U52maTTkvi^G@I%itd?f-PbiK6fxXDN|#IN;O$SIfF%q5*SFf>@vI1mu@|n z+Hhy!Z_CuwrAICIf;8l@Pn~+yq6MiT2gs~Sa=A(+ZM}E-;vv=B5J`gKo-A8(DOWiI5E###lSJYJAQi9JjQln7G zlUWh#AkWm)r7ldB@3R-O#ywRHIO>AbvX>OMH054X@{HuN!E}P{N{%`qb;(qkK=~QD zqUxob3sOgmz%k- zn0Bacl8`B26f3(TyBR~Hl~9SNsmv13Ls@SxSXp1{OZyzGbCi!y1-zghpF%|BUd`fai!`t)I{5W1dyvPy+ zehCr65W_Y^WS`y~#ur&N+=l*}Oo|#JkUL>Yet@%}NSLvV#9$jnbhnnm5#juA!hh!q zCj5Na0ROVV$pO&m0On6VMvaw9CH7$VFI29;lzMd^C#!WtI5=3pLKtmJT@HPg1)ieu zqG-xR*`BW=QF!LAnG2k+OD`i8^xfoIzWk~fv`nt%e$4Kx5W?Lai92?X61MrtVPI)<$u zAVezD38mXcWCi8QRZzL4%0T8+V5mslEL~KfnV~_AHLm%hH5&Boc?Cv%W4o_MyHC3Q zGCV%V=lSvSrWb-)Y{&Li$qhgGgI6WjBnCBGZ!$EqGAjKm{}c~2n`TL>&HDlcN{fTW zWoLxc8v&)m1PVJ4Nr(C7Uxk{%&>tZ!G)%f~JnMS$1Q5xmII;71ad;LeDT50-NrY}%Rf1R3aiS1G_`43b3D|Vs(#43hTsGb z$-Js@f$vYiiREY?y<5d`h9w?^^_W|h9$~!Azg|z$8QpkBXOjC`{Rz0IGjl0D{gz#u3=13pcKeO5!~3nLe!^ZS@55Hm znQ4B%#SmHFZCS6MhS}h?xp9V=n8wdn*v#Mx)}}0Wt8y89QV>K{c)sBJBX_Osm*wf> zG|o@@QCw#~x^UGt6|oxGK2{~ychWD95+cmG z{=fS(dC;hFTJ&Xi+aiR>&cET{U$ssM5v_F#CxZtE7gpnEw%jMSpg+(Abf)1pA=s71 ztAd12f81@*7AxxpgVx|d-Dfe^c0X?pu&h@sR%TgI3d#tbMB;42*y46ZV+w3Yo0u4OZ)P%)XJr8XFz*y10Ia5W_t= zx}q9q9bj*Hlf8K;xYHap_8^Gy;P#Jp)AcO5;KIqQXy)w7lmiF3%K1k5ZJC$3B0MPn z7Sq3QBC=yx+`u6_eSW?O>SFEt&Nv}?R~TlK?eKiL&^j*mZ(2C2GfwvJ=E3GgyDSBv zxT3|^%wI{T&8eK!iP+vV#K$pC^Ptv2P7q)x)&*ykJCHo=_d|4lCp+!>?|eaKRjJc6 z=$VmmBKXuiw@Eda>9Tx5sg(rKqH%ylHK0s;YqzT@ zf)Q_N(BLr{pJ?e=Y9iy8}kJ|FU-S3ca#!dH@2O)Bz1r;na?NpuV zixYn>G~`Vo0@UMGW%rDL+eA22E&6Kvm2)(cyuHDbINySKlQ5D}9Zv<|?7z^R5Oj%J81 zD^^9}O))AAmh#F}>frqT?Yx>^kv1~C6gVOr#8_DD#7P3Yw@&p+cZB8FXbSX@x>i4F7fSAH$cWAe_4B6Km{-5Yd}M? zS$i{Gr!pa{C$d>md04KUrY^~5Smp6qk_v5YU>>JsgHVx9xwj;lrxGXP7J*qcMt=s22@cD$EFabK&uYE^l%A`1%y;(Fvmx2jjL=lLsLVzU*P~ zsGD%G!62Jqa1|FgNzZ|!NQgw}UK}$E2(=$0POz~<&jjqSzf2~{_r+D`-b*HEVFDz9 zoYf9}4M*_pV-X{r3p6$z!O@SiErQ0`HV8yf;RRmuno)k+JgA{DE~7@;bYCn#5ynvD zWsGixk`I~Lz=ZqASQIDq)43?M27-iZf^A+gzG0Wza<)i2KpIw^WQ#{$?3#juMqLcc zP}qaEeWw}9j=C(>cA#XCISm5kPy$0_p!CNGkoHSLaWkQcDon(&c;TM?qYNID>9$s^ zonyOJMXg_+$`YkSbw*QOISL>f4>rl*eLusT)ld0wSx^XR}RFPoM8Yhen z=1$fU3M`^$!m-*kud+F9{qv%WT(;jX7pI$6N89{WE%Uz5XS9Ou@)@Oa0PlSG#SX`{ zuvF=?I$|;LP#}E^56wl^IXJ{vG;Ln5+lI-UCLVuz$;S5Slp$*!VcifNWYdtzg9$Pi z{VEVP?HFWhU^YwE5Qq9jR&93Nhq7kE$^v}G4N!sSgx4}j zPunV3e*3(-9!2)ItL;U3Eg*#5z}0*JkqKTeT-6U;g$9ZV0zUeZBoq<83R>TMd0u7l zSyr=Ty>Kd;rs(m_GTW@JDcfa~pyko(wF|P?p`MILRHoYI*@= z3P6-InQq>1e%k>pM5lY@N~h4tUb$L9YKV7M5Ra*ctyhZ)b$L13$bC1|{7|X?CCLRJ$$#ZuUvl*<*fa zv275YcXG^+A)ekA(d#u-!ZeGFuX68{LPd*D(isLN#*Co^Y^{x{ac67n$F)=Sn|Itg z(jN7H?XI=Z{i`R{nI#(R6M{wp=k8zKTk77+S6wVYj%nIO7gsCLyk9!)mI`zkK<7Ac zodLE*IP%^_S*wx1GTRny$I;E8(M;JJN#$1az&Df!*)h!@Cy-0Cc&Cni*|sBF0CwKzFr*{~&tO3Gf}t&8}|xGYV)?{~8b*WQhyjS_79)tmpz zD>fOpys=aL_`&2Qe-wplN-QkpM7wNe>y6x6Cdw0@7SOHI*%cl8V)$f~3%3(5w)- zD%6mL_Y5jB>Y#GEw5)?z*h0mX2k)$tbJA0h=1s78Um}Orc)bF&K>0XJ$-i*XtPM)! zV^NKel@Z^JFfN#3kuy&Pm!2<iUgk`4vYP=kTW1D>&a@o|A&&RBPmaO~PK-+{Rn@mMIH|KUcgooca z0Mj&-xVp~2<7*M~$3fP;b0oIP} z!evlUuuoI+#8<|5na&cs0P?8<{-8`bTv@E>@ihxg@zmlAOUxEo6w4@wk$FBG>>9-Y znx#y3^QIQ5){A#qN5tK3`tcU)Xff-E1Ua;h=IhoGIa6uUjyh)@l}^jrf)LrWj&>{} zXC28!vQ&U)dzY2_uw#_=yk+;pQr$*u7>43p? z0cxfS7Ka~#r3~93l2L235~~#POs0~(n8s7kZE3>-Vvc7ap~jb&$B@m1Bp&1}S6mHp z4@oE+kCsx(#D>u3zZIJ>m3=K+<=8aIXT;t}EX`4dKPwQMO4U3I+C-^3|B6=}PUX9r zjZWsfnvDb>byALEE%NEOMT#31$(Fx(^_mNvd77d!Q_iVu`5UtM<(O=~G>!FO;dbfi z#8VdcXc7Hct^wmsE7Pj&UO6t1>=zKe%&WB+UrNl`=y)UT=Am?db9nl=KK=Ca`n-A% z4-0*f{;!AO@nN`xhp)c=3pub?KH#!6|3wZG+YUuVS$*P8y0> z3b^#KI6K2(*@v>FmtjU;d2Cc0yTywAt)Xtd|EG(w=uRDMP_ZU?y;Z#Fwx`cW9H}8&jd_AQt0Ak7@}a=86fsq+Q-QDHMhh(GSsl~X?%+F>*M(LLGA(( z?LPp2w0*?9k_$moo{M0%3#ptJlH*6^B$0iSru3QKV=x&%Wrm!A1s=mQr8?2e4K{GP z*G@WzP{J}Eo4Xj5OS&|1JjUva=MAH9ka!D5_dZyE9h~-?F&<(IVEAFt=onsRB&|w7 zJx~lxE_G1M#+rQqql95(bg+QUTBgYq$_v3oQ)vQqq+w=$6pbJoFXOPv|Lf~GzS4)6 zmvsQ>6lsMh3R7TtVQG$3t^}0WUe&!J3t)Ovg1Ax>V0eyD0|TlFVVgiu+kOX|0F#}~ ziq2sVD_+_Ry8 z;Q?Sd9WwyW_Td@4&8ncF1=Vk^YZtG>6ks3zt?&RaAu`T=I0Bpu9sthYs5H5lL0+ak zcnmCfvPLOx_!yM6i<(E}8D?A-Jw3h7hJtK0Dy}2FvLR!(h|V9fZ)1_71587DZ3w@K zJQ`p&vppb&>7-}_g)#i@K?=cRFz?0|n*d9AwZi+12|Azk7$Y7B@>S3yjOSOl5FePq zx9Y&%@{Xy|F+94?@cl+n{9jA-T@03GfO5qC6}`)I2v5`?a3Q{_Fy@6mv(1K+00Sgf!FSeYou?vBPpj|GqE2?{PJ)8t+rj#lf;83Jrc zI>PeW!}_Bt_p~qvx$($(Bu9_)TnI{}zjYCDr+S*__*8TVd`~Y||b_M~Sy!bey%Wsc#fwbj%xL z>e6MgVDKq8yO^?ILgLCmK(6DX=w_=?Vmkr;c$xC@R!PWEK60qWPW=fyk#LzT$7DW= z57&uNB&ZI?@pka9ATbMW7H&6PqtX6Ex$OAd$ z&w+BE$_h%hB6IQOVqjf^jkAaH*7d1^Vh4wmsDmYVTc2@z}zE${)q6waU znj^qhG;MI(XE-Uq2Ra9n+abYQg~#ZjS>U%+xorQjpu2s$p$ny5h{{>SQjCMx=9u-6 zmc)wluFev|*$my+Z0>FY%|vAha?jMP9_{6^XvWJHy|GW5l&j(GM;@I;|Fe|`tNuwZ zw4O>uc^_E*cV=5nZxr!Z0&JVVmX1Xy_;2 zg3;$GCWu?~-eRJ^l;=H#(&s4#mI~fgjBU1Y4h!oAkI7+v=jgzw*edRi?A#S`U-Fa0RyapG7M>4}Xu-l8JXbEEL}r7{ zd%TW1A#KFySOQKBNTZ4y!%Du8!Eszn!U)v(I8EO17>JecYQyPq47afsm6O1P%iUeJu ziS}8zX@$t8b4woIR$cM-dU+Z@48zOA+r#qkMfx8PPt(J2xorEZLLLi2kKdETVB~M& zn7RC|)A#;11>O7i_JVR#7HyI?P1?F0{f|$tZ{zr3%)XKwi<^V#l-YC%hkg|%_Q3ee zslhSXvO!$N7S4mnHlO5=EeoP>P&oDBqJ7ac;ty8i%~%& zrm#KdU{TN(nUf%4gPy`cMhm1tgE!|t2iz7l&c zHn@3jWCe)}RApUD^bWvf-kyVsnZc_Qzvym*RU%nf{Je%yh}viwwzvKbRi5=t2wK;!Y;w`?-kK4f2?PU`2 zlW>CsCB$Ok30}T$Y3Z(4rL(GJ3nA%L?>!p*vaaOcea`(I(=2i9>9D-LzxW=47ryK7 zaY0vVWPhiA)5eMi_ILkHG||uxUr*@#sBT z9xbPyyRKb$h=@{>SY}uI=*fXi0JpmcSNW6R-L{2x(NNuz4u6) zP~WuW{x!Tl&(p_TA!c>JgAM}OBx%stfgl2Gjxp>8704!Fg!(1Rp${wi&XVbRrO%a2 z*DHOlWV_S#JUhB2(_W=Rw`B7d>yqg-JFb*W_cPj;Oa|^%`drCwdPE2=9Z~4=ybjfm zJXBq2Ys*xJC)^cS0vsvfa>i1!O|S#aHntOgH*n(4=8Z`ULZyVV-oubB018BK(cIjF zR1Yr4kbUf8QQdX;5zsw|-t#a*dR9R=ezqx5LtcdqRP%B8S6sw)FlY*V zF^_;H*&ewJ;}V}i0wJzuB3vHOdwG(k@Lq%2*CW4w5|Jy(b|ffpBwN2+WH39h4IFA% zmS2nc$Yu?={`+x>xeJeYL!b|QOqM@c{sbdc_`aA-3bB504JE{RCv6EEWZ9=Cy7$UX;yv_D!TI zmp~G;NWxVVLbnmd_8^qU!R!@F^zw1l#&aGEPs0|)u!6-XpN1=t9Ifrz)we+Oo`(^l zqhtwK%`x%*mN({&G|C{F6fDP;%QDp<*=2wvzUU!$tIoW0&>31*M7kR|0J^ID)D{_Z z(5aaNkeL|fz5%poPtJWc`U8`!dCE#oQ24AkvTmHW^9+wWbc|FHse=FI=|{Ba9`JWX3nP{ajHp?r2m z;e>Y~co-iT75n{MSU!CWAlP$c8bhAC^@Np`aG+8S%? zNqh%2C(5+US0>ClO(Go2v=|JwK=$BZUB|NX7)N3k9bCjW@Z1lhKj_b`#`Zn02Bupd z_&f#eI7Yiov-1f?NN<0E#YU5CooNt8e+-8LY!i;+sZHwC@fKwm_+B|I0K?Ue z!Y1Tj1CI`#6)xXNwYVX`iy5n@U;-qjcfjf}zq{ZeK|$StvIi^Y1*W}(D{%7yx?T%+SHo^IKNF9NtA-b5^siMo0n^zB zK)z;>JH4ub5UDf4!_#Fs6nWa0`P)RE!tLVRh3kSq=5OmZI+F74vu zhv}UBe!qP`1^a~z&(-FA$lR6`t!shWH;&6FYJHKpNnhWa4Q%pE9PdxbzfEAwL-=rzE40ga z-av_B1#JmmB;w#u4$O&uikJ&bmTTWfJC3FTv&W$+!t`Ex0Y!i0#yJO6Hd5$6t^ z@?}tpD4cJim;~NFz0$&AgM-2Z$UJu8?Sq&GM}rrG$A#i-NFN^LV^MJ_wr&S5rzjNJ zf$_h513$yL3k9aIlW;5ixeFRiAZHZ3WP+Cvu7q=zQ06>_Rpt}Rysk5;<6w(md$5E= zdBPd=xg13;vsgrU?!)DaK!PmS;c|~!2PT&=*<88{&wZF8AX$Do@Q85g!#Tbph<i}o9bi{I=7pDw47>} z*jf}N4i8z18ZR@;tQaJNyiVbf4gHx-kx5M`-U%j6^sz!Qv$`($3k~% zfwi+*lWz_*6WXtQL4FzQCA6MdKUf}tA>>|a^`9mp{;q=XcE++H9k=#REsVR?@nN@U z%0~?*d)nz@3)MxvG)Y)Najb zn&yMLGr$wT(ha0NQ6<0w6HlP}GJ`d0U20^^4#@0y3rx2@aHXJev4ogzf~NxWc7bxl zb}sP3(wOaAfM=0Xgd+$3h&4~p-&)+5owh9h)1_=2%R&9}Ni^9tR{AE*qK-D3$>co& zRCcP(vaMWf5ethI2eybG=|1(mvza@8wrE%b4Sk4!DD$Ws^isj)_b`N4)k?;}j2d04 zK-FP8)j{LNide*!Oe6a}@}Pt!eU((IQ#}$X&p8cB2HCR6HCv>Yto8}+9iKcj;Kp-G zvJquUdNiuEJVHEULAjV^Nb~yr=B7-fMzJc6o&}O0>}?YenD!DBUV={7M=+hq)Psxb zm}0y?Uf}^~{>(5XOgX8DUG+RV-ac zkZAoL%ob}f>or@ZgvbtjoGP(ZmjLfPx1u(wQYYR@2d(YFv#QMePMa?jroDviubQ4G zV6sWP)cY#T0UD{cOk9MBgE{!wf#=kLH&b*rQqID?89heAPm=3Hm4CHB zo*&9)-H9hA-aDa`7ABv75A*B9U0Uq@#iKJ=4~XQgH0lr9-f25cCa^UqA-vdq7uMJf zwi*7_!NM~;7pGQ&tYa7hhXjhq!{#O7b%C;G2uKUZornDkc54LHes!=lb8xWnXfb74 zG13_98E+jj`Cjv0ymhU-fvsYLW0H5}0%|8~=Y+@6fsG=4gh>CyMzy*#$=R-;ia(#~ zG6gT;!y4`?ey>L*!V+pNZ2txax2kv+9_#hiBj{gTvmTt9T|k=}y#2LHY7(r|XN{=v z1lrNTpaf?^#J-6QS`b2(VOz>Vr3TRixGEP6^DsE&)l+04c;kTf5^jb6#0nbZ8sP1R z`w`E;gP1_Y){)JN8mn9&V3;6~cm*ay_%(~4@jEbnY$PWBM?!~{QZTq|sXV-PG5<)# zzh>o``y`=E^5yg;Z@6BLzdK>KoV~h18>wH;UYYxCBxKxCIfKpNZRkg?x$~-M(sByC zC_Dg+SF5?}67kkK6E-}ZdJK~wD5;m1RpwuDfya!jYPg7N0R98NgQLtza}7KQlwSkD zLBhp(>`SoECbV)Y(1%u!H301;RIoK-um-`+0cAzRhp;nTf-EuzsrAj=Dua+?u4Qo2 z42VTDaf)|S|1L~B2|KF5Z+tB|%REWi@j*3PEhQygChEisHdC;L?2tVX84Q=V%MxEd zZo-p4dkq&bK@nf7GIvQ*dDK>q4-4x zgs`+kf=jQJI8_+~%@ENrzx?Cj2!>>=6pO;-68f&>DVW~zFuu5|(FDzzs;0T-<6xSR zN2WkLQ;UH+wVj}&n|qCqhC+A7#iFC*91_@G90ASlVS?ihDcR+%PLUL|U9_d@8M_B2 z4<5T^BIJUk5hgF9N~4PAh+l-WS@o|$$VnM2`*pArSeNLe z&b{`QbZ5}oow^NTkr;^>1EuynZnJum?LR!JX@n4C$lPzLQDt=DRY|%0IS<)|m0T31 z7Gzw@yYX(T2{)C9B*=B!gRascgnNPvT8t+8zx*KsuR`;_Qe&H@Cib*)l(Cj=3AOCh zrXV!|MWsb#>~wvpGP^KtxiK!6IJSMJ4hlcIuSk4C!$trw#Bwxt(h8JWm5betXCBZ{>Ost%ue6{=m;BBCcsoe3!ce+eL zeCX|;=ODpEaMjwCAdEy_Su-jM`&y=@Dzul%HLBv9GKFY!3R}eSWS)u}kl8?hOaPbe z%L9Q&qS{y)4*Fx5z0esF0JhMk51K?(LMkMLp#o4JXr85F!|7nQDFLD*RIVZ`FT$Ao z#zd`9gcyG8-&jS69Kui#zQ9y}T6Ay>Nk4I#J8aM`a+CQarm|6KCt1Zb9)UWS1nma~ zFU#X&eEaw+TN7ruzVVtPe$5a654u*U_QQ1^&mkzAJd;kEjU&#Eqj4{1j;dZJ0$ro; zlub$nE_POdryUq3el=ujcsU2ttq=UCD~Mz7xlhVIgJM>caRA$D7{)wuQ9&lT^BCJp zK!$r>@LWR_a7$DBNB|K${C$S238iI$b;5id8bK+5dj~sLhrQ(U5WVMNgy{1W1K-4a zZ-K__gc}7Hn1@5LiY#OD~N7?qz}>ODJIzG)!tj+e=M_m(IvY88X|?3 zY>!;xDt8}oooE)nc{>M5WnG6?1o<825znB%OC;AY8QM7>oA=~K_BzU1mt3kbXNsQQ z2;~;r9+t-?dGrjH&4*B9bd-1tMxUpcZ1sPRV#-I1KSME#UM(}RC!Wf7yrQ$@R+9My zMuPfc#>-`%0tW}ONV7LRa^9kt&6&23>y>bYdeiMGO7x~%XIhV$-R4cVFjWc5y)eD! z;Xl2C(3z(DylHW!FRN9NWrQpe8@M|mh!@JM6Gm6cq@9w+u}V~uAXv%bBj%7@sJ;Db z@lV@NZ;-5CN8MgI_8~Qz0^lKw^#imF^*C1&;NocEFwWM4!&RLKb zwkj#xwJQkw2ds1M$IJ6&i63#PM_1L0`VbvJq6E4+K#Vb-cGw>3?rh0c-~QGXgVmF% zVcyvyVeCmK<$C{|!L!1jd6D-@L2heHQ#)zSR5gtSdv|M#!Pj=SP&(Y&Qecda+Rm0F zUv*(-+edrV0=fsLTOar=1>wft`UDhw=T{r=GU3CQv6%+d29bf2hjIX{Fk%U+?b!4zUvf(P^|WNnww^3sV#{;xtlcw-R*d?2#ml4H z0Y>_{K7O_SZd7ca>0IcTU{jXP%!|qF5scRS{|rZ7KSx1)y*C@OQw7srLJRBHWr7C1 zqS-bS%DFH}Sj3NVYrSx0Ke!;CZZ4HEs~Vrj+~sAQ;_~_o>%oKc`BGj|tU8Vy1VRKe zjk-_O`1Pmn7yW4!lqy&X8X!U7qQ7{qDRV}F7hT0suWlhglE{mls5zHi4aEV2^ z5{t|CvABFZ3FLyJC{GEo@~aWXJoYR}AS1xGFKzm~^DpG&r+6il5#^_DIg)>Bc+$aE z!YLdM7bM7w)D)bh68mu8YCXxe?nNq8!Mi7>5uR9#N|mmjlHNZ6UG4CI;opak4?T;C z%`dntm)E#t!6=9RIr4+IHFxiW)=y(dCVbfQ7WzW_#rlYRhaJJt^BswxuBfD0;USN! z3LdgZtN3j&AtLa`VTI=o92lOZ1GktlAYFKh9k>+3L*Z>Jo)p<`18m_kQX6?oQ91=j?Z@r7_$n7~YYLWHfvbZ~SV4NdpvB@AJygOlp{v1KoDMFfRa1n=TP#Z<6TWl5r3@*jtGAzWLOJ$q%Y<6?WTC0oHJjUN0M1%gel;?`c+e z1UOpHyx6DW&FVIvSzuHy)^w*tY+KId?eP-d%6NHL9==Gwdw9A$442EpS6}~yKB4;2 z#eG63LpF(}m>g5fPj*A28seV?A@mfM{9g5GD{Qg&h1EhE&(X7yY}b&qpKIvOiOg86 zri=i&XV+|Sebk_lhO^YJ z#dc4@bW7$VDjj5o)0V0JHB`DK^AVNyGY6ERp-Xid@>-6QU^_wH3DRe((4q3b3G7ct zcYyT-c{kX*|NAa*-(tHF?Aefaf%KUwbj7K;0sMK0+%-gc7RPmv?tHERd7sDAz{br- z+r1SaN@~-*&1{;uLA+`xts5cVg!$@hi)P~oCx|3c}_e+qx5KC;?=Xi;UR;iDqUrP()~|A^WOHNC?$c7_LDw`0WV$S;&R*gh#Po z{{tvB5cP=nK=gTv(dz~7dDu6Zdc;F$rvs=*^pF#@T_NCUEoi+BYUHaYh+0|dO0gSA zO~u^MQ&*%nrg~$q9;953Le$X{Jxws6wlT=G7qz--NM83<}eFkrFGYz}+ zG6qeatsvFUalSmigg7j(WBTwIwi~FYPA)AOZW6!x=A$YOCPzCA=&_>9dl~%!?;SYT za~%mFFJ2}K=cLPInC=kb-av|UG=2c;4J0?h>CVC4K%`)AGPwb&jy>NG{b?)zd$|Om zk(>HQAm852T5U5P`B_IiUoh{qs`d z;3$4U+$v&L4U?kFwBTD6Opa2uky=ur3>9U%LZLz$t4y-d%zEY;g=}XUW?F8h(0z~D zW-M1wa-{nxKNs|;P|&}t9CZphJ0RL&Vd$2^n@HPHl!MI=o%e?BaAS!&#he zec3i!=_5vPf55d^%2={gZS~)?!M6KF5>)WaUogo({ZU*j-hUZ5(rHw-Q_gp-c`IW#GwzfG}Hc*!d=# z;>)E9GrZB78tg9lWhaB@O@nLiQB6P`EW0iazGeAf6KwM+Ou_iTh7Y(4#Mu=3ZmJ1= zeDK24!6`X$lYVtLlpm}284h>d!3#oF-cXe{R^^Ql7}FQjPJZw>K98eE7BgJ_=<|oF zxMb1(X zeDZuLRc`kSVK!};^4ppQ#mr6x$PxsMY8DU} zry^li1xw+Q211P}n-8EG49uS5{I^_oRJU(XlDt+ct_?V%(3wn9+4mNtj2MwJxu#jn zPf0`5{?PD9CxfUIRY$cl?opgbRQ)Ig)k@h1{eo7=mjjv^XQM0S)cV}dSiEeUTAkl$ z?W;qi z6ompGd}R{M!b}12VG4+syWa~w*f{#D2ZyKVifEN-Y$&mzq_&_`XBrwxUprO>#U@j#DYbnLGl~@yn@lYw z)s)(g%r?Ypl_{pWAuXlcSBC`+<$g8DH0nJhR!+^MhQeQMLDEpiV#CNnjFg)F6dO_* zgHDL&io(}cAcYzd-+Y46;iaO;seX*)=M96NR8ykZNahPJ6-74aIMP&6#)hKTa+sQ0 zLRFC6>|m6sqKpkiZ%{E3>a-b4KBCwcyOgAEgi@*yeE`=&D5Ydq5PnNUEVEEiOid{z z|9Yo8@A_RSDCn=%5U140?Ues%3W^OMsU_V)>LS$PJK%8dp23vOTWjRVX!sv}* zFf}s{TQ)V`i#1C1fanYbcL1`@%j;=32N+5ojGN2chbIT27%8#YovaVM)N6(Ojy*C z?xWmT(6~#XGntGsbSV{&U>u1Szk#9!1yD5|id1gwQb3MSpgqOJ4@KN-I4{{?^(iMr z=|D6o)yk`sL{1Hj+F6w;ZBMmx8&9+ep0Zp<;G&x7v=AhYM`f17JKng&mXcbERsk0d z<|>o_)hl@wh4wNHH&RMLq0Y24lzJ^vB^V0|O{QweGf8C1mFy_ILSJPHHKj-}8VzoJ zHiYt_W(JT%1df4H)`g*DUZDq4Y$!6NC`CLAJSv7?qhr!sQSj@1u4FtpoY2X# z#(T~QgT($&`9slDY>S^G>597TD&2&5ASLz^{pObQ=!7RFSHe&DSU+ z7{`uORcN|KNvdfp(jlcWriD^%(C94rU8BUdptzUmkdhjM%{2Zkl%XxCT}-`&a#&D* z;j&Mm`yTt#DX3k(5lYo-l{~fs-35FepPt9KYPWT2B=G>px|C5gJia=)T;hgW%i8(9 z2r$|JsoGMZqqhtZ35r*3Q&pkE7#0%LV1ba2m|0Shq&5strr;D+>{8e|0@Ca$;~ph> z3I*3$mgx*-ZVOtPOb{RtikWBX@A#%ccR?m)vMtZ1mf|9H5$Ah#t<@$#%3`Y4#7etg z*KN&p6tOY|o$P0g=AVubGdx_>2e^PGsbojYR7-aI92f<4;SnCI3u;1ZZM+AZO#3V! zTr}ZIN@%lb&uD3LKuOJIll3_SxuuQPpF$agQg;kZ5G)pNesp7gqd;_y7lLEeGvr9D zH;__%3hb;wC;=(tNukN447@_A3mThDUCPu@X!4qRWR5sWWr9SexMd`&Y?;WGIZCvT zyRd3Rt7Ndz54vUwq3R?~NLX^^Th4m;UA;sL(WgCpQXLj;X{u6B$XM*~PcFvp_e zot+$wH5KbZM_)N{Xk9L?-i*aeFr$gW28Wa&mS+*~h09y;Dd?yHDJ5)a+n|+%4EF&{ z7YuWh0d^iSTOg{zMUGv{Qc=i}X199Lr!0;#2!(1=@|n_E3SG<8hPnC_y6>^81)U43 z0c9+dSTWnj!fv39?G85RTBZzLN^+C{awLNyH53)dK`1J;N*PqL&GZ!Uk8eUzfmVu2 z?~71q5YZfjRO&SF*NSzJY_LS=Q?J>0!@=EMYXS8z8?XY-OF^A zHQhN{ibQ0op6tX`vGo|%!mh{Ep0im*rwkJhCP0UArU1X{nNZEx2ip}p@lE~d*L(Vj z`)On2Y0^>RxwP1ZYU;%PuMu~wQ9to&8Vts^M(s$o-`EU(sL?^9>d_Nvbd>0RwP;mo zw3j%vB6yfaM~U(?FK|Dt+x@Y8T%_B^y)^cXhTRAC(^9>(#nS2;ZQ<0tHXf!CPr$vz zQPMihDW714G#~PtWGZid$7wv#_7Zg!pu-%U;4qDAB;TCwwedJjlxasFjpnH~%o%ku z>gb~=F`P~tkJCz>QcoY5YtvabT6WX)%DWx|BtYL!v)QQ8M>tirp}%Q2jn@&V%K&#? zwn>v(b@Y*2z^sGC%WfKO_Mn;p{%~%c=3)pgjjA&rRYY^zO~WhxQiA&hHM{-hvuP>$ z9kolHxRchb`{kw`B{r;SnWLQ1P~C)yRa6(!aP`RvM=2=u6RWgZA7j-9*y>Rdkr_U1 zXwoq3<1SyX6EV#*XKd0415PL?ekPA?8pH&)f9WXE>V|2lO|vzzY7ilF>h(TsKeWEp zNA+p}2tcy{p=m1FZlE%$7V6_Py7$34jrxht>G?6fE@S-o;d#5JE!98Ivg8Ks)D~DJ zby75()7eiy?>%I_P`Q=~>?7x6I+^m`C$`I_kQ?9rAuP|Q3~h=#Zn*Rw!clgRLOuyb z*l<(IOuZzi`RNpb>8kh8=@h!>pp)DW$aDp!M?rq@-jukQDj(t1l`GptD3Pzvh{aVrvbGt{$oo zKWHJo*i#%`hKEx3?x3xL!?g@`Q>YHnosnbc+a^|L{ZT&@k5eRMOQBN_jn`A^L3<~Med^ptMW(o*x&uamS%QbI2Q{AMT!YQ0xjFGbq+JU(CI zM_PtB`dI1!xyHte(15t}eDL#4_%UO@M`5UD0;`d^B;=}`*HU0B2)8Wqa+zOLl{an+ z8cx$%-@-vQ7f+07vCyky(|~((E0+(7sN7G;XncVhbq7|^lx9UFN=lw8>V1B;eBhiH z@n)L$=)wSXBDK+$nW`v_?F6dh9wX`cRgz#JK^lThth5e?I+!qLiH98n zk)Hz}j`CUy5Jo82S4@}Y6`a0I`G*zj41Ui1no68x)+-vF5Qc=0MLfe@DmFc-+1Jf-K3{+2zA(NDeEvFW`?m+wMl5jT z-k@ZC*c(U^O4sNjQft;uY0J+i^lRT0uofHNk|iKZ__=%z?@Y>E6-A~YOQiB7yy4%K z)M5#5d~(}5aSgxgtYuDS%s2T$e*pmbMuM0b;A~K&?0>(W4ZkPmukrIj3JELO;Y8F0 z3+Kx~7yR|-Us(&MZ|#UhUc?o1Cl}q#(|<0yC5w+X;7{4wTx8CRX2Rj+Ekg7j^?jI} zXW87Bq`z{SPqaNZ1~(sHZA&~6PvG11bos$u6hBp}G*rK!^QRWCUht=HMzhM|=gU5F zx4RDo9T$pP)&m3Ew5S{hvEk0sVUK=?zj0ocFl_hS-NI?}Bg~fBK(AhUY>-utp&iDQ zYO^%6og~^76u;qjt|FLb+O*j?tK1T|r@2fy?khH_Xns342DGa@#969?nrqqi7})`e zr$m89oK87F5-iTXsr7NHQWM`ZldkxabNZP1#k5%PQ~K{->=@dca}`HAIN5^;;F+@5 zoYs$9_h>1BnDy~eaUz1UH71StCpT!&hCYngQsHP;+daW7$4ZDqj<9-Kl<;1zvb+-X zg+T|UusV31)bYdpl!@o74*t&xWyvjaL7%#zFTxwSrVBGpc1cGjAefWqk>zuU;O8Tq zFI=R}Ladl3E`RNzfgE@Baz_gHw?FT)kIYggZ_YPJW;CU^OQ&aQZ+cvLYdN7DYzVoj zTi7u)=v}q8PiK7(ZN~9*v3a##X>TKz)!w-jp6Kk?o*kxgb*;{pDNu>z*II11=$!G3 zb7!nHU*#oWS#+nWRQ3He?~36zZ$njSR@035CXx;poAMTTt^#B6GnlFPb^%3ndy?*%T^}k=DHU_&DwO~s5>t)C@;n(-pXTOWLN1WT%cK+T= zwFzqeGK?V`?dOiKyDy|M_s2tl_L9m4Uf&rZ-$Qf08%dmR4%>_Oo>yYIVXq+rHkFyy z6Q07Pazcgq;bW&1`~YNFs{Bglp%S6q+>M+f{{S0Uw4L2wI7LEi546FKq~f9g1L1LY z%l<5O_SWL!auotH;s02?No6B{ZY0xwb<%p@73^W>D;?80exN?z#j0!&hB;I{I==CX zU9}_`k(0%lHK%x1@t~yE`q>@8*KbxMlU_SKG~vXs{d7`tQPEj7lDHD!nRY=3eh7ho z{9BU(4_zRf^69%ToE3lUkE6;zg+EJJJ-h{-LO<=O<_N(c4y? zdqkb5tw1HOjmAP#4_5v^suV_GH2~Y9E}Ws&hH!I$;qK;aa%1EVK0l`^Zfdn{z4Ut4RzJ`ggrx**!dA z5umLnwnehrk!WS`Rh-*l=hHkr5EFw$x&J@qtXZJ1kX4#e(x89*8gL@>h||D)EDEyd z&ZU*_CTZs|`5%!x*x8z0q95nwT9QjU#f_-+;{;5SQ<^iBMY&FAUQy7N#bbMnf`p%N zi#g*W%Yj5rr}nl2)V1_MPdS#!?hM#lyHXTnoYE6yG1L{rxP15X@ZT5I2NA&7{86t+X>(RS{;Byo-3j3=#U&!2;zpgi&uCUb zikhrVwEpFiS|mmE>r-1IVOhkzk^yp=loowG)?RRA8;j%jQj4gEbzv z4Sry89!+qgS`DBm&7ss+8oomEBq?N)BkqYm4rNJ=rduX;Ybuu-Nb<5)ZB}TjA@_K) z`X#T9lD?}_D2w@Hg7m<~gVJ)_5i7x?@;!i0Vm>&1J@k)5S?2T^i?&|pqy3M0U0~YUp(&LY+XXpGFzrBRK7wo;QpMtGU z(Ff20%4dVlobR?L$5}8ZTD&vh!Sg)qVqixj1_jpf1hau^+3QjB$V)z^abpY99YI#` z{Afu3Xznmwj?eG42U0>eSIKHejjC6vs8zJX~&^#-Tb=Pn;n7aQ5 za|JcoJ+QJd;Ds*;I8WtP#c#1j|Gp|aKu}HehZ(LjMJo%#HbMCIZX*e*bo8uv^N})F z&obyrk*bvs-6d5^A5BHQBJqEm_GDRhQ$KbVvQDK`DwT~w%qMZ~eYVq;pRt;HQHhh~ z{ZmJ}Gjw^~w=i+tTea(mFSBA$K``K~yeN7_b*%5O)uPjn&sbNv! z1<^0_s2(}@USDaq-KULuml;0vv~})yz0!3PP7uDDmv5uV;+jsOn2nem_C{R4MJpbz=4y-T)>d+jaog0>IpoWJj#!-&=fn~J z5K7!simOP{dd8;t+UsSdY}HJT%45l%bvf+}*Oa&g zD0)L00!)SO$Pioa$0K&Gb!h2XsfL?YH2i`lP%0G6$DmRk9Bj8F*4-8ytYUHG8`TwU zq#E&~M9wK2RGp6MA(Y@}bQf>-IGTrrStRn#7ya*6GjM}t`axg-kw(f-jKZ6RolJ$7 zi`@b{`&>-sF#^)Gr95SUl&17NkNYFKwdb_-BO`zrWb2`UHdPAFTF_rF=Y!fPY;*TL zXVvPkKVoVJA|9(>eu0QndHl?h>NL@?63d8yMq8B;N^cuc(rL8p{;NU_mbL9@T!+k> z`q?cgQC$`<*Z^K4w6f0Fl;%wL#B(XRfCmee5mlQcsxgKa(g0e63cKd2 z!^!>}&e85#sRtW?mnZt*cOojcg&_mL02@#aMKyy!x1`j>Pd+P(-XfOp*NFyle5diN zLB=Wep$br9kYlKAfRVjOiWqh%A?BtTl1^M~>A_o?1+LNM{7x!9#)K+5cWoqdbbQ(X_ z&lO}56H)0)&Sycr-aAWNy#8b0L*L)@<)vFLd5B(lAwXS$6wf-;y(7aj6z}fUGhs?dQINgccayQ`K{emdtKld~$ZHfHLi;ryHn$tmQwTG;Mvv1pfbqV?kr zw`2V#;|d$&7G)NoFN+c8thwKbY+Ttc8j41j3Si6KLdDjj7GOGv%Zgxw$^I^(yUjzA zCA=8)MGFAe04;`CGeGiN4wJOk@4DUY0M@fdDGDzHTmJygQ#HI?i==)xsIh3~)c~^O znfxDO(u}H)ri!r|kASbtMFSDVu4?NOyjnNs&TPO8htOh@`!F#ALYpdcwbPk=@deXW z8jFr_J4uFK^5;{%Ak%R`qTl_)ovQ%fwQ>iN?OYVCTv;;ZgT>7*nd2KH9D2>Kx-jmy zZ!r#>A^?5B(9fwqyEoeu$*61Wy<|VO#0u*1s%Ol97-8Y>>f6cSAbT46@`x8a@`YG*T&l++K`lVHDeZLA$h$_4EPc}9?sc2Alt#>{hHK#bf zL)8{&gl*yfnP(X9!Qk&eG7tJTL!I!AnM3Q>jY!&-)rupC@>UDvkiv03b0j6EHnwQt z%wc3P#uq(l+9CGwU>(-y^~Tb)&XntVmgkz&_+>A=KQ3D6lQG7NO7vG73NuXPF{K+n zp$YyB7A#}%*h_ZWvz7NB4{h^DtLhHrtN!g`{?=|+ijD9SpsZO_0Ong9Uu&rP&ZE^d zboYX(pK>kf!{Cfw=wr*V<5r-Y7)5kopGK=Q{!DH%@#`SoU6ZvN<+zL=FXhAHi9y3% zy!fT#!m!(jPI!N7CnMqI?t$U1k0n=xKw_mW&b^$ttq%X<6O?1w_W|Cz&NA6ub?~(N zbXhS&1DKZ)?J#X4coDjaWC=^LT1R&aIXQS#OYxVRv?S6%k@?}#8iug;ui{w~vJV0s zn&X^@BY|xNM^KC^_h8z|Bh7g={C|AZjp`L9c!iE}8R-LBYx+bk&NrvZq?kiDe;ek9 z{n{2UcIm$KgzKxgM}$rD0aPrU`ERa10O#!ZK9pAL!7p!4%-kOm@_U;^yirpW(J(c% z<>&OM)&X;CBWurzFtUW@-WtiVuV_wWrgW(VxSq=XmEIJYsq_gS#qx@X@Ute|qs1;ga4kMoJEHn2)0eaHjZ+ig z+`ZXeWR+w^foLv8ZtILbG8U0C)8u$o&B-t>Lv~nN3R540-mim7INY+WH`nQ&;Q?&? zn}7M!^b22N>q}~X#osJnj&zFeU`t?B2b}*MtyRFz+n?s7Y*>DvJ^Fb)B|hT6CgET=53Nm%4+vwU1F$xh)) z<0Sft&P##xD#~}mdnzZzxLzoP zYkCExjNAKY?>1?%YUpeS2yA|;PcVu~8ra(HiKN1&rx64%@4I@AsIzjF%hVUk0vbt( zEb}*Ub}eX7Yb3pznSR9)37dSC&R3@f(3x=Hej2sIgvGf>P;*N(2vjT9$gi(3%8={E z`7k@jw{X-ql))D`Y_M#b!HPd->{x$>&OB8N;re77Iurcyirl{N^J5NFU%sSC`alia zEA4g4zQ`}UmVpmXakD0Xa^=TdKVo;M`}3Mw;bH_ze^h>2d2wPbp8RIof(kUf6{nlw zkY6>qTYWd9i9yPOH~XE8!xY`$HqNkRAz+EK?3oityPLlXUlOEAge-?sUW-*G@Y9dE zJm11=LI)n6A~Uy0MRpzXeFOT#Y=*0}-W)PZKQ!M7eMns@P|Y%AD+}w&2<-Cg?M`KU zswLr6N2Uz>L>wZxYi{csP1~0(_FK*l>PesnkM40C4W{Tw!(MNIG1MKs9GPa$YmH@vvt=SfS)gPBFl5f z1K1=Szh5igl_k||Ty-q5AK)1&_Ie-0OZ%UlzhXV!)x1T=y)>zOoIEZLnN(|eM&xVE zdix5p9q|@fUq&=jss0>vBEC8-To^UEZ1!LPpLbH1RABqOUK#@}Crso%uCnSB0HHpS z;5;y^(>%sd^09CK{ewf)f#C+tL)X%=m_v`n92e}!yk07Qq4WnG#N%;n$|qd3JAP~N zzju(j{E@G4{3?3n8up0YE4+7WFHzO?_v=K#BaWiL2dFFIXTO;_zz2ZR`;lEp_2jR~ zLZtBD%c=<7%(?2*o#hBqly2K7NYpfbMNyIijP6gBkgiBxVCl~HF2>b-D6J)jzOTxL zIi0!SnImW9{7&PBbL#+nH_{KK`ajt4U|QT|;D0#lRrO8nt#1B$6XPjvG$IOpMvsZ- zYZmTM%F-G<5ne;oHp(3S$G0irb2C#vInyzBJ^V0F5menlL(~e|Utbcrf0btM=mF`3?M&Ca;cYdRfG^N8D%ZH|8laIT-*uoe`aedWtB z9WZKOBuHvrcLollRs24-1UBL^aS#Au@8*s>VI@2lQSV_hr9UG0TQ8xHnA6)c&*^_z zV|@>1rg;7}A1X@emgpDvy2Qk7hC8<~GEp)HeA%FZaT)nS@e1kZU^?02%T zEG!yd;_${l7jOt}8w#^2ZhP*at~fjk8hl$Cwoz$GpK2F05W3M*`TG@=zWiue(Ly{^ z#P@`KGQ~Vx5Al1(-K&J;{JDD{H26o(UtJgP%?+|;*!iezNr0+4bkI(Dllzu#!x4ye zT=)Bz^qgRHM0TT0A6RjNll*!fwvKJWhMPCv0SA*d30m7*(4pRqLx>KE!gKVq3rK@Q zEr&)}HZ6wQa{w(2VQd^PLe`yyfT4EMIU*rFK&gy%UE&FY@rV!Ll8SswoRjlY_}JNx z?z?$5)0&cshzIm;9o2HS;U@&yq{lxD80QN@43bC+J$iky=wFddKh?1v+jT!j07_9Z z|Avxh?73%-E{9syPpWbC9a6zr63il+X<-DeSR1M4$1z_oO>rSr>a-ZW>Vq)v@>zU~ zEQiJ;MlH|qybTvEelxT@YAOKOL+^xQLl)w(ouerQu?rPrsjtaOhN`BA(GTSYSsRAc zTxF9!IUz-V7=E)yP7q4gXD7(PEi?Ch>iY3P_D=}g>K=}k;Mt7j_TGYr=`Pw4GH4|@ zB_;S5e2Trm32On0y|QZovb_%(>uQH_SmuGFjN*Zooy0xI^4ZzBucF#k!f0O(Tj?}h zdE&K|bNa&lMl{<;qQ71pzpzb*ad&;s%1B$gz=!^oYHB7+bPd3NeP4clJRFR_6Tp1s zeMb}(&HT9o0Kt(+d_Hr1Y*nQj_TBw{I12B?3G?SA6D9fOfj zEUR-nvk-MdPlNjQ^s*6K(#Zo3wNrZdltWS_4?xli%a6$j)GezmjI5DuoTO|>l)vTr zb&WQE?J#nj8dSYg((|2Cl@Js6i|VFHLa`;$+*cF*;?CTWh!rFHYzPg6$yvT<=^!&? zvgx!UhN9kuMO?|^m&$^sdPtjv5(&U)v8u(2WF zNsvDYBWFXP@gO;5X6-;^jKG1WQMJ57Uz7x6EdDx?Ife4>cjeg6FgKkvnX@#Pi#2%S za;MFFPEANbwe_@5jLp_zGJ>Wb0`1$K2ZhF5)+TvAH0FBD<5jr3YG!tyB0E{<;^((_~Za8BBa`+ zRW=w0bP$VDx}0n&-CnAta^chY{5Q>>yyPzN<1h+i(hu zus<6-sDO0Yj(EE(+$AqJ3F`17prW-W5ny@(2S#BegNn`ax~)2 zFwv28#KwO>6TO8%83aY7W60HjqiKndVn07H1(-2@vpq$n+Rxyuj#vqlR5QZ=ZKqIV zFic$OBJc56rBZMy9&@lKU4+~qt`1%A#nRnb?7TWeJa+NT1Q)Qp@@~J{aiMH#xn6*p zaySncKgH%eEYPQBv!SUiQ;CqMmpoUcMJK6}Mni*TnL(7%2!N>wfKo;eTwT!f=8S_fo#F&NoiZ8%B>g17lVFGkdF;Aa?_55U5{Cp6ab~amjC>|b8k!_Dg~=( zx!e}+3goR9_0v z)QVfY%G_@w@no3Plg&8>mnURFxfqR2kDARgfF;;%n3pw0R0?ZJ zfhcA{_h;&c_DQ1Vqa)P72x+>xXkTl3$w*Ok-;64R^3#Q1j&iK{A<5`0658~@ zlhHOySv*Ar=HDup#H_e0%np$aT=~F}e*nMYqUHr%fRCsvMYgYnl56g3)I5)T{zHB> zNn1GmcHOwdapM5lonUNy6@_wGZD*VsU&&;ar+q3ikM8d+=1VgiJ2?f=WDPuCa-Tpwv1IB1?+>aZ_Dlz**TG<3UJ3)W~RIWUK;6 zklqyqHoQAlF|KlSdLS%-M@&@?J;Jzhlo64IY&u3cZaEFp*aZdFg06s(5_6BaY<8Z= zcwZk4cKLqGj{>poxmOK75jYTAsb?l?rduAtY5#$)JfH06Zoym*ImgR-1BV93d0*FN z6cbr5a`CRs*zzjkxG0=Zx|*xt$&=0x)T=hda$Lnyi>3{N?#XrA@?%u?|E45*$l|8^ z0(Bqic7t5beb%}vuME$1gRq6hqjh9{_$qIWAB9y`kv?V;WYNk+N zO)<1r+iY#ti#NV<7rZR~6z_gG+RPJgt>a4j2hD;*S@L4sn)+32-z1=g z@t+=4Xb)Ct?&mFxd;eTyUXh<+;Vf`;dEJI}%EvAv4}{0;e`v713AB>G< zYEg0xg!x2*m=cm>C_N|aLt15OW28euCiL7wBPV@M=Jy$7Uf-U#Ih6D2PEtPoizj#) z)X!{|l3s>l_h)&q3`!%t@F0_jipW=BC;%@fAnLhVVn45OA$()bU)0WQ}T#x<-;yfyBhIpo(AwWBVuR0 zoO0|E+-*!-MS0a%m;VXu^ATUuzzglWgJEO->i=Z*?Wr-pO&ut$$7`oA_vQ0oXK4x5 ziH;7JyXNK4gI@17GfA2Q+?iuYC&Y z{nBc%81AuaxkB!q=!UpKekZe2d6af#Y@5UOo}756|BY#+UDbW@MSIE5#%iQn4(WV~ z=UD#{f+NQUxsYD5d0Ir7K;f#!a}CC;y&ZUO>NZ*(dx~N`W3icD!7?343u2ZA@b;^q zN{?79PNVE_CwjBtCl!p8^C=jpxZN43H8k_zO6W`7q9M;5 z^S^F(b6ea|UJu*8^c+t6RpEPmRygY);Ca2?d{O0~J#|>dnO%=3O(BAM{c-=3qPkAq zWFRMqPs2+roke~n@*3t>ga8B+$5p`~}`Fa!=*T>8}!_L81S) zWHq6#62Je%R{3$Ox+NYCN&lj5tKD3b*xg&3>8REK@Nf!Q{`zj%SlVP!Y8mUi#c&5t zBCLkh{7HePh@1M1wtlWJ#aP-gmHKb9J~A9y-0ZClv|NKw$1lpl@KFShZkRkBh6Qx; zpXqhEk__0{SEt$k`l^wA%0mKDrpTwi9^3H~-0_+ruYQc{bu1qLk|G;Nj?k&FbrHl6 z+RwOAjI5(Bbny>fcap$`mInj3OyV&WJ0vE?`0Knx7mevGg@o*IdShW+(2VN!5?0P; z!WNl_*B1wy>aqpZRlNoCMBSR;TCsE8a}x`?G!NBi>xdZ6fUG1ALhIUqMl645;0vgW z6z5Yn&ZtV#h!KX**Hm_2+x$#h3U4GG+5HH-t&VT|0q@%s3|Gu&ifN@bRjspVSpM{? z3HK7FrEgz+2As=78o_~d{?=Ma^?70ClkV&h2wcK)?o94}`1&Tj$x5Lex= zj*6nA=SNZ*rS2QtP0FVEvT3}omPRf%7`*jjx(Y!Dlt5{NuD**eh;T+_I9-M|PJe8_ zYEL;+d22ly8Hh~-R5u)|Ua&6v33HRSG&L7*tBWl&tJLMPGN2gEn~{hRFO4(~!co|) z2xP%rdw#}9;(`p6g8EAxA|7xVY^qJcJ73~ z;6TPI&YRVV9dE4tGI{P|;;FL4U{xi4;Q6h}$xAma56jfNs1=ZQD%tX1?*W*}NhSR> zzBN0hqc(t%61?_IYf%q!COH*Gx0113)8<-+6u)Vvdq#5v|u zjSY)`8r(st!=m=-fq){4S)XYP!)7tm*T-y2&2Si9h*%Ifscv_xr3_P8I82dXjnEJac%n?13^enP-Vc;N4 zCLlaUmkwG%UA&6RF&xUFD3($%E}o9O;Q2Gtc9%}d^miJW31oE9Q3=XCqSoZ5dKt{a z;;bsopJG!3=RqMcjYd_mGr>#6D}iaKsi$jt8FpBeR$TQ5VqC{vRqfTcQ2)MHk+%7M zCM6vPXQcb|HNu#9Bw_lV{qsNuxgwiWdJAXg<9TLIt7lFpGs3^0urL(~wq(Y?Xkw#X zqHz^|?vseKu5NjidI&{I8LBFaDJKcuqXR{3sec(J(aE{OO8bZ28}bEP5kCM9z>G{) z!juS-2}AoLl!Z4ogmxjHzlO$9%qW}HPf0z~_w7@D|EGzP&lg`x3Y;h~!p{H-xh9kh zT9s^(#A2k8Y@m9BsuLDxxuuk?102TN+iDo4V{t@&w$3CcZ3$F|mXVX~d>LcKfILb6 z+?;q(jTV7W_^+vMxk2=a>N>8*&uU`bQ_hjoEP|>^Cw!(#FoE3iJLN(NM&jxq#AB{B zN*=~K-$LlZx?Nh5cMSE5;q#9TrO%`FxU@@01GPex4i6$ibrel*ohSm*@xzqGxh?RQcF2N}T zopm#ot5*}2QB3)s9uooIV}aXWPUo|DKNU+!RP6c9d>(tJ#*l{e-slv5^WD$O&iRrV z$`~0&F4dn%Pu6vh`I=&Rlp*^-GF5(aU*7v6n#CW3B8DKY;`}9IG!lzdK*hO- z2UKRZ82+Tt9)2^3xv*%m?Skuv(w%q3IcwYQ3VM;P?{DUjE+AUQx1h!$0eEJ?$(?{{ zK1;af(c($AnKNQfOyCg^L8Atc?k;JE&KqkSYaO?`Al31^`UniXtH0Ew3 zkU7c8&5xQ1OEK6-J>j{CS<8zz&(noASS^zdCj(D_ zWhYhfa-p_E?1k+WLZBL6$>{l5=J@hk1^ZOIv`3P4z7o@E;-vkI5p$+D2R3KYxcDO0 zKj|xuzY>yu?g+^*{6spMKgl?gA`-U6Y? zNTF%jw<+t$(&VLNT^&V;g@>o<@`tW!WP+pK(4LvROBzD+)6T5qxgv&&NaMnMUJAYT z=veHNorqRr-Y4DkDCYqy0(9l3#G`rn)vFPqS&xLN--lc!c`R?fVtZ|$UW#P zl*$B`iRFmCi|GMjG&dbxdtb3bsP*@<%-e}YR9q?H)pvy6T}KfCOy%cK=tdUkp{Yqm zJEoQlLEwD%25ZUq_|IQb)H2oO!pZNe`QLh4O+M7Fw{_2?am9u7z@Y>BEN7hRQEKV2 zO^<_eT&uzPE$n_hhe%c{{nEyRqWH}(qW;O&H@e;Mxt5Y(3zj+!vIr@(&E zT){C~%P>4l0!soAr4#e}aX%Fzo0w$r6;e{eMaU+kS7a9>ho??Ll3d!0r7Rd?b6kE=>xX#~NzI`SDnu#T(KN$0S%>$bxyoif3Dm!0S zwAm

SMCj`Os)`g!KLHDjH+VpKskCF}635iLo7~=prfeHtMVZ^6s-ewlW>D zQ*^Z)>Uend1_Apc(4tCJvRpd{GRs6>BsRIH&>kxV%DQ}3!|P6L9$H-0GO=|E*K2jV z!?189Mf+#YTr1+mWKhL(`#kH*7NYS6@CPyNcM#&}@0kqrYoJ6rnvii-EU`*jJO%>FM^34yXbZ0P^3F2Tm2Qo) zw`OIiZWquI*e!$)oEV5=7YYN^S_%fYfsO?+63gYkVkj@^#)tcfI~0cVLj*#MDH2RA zeO4;VrPza&<{#|L<)@<~5G$x<*?AEkYmiI*KNo4c<7ROhkeu`13f%gf z{&AP1ykF7t34sj9GU<*n1~`Cj<}AlW^kJ?4UcN|NBb=07Xe|&8CEXZ?JU5WL1q(r^ zP9l_NJ&;v97ijUOUKB2ti^OC z!BoSKCHYmwBKlsjAl*Wm-f%S!|5Q?pePJ`)k_?5Q{3fg*FrBR8=VFqH7AwPvvH27E zY*z|eFKH!>(c8?T47%ZJDUrNkOj|!aT=GcdZiey%j8$f3-(~952!3!=>;h($oLltM z+6?|J`?|Qzxodkdc&-z$vJkSOy6t4rA~B^^K4=opuKm)A!7g@R9Lf<^@<%wphe}J? zDNxx4_+^BOrvV6)nr2K!aa4wfDRV9xkR8K+v16zcLE&?aT@T$8(j|9+EA<|3oG;JN zNH~_GZD}!|qy66uXng;+&eN@8I&(7wcP0M=&*~PJA3(!QG{cyut>IkHqgZA`!vW_K zzi-0W5SR&BI^I(mnTqG1aW6zdM-lxCE)J?XXjss7!Bi%f2lQj`*M06=g#5b(rl{Z( z(i*q}Pqhp@bkz6bfkvtv7V{roHf6Il*z!add}dOZ&-f;X3uz?$!KYeJby{AwX5@^W|gY8&@NH1mnqWZD`18|mbY zTEMgr-&!U;wK? zKuKCti$@FFCsoZpQhe+Mmcc1zJ10VWF#)x6eak zwr;}3DL^QW^;KTxs4t)>=N~Mz*%j&-V2RiGy)vN6w&0gcbR$6v;c!7W?gNghODo-=EGE2opNt5j!#ty2P^Qimp+p>|&87 z8zZ@6($Vdj`Pprml7vmzEcGhh!1EylMla3K9fI(j39ncha~_WmdR!P0p2x6SyVZN< zpHo<@AgbnKUe+zgX7E$lEj)n$;?t`-vF(OwBNrc1@w^u=n#GOFvYK-)e`+2?6~gU{ ztr7F#kjp0S@6WRwdhncxB-t@u}G zb8=S&nGWRCLsX#w2??|m>N-Pe)nbu<5ooWV2(5H2%|9bbHUoIf5O=i^mo=WZS27ye zGmZ#&tXTf+4D$_g3^3$CjO5*SMgcirQ^>gL><0|4wc+JmijkDrWcGUQsI(B7AWbFRSc8kdge-=wCE8>lzh?Mm@dIsj2(*!YTGISP zCSBREs#=#^_dQQ{;xGmS`NT^s<<8E`I3aqCB@pk#5HSZ~7s@)QWzE9{85?1dNI80y zKy43W=-=$1%1>u%Dw>$(J>7QmbC59$9bsKO%s_OZzLTLcdFCRp)i+=GYv$(Pd(6$K z?}-u30e)BM(ZfU3ep_+rc3UB5WBA1T$;fmX<$Lr<7QPCE_=s2ee~p5MqwpNZvKEUv zT}PrVhFs!x7Iha9CSx173?dlRYMgQwdlO=&rIrly6mW6lwZ9nphvIGkol01$k95*v z_grRiRt17Ij<$ETHtzM240f&3(XHhR8L@Ov8(4eewr?JnEr^{K#X}O= zOBDHZi(89tOC%?gDwP%4P_^oMSQWWxzVj%^=xCwubnn^|GejQAr_xor+iLd9D=~Z8 zYI> zNIH5W)kSo(EXiV+a|`lGdkfKGt*8>v`Xa|KuRGQMZqdB}5WP{jc5iZ)S3==Z+^JL2 z$706>j|Xwjjce&JCe|f{t|?SCjP-jnHGq3VT428`1asfCNqNeCn(jsa<6F<@G@agTIq*|JuyI|%d=7I7wm0fw}QfUk{l-~wR@ky(p^Tx%P7EGA1 z$jpbwq2^W{_#$OW?`b`#@4TGz#eiI(kV%Bs(yX(XTMP2e8DW3?Nx69@P-Y{!v)}|? z>2`r!Q){@SxrACz9rj#HWEQ)~Cw0vTGvVr@&8Cnx-jgZ^E!eEuq2K>_xikZ_lKyO# z)xE}8hKWzMS>?w$GS{QN0(h06v%cChlp?av6%M*_U*wlEch$R~^Y3xv#X?)^QL&Hm zwY$!Tr!R@}lsQlVMo&*>(&>v9W#&uMndSriLxYrqgAyo!n1DkSRd@LlSC__m*nBOt;IC+8S?v6vYgw=;T61^DgNzHaw z!?E*D4wiZNnr;q57)!1l*)V#ExHnu_-_;SIcnUW9gs2h~3!$#U?@xXKrTq6WPF07v zEd*yv=DD+ZBlIu?wY?c(yfZVa++O%h{4-W?hR4aQ1h+3dhc4%$zepsur+%gU-E!gS z_5w?P;(*7ibSc-d6}jCa11NQ)(oa2d2*0sf{x`}7@qj)p$^6PSh{Q!;EV|T7)E=+d z>K?!&gO-z<&E>kQ;sakHo8{B-Lg}|^)}}$-#}Kv zicR50H$GyiT#UU+7qs7)=U}zaLx4s19YdKFL7d^F)P+pB5a-q${W}S!ac;5+XL{0D zn>E5ID7a=@>TfVg5LeQa>}>kl@1pRhw)z~TcCB^xMoJdJXv=^N0e+mEMf%7u5@j?y z1-M-oG9U_)z&fdS5mv0ly-%~FhlV?79unIA`jJ+UX;wBERRqu>Kc}SCAVMp)V<=6h z8swVvsBGl)`1=3*jEOfDG)u1L!q9Zu%XC?P)7P^n9L)dy%$DQr>&VBBD}^ry2Z*5= zs!eam9!pCk$8q4(E4ae9jNGPJJ2>F9xcvKLnsArKkfBDa3JUe$OZcunOvyn?bSb4J zC#-Pj@af!3G~Mk;?yw)>qr;h+gc9zw4W{u{Vln=z3^S;po;KYK=5DUWoZQ1Oi$yXk zMyLLxe4cZ>vD|5nB}l3910Fv9bHd=)xcdaS6KL2!USY4~_w=yU$piPo>GWUAd&|JJcR zFQ%%gsy(z+8U0{bj$B`UcEp$A{I69Xbc=#=0^!`YI5(``q$GAS-XCrl>%rB0RFS^~ zsQKb&g$A(nw6|%gom!D)|T2(N%%XFTcI=w`-qLDHNXP~b| ze77eA1LHKpRwO?#-_3Y+s= zR6awa3U79#-|2^L6`GzP3b*Ep6`qQjY8FVOt+{Gg$j+38zJ0I|Qn_!M@7XL`c6h>! z0bDQEHU6WQW;*LEnXrhY*nyQdK6=Aa9wOI0L5$X@Vzx#8iJEASIGXOO&o?$Z|BwIvNH^911#+u}e%g?=pTE^XMJ( zx0aXw0KU+4emst%Q{rz^x(m&}RBK<7vO7xda$82*WqK)1Za{ROavR4zoK>5vbaY1i zX0TAtxqdgDx7_}GasDiDD+eW|F1}4Q$=fTh^9u7@4jQVm-@3X6j^@hlGruCaC3b2s zv&7i!a~-5{oPS1Fh&OyWDp48;IH!Xnor~Hh_Ggc-{!ojZs@9zJP`~lO04Ft=J~FCctVEZaJ~c$B0K~k|+?o4c>ozxSCGH=EIZJnf3Q5k* zgoI=cYrOF9=YC)le|~#ys>sW{f-feXE3{ z4xvk}kS0^A#Z@aconWYnYN&dh3-$G$Lxsr`lB%~S$;9h=K5}3{l3`<+od9NcaFKhT z53L*j*p~05Rs6#36}?t>=p9Zp2h_~BtN+OUcVwKS1z*4E3erMzz%HO*M9MWHOkFI7X0L5G2rCqXLI!T$CkCV1e4=L zLduaF{`rK5Bj*7Sn3%3@cZ;w+6R)CqWhe?a z{i;q27)L%kIJGcu$UJH^GIDHVX4lUMI?YF1EIOWH(>dIkhlQ<`6}b8HH4kNx)&=UO z0Oc~|NJPLF@(&g&-m~&sG>?Sg(I1ziMshD>7^9G41?PN1c;vYnV^Ny}YR=#|H=4v` zcetyT+pvylVG!BWnf2OJN;uKG6NfWl$@guL ze5NMTa);)iH090IRu#M8)EupHKG6x{#$wUnF1=#bmv$vVs0wadT9&KCiRTawvAiH2 zv{$>ZHsM{keKorEEWzxIV?CS2kG)POoOgL!Z4KkodFAD+ob+$&D*wYkvCBWGKZ>Ef zt3Mv_|1mzb-NjHe!4_H|H%xO~_`qHCp1u`&-=5sp{D5axZ_TJ1v1gz{8DJZwN~vfz z>-opVv0UVRuF9PuTg)3^?PyIVGajk4D5wz zpI;x)TfsLK>*#e)-i7T!g=X(4Kt_YFD~$sx?0p`p%3`2S{1WK7)A*d868k`XL|J~e z6nEaS$D*&C92OGZU|bz98DYr+2{Y9w*&L8Yve8ggS;CoJeWRPzg>5YwCbp7$3K-LY zHzH|KGs^uIlJb`{|Hb)^->D7Qn-RG|9QUU$Xn3nOW;c(i}Q=`hdG{ASNbkQ$lj zyp%P#t|~^S#EFgFX~^k>GNiwDQ5EXEL#fTkyP!?dNymgT##=Dj-&x-{564Wmy`<}k zfq>=V8Rkc8o-iS1}@sv(}x?LOnd@N6c8qPa)zKUkY!Tf@8}6w%jfM`T%8n&-%3g4GJ?|diO}Hp_$b_H#IFA-#@@76A;_`wu^ToduGowY zU+en>Ch4zqvG9Ln6Ti4-y)nMXA_?JE{`<2mr&BrG<(ixIa8IS{dAmWq%=2do6 zUtRcm18mzsLhqcqZ6uX1Se8=6G#9t!l{VhcG@h}%8KUV!sU(Deyf9&@UYs zRI$_X%0e(|&wQ$M6Y2w1I%D_=GuJCioSmjq@#Ms&t1K=V=Bk1fkC)8<`s zD;G)$IQra17W!FdY90a0jWK|_c?x?pxaHoZ*9~Z{+2_|*)WDA2BHLWXTKb(vAdQiC zO6V_eyxfi>(5G^uR(bv2Uq*0kLFA|;ubho3$BlUFoS6k0>yWAhnfyIZ^4( zV8}98@v-1A#Yy8k5%B+&?rz6H{?BIF>5VBpojm)ef4E{6(sRt@-7ui7)ty)8vZ4nIC5ZZaKI% zhX!F?UoGP{&SD}RO!Z`~e6s4PJH{#0;4%%WEb|^=XX1@8IFw%A^9mc3`xA2z&Cj(B z1ZqZ0|ba%kMBRDK6?HAuG5NJ>=Ld{bwMm#vtUeR0_PL0Er~-%lxV1_ zC_S-adZQy7wOZVUsNjl#^3ht6aIoVti$OZ5NBLFz215$mHu(3RNzz1;bxCqg%klLN z?bfR#&Y_*7U>6!ERQAvdTpiS0>O;I@_I!I=&&@f36oOTWeZ2GaqU!H!k$y6D1dN$r z;4Bag>PDC(UIn|O#a2CCRPjWhmLEgl&0O_*g&PXK|7PnY`X@%}g#-@Du=^{yus@;= z{aEzhWwGW`NYvG)AuCx7hws&6g7X-LrT*J&W5O1-;BJ&&v*n|z2>Kb{>iq~5&f^;f zIZm8rvZ;k9r0$1bXd@h?Uju=j8WV~#u_9(42W2CU>!G&OGpW$^ z%NDdLHLZ+0r`&cY(PLehN*!L)eFN4LtCb2msh+dW@aI(pp582i-cHKlK`=I=KcC62 zo(Kb?vNG*V=n@-KzUW{2DT2ZJ`)H_aKdvgCOtC)eJqeWYegK~Dwu8m3>++O1R4_YM z?O#QCs`Q@jh@rt(Zxyf3X1=6_EUu>|IRl>v*_n&jj2p@|D;usaYRfGxG|Q;z3$dLz z(yAjPQj8Xj>;CWXxYA=&zxltV-BL`{QGchtziS*RQpc*;a`4XoJSt$MMIY(Ri8~Xk z9&_A;YB%@9jn%UGuazvSFwQT_>=;d3u&iam1P|>-{0e&hc$|LmcN=SA=z}#ZsgZs{ z?kugy4+7QLM}UY0{QD0le{^Sl{)2V?PX$Z7A46miZ94$ae__-Djz7{PnL#8^nVD|Q z3E|I9x|(I1NsVEJb125(BB#3d9Pu5u4Hh*>P#xXUG*3M?|IBJ%&W2^T>}^lZjOOy? zhue;IPEqe^xiuM?MQ5L`pDtH}*=vT_!n;ny1tK;(QG{WJ4x_R3TtA$%9^N2Hh zPvgk3coO1eQE?5q`kx`+)AP-lR2gxl?=qqxt(cJm{r1TDVDBc`DT()_5~=x4UHB7P z*#E+jR!BMPE*j=u17K)!ut*4oG*^;mtg&}8&am%MFkCaxCkol;^Cn_a3{EL_k$5(< zLX}-~Cv_TN^oDfy#>#zk&5h7NLLz3h{lZ>JetB`lE-MY>M#Y&4>)pftT?MG zhx9-pO5-IY;42R!OEL4*;*BA6tp`K&Nyd&Kq&8j%X^ zgxAWyM^>PS3jr?I0FbPr1H6z%(8|;SCmEt{q~}eY?h`RoUHN>c3pqX&0k4addUpP* z$vAwMTH(6s9+8#K8B{EjAZx^m#)@R|akRK4h{3Jm)#qHJPX0c4rA)~=)?TYu^j z!K7{iI{X1JG79HRi*Udt?*0|W*Xf^@@XQ^ew2uoK8H=J)bw+g2$e-+l(kpal-BVXw z-Ei`dYZ~k#w<;5KS#w(m+#zD4*q?cl>?m9;`RSiw!@79S`72lwv_2s8|$m@6<&92|pYl6Y3?lIn8HC$}MCTB@um z=ultO7+X?2j{rfk)%dPQ%w0;u|&mOUdPTMlZ%1SL6tyK0DwQ)Odtyu4E z>tikze$JMbGUvDnHQBAK^OiDTP3ue|NVi3F$s|*Vu|dDJC~BDCB$|SB1u0mo46u}> z3$Yox*s$N{fr=PQAWrO37CI+>C(CUS5^vUp1g~>HoCFs6cp_(tt`k2_&CCR$td5bDm=NQ0u;nGzfNqM?6@>Z z8{OqK+y4b%0@}EYwlv*s+DDUpM>6;4TW5-ad?n&?9@R5~(7ES5;4v9jbf zL4KmpDK>jtI;2ZtSR7JwGSj7kAYrKH^=5v^1x9 z(e6m4XVe?2jjGg$7`k%iL3T+bXK?YfG5KBOfMigjHXDz_x`Ld=t?#a zG3wM)2ft#*kRDG@o-hxog@DlHYFqROp%AT=0SyqGq>_A zTz%TF$!GijNrSD0XKgGn>Nsqc*uz-JvJQsURyb&b&RdmHum^NSsdaW!ATEPf;K==G z5#}bX)dJDRDoO~8EWtW|j!#7VJ50nDyWJZi|M21lv1-4bGZ?61hw{&I!j7h##SNzJ zI(&wA!=w4zS@BwSE;$gq&xgpjiELe(0TR^k;`DbKjT7vsLiWhVx8F7!25rdx7#32r zoJwExvgih+zJu-6M&JSAXyOpF_OYier}ySCOJh+XzsI|jcPAI7>>#O_I+e+e48#|W zK-g!pjS{{zFo`N~-^jMo%7L53$gCvhW8$toc!8@Jm=$X43YuNR@AYATxdnDaF%gY$ zHEB!A6;(-Hzp9zs8-uh(Q{De8L)}2DTICTYf`Ds-myoWsz}*)nb4A_gn{2bw9Z%Ow zYrQzCH|=|>tn+%?hGD!rPsC5aTEE=m%Hy|UT08~pLlcmppCYhO>~`qs8a&R-L~Y-+ zhY}=5*EhF~6TC#d*wvE7SmelxpR5pgYyC%j5PecfYB^(mrofN1$~AWhZj}728PTrL zjg0@S<{Na@`Xf$+U$j761R7O!^!Dt$aeaUQ*J>heHmP7CPKynNKLA*)(Th9O|=|W*~91Ctm)(SGW}fSm{So7*IcBQhZ`$5n!b4ZtmH6e zP@;-RZdn>UctcsdptVv}*X*W;##&Lc-a5!5z#eiSJ?hxwmQ+K-16)a7KC$r0Z6&02 z0EujYiah{UJbd4qaj{Fyb<2IVFmgCXl=^3^a1GkVru6zkAF=<=PeesB}n60(0 z8}QkDfP*zG#N(TcTuoe!3N$0fP}#7IGu@*pheIo2V19^us+$6^7;`!_PQZ3TdMY!3 zbo{g0ko)kET!>lE_o16XYGgeAp9H0a+M3r@Q2rztBSSih;teL5jfD`~u47^-rc{{1 z-S#k*p3UEsz5S+VV^j4WxNlS$O<#IGk0ykTjm<0VX)BP}Kcd9UQI2-=N-5TxX1Is} zNo}4C5*HbKVIZ@YQ#wSWkCFXr^cM;d8-MLW(d~|rLd?WUJ@g+S@hIhl-9c-b2TR$$ z<3sex%St8!`u~ax`n}>%I?De*VPgd`J0Zg;MXx!~G-3j(_8*&wnti0b z@7MFp*0PU0CU_Ec>6(7;@WfqiqF4w z(v3}T>rf^Tmzz8%G~&)DTFI}gbf;A#)#ou*Q9#rDvA#s#-)~kEg`_4svaqVwD#uPk zEdJ$6M;ZgShHm@TiiM z#jbPaloXB82?63|=_-s1Djb@pt+@ba+jt@!UQ3Q=yeZVFD?cdn4aC4eNpaav@f_8vIxnfrj<9Esf08`B<>y4+yPN+#`zP=L z=1G_)9W5?%vJ(dV;ip|w!FVYlj!Sz3E@@nEm+qZ<2>T_}ve34Ms3{T_)}t)9LGGag zPWQwN(ZM=FxdP)X9j`X3@yx5_e>%0e)U7*wngycP@yrG1zx?GhYB^e|ZfiF6S_3eu zB5!2PZ)MpT2yb;}(pu4$mdA20h$0oK`$i@kWr2-`sg8sZ2U_4_s)g;WCIg?zIB)nk zQz~if)hV%MhDOwPf_RmC&7amV)vM~;-=i#LogWeK_QwI4-GYcqhPuhX1G%HX5a+PI z>a!EAFF76P&2u@!yK)_I#N6PIcvL<{n`1JOiHN$%1i9!K#}a{~$%O_h8`zmu;wMsj zYvJsJ{-nud@zMAN*i0?aDb(}X17QEn>X$71TQ^9Yq#}Fj@>rTs7C5g8;tSj$*vX_x z8t5!wpN@^xqFP4givGMuq~_rZX{f;Colj?f$>n~&b0z76=N2Y)voc$rR`W$G)8*j# zZN7k24hE~#1eTy>M#VyhOQBLZ^9Br)RQ0li{a*d&IAXn$$}$pd8o%{;Zr7S>76EPJ zU}Uf6DCGuHHnezZ_W7M zkw2oo3tBfq<3VvVNxZJg(u&28!KCg<#oMcK8$y`#agpxmSL{FUDq*tpsdLJ3=%)H= zQtL`_=0Gz!2xL$PdOlfZqvSvj4sJ!<&p@k(Q>gcyo(rVu^R(UcHwN@A@PJi6F|a*MX+FH{c>xZqdNY;8Zrt6 z#~|LN{%}2Y&)-3rE)o1R?gvF*B&6T_8+H8onaY7-yK#16|C7jz%1i0VSt?ZrJrMOu z(##{3MAbq5;a3|}u?^gSP}#vihg^wd4Ynrj+w58SL`)^;I^imc9?Y>$*3FB4Y&BuY zi3a0yKn!)^HeomDp;FfX8}+jh#`G^nv0tb6A=nI04;sC^55H2^i1O;(zsr*=VSe9F z@lCYw9xxK3?r1jYOVIp%Uil!7k0%!;#N(+O12=oDKwuFT5N0@$vg@jZN~0I%64iq- z6gn&$FgfC%GXj%Ms!3uY{E5nF?x@Lki*X;{66B{0_v{$-Bmk}U)>`O48YZXEbNmI6 zzWKY?#Ogxs6a|r;IEL7UAROMl5 zBu%3a8-#n7^5c(JN=HfXcmi_cp9RWEOZlYlpT!+5y##)n%Q&PaZm+_VbgeU?mP z-3YsNt!znKjC$kTVf`@tUqPJ-iWN`O?et}nCs=C?CREZShK1}z4CyEDd2`nfAn}&G zKRI|U80Zg=5Dr>R(8nMTRkf4`I6nThYw-8`2H^;vRVfH1qdQe9bxL_eQkY*h3eRc zei!yH{%_&E5|UANy`A{a%Bs$&Gjtw|kYpRp>MaUQx+-`*pB^A?A*as#P=<5{I-I3| z9g+7oJP33G;NFo+z-WQu%zld-yBjDA&y!SsKE8cxqrTSX(D9_wKN6kvs1XlGiD;(6 zis3hpP1h9L@_H6XDg-IlMOc_3S*ny2C`0#WTNwJNw1iT^XwhZLh&Z>Nd_T+ct@QVIU|r$Pn0NF_9r{ zZM)`5Y0ERgY(P3xYF<_?H8RCvNpK~$P2J;_IKYH(pAiUo`Jc!=>l8+3T z_5k0*YgK~gr063bB*t;c);8K_m%XZ3!Wd2wR5mZtMgKV-1bxVUvm=;<9|TJ`9au3u z^&5-4Cyduprb%!oePkOq8>YQzIRO(+#h?A|kmA~(QSTJx zpm^9>DCFlu%RyXfK$aZDLCe30u|A0>ysXf(+GoPv1>6efr2BzZJ4#)Yo;B9ezr3Xe zjUC6yGc(^?#Q_o1fzy4cVr*eK5I+uw5VTm{1 zwOl*!=T2tVL3`H~#+JJvv-A0KiFZ)krXq6b?TNJbr3vb@^;NO6p4FD6y~Ha}uVlgf zyYYM0b&$6c{v5~a2efnt9G!ShciOWFpP}B~AN>tJEz+J{6HZnWp0!6pnnLX^Cog9Z z@M66?2Ao%mV4t`=yaLZ_gbzR6GmkOM_Xh5j7jcqrzr}HzMCbw)dSq* zrRxsNdLpCMpsVV<>mF>pR4%5VLMq8}iKb{CDA1@BD(9!oQd2 zeHne)yL>CquJ~OkSkU^cw_N-NK$vJYz0caF;Jrd+v)rF#X0vcTX-$4@-W~Pr7lj0dm6nlLx+da6%!oj=C=YMM@jy!bLUK-zHS*FDKJ!-x; zZAZIanT+UHKFet*EFcg)?2jP8`Bz-vKyiwOXGviV_w6sglApZG3T_C0)6Ex08@|Ti zN6+qz5f9CKP(BE)X?P!pSX8XfwltYi;#dAlJ6sVl#VzoZujKj%;1=1=bX~}$yRcm{Yzp9#XuXk? zIJek1*~;fY}kX&j8$*K-%P5oIZPzHWzv$?pJyv zHs@}#}zH@2PnCJfp7-K3sGIKhas4TQgZdR6u z+faZYagyS-)k3sSb7k0O@paCbS$82J_;KIV^z@4e(rJ&iCd~LJ#i}@wp~3Y>2E7sp z$G6RjL0Q*b>6cUUU79opf+nXIc$qO#OW(^)c|VJ8)K}I2i_Dve+Np6AGMFG|4G64o zo2sJim5PDwUH3se+owwJfh@s=9WqwOJkG9&$cInd* zbuI+8xJ-8Fi@M*QHu7FR*OqTNdeH;@XnF7Ayl13!-MdnL&sxot(Z<=Q%6A{+-6^dQ zRl?R6~7`7{*bquA?$x>;T=W2K+gu2t>n+^v>VQ{WdZqGW-b5kEBVBryq<%; zX#~R7a}TfYCdW;47SP}=e??~l?<#Zp!^?uW{|;E_`#w`2{+P&XAi8wXjC!&b=ML!D z<|}JH(EjhnpMNXp!ViI73RnVa;EMMcw8a_i>s1Sl3hYS|p8+Y6h5c+uABM7LT1ebe znv!M;K6<>Q7ZAu5bd~dIDH2G5Sy2OI*9V8y%me>?Ya97sxG=C*ubG!#IG$0zy>`tb zBSzUUQ}Wa#az$uggq4op8m`4KnS0!K@-%le0hA>S4$($f#6xyy+i8soR8ToyBV7xL zP(H)tmNZ>;-l^hE*5r?|s9pOm64xFQMpl=@@O2K+07`Kyf)IqP5#(G|Woiz_=D)bU z64qzV)|;o{x9A*=e;n9b_FVx9O4j$>N*1${8O5~A+*7?G%GvTSg$2*WC*f#z=-)l{ zuoXNX3>GEj%3esp!+i0v;EhDK-pa*Q9b9jPENxW0Gqs{A?ZaJJp1dBS#U$_?;`jtH z&C#HxdyyrG$WjdlrlHH?Z5ovWc*W|F9w+p-o<75vfZ`{%@i4q=BiwvNN$`R>;=c;z z+S~UzHJD%)MW1T;rd+s5YR0pegI6*8{z>AE^B({iaBV)&6d zAB*6}Xxa;S~UkkhfW@lcagHqIiE`Nxt;<&iW$OTOHh6ChzchG;;;h~ zQ%0Efd;{JWpxY;-H{2|2XduZw8s)Ps8iUZY-MMe*zwb$ix7!Xxh|z*`}r)e2(fl zVO=N`tc$dPN^!1D)A;rns}R427IE=ZJOc-QTa|+TaXPzTF0G3w5B96=O;XYNmtehB zqKbX0mZPS28vK&!Dr(YzM%Nq%t?g=BJ>#lMRGpny!N%Kt>2dx3|F(MQHm4U`NMd?>h-AdY&<0#6nzI#d1D#y21UKnuEw%G$|L;?~8O2x4b*I_I== z(Gmql8KzGJs>2Q+)%0O0W#wa6hNE#_Q^2EJOAMvIq8(cRyHRFjZZN}dwK(d=G%wGY z@#q@{y0E#F3$40liwjdm9Th`P)JK*d;z5;?>?%^SL_DbNlb$Md#$p9^jWeZ~HUkf)4x_qa~6Z2J)e=40I(2|?11rnMHav3Ur93<$d zD2f2H<~+Vn>@MS5tIU^Z^Y(nOsRY%w$1vr1kJx}v}Ta3cG96osV|28&#r~1Si zxA!}$|L*L39}Wao6D*Rgmctxc)CZeK{q8K+o@!jT{U40;5G6rxrx4AbvT1r)-`;Wx zJCFOGt^b2y=$9AXyBzGOG~BV*!Ss9Ts*!LHFX-xKEaZwmvV?IS80X`k^>9K25xA4< zM{8)Gq4WsM+bg7(D}twhMy#2!DRj#B7|E%rFG@9tRa+B5vLxQyW+QfB$q_AN$Uzc7 zy7gi$6W#5W=}_yW?lsLdstyuNtwW^kfnEf`=G(EadF?DwNv>)D%tj58!eXfjpQM*i zmmYvK(5$^ax~`k)dX2Q&Qi|duny=@$Pv4fj6JPS?9JfvPZ&%-?ulqe{u~K%9V?V#R zyp7;x_b3tfQ{Q*R^51`pUvBZ8z%`J7-|uUM$LsymzmsR;$JFI7l9A5|n87Zs5+kui+<9tgN81yw0mT2fI0Kf50yKU_=8 zeuovI;e|(5ST^<_%{ldAA7VAwnyHwd6O1`FMsZoiQ#vjWn+lXfN3k>Uc zF>+cQow*7DT zY&HHv*%5@KRGFcKoVlAWl2g*cK+pJ>XSzfsfjFNQQW$TidMJ%Rw4>6ra&5eJA8i_1 zz9F-OPs$0o_RltDvU%FnNSyoUg48qce_*r!vCrBqM=T!u>N5G{U|u+4yZ-xo{?9Xe zpL597P2(^xL!jN)z)7P7TATlWfJ7T~E zyZr9~M~WbW1iU&@hMaCb0AVFIr6MK+B|4@-WSv4r)HJDtEI>-i@S%{qvGi4=@|>=- zm#z?2#n0BFznpzVOP29#}fkGIv_otMV4pR)8atOcq{gKM1xR|oPQYO3uBLdGG zT`IX*T1W*wN~ejJk`c3&G9#+e(4rG(Q5aQHrbAP@fg|_`S;`5IAg3IK4_Quuf85%a zH3x~p-O4P&;ti?{&0EAm7i2s|5AD(N#4>UGpDPPy-Q0*Q+qWV4NLi}Rwb@VayS$K@ zuS>U&Qzq)$qmoCXnZy~*J+oj@vH+OLSv%?d7b12gS}sEZ-hBfpsF<%C<-P9K=VLbR zOP)n_`I|Gmq2?$^m_67SdiLg~R|edNDc`tN9!k+%6Q2uk}}EY5oD z186!Gj@eKK^OVuyX%~g#Fo(6L4j{REHc`sQvOK4U#$Xa3!wG+ePtw%S)%w~8k zUO99DFm8wF1g8EZW_VChz*ha{hs}$gK#M0wl6Nfn)U&!ak%WyQed6y41&+7LIapwr zMKVc$gGv^XHkVFMyB43&`p5bo|H;U&KpL89bC~Q?<99k0um`zNNSrQ6RVg=G6&U`~ z#X3vjfJ-ghSr|R5G9eM$v?mvu)Ue4tvZ(<8b1HMF$YG5o$p+fa?yuz!ChEHAh$N{d z$#T`4e3N2CUTwvlYLJlqaaEgO=%RsVYJ{#EZgZ}xw;D>?t~J>NmCetnuY^yp8e8o2 z$$)#t2yfW~>y;%~OL5_OIc-BTR8zp3_bSj&bR*0`ct#d@U2I;?)#N*@sHP?v$IFX zQ|NbgM&vf_j{S8(ob~bexYw|4$0(-B*{|QNpXKk@+{dHS6O8@d?4Pm4)t$fM|0q8T z75ClI`DE69J$ol}e-U-NvGew48H2(WP1B!ZJbRnm>3b>VI@j;%-eeCS&+cmP=vlM# zYELKr_(IefaDVDuIyCAA{@6O4zIe5Bv+()j3!OZ2e*4%owWhOCo{$keCv`E+6CrZLU=$=kohF?n$*%|wh8j(tcIPQ{tt$rL|| z{mjnpdDwsN^)~bMX2#Why45+UTEBCJJcR$D*MFTm$7Po7Y47=R_d4@b(D@G8_0K0A zDO@k{*!QipNuUb{8Cs#I+s8Yz|Kb5IBdR}l+CH19FEKJduzPa2uJf$7|9$jp-uZ5F zhYNZqz286kusHk3+yCXU>1xNC@r(K8BRjvKz5i`%=c6ti`JLolN*!B2=Sq27pvD65 z`8Eti#Y7xmt>=~)|H|#!&TO~S=$`KHvAXqyl)mBHeC4@bf~T_BtFH$1t*7}we+wqw zvw;rI;p!s}j`er(<+<}+eVh5g)DHiJT(h66|A$vY)O~dFggD#&3VC`{peIap_3~(X z=IOZM;cD}$&0V{{EqZbo4$wI=&&8YdxMsE{m_ACrLz#29^gS_nqn1nB!UEKJ&)XU3 zeIU?(x6vN>(@*QIkf~9$-`6Ak2b2ER&Q6`Z+1!teo-Mb|?5(YrJ_aWL{aHpcuGsvi z_VoJC>xGT&*8Dg>D0C@jmZ=M!=b1}i-YYZjula|=VPAXy&QQ>OggG}Iy~ld!IoKSM zv+>q`ujU8j9lp(ZBk6A<>>RMwW!=9YnV0Z&0UzQ0_{!`vmg$rJ$_L)h=QDwC_j0(i zG<~P`$ z`}D~~yTEX(FVFpxIdCK0PxBS>ykv5%d!tM}#^+;*$*jJ!Lg2ZSm*BGhZQA~)@2ks4 zgXVB&uU|*zg_nO=%j4OGEfpB))Yah-DneY1AX+`Aqhte4l?a|I|~@T77zU12KM(USEAN zi;KaUP1aV-%pYF8g&jG+_^a(%xA2GQ;R_r0w?ABG-YtDS+V)2)y=ZE{_#T-Sx9%ge z?#-GfL_1Fb!VW|?+1Z!k_Tb{f?(U!GzQUhlz8&u^MYPqo1ZuPKLj+gZYf$os=OgY0 zZa{$@50IWz=iawIt^h)i;rsmA>$|t%8DppbFywAu=V;{1(}3ZWOJY2N#qy=~{4?nc zK`&oX?`OBIU3x#FcYGaKLIUj0H?Z*i-m5{vLA8v-C%m>3RA+zTPhWnl|J~%;`f=H^|H2se@`#AAjWIu1-zxfmbl-jz z|E~se6GyJ$z-{o&@6Y1~_}dQ>)hb`<4&CgvOE>^a_M2z0rN)5>CkTJ|^f5=yZfd}z zD_pD^zH&sn`aNAKM3L^lJ1%~XwS|te^?6fcIXK<@m;I%{8wQY85OnQ6@QFf zUR+%v5yAU#N0eWq=^R3}U^ry4N+f)5}|21OBOn}`D!mFxqQ81uyb$c7^ z`2~XeYb3My=uI)UkQn606U-{kKn>323vNl{_LxfzV`RyF{nVv;_Zkr>W*+~m4$kug z85E%SiI{WH=wT|kjVH+&q{IvV>z@zPi5bN3bFcpM*19`LJK#Hn34Yy$HV?E()1-hA5LK=dDU9Y3WT-VA2EnH+~nQpxnG zb}!2b6H=YIeysWJYY#zv5?RJori0hoGn;#T{+#5EBY9rVT zpMnLsL#v?gA6f4^P(A<15Jf%<>-cmAIgkBo_0*pCPh>Z6=eq|;h!;~RJjeUH6ZCqM zq<14o)!L>B*8I0{Je{;x;P=olnbHHlJB*Fjrp|8t=|@4)9s(~9{&Hw!{maOI)(*dz zJlbm*sqOusbqNmV{v6$%UlJLZD0a3u_OHB4cIDa{{0-YQyR&TnNwS9+s>ecQ6)#tt zug`7$sWAGiBz8E9iEwrmU^{?J6j6YiLS&OMb8jjC-0V7AGBw183?q%QO4jpR6vSxS zI9r?K*nQfT+=Nlc!^#|A4WSL15b>*ry6;14XC`uv@yyiqrC^;MyWN5{EvmM@u7ADw|9^SkRmU+}bCk-O^ zz4E<|YhU#JWpBj3=wp8`4Jm4;Gy7u;uN#>cW@7{$+yvCkoGqM7|!v>L$`9K>uSL+U=$-(|(9ish2gRQGYh=R!!7&(1y%SBZAt69(1U~M@{=I{6(R`Zbifh*cP8Q7 zHC_@dhx{o=LZ9*|uxAFcY6Hda;5}|39zV;smX_bE{}}gG%-g+JCA!V_3ahTSTY`|C zkwuZcodzH!`%X!ySv{715Dcfy-5^|$WS=OLEgDv z8GqV^z7xu@8dM@vzW_jBf)@IKhf8C#^B&wzG$Ugl%F~=Iayyj2GG>V&0E-~IwB>vdxveqFR|L!+93kcy z!rp>p>=~Aul=0fwxw|)=!HM< z=SFUdvCe3_51Ik$_4e7*Q|?U8o_WVV=0H?%Y#h_z_9D^Spd85>^y#4%VNkcCxHyF! z<5SJK-;eGJ*GSSdbf1MiFU<314croup+bk_FgF{y0)!G!`R?ubi7p8RSwUDOkZb!w z(BXHZ4Ad8Z>T0Z0D)@ztAz8f~KC4W_NUKS_3Cq7ii9@Mz8kp1tU3@xvw!BcLFrUZv z-GG6x{URXVNDp@x;2=e<2_$*ZU66i;M~?qKKm#zC_dGR0xu7X1m0BJAh%fmJy}`35 z9^>&zZ{kOFK%xyzJ}vr!uX^A1K|j&@!iu#s+UptcgBN=QG@`?&5pP^1nnZM_(b`BQ zB$~kv_2+-Xpk%ShmbNe@-A07inNh5fF$?Q%gXBh|LCV8UUB0((*QO(D8_EX}I712^ zVe>WS*{`GQr0Cu^e-4l<@*fiiUV*Ri*+9t4o`;m&mq*S_5?z6o(roGL2k|*kN8klO z|DB~63#WF(x+;WbSh?GnUE^2%lT8#qo0euVj*b@LYeHBcK;s5=gPl}PvQ8#qmo~Q0 zER(vom}%Th;dMb+AXgNZfeCyPAE>;vJXAp0J&hH^wh{HV4~N(%dZrSOali^u()~px za&oJ=Q^>p8=G{zMNo>FKS5Qv}oz+8>SwfszU#!p$(g(NZHMm?M!BW);vto*uP&hbg zMw%youKZ&1?Hchht z_pd_)jW9Y%--R&OyHE${+TSg`MdCbeNXhga8QzroMk{DFENT?3;x#b3`+-f!UuF4? zPyPA6pMFi9>2T4rJ8v0}$y$ zlRsB4MF)-zNnq~^{UWdoVA{2vZnTqr`PvXX?3SkoOE4+Kgy1T~W4Ap-|R<7|W#@?*U(8vm-j0-2+ zJ3bq5slIoh@glMF)HpCN9WrMS5pY3rqFbd@H33m6>7OINp^bw8|{I-e)dlOB)T^wwHVf7CQvEYheebeI(`RS~U^TCU1f+YtT zLO>vo_bVt2VO@ky2Kssje+?p5$B>Xi^OYr=A83l3;5+@$T7&t zboT7?7C-X~D`GqHKzxA#2qxOAqMFLv!H5(A0NGEDuedUt!0>rHm}a&Q^duKG8N@e| zI?J$7VY^WOV;=%BoyY}9Pd&yHgw!WKjNnxQ z-FI`1LRrL|?$g?p65Ov1z7#fvb55q&%p+iARy)hEK$hgo{>g%!XWWMLJH@S;{2wJL zKcT4t`DLO7QOs+>k5}zp$?%UT50v7gb@Z(rsANAX7@A4DjbS^KaYTCDJ!c_qM3c2Q znEPZvrl|o5Ey4Zx?>+NNl!)W-Kr$&&US4(znC_N8GufggNVp~c_+B0^-P&HRopu7Y z11ad*fggrn`*@+jBw1aa35mVb10d@mkk$i_G2Pn-eqm4s?&rH8p=s*ttrulN z*6=FG7Ydx*+5P!j%zAfBTfrxpluIU%_91;-TRU1df zy=H9{qwKSCBBr;ssCB>B<$IVxB@8@$#t)}`J~Dwm*6S3hg|j8ov2nkdp|`~i zlWwt`jkvdcg-Wa%TbkZJ4+DwW|B4Bh#rk=7`yRIsJK&(he9ogxbY2N2>f z@F7{|f#x@fL0B@?#z_v|pjrIdG4^@lp9``^ll6YQFh#L{5qd&jQ*%`8*e}YTXv;Vj z{_Q{_L^793=Rcr40Hx2M>BbOc7{#wRK2&vJnGLC(8k&?d*4xt*o^7fQ*LMB9Lp}k? zS;NWP@Rte&A6oU{-4_hVNjaHjL*P%=XpT>WE z^_|6B-<3}ePy2wn3UYmp>-LEZvLu87;^7tPIK)MN`*PP+aFuYVZ4{+Hak<|;JX~BH zJ$iM1z-S%5AYDYM^}k%;f89l!mqu+CV?9^o*9`s<7r2dDgq&0%ni+&v1R+yzYU=*dYt|;(nTe!P$O|`}H$4O#2trPp=UBgVTFn zb&B^M1nMjzzXP27T68CGRwu^7C?9~tAEPK^6_CMpnW7N5s8JvsO+V)b{pyE~u-je2X8-BVjtk~aAD!L9#c^aK)?tv>bB{|~1!GVeb`bwUd2Tpd6J^q;R0m=dA$M6@h;l$k3t z321MK7c`S)auOFNT=gOl1E_^AU+#w10Gv6f+}6hLxfQ$P-lR~W$DkOF&ToPi$Xxi% zI?jb1l7JYAGkbGq00W`~?p>0F9$%tM+IJ)-cg3f_sRt?X)w*LThJMao-HC`aU5tKn zM13Tm_loWFHJEVjBriqC5h_8ds%p4ai$PPMW}z78w?t8JA?}ODnO~q4>1%Fa=%KU!-GeFG0J|HS=IC7RU`v5&LqMHUKiC{`SfbOqmADOX+)7RV!b5E)W zq6H1z$2`dTfMl=ViVn9TKE)ez`-!?lb8BD_kIXO78V@J4W9Pz__%+J#ZAIo!q`+&0Jut>T^B@Jg?Ai5Hjs!P_2K z&BjL`%~V)4jc0#>5~(HePK~9!L1K=AdjthwQO-}@L|yric13NmLCSu=YYxbFqd0Fy zGN{Gu;TCWh$uwtZdwE!+xOw&&m`$`Hu86%G0ej~^4*nvA5#}IJC$S=s=)XbM3cD17b?965XS3!1FRQ*kyq5Heei~ znkJOVBTJR^W2uw_f-!fw_xG<^`^wB8k}y9Vhz}+6AhD1MLtcyhP!Olh)FP;R@u5>A zwAR_aL#wCl!yZhSoRLklNOol+3dIZ98jqn<|eP2>@=WGQa&S71JJud)AC zeUXRX1)d(M*0*oj4u0VK)e;TwUtF<_6n0Y+nqX>;o4DVt??9@)aIN=6;Yz|ZHMbfS zC?o@$NT5rI85Z`5Mg1rp9`@-#ptbbn?%@_UDJ(^kY-4uqddn1{tU3UK0R2axkP}NW zIcKosPUfmL>7|N@S;y`v5)?R$O26utK%A`i4Vn9k43!k=9ea;;h-;9fEP}6>3tB*g zwlV{WHc+$?_bAH74zF9Y0ODYDX#gTi#NaCg8m;|7)>dF_h1yQ^5W5)W{UPC)insG* z3+{A&oEso0hDF!DcEe&Yi1u}+48jN*#3s_E&t;##Z-FBu1NM6qrmbg$Ct2!)_M56N zW}DG1x&x9SJc1JtgCecL%wua!hr(q6qlU12MjZE+j0Ab4>1ls)O@n_$y8@mL52yQ) zzVc~B?kra0J{$r`Ba9|F=tMR!j%Xv|H4)aK9|$o53{N`_;tUTXJJTdXNhjfjm~RV= z_T4IxtO)`=G`jq>BAv*ANEi@of`k)}yO{R{v`7ywsH< z0YD1|Ko%F+qX%qN2LNJWu|Oe&h!zkzRRed-OvmmWMG$EfOr+PsA)Nt-JTr~tk3b^? z6~Ty(;StkTG83y7JaPsg#U1Yer5Q>2Ru&|I;&it*exn40P`G^2?Go{wx>Krsoz?!S z5}Z!8IS=Wqn!}1qCq_c=m8Kpfw}R|x|8RFt=*`gyT^L>y+zO%pBBn;;&!9&F z!}U)yHTG1~M>WYDi-nD6KnVEJAPgv9qb(%uUSJA(OM^=dl0a!-O#6&V!4c@1+X=2o zm#ca^W1Y5E<+~L%IFvOoY6!}9ROzFrg^0vIH@hvG*rTK#?HOF&M;gdz=x=mStZ>Qa z_WK`NE6gX^b22PH2xf$AaHJ;1OTQ@O?iws@Q80ki`kqAM=;#XlU9-hS?pXvt9S?W! z__ZC$3=v^0whw6342+Lw*~kfFzaoNfW!ZZqx|U|2pph?L!^q3p>Fa2OBMb3@30W`=jLN05)O7NQD2;e3c@#ZM~^9%H_#?~r&}tBO5t%$S;N-|O99hc)csBBSWSK+1 ztkL`i6`!b69^@i>LdlEYY_~z|?&*11&m{qj<@zIGLf(%Rx`cj=!05~#JOuC`Q8FBv zkJE73$mfr%wY-e$$tMCU{9dHvH~;7dI$Jj4Kd=Uc^!ok*?AK8KwSTx-Azm96=m?Hr zPP?`sV&|^$B;psK2|kfvNcHvrfi_zL`g%&48kKKLD{GbW0w={*R@ee-=EGQT{QwSM zm^0ttL1Y8R482a7492x4ZWp2X6>r=H>>IJ~O@)XEE9h15DH3q{kLbgNd*!Ct9mu5r zX1((}4@iX@5|04c+{i%yVd!|kj?jcb$;g6E77^LeveHDrS{rWJAKcvXHf$jqgeUE~s==mk2DxnHJ7vO{Bn z`_ndIPefg}eFuGDR|jl-jutyOX9M&|6@zvn&F#VAHKe*OuB3NYtF^$;~{B0mz^-SDMP2p7QecsL|U zEG}?!r_$@FC+z#OykuOng#avg96pO{PJdL6S^~!5h_)dp=Eh>L`9*}@B-b=#pg}0? zA2|!6i|1gMbeCi^g?15Un9O@@NHzDbX{G7AR<5X@Xx^ai zgJ6t}UaDYyj?Sx5(1iuLJ+f2&oGiwTUz8Zwj55-$fw3o2GiZm-Xk5=b+o4|(&S1&n zomjVT$O#OTC;v>I90-VMf53-!Xa`0jaRbQIjS4xj*-m^kJnbVE<7l^g5Bmzw3h)tf z=FGCJjftO773dAkR&?x1>W}Gyows~%NEUuTiv;@#N8gCneT{usGXYknD;kEO^C?LhI|8(t7!=R$uN0)Tl(!8D z2NK|pu>H#woiBu;tb4m)-5X&sf@%9^hmMm&2|ZCcK{_fva3#`C#UQq28orr_00Hz=v6Y_vjBZT`aTMFAu5+e!xrQi24lAA z>NXGuAQ~c6Jo~)fv*)JK9W!D9Ee#s*6M7M!AN18;1)B_Z6%N9bKtNri*`0=@*|N>f z4F--`s;+dY4ouKD+*mHQr5nPxe__|-QPz)ayvQ+NHQHJSwI334JJ@3AifApl#0MAW@)zIkzxrAF z7oJcc^$iILr_00(T#2FQ%s6D%3*6uySQWreZV2616a#$uJ7^pbhe!7F z+1L$GqVHe-A+NrNnu}qHHQUV(#fWM~D7prhxBbJe`5_--wX0B2e6!vB8}#mn{CoV; zxab>^EPds|-@w>?Xz)3NiTeD>IHQ-?~g+tkW+d~?b9Ih0ZRSYCKYDn zE&kKvw}*wjEtBy!TuD14ly32ZPDdEV3QvOUddTdqao$2L4T}_5rib1tc~1-{d4mN> zZ7&L4B|(D_FS}SeZ}c63ZCNmdyjMBD>Ss$}bMpbijC_dt>=BS;{h{_pBKe5Po$N?t zGi6r-4gB*0H&BQ5dI=fjxiY_eL$q0rFSr(FB1Z%EfD-0pg##HY<*dw z6xy2IErmNl#VFdi4l?(}O2ls%4QZ?nVo(-VipgP3%v+@ zE!%oX#CHv1a#o_=vYXQTWY$T+BSuPLuRFVhIWOV+$TD+*^e~?z2xZ)dNWIG1i8mc; z_pK0xBOuecNi7pP9~))-yi0ePK3w*u8R17bRg5i%a+OtfK~6=K?CCBRf6 z7-%5{_ma6RM#9-jv(%F zg$P?!$4Y8lD3N9)RRsrXTpC~CK+OrTghhP$5x2O!2PBI07~G-BSmtM4Z``Vp^r1W)PI9GAJ&-AA)p8-NREVKfb^q!>1n&M zoS&k4ue<#@ED4G|xkBXkuxA>u9GBfhU@^{1ruoK9^9@L`0)%1N=c*hi7DESqkJiIL zn9ZY|Fvk$%7C)$6o$tn=UqVpGghCvUIXsa~q!k4JnTW)S2}5vz1yYJBLzagUw;j?V z{Hvd_Kp}c+7bLo-b*2-QOuQJ7&|JvGZbHwuThuL-ZJ=iq*G2dMNfvaV` z1%(Yk!|hx3|Nh|r_pO3t;+>}lz)-e7FlrVK6h?I8dt`ekklRWkzE$Hfz2tEwIQ*ic zEn0)FBQA+TJ4E&j$ASo;t}voFX`APLIJm>Xct8jDXLJ?hlg5gM^)4LJi{y}ghhB|H zQ}dhL8Q2|nji@{Dw0Mra2SH|#zq}=?MRh%V_&B&DIk?1)Q4~YmA_kPR6b1O@d}ytY zbm+>Tp+}@AsIP~Be^G9>THmaYWe5=3z;zaEdhYrYsao+|$H6&=31>UH7&(6*Ia}CY z%-gQg(}!w6;m&}wjLskR`O3orzmuwk1A6|ATovvNEKB+;G38|I%Z2PRBai6Vkc=$x{p;Jni{z z`w^6)B)MmoRqh3Pl1l^Jfjc_Dj$FLM`DN#7diV#w*vR_z2j&Q3Y^LTj$TiSjL5c2M7!B-Dcix~a=Ba2G007y%m{wr0M0~n*)US8 zOMT1XYozz;+1P?c6|oHNxQX^>4su980tllALGgz4Z4Un6=i$@k{Z_(#Qp5?P5?Te7 z5V+rxxdAdK?_2yO%DKdieeq?jyq{mnw??oa)SIVi=cy#%y%loE3@o|-kf;EcjNB*- z*oMG9x75I*g_^zFYjhv7X`(Q2PQbIj#i94s0B`U#3D3)7$cWDtFXg}%lFh4UT(8F2 zHF7d&VWCEPWNQ#@9K=*KqE#3WnL`y-#P^BP*!Z@Af5Xdb!P%GSpVsI%LcgUb^+mkT$Y%eTM9&7Z(gDvDd zgd`}n!l;Dytj=Xb9LgnxHSde{0R&{=y?%upZb`-l5Q;I%$i-Qxz%NvyQ@wZW@I}g_ zwVl6^rAW=dKCTbsjS!Lqtcjs<969g6eh}zm)*mA_!Tby?BnaW)J@>!c-tX2dk+H^2 z!5|UII5g%}1B)~?6M`Ew)fRMo-9^_-&k7sZG|D^dW?HF=0IH+N)GBjaKY<+maWCefRq)-7mz zx*L+d@r%%1Un!BZd)wtLAerG%AjpIdsN^qr05$i zxN55wHB#|nZYl(Xj6L4lz!c)}$;vVf#7d4WBnI_>;`V_&FhSae8x2&fne#k&5iBs! zMvh&&Q3TzyVYWdDwn36o9K4Vsaw5&Lv9nNqHph7cs}Tu3!=CO3_D0be^d0G=t_8a# zO5FE2t&O@Kywb=TpZzB$B{DhiomdnJ^k@_O}x<{iXTw8!E3Lp9xq^u9dfqQ)`_H==5zjy%t z)WViKks`K+!ZGE*`55(4LFzUPY^T!mYtSbSqPPWXj%u}IjU*=#iNfmkRxxuCA#vG) z?D^>z3eOVIk0!=B*xjJ@IU!Lv9!AkEQ8ZPq?@VS%PW+=8I&PcqDain)p@Su#GsfNE zRwO6@2AXNyTZ*G5#1^<7qwY=pyg?Z-sLCjU$D$L)bmp9J>Ukcg=@fanWHV$sq@lbo z*2r+X(D_qIIGDPO4I8nCpdE1TX6R!hetM(|2J8EMbD&6%EKeLG*V7Ixz_iQkHUtml zsK6Tx9mV;qz)S4F&;>WS=RHA=#6enzIU0T=rBv>X5ABO1384c|8L7fsiO z_VN?uZ|GA)BiNB2mL46zW`;Ddkb9Q2H2@2<48V1BS@ilvHR8UY-SLJLVxERk0Vy38 zgxNW~iJ_uH;{`aYLF2B~croyNOlbV7H-c9PmqXSXhj zfP^++FAtU-#@K)1JBNbGyO)LU^#8{%7IgfD#NEeU7@s}#$=S+yLoCc?i1%{`^pI$E zNW4s195*b6r5w*2$(jf9cpHZtfePQl&Dn%F%=o2XOU{W`IAY9Fj{q`U>@EQ@SYRu$ z)es$GL>dmU#>H^J!B=pf`kk@StU(-YQPMIIMr#GoahJsa$O{{_bCd|Z%8SyK7o*sm z!&4!W-{TJ}@E9eP8^)_at=lbX-EMIX*bRiDAtyeh;&Fl+*f5#E8gTZ-b}AYRd)K%` zA?}Kes?M%u%Y+Tq=%L5MCs48haX8WV(nNAT5?27o8yMv4tr!K1K>_AJK=Cps1dy;h zOB1~}1fU6d=lmp4%DfVTdRW(Qg3809j5yVSq5TT7{$r@w70AeOmPn#CFw7kp?FHUH z$xGJ+dv8s<0ZphTpg*-0eoj(`%F?x4xLqL_N$OmYK zv4tojmc|xZyhM+EVEfk+V~Cs@)WOE-?6CU;TGfX9(up9$hR$6`;=PrGY-zi`?t4c4 ziVrual?A_SOy~807I=vw9tJfWC`%f|6Ea_gvE@;(gp~E4~<|!Kp`*JZ2OJ^OEI9wI*Oq=dx`GpsCgorOYG~-?e+naCI@#5 zx;eafcf><*#5V7cVF`?KHsAL&nw?q=eQXCxTkk~<=7Q=6vov70TDl*QVpH0*wKyrH zd1XvIIo+*mF`3}FEPKJS5_yHGAOu~%!I>2r2>0=q-ZC*#Jcfy}HI+}Vh*0TR6A*9% z0YZAzXi;!#j5yraulC#x;CY;-2uL7LK{s{CKfK7U^F=`LLmW9F_GCL5Q97qW&4;KA z^c$8~ZJnyMI0hr0()n3n$5!mmWcrX_XTA>DOl?yh*K$-w5MzfHBE5%w^ zgaEYZC3N%%t!2IwY=;P0TYxtnYC*sgr(UwD5u8n*_m9WR^?>C9CF8?3)5GJDnz~^ z;A<1ZSjgDyA^INeBM_q8&p37mkdK^-?Jcu@w(=;+-tRb3z^WP?ALK*Ss?TyfCjr5 z+{Dp!gCj?aFL?+veT4|g&rgW5it&&DeYq(9EC=uRoYTWiuxB+XaREuN`B|6AWpi0U7x$wx4z6{ErW}C%u^4*%e+Bs=ya)^tDbWi~HL$V<MK=g3;uqX5Ek0A^^_a3$r`h!y7Wz6R$KFE-msEH>ja>K(W*m{}iXr~_$ zhvf@1qaoI;vSu)+WVHk=g5Ad;bVnnRu$w72xj97z+eD&+Uqx1D`z=I(9wbi+n*h@X zs_V+wMFIzeB)YLjUl0Yg>|h0M=xxRfD3ZW{&X=4@?jrFhLfD3UNytZKtzLt9Ea&<8 zqoiZr95|~Aa9z!32`s%~d)PE154`gR8IGQK{cM(7QoHjfA$`1Z{^;fk9asBMGXE8N zR*gc{SsVrFq+|U;;=7%3Z?_7dSk(&g_##o~C!;+i`T<*AsJawUKR+CdpOx^79K1Z2 z`jWh2X%&#p#bv-iX!tD?p`I-KFc*5zitmX(_akvjo@zRTfj5pbosNVfUFj52WIg7z zFF4NRlEO!#0OY}wY9hqmVOlB72l>fDtKh|Mm^Xx{k$j;sFmF7v;=_@3Z0QqgwSRaA zmJ+Fxr34dSJ1a`+V>EG0?xTWQRR}IaTu!6SYjz~ybQ08Um{lbuyA|?0=gJnJ^F3x2 zfmsScF37cLgFy;$^LIx*y&|Rpwoe+Wu|pm?8ecGt*>{%VEQWrMw)zGl`>Z72K0I*) z7foes(zHh#Y7V^+1euB}tdP}33G*Tss)UqWQMS{=N`+(5%RAf~uIjL|M;(nW>m13Mw*8v$#Jf3C#({00OWi)V&41t-+Y*p|dPd zalWY;v6b8=rHnfb#z^byd2y;}^xj%o(ZI=r&k{BGBCVFNgH#yauFv66wCCIHVk zK5Ut&AF7MWRTpI%A2!)9;y%hMXFR;gxi zk9&$#codT2VYk6)zhBMBclT}Tm55LGOI2)uS9g+(OtMvz%qkd~K`Wc>fqnB2%8j(r zlwp#p`tSeYmw$NmcPRB453zL-ae)Ms4l`$fGKav1b#$UN!~Cgi1Wl>|jD3K_St`Zh z=%!uXTN0xrHt{P5YT}1F93kXkWyl&8?I!jPlM&4uW=O^CGq&D0QNTG74evaKk?)-# z4PayZg_sUF6{avr0>-8Fg07A1#cB02*5wd#kxb zAC~YwMCbi>@_XF(v*pGgcOau3nU;IN1%@Zqk!nq`{`0)(vrSdO+RR@d0|y|Sd$6It z-7de7j+u0rNM3%XJYN{R)mn$r8F{slQ%V+XboSrA#fXiRgWy1!lA!SyqAQ?42|>BL zMDQFC&u!!35dZ~%U^l@uqM_8h6!npVwC-2y1$xa=_(3!&A~(gzr+2|0#W)}^<46@O zT;w5mFQDM{!!KrCXh!sK!bX=F~4M?OhQ5QvP?F#T~4+)D4U@gH8u zM>R2G$?CjU#2Ew>BYVnm09FwZ_!bym-!}W;|G0)Y^ytb6xoVUL=42sjig(5*iUQLr zQOrp!ND}ZsQGAgeB$FGY>`2{8Ti~4;F#D{6XnR<=C_)fLot3O_lKUz-0O|^G;5^T` z|Jj6Dt#0l5%*COCT1L`@!iw<12>9gStGHGAZ7+s2obeSU%A*illD{z%I_4IQ#%PMwR}`H*#MRTxU>|Aq$>Kw-9F*z=n3(Cjfl|axnDF=9+>xZsziBB{FED;fMnG#hr!h1b`dVoF*W zAiOC?2iv<4)L89Sn_JDNz-hT;HApT^na#sI&CKI8E1$Nz+c~(rprq)J6m^!vqu$Lv zAT`iuAxnsy4B_k80rS4M2**jBxEG8v|kR_yah4J_g*ahDM{38A!Dh`vR^ zLOK*pIfn7j(T<&0qM`0CjvNU}JvtYe+9&(2iX#uk8k0aYE^r`_Ie!a=S3&8s|I{86 zYpHF$##V)AV%0 z1P)M&NpEu`ZA>>J2NZxhyKLb1kz4l%jKZvX{QSYj5Qt)nBKkn3cp@SIQ96kT)+TX9 zCg8&U!GzO;Dw>Cr)toxd=hQHrPB+O`p)-hZ!tRFTxM|9)E;nhX;<+iYCi* z%}23{(v{+5R2N2Nxj5zzBhhDjsk+i}%txGS0m;%t>=AGR}5Qq{Q*4Z;<}>!uJ0WtlPi?L9B_)q z-{~npjfV})LJ87ZemuliP9k!Muys$JJ1Xs^^{_3JKVOY7>ROF z4#s&Qw*DbKl0lbv=8y=TH_KhqfDNDD&-^AQU2lN9=J>UEO{UEtOhYdZ&pGYaY}qB5 z6fR$ZV+D~3!C{F5*M{#wzWHH3Y=4er;h?A}3R-7*4y*ZvsXRRpY?h8F8|2uK zi{Ai0++SAeN41FG^3yQXw?G#GcvR*Ad7t)FL^Hw0%wlMdOY;&vFwn<~SBaeu)bVD; zo*Z`^7z)kjHqA0b+>cUjky*+zc$biEO*9l&fL}X)wl&9ONn`l7v3TD@)KTPO!+z;k zEmpQA_&|khFx`sdkRk0Hr-Bxx)S#^BgGTf7!wZwnfV(RZ(hNibX9%)1xx4(C9a%V4 zOe-E-zkeZN5hN)dU;vRMa^z*UOxND!Ah^Siv@O~w6%TqarFPpv`p<#FtqX4&Xd8wx z%t;3f89mlba2VDsqU7&43#kWKGxur6YJ{Z=LB`wE@kbTd4%DEL{DBj@6R3OGHzIw2 zfG5R}WU8t~98e{g4m)c)%+}fzVqwumhU`P?NaS5<(lFdihbH8x1u!~r?J+93YLw9* zk1QyKy@_L_BXP^N6S^GNX$a9OQTmK8#%_F3QJ@vS1zGKrCKC?0bfAh6^~G!OFt{kU zNRu81;@5xh=wG24@z@#^nnUh`eu9ijH~}C__us4bEXf8fT4ErU<_-u9&!>)MgF@hR zl+!2j@Ac(NPdZSlL)8b(FNcGS@>wY9&Bi<|zaJ5qw4RgYsWEK0rybz8sJKLC&hQn> z%bzyvH^erUZ$QMn-UJ`oOc-z&g9zrGZWjrYb|Aq&cVlt90~k4t+6TFJmc11*aJh=iG~D>vo?8AR5GBMO^p_yFp1&Fx(I7<$7UB4P6_Gv zBuc?Pq%Xw7%qK+D2NC`s6(+|uVl!|w-e7h<)Dw_;2Ta+}0MhdPQ509sqPX3mYSg45 z;xJE~ObxLkJN7R_3di6}8IpOn?QY0hoJ(gkGm|sg?Z@aTvz( z%=1xhNiF*@^ z<1mUti4yox)e1|oX!Ejo%-VTzT?a3#4i^|>y%&;s9_3_R%F15!M^$B6*&=JPvJw}6 z8TeSIjBFI2mxxYvfPajSojbLF$M7cy4hn}=O> z@UTPGxTQR-ux1>&cHEB=9wCZYGpQapw+i+TJQ}$zaf~Y>%t51zEGn8nwOG_e#Blb( z$O{@2{?W8`MP41@*>P3S$NAa6BM&9D>x#Qtj7H=1amGuYYa2vivPXRljE31;WfoF3 zSu3p_3RjP_WR?b9T8F6duz!J_h*>htM5h?E$VcPW!CG2&Ocg<0+#)FgTkP0E!ak?m zSie6UJNV`)Zftqc;>HHSMg{v5XgLu1kFyXwH!x*r>!Fc_)|zKJ2oTxkdM3evR+c9t zgJEYAKOwG_wOd)1tgkFt)@I3k7YBJpp``Xf3c55b*>F6fhmNsi?~x_zB=Jm>tH6@2 zm?dix5${o!ta2>rGAx;J6|18x*;rUI&=6pfWW$+(Yekkvk$yG4FDP{a8Y{+G2O2Yh z{xQa6)*m{AArTsrCjWFPu_*JwkYG)UfZ+6_cL`#k!Jw=d1C1C1{S%n%T)b1~JM0IC zsf0KLY|-c_=_4xCBdIF~P*g*nLRIOIy&Em1*Gq75i6b!Y$Xe|I z-P)EeruLATvaW-lFfAllMdy^xaHpA8)<*S@fxIP9z!e7%2`9W^^Gj76#a>AvO>z>2 zLgKEd`<2dQ(C@9;6>W#Y;4!jEVo$%_a3-p*I1cQgPE5cdZc@1hJhU=7d{3&I&;sCyDOg4MWvStA!T zmvL?+q)cq|m?Voqf5;TWJwa~}C?dXskhDQ0o0ps*5qF|k!zj)t!gehtIt?bik`2U3 zwqjW+Flp<^Ga+cr;*j)iiQ0384fR?~#_K5P2L&Jj4jkE(9JW@}zMhPs)L(aQ@$MJ9cjpy_=*vQFa!H zL`zbYB+f5ep$6!8BD{5BgMFV9-Fk?@9`y~`d>$~u_E5k{@psc0*ei*F%{GuBgf8)b zI=*U{0VFB23`nLS{6I=)6hZYf6hY=nILoF%ikP%Stsp1Ct2}($>NH?zp+OSsSkc$49adP?ZSz9nyh~+HEvgTt0l`?j% z*g)mDX$YXXp2Zkgt&tOo{b32q_A?{{;$|q=XyV@sNHTI=Mafz+M%a5W8nNBRV!P%gXP6 ziR9PY&Byh|C{;2!Omf1zu#(Eg)lHVM5{P3p;-%FENlI=GW*Ti&142Sx80pRh^4=&= zM3O2<=m)H0i?a7()VGHlid_nyxTZJ>KOHpp5xHKYg*(OhS$o`!!7M2`98u6An)n)9 z7BwM zras0kWgEiYFh+6KRF9F?Nk-9Qj^T5F{Tz@>@fWf1Kmia%UcrE(%Tw#5AXGfKmPjM~ zzr7NcJJ?n;_)YT1t26UF2$9$+nKFl<+z``gZ9-s)rAP-3fr&NX2olU+32(2#^tT~C z4d$R;TKKU8tor{FlM&J&t%DGZg0{HJ51v0H5n-QntOIg;qKFJZCPivR534?)Xy9Ql zTsXjLjl_lY8d7VxXGBtfu`9%Y72@dc3DvbB4pH)3Yjg}|oI8{=^()Bp^>?0p@M8dc zcBM*I3A=UNs#0?^U0C8VHoWJ^x5teW1|kVtMaBsO5bR-OB$Egpc~a5SbEXgqHDFYS zSckMc&J)ZH5n`;%T7?MJ&Bj>fgGQqt;*y}Gs<*USzBn5R(t;=h&DeSrWO@BU4C@gU z+D0&>D_H-nXqmB0A%B8wm8r!fyH;>xV6iKvdUurS{X33Qr~%OKUWoV;?05%7YQ^HP z@}C0f0Fu6Yq}$##5aW3s4Y<+f0D{}KCPr98t)XGwPm3!iw?ljb;Q_#q_)4N2VE`G- z@r>Vo-`%l@IK554Uo1VG6BtqU;f0%=MzU;!;NjJDi5Sb)SYNnIC0Go z4;JRn!H_d17Gv)m~HVHNNeJP_pP5eI2*^P|hka7|cZ`L_(6jl}^(Lz+3 zbmzr@{LhPUG_8y61Fl`-x-fp4w9qdVQI`4vW4+kpcz2@Fyd$!0wo$NaR^6 z0nYt@4*t#3P!elq0>dWiNFSUz0~*NmlMrcZWUV#Yqm!%nGy~Ns2`|hMOB-P;!}Q3J zo#ReoG6(eu1sZ&JJKKDCxo;o@-~l@tf--{rAgO8LTbSw;!cW|Z6~avvRa7kmGZua0 zSt#r}sf0!?Afjl)nn-#E8mGr5i4!>_F({3m8g~EO1rd*iNR(g@BqwI)(N-TzQbzy% zZ{Pevv~zH^g3v|GD`bZIG#O#p2-K3Xr-W8}%%byfP~u#jKaobfKd7Z47?-hITkOGN zk~}u-5{JR!$^g+7q&%{uQSTK^o+$r*Wh2D5oVUV&ci^3oT{LN%~y} z!?;odjrs23&<#l(jnCv41LC8nhTn++9on;HW6bfey!_iQ?;ze28g>bX&em;s=DC=h zh3_ye1qV79npZ_#Q(PSN_o^_^%#Ej_`#Jj<&8aYl>iZw=J;L*VF$WFa7sx)5d3xXP zap*Lu{{A0Cq|``J^u~RS%ZQYf$;gZp8KvXcatLm;_=rMELRb;Mpd|&_-dvh{c(kkw zZ^&63CvV02qbV-97i+K`zhj^|W~$)rB@p+1;NSOh=m8GtfOmSmaZb}Lf~PgOw~9UD zWI2b_}d_dx`;Kt4ybyG5kui91`vovfiHziW{gyo7hykr08WT7l_8$N3P0w9KvGop6mP z|KxR;3wU@#{ja8U=YP|iy^uc zs~zX$nA;zWAm4QqAp(P>qPiU){VbB#8TIMIX8Va`H3IgQ%~8cW!6TK}$q>=qT}F7P z@}VF_N5^b@4+lry-f{9giXMuAqqATuB1r*A>c_eNMDN}N8E!#2L9Z!#grvb6ixdMg z{T(RN-<+ZcwImsyMlAX^3O^fpL9`|ei=ZOfCApY(hF!eU0Bg(07Hms#!FD7=Q=OYj zh>@!h(=0U$rQ|&rRk)AQb;MMMLr1bFw53z*Vhq%gjVXq>v1v7~5m)SihVY z{+J}qP z$N`S9h6s@SR>RXhM&TgrI@t9sFCpki1yO1kUvv&bf}4nho}dEJp7tz?1{oxQdI3ko*ULQ}SLSZ|AI+LhYnZu;3mJ`E)ANI}d zqDMIsNc@;UgNxWWv?Vj9xDX?TQFC#lqzS;zMlJ#@?m=no!dT6&Ng0DPMk?VH%9+Es z?2dxTR`9RT4tju+v$zoy>7)$%D3Yi~h%A!+7*3r~4hmo0Wf}u`B|+0xN8mgN{L9jD zUXZ?I7wl=?2gxvQ`~JOd(0E8K?8YJEW*fs@$7x_malUxih;=Zfu%a_a@wmu}{E~80 zyysYoA;~pR1?*EQzCww-b1aO|**kH*Wi@zY6|D{gbdW4-@7TWlRWGC-SS7h85JFI+ zEZxztrJ9A*h4rEk3KS0UASj<#3PQ`rUElD84^g=$N-B*2s)fJUU&Fg z8D7m$DP|4}$`vmr-|sIu(eg0Qs$+)8ln)FojwRT@@6p~S6F`?h;C>_*>~S=9pvUQ! zi4lglr+KM|17t>XFqRJWmx1egB_>e{)nUKoe?Pjtu7y`4~;HS)w1Akh5^zn2>+|mx52o4P(bQIe#0;C z4=?`pUzoKf*KuK)6Mr5@!rhbZJZMBe9=G{!`T-AU2Lqb(^83|WU$VA)VFd)xQIeA4 zoA!$&3+Oo+4|@#LbBToj^@v?}A@{n+05mnYJxsrlH1$2Btg;xgcopp-G9Q9&HemJZ zR?=_mL%+(=*ioCdJ@#|s@8B}@hGQpt6F)^wi$CumOPetoec!-6I6}F6M6!$J<0P}SN($1_|!yLT~RWxAW_@%ZO&T0e|RT32f^w5KsP(2HDq)XI0Vlj4n~e+ zmKndp9XCOmP))QZeRgP{Ly=7!6FIHu%5NVg>Qstl}4GW?v&Z;ENaU2#_oj1(c z{WI+b4)l74;|LiO!Y9RFZK)O_{51;Cn!1Yofn{orBc@WvSC@Zocb6<1{%kwJ%Zz+> z`^!LHpy>)sEznRntHwkgw>giDXOIf%6xb90hE!{l^u!PZeM zOKl(noMsayzDQz85P~fRqpBOh7Q+fH*=JvV0M8*ba6%Dr4#rn|B1zs;G;oAAuz0{@Cne&nWnl=f7rz(S9?&T;DVlL3(}jVMC4&cg+k=24_JQ z?X7PJt&Ml<1*n1n-;RfXZJ=d)W+7i1q7XlQ;?|m#;Yq}^zwl`rku>c{L6%!&GzMB}}Na1pBZO>M#Tkd)vXvqE#Z0Y!cME zVLL@i-4JWJB0B{Z|3`6#RJ}RVuVcfw{PLf;N?raDV%yc@CCQ1@N23gc_Kv~BK7!4%x z<7O`ogfsR!ARDS}K&!~RHO9FO4dq3!?03ds{=;2ZpPys3;^x+7{(H>hndiC^n9arX&ietW#2m zhw2j3f_6EhA}<=8QAyg}TN+2Ry~JyiIMpsBB0OklMBHu^Yk6f73X2865etKUjPQk? zmc^r$mly~={7;nSo^O`M8w!IJ`-f#B-g%?ZvHlSe7 z$r<01u$dgP?RhP_&MT&mIOYf|$+_BaghcisT5QrTSo}#w0`gPH!fKxMfR0jgVDy!J zP&aV2qa+I`9zK%I!JtvNdiQX9SjX({QI6D__w27w&T_LN6%)p(MCZeshdomjnh0dm ze1;vYGLY<|8dPoeopV*ukz)R|!*p;jp6lL<-@kZ@;11Gnf;mL+B*U8~<5_ASDZx$<=8(G> z$4VqHb-3HUL_5t2mRqb5KU}UJ?jQ#vYD*1dalw=Cqa++((7v7ZB3IZYrGO|eI#0V> zTtfQK;6jM8Q13*_*GEn6!7EnDpZ2(+&A(fpN_WFB;6`zmv zfxgid#}o}6ZOpVCi6rkCUQrOU4x_8c#aR>08%6^H90pI*FE4u%MNLMc;j1O*We9(9 zzgy2F16bQdj!m$^YEv~qw#^d3;aT!8{*3lVoz4m!Ye1M;{P0e$Yd50_uWNO6RT z}BxOsThs25?anp5Od(`PqKJ)fgqUO52ulOLk^;FML2eFJ&dv%;PC@Wi)6$xZ>=zWl z*?0T@$I0X|uv21c1`XyJ_`O1#vVFl02$RV}4wJs<2ub9WTFqrd&je+_{)@!K2KiI2C;z5=#h4`_yG6PWc@J znKEvZCDJzlzH~wj(IiQkR9FjRDm0r*l<3P5g*+%LkkpZ`m>m*Y!WCnFX-!_xz|e^y zm5@sx3F)&?x6=>haw#&DdU#>nIjEZMh zktF@ojr1~IDhRSd?WxIFtb%Z7s30gOh!_H^?jdud zL7(~`$QU@*ExTT*L>&YMS|EXU!J*_*fjY{#*~3Zckk2{#eqt*%3_W8~+QBbo`NJv_qIhTaIH{3%TLJ zO+n;F@rKoS2>V8{gn}K2M;%~0GF%WdZksTXR4oy1GTw?bbqh9rX0toJIglNZXap(H z1IO-aaK@NA%<~enI2&=s2N>J&xO`=hX<~5}VKSNR z{QRs@Y(pZ3l?1dx?y-*!9XDTY{o&pjY8j<@*hLY7L^6~k`~Mo#{}I4Z>yaqL=qj3( zx&BLjg$%yU=(JfFxekMgn=ncJ`1 zeY2Y#kOTem6@%apWbfu*Esi0?(Bt<39s1n^7KElzixt5lX$B1E`e9Fr0S_~dkh@w7 z!l{N9FZk#E!=`6Aka20B!FJ+Dod4gN8v6Fl0~ms?$gJnrdyeT3d*FTs+!t;{kG&c+ zDBqw#Vr&-OaB6wdzDLyTjY|xlKO|0C;I@ZxnZ$90k10BmaO^2WoScGr78IRe){EwW z6g!Heb-+1NtKNqVWeT20Av%q31LbF-pxQVtcADrzgzS#^5)=l&Sd)h@+z@1Cmnsp( zOFReXLHA~PSncJMK->}uTe1Kf%MA&n{9g4N_yZ}o8??J8&w$MJMt!V%IK$@G>XVwkYuk-gZSVvDqtzY6FJ zG)Xj{CX_*j8J`=ndlJy{fKF_w+;42Vuhf*3!mcz-5qZbAHaq%e2|t zfSzjP#C&bl1jI4mfG=QJhjNP!J=6frW^tEYlaU#wVD2_DBTbpF{%YhB!5<+@(#t*z{04(6NHiBHA@(0` zPYtprVn{9Xv*P*jBoszjTW>JZ6pep`;r6J#);xJcQe=04PBv=7?gnMwz7kJcH7LTp zScchvDEENp&2K=I%;H=u(y17aOdC2KAHsCFVO9k(>im+|G@T9;(!Gz8Ab!sF3l19? z^MV8Om`|E;lttgZBmv&VT7NJY##@yh%>4959f3{16Jh9Zh?0o#Xw(-Fc$@CHxI3dN zc~uz6^~N%~lwlzsq>S#+|L5S}L2UDId)d#}q5Jj2W={4iIRygO$TpmL%p*3^um!}n z7UP=PZl{T$IGPq))p9U%M-1vBLQqSSEQ4K}BEL4Gxr3j#$5m*&9EBΜHIUnnF6V zhJE+uYm`+#0Le~S{Llwnl&D7>!6a}YA^QW;A&*3{yF$Te$>KW_ibkH5OPkDgvu5N? zK4Xq@u@RkxiGV_KWSp`_u}p+OuSungqAX;Gws&_#JOXxEhbUeUH?2gw3-YW6{Hd$8 zq--9Lj7pifrA?)wzVsd5Wsoh>v18MSn1bMtZ-WnFv&f^s%48R#AsZO^Y1mP=xEqO& z1F^l{;Z_99eg8tmzFqyl?m0C6lnhK93WXi7dYGGVf7yt49r(HIcFmq5^#t#DV9t@` zZ3TiuN-o~A(Pm)(I>`<&yTQOedq`~jh5YW<)E|^f_M$8$G|Kr)NJQf7n1Ft+2*ql^ zx{=B40!v`GXO!L~xLflcl4)B+hqN;qK}uDd_>H9d#vt8z2-09*?9Cq{&GYz@Soa}w z+ZU#iS1cYTVK?nYIt6iFR}>|+)(1Fhq$Q$N0pif*MaHR3egT_y&)GG`fe>RHz>fUe znTSBXPy~fZvk8He@sZz7ejmwBu+oZcz84H zAV~YTo^OJql$lxjy`i}XagHKWWX)eDlL#ly1#1KQM6{+@&u$*XT`=s@Xde=& zI~ga~?V&FNoltSfp}F_eBpgH1pvKCSzBP!9g*e}n`WDR`DXKRC%wUuU6M+pFSsbZ= z1BQ?YJM^5#6e8>i$KVQ6IG`6r@O}-7TZWeQqoifor6MyF$$~adg~o@*h=%Ft_Q+#h z)_cZzbi)Dpr1M~NGYstB3P~6s)D!4p`u*aFs>a_y*D@Rm+FhWW8`&dvs=*-eAHWz! z0>k2qnCgY*3GFx*?4ToiaLs-|`2FFyL*3IaO(-Zo=FRpKC3@rCOPvbl8?Ijw6(SGD z>Ky;z#-GYtWDGnz&;U^^&UoWt@+NT|FU1}rL-k6eM4``9-e0hXh`nu)VS}&iSyf`R zFzjd3(LnGTNZ%)2%MBQ2ueA5l#?{XlqvrR(3d}zL4(vhv40~p1)#lkKnhg;+UxW~o z2KQ`of%tB|eIScnoNk_V^;*V2fwY873EKUTK{RmIp!}@Pns|(21qxR62fZ|w<_;=TG=Yy@Qd(`?9AA1>(W8 zSP(@D><%LMbj1gd017}*f&xw*@Wls;ceJ=cy$8rO7s$0TR34@^9T&yXylMkRr${ri zso{4hp<&YqGJ+sCq^3(*rPz_v+#p?%CT1C3w5(bLwPPBB=Aa~rg^AzOo0R`%fC`vk zH%Sa&ClDk%(Zs3}4F;M!&36GH@*rUayG<_1#z6WF&OQUl`V!ftlltw7V+8I3pJa24 ziE#tbik?Y>(Fi#)UvKxG4QN8{F4`HWATP?Y8|n@Q7;}UlCuhnLIaAJ~W42D9?qHPL zY;-^5Ics_6TGw-H31x|$D2tB#cLOWJ_oP#Y0UdM>1UHcZo$t4Y9In%X4xK1rpBkMy ziTK0Q6tLi>NMZVvv1fA-*o8xVs~4$fKi30Ehc1=`&W0EaFb3YH1p>Yu!21v;)(r=v zEQQ}4ZKUu!YGPPfSIaJ=!Im|_uy^kwAAM$CZv$ifgYb^T}3Dflyy#un-{%puQE4^Cfd(Cj-=aEWgo z8o3#>0yM!2n?AUP9mo(a#*lXHHT(@(t*MmYxZ6V+TUkkv?BL~gbY$@DC{!D>B!=5Z zm;#0uLA85}UNdAovqwRKsDh)75G4pJ#uum4LgBLQ)ooqPu0lEx*XlVkU6b!I!Uj1w@l;p~1#xDM zZNa~%PDLi8cUzq?oXwb%wl@9EZZ*5#YyKJ$4b0i-4UMe;_J}|+WH)6KzZrR5Nqp9s z`mw^@lmv_2K9z?={)+4kQL59YfTR6fXs5Nh*WEd z^MLT#!H(LzjPLke)!g1jQ96eg6q1tWW*w&xfdl~XLT?drut|h6&1l`@^g@#D+E3&V z4ee)<$+Q_bicz}-;v5lGQxjx5Iud>3V8pJs5qBSsVS_X@$RLC0*d+gs^#7QN6FT|! zZ5aI*GM$a(k}r&L_WMf;hO#)kT#+g?h&s?u?`<#>UK1Zc?r%mO0!KleiViyx?^3dI z2qf;8xxu8;3(C!jigmykBp8yxTw~@UvWCfA5|mDJ$wyK58H?KVfRs3ZB+DLuWC+Mm zG#avvh(>q|A%r1o)%x0FaY{{bIl(*=Na1g`LNGDxqbx#ek>w_PEXJYxyV>n+vukNe zR?)z&_z;tc6>flQ#}V>VLQcn;(?7f=<^Da1XE{lQ2Ob^_mA8kVqHcLeB4Na;N?orwz9b zbk&zA=gUeh1tl-}PNdyoT30Hpr&zR{#axWCF-xW-qT;+T;)n=2X1<`VjWZ%6K;px% z&}=acKHW;w>Y+u1#fseW2;QZ%Gj!q_s2dt{F%absZ- zX$i?=_fHhfRO14yA;zGm4AyhB9`wvQHb{x(9VV8)0}2ohTa?~5+JLQ?mT1Ho!Vj)5 ziyUn6*J5`*0vm~t70gOjlYs;S;oI z6dxJ_ZdD@^5WNln01t>Hdu#{JiXxYE+q;T!CCv}p)$}P#M-&|`aXz~!2IAQF9rgW( z7xepYAQ3=Y3Y0z$O-Q2UWq9Q6WnbbZf-J6J->siEeo7|dV!9oQYP^7w#~X)5_){9++O_(41thq@dSPM?Aw!v9N7F@+7bY({u&SDM+kO08j_lM zqQP2#C)$BU%8i7u2nNs?vH`Dw~;38_9 zf=VI*95gH}0v%RI9D@fX1B<(N^Vtsr8%-__pBBvE)}97!$QTje2K0*pvQ_9kte-}C zDoaI^w&W22ejXkyw+J&qMKz$vaH&U9W?YSdoO;*Hw{N-ESJIpZZLElKK~{+fpL`KC ztS`P!lF@ve(_;A=>u&_wsv!k_Yz_@O}=byK<wV)o1W(Dl&{tfhya7g|1z$w77YDC`X zT{Am(GmP^|8{^*ZL);lKvK~fU$$IXXc|1TiHMT34O%je9$by!Js(4AgE!FehS#j2e&`WfO7Q63u0am zOH7XrP}*a}+Zl$gp{QRQ zjBa5GA$-9QLf=bL2uw+fywQqWR|&7u3qweueYu#0Pypnl%}f)rllW!mG^L^fO#cT_ zxW0qqD~VV7I1n^Iw#c6KEvTf`z#jMFCPLAV?!ds8Y&0$;rBw@6x<2Wq?(a!ch!0aX zrj|iPmITJQpD*6Ri&tqt;0oL~;kMwOfB9!Ml^DpaC}i>W9gG+>f_|KM1Z2=dwoX~F z0eg|58eOx4W47o|9^LL8=M=t*cZ*wV;TV7w;}jQZmCGLbwc~g|-_UC=>9YSRD*Tar)JajMP<2K#Xu| zqe?o129q|km;kLy*J_sQG%CRnNfM<@mrNNXQu72n63Zej(^!r7vzulP7<0=Zp2CMB zxiY709COl4atNgfUqfmkB@PrvhWn+WhUwgudr>#Pqp1Obj&wOsyjkYitKE&Sfl>Up*#> zyT&9agFF&oeIa37L}`$@=!!Dh1`DMTSty-OXIlhAZA#E+lE5}^S_P!T6f9RA4N|VAag|PKYS=~h|iOu-pnEfwoA96iggRT9tZB)(%wF}Jq zrx|k^(!a%edBiDsb=|VhDO1MPQ|*<5e@n8@@n~Ba*HpxT9BXr!{ZJ4mMf_;$u^3`( zXns@WDjg*@mU1~yBO;oPM1_l!*OCuDxl12Ojw6neIfu-^$CAa- z)z_GA{VXf&#Oj0ibzK8wA94ulZRps6jku)} zV+|#X6mjk%c`ey=VHm(WgKmQyQrOrI3P_KG0mS zxWGSMRiIPJ$_t*Q6@3|JJEQ0ao*m3g^n(#Y_&}>zO{mf;S)f&dj&`|b1?*A;k%^1e z9cr7hCFz!#AtQkZ)(dA`MF>`?0nX+?#5k!N8`hJDT|(CoCF}-+5o}E~ zT`&j%T_o6l>^QJ$QF7iYf4^<+F)-#H3sziODk&Eb2fMQO&(I@AF72rvv45qJ zss^KyrDJ~67aXaIo+j|MK?4MjhHd{@ zH9cSadasrA+-=a%VL%C$Ak!9{7>Ly(GOf=@1VNX%o1!?2N0G}YFxb{%#rAsCwFU3V zRrFm7JZ_rF&13*Slk>EZyv>o~k%ol>)jS~KeeK4Oqd~Xc#h!5Xa|yYgo{p=)oHiO4 z%wberlb(T(HBS>|O~}weh~s04j?`Jb<=e~A)$l6P{=EOIh((qv1EahaAB0jVAXj|G6QwAmh2&k9TBpw z6r4#Dy4HlFDu<)_g+&Z{FHHMOo23Yw-d?4hGTJ6(@;B zfgdqN7+G$;SGm!qxxw5_U%*ab3wuCKJl7%U>` zAPkupTB3E|onM2yhdZf5E6uz0hU%PoZasgwLyf9^5_GqoP=~3WHy?SS`L!D|`1H#U zBtg{vcCHl-^=b~$0r|y@6XC)gwEpBr$=K=xd9P9>do4{M$tx*jP!(_w%1@2EO(ArM zX}3}2*p3)xQurrbbFv8r2=u%42S%T-?t|5lpKxTI-e7mh1|Yo;yXiIT;oa5w+MW#_!ZILOu zTc#6$$1@XI8ph@(Fx!0g80D7Zg)9sw>{88&G{fc(G>h}C?%Ek8(Yi$fRY08IvlIle z1J)Mga+VO9n2Z=>wi?QD*&-Xk?SN5!h*5}L_cQ4^wz=^YL^p^qiAz1R$L3H-lgL#G zoh338mbb{I|NgtXW;x@?K5Wal@U+*Kpe2#j&IE}97`VM{Sht4zFp&H2C7O~B2Kscp zyhgMO5v(3S(@QQ2B%S2&JN~#9w~>aT*ghNVFl&`CV>kJc z8YY-sCrzZ#0W*C^bnwN%g-pw+LS;cN#IWf|7qnFQgCqF^`xdO!Y?NkB27**sQ4)JC zXbo+G7wA%dFwvb)*dKAyy3B#}7#v(Bv?a)3`e7C&;)_UMALR%o$boB1+!_4|8O44lKj!=+FTKv>6JtISkL*bcckGelHQ5L2_y3GeA7!7$TA` z_ML|=H;e(m0g1KYcK32rTMWgq*=v{FF{OGFlfg5n=uE!e*pZoU~N0e?3o&d(*yz^BK2|enrK$a8_aHj&pAjc4i>=s z(PGHDBrn3AoxTC)Z^#=fIH{J%#qf!1ZiQ!JruM@cSDvC@%$5%gXOzYTe%Q=DZr5{W zjW`tgpyKm(C%Yf>&9m?yb0}>=RlMhQ#e_`lW46%9-kgr=Sd0twuAT*i5k`H*feSaC z5kC5(B=QNzy;@(+#Zd1P<`qXZ1aE^7mT3z)8_~h{7EghJ^P%I0!~)GjcN}_-x&Cn1 z2)iP#F+60K3|TfHqIAZ>@wgRTXRavFo+V|vih~$)Ag;dMH8=d)dPu%u-h_ ziY~VZiB%p@YgA!6KF`kcfhU^sqmY;yVjm2mz92fGo!#~*gWnx6SPAYX6Zj$OhU`cTQ@SqDn}dgS=<6jS}?omFjN8t zB7**kL=s(++c{LwRjKg68TETV`&~uJAc>&Hm|Lr`xq1>h)s8|drWx_x(Zbc6t1ztGEqy^^CJXa{r z3MNibNGt6hrL#Z=#fDT5WIW}Od5KA@(1Z@CSN|WI5ut|Tiga$r@VF|LWv$Vel-<3Q zs*Z?w`99jcB3ze-*V~>Zlrhd-HX&NCBgTnRNkXPVZY%Oc+RP=lF$tj%qVDb>rCGB> zVJms+9a$e#?veO`ci_##U7+9w!5XW#3c-r_z2)p-3F&+zWgts>(c}{YH7KbdgI8VP zph8su#BE=-Vug$VVqV`}v&#N5AA88>vEHlu-TD^GUzrl@AVlEq*4D)pjM(qHOjx%d%1`joe;@7WmNZ>ei zz))g$jS1!mxk(x|jU=jKcFS@$@9YrET zL6z8`jcqKp*hVc)pLGK#4j@CA9GZ3r>F5^=49_CdKY8BH3%iAZxt`oLgG9LVClH5& zciO{_vxXWQKY2+9DqWBsd`A{i3>|fFThjs9E^nG3qj6zF-C1dqSmWUQAXV-w~^8Pg#H1-GRR#S|~hp)Q+-dY$M zKs%zhD*tnHS!DEh|C((F(X`ODd;A011+rNnesUFkGS7#(W5EaBUefIq-Cj_SARruB zmyncld?8V(s=|Uu7Vi~1GOx4ia8ikQm&=?~wDJbI6G8tv(>7xiT=#1m9eW(9L8b;b=~Y_fcgEc-eJJ$znsOVczrE%oe-{&S zGmvQrfxA>kmX2&Yk)h=kvhBbDsJ~7}vjs6iASs-Ka|g|~NQ3*7FUpw;@{3BE(r&*`t$?d40PC14=Ytl}NLHXXkJxHT3 zNYSb}mAj==eNFGH38|y+LyQn4aX1exYx|NTh4LndGvC3-bDja~khG^})4LfZUekp!FUh=%blS}D)PTxIG6V|{hvDMFW+nBVMYNmS;3?*DV}Zx-8v zt}p-AI<{p=T&H0$86`d8z)|ldb#1>)LQdsXN#FS=58mvH6d0;tGZ~}zkWdIt zL-EsjjN&B(YZ0;y&n}n=Zk_buMHE>fqy-WaLfu40m$d?$X}tl1{BQ7|b-T{T*4}bh zTu5dx#2|N)$V!thTH7{b@xVIZZBP7MC;}wfJ)om8k~7ORY$>HM+ANzng;I83tXC$v zP%I)tNgL$1+ki5dSYL@S7)7>{#aLW3#3rY03q`6dV*Hc*oS2IegKsvM6!mwUGxdrx z+#SbRe2JEtb9c!G8Y<95x&jqH^FWoF3~4woM1rtObK@j}$h8=HVx9-xNZWlA1h6JQ z&&L@{fmzYu0zr%Z*(S#|kz_1@ej7R3aL#9NaiFs?hl7ESO6G*MDElP>vQRD_O{ht> z`UbGA#6OW_(d4)iszNsZJ?XtPv=#4j|K*plDl4`&UtInfonn?w<*G|7m8Dor?_vxT z@=iE5_|%O~$rC}QNLD_6n;_BTCj}};Mnog-AS7RfxU5K7A}XEhUcHshEHUD{`dEv% zyqUd0exV!@ZTO6ZJh%V|6_^4QFJP+F#oeskxp>_SLTIb8*wZzB>5i5kCS7e!c|7ZG@4n@@&s zWv=jz(%FNKFc`9_aX1^rBHLJu_+r0e2HcrV!cV3Wf(e=xg6PiMecKJiQ&Ax#qN3_jkX+WIjaVp zVvNShbm+ls*IQZ|Q+(avlhTb8QmbY25yMFmRA71#wVgD8S}iom3Mhrzp8JFKe-l26 z+=>Yf%FdU&?E?;ht%pUAXY6Ixjx4|C5ZYs@?k~jmq08f*H^iDuok^}35XB8!iExbO z!>*Yf)W4n*LP}1!TMYz4er47er(fyWzlo?C5&T5_6Vwy&HnK`oLQez1uq#0DV# z8rWWhAosBLC~IMsWTjSI$nl;qPh;HjnO>LV2Bc_@Px0&^rx*`(rhe68L_HkZUh8S* z>^^HYAb<$?{&Qb#Esa1B)4BxY3Wb z=f2}F$s|cEY;y$wHV_FV8FNvRG1)JX@dcZM?a5Y=!B^@fKr2%D3l`c-jyY zqycccSqH|HeoUSjM;yN9U>r@@OIf6ceTbg0XUDa5aHF$6VH#LGa#}MC54)tY<)fiP zSaaVFXJou&=4gKI91)ScdV#skT)hbPGHvHSgbTM6fF{sK?r~gvc|>jEQ#6{ zm!E7XZJy4-0C;9`ywfCX5huu&W*~aQtxuEda{n?m5J^O2cG-{Arc9!hB>U?HS<(jU znk}QF5N5y}cM6dGfFUq}!6sapIW}Mna(x5~ChwJ|TEcZ}dd8khj9z-8P2#5y!9ilx z(x$ms22D=)An`=WzsYVM(o3$AoN@Had0e3yS-kd_e~}#OguU{DNKcCtLK&-+^E1oH!=KT7W$raCi!k zM`KPBMCr@e37+3xir))TkCcqwx(e>W5(#!&L(CTA#pM$0^^u0ZVO}csN@O)4^9T41 zu&_g(4lQSIC{YcjFR{E4F(G$k2tEVwBbiqgY2Fv&I+w)3>{;6dF&jAZ7HkAeS=^wD z@XZVW&3oPi0}9R+zKiS@`!n65M=vD@2Ux?|N;pHx>KM#(9#f)>L!LSo-Z+CBa*gnF-W=jcON|)LQSFXb6-hUfdQXl;lz|G^(M))6Ryi zxfoPzAR_KsH;m)ZksY?%YJE?Vkacu!TpCOO6-K(T)-=($jGR54;WhH1Bn9;J)<*I* zC#^yB4`kUSP~gf_ChdygZg&O7hD%brwSfe{_V_!Ce{M$_SG~ zun+X9gLX(N^>6?K)!s7sUL>`pz05)*{`ykBE)P9y;vR^h!{X97j>+dBj9~`sNE(q6 z1!;>`cUBypG5E$VIZVPvi5S*5cP21dC6J)6kf1el5XnFJ!j zJ^ZqLU)O>`$9WcXSYZr%8Nwb~nh)X{Ppi;L;h?Uqi;`u_ssHI7R+%XyCskckmdWcnUdGDTAo1qnH$&~JgC z-8;-$-_684A6LA`NvX z^f;t5SFA1$>10*Ft(0BE9F8N+*bIg&!=__i>hP6i*e6S!!%RG6Jylm8*~Q&JZps|3Ir(X?(TDLe%-A`5Nh{Sb-G$|m zIxLfB!KEFoRMHESXYYzl@^z+6GX%_;XJXwMU*UPD%=4Zv8Gv86G=!J*(&sT zRdGa{6Dn=+r0$P{aP_$l%GXWju@H&=SxymZ4y6Kbm!Lp2?sX$zKwKLF|K({JPqDeG zD?%90fxy=U*3)JoHBNU>@E3mZ;|F%MCqEp_C_4a>eGL2pAj`%W9Z zW$@7pL4&6D;%PCY7qyrZ0WiZUlh?;`oT;0n(6uG;n=ukD@UMW+Y~da3ZSCsABlr~WmnBLsIcZ|!U9@`^3`y# zzV5HXmNBMMjc63WXdku!d{QZ{sUh;?bLxU%++16&XO^kUFid6}IIIT|_H18Bz>mTK zGj}9-qX_50z?h*Mdt>gwLLdlIrnV%-Ig~!#^eMx8lVeBRBs7HWJF-)1hIx_Yr#5Tp zhV=tZ>~4&#OLC7S0N#IsRHohy&y+jahj-Gd*~^!YC3-u-qxU zo5~O#32rT~TaN-9hX9lKTyMuY(v9tMIYp&bNm4l)1roWwP$eP=z<*dk#y}l9E(TbU zs?atp>Ic*wTRnVFK`ER?k|uiM0IYj+;Ar9 z*I98C`Q^0~Co6#nTFkZ7Pxr^ftK`+lbCpBXM+1}+7X-&Ff3kCd1A@N@8<3rUq>L!M zt5`UA@G(l&$G2p~rd4>1`;BEhyKB!x5f%@F$%>cu4hXIYQvjRE z-p&f>pfG#gGD<7 z-A`-$W+mK*(g2L>ena{iZD<6=L%)#wUG8uCVQL-e*!ds46 zcHvvMGe}_pNOu!dq1VM5L6-)`{xNft&G4jh2;mGmaHbl2Kgi$!RY+KvMk-tn`{xtl zF6`~}I0?n%JZ-O0odsc7E63aM(QlFZ2US%2%iv9r{Zho3N9GJMh4X@IfBKN;e-D6y zOhK*Fc%D&3vdBe{SyUo5;f9mt261jNiBPZMaunr>Qjkvj!&|~eiH~_W>qB1mlg-Fv zG$RRdE(zF>fioef%^7A1>8+{e1N{Cz_2&cck!abmYNrHhYqRte6j6Z9Uk-?FN^(}(kt>4|c3HZL zXuQ?rW-D@mwyz^hcqwhg7`Zw+b;nM3=eO;BkpD?h@qF1u$fHsMu^aVK>Zc;_4&KYV zymHuS+Yz8C4RUgEH`6J(ViJGfQi%I#TMZ!hw!|k}E=XG(VP*EDa!Q)P)76_R$Cg#B z;ai3N-%vre#EmuvL`*AMUWtOmMkB-=7Yt{S|Mw!8>RP85a@jDr z)pzku3eR&Q=R#^!;9p^~^y`cHa-fh7Apn7r)VC3SzbevSxI0RhB{{sE25CWz;OoQWzGfJF1-7bLG)CO>sPZ5QWckOrI)}5zv=jRk;ug9nKX}IN3u`pD~pD9Q@{6Il8;wa@A zpi$yUkEh@9bg%utksfK;tMINLPZ@n%zaYL~R-g&Ygpw7h67S`p7%v=Wqx-B{Kw9{C zEhzEJps)Y&nyb>~UXG8^y;AnN>NnqiNUVBBIk(ncTlcBk4-7Ts#_1s??hDc|P-_^3 zN4qtmV7mSrHtFB4;+F&nlk@TEO}x!f36X>A=mmP}5I@>Ft_+HUE8Ck*E5-Y4^MEL3 zm3oYawiOZW=|kOA&}D!@T6usSL@0t?pN8?!_y0*RwLcJS%vJ}LqXImrTPr5E!L4{z z#5-t&6a62%g}wh zWAkUqu& z$Q_yAfokt|HoN1_%%=2@jq6Uv_h((9d*%e<6L0nXm~%ho^bCBprDvi$atZZ=d|8!( z>iC#}rOklFO@K+mr}3fF-+1HkJ-QPz;}VvxLT?V|?FkUaxkH8@6rAfMiWTIM7mWh7 zLizKaLTLLQ^O0lNKmGI;Auddet>1R0Le%FvMAJGN)~&6Zd+W&!J%rAT z8$LO1_(-Lb^${m4EAtXdp#pU)Kp%Tjyg z;PZ(=Aao&1)pSkDGq=L9Cf1570E@R{Ji8dbsO1vMru}#N@kBpXcl2|u5F?Cua!D&t z96Fg!hwF$N8SlifNmQk@e<-91gPiDWysv=m2~tJ1LqR^#ySy$$A>}yUCqh$M37QHi z^}%7D)MKSSsvS+BpY=J$rZ)1lU%fo;NvU3odOJ1!Dt>+1Kkpy+o42^+hQpFtCm;fw zb5`p0;#?U6VC3R>O7zAaP17?>+bAY&jD|ANgx0gk85cAEx^|G7zs9m0u0CX%|5vD+ z!Tmj-Prky%Z0(aAVZAji%R|2)Uo$Cns>7`~mFqt~6-qK~a1)6hI;4 z_kYwgm2$Ldp3m-D5|GRIe|7j0eKM#D%O zg=)DBsYp@flnmH~5!MUGi1Xycqm)Z&oP%i%i%?16lcCEVffexfD``WLpYxiSkAp{w z%Wf79P0KgZ+f-*d*HylefTq+_Ur=gpQccuaU=DDF?>rY%*VR-ICGhs%dPp&s9DhE} z!}{fsO9lWV(wVYvO6exZTf{^DXbEYYj5d0l*dA@ZVu{CqPMwXi_?fGDaW}nOtFWoB5&_HgZ5L@LDRr%N?Ibm%GP|42Mu8rI zdZ`)oSu|ncs-cGD753ENRAFAsC0ZfV zQa7V$Medw_57r7*s|dUyuxcbwOb;4c*WgbUQ$S`stTG=kfpsk$m&mz(u-A~1bWp2v z9Pwg~ckUM+`>F3(nfu*~kai7;T3x0A&8_!`KA#OClLsF<3;F{wKko8j-NI;xE_|Qk zD)=JKaK(_3=JU>r^QjV1%|XM55H1g?v0gR(j49>- z>Mcv}Q7M)sGjHv)Q^@c}B-6Ld6~LpCc6No2z#!VOgyntit;sP_n5$qeAuccb)L(}c zld6*My;U!NcHwdC=EtJr*r(F~aB8%A(j!;-;EDv?Iad>JHrWrZ*jq87Cdr?q{&W`0G8lA9P4T+y%IAl>oig z7xLD`QMA5ZJKfdtTi+Rg)MHE%KI4$SEK3#41!_tbTUbj861}KT7X4EQT z`((0n<3dM1%3;57b^PnA$9{d%5tGx#1KJRpEt?aMS5+fe1An2l2y-@yy?Hvqjtm0D z4$T301aaWvC6J6Lv+I~_M9XOvNb@^ZvF`5$Eh0b3KNw=kAp6}VD)_S6BRf}mRG)Ed zlzdm$Yfs7iDuaSUd@0pg1J}1TGkA0;&d!?i^6KFQ$^w=w_wR!gLw+ZV^Zs5~KM!^a z$N{q!XWa)XEIV55YleZgc_&LHSegnFj3i5EUS;E_XWNe_p=+aCG7P-H00D7nS@!?N z7uKthq*WvR;mN3ik6_uQLSaPMcRI*~R{lnPT^wlhJn^}{y+M6D{7kEb;2N-2NL?e| zmK#J2fY@-tv>fA+zgwY-Q+ReLVgK~~55FfOZzwD`3;@KU)Az_abUjW)I&#huj=h{J z6`w;}q|5|(wm4UZhC+(Hul_Isow*Uj{S&LVP_dp^4I^8iTMpI_J{=hs{7DKRw}NRC z2-M~X_4CkceiWx|?DPp(hO2B%Q{oir_2FOBW8f%BRw6k!V;GDZ_VzodmOy19QYY=# z>m;pLtve(~!LmP`IODeCG(Fa9AtBaV0YBu7cxH5@`p-!cuk z)aWxfu;L#q;XruM{l|zHKpB=@Xg4z*x8zQdpxFCI)ZxdR7f2D}4Pup(qaTWY0} zYJuF3Kmu_(_2RIz(&`Fv5&vcDU;ttM=~%2i6YxJqYhq7;LF6L5SIaS7iq5^PFqve2eTG5X zOT|9D1uGQ&BlJS}n^%-VD%2y7hS*6{&kUoC$UJSwP@0kkJj%(nGFJ8!S<*N6u#tK* z)gJz^5YCkTjU$EM_q$Ymhfn_T`#;|JnhCAGXML5|l4S4J1+4qGCo!1~c4_pOoaCsHX=RseC)st!;;g z^S9ngDq%Z4vtpS$P|9Al=8%I7DS^{Uta5r3$~BWm^`8zegbl)rNr6j;=CvQ867*M>`5bcYri{5jW?8;0hPjA%WOJ|{O?TJ4F= z5u~q7E2(eJ@I)e`LIXm zLk?o-vZ%l39rK7@ArGO~Jd0`;MM)7PgvDzUVoCw7Y@d_&&Td@CN*BP^FPPrR^5I+pCqE zzBo%hD|LtMg|)L@JPnV-f=qmST~)8BbVvKWQJZ1UGeDyqz)3mEf9DSkM>z+W-GNZmkW^^J3&liL&xEbegTQkAh%Ocpk*asE z-!L}opqKSfv#n_T`M}Etq|;ed8SEV6!a!Jhv0P$*Z+MkhQkOGw@hqCgtrwgBx{EeF z^svADkxDgaiL)g^ohGct*wP(JyysJKM76miD)=fciqn)G4|&`TyEw3`<42<7&_PGw z_Ch67HVKG7(mpv}P@_lu)j+t`8kb;OG#A4lJuGUekBzJFGnXIk_Ha_K7v%5GK&0+O z#+Y?eBn_CgVM72e=jANm=|4O4-*qkkm6VOaMN{b+yfjjRDp+KicUQL4Gk7`do-x9f z!Sp%gZy#h5?1@jZs%d!EX{3fMq$BTCQ6d?7;NR&vAkIUf6pr#dl0@L;phDm=whvFF zzQ&XaCsRNuIUOy>us7MsrE1Dv=Y^c`bzE<>@G9XD^@a_4!L_0)TMt)GKQ31qSim9y zb|!_COQreyhL1HKrU#hY?A29G1bg<02krxJbR>uZTNDC9ygKxaxpHomn8t3pQ6DEC zUDt-NM^QN{E=K?6tz`W$Cd5=-^R|`YTkF+eCERFoH;1dwsn&x%y)0*=Xg3ciOMG^W zEhUx1TrBoZQADiB%;@i;#`eHW+Xcdlcw=Z#Or^lq8jA-WgALgR)ooea&L|+(evJP$ z)mIj#=snX=X+#m9dA?uUY*1Ps`I@%^3lvrWWCsr`1Zyek)GxV?&dA zK}x`|L(D+^O6XKPOV69;s@YViL{93Rmc&y`wbZ~^oD4!bhI8e6y&@&Qa4^P%DDgNU zG{ZZ$im+CjZWXG_M^%|6r9aUlqG4(N>wFZXxVWJMwa}z455eW zbcGDbc4DHyL=k3a8U!K@P{_qF7+33&wcbGy&&9XIM`F)bL%3izd?TM^@AHl`R;g3Pm4Q{}N=5s!T6&Z=ipZsfPQ`F#t3;gWQ17^5i~j>e1sgU0LB_)7Gox+`v1AbQ9xepVvef}(mMDUnVS zU~$U19X*DdW$#^FBg97bRg!k6tlXlp4LFG5?kq~HH5 zzN-xDQeC7;l!^0HUAmOD9%Yiqdh+CZYK`xn#RGsa&-d4cf((bL+E+JP0xkZoz!kY` z*?es^Z6sZ7t@YNoR&D)_(RQ!wlv)QjMojRb?)9z4s@D|!JM6W{T)zK6p<0xQM@6OW z=2d$)pVAQz$8k4p?uwdv4piegbqm5kL9b}9)*TWqxVW2~ix1)oSa_$t^rQb}u{^;nC4 z^FWCRg4X0%@v)j)+q8pd$XDU)Lf1(FgadR3)kfDPXD`yw#?HUdiv7yS*44x$=G8*Z zZ#t3+iKB?Dt#4~-JL*T)M;HdEbPsxAbFq&@LiY7Od$KZY$Fa5st(D>UDe1B7kc_Ve zEP==IE0NBBZ;M|eEVOeZ>jQ=vC>CK;nfv9$yhX?X46>$Gp=4q9?GI~Nsj5`T4v z)cKl!HFxbdDEHe7$fgf&(6TRP;wi=)fA@SC9yPWV$Eo0{t;!NqA7^trsjAv0x8}*M>h^qg zL)AcEwAr0pn%$~Rb2lnMK@m}~8s=(RetjKGs+rARz7B$42*qe2;GO1e<14*V4^Y*l z9YzgtT}<;uQ=ss=z>13mZ$ zZ{orzyk^X-DJP>t4NYQC>UpsiqDtMK9X~e+&e}EgZP~%GG9&Z5tZ?Bfuv>KblSiKWS{pW z_jw~9;P=DfczQHm8@L(xZO*LdqgbbLuaNp)D|s*%keErr8R81dZ=M4Q96SlY``G{u z$!f_gaBq5R#U++_h-N2=sx0XED1@sk_G~Xrgs;ydnnx|%InU`MR?}XMnNbhoXDdMo zB0dgEDL=4&LN;gY&`+@2+?aB{WLz6m9{)j-V~&2P0;gfwh>uW(_EA4FFT6|V<4hY4 z4&L7%drA}L6rJtj*ndEhACgS(EiytY+Me*hA1{?v(~#!vQPIeKTRx(i0L|%AzrSA% zFTm7ZX=8)6A|#xaS;(W%B`0pB~5Ns747ml`sSLS}(jDXl&kT!;`RmSVWj5zAt#xuHKZxJ!E6(JeoXHFILes~`*Z`MaI5)D{lTi9e z8dw4<;mA^86uQZQ8$euy)0(((A#urjQ zstUKn9+a~=cYz(*u?xX<;g8w{9w)x-LKor};bIpiqb~3hs0($}TYe7Ig}RU~z>|fO znNv4o(|^c2z&B0fgGL>|7hmr{KnM6j>HsCj^B6SovIPe(-$w1973J$}@ZD`gmaZtU z(iMGf&0T?yUK;1GsPT5Djkm39d~>m7ZIf@}w`agPG{_&UTIGK_~pA1OoellB3s33o#?yPTxY`p1u;{>`eX z^lv&Z-0wr6h@|<7atb8F0 zjc%s9wTMHv$dYWHL)R4ih4;R&ukO-f9!r(@a1}`XRo2Fj~<(3Dh51Fb-=gpZH}+>p7*QVjo#OVO+qz&r}y!T z@N1Lsly4q9mz*i+K2~Zj+@)PzFke}$aw&RChmQ2Np4XR`41Mvdy|sSECk7qnx}D}b zO4|D^TS_iPFG+k#FWd5+UeXt5ZpG2M&W1GLvH6M^Eop7{|87b1-FN|#l=1>yef9#p zDCmm|FOaQh8|97;5b1hT%toWk&0Si|+Bdpi`Zk>xI=Iue2^aeozxdkVjzk4LF<%1t zQk52e%Tnb_Nk=NlYBME?JY}&@i>se3g_D%uPX|TmI$j&C_2bI{_68}w66H#ervOnV zU)52v);36V>Yq-)!(^q(jk#XM^`(`7j~B zrJU9+-}AHIZk%a!P_y2sLe;Ne(Jy1U+-MNmVVPCz47<<5a6s?)o=hfWv%wLTERebL z@2e@MW`R&zuBhN57;U?8dTTww#X_<^#URC=>+x;SU>KL*u;EuV6fV*WAM{oYIA)1G zR`|2>-jECBQx686!HZ$NIBpFdO^^yCd9BynwZU@N2&h3rL~z&EY@Sg{-D%Bqa+2D1 zyg?gCHTxK+xakN@FdhyS@$3Set|Mb*UYq1>qis-4i z2ZCpg2PL{eE&Yr}C6$eutBayc%{^&zbcJv3FYrp;u7(Y>qNKp-{Z&zZ`%myuk!L-% zCyk&zPBTM$^rqI{%G!%pyxZP7tvOkRE_AD{aT92*n$EPfs;N?If3Vi#6~47UP`c3d z)3xk8`+M=N(h+R;~9;TzFPTUsm4tP0l3 zPu7ZDp{@KRDOvg+>%0XSC%y?SkdIti=!!bM$J$zmSNL)IJsl-+h21(}t&+Tp@0kaklv$Uo<3tCT5IB-er)AazjJlco$0NP~O6dm{Kw=1Fg{ zT{)4Jd3zb z5b&;nVz#ft1-fZUmbE1910A%L0jjwe`iXR7t0gV-I7netvJf|MT?0z zd(Z1C2t@-*ZSg2wc3W+-$D~?CxSjS9Lp^&14lMDw8ly@*+n|0|sq+ zNgIHKh*7uow83-MF7Gl}zSia}Ukh5`ze!1~q)O5>P=F>)%w>*A&PPuCe9aUk>Sq%K z{F1~GqeOK*um6tMzcy(4tkf4`)KU9#q*NiR67ktZ#OyYHHBLzAQlTb7j1&hsq}n>* z_N5I7$Oe5Zm*e%L$~lfNCa?}J39~;R2NbB#9|VoW*lv5DCnvEO3QH6nCxIS&!!9v+ zSVoKC%Bqm9!2hCkN26ucpTraUU!d%wkc^PzzyKL>w#j^Lc%!W^ZYc5Hm(}zFD3J*k z1QzPSG&H0At#HzRHL@ z*gQfA;DfhEnm7cw^MVgR{Lt$&x6y~|P)n7e;GX4At&3>L;i*4+i!yU6|GPelGV3^n zhfN@qo=I5D!Z$APMLDu@jHz;C$5cC>0^4FS52N6bgujSTH+(X1MKqSGD|-<6w4Apy zCUS=*<-+EP_eB$7FO^@@T@t09-D!rY8>_19T5i!pS!2LZskI52ccYe~?+mt?lh2Ui zk|OPOkQ7BiC_vuH{r$p-m<}fhF2}(2x?rnPsRJos${s_UFyWwoU31{=+ zq}FvSekaO3fnj9iQ;zHpsv8Tmxguk&&hs)5avn~cGr!hFdE4Apoc~pGX0_ykMtkqB zPjCqHl2-3VlQBJjR%IxZai9+;ACKRk{ixhJ$J2hl~TS_nNRb+W*X>1 zV1nalRjF3p@571gLj+t|v0n3jlo*kN^0fVE7>?JN68vz}2eEC-|~-w38yVnb?a#j~c16wG=^DV_plKK&jjcd z7Ny_V>lcAR668eeXeNJRf~ACPyS73=&1Sht>a(xUAwMDA3c$?c>F}}?ro5id44>(~ zzx63Tfs0!N_<)OB1eMHHm)}XbgyJES2HF3ze0)K<=++i_|8%s3&}8?-fkuI27cN8j zVzCO*qLNKI4y7;sF}pV^pq?ezHrQiXeXoRZ&*F(g=$1hRI}8^-4RgeYizD4y74+ z`HSo|-H2#w!9e>~zB2ih@|ED~SE{a@T`XXUMjfGg`B7e0g zahcT9GG7t^8yqEneyzGii(3Sz8NOaD6X$*}RnXJRe&n~I0N%x-qDFu(T;WMmUYe7J zqFBFHR@BsUo#BgxCt7zB8Vc2Xb@CdNVf>#@9{M`PTuEQI?^T}A0&KYcAkA}I6n>tI zqAF2-@_=%LqAoj|=AvlgU&S}q0XI^3GA~8p`;9+kkVh1#Q0e6n{MRfh2W{M+R14uQ zT4xt=Nkq{y`wIE9nl>Ps^W9N-yG^_$Pd~f&=M$@*`XVTu<|ZPC^ClIzsORW6*0&Uy z1QA0+!jy1ewxE2<3R9{sTqsPLoEN54zSy+Q3OhzeJ|dS~KMgDdw+|9$RXhAB=24s%?`UH}K& zKqeD)d{8UF`xGk5T1rH0$6kfO$W4PA#n3rYj;%SiV!y{8x7pa^M$$IW&)L}H#(E-- z)fZzgQEGNBJQ00jK zftoGq2!Fd0p*^BPGl?C9Mr!)oL^fH4GcU5q6r1HDn`u5-kTTOJi%~j-(OTD#v+G<| z-+?|#ZUphD+$hC^QT?YI(*+xMf&9B=;atv@5Qe=QT=wT<(eSfH^YRE+f^%*_dF8aN z6iZZb9Nrbgb(@=?Aevk}u5Vu#OtF$h06G(L8Y4NZ`T4`KIj4NoiG^=mGy19!T zw+fechYH-;5}-b@C2lVP#k4W*>cne%w$5%i_`%B|oz_An%ka_Xi!80p8>yZx8eU$N zUtZ~X>EvfF*>0)}3jx(;Vz9j8yrNi1SaX!s%y122v^vUsu>~l;fG|9WC&r6qVH}8N zbZt@-*37erLH6i2suYr!fD;=TDhf#b+TNSW&cjFZ|Iv}YWB-V-dnLtWd-Zuh1Uowm zyjNQg4BQz7tzhqz3ULhb8lX3+TETl0Pz~=uOVlj;6lz(%99%+xAW(_Pxb}W>j?!Y3 z(Q&OSrE%mFpl;fCj)1zIlHexst7kyn%`Y4R^^6pHWDi5E(jE8wQ;vXo4#l268cTN7 z!Hd?e7)$RrBGaVM?wl7(lHE$H-x`M zKkAz@wlybmy}!{ak3U69H!Vy6JpJ3^CSHA0+?3y-7IY>H@!y%}0tjvbZoXocQ53HpfnDk^+P{D4111C+Fp_D2mG5o#qX7x=V>K3>UJEIH_tYC6UHt= z$)h@LU&OLy4p!Prqa7l=s$Nez-z37JiY+iwl;UTkC>=gzEUK5b({E6egiZCUTkUhc z&yqJ{M6Eg>Y4A0nSB9F-1!(ZJa}3$W1jL0c3eDKTbYR2U{xlV;GJ7{UIFfS%j*XBQ zRnzFnV^pDnYPI6W6!sW3YGGmrIs!RzsyD3dhOOPOwi^UCMX(?k3;tYMOr~9B=Ox?0 zjK;1N_N^eaXoBhnHY#7~xQZmbKpfblw#m6OIxPn#O04JOLnuWeIGpdYy0#-nzIr8# z2|I<6sy=0_lh4%T^m7?n##qRK9%pg$-KGtW@<9ahLBRko!me5JLz0gI27#{F#yuQ? z5?8V}_@thC!tZjPaI}So~ zZkcDwk`$lc0b;2vD|5VK6Yn9(Kn}w%bp*?raLo*2%Jxn+fpFbLdcb+Eh88n3>!p8`cHsPlIY^p9oquVa>H>5VrbQwdDr;b`0|bXWEN7Y?oBv4h>PPW@9(tj4bS8;1GVrz)4IBKD{2j83)eNR1@p>F9Ar zB_K46d#h1A*)LWO!6sgMZT8}qaLixQQOo(%4vJR1`0R=wd-1Ds{#6|{RP1DbUq>Z3 zP4R07XLwIZjRxDSFeR0yd(`|?!po=3`5|XG@ryHzkUPmI)vPLm+~bngkwv$3xg`=t zsOC+X`6}vy4kI4Dswhi3WaDpGf>qS7=unHlXF$m+TH(5mXKN)r{kE&-bWumwuH!Ed zTTjVc++IyfhA^s0Jq~)}&2)kKgGy+bU4HllV+UV`+Ps|A++&t;{|W6K0m1ZXjox>N z?)V_ZS?_#?5dk|DNB@|7P$2-5rwn(?!mBqBzoBLl;#nq^EV_`w54W+SkNXO+lX^l?2JAqh~^2!&yUwo|Fz z$FE0~HE>ngTe9SLFAB>uY&i9lM{|2NnkmCcfPwCiZ{S={G*0B@Fe0a~#)>%MDtRc0 zUkc>NtW22iq)6J+Mxvm4m?s_DVkS(uMfB)+?1p(>C3Q;oB`49TP5YnX$R1+aA&r*Ef~d{iSPBObWl&O6p0TEs-g`h$cGxd@$$kM@ro z=eUUvyopcHlU!nR@q2%ee3N3oCvl!FtIjbv_GD%w*qsK~9NiKh}a0i9kKM2qfwgBrVcA znOvXqdk&n5*ndO-(LOm z{!8y1$TiNz^fbi~%c!dub|lXK+ovNMY6cO5ddjN~9ANekxYI&3921+N8AE=W-{KD*%y z?FdOsMFTaRNj&TxN){Ky<=JyoPP9Eb{+Y9)@)V~pJ~l@=9enn9;7*#gjm0$JU&-NY z?|8KtI0kha+0%j?uZt@03j*!J5(!6G!jY7mk@pmUQoE6-0pv~y)O=qX55|JId84+p z-wjLSLt&md&_<{`L46m1%%E3s8ZxE2t4$N%nhw_Axsl?W@z&l$((*Xe$izI$K_`j? z;Ek2tXsN4PFi?zK8IQ0$!~cREr3?6kQBkMcTGLH7T=i1b2=nZB|6oi$P+ z449<fF`~79QfrB5m43)j{8Q?MZVKhesYkP|@NAk-5{Fka>J9G@;RkYC3->j}8zS zzm}F!7xh;U@|2fRa}M6SNBQ_94|O#?JId)?!o%uxLVx8T54B6md(YXjwR7Ab;i2xD z-*$!*fI@bLhk7=%GrYK)#&N_cepfUlJRId+HP?T&@iU1(7R80LytwPS4%ZfU^Rkn> zN_-Yi?z;w`&1Mw8M7Ew@wzbRFb{XA3kNXC83VA1k`4uNz0$SO=wY|=9F%gN**$`5O zeik?6x*E5CVEU!dfRuD|_aAFhp*vX=pkM2O=aYVFqol+4Aqo9@kYRW-A&`Gzk z?qi?@Xq!_@wp_u;pw3pGcJYK)-sHjQLul}Z5hvKG;XKBpT1TKJvzus>iXB@%Dr6Ln zj}$>b<^7qMzG1@Ij7k=Rb-f+e_fqD8_g5o~t!1QZSKdD!85O85lmK3zgduPT+}=ug z{bqu&ok+1!c;L7|Io@jSq^UvRQZY6%oe~SHfy#=P9N~}j=-v>MwRpf0EJ3^Q%r}9t z8jn_pQD|8AGQubL2Bn5UvL;WUcws15ra5FW4z%8k-*rmR$Pcl7u*EPcK-6ympbLjg zSPHblQHV#6+Y4hG`@%74fsKLgol#mUABe~4W*6cKb(x4U2q)b}Q9NvB<>h)0XQC{0i5JEr z0oUc&`mW-Cfq^ea%=X}}bJTT`GwLw3M^u`9P+0X89W~B95=FRFVaMz8a01TpC&%~qE6J6wdK7rnwcc9IHlu*7SglJKcXZwPg4mu- z?VS{H4_E&x*Raovjt8z_(L|f4;g^))GKa8omM}W~-~Z|5aR71xyGp!>W;^8V7)8FA z7(lI7BPY)?2I9;=>^6@JqU;hR8#Mt};t_w2XVif$N~(})XJ9xifq#@2+;GiQ3XF(f zX3q{&vP{$tlNY#vZbN_l2`)_Bblk5`uOnI-E6*e$bD;pw!s~(LKD( ziBz!G_)C;u@$-#B!fVy`SjlcDIflc6DF7C4AmQFizBQRhTp+vu^m07yND|im6g^Mw z<6Xu*M=D~%Nl6!vaD6c190)cotq_+M5OhqQ?%?~GRRSbL;fPwy5FEdRhoh`VC!AhS zOi`O>38XYdiPa+c>$p9$An8dvDz4Xc=j3|YE+@Cf7bk%XDdHdB{Se=SJ!_EEXNp&& zrf3;Q{JS4Spsv)V3=Gl;%kJ6% zSpa#Bs~8BhkNCDN)i;AfF?6LZuYy!=Y&F{Kb)xca(O-M$e?m8b5{X{cm3f0eI*lcl zW&0R+UfcoVVEfbkRS8KWVqqW`v`K{XD4*u1932KJqs^QTS5~%-PUPhX;_@&AIe826 z38)H8nFX7;FkY|=UiV1F8FEz;NDag*;c%U>ku`_lv#8XM!)Mx;WpQ1S+FKB@BM6;= z$z;XZ>U&|yaHW&y!#CQM0Q}iVgMk_x=gPoM z$2Hsw_6xxalfg)dEP+2lISxqaxvsdAmuJDcw>avb;Fy0kP zPeV%Y&ip>Y*&_nNdCj=CJMalYUw+oNdQQ@Mn-YxTrg%z;Qw{78P78xfm~z0!I4)Je zSQ|nRcphNc5M)LMjXR7PFG%@)6Ly{b^)|trC5mk;=}mYHIW6iGBZVXw$&ZtwCJ0$QXH89E*?5IU_}r9T{5lb19{baY6ZNS4nTqk;yTjBV zNL!FVs1^kb;r)R>-B@RgKf9_q%l?saJ=yH~=l$;1n_iPoj^DKiRoi;--%hIJe?0M~ z;KMQ_^3b{xl`E(#;Q;8?J$kDL)7pk2;|W3`wll`a0XIeGM>g4ebPKiM6nn*&_(j4+ z&G(``CID1_cIrA5AbmGlC)*M6V8L*@!b@C_wDe6P_C||%b*$L@y^Kc|K!{DlW&igX zcIuM!rU#s{3F%fWSI1>ql3iJl7m<2}(LhZ9+!FBisWSJ{j%5VWDnWL=%lYjozd62) z%*o8+6wtd2%n2DgTrQlEFo8&@|K-H~P?FmN$Z`RiJY~ap#XK0wjLI+9&_x<|M8lvs zpDjqqTp_?-FOUiz`R{;)3URtV?4NAs+s+~Ho?qNF>G8Z|e)vyt0$fYgNuSJ@EuoS`a5aNTG_!qMEv0qS* zgx&)$1a)Y)$2Yz9qzE;aU@t`DPyN86T7#Q&qr{xo!n53C0uFD?Hx(H6dy{}8>aJiY za{(>s7XsOraEf|Ktf`CZM4Ef{{DM)0eQr-9$2P+^h<#{qkoE$S4;t#4e zn%Am7E)T=AV6K@qy?ACAVGKYz3o;>{82{WkC7I3a>qCes&fYe6+^p1H^I-I6o`v^; z!OTBh2Ns;@scZogqKP2IDA$?HFCN~0YbQJ%}SV!KJ;MANAi3Ct7*;Fk|=B`%7aWi^Nu}?`PO?TBOnAiRH#ESa-Ha5_Ze~6a8!LME(azB zq@>b5A)g6Nnjne_kzQ0p=me1&>J3Uw6sRHV@W!2#8}VEUfS;2^Tu;N0V*>0S5`r|7 z(=YdrkBeuLl7biS9$Ui&7AGpG_OVU`i*&WPZna|b3{A~t%JtO$;3&mELExkexSmt*xH>DI{(E1XmjDl;!P3~ z__HSD#T-pCYU6hKFN38^9D+!)AaF(8%f1B0fmg%<(p!XR{`)^&@2CB7cbMn6Wp(U@v+bH`}T}qsfAhFr6IO|3L+v2>_ z1o-S5@|v+*`a%%3f>_oOLJATx4t;wDm`IIW1Was4y#eX2&gg_^w3Qj9r9B)^|8?@4 zspT_%&1bxeOlH33)2jx;>Gg><`nw-e(trHd@4#vkt)Ya`!P#Smy1~m$VxSwX+TC-} z0?JKA;-k*aKvN9Mi_5){yCT_aY>#D?zDkj zA`k)S6Oy|3F~N*nWSmENMpPN4o>S8`wU+hUx{w4`h7m!J=rSa0GC0VXoxYl-Tvw#Y zJVz_c2JPNjyZ+nN4W*9;0TH2;U;IK!sqxq|7D)Dtk0kED{c!u2e~(xGFNG>;tNGc5 z$Cpo=!?61HL%fivX8(4jc`hG0XPCH5Ins?+`$(f6Kl+eRM5MfEZT(JfpxL{G8RCX< zl=(xAF7ZzP`9`U{J|U-^%?p2LG%w;V$qS^4GP^{5dNd2DaVqQV`@GC5Hf0a0TH>Zp zSLPW&bftChT~vUoo|aZ{CAm3iV5*yZkngYKFS)THOGIwW3&l{#)e+aXr59l`?&X&i zbJdkD!Ethx!=kY=E>bZB(&s$Gin=Fh<5}_7eCqBxOLmTT#V_K|Et?US)13*pt~t+@ zl6R_Z?5&V3RROWjWlPnlL`-{fJ{7bCMwJ+k`nMzqZsMy1-u>Sw&P1Dvz3Bp~%M*Z>C`j0to&PiD^=eW5~D*?l!lC4VxQGo17s%}L(F17iu9HwMyr{SL7Tybw**@;U##saNaNag+d$5>XML zav)rsVI+kvJy+Gh7s6kxxO*15=Q!mS;Q};*G zR$%P|5^ZIQ>q^R!LNlMehpSc=4BWSjR}-ub7bc;Ulg0hFmDC|IX2`e>iL=SxS6TSU zH%FX~M^ag+5c_YGKY(92PYiN|Q^0;*W0Cx<_cfN-ES*y;I!r%CaSLT-DX;VqfOV{^ zjIE%0Yy}Zrcc7p3R}e(_E%j~n*oB$FAB|7l6^+Mofu4TANe_)}J_~loNL!JNA^CEA zzQ#w0$`V(5xLongX+4?#q5ZCI@Gt6=s1<@3!dLUK-RtjbR|BO(I(olszS9eODu9pL z`KVYRoIt%_*|mb!FV|WHR#6x@pZnTXn}2o~s8(+7>Y=jt*B7g9@GsS#p`vrbfvcwy zz;)g#Q4{DG1ZLVXhTg3EMiKX_bm*tcO9p$HmW-S?D_Mh>3OL@Z$TbQZ+{)_fi!z(Kp7q+HpOHd_(<5@mKmU4$A-Y z*Z=r`;;+9#41;#$=u?~}5VwXa?+2ekq9rze|2+jdxT=PIErK__h2j2#Bb>3S901p~ zVxXY_p32|z3LVK3Qj-m!)D9rt)4`A|5JZ)(_&bd|P_Naw_z>Qy4YK?dx{sy^S=ifg zN1{buLuw8?B@mhxx%ZdIK3v?j2zmWYgXM|chuQ` zML8Jm$)0FWp&65yGay2so{?YW;%qmr{D6oW3hxeLROw$5bpKH^N~#X$mQ+%4JsY9K zOeYfHFV$sN0`WMY1djrHQ|=~tc3z#c2!(l<%f4=tdbT&FFq}H-{rxD}X{Kvpjg3P)A$Dl-Swaw752N;E{_UK7 zVvF2TNP;;^_AOOTtO=eHT<-`LK&T;QRHWB^BasAUU*tC$6lcMkRMo_4rtydS zU5LKrD~-Wo`Pp@W%_rLC>FhyTKqFYE$K|$9WT&1Vb3n5H;`Os-r3xucTr`8Tyrfi4 zi%&Vtkan94rLU>mP}l>K%fnWx+Gjy@C|4NYg4K5#Zxk_F}Y!5>0PKK zw3TJq!8abIXGqdh&_|9a@{_O;1@?EF(>6}Q%um4>i8%JUnvq(u7Cz~QvoSp_6+|t) z=<3XfK*pg!OJH8i6?@W(LJ!=8$RjfaWks@*Wf=QE^snOiwlsI&rtm@iGFZ+cneTC# zG(c%(N|gr|8{5Sf-YYhcc90T60ON?tfGnw;_GVkfeSM?m1L4fSad)Tb_tP@lvSWgSD!*a>_1 zHAgTNRSFAPs>H%Tr^2a8J!ATVtyH7C*l<7gN;IvC)U89OQoJfbsXFg3tn2%8&psYU zW$BAVBg}g%(D@387wEQ>>sn4eoyEZz9nDPXP4=)gIF&J7BV(Mob{9k?zV(Mqqi8v2 z;iWz`#6lwKBXlg&oF(#9Z>1_9B#wy{uL>l<-i*Uz zoDF+aul)Ve_nY7vP6OrMeYw|Gkiuo&w831a^K<@Q!#DcEB7mzd$Zq*6v!BFSe+dti9RObe&Gm?(|Hq)YV*4v-?#^tp^l$GefCC`7 zFEb022KM-$DROZ{>->mw*wDHe8~ttkFVN2hMnAP^M!#!vY@!1-)P3)L=Wa?a!>0vC zI&rrl343?iaBj{aopQK%ri4j8)9ti)1s4s zh7~WsE?VLuH@U%}EE2M{5l0Al9D!W~dXo&nR`UJa82llN8(f4*v6n$`1wI#uKk_hq z4KhVs|1+T_nMnv>RxxDCtQL#C$)Ho580iW%(;Hujon1k|&g2p3$|Ca{sL;8s?+Od0 zRry2ky@IjKIg(J~M9U%xBR17*A%CYTtMwJR-=9?V-HER@ezRpl>Fb2~&JIguN!h*w zqoVxmW_vH*G_UY)Fa41;4iI_q*#_eu!hKF_+fKsm<+H1n;tEl${&}CF7jUeQ9r3yM zRsju?jvGb=;y(DEsZ4OdBGHOe0tSsZF7=!|F(gQ*xpPD;b(0`w;ndB|IjaD;TE86d zZP;e&*xI5Z87gV#JIEMeD}%ZdFc*SqsE2k{)O|`siZ_Sz;uXQtiiepc&nl$aFUmPX#!;oH;(E$gt!e00LMyACCAWvaVkkENIZy_UQ|?OB&GR+wpm^jO~$pMyBBj+;#&OUbM5C15rOf@kQiG09lms6d5ELY+Ex`vnc8)f+)|o zcXqoL&5*jD_}zE|h*)wimlu^q!G%`jQ;=Ou0dgV?aAhVEV5&>G%ECKDEbFl*!3goU zb-uxE5|oY}PBd_)?_mcM{Erz>1X)l<6*}JY-eFuQ?L>;xCz>!+j;dkE)wgf!RypC| zkRB<=b;GocNjZK#U=k@-5_g9^nzRN(g6W`^>y@SfD-z2TV`yJH?Uyy*_&l8SXL%~= zMLtFg`xz@-1NBWxJLuho_x|74r06Jv1hjtl#ILHKFY)ImR`rcwVm5k!DU-z==>}E1 zwy;S9M|kj8du=LS=M}1vTG_B)+TCmWJZe|*E;=0lcJ-t9`di}GBsCAZesMR`e3Z5+=7Bd5?en57B#4B6W8Wji zWX&IS#q35zLj3pq8ziC5gy;q&6`YGNf9FLau z34SD`iiS(a-Q9JkK@x^0-vQhu0~I9k@6dDbiqIO`p7gN(KFwdBq0xw7z;QvL;4e(cC+6%&D|A!9|=QxkdS2kTU3YbBFgc&>;EU}9z5CU4B&N8;QMn@TlGDe9G=_Dgj6-d@_ z(DnRqT8+;m9#qPjbO%uoW(aO(b0!Iejp0|0op9xk8f88BO-D^A>uz+hry!-1=Bam& z20&i;&4)|b6{Azn3*-x6_3x0FmeZ=eaNNi_BLl`z{mgXZm|VJUl1pB&xdaowOymOb zyE5^sj4v#JHNUSaW?iDn%bv+{e@IUSkm~rUMQWKI8~>a}wV=K|U z$@!8Q@S1B*#s)$QZZMg-0R-0RC4HEs<)Dx3H#ykMywM&-k;<|vgEEl(zYHt@j(`ME z?E_efv_f&;w&bWa<$F0Pw^+T&eR7^IQjJ#%)xx)J9HtyPgtITmseH#KL-`IUcFQax z8Rh)P>#j2ta!vg+S{}x5n|!jW#dH|Lyg4J-YQzq;LnvPv40<6wD3sBGbMjJ1?423_ z43I^Qdk204S_4jwp-hcAyZ$cnn}jPZIS@AI@;wUCs0OR-aT)u8`5^GIrRZ}SFcSR9 z8vJ%9A2G!N0d_Z{`3Zpl-Tlb%19W0`DZO8i`@J#rygPt-Lpue7$mL4kP+E5Efn>~) zOc5nIrTXL85x9MMpytfo{^j5Qm+HU_daoB`Micqw!#J@YR5Us$n%oa$a66M(R`@C zS_aU}+tPpko09uOo?92i^#5IQe_PJWk^fz?-P1oU07HrX!OcoG2;|xS&*bZWLW`%Z z#b)ZtFZ28@(JOyfdebUgf+cU;HmO5{e`Cwlmshs>f?zR+Ni-I{mP32&o_}7SXJmHQ z&mq&Wqa`y9^s_q;&d@`~;feTJs^#FMOk9?rVAqy@so#AS3bwS%e0ZXjY@g+Mc!GkD z;VeACuU*~fpS;y2Gy@Z~oo6g`DV{PaZ{V28d36Kj7D{!)bnfxwleC)}?JoyI#xwaX z)Q?TLnprr4gCr1)V8UBoHLNp<>cn94TLGKj2-y540Bs)ZHgQVRt@+6*x$gCw+q;z8iOb7-F`d=khO z1A}m|j%%(v-oKoVR7a}i(}sPZ&pe>0V#{V&4~%G^EB5&4Z+cNEU(|0$#k7H#V``ff z;2jw3hZ_@Mm0J&sXlmFI{&xbSK)p37b5sf~1bg4@jBE5M?{u9HgkGlfIsi0BrSW1V z*&LjZDj4!K8-(~LMdyNfbJ)lEc%s!|L)cqVSQuj%Y?v;urW*|0wrO%f>l;N zKa3D9j{KYipg?8-s%!o(9#&!#a0o)tY6?JWDpM}fjoxFv$5josYEkwMBX z&F11G><|SbkabYXnLa2?8J6Y~fx;ls zk;`EW9AbO5;|K5Wy9N_5APC&;%EOZ`G{j5`JE+^TLHlV0)&7$#XZz3R5w92 z*;W=<{g_o1C7V!_G*5w-7x1;;A76J-I!z(e|o;f4aZeOP@K3v-3oo?|v_7h;!%(8{IcD)Q$SIxOu=_M}$>=?`m z;5674S0HQc!3rsz(9;ml?gxgt@}tm29Sqm#Adnw7Fm~KJ-vLFsaxTNQl||U6Hl^n8 z9|g{c-_0W$^@^Mn$LG*PZ_VL9=5Se_0Pf2;Q#gjzaLzW<%p~R@d>x}h?~c78*xy)U z3FK_U+eBfehHTjDs^-?SVHXVLe96sg?r3mBxYi|?jbYu+o9wAHhDBhssvN?Q>+}?2 z>S=M}PO5(4vDz4*T}~*mq@sC2ND;9d+RTDuJh@tl^M& z(&9%z6aviKw4I-!TepqU6e4L?4y1MN2NR?P2u;1{SUvVkcxo$1#?y5z|rl_Y90x__2T6bwai69=8truf3Fj^>-?Jw4*Xcj0!9ex zg2Ax}=zG@Nc0D2dkrpICXioSN_#0=tiC+0k`Rz0AV4R<3JU>*2&8c60BO(H1=|B(m zrYr7;!ozuSj=(YTIy1-?HfSI6+39mlWu_a4qgDw7Y^x|?nGB^GY*?95``#QDPZY^% z64WhD{ljMy4bO+3n-xSEz{DEZ!9yGo zBFm#Xeq>Q4uepefg3Pzn%E*hZ-@T$a^PW=h-W-uSFGM;dQ@tH z%AJv&0;K*!!UU%|;jt1eh}QP=WFEy4?)Z5URpQYWs>KGEgR8gqRa@Hls`)(KMYx$Y74O%3+MpRb~s^8h)+`G&cDbay?SF_{AU4Z4~64dA3F zI*TI@;6Z_*1w3#a#&TGpZxib2(-mIh<(T$boPD9Q{EV{%M@8H__*683^#ZCp312%V z2qAW{q7k5Mc7OGLy{3BlZLY1NdMv@_)eVZpc*J9iZ8Oz}Ov~}iJ>*hvX{3RLF>Yes z)Q~VcLU_{r>jMsa&(Z@-y$NR#>mv&u8VK(FN~_%dU8QAuBx`k9x3{0 zI8%*TnA6lK*yoyu`4-eRTq+?Feyw|midoX3wvc0ea%N8eTn^BR+1fK#R3FYy976Nz z1^zbma^p}ph%)(Lk1mKr6?u0opUJXDiXa=7t{I4IuehKFNdn7yP@JQ(<4 zW;P`=l$=edp?V#l4KgI0-A}_ri9HBYWut|tj2XCM8~mP8I%)Tn;gdomZn#3rloOQe zGUDL+9I$%i>RPH=jta?3WGGt&WIR)n;!zM zNMY@sfiM}aT5&Mf_Nw8^V@76;(r8hZG36IO=pa~_A-rAH({UEhjLdNQIUUpMIHXff z=0WwEG7Dt*pBa*=`c(NUc>XTSIx@}=!F+@!s67#Sr;w6+QN#$R7ilwy zeRDwOFI!1guMMyrdGzOoV0GCLG@zZeS%To`fW3mUyw%=}zjtgn%db4v5E@R1_P{(lPo{jH{L#Cgb5~T8IgvLb zCo((176?V2`;db-Z}6&nD}1&ie47s8$!v;BMpAhnSWU-^gt5gBbYL+=n5Fx(xzOoU zD>`~qU=J*7Q&P-5rcg+l=T`iS&XAiq4VBiaSfM}%pUmf44V3Z8V61}o^luro_WoXa z_`$e56xi2!#e9aG;a$Unlvn~JnlpIay60ep9AITxB2b1u9VM!nE8qRgk2i=%JPx1J zDHV;!vYt;Txc!pxIAWk!IZbSPRhNiPG$GLi=ko$ruRPn$39+@rt^MXy>O@zkDz%=y znWUyjgsKJOo=|}9xL5zl=3V6Rz!>{7$b8Jk9@Pl)Q-+J=wmb^Vwa`Oxv377i1S4D+ z`K-9(*+_({E^2P$=dY~J^##|Mmfn{Q@T8^n%ECtPcEE{*3>e%f^B&mm;395E&Eec} z1G&wxRwKtfcrto0GdQ;ZB+N_)xIe!PU`E;sGld!rvw6$+Lu>TWXV?6|sZHTPm&7E~uySc|7w-PDe?u zVQj}Y#&nROkB2AS_tOjC#)3gzH1sZ*u@a{LvCP1-Aq|Yu^ z4{>-T;)QS{fX20sPtZ%2)ZF86+yaYSGCUKJqyXsAE{GSMt{u?)kVYxcFo{P`H`Osp z3q)v?FpIJxk;aP$an|~gv!{eIcxA?4@@nbMAmMxbjO<4q_G?9r50J-Ocm^KW@|U>y zFR;%}d*>U{!3AedsOHRGy3+!Qc+#jj%kYWX7>zSi7OXMiTTjyvFm6;eMni|_Bj$Ii zZf<06;(iw75v%P!5wh-tj22`(ZL>I_~6j)mfStpJZ$ zR8NNI)14=?7r#3w?HVuwV9%@H(wmPJKNTti(& zX5w_Uai_3@7rUY7sb*1ey-Bsb>t8Ws626JDVdi7X#Ze}z;ycfPmYwW`eAw@pqWUje zM7eJiF$j1@ylqGA3%L)HWH^doD`b;rs0%;dzJM=c$)>0ZaH(;2E35YmKD=i(Nq)t>{{0XClZk^KscV)lKG;5p zPXi`!IWM$chd5k#Vn*|sVL8#I$y3&A9h4it01g5gEX}SlIJ2|>oMMeRS#VSiqjJYt zY}O>*c7w=OS&B!Sj_F}Dko;cy5q^-s7_Iq`7X3Ilc@$D;F+Az6Wz@eYBFWh^4=Jr` z0k+6|6B%#N`D3e~1#bDYS?uR*0K^XjL8Sy>>@CaE(VXRz$-N>poQ!<-)LJ|pgGPi@ zKjWt?4j66r>SFPZRK7S5^NEWtXU&S4)Q`5aI9QfUq|6*ae8e%%QV$(vIK7<2Wz977 zOkNRK?hm#jHsTcT_3(Cp5jT*m%l`QUuO|wZFyMO4SJjlsG!uqn@;ggq?F${2f=}v{Faune*9rtmtToD$N z>SdH7Ci>&kTiSRsPJW6c%jw@_K^t)mYt4zl34%(4J3xjdGH(kI-IWl)WcHNp%o=ps_%{ZKSvd>HlLCFrX+w)q}t6@^E=9CHwfJ3YBj^P>B!n_2Yx2 za)5LGUD+w!M@#Z_&)vG=2T^4vYK9DadakYDYAFf&e-%cI>j2W%9;ylt6KE!fGp041 zTquxkl_H4$;c%8i)@Aflx~nB1lKHPCAo0!b@k3ClB#qy`CM2uJ{m#ld87s+^buy=K zn!=ih8Xm6duAu^jx@)Cs66yUjCeIl_45}NeEx3fxp79^WU+FRIkGSEA7}65~xx-c0 zdh-px=qv5&C_rDY72Z=Hh>;@P-{P;=pXqtgK0Jb)Q}F-xaH%k5PTCNBXiz6IS~ipJ z$_^M*&G)2y0Ko*%Iz~?YDxJBikwUCHH|GZu;FU224`>J@Ot^H_0wB#8{PSlDSRNFF+&M@F>hl8#H>i7zSk9W(1fM};-Xea9iHoLu$JzdJ;n{Hk z^TCcWxhmrTpzQrQ5HZqhbHZwlFl47hNT&p%g`eK8$GArRl#tx+quU&nV1JDous`{R zb6NsQWFj%L?iq0^YC8l#yS*^E^0_WL)Y<7ZT1}}~G3BgMS-6<{c?oNRjG@=p(};V`SJPvwU#qQ+yS&Kh}=I%OV>I ztc*NBFT1l9ChCfKoChWY%yh4rngfLv&AOG&$yy}Z^!zgc7eov-H77(q613wb$1y4M zSVIG32L~!RC5^bgY2T1Xjd)G!^E_Kky<#CsI#Vily6ShU{ei+NK@Ij7h2B%ovN6G| z&g-SJtbEG9>xv5*SoEpJLR~~U7*li~3p z+ZtxeZM>e)q2c`tE^h?J{v0{JK7XWU91b1;1hJpEiY}$o90^&0!?a{1X>`y`M0&mN z5Y{oBcQcO7rO3NG#m5#IV_=)VhceizQim}buUs|IFE~nn!_Rrs2(=?Jz1G|{kl_wJ zXEd^*FF*rsJyiu6QB|p>C$c~^yG&BQUB?h$Ot?yW&N7t@F`PMy00bxe@#WL$5QoHq6xJ2;qbb1R4j#BK9C6o7I}zg|^>|#PVmJ>z zzqIs1u7u}{x#?FgN?#j07AK!EZ<#q8F^xy^@QOl8+jBu>Lmjxx#Y$i85sp@;L-PBX z^-4y?wPT6eo>No@a?W)^RwZ;Cd*wwl^?U;LoU+q!nl3oW-~yr`(xP=UTJ%;R4mZhH zzs|YIXxT)$a##GlEa_ML{$J=!xd`IP7JO!Uf9Vfuz;~)~$9X@Te1;mV+xf`4ZOEM( z8agA#P*W7+y%wmJgGe6pm~OHc=-_!-#kEy*tjI@!0QQYfZ^}0HNt3EbeIw0_?k?%7 zUhxTOiE+H^6FIQrKol*BQ6_nLVo`+BI7v-|4jP*p`&XF{HlxNm)B0t+YCmHh%ml8G zfAI~8FlMISybPy-K|m410EN&kIPbXzZok8)D7U6zIrNW``-Brq z4&SW6jbja8=nRWv0y$55uuH4&7 z-UiR+I5UdY*Zn9$W_&4eNP>;Y1D+mjm6aaO&@5S*#GI@ojb)>u3Qh?Ux-%6qE~2C31Gd$x`hgOhdi6iH~$NlN2a>UEq#r!+{UnP$%jOF zizv1t=8EjXUhtE zbC4XE9vBDg&rM_V88ay9Et~?SDD1G_D3(Y03tWVTBS!+}kNe~Cjsg>cDUDNA^cbyv z*zK7;Xv^S%T3SqrFu)AeH4Lq}Q2hOcL|9&i$I~qmYBKozMjf20LM%V$_#dJ|%6Y%G zJw*d}p zS;h@Coe(}`8UO6!R%%PGaUi5Qi)TMiW+eeBPNOt#iYjiMpE&XI{^|&?a@j^3m>v~( zR($B|)ep!xg_8qGfn6(K#~u*e6?I`EU7$N$7pf@D7q)9d6_*8lTPQghn2;yVT(ooP9rz*CV2v`unldY- zx6jYLK9*9&1>#as~wSBXKn1r&zxYu8V>h(>;y?09R)*@2c#{bTY(=LyG6m~ zhEN}QDBH4=&PA;PDus+$yp9gmrs2Mo^SHCBLVHM0(V_Zrl*hoR?V~VzBS>-C*|&H+ zMSF0Obly0`M}D2MD~iz747%r>P5MI8gZM&Ia$TjhF>R-TUx;#Y_+99d*xPvi<6Y)L z@kp@hAV@I_74TzI7{=YB|K#{+&DS_ihP^qCu?U0>AF>G@nbtu#H- zZMlM$w=7=_Yt_VIt!A;cYHs@;my(A)_Wyf*%*Iz%^i)(tGrbbMW_YrkN)->>AayQ2 zl{;6K(F$2eG7{=|L=wd-NE9{QXm9DQ*j&8JCBZRwYcF`L!1?Lw|9TxLc@A}kM!;>U zvWgOD5hOoqzjK|UEB&ULS+Dq;xZuP21jCm)-cW6KVScGWWaR96$<&jC%>_0T)gH{w zvphsZNFV5py1JF&ZK64`80C85Rrr#4OL#R(WPTOBnbzu``(@m@(%Z94ua&EV3z_rK z+?jZ%Tj_lZJHpbinA`;_Y;dXFIzikIdYx%(-k)I8sLJD|z9-Mhj?E%j)O797#IZ46 zH~a;!oAC^7)H7%2w&~#78@`>7(@q9=7LR3FC|3;2L8LxaL zMYpW1r}}*AsTH__vXUe^S4ln}STjzg&_h-MeYDEVt(6;}ZFn>l_@ziL@JrQEnZip& zlQIs`n|7P8@bZ+`gx-pM@ej0pHL_BVI6*~y9vqWlE>h)D)SqUm{b?HM&mb9btS_QZ zeAW40)!Wghf7|Y1rLC-N;rQ~*0gp}aUDICh&D4u&IroCSYCG$TaO6ry=t6fAR>TmK z4sZFIw+$I=pT(Sts}Q;Ck{inf7#9kvN4eLWG>sp$y`U;p9TMD9=%IgSpGjxv^=8o= z_FDCwAbk<{!zUTsIn_pv&uqC~~ypk7S&{kc#@q*xlTI}E^c942EwcgDo z_6gnpFr2ne{RtQbPtZx|Y3sllX^OgWVMFm#$Y3dE(y6*=8dgYm`@?BVmsSlv&4%@P z|ERJI_hMLsGgnGmPWEi->p%8{?DD+DU_oplX@04ziP5hQ`_qC`N&gw0*k1h$Hmdbp zo=H8g3Vu4#{5CwC&x~te9pp=FxtE=FEq)WUYd;JJslDsf<{cpfyd4Tx#--e(z_uP#L=ihE`{m2hCg8l62*#7?(GDZ z`>7vIqy14$pKF}Y3AUh<^N=+WJ}YY15b2S82aU(OWeCRk+rQ9`s!vy^&sV#ydf z2$7LElP6KPyIE%X)osPEOZMdTpiDg&AQ3 zPxaQxEogdm*9hGD1qmh{pf{oo&l552#Ew;Le|ZlFNAeo6br*e9H@E zoiCKkazd)mp-<6b}4LM$%@npVF7UFbJp&4=ws zsOe=>GAgvJ5wn43xokV+J^wOr#jIgjOZaQIKOt=fy_IKDZ_7>wzbPy2=fF^LQ~9ti z6PZw^1>t7f1g)N%y8eSb{I_dCaJ7ojNHM0sY!fod@<8f$H51!0+WKNUayXF%h zbnDqUToXD=Wg0MDWZU+CE)JU>Btd5oC>ezEnGzY!&OGTiy#dJ+euHL~zX8M|KJ{Xe z;fe4DbnB%z;LVy|f}7x%_=%Em=q)Y)xny__zd;q22S)>`=^A~Du9#n#hYs0&=`932 zlQ$;ae73BU8DbC5`j5gEj*vgf#9RQx`4#p4!pv}ECA|sBIP<5P(|B>!lq1~ys7p6h zpr7T~WBfY>BLS<|>1rhuJXr81{~mcil2)rMD<0llBwy@Q6kguNThZ_ZXQR*A%gn9C zvlz)f(dMTM$vPw-Gw-=)<|E=S$W9zNm(_%m(Ws&V?ETxd(_*2pi4sX5(b{MYZF5%6 z$go^qz2Fr({k)pWO#h7YP1l^_gp)SG_0+ughv1{9uiHIRNzfe0-~MhR=?+ByQcn_U z(_!(LByp_@aD!Mg(SPjdRjufZT{J9%Qn1AW+L0jYp2MAx=fqCX z)+C=@{A||gp%TR*m56Gq4J)%{qpg;ds#w-7(4hLO%DTZ<#&0`&UK%QWs`$+1T5{!9 zU6QyIuq@%e`ovQ7BZ>$D8djuBTA2yngq5VV=Q|!LH5;j}-zMRo#U)Yzf1n~%uQNL= zhThT{reLFa(i7m0*|ddp<9j`ha27Egm!SKgxQe37lix3{URLP%-3oZ1RW%c0ja~nm zcr^^y2N{LE>V$6w0=~ZT=%fpU^EYSqBe}-=z~B-*DDe%MH>2kyjs30{%7I4V9euiZ zm+~1JO>vX3f0X<^K?h1-RuW0~m|Qp43N^A^a4I=qNS1(*LX0R+#9A_jD#4&sO`yt1 z3TK_9j1&-D8Mnp2Rq?49QMBw-jEI-_Y2jlp388m!9lGX<@?cw@CwZ}mmkY0_yImj+ zwu(*x5Aw?)QEssKJDR&A2P4)Yr@i2&y-d^2ndeyBNQh5SE|wS-M|Bfa4H$1+`pD=e zaHr!Y$w&qTEoUVmyXYku3Ts#_V4L{_G>_$C#|5^E`CHAD0oWux?G@H9EIUF0rzv;m z#6WoPZcF9107~X`e2t9p4|B`gX}elTmZ~xPi&%Zk6bg#3x4Z07Fw>?BV5%7 ze)(mJ1QMWi9H+QOW=dIsq9*AuZ3B#Bo{i|I0zOX~JW7KE1zS|FEuU;9AS37jZ;^pB zga>Oc5nvPi zVUvTjEfu#~>q#gv9K^H(%KuaJlM+eGsio%~IMq~u5|D-XF-~3=E$Fr;`3oy8YWq>> zc5MpBOye^;iVcpkF6e}g%4!FqifCYjSLY=!>p1^(*SVLR$u+Ho@MSWaQ3PDw4K;z{ z@l&a9RDa|9HlYF;wTY;2Q`WwpuFAVYJ2hn$dtNYTA6GMiE9es{4K*G8_;Oe;mewi0 zt&D3k3N-CX{E)o2vE_LIB*56_tZUEOWCeQ@}6?RjBu!}CjmVv%aLceCw68T3f!0Vr0U4giJnwdcRnOkdg6Mbe8M&) z{5=oJzbvdB{unLPtcOzzb<>5j7WjK>0VoF>U+!&f@s({!3+jJZQ~W8KYKvN%YPt(e z@%O2zXQIgga^vU^H9u>Qzqj_Xcfv#leMC3U z-ibd&8`H+#X^)&wl)%8P7Q!h& zYvDwYMDZ*UjA{lw^6*w={z!@6Y8Yf2b%7 literal 0 HcmV?d00001 diff --git a/GDT/1.2.html b/GDT/1.2.html new file mode 100644 index 000000000..db1bcf257 --- /dev/null +++ b/GDT/1.2.html @@ -0,0 +1,3140 @@ + + + + +GDT 1.2 netCDF conventions + + + + + + +

+

GDT netCDF conventions for climate data

+

Version 1.2

+

Jonathan Gregory (1), Bob Drach (2) and Simon +Tett (1) +
+(1) Hadley Centre, UK Met Office; +(2) PCMDI, LLNL

+

10th January 1999

+

+

1 Purposes

+This standard defines a set of conventions adopted in order to promote the +interchange and sharing of files created with the netCDF Application +Programmer Interface (API). The standard is based upon version 2.4 of +netCDF. Documentation of the netCDF API may be found in the ``NetCDF +Users' Guide'', Version 2.4, February 1996, available from +http://www.unidata.ucar.edu/packages/netcdf/ +or via anonymous ftp at +ftp.unidata.ucar.edu. +The standard is named ``GDT'' from the names of the authors. +

+This standard is intended for use with climate data, +and was designed with data generated by GCMs particularly in mind. +We recognise that there are limits to what a standard can practically cover; +we restrict ourselves to issues which we believe to be of common and +frequent concern in the design of climate metadata. +Although this is specifically a netCDF standard, +we feel that most of the ideas are of wider application. +Our main purpose is to propose a clear, adequate and flexible definition +of the metadata needed for climate data. +The metadata objects could be contained in file formats other than netCDF. +Interconversion of the metadata between files of different formats +will be facilitated if they are based on similar ideas. +

+This standard is mostly additional to the conventions sponsored by COARDS +(ftp://ftp.unidata.ucar.edu/pub/netcdf/Conventions/COARDS). +In addition, all Unidata recommendations are supported here +unless noted to the contrary. +Comments indicate the places where there are differences +between the standards. +Comments given in emphasised type +and CDL examples given in slanted typewriter type +are not part of the standard. +Note that examples typically show only the details relevant to the +point under discussion, +and hence may be incomplete with respect to the provisions +of the complete standard. +

+ Successful transmission of data +depends upon the receiver of the file having +software that will correctly interpret it. For this reason, a +strategy that is as conservative as possible in the use of attributes +and encoding techniques will best promote portability of data. +

+ This standard also refers to the udunits standard +supported by Unidata. +The udunits package is available via anonymous +ftp at ftp.unidata.ucar.edu. + See section 11 for details of how the package +is used by this convention to define units for physical quantities. +

+Useful comments and suggestions from Karl Taylor, +John Sheldon, Jan Polcher, Bryant McAvaney, Harvey Davies, +John Caron, Steve Hankin and contributors to the netCDF news group +have influenced the development of this standard. +We have made some changes to gain greater compatibility with +the NCAR CSM netCDF standard. +

+

2 Filename

+ NetCDF files should have the file name extension .nc. +

+ +

3 Data types

+

+ The netCDF data types +char, short, long, float, and double +are all acceptable. +All numeric types are signed. +The byte data type, which is functionally identical to char, +is not recommended because its signedness is ambiguous in netCDF. +The COARDS convention deprecates char, rather than byte. +

+ NetCDF does not support a character string type, +so these have to be represented as char arrays. +In this standard, we refer to them as type ``string''. +A string array must be implemented as a +two-dimensional character data variable, +serving as a vector of fixed-length strings, +the second dimension of its CDL declaration +(leading dimension in terms of Fortran) +being recorded as a dimension in the netCDF file. +

+ +

4 Attributes

+

+ This standard describes many attributes +(some mandatory, others optional), +but a file may also contain non-standard attributes. +Such attributes do not represent a violation of this standard. +Application programs should ignore attributes that they do not recognise +or which are irrelevant for their purposes. + Conventional attribute names should be used wherever applicable. +Non-standard names should be as meaningful as possible. +Before introducing an attribute, consideration should be given to +whether the information would be better represented as a variable. +In general, if a proposed attribute +requires ancillary data to describe it, +is multidimensional, requires any of the defined netCDF +dimensions to index its values, +or requires a significant amount of storage, +a variable should be used instead. + When this standard defines string attributes which make take +various prescribed values, the possible values are given in lower case. +However, applications programs should not be sensitive to case +in these attributes. + Several string attributes are defined by this standard to +contain ``blank-separated lists''. +Consecutive words in such a list are separated by one or more +adjacent spaces. The list may begin and end with any number of spaces. +See Appendix A for a list of attributes described by this standard. +

+ +

5 Global attributes

+

+ The Unidata-standard attribute Conventions +is recommended to reference this standard, containing the +string "GDT 1.2". +This standard is registered with Unidata under the name ``GDT'' +in the directory ftp://ftp.unidata.ucar.edu/pub/netcdf/Conventions +and is available from +http://www-pcmdi.llnl.gov/drach/GDT_convention.html and +http://www.met-office.gov.uk/sec5/CR_div/GDT_convention.html. +

+ The float attribute appendices is recommended to record +the version number of the appendices to this standard used by the +application which generated the file (see section 12). +This information could perhaps be recorded in the Conventions +attribute, but having a separate attribute +for it will allow applications to extract the information +without having to parse a string. + The string attribute quantity_table should be used to +record the URL of the quantity table (see section 12). +If this attribute is not supplied, it is assumed that Appendix D was used, +at the version specified by appendices. +

+ The string attribute comment may be used to record any +extra information about the file. +Additional attributes to describe the file may be included as required. +For example, GCM output might include an attribute +to name the model integration. +

+ Although not mandatory, +the Unidata-standard attribute history is recommended to +record the evolution of the data contained within a netCDF +file. Applications which process netCDF data can append their +information to the history attribute. +The global history attribute is assumed to apply to all +data variables; +individual data variables may have their own history attributes +supplying additional information (see section 12). +

+ Use of the string attributes institution and +production is recommended. +The attribute +institution specifies who produced or supplied the data. +We prefer this name to ``center'' or ``centre'' because the two +possible spellings could cause confusion. +The attribute +production indicates how the data was produced. +If it was model-generated, production should name the model +and its version, as specifically as could be useful. +If it is observational, production should characterise it +e.g. "surface observation" or "radiosonde". +The global institution and production attributes are taken to +apply to all data variables which do not have their own such attributes +(see section 12). +

+ The calendar attribute (see section 23) may +be recorded as a global attribute. +The global calendar +attribute is interpreted as a default for all time axes. +

+ +

6 Variable names

+

+ Variable names should begin with a letter +and be composed of letters, +digits, and underscores. + Case is significant in netCDF names, but it is recommended that +names should not be distinguished purely by case +i.e. if case is disregarded, no two names should be the same. + It is also recommended that variable names should be obviously +meaningful, if possible, as this renders the file more effectively +self-describing. +However, nothing in this convention relies on the use +of particular names for variables. +

+ +

7 Data variables

+

+ The netCDF variables which contain the +physical data are referred to as ``data variables'', +also referred to as ``primary variables'' by Unidata. +Apart from the general naming rules for variables (above, +section 6), +the names of data variables are not standardised by these +conventions (since files may in general contain multiple +data variables of the same physical quantity). +

+ +

8 Coordinate variables

+

+ A one-dimensional netCDF variable +associated with a dimension of one or more data variables +is called a ``coordinate variable''. +A coordinate variable whose dimension name is identical to +its own name is referred to as a ``main coordinate variable'' +in this standard, +when it is necessary to distinguish it +from other types of coordinate variable +(sections 17, 18, +19 and 20). +Apart from the general naming rules for variables (above, +section 6), +the names of coordinate variables are not standardised by these +conventions (since files may in general contain multiple +coordinate variables of the same orientation). + The values in a main coordinate variable must be +strictly monotonic +(all values are different and either increasing or decreasing) +because this assumption is frequently made by software. +

+ +

9 Axes and dimensionality of a data variable

+

+ A data variable may have any number of dimensions, including zero, +and the dimensions must all have different names. +COARDS strongly recommends limiting the number to four, +but we wish to allow greater flexibility. +The dimensions of the variable define the axes of the quantity it contains. +Dimensions other than those of space and time may be included. +Several examples can be found in this document. +Components of vector or tensor quantities could be contained +in a single data variable by giving the variable a dimension over components. +While there exist advantages for manipulating such a variable in memory, +we see no strong advantage in introducing this complexity into the netCDF +description, and do not recommend it. +Under certain circumstances, one may need more than one +dimension in a particular quantity +(see section 29 concerning multiple time axes). +For instance, a data variable containing a two-dimensional probability +density function might correlate the temperature at two different +vertical levels, +and hence would have temperature on both axes. +

+ If any or all of the +dimensions of a data variable have the interpretations +of ``date or time'' (T), ``height or depth'' (Z), +``latitude'' (Y), or ``longitude'' (X) then those +dimensions should appear in the relative order T, then Z, +then Y, then X in the CDL definition corresponding to the file. +In terms of Fortran, this means X is the first dimension +of the array. Non-spatiotemporal dimensions +should be placed to the left +of the spatiotemporal dimensions +i.e. as trailing dimensions in terms of Fortran. +

+The reason for this convention is that these kinds of axes may +have special meanings to particular applications. +For instance, an application might want to plot a longitude-latitude map, +or integrate vertically, or extract a timeseries. +In the COARDS standard, the indication given by the order of dimensions +and information in the attributes of the coordinate variables +have to be used together to identify the required axes. +For compatibility with COARDS, we uphold all these conventions, +but we also introduce a new attribute to make the identification +straightforward and unambiguous, as follows. +

+ If the last four dimensions +do not have the interpretations TZYX +(in CDL order, omitting from the left if there are fewer than four dimensions), +an axis attribute should be attached to the data variable. +In other cases it is optional, but recommended. +This attribute is a char array of size equal to the dimensionality of the +data variable, +having one element for each dimension (in CDL order), +indicating the interpretation of that dimension. +The permitted characters are T Z Y and X, +with the meanings given above, +and - as a placeholder for a dimension which has none of these meanings. +Each permitted letter may appear no more than once in the array. +If a data variable has more than one dimension which could be given a certain +interpretation, +the axis attribute will therefore clarify which should be chosen. +Note that if there are multiple time axes (section 29), +of which only one is not collapsed, +this single ``climatological time'' axis would normally be the +designated T-axis. +If the axis attribute is included, +the dimensions may be put in any order, +but this should be avoided if possible +because applications not able to use this attribute +may not process the data correctly. +

+

Axes for an ordinary time-mean longitude-latitude-height +variable: +

  dimensions:
+    lat=18;
+    lon=36;
+    pressure=15;
+    con_time=1;
+  variables:
+    float xwind(con_time,pressure,lat,lon); // order T Z Y X
+      xwind:axis="TZYX";
+    float lon(lon);
+    float lat(lat);
+    float pressure(pressure);
+    float con_time(con_time);
+See section 11 for the description of the units and quantities +of the variables and sections 14, 15, 16, +24 +for details of longitude, latitude, vertical and time axes. + +

+ The coordinates of points within the data variable are +the simple ordered tuples formed by associating values from the +coordinate variables (section 8). + If a particular axis does not have a coordinate variable, +the coordinate values are assumed to be equal to their indices along +the axis, numbering from 0. +

+ Dimensions may be of any size, including unity. +When a single value of some physical quantity applies to all the values +in a data variable, +the recommended means of attaching this information to the variable +is by use of a singleton dimension (a dimension of size unity) +with a one-element coordinate variable. +The advantage of this method is that all the attributes of a coordinate +variable (quantity, components, boundaries, etc.) can be used to describe +the single-valued quantity. +Singleton dimensions also result from contractions, described +in section 22. +

+

Longitude-latitude field of temperature +on a pressure level: This would use a +singleton pressure dimension to record the level, thus: +

  dimensions:
+    lon=96;
+    lat=72;
+    pressure=1; // single-valued coordinate variable
+  variables:
+    float temperature(pressure,lat,lon); // axes in order Z Y X
+      temperature:axis="ZYX";
+    float pressure(pressure);
+      pressure:long_name="pressure";
+      pressure:units="kPa";
+  data:
+    pressure=50.0; // Pressure level of 50 kPa = 500 mbar
+The units and long_name attributes +are described in section 12. + +

+

Surface air temperature: Surface meteorological +measurements are made at +a certain defined height e.g. 1.5 m, which can be shown thus: +

  variables:
+    float temperature(height,lat,lon);
+      temperature:axis="ZYX";
+      temperature:long_name="atmospheric temperature";
+      temperature:units="K";
+    float height(height);
+      height:long_name="height above the surface";
+      height:units="m";
+  data:
+    height=1.5;
+The explicit height should not be given if the surface of measurement +is included in the name of the quantity e.g. screen height. + +

+ +

10 Coordinate systems

+

+ A coordinate system for the Earth's surface which is rectilinear +but +based on a polar axis other than the normal geographical axis is referred +to as a ``rotated grid''. +To describe rotated grids, a two-element float attribute +north_pole is attached to the data variable, +specifying the (longitude,latitude) coordinates of the rotated north pole. +If the attribute is absent and relevant, +it is assumed to have the value (0.,90.) i.e. the geographical north pole. +

+ In some systems, +the axes covering the Earth's surface +do not define a rectilinear grid. +We do not wish necessarily to exclude non-rectilinear systems. +For the moment, this standard is undefined for these systems, +and we invite comments from potential users on the appropriate definition. +The COARDS standard excludes non-rectilinear systems. +In principle, any coordinate system can be handled, +albeit clumsily, by replacing the relevant two or more axes by +by a single axis which indexes the points, +and providing associated coordinate variables to specify the coordinates, +point by point (see section 18). +

+ +

11 Units

+

+ The udunits package includes a +file udunits.dat, which lists collections of unit names. The names +given in the most recent version of this file +and their plural forms will be regarded as acceptable +unit names for this standard, with a few modifications which will be +listed in Appendix C to this standard. +COARDS lists some modifications within the standard, +but we would prefer to put in place a means to allow future +modifications to be made easily. + Users of this standard should not define their own units, +because this would make their files less portable; +requests for new units should be directed to Unidata. +

+ The udunits package also defines a means for linear transformation +of units by a scale factor and an offset. +This convention is allowed when it is natural to express a unit in +such a form +e.g. density of sea-water in +kg m¯³ +in excess of 1000 +kg m¯³, +which can be specified to udunits as "kg m-3 @ 1000". +COARDS does not permit the use of this facility. +This facility should not be used as a means of data compression, +for which an alternative is provided +(see section 33). +

+ +

12 Physical quantity of a variable

+

+ These conventions standardise two attributes for specifying +the physical quantity of data and coordinate variables. +Both of them are strings and defined by Unidata conventions. +

+ The units attribute is formatted as per the +recommendations in the Unidata udunits package +(see section 11), +with extensions for time +(see section 26). +Case is significant in the units. +This attribute is mandatory unless the quantity is dimensionless +(a pure number), +in which case the units may be given as a pure number. +There are a few defined dimensionless units, +such as percent, but +there is no need for a wide variety of dimensionless units for quantities +like sea-ice concentration, cloud fraction, probability and so on; +this descriptive information is the long_name rather +than the units. +A scale factor and/or offset may be specified +quantity e.g. sea-ice concentration in tenths +may be given as units="0.1f". +A dimensionless quantity with no scaling or offset +may have units="1.0f" or units="unity". +

+ The long_name is an attribute +containing a descriptive name, +which should not specify the units. +Case is not significant in the long_name. +This attribute is optional for most variables, +but mandatory for coordinate variables +of longitude, latitude, vertical axes and time +(see sections 14, 15, 16 +and 24). +If the long_name attribute is present, +it must be a quantity chosen from a standard list, +optionally with additional information enclosed in parentheses () +in the case that sufficient detail cannot be given by +a standardised description. +

+ We refer to the list of +possible quantities as the ``quantity table''. +The quantity table defines the long_name and a permissible unit +for each quantity. +Any legal and physically equivalent unit would be an acceptable +units attribute. +There are two options for a choice of quantity table. +One option is to use Appendix D of this standard, +which will be made available on the web. +Each quantity in Appendix D will be labelled with the version of the +appendices at which it was introduced, +enabling an application to deduce the complete set of +quantities which was available to the application which generated the file. +The other option is to produce a list of all +the possible long_names, +giving for each one an acceptable unit and the long_name +of an equivalent quantity in Appendix D. +This list should be made available on the web, +and its URL recorded in the global quantity_table attribute. +The purpose of the quantity table, +in either form, +is to allow users of data from different sources to decide which +quantities are comparable. +

+

Quantity attributes: +

    float tempt(pressure,lat,lon);
+      tempt:long_name="atmospheric potential temperature "
+        "(after timestep)";
+      tempt:units="K";
+"atmospheric potential temperature" is the standard +quantity, and "after timestep" is additional information, +which a generic application can disregard. + +

+Whether two physical quantities are different or the same is often +not a question with a well-defined answer. +Certainly if they are the same, they must have the same unit, but +various quantities with the same unit may have to be distinguished +e.g. atmospheric potential temperature and soil temperature. +In practice, +the most specific description applicable should be used. +We intend to expand Appendix D on an ongoing basis +in response to requests by users of this standard, +since we cannot foresee all the possibilities, +and we will err on the side of expansion, rather than restriction, +when it is unclear whether a new quantity is needed. +

+ The subgrid attribute (see section 21) +can be regarded as a modifier of the long_name; +it applies only to data variables, not coordinate variables. +The long_name and subgrid attributes together define +the physical dimensions of the quantity +(through information given in Appendix B and the quantity table), +and the units must be consistent with this. +

+ A data variable may have a history attribute, +supplying information about the derivation of the quantity that is not +possible to include either as a standardised form or in parentheses in +the long_name and subgrid attributes. +This attribute should be used as a last resort. +The global history attribute, if present, also applies to all +data variables (section 5). +A data variable may also have institution and production +attributes, indicating how the data was originally obtained +(see section 5). +These attributes take precedence over the corresponding global attributes. +The attributes history, institution and production +must not be relied upon to distinguish between data variables in the file, +and generic applications may ignore them. +

+

Optional quantity information: +These optional attributes might be used thus to describe a gridded +observational precipitation climatology. +

    float precipitation(lat,lon);
+      precipitation:history="gridded using Thiessen polygon weighting";
+      precipitation:institution="Climatic Research Unit, "
+        "University of East Anglia, UK";
+      precipitation:long_name="rate of precipitation";
+      precipitation:production="surface station observation";
+      precipitation:units="mm day-1";
+This would be appropriate if the data variable was in a file with data from +other institutions or methods of production. + +

+ The optional modulo attribute of a variable, if present, +records a number which can be added or subtracted +without altering the validity or physical significance of +the quantity. +It should be given in the same units as the variable. +This is most likely to be useful for longitude coordinate axes +(section 14), +with a modulo of 360, +and climatological +axes of seasonal or diurnal phase (sections 26 +and 29). +

+ We note that the Unidata-standard FORTRAN_format attribute +may be useful for both coordinate and data variables. +

+ In addition, +other model-dependent attributes may included to define +the quantity of a variable. +The Hadley Centre model will give each data variable integer stash +and submodel attributes, for example, which are codes identifying +GCM diagnostic output quantities. +

+ Variables may contain quantities which are discrete rather than +continuous, +meaning they can take only certain prescribed values. +This is more likely for coordinate variables than data variables. +For instance, a data variable containing the results of +a Fourier or spherical harmonic analysis might have a dimension +for harmonic number. +Section 26 describes some time variables which are discrete. +

+ +

13 Topology of an axis

+

+ An axis with ``circular topology'' is one +which can be legitimately +transformed by shifting all the points one place along the axis, +moving the last point to the beginning, any number of times. +The main coordinate variable of an axis with circular topology is distinguished +by the presence of an attribute topology="circular". +A longitude axis which circles the whole globe is an example. +The value linear or the absence of this attribute +indicates an axis with ``linear topology''. +The topology is indicated only by the main coordinate variable, +but since it is the property of the axis it applies to any component, +associated or boundary +coordinate variables as well. +

+ When a circular axis is rotated, the main coordinate values must be +altered in order to remain monotonic. +Therefore the main coordinate variable of a circular axis +requires a modulo (section 12). +

+Note that the topology and modulo attributes convey +different information. +For instance, +a longitude coordinate variable limited to values in the eastern +hemisphere between the Greenwich meridian and the date-line (e.g. 0E, 25E, +120E, 130E, 180E) does not have circular topology. (This might be from a model +of a limited area of the world.) When +making a contour map of a field with such +a longitude axis, one can interpolate anywhere within +the eastern hemisphere to draw the +contours, but it is not legitimate to interpolate over the western hemisphere +and draw the rest of the world, which is simply missing. The +implication of circular topology would be that one could put any longitude +at all on the left-hand side of the map. However, +this coordinate variable does have a modulo (of 360, +as required in section 14), and the points +can be labelled +in any way which is equivalent under the modulo to the coordinates +in the file. +The coordinate values 0,25,120,130,180 are thus equivalent to +-360,-335,-240,-230,-180. + +

+ +

14 Longitude dimension

+

+ Coordinate variables representing +longitudes must always explicitly +include the units attribute; there is no default value. +The units attribute will be a +string formatted as per the recommendations in the Unidata udunits +package. +The recommended unit of longitude is degrees_east (eastward +positive). Also acceptable are degree_east, degree_E, and +degrees_E. The unit degrees_west (westward positive) is not +recommended because it implies a negative conversion factor from +degrees_east. + A longitude coordinate variable must have an attribute +long_name="longitude". + Such a variable is identifiable from its long_name +and its units. +The COARDS convention relies on the unit as the only way to +identify a longitude variable. This standard uses the long_name, but +requires the units to be specified as well for compatibility with +COARDS. +

+ Longitude axes should have the attribute modulo=360, +indicating that they may be interpreted modulo 360. +Thus, for example, +-180, 180, and 540 are all valid representations of the International Dateline +and 0 and 360 are both valid representations of the Prime Meridian. +COARDS assumes that longitudes may always be treated in this way. +Since we have introduced the modulo attribute, +we require that it should be specified to indicate this. +A global longitude axis should have the attribute topology="circular". +Note that the presence of a modulo attribute +does not mean that the axis necessarily has circular topology +(section 13); +a longitude axis covering only part of the globe cannot have its points +rotated. + The sequence of numerical longitude values stored in the netCDF +file must be monotonic in a non-modulo sense +for a main coordinate variable of longitude. +

+

Global longitude axis: +

    float lon(lon);
+      lon:long_name="longitude";
+      lon:modulo=360.0f;
+      lon:topology="circular";
+      lon:units="degrees_east";
+ +

+ +

15 Latitude dimension

+

+ Coordinate variables representing +latitudes must always explicitly +include the units attribute; there is no default value. +The units attribute will be a +string formatted as per the recommendations in the Unidata udunits +package. +The recommended unit of latitude is degrees_north. Also acceptable +are degree_north, degree_N, and degrees_N. + A latitude coordinate variable must have an attribute +long_name="latitude". + Such a variable is identifiable from its long_name +and its units. +The COARDS convention relies on the unit as the only way to +identify a latitude variable. This standard uses the long_name, but +requires the units to be specified as well for compatibility with +COARDS. +

+

Latitude axis: +

    float lat(lat);
+      lat:long_name="latitude";
+      lat:units="degrees_north";
+ +

+ +

16 Vertical (height or depth) dimension

+

+ Whereas the two horizontal dimensions are usually longitude +and latitude, whose direction is well defined, +a variety of quantities may be used for the vertical axis, +if there is one. + The axis to be regarded as the vertical axis +must have both a long_name attribute (section 12) +and a positive attribute, +with one of the allowed values up or down, +to indicate the sense of the direction of positive +since this information may be useful for +applications which display the data. +

+

Vertical pressure axis: +

  dimensions:
+    pressure=15;
+  variables:
+    float pressure(pressure);
+      pressure:long_name="pressure";
+      pressure:positive="down";
+      pressure:units="hPa";
+  data:
+    pressure=850, 700, 500, 300, 200, 150, 100, 50, 30, 20, 10;
+ +

+The COARDS standard requires the units of +the vertical axis to be selected from a defined list, +in order that this axis can be recognised by its units. +It gives special status to units of pressure, +for which the direction of positive is defined, +and makes the positive attribute mandatory +for vertical axes with other units. +

+We have adopted a different approach for a number of reasons. +Firstly, to require units for the vertical axis +means defining dimensionless units for any +dimensionless quantity one might wish to use for the coordinate variable. +This is inconsistent with the treatment of a data variable; +the standard does not require that dimensionless units be invented +for dimensionless physical quantities in data variables. +Secondly, the long_name attribute is more informative than the units. +Thirdly, the vertical dimension of a data variable can be identified +from the axis attribute or the +the order of dimensions +(see section 9), +allowing any application which expects such a +dimension to find it without any further help. +

+We are not entirely convinced that +the direction of positive should really be recorded as part of the +data structure. +It is mostly an issue for displaying the data, +and is to some extent a matter of personal preference. +If such special treatment is given to the vertical axis, +why is it not also recorded for other axes? +For instance, when latitude is shown on the horizontal axis of a plot, +is north on the left or the right? +This is the same kind of question, +but it strikes us as more a matter for a graphics application to consider. +Nonetheless, we have required the positive attribute for +compatibility with COARDS. +

+For example, if an oceanographic netCDF file encodes the depth of the +surface as 0 and the depth of 1000 m as 1000 then the axis would +use attributes as follows: units="m", +long_name="depth below the surface", positive="down". +If, on the other hand, the depth of 1000 m were represented as +-1000, +we would have long_name="height above the surface", +positive="up". +

+ +

17 Component variables

+

+ A continuous physical variable may require +more than one number to specify it at each point. +We refer to these as ``components''. +The values of the components +are recorded in variables +referred to as ``component variables''. +The variable to which the components belong is called the ``head'' +variable of the components. +The names of the component variables are recorded as +a blank-separated list in a component string attribute +of the head variable. +The dimensions of a component variable must be identical with those +of its head variable. +OGDT restricted components to coordinate variables, +but the concept has here been generalised, +since any quantity that is used as a coordinate variable might also +be needed as a data variable. +

+ When a coordinate variable has components, +this standard requires that a main coordinate variable should nonetheless +be supplied which represents a combination of the components that can be +used to order the points on the axis. +As usual, this main coordinate variable must be monotonic, +but the components do not need to be monotonic. + The definition of the main coordinate in terms of its +components may be given in parentheses in the component attribute. +This information is not standardised and generic applications cannot be +expected to make use of it. +

+

Hybrid vertical coordinate: +A vertical coordinate +eta = p/p0 + sigma +is used in some atmospheric GCMs. +Atmospheric model levels are specified in terms of +(p,sigma) +pairs, +where p is pressure, +p0 +is a constant and +sigma +is fraction +of surface pressure (which is variable). +The eta +value is a linear combination of the two, +which cannot be uniquely decomposed back into +(p,sigma). +We would record this coordinate variable thus: +

    float eta(eta); // main coordinate variable
+      eta:component="pressure sigma "
+        "(eta=pressure/p0+sigma; p0=100 kPa)";
+    float pressure(eta);
+    float sigma(eta);
+A generic application would treat the component and main coordinates +as independent information. +The extra knowledge required to relate them would reside in any +specific application which needed it. +Hybrid vertical coordinates are the only obvious application of +component variables, but the convention could be used for other +similar purposes if they arise. + +

+ +

18 Associated variables

+

+ An axis of a data variable, or two or more axes in combination, +may have alternative sets of coordinate values. +These alternative sets are recorded in variables +referred to as ``associated'' variables, +having their own units, long_name +and other appropriate attributes to describe them. +The names of the associated variables are recorded as a blank-separated list +in an associate string attribute either of the data variable +or of the main coordinate variable of the axis concerned. +If the association is with the data variable, it applies only for that +data variable, +but if it is with the main coordinate variable, +it applies for any data variable which uses that main coordinate variable. +Association with the main coordinate variable thus may be more convenient, +but is less flexible. +Association with the data variable is the only option when several axes are +involved, and when there is no main coordinate variable. +The examples illustrate these points. +

+The associate +attribute may alternatively and equivalently be named coordinates. +This possibility is included for compatibility with the CSM standard. +However, in the present standard it is deprecated because of possible confusion +with the normal definition of ``coordinate variable'' (section +8) and because the use of associated variables is wider +than just coordinate variables in the usual sense. +

+A variable may be associated with more than one data variable +or coordinate variable. +If an associated variable itself has an associate attribute, +variables named by this attribute are also regarded as being associated. +

+An associated variable must have dimensions which are all dimensions +of any data variable with which it is associated; +the associated variable can be regarded as a function +of the indices along these axes. +The values of an associated variable do not have to be monotonic. +

+A generic application is not required to make any +use of associated variables. +Associated variables are not indicated in the axis attribute of the +data variable (section 9). +However, to improve legibility of the CDL file, +it is recommended that when +variables named by the associate attribute +of a data variable have interpretations that would be +denoted by T Z Y or X in the axis attribute, +they are listed in that order, with any others appearing in front. +

+

Vertical axis: Many associated variables +will be one-dimensional, +giving alternative sets of values for a single axis. +One example is a vertical axis where one wishes to store both the +physical coordinate and the ordinal model level number: +

  dimensions:
+    lat=90;
+    sigma=19;
+  variables:
+    float xwind(sigma,lat); // 2D data variable
+      xwind:axis="ZY";
+    float lat(lat);
+      lat:long_name="latitude";
+      lat:units="degrees_north";
+    float sigma(sigma); // physical height coordinate
+      sigma:associate="model_level";
+      sigma:long_name="sigma";
+      sigma:positive="down";
+    int model_level(sigma); // model level number at each height
+      model_level:long_name="model level number";
+      model_level:positive="up";
+As the association of model_level is with sigma, +any data variable with a sigma-axis +will have the association with model_level; +it is not a property of xwind particularly. + +

+

Trajectory: The +value of a quantity along a one-dimensional +trajectory. +In such a case, we might have a coordinate variable containing +time of travel and +associated coordinate variables +giving the latitude and longitude of each point: +

  dimensions:
+    day=10; // 10 sample times along a trajectory
+  variables:
+    float hice(day); // sea-ice thickness measured as the floe drifts
+      hice:associate="lat lon";
+      hice:axis="T";
+      hice:units="m";
+    float day(day); // time since the beginning of the journey
+      day:long_name="time";
+      day:units="day";
+    float lon(day); // longitude at each time
+      lon:long_name="longitude";
+      lon:units="degrees_east";
+    float lat(day); // latitude at each time
+      lat:long_name="latitude";
+      lat:units="degrees_north";
+The main coordinate variable (day) must be +monotonic, but the associated coordinates are not necessarily. +An important application +rather similar to this one +is described in section 19. +Note that lat and lon cannot be +indicated as the X and Y coordinates in the axis +attribute (section 9). +This is reasonable because, even though they have the interpretations +of latitude and longitude, they are not independent dimensions in +the sense normally expected by an application which might look for +such axes. +

+Since lon and lat are associated with the data variable +hice, other variables with a day-axis will not share these +associations. +If they are wanted, they would have to be indicated on those data +variables as well. +This approach allows the possibility that the same day +variable might occur in combination with various different sets of +associated coordinate variables. +For instance, there might be more than one trajectory, with the +same day coordinates but different longitude-latitude positions. + +

+

Transformed coordinates: Associated variables +of more than one dimension can be used to +describe alternative coordinate systems. +For example, vertical profiles of atmospheric humidity might +be available on a regular longitude-latitude grid, +but we might also wish to give the national grid coordinates of +each point. +The national grid x- and y-coordinates are each functions of +both latitude and longitude; +the x-coordinate does not correspond specifically to longitude, +nor the y- to latitude. +The appropriate representation is therefore: +

  dimensions:
+    lon=10;
+    lat=20;
+    pressure=15;
+  variables:
+    float humidity(pressure,lat,lon);
+      humidity:associate="y x";
+    float pressure(pressure);
+      pressure:long_name="pressure";
+      pressure:positive="down";
+      pressure:units="kPa";
+    float lon(lon); // 1D main coordinate variable
+      lon:long_name="longitude";
+      lon:modulo=360.0f;
+      lon:units="degrees_east";
+    float lat(lat);
+      lat:long_name="latitude";
+      lat:units="degrees_north";
+    float x(lat,lon); // 2D associated coordinate variable
+      x:long_name="UK national grid eastings";
+    float y(lat,lon);
+      y:long_name="UK national grid northings";
+This tells us that humidity[*][10][5] is the vertical profile +of humidity at the point with latitude lat[10] and +longitude lon[5], +which is at national grid x-coordinate x[10][5] and y-coordinate +y[10][5]. +Because the associated variables are multidimensional, they do not +correspond one-to-one with the axes, +so the association must be with the data variable rather than the +main coordinate variables. + +

+

No main coordinate variables: +A related situation is when the 2D grid is staggered or transformed +geometrically in some way (other than rotation--see section 10), +so that it is not possible or easy to give +one-dimensional coordinate variables for the axes. +In this case, there would be no main coordinate variables, +and the default of plain indexes would apply. +The physical coordinates are functions of the 2D gridpoint indices, +and would be given in associated variables of the data variable +just as above: +

  dimensions:
+    x=90;
+    y=45;
+  variables:
+    float orog(y,x); // 2D variable on a horizontal grid
+      orog:associate="lat lon";
+      orog:axis="--";
+      orog:long_name="height of the surface above sea-level";
+      orog:units="m";
+    float lon(y,x); // 2D coordinate variable on the same grid
+      lon:long_name="longitude";
+      lon:modulo=360.0f;
+      lon:units="degrees_east";
+    float lat(y,x);
+      lat:long_name="latitude";
+      lat:units="degrees_north";
+The lat and lon variables are not indicated as +the X and Y coordinates in the axis variable +(section 9). +An application which referred to latitude and longitude coordinates +would not generally expect them to be two-dimensional. +If it can handle this situation, it should identify these axes by +their long_name and units. + +

+

3D associated coordinates: +These could be used if one wished to describe +a field with alternative 3D coordinate systems, +for instance on both a regular Cartesian grid, +and in cylindrical or spherical coordinates. +The values of the alternative coordinates would be given on the Cartesian +grid. In the spherical case, for instance +

    float temperature(z,y,x); // 3D variable on a Cartesian grid
+      temperature:associate="radius theta phi";
+    float radius(z,y,x);
+    float theta(z,y,x);
+    float phi(z,y,x);
+ +

+ A particular technical application of a one-dimensional +associated coordinate is to deal with the limitation of netCDF to a +single unlimited dimension. +If several data variables have unlimited axes of different lengths or +physical significance, +they can all share a nominal unlimited dimension, +and each have associated variables specifying the meaning of the axis. +

+

More than one unlimited axis: +Consider a file which contains data variables with unlimited +axes measuring elapsed time with different sampling frequency, +and hence of different lengths. +

  dimensions:
+    time_counter=UNLIMITED;
+  variables:
+    float sw(time_counter); // sampled every 3 hours
+      sw:associate="time_3h";
+      sw:axis="T";
+      sw:long_name="vertical component of "
+        "shortwave radiative flux density";
+      sw:units="W m-2";
+    float latent(time_counter); // sampled every 30 minutes
+      latent:associate="time_30min";
+      latent:axis="T";
+      latent:long_name="latent heat flux density";
+      latent:units="W m-2";
+    float time_3h(time_counter)";
+      time_3h:long_name="elapsed time";
+      time_3h:units="h";
+    float time_30min(time_counter);
+      time_30min:long_name="elapsed time";
+      time_30min:units="min";
+ +

+ +

19 Bundles

+

+ If several data arrays containing the same physical quantity +have one or more identical axes, +but are distinguished by the values of other singleton coordinate variables, +it may be convenient to store them in the same data variable. +The common axes of the separate arrays become axes of the combined variable. +One or more additional axes are introduced to ``bundle up'' the separate +arrays. +Such an axis does not correspond to a continuous physical coordinate. +It acts simply as an index of the bundled-up arrays. +

+ The singleton values of the separate arrays are recorded in +associated coordinate variables for the bundling dimension. +They should not be interpreted as continuous coordinates. +

+

Timeseries: +The Hadley Centre GCM +can generate timeseries of the values of quantities +at individual points. +Typically, timeseries from many different points are produced of the same +quantity at the same sampling times. +It is natural to contain this information +in a data variable with two dimensions. +One dimension is the common time axis, specifying the sampling times, +which are the same for all the points sampled. +The other dimension is not a continuous physical coordinate; +it is simply being used to ``bundle up'' the timeseries, +the points being +irregularly scattered in a space of two or more dimensions. +Thus: +

  dimensions:
+    points=15; // measurement locations
+    times=20; // sampling times
+  variables:
+    float snowdepth(times,points);
+      snowdepth:associate="sitename lat lon";
+      snowdepth:axis="T-";
+    float lon(points); // longitude of sites
+      lon:long_name="longitude";
+      lon:modulo=360.0f;
+      lon:units="degrees_east";
+    float lat(points); // latitude of sites
+      lat:long_name="latitude";
+      lat:units="degrees_north";
+    char sitename(points,StringMaxLength); // string array of sitenames
+    double times(times); // times of measurement
+See section 23 concerning the time coordinates. +This same form could be used for observed timeseries from stations. +The bundling axis (points) is simply an index. +The long_name and units of the associated coordinates +identify their meanings. + +

+

Vertical profiles: +A similar application is that of vertical +profiles at sets of points; +for example, scattered vertical temperature profiles through the ocean, +or data from various radiosonde stations. +

  dimensions:
+    station=10; // measurement locations
+    pressure=11; // pressure levels
+  variables:
+    float humidity(pressure,station);
+      humidity:associate="lat lon";
+      humidity:axis="Z-";
+    int station(station); // station numbers
+    float lon(station); // longitude of stations
+    float lat(station); // latitude of stations
+    float pressure(pressure)
+ +

+

Several parcel trajectories: +Consider a set of Lagrangian parcel trajectories (for instance, ocean +drifters). +Various parameters are evaluated at fixed times from the start of each +trajectory. +The trajectories are identified by their place of origin, +and the position is a function of time and trajectory identity. +The position information is therefore stored in multidimensional +associated coordinate variables. +

  dimensions:
+    parcel=15; // number of trajectories
+    times=20;
+    max_len_parcel_name=64; // max length of trajectory name
+  variables:
+    float temperature(parcel,times);
+      temperature:associate="parcel_name lat lon";
+      temperature:axis="-T";
+    float salinity(parcel,times);
+      salinity:associate="parcel_name lat lon";
+      salinity:axis="-T";
+    float times(times);
+      times:units="days";
+    char parcel_name(parcel,max_len_parcel_name);
+    float lon(parcel,times);
+    float lat(parcel,times);
+The associations have to be made on the data variables in this case, +because parcel_name does not have a main coordinate variable, +while lon and lat are multidimensional. + +

+This section raises the question of how best to store a single timeseries, +or a single vertical profile. +Following the scheme of this section, +it could be contained in a two-dimensional data variable with the bundling +axis being of size unity. +The associated information such as latitude or longitude would then be +stored in singleton coordinate variables, +all associated with the same dimension. +Alternatively, these values could be recorded as +separate singleton dimensions (following section 9). +We have no recommendation for this. +Either scheme could be appropriate; +which is more natural perhaps depends on how the data was extracted from +the continuous axes. +

+ +

20 Boundary variables

+

+ Along a dimension, +the values might relate to points (at the coordinate values) +or to contiguous or non-contiguous cells. +The boundaries of the cells should be defined +as well as the point coordinate values. +The convention is to define an additional two-dimensional +``boundary variable'' with +a right-hand dimension (leading dimension in Fortran terms) of size two. +The values for which this dimension has index 0 +(numbering from 0 i.e. in C notation) supply the boundaries +with the smaller main coordinate values, +and those with index 1 the large values, +where ``smaller'' and ``larger'' refer simply to numerical comparison, +not to a physical direction. +Supplying upper and lower boundaries separately allows for the possibility +that the cells might not be contiguous; they might even overlap. +If a lower boundary value is equal to the valid_min for the coordinate +variable (section 30), the cell has no lower boundary. +If an upper boundary value is equal to the valid_max, +the cell has no upper boundary. +The name of the boundary variable is recorded in a string attribute +bounds of the main coordinate variable. +We recommend that the it should be named +by the coordinate dimension with the prefix bounds_. +

+

Boundaries for a one-dimensional latitude coordinate variable: +

    float lat(lat);
+      lat:bounds="bounds_lat";
+    float bounds_lat(lat,2);
+In C notation, +lat[0] gives the coordinate of the first point, +bounds_lat[0][0] its lower boundary, bounds_lat[0][1] +its upper boundary. +In Fortran notation, the declarations are lat(lat) +and bounds_lat(2,lat), +and the relevant elements are lat(1), bounds_lat(1,1), +bounds_lat(2,1). + +

+

Albedo as a function of wavelength and snow cover: +Characteristic values of albedo are given for various wavelength +bands, dependent also on snowdepth. +

  dimensions:
+    lambda=4; // number of shortwave frequency bands
+    snowdepth=10; // number of snowdepth categories
+  variables:
+    float albedo(lambda,snowdepth); // no units for albedo
+      albedo:axis="--";
+      albedo:long_name="surface albedo";
+    float lambda(lambda);
+      lambda:bounds="bounds_lambda";
+      lambda:long_name="wavelength";
+      lambda:units="nm";
+    float bounds_lambda(lambda,2);
+    float snowdepth(snowdepth);
+      snowdepth:bounds="bounds_snowdepth";
+      snowdepth:long_name="mass per unit area of lying snow";
+      snowdepth:units="kg m-2";
+      snowdepth:valid_max=1e9;
+    float bounds_snowdepth(snowdepth,2);
+  data:
+    lambda=250, 385, 570, 795;
+    bounds_lambda=175,320, 320,450, 450,690, 690,900;
+    snowdepth=0.05, 0.15, 0.35, 0.75, 1.25, 1.75, ..., 450.0, 1000.0;
+    bounds_snowdepth=0.0,0.1, 0.1,0.2, 0.2,0.5, 0.5,1.0,
+                     1.0,1.5, 1.5,2.0, ..., 400.0,500.0, 500.0,1e9;
+A first index of 0, for instance, gives albedo values for +the wavelength range 175-320 nm. +The deepest snowdepth class has no upper bound; any value above 500 +falls into this class. + +

+ In some cases such as the previous example the boundary +coordinates are well defined, +but the gridpoint coordinates are arbitrary. +In such circumstances, this standard recommends the mid-point of the boundaries +be used as the gridpoint. +Two advantages of this choice are: firstly, comparison of a gridpoint with a +boundary will always decide which cell the point belongs in; +secondly, it will probably be an appropriate choice for plotting and calculations +involving gridpoints such as differentiation. +However, as shown by the last snowdepth cell of the above example, +which is unbounded upwards, the mid-point is not always a sensible choice. +

+

Probability density function of precipitation amounts: +

  dimensions:
+    ppn=10;
+  variables:
+    float pdf(ppn,lat,lon);
+      pdf:axis="-YX";
+      pdf:long_name="probability density of "
+        "depth of water-equivalent precipitation";
+      pdf:units="mm-1";
+    float ppn(ppn);
+      ppn:units="mm";
+      ppn:long_name="depth of water-equivalent precipitation";
+      ppn:bounds="bounds_ppn";
+    float bounds_ppn(ppn,2);
+  data:
+    bounds_ppn=0.0,0.1, 0.1,0.2, 0.2,0.35, 0.35,0.5, 0.5,1.0, ...;
+pdf[3][10][12] gives the probability density of precipitation +amounts between 0.35 and 0.5 mm falling at the location +lat[10] lon[12]. + +

+ Boundary variables +are recommended if the main coordinate values +are not evenly spaced, +or if the dimension has a size of unity. +If the coordinates are evenly spaced, +and boundaries are not specified, +generic applications may assume that the main coordinates +lie at the centres of their cells. +Boundary variables may be supplied for component and associated coordinate +variables as well as for main coordinate variables. +Their elements are ordered so as to correspond +to the corresponding main boundary variables. +Hence they will not necessarily be monotonic, +and index 0 and 1 (numbering from 0) of the dimension of size two +will not necessarily contain the smaller and larger values. +

+

Boundary values for a hybrid vertical coordinate: +The atmospheric column is here divided into three cells in the vertical; +from the surface to +sigma=0.7; +from there to 20 kPa, and finally to the top of the atmosphere, +using the hybrid vertical coordinate introduced in an example +in section 17. +

  dimensions:
+    eta=3;
+  variables:
+    float(eta);
+      eta:long_name="pressure-sigma hybrid";
+      eta:component="pressure sigma";
+      eta:bounds="bounds_eta";
+      eta:positive="down";
+    float bounds_eta(eta,2);
+    float pressure(eta);
+      pressure:units="kPa";
+      pressure:long_name="pressure";
+      pressure:bounds="bounds_pressure";
+    float bounds_pressure(eta,2);
+    float sigma(eta);
+      sigma:long_name="sigma";
+      sigma:bounds="bounds_sigma";
+    float bounds_sigma(eta,2);
+  data:
+    eta=0.75, 0.45, 0.05;
+    bounds_eta=0.7,1.0, 0.3,0.7, 0.0,0.3;
+    pressure=0.0, 10.0, 5.0;  // does not need to be monotonic
+    bounds_pressure=0.0,0.0, 20.0,0.0, 0.0,20.0; // note order
+    sigma=0.75, 0.35, 0.0;
+    bounds_sigma=0.7,1.0, 0.1,0.7, 0.0,0.1;
+bounds_pressure[1][0] exceeds bounds_pressure[1][1] +because they are ordered to +correspond to bounds_eta. + +

+ Boundary variables may be given for associated multidimensional +coordinate variables (section 18). +Each dimension of the main variable requires an extra dimension of size 2 +in the boundary variable. +These extra dimensions are placed on the right +(left in Fortran terms) of the coordinate dimensions, +and in the same order as the coordinate dimensions. +

+

Boundaries for a two-dimensional latitude coordinate variable: +

    float lat(y,x);
+      lat:bounds="bounds_lat";
+    float bounds_lat(y,x,2,2);
+so bounds[4][5][0][0] contains the latitude of +the lower left (smaller x and y) +corner of gridbox [4][5], bounds[4][5][0][1] the lower right corner, +bounds[4][5][1][0] upper left and bounds[4][5][1][1] upper right. +In Fortran, the indices of the boxes would be (1,1,6,5), (2,1,6,5), +(1,2,6,5), (2,2,6,5) respectively. + +

+ +

21 Representation of subgrid variation

+

+ Since a data variable usually +represents a physical quantity which varies +continuously along the axes, +in reality there will generally be variation of the quantity between +adjacent gridpoints. +The data variable can give only one value for each cell, +despite this subgrid variation. +For many purposes, this can be taken as a ``representative'' value, +and it is not necessary to define precisely how it relates to +the subgrid variation. +

+ To be explicit about how each data value reflects subgrid variation +along a particular axis, +use the subgrid attribute of the data variable. +The most important application of this attribute is to +contracted or collapsed axes, described in section 22. +This is a string attribute comprising a list of blank-separated words. +In this list, "name: method" +indicates that subgrid variation along the axis with the dimension whose +``name'' is given +is represented by the specified ``method''. +The method, which may be several words, should be one of the +permitted values detailed in Appendix B, which include +mean, maximum, minimum, mid-range, +standard deviation, variance, mode, +median, cell, point. +Case and punctuation are not significant in the method. +Like Appendix D, Appendix B will be expanded on request by +users of this standard. +Some methods imply a change of units of the data variable, +and this also is specified by Appendix B. +In the above list, this is true for variance. +The method point indicates that the data values apply exactly +at the coordinate values, and do not at all represent the variation between +adjacent gridpoints along the axis concerned. +The method cell indicates that each value +should be regarded as a property of the whole cell +along the axis concerned e.g. a sum or integral. +The method can be differently specified for the various dimensions. +It must be remembered that the method applies only to the axis indicated. +If a precipitation value in a longitude-latitude gridbox is given the +method maximum for these axes, for instance, it means that it is +the maximum within these spatial cells, and does not imply that +it is also the maximum in time. +

+ The absence of any specification means that generic applications may +regard the data values as representative in whatever way suits their needs. +For quantities calculated at gridpoints by numerical models, +this kind of vagueness is unavoidable. +If a model provides a longitude-latitude field of temperature at gridpoints, +an application used to draw a contour plot of the field will generally +assume that the temperatures apply at points, and will use some interpolation +scheme to compute values between them. +An application which calculates the mean of the field, however, will +probably assume that the temperatures are gridbox means, +and average them by weighting each with its area. +Both of these approaches are valid. +The finite-difference scheme by definition does not have any information +about subgrid variation, +and may itself treat the values in both ways; +it might calculate gradients between them, regarding them as points, +or enforce conservation properties, regarding them as means. +It would be unusual to regard the values as extrema, however, unless +this was explicitly indicated. +

+ Except when the data are point values, +it may also be unclear what the coordinates of the gridpoints should be. +For example, what time coordinate should be assigned to a time mean value? +In such circumstances, if the boundaries of the cell are well defined, +this standard recommends that the gridpoints should be +defined as the mid-point between the boundaries +(see section 20). +

+

Subgrid time variation in timeseries: +Consider 12-hourly timeseries of pressure, temperature +and precipitation from a number of stations, +where pressure is measured instantaneously, temperature extremes over +the preceding period are recorded by maximum and minimum thermometers, +and precipitation is accumulated in a rain gauge. +For a period of 48 hours from 6 a.m. on 19th April 1998, +the data is structured as follows: +

  dimensions:
+    instanttime=5; // 5 instantaneous measurements at 12-hour intervals
+    periodtime=4; // 4 intervening 12-hour periods
+    station=10;
+  variables:
+    float pressure(station,instanttime);
+      pressure:axis="-T";
+      pressure:long_name="pressure";
+      pressure:subgrid="instanttime: point";
+      pressure:units="kPa";
+    float maxtemp(station,periodtime);
+      maxtemp:axis="-T";
+      maxtemp:long_name="temperature";
+      maxtemp:subgrid="periodtime: maximum";
+      maxtemp:units="K";
+    float ppn(station,periodtime);
+      ppn:axis="-T";
+      ppn:long_name="depth of water-equivalent precipitation";
+      ppn:subgrid="periodtime: cell";
+      ppn:units="mm";
+    double instanttime(instanttime);
+      instanttime:long_name="time";
+      instanttime:units="h since 1998-19-4 6:0:0";
+    double periodtime(periodtime);
+      periodtime:bounds="bounds_periodtime";
+      periodtime:long_name="time";
+      periodtime:units="h since 1998-19-4 6:0:0";
+    double bounds_periodtime(periodtime,2);
+  data:
+    instanttime=0., 12., 24., 36., 48.;
+    periodtime=6., 18., 30., 42.;
+    bounds_periodtime=0.,12., 12.,24., 24.,36., 36.,48.;
+It is not appropriate give a subgrid method for the station axis, +since this is a bundling axis (section 19) and not +a continuous physical coordinate. +The instantaneous and period measurements have to have different time +axes both because of their different dimension, +and because they do not coincide. +If the pressure measurements were made at times half-way between the +others (noon and midnight), the time axes could be shared. +Since the precipitation is given as an amount, it is a sum over +the interval of time by definition. +It could instead have been expressed as a rate in mm h-1, for +instance, in which case its subgrid method would be mean +rather than cell. + +

+

Thickness (geopotential difference): The +``thickness'' is the difference +in geopotential height between two pressure surfaces in the +atmosphere. This quantity is by definition one which relates +to the whole extent of its cell in the vertical dimension. +

  variables:
+    float thickness(pressure,lat,lon);
+      thickness:long_name="thickness";
+      thickness:subgrid="pressure: cell";
+      thickness:units="m2 s-2";
+    float pressure(pressure);
+      pressure:bounds="bounds_pressure";
+      pressure:long_name="pressure";
+      pressure:units="hPa";
+    float bounds_pressure(pressure,2);
+Here, bounds_pressure[0][0] and bounds_pressure[0][1] +will be the upper and lower pressure bounds of the thickness field +thickness[0][*][*]. + +

+ If more than one subgrid method is to be indicated, +they should be arranged in the order they were applied. +The left-most operation is assumed to have been applied first. +Suppose a quantity varies in both longitude and time +(dimensions lon and time) within each +gridbox. +Values which represent the time-average of the zonal maximum +are labelled subgrid="lon: maximum time: mean", +i.e. find the largest value at each instant of time over all longitudes, +then average these maxima over time; +values of the zonal maximum of time-averages are +labelled subgrid="time: mean lon: maximum". +If the methods could have been applied in any order without affecting +the outcome, they may be put in any order in the subgrid attribute. +

+ If a data value is representative of variation over a +combination of axes, +a single method should be prefixed by the names of all the dimensions involved, +whose order is immaterial. +Dimensions should be grouped in this way only if there is an essential difference +from treating them individually. +For instance, the subgrid standard deviation of +topographic height within a longitude-latitude gridbox +would have subgrid="lat: lon: standard deviation". +This is not the same as +subgrid="lon: standard deviation lat: standard deviation", +which would mean finding the standard deviation along each parallel of +latitude within the zonal extent of the gridbox, +and then the standard deviation of these values over latitude. +

+ To indicate more precisely how the subgrid method was +applied, +extra information may be included in parentheses () after the +identification of the method. +This information is not standardised and may be ignored by a generic +application. +A mean over latitude, for instance, may be area-weighted. +This could be indicated as "lat: mean (area-weighted)". +

+ The subgrid attribute cannot be used +to show how a value reflects variation over a coordinate which does +not have a dimension in the data variable. +This should be done in the long_name instead. +It is generally more informative and precise to introduce a singleton dimension +specifically for this purpose, however. +For example, we could describe a quantity in its long_name as +being simply a temporal variance, +but it would be more informative to record it as a subgrid +method, by giving the variable +a singleton time dimension, which could also +be used to should the range of times +it covers and the time-interval of the data from which the variance was +calculated. See also section 22. +

+ +

22 Contracted dimensions

+

+ A contracted axis is one which is formed by aggregating +the values of an axis with a larger dimension into a smaller number of groups. +In the commonest case, the dimension is collapsed completely to +a singleton dimension (i.e. a size of unity, section 9), +where all data points share the entire collapsed axis. + The collapsed dimension +indicates the relationship of the data variable which is +being described to another variable of higher dimensionality. + The boundaries of the cells along a contracted axis +will be the outside boundaries of the groups of cells along the +uncontracted axis, +or the outside coordinates if boundaries were not given. +The main coordinate values of a contracted axis will be values representative +of the coordinate ranges spanned by the groups. + A collapsed dimension has a single representative main +coordinate value and boundary coordinate values supplying the +complete range of the uncollapsed axis. + These boundaries will be +the extreme boundary coordinate values of the uncollapsed axis, +or the extreme main coordinate values if boundaries were not supplied. +A very important application of collapsed axes is to indicate +climatological time. +This is discussed in section 29. +

+ The subgrid attribute +(section 21) of the data variable +with contracted axes +can be used to indicate how the data values +of the variable with uncontracted axes +were aggregated to reduce the dimensions. +The new subgrid information will be appended to the existing attribute, +if any, +indicating the name of the newly contracted dimension. +Any existing references to the uncontracted dimension in the +subgrid attribute should be modified to refer to the contracted +dimension, since the uncontracted dimension will no longer be a dimension +of the data variable. +

+As explained in section 21, this attribute will indicate +that the data value is the mean, maximum, minimum, etc. +The allowed subgrid ``methods'' +are listed in Appendix B, which will be expanded as need arises. +As foreseen at the moment, the idea is limited to operations which give +a single value representative of each contracted group of values, +without reference to any external constants. +For example, the number which exceeds 20% of the values in the group, +or equivalently the 20th percentile, is a single number +representing the group, but the +procedure of finding it is not treated as a subgrid method +because it requires the constant 0.2 to define it. +Instead, the relationship of this new variable to the old should be shown +by changing its long_name to indicate that it is a percentile value, +and giving it a new singleton percentage axis with value 20, +or cumulative probability with value 0.2. +This kind of transformation is analogous to reducing a variable on three +spatial dimensions (say) to two by extracting its values on a specified +surface. +The contraction or collapse is a special case, because, +in general, the percentile axis need not have a size of unity; +it might be a new multi-valued axis (in cumulative probability) +replacing the old one (in some spatial dimension, for instance). +This is like regridding a vertical axis of height onto pressure. +Having said all this, however, we note that median is in fact a named +instance of this operation--extraction of the 50th percentile--but we +allow it on the grounds that it is a common method +for choosing a single representative value. +

+Singleton axes are not necessarily the result of collapsing an axis. +In section 9, we recommend singleton axes as the means +of attaching characteristic single physical values to a data variable, +for instance the height or pressure of the surface on which a variable +is supplied. +If no subgrid method is specified, the application knows only +that the single value characterises the data in some way. +All information in the subgrid attribute is entirely optional. +For instance, a time-mean quantity should generally have a singleton time +dimension +to indicate the range of times to which it applies, but +it is not mandatory to indicate in the subgrid attribute that it is +a mean over time. +

+ On the coordinate variable of a contracted axis, +the optional old_interval attribute +specifies the typical spacing between +two adjacent coordinates of the uncontracted axis, where +``typical'' is not well defined. +The old_interval attribute should be given in the same units as the +coordinates. +Further information may be given by the +the optional old_spacing attribute, +which may have value uniform, indicating that the coordinates were +evenly spaced with the old_interval specified (if any) and the cells +contiguous, +or variable, if they were not evenly spaced but still contiguous, +or disjoint, which means there may have been gaps between them. +The coordinates of the uncontracted axis may be explicitly recorded +in separate variables; +if so, the main uncontracted coordinate variable +should be named by the attribute expand of the +main contracted coordinate variable. +

+

Area-averaging a longitude-latitude field to one of +lower resolution: The original resolution was 1 degree, and the field +has been averaged into 10-degree boxes. +

  dimensions:
+    con_lat=18; // contracted dimension
+    con_lon=36;
+    lat=180; // original uncontracted dimension
+    lon=360;
+  variables:
+    float sst(con_lat,con_lon);
+      sst:axis="YX";
+      sst:long_name="sea surface temperature";
+      sst:subgrid="con_lat: mean con_lon: mean";
+      sst:units="degC";
+    float con_lat(con_lat); // contracted latitude axis
+      con_lat:bounds="bounds_con_lat";
+      con_lat:expand="lat";
+      con_lat:old_interval=1.0f; // original resolution in latitude
+      con_lat:long_name="latitude";
+      con_lat:units="degree_north";
+    float bounds_con_lat(con_lat,2);
+    float lat(lat); // original uncontracted latitude axis
+      lat:bounds="bounds_lat";
+    float bounds_lat(lat,2);
+  data:
+    con_lat=-85, -75, -65, ...;
+    bounds_con_lat=-90.,-80., -80.,-70., -70.,-60., ..., 80.,90.;
+    lat=-89.5, -88.5, -87.5, ...;
+    bounds_lat=-90,-89, -89,-88, -88,-87, ..., 89,90;
+Instead of an area-average, the contracted field might instead have +represented the subgrid spatial variation of SST. +In that case, subgrid="con_lat: con_lon: standard deviation". + +

+

Mean over time and longitude: Here, the time-mean +zonal-mean humidity is given as a function of latitude and height. +The means have been formed over the complete time and longitude +intervals of the original data, so these dimensions are collapsed. +

  dimensions:
+    con_lon=1; // collapsed longitude dimension
+    con_time=1; // collapsed time dimension
+    lon=72;
+    sigma=6;
+  variables:
+    float humidity(con_time,sigma,lat,con_lon);
+      humidity:axis="TZYX";
+      humidity:long_name="specific humidity";
+      humidity:subgrid="con_time: mean con_lon: mean";
+    double con_time(con_time);
+      con_time:bounds="bounds_con_time";
+      con_time:old_interval=0.125; // originally at intervals of 3 h
+      con_time:units="days as %Y%m%d.%f";
+    float bounds_con_time(con_time,2);
+    float con_lon(con_lon);
+      con_lon:bounds="bounds_con_lon";
+      con_lon:long_name="longitude";
+      con_lon:modulo=360f;
+      con_lon:topology="circular";
+      con_lon:units="degree_east";
+    float bounds_con_lon(con_lon,2);
+    float sigma(sigma);
+      sigma:bounds="bounds_sigma";
+      sigma:long_name="sigma";
+    float bounds_sigma(sigma,2);
+  data:
+    con_time=19960901.0;
+    bounds_con_time=19960301.0, 19970301.0;
+    con_lon=180;
+    bounds_con_lon=0, 360;
+    sigma=0.99, 0.96, 0.92, 0.8, 0.5, 0.1;
+    bounds_sigma=0.98,1.00, 0.94,0.98, 0.86,0.94,
+                 0.65,0.86, 0.30,0.65, 0.05,0.30;
+This is a mean over the complete +range of longitude from 1 March 1996 to 1 March 1997 +(see section 26 concerning the time coordinate). +The longitude axis indicates circular topology because +this was the case before it was collapsed; after collapse, the topology +is not really meaningful. +If the humidity was subsequently meaned over the depth of the atmosphere +as well, +subgrid would be suffixed with con_sigma: mean, +and con_sigma would have bounds 0.05 and 1.00. + +

+ If the same axis is contracted repeatedly, +the methods may all be +recorded in the subgrid attribute of the data variable, +but only the most recent old_interval +and old_spacing will be shown on the contracted coordinate variable. +But if the axis before contraction is retained in the file (identified by +an expand attribute), and was itself the result of a contraction, +it can record the previous old_interval and old_spacing. +

+ Repeated operations of some methods can be regarded as equivalent to a +single operation. +For instance, meaning longitude cells of 1 degree width to 5 degrees, +and then from 5 to 45 degrees, gives the same result as meaning in one step +from 1 degree to 45 degrees (apart from complications with missing data). +Similarly, meaning a time axis from days into months, then into seasons, +and finally into years could be represented as a single operation of meaning +from days to years. +In that case, the subgrid, +old_interval and old_spacing attributes need not be modified +for successive operations. +The choice of whether to take this approach +is left to the application. +

+ +

23 Time variables and intervals

+

+ A ``time variable'' is one which represents date and time, +which we will refer to hereafter just as ``time''. +An ``interval of time'' is the difference between two times. +

+ It would be possible to describe +time in terms of six components (year, month, day, hour, minute, second) +in a netCDF file, +using six component variables of various data types. +However, it is more efficient and for many purposes more convenient +to represent a time as a single number, +giving the elapsed interval since a certain reference time, +which may be either implicit or explicit. + We refer to conversion from the components of a time into a single +number as ``encoding'', and the reverse as ``decoding''. +Encoding and decoding are complicated because +year and month are units with lengths that depend on the date +and the calendar in use, +so special provisions are needed for time axes. +

+ A ``calendar'' defines the set of valid dates +(year-month-day combinations). +The standard calendar is the Gregorian +(the calendar of udunits), +but climate models do not always use this. +For instance, in the calendar of +the Hadley Centre GCM, all months have 30 days. +The elapsed interval in units of fixed length +(days, hours, minutes, seconds) +between two times will not necessarily be the same +in two different calendars, +because there may be different numbers of valid dates between them. +For example, +the interval between 1 February 1996 and 1 March 1996 is one month, +and equals +29 days in the standard calendar, but 30 days in the Hadley Centre +model calendar, +since 30 February is a valid date in the latter. +Therefore the encoding of a time into an elapsed interval will +depend on the calendar, +and it is necessary to know the calendar when converting. +This standard permits the use of the standard calendar +(below, section 27) +and of other calendars (section 28). +The calendar attribute, +described in the following sections, +indicates the calendar in use. +If a time coordinate variable has no calendar attribute, +the global calendar attribute (section 5), +if present, applies to it. +

+ This standard permits two different methods, +distinguished by their units, +of encoding a time into a number. +These methods, referred to as ``relative time'' +and ``absolute time'', +are described in the following sections +(25 and 26). +Relative time is a more familiar method, +but absolute time offers important advantages. +

+ Time variables may have +an attribute time_format, to specify +a format for printing the date and time, +according to the conventions of the +Unix (TM) date command. +

+ +

24 Time dimension

+

+ Time coordinate variables must always explicitly include +the units attribute; there is no default value. + The long_name attribute is also mandatory. +A time coordinate +variable will be identifiable by its units and long_name. +

+ +

25 Relative time

+

+ A time encoded as a relative time gives the elapsed interval since +a specified reference time; +units takes the form "time-unit since +reference-time", +as per the recommendations of +the Unidata udunits package (but see below concerning +the time-unit) +e.g. a unit of seconds since 1992-10-8 15:15:42.5 +indicates seconds since 8 October 1992 at 3 hours, 15 minutes and +42.5 seconds in the afternoon, in Universal Coordinated Time +(time zones can also be handled). +In order to decode the values on a relative time axis, +the application will in general need to know the calendar; +the encoded time values are meaningless without this knowledge. +Furthermore, a given date may result in different time values +when encoded in two different calendars +with the same units. +For instance, 1996-2-1 15:00:00 is 62.625 days since +1995-12-1 0:0:0 in the standard calendar, +and 60.625 days since 1995-12-1 0:0:0 in the 360-day calendar. +

+ The file udunits.dat defines +second, minute, hour and day as units of time. +Units of months and years are disallowed by Appendix C +of this standard, +because they are not well-defined; +since udunits defines a year as a ``tropical year'' +of 31556925.97 s (674.03 s less than 365 days) +and a month as exactly a twelfth of a year, +use of these units will probably not give the expected results. +For example, 1 month since 1995-4-1 0:0:0 +is treated by udunits as +30.4368 days since 1995-4-1 0:0:0, which is approximately +1995-5-1 10:29, +not 1995-5-1 0:0:0. +Also, 1 year since 1995-4-1 0:0:0 is about 1996-3-31 5:49, +not 1996-4-1 0:0:0. + The udunits unit common_year (exactly 365 days) +is permitted, but not recommended. +

+

A relative time axis for instantaneous measurements of a +quantity: Measurements are made +at noon on 2nd-5th June 1996. +

  dimensions:
+    time=4;
+  variables:
+    double time(time);
+      time:long_name="time";
+      time:units="days since 1996-1-1 0:0:0";
+  data:
+    time=1.5, 2.5, 3.5, 4.5;
+ +

+

A relative time axis for monthly means: +Means are calculated for February, March +and April of 1990. +

  dimensions:
+    time=3;
+  variables:
+    double time(time);
+      time:bounds="bounds_time";
+      time:long_name="time";
+      time:units="days since 1990-1-1 0:0:0";
+    double bounds_time(time,2);
+  data:
+    time=45.0, 74.5, 105.0;
+    bounds_time=31.0,59.0, 59.0,90.0, 90.0,120.0;
+In this example, the main time coordinates are merely representative +values, being the mid-points of their respective months. +Decoded, they are 1990-2-15 0:0:0, 1990-3-16 12:0:0 +and 1990-4-16 0:0:0. + +

+ +

26 Absolute time

+

+ This method of encoding time refers to the separate components +of time, rather than to a single unit of fixed length. +It offers two advantages. +Firstly, the encoded times are meaningful and can be decoded +into components of time without knowledge of the +calendar, although to calculate intervals between them +this knowledge is still required. +Secondly, ``partial'' times can be encoded, +which omit the year, or the ``seasonal phase'' (time of year, +time within the seasonal cycle), or the ``diurnal phase'' +(time of day, time within the diurnal cycle). +By contrast, relative times can only be ``complete'' times, +which include information about all three of these. +

+ The units attribute of absolute time takes the form +"time-unit as time-string", +The possibilities with the recommended data types +and their meanings are as follows: +

++ + + + + + + + + + + + + + + + + +
+Format Data type Interpretation
second as %S.%f float Diurnal phase
+minute as %M.%f float Diurnal phase
+hour as %H.%f float Diurnal phase
+day as %Y%m%d.%f double Time
+day as %Y%m%d int Year and seasonal phase
+day as %m%d.%f double Seasonal phase and diurnal phase
+day as %m%d int Seasonal phase
+day as .%f float Diurnal phase
+calendar_month as %Y%m.%f double + Year and seasonal phase
+calendar_month as %m.%f float Seasonal phase
+calendar_year as %Y.%f double Year and seasonal phase
+calendar_year as %Y int Year
+calendar_year as .%f float Seasonal phase
+

+Standard abbreviations and plural forms of the unit names are acceptable, +as usual. +The time-units calendar_year and calendar_month +are units of time defined by this standard (Appendix C). +

+ The time-string +codes show how the year, month, day within month +and time within day are encoded into a single number, +after the fashion of the Unix (TM) date +and printf commands, thus: +

++ + + + + + + + + + + + +
+Format letter Interpretation
%Y Year (including century)
+%m Two-digit month (01=January)
+%d Two-digit day within month
+%H Hours since midnight
+%M Minutes since midnight
+%S Seconds since midnight
+%f Floating-point fraction of the specified time-unit
+. Position of decimal point
+

+Since an encoded time is an ordinary number, leading zeroes in the +integer part may be omitted. +It is not mandatory to use the suggested data types, which are recommended +on grounds of precision. +If an integer data type is used for an absolute time variable whose format +includes a fraction %f, +a fraction of zero is assumed. +If a floating-point data type is used for a format not including %f, +any fraction is ignored. +

+In absolute time, +3 p.m. on 5th April 1998 is encoded with value 19980405.625 +and units="day as %Y%m%d.%f". +The advantage of this method of encoding a complete time is that it +can be done without knowledge of the calendar, +whereas if we encoded in relative time units of +days since 1900-1-1, the value would be 35888.625 in the standard +calendar, and 35374.625 in the 360-day calendar. +We also know, without reference to the calendar, that the value +19980605.625, with the same units, is a time exactly two calendar months +later, and 19970405.625 is exactly one calendar year earlier. +But to calculate these intervals in other time units--days, hours, etc.--we +still need to know the calendar. +

+ The only complete form of absolute time +is "day as %Y%m%d.%f". +Note in particular that the forms "calendar_month as %Y%m.%f" +and "calendar_year as %Y.%f" are partial times which imply +no information about the diurnal phase. +This is a very important point. +For instance, 1998.25 calendar_year as %Y.%f means no more than +``a quarter of the way through 1998 as regards the seasonal cycle''. +This meaning is the same in the standard and 360-day calendars. +Because this representation carries no information about diurnal phase, +it is not permitted to decode it to 1998-4-2 3:0:0 +(i.e. 91.25 days from the start of the year) +in the standard calendar or +1998-4-1 0:0:0 in the 360-day calendar. +Similarly, 199804.3 calendar_month as %Y%m.%f means +``30% of the way through April 1998 as regards the seasonal cycle''. +Examples below show the use of such partial times. +

+ Note also +that the only form of partial time which is composed of the +seasonal and diurnal phases is "day as %m%d.%f"; +there is no method of encoding the seasonal phase as a fraction +of a calendar year or month in combination with the diurnal phase. +Should this be required, the application could construct it as a +two-component time variable. +This exclusion seems reasonable because data which resolves both +the seasonal and diurnal cycles must belong to a known calendar +(it will exhibit a certain number of days in a year, for instance) +and so its seasonal cycle can be labelled by month and day. +The calendar-independent representations of the seasonal cycle, +shown in examples below, +are more useful when portions of the seasonal cycle have been averaged, +in which case the seasonal and diurnal cycles, if both present, +will be on separate axes. +

+ The forms of partial time which do not include .%f +in the time-string are discrete variables, rather than continuous. +The interval of time in days or years +covered by a time axis in one of these forms +is reckoned by including the both ends of the axis in the count, +or equivalently by adding one to the difference of the ends. +For instance, an axis with units="calendar_year as %Y" +that runs from 1930 to 1939 covers ten years, not nine, +since both ends are included. +This is a partial time indicating only the year, +not the seasonal phase. +Contrast this with an axis including the seasonal phase, +having units="calendar_year as %Y.%f", +and end points of 1930.0 and 1939.0. +This axis spans nine years, from the beginning of 1930 to the +beginning of 1939, and does not include 1939 itself. +Examples below illustrate this point further. +

+

An absolute +time axis for instantaneous measurements of a quantity: +Measurements are made at noon on 2nd-5th June 1996. +

  dimensions:
+    time=4;
+  variables:
+    double time(time);
+      time:long_name="time";
+      time:units="days as %Y%m%d.%f";
+  data:
+    time=19960602.5, 19960603.5, 19960604.5, 19960605.5;
+ +

+

An absolute time axis for monthly means, encoded in days: +

  dimensions:
+    time=3;
+  variables:
+    double time(time);
+      time:bounds="bounds_time";
+      time:long_name="time";
+      time:units="days as %Y%m%d.%f";
+    double bounds_time(time,2);
+  data:
+    time=19900215.0, 19900316.5, 19900416.0;
+    bounds_time=19900201.0,19900301.0, 19900301.0,19900401.0,
+                19900401.0,19900501.0;
+As in the relative time version of this example, +the main time coordinates are the mid-points of their respective months. +Although they are encoded straightforwardly, +their values depend on the calendar. +If one was comparing means over these months from data sources which +used different calendars, that might be inconvenient, +and could be avoided as in the next example. + +

+

An absolute time axis for monthly means, encoded in months: +

  dimensions:
+    time=3;
+  variables:
+    double time(time);
+      time:bounds="bounds_time";
+      time:long_name="year and seasonal phase";
+      time:units="calendar_months as %Y%m.%f";
+    double bounds_time(time,2);
+  data:
+    time=199002.5, 199003.5, 199004.5;
+    bounds_time=199002.0,199003.0, 199003.0,199004.0, 199004.0,199005.0;
+This method shows directly that the main coordinates are half-way +through their months. + +

+

A partial time defining just the year: +An axis of this kind could be used to record +the number of occurrences of a particular kind of event: +

  dimensions:
+    year=3;
+  variables:
+    int year(year);
+      year:long_name="year";
+      year:units="calendar_year as %Y";
+    int count(year);
+  data:
+    year=1991,1992,1993,1994,1995;
+    count=0,2,1,0,1;
+As discussed above, this axis spans five years. +No boundaries are supplied, because each element of count +applies to just a single year. +Hence, both the upper and the lower year boundaries would be +equal to their years. + +

+

Year and seasonal phase defined in calendar years: +By contrast to the last example, if it was appropriate to indicate that +each count applied to the whole of the continuous period of time of its +respective year, this could be done thus: +

  variables:
+    double year(year);
+      year:bounds="bounds_year";
+      year:long_name="year and seasonal phase";
+      year:units="calendar_year as %Y.%f";
+    double bounds_year(year,2);
+    int count(year);
+  data:
+    year=1991.5, 1992.5, 1993.5, 1994.5, 1995.5;
+    bounds_year=1991.0,1992.0, 1992.0,1993.0, 1993.0,1994.0,
+                1994.0,1995.0, 1995.0,1996.0;
+    count=0,2,1,0,1;
+The use of floating-point years allows us conveniently to represent +exactly the beginning and ending of a year and a point half-way through. +In the standard calendar, of course, the interval from 1992.0 to 1993.0 +is longer in relative time than all the other years. +But for some purposes, it might be more useful to record that each interval +is a calendar year. +This could be especially helpful when comparing data from different +calendars. + +

+

Seasonal phase as a function of year: +Here we show the date within the year of a particular event, +such as the highest daily +maximum temperature, or the onset of the monsoon, as a partial time +within its year. +

  dimensions:
+    year=5;
+  variables:
+    int year(year);
+      year:long_name="year";
+      year:units="calendar_year as %Y";
+    int date(year);
+      date:long_name="seasonal phase";
+      date:units="day as %m%d";
+  data:
+    year=2011, 2013, 2027, 2028, 2051;
+    date=629, 627, 626, 703, 710;
+The event concerned occurred on 29th June 2011, 27th June 2013, +26th June 2027, 3rd July 2028 and 10th July 2051. +Clearly the date variable could have been encoded as a complete time, +perhaps in relative time units, +but this would have included redundant year information. + +

+ A time variable which indicates seasonal phase but not year has a +modulo of one year. If it spans the entire seasonal cycle, it also has +circular topology. Similarly, a time variable indicating diurnal phase +but not seasonal phase has a modulo of one day, and has circular +topology if it spans the entire diurnal cycle. +These kinds of time coordinate are particular useful for +representing climatological time, in conjunction with other contracted +time axes. See section 29. +

+

Average seasonal cycle expressed in months: +Data for solar radiation as 3-monthly averages. +

  dimensions:
+    time=4;
+    lat=72;
+    lon=96;
+  variables:
+    float sol(time,lat,lon);
+      sol:axis="TYX";
+      sol:long_name="vertical component of "
+        "solar radiative flux density";
+      sol:units="W m-2";
+    float time(time); 
+      time:bounds="bounds_time";
+      time:long_name="seasonal phase";
+      time:modulo=12.0f;
+      time:topology="circular";
+      time:units="calendar_month as %m.%f";
+    float bounds_time(time,2);
+  data:
+    time=10.5, 13.5, 16.5, 19.5;
+    bounds_time=9.0,12.0, 12.0,15.0, 15.0,18.0, 18.0,21.0;
+The first time point applies from the beginning of month 9 to the +beginning of month 12, i.e. September to November inclusive. +A representative main coordinate is given of half-way through October. +The second point runs to the beginning of month 15, which is equivalent +to 3 i.e. March under modulo 12, and thus covers December to February. +The use of the modulo allows the main coordinate to be specified as +monotonic, as is generally required. Because the axis is also circular, +it would be permissible to rotate the values in order to begin with a +different season. + +

+

Average seasonal cycle expressed in years: The time +coordinate above could equally well be given in calendar years, thus: +

  dimensions:
+    double time(time); 
+      time:bounds="bounds_time";
+      time:long_name="seasonal phase";
+      time:modulo=1.0;
+      time:topology="circular";
+      time:units="calendar_year as .%f";
+    float bounds_time(time,2);
+  data:
+    time=0.7917, 1.0417, 1.2917, 1.5417;
+    bounds_time=0.6667,0.9167, 0.9167,1.1667,
+                1.1667,1.4167, 1.4167,1.6667;
+Here, the periods have been constructed as exactly quarters of a +year, beginning two-thirds of the way through the year. +In the 360-day calendar, this is identical to the last example, of +periods of three months starting at the beginning of September, +but in the standard calendar it is slightly different, +since a quarter of a year is not exactly three calendar months. + +

+ +

27 Gregorian calendar

+

+ This standard recommends that Gregorian times be given in +units of +days as %Y%m%d.%f with data type double +(section 26), +unless compatibility is essential with applications that cannot process +absolute times. + In that case, Gregorian times may have units of time +formatted as per the recommendations of the Unidata udunits package, +which specify a unit and a reference time, i.e. a relative time +(section 25). + The recommended unit is days, with data type double. +

+ Intervals between two times in the standard +Gregorian calendar can be calculated by the Unidata udunits package. +Udunits implements the mixed Gregorian/Julian calendar system, +as followed in England, in which +dates prior to 1582-10-15 are assumed to use the Julian calendar. +Other software cannot be relied upon to handle the change of calendar +in the same way, so for robustness it is recommended that the reference +date be later than 1582. +If earlier dates must be used, it should be noted that +udunits treats 0 AD as identical to 1 AD. +

+Data type double gives a precision of about +16 decimal digits, +which means that it can resolve tenths of a second +for years of up to O(1 million) in relative times. +The precision of absolute times is an order of magnitude worse, +since a year looks like 10000 days, rather than 365. +The larger the year, the worse the absolute precision. +If very large years are needed and the precision is not sufficient, +the reference year will have to be modified to keep the interval +small enough. +

+ If there is no calendar attribute applying to +a time variable, +the values are assumed to be in the normal Gregorian calendar. +This can be made explicit by setting calendar to standard +or gregorian. +

+ +

28 Non-Gregorian calendars

+

+ It is recommended that +times in other calendars should be encoded in units of +days as %Y%m%d.%f with data type double +(section 26). +Relative times are permitted, the recommended units being +days since 1-1-1 (midnight on 1 January of year 1), +with data type double. +Since the Unidata udunits package can process only +the standard calendar, +an extension will be required to process relative times +for other calendars. +

+ Apart from the Gregorian, +calendars recognised by this standard are +julian for the Julian calendar (in which all years divisible +by four are leap years), +noleap for a calendar with 365 days in every year, +and 360 when each month has 30 days in every year. +If any other calendar is used, a suitable description should appear +in the calendar attribute, +but generic applications cannot be expected to be able to encode and +decode relative times or calculate intervals in the calendar concerned. +

+ +

29 Multiple time axes and climatological time

+

+ There is no bar on a data variable having more than one +dimension in +a particular quantity, so long as the dimensions have different names. +A particular use of this is to decompose time +into multiple partial time dimensions (section 26), +of which one or may be collapsed (section 22). + This gives a method of indicating disjoint intervals of time +belonging to corresponding parts of the seasonal or diurnal cycles. +When a variable has two or three time axes, +the first interval of time which they cover is assumed to begin +at the earliest boundary values of all the axes. +If there is an uncollapsed axis in combination with collapsed axes, +it is a ``climatological time'' axis. +There may be more than one--see below for an example. +

+COARDS recommends use of year 0 to indicate climatological time. +We do not favour this convention. +Firstly, it does not provide any way of recording which years were used +to make the climatology. +Secondly, udunits treats year 0 and year 1 as identical +(which is reasonable because year 0 does not exist--there is +no year between 1 AD and 1 BC). +

+

A mean of a corresponding months in a number of years: +A longitude-latitude precipitation field with time axes +to indicate the mean over the months of January in 1961 to 1990 inclusive: +

  dimensions:
+    con_year=1;
+    year=30;
+    month=1;
+  variables:
+    float precipitation(con_year,month,lat,lon);
+      precipitation:axis="-TYX";
+      precipitation:subgrid="month: mean con_year: mean";
+    int con_year(con_year);
+      con_year:bounds="bounds_con_year";
+      con_year:expand="year";
+      con_year:long_name="year";
+      con_year:old_interval=1;
+      con_year:units="calendar_year as %Y";
+    int bounds_con_year(con_year,2);
+    int year(year);
+    float month(month);
+      month:bounds="bounds_month";
+      month:long_name="seasonal phase";
+      month:units="calendar_month as %m.%f";
+    float bounds_month(month,2);
+  data:
+    con_year=1975;
+    bounds_con_year=1961, 1990;
+    year=1961, 1962, 1963, ..., 1990;
+    month=1.5;
+    bounds_month=1.0, 2.0;
+The representative year is not likely to be particularly useful in +this case; the important information is the boundaries, which indicate +the range of years used to form the climatological mean. +These years are also given explicitly, and optionally, for reference. +Since the con_year axis has a discrete form of time (%.f does +not appear in the time-string--see section 26), +we have to include both ends of the axis in working out how many years are +involved in the mean: +bounds_con_year tells us that years from 1961 to 1990 were used, +which comes to 30 years, counting both 1961 and 1990. +

+Suppose we wish to include in the same data variable the mean over Decembers +from 1960 to 1989. +To do this, we give the month variable modulo=12.0f, +which previously was not needed, change the dimension to month=2 and +the data to +

    month=0.5, 1.5;
+    bounds_month=0.0,1.0, 1.0,2.0;
+According to the convention, the earliest time indicated by the combined axes +is the lower boundary of all of them, which is the start of the 0th month of +1961, equivalent to the start of the 12th month of 1960 under modulo 12. +Alternatively, we could have +
    bounds_con_year=1960,1989;
+    month=12.5, 13.5;
+    bounds_month=12.0,13.0, 13.0,14.0;
+This is exactly equivalent. The last January used ends at the start of the 14th +month of 1989, which is the start of February 1990, as before. +

+There is no standardised way to indicate missing months from the mean, say +December 1974 in this case, +except to include con_year:old_spacing="disjoint". +The information could be included as a note in the subgrid attribute, thus +"con_year: mean (December 1974 missing)". + +

+

Climatological seasonal means for several decades: +This is an extension of the previous case, +and of the +example of an average seasonal cycle in section 26. +Here, the axes are set up +to indicate climatological means for two of the seasons +in three successive decades. +

  dimensions:
+    decade=3;
+    season=2;
+  variables:
+    float precipitation(decade,season,lat,lon);
+      precipitation:axis="-TYX";
+      precipitation:subgrid="season: mean decade: mean";
+    int decade(decade);
+      decade:bounds="bounds_decade";
+      decade:old_interval=1;
+      decade:units="calendar_year as %Y";
+    int bounds_decade(decade,2);
+    int season(season);
+      season:bounds="bounds_season";
+      season:calendar="standard";
+      season:modulo=1200;
+      season:units="day as %m%d";
+    int bounds_season(season,2);
+  data:
+    decade=1966, 1976, 1986;
+    bounds_decade=1961,1970, 1971,1980, 1981,1990;
+    season=115, 415;
+    bounds_season=1,228, 301,531;
+Here, precipitation[0][0][*][*] is the data for December-February +(i.e. 1 December to 28 February inclusive) +of the decade 1960-1970 (first December in 1960, last February in 1970), +while [2][1][*][*] is March-May 1981-1990. +The choice has been made to give the seasonal phase in months and days, +rather than months alone; +hence the modulo is 1200 rather than 12. +Under modulo 1200, midnight on 1 December can be expressed equivalently +as 1 or 1201. +If 1201 were specified, it would mean that the first interval of time +began on 1 December 1961 (rather than 1960), +taking the combination of the lower boundaries of both time axes; +the value 1 is a year earlier. +The drawback of this %m%d scheme +is that it is awkward or impossible to give accurate +representative dates for the middle of the periods, especially since +February has variable length. +The absolute time format %m.%f for seasonal phase +is better from this point of view. +The season axis is not shown as having circular topology because no +information is implied about the other two seasons. + +

+

Average early June maximum temperatures for several years: +In this example, the dimensions indicate that +maximum daily temperatures (between 9 a.m. on the day +of record and 9 a.m. of the previous day) were recorded for 1-10 June, +and an average maximum found for these ten days in +each of the years 1980-1984. +

  dimensions:
+    year=5;
+    con_season=1;
+    con_day=1;
+  variables:
+    float temperature(year,con_season,con_day);
+      temperature:axis="T--";
+      temperature:subgrid="con_day: maximum con_season: mean";
+    int year(year);
+      year:long_name="year";
+      year:units="calendar_year as %Y";
+    int con_season(con_season);
+      con_season:bounds="bounds_con_season";
+      con_season:long_name="seasonal phase";
+      con_season:old_interval=1;
+      con_season:units="day as %m%d";
+    int bounds_con_season(con_season,2);
+    float con_day(con_day);
+      con_day:bounds="bounds_con_day";
+      con_day:long_name="diurnal phase";
+      con_day:modulo=24.0f;
+      con_day:units="hour as %H.%f";
+    float bounds_con_day(con_day,2);
+  data:
+    year=1980, 1981, 1982, 1983, 1984;
+    con_season=605;
+    bounds_con_season=601, 610;
+    con_day=-3.0;
+    bounds_con_day=-15.0, 9.0;
+The diurnal phase of -15 h means 15 hours before the beginning +of the day in question, i.e. 9 a.m. on the previous day. +No bounds are given for the year, because it is a discrete quantity, +and there is no further information which could be added. +But if the five years were averaged together, this would collapse the year +axis, and the extreme years of 1980 and 1984 would be recorded as +the boundaries of the collapsed axis. +If, say, 1981 were not used in forming the average, +the collapsed axis would have attribute old_spacing="disjoint". + +

+

Daily values as an average of subdaily values: +Instantaneous pressure measurements are made at intervals of 3 hours +(first measurement at midnight) +throughout the days 6 May to 9 June 1937, +and daily means formed from midnight to midnight. +

  dimensions:
+    con_subday=1;
+    day=35;
+  variables:
+    float pressure(day,con_subday);
+      pressure:axis="T-";
+      pressure:subgrid="con_subday: point con_subday: mean";
+    float con_subday(con_subday);
+      con_subday:bounds="bounds_con_subday";
+      con_subday:long_name="diurnal phase";
+      con_subday:old_interval=0.125f;
+      con_subday:old_spacing="uniform";
+      con_subday:units="days as .%f";
+    float bounds_con_subday(con_subday,2);
+    int day(day);
+      day:long_name="year and seasonal phase";
+      day:units="days as %Y%m%d";
+  data:
+    con_subday=0.5;
+    bounds_con_subday=0.0, 0.875;
+    day=19370506, 19370507, ..., 19370608, 19370609;
+Note that the con_subday axis is shown with two +subgrid methods, referring to subgrid variation before +and after its collapse. +The only point here in having separate axes for day and diurnal phase +is to show when the first and last instantaneous measurements were made +in each day. +If this is not important to record, the two axes could be merged together +thus: +
  dimensions:
+    day=35;
+  variables:
+    float pressure(day);
+      pressure:subgrid="day: point day: mean";
+    float day(day);
+      con_subday:bounds="bounds_day";
+      con_subday:old_interval=0.125f;
+      con_subday:long_name="time";
+      con_subday:old_spacing="uniform";
+      con_subday:units="days as %Y%m%d.%f";
+    float bounds_day(day,2);
+  data:
+    day=19370506.5, 19370507.5, ..., 19370608.5, 19370609.5;
+    bounds_day=19370506.0,19370507.0, 19370507.0,19370508.0, ...,
+               19370608.0,19370609.0, 19370609.0,19370610.0;
+If the 35 days were then averaged together, +the date axis would collapse with bounds of 19370506.0 and 19370610.0. +The subgrid attribute would not need modification since +it is already shown as a mean over the day axis. + +

+

Average diurnal cycle: +The following axes are appropriate for +the average diurnal cycle of +precipitation rate in July 1970-1979 as a function of latitude: +

  dimensions:
+    con_year=1;
+    con_month=1;
+    hour=8;
+    lat=45;
+    con_lon=1;
+  variables:
+    float ppnrate(con_year,con_month,hour,lat,con_lon);
+      ppnrate:axis="--TYX";
+      ppnrate:subgrid="con_lon: mean con_month: mean "
+        "con_year: mean";
+      ppnrate:units="kg m-2 s-1";
+    int con_year(con_year);
+      con_year:bounds="bounds_con_year";
+      con_year:old_interval=1;
+      con_year:units="calendar_year as %Y";
+    int bounds_con_year(con_year,2);
+    float con_month(con_month);
+      con_month:bounds="bounds_con_month";
+      con_month:units="calendar_month as %m.%f";
+    float bounds_con_month(con_month,2);
+    float hour(hour);
+      hour:bounds="bounds_hour";
+      hour:modulo=24.0f;
+      hour:topology="circular";
+      hour:units="hour as %H.%f";
+    float bounds_hour(bounds_hour,2);
+  data:
+    con_year=1975;
+    bounds_con_year=1970, 1979;
+    con_month=7.5;
+    bounds_con_month=7.0, 8.0;
+    hour=1.5, 4.5, 7.5, 10.5, 13.5, 16.5, 19.5, 22.5;
+    bounds_hour=0.0,3.0, 3.0,6.0, 6.0,9.0, 9.0,12.0,
+                12.0,15.0, 15.0,18.0, 18.0,21.0, 21.0,24.0;
+ +

+ +

30 Invalid values in a data variable

+

+ Invalid values are any which fall +outside the valid range or equal the fill value, as indicated by the +Unidata-standard attributes described here. An invalid value indicates bad data +i.e. a software problem, which is a different circumstance from unknown or +missing data (see section 31). +Invalid values are not permitted in a coordinate variable, +but the attributes which define the valid range may be used in boundary +variables (section 20) to indicate unbounded cells. +

+ The attribute valid_min is a scalar specifying the +minimum valid value for a variable. The attribute valid_max specifies +the maximum valid value, while valid_range is a vector of two numbers +specifying the minimum and maximum valid values, in that order, equivalent to +specifying values for both valid_min and valid_max +attributes. Any of these attributes define the valid range. The attribute +valid_range must not be defined if either valid_min or +valid_max is defined. Generic applications should treat values outside the +valid range as invalid. The type of each valid_range, valid_min +and valid_max attribute should match the type of its variable. The +Unidata special treatment of byte type is not included here as we do not +recommend use of that type (see section 3). +

+ A scalar attribute with the name _FillValue and of the +same type as its variable is used as the fill value for the variable. The +netCDF package defines a default fill value for each type of variable, so it is +not necessary to define your own _FillValue attribute if the default is +suitable. The purpose of the fill value is to save the applications programmer +the work of prefilling the data and also to eliminate the duplicate writes that +result from netCDF filling in undefined data with its default fill value, +only to be immediately overwritten by the programmer's preferred value. This +value is considered to be a special value that indicates undefined data, and is +returned when reading values that were not written. The _FillValue +should be outside the range specified by valid_range (if used) for a +variable. + In cases where the data variable is packed using the +scale_factor and add_offset attributes (section 33), the +_FillValue attribute applies the numbers as packed, so they must be +checked against it before unpacking. +

+ If none of valid_min, valid_max or valid_range +is defined then generic applications should define a valid range by using the +fill value (whether defined explicitly or by default); if the fill value is +positive then it defines a valid maximum, otherwise it defines a valid +minimum. For integer types, there should be a difference of 1 between the fill +value and this valid minimum or maximum. For floating point types, the valid +extreme should have a magnitude which is half the magnitude of the fill value. +We recommend a factor of two, rather than a difference of one bit, because +it is easier for applications programmers. There is no special treatment for +byte as we do not recommend that type (see section 3). +

+ +

31 Missing values in a data variable

+

+ Missing values are not permitted in a coordinate variable, so this +section applies only to data variables. The missing_value attribute +indicates a value that is used for data that are unknown or ``missing''. This +attribute is not be treated in any special way by the netCDF API, unlike the +_FillValue attribute (section 30). +The missing_value should be outside the valid range +(section 30), so that generic applications will treat it +appropriately. +The netCDF data type of +the missing_value attribute should match the netCDF data type of the +data variable that it describes. In cases where the data variable is packed via +the scale_factor and add_offset attributes (section 33), +the missing_value attribute +matches the type of and should be compared with the data +after unpacking. +This standard is unlike COARDS in giving +a particular interpretation to the +distinction between missing_value and _FillValue. +

+ +

32 Compression by gathering

+

+ To save space in the netCDF file, +it may be desirable to eliminate points +from data arrays which are invariably missing. +Such a compression can operate over one or more adjacent axes, +and is accomplished with reference to a list of +the points to be stored. +The list is constructed by considering a mask array which has just the +axes to be compressed, +and mapping this array onto one dimension without reordering. +The list is the set of indices +in this one-dimensional mask of the required points. +In the compressed array, +the axes to be compressed are all replaced by a single axis, +whose dimension is the number of wanted points. +The wanted points appear along this dimension in the same order they +appear in the uncompressed array, with the unwanted points skipped over. +Compression and uncompression are executed by looping over the list. +

+ The list is stored as the +coordinate variable for the compressed axis of the data array. +Thus, the list variable and its dimension have the same name. +The list variable has a string attribute compress, +containing a blank-separated list of the dimensions which were affected +by the compression in the order of the CDL declaration of the +uncompressed array. +The presence of this attribute identifies the list variable as such. +The list, the original dimensions and coordinate variables +(including component, associated and boundary variables), +and the compressed data variables with all the attributes +of the uncompressed variables +are written to the archived netCDF file. +The uncompressed data variables can be reconstituted +exactly as they were +using this information, +except that their original variable names are not known. +

+

Horizontal compression of a three-dimensional array: +We eliminate sea points at all depths +in a longitude-latitude-depth array of soil temperatures. +In this case, only the longitude and latitude axes would be affected +by the compression. +We construct a list landpoint(landpoint) +containing the indices of land points. +

  dimensions:
+    lat=73;
+    lon=96;
+    landpoint=2381;
+    depth=4;
+  variables:
+    long landpoint(landpoint);
+      landpoint:compress="lat lon";
+    float landsoilt(depth,landpoint);
+      landsoilt:axis="Z-";
+      landsoilt:long_name="soil temperature";
+      landsoilt:units="K";
+    float depth(depth);
+    float lat(lat);
+    float lon(lon);
+  data:
+    landpoint=363, 364, 365, ...;
+Since landpoint[0]=363, for instance, +we know that landsoilt[*][0] +maps on to point 363 of the original data with dimensions (lat,lon). +This corresponds to indices [3][75]. + +

+

Compression of a three-dimensional field: +We compress a longitude-latitude-depth field of ocean +salinity by eliminating points below the sea-floor. +In this case, all three dimensions are affected by the compression, +since there are successively fewer active ocean points at increasing +depths. +

  variables:
+    float salinity(oceanpoint);
+      salinity:axis="-";
+    long oceanpoint(oceanpoint);
+      oceanpoint:compress="depth lat lon";
+    float depth(depth);
+    float lat(lat);
+    float lon(lon);
+This information implies that the salinity field should be +uncompressed to an array with dimensions (depth,lat,lon). + +

+ +

33 Compression using a scale and offset

+

+ This standard +endorses the use of the optional Unidata-standard attributes +scale_factor and add_offset +for data and coordinate variables. +These attributes +can be used to provide simple number compression (packing), +to store low-resolution +floating-point data as small integers in a netCDF file. +After the data values of the variable have been read in, +they are to be multiplied by the scale_offset, +and have add_offset added to them. +If both scale_factor and add_offset attributes are present, the +data are scaled before the offset is added. +When scaled data are +written, the application should first subtract the offset and then divide by +the scale factor. + This procedure is concerned only with storage. +It does not affect the unit of the quantity. +For instance, a pressure variable with values in the range +900.0-1100.0 Pa +could be converted to short integers in the range +±20000 +by subtracting 1000 and dividing by 0.005 i.e. multiplying by 200. +The units of the compressed variable are still recorded as pascals. +

+ This standard is more restrictive than the netCDF Users' +Guide with respect to the use of the scale_factor and add_offset +attributes; ambiguities and precision problems related to data type conversions +are resolved by these restrictions. If the scale_factor and +add_offset attributes are of the same data type as the associated variable no +restrictions apply; the unpacked data is assumed to be of the same data type as +the packed data. However, if the scale_factor and add_offset +attributes are of a different data type from the variable +(containing the packed data) then in files adhering to this standard the +variable may only be of type short or long. We +exclude byte on grounds discussed in section 3. The +attributes scale_factor and add_offset (which must match in data +type) must be of type float or double. The data type of the +attributes should match the intended type of the unpacked data. (It is not +advised to unpack a long into a float as there is a potential +precision loss.) Users should note that Unidata may provide a built-in +means of packing data in netCDF files in future. +

+

+

+

A Attributes

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
+Attribute T Use Section(s) Description
add_offset N CD 30 33 +Additive offset for packing data
+appendices S G 5 + Version number of these appendices
+associate S CD + 18 19 Identifies variables + containing alternative sets of coordinates
+axis S D + 9 16 18 + Identifies special spatiotemporal dimensions
+bounds N C + 20 22 29 + Identifies a variable containing boundary values
+calendar S GD 5 23 27 + 28 Calendar used for encoding time axes
+comment S G 5 + Additional information about the file
+component S CD + 17 20 Identifies variables containing + components of a variable
+compress S D 32 + Records the dimensions which have been compressed by gathering
+Conventions S G + 5 Identifies the netCDF standard
+coordinates S CD + 18 Synonym for associate
+expand S C + 22 29 Records coordinates + of an axis before contraction
+_FillValue N D + 30 Indicator of invalid data
+FORTRAN_format S CD + 12 Format for a variable
+history S GD 5 12 + Evolution of the data in the file
+institution S GD 5 12 + Who made or supplied the data
+long_name S CD 12 14-16 + 24 Long description of a physical quantity
+modulo N CD + 12 14 26 Arithmetic + modulo of a variable
+north_pole N D + 10 Long.,lat. of rotated + North Pole
+old_interval N C 22 29 + The typical separation between points on an axis before contraction
+old_spacing S C + 22 29 Indicates the spacing + of points along an axis before contraction
+positive S C 16 + Direction of positive for a vertical axis
+production S GD 5 12 + How the data was produced
+quantity_table S G 5 12 + URL of the quantity table
+scale_factor N CD 30 33 + Multiplicative factor for packing data
+subgrid S D + 21 22 29 + Records how the data values represent subgrid variation
+topology S C + 13 14 26 Topology + of an axis (circular or not)
+time_format S CD + 23 Format for printing a time and date
+units S CD 12 14 15 + 24-28 + Units of a physical quantity
+valid_max N CD + 20 30 Largest valid value of a + variable
+valid_min N CD + 20 30 Smallest valid value of a + variable
+valid_range N CD + 30 Smallest and largest valid values + of a variable
+

+

T is S for string, N for numeric. +
Use is composed of G for global, +C for coordinate variables (including multidimensional +coordinate variables), D for data variables. +

+

+

B Methods of representing subgrid variation

+See section 21. +

+

++ + + + + + + + + + + + + + +
+Method Units Description
cell u Value is a property of the whole cell (e.g. an integral)
+maximum u Maximum
+median u Median
+mid-range u Average of maximum and minimum
+minimum u Minimum
+mean u Mean (average)
+mode u Mode (most common)
+point u Value applies at gridpoint
+standard deviation u Standard deviation
+variance +u² + Variance
+

+

Units: u means the units of the quantity whose +subgrid variation is represented by this method. +

+

C Modifications to udunits.dat

+See section 11. +

+ The unit unity is defined as a dimensionless constant +equal to one. +

+ The unit degrees is not permitted, because +it creates ambiguities when attempting to +differentiate longitude and latitude coordinate variables. +This unit does not appear in the current version +of the file. +

+ The units calendar_month and calendar_year +are units of time, but cannot be converted into each other +or any other units of time, +except that multiples of 12 calendar months equal integral numbers +of calendar years. The units year and month are not allowed, +because they can cause confusion. +

+

+

D Long names for quantities

+See section 12. +This Appendix is not yet available. +As well as existing as part of this standard, +it will be made available on the web. +

+

++ + + + + + + + + + + + + +
+Version long_name units
1.0 depth below the surface m
+1.0 height above the surface m
+1.0 latitude degree_north
+1.0 longitude degree_east
+1.0 pressure Pa
+1.0 soil temperature K
+1.0 specific humidity unity
+1.0 temperature K
+1.0 time s
+

+

+

Version: The version of the appendices at which this +quantity was introduced.
+long_name: Case, spaces and punctuation are not +significant in the long_name. +

+

+

+

+Jonathan Gregory | jmgregory@meto.gov.uk +
+
+Robert Drach | drach@llnl.gov +
+

+LLNL Disclaimers +

+UCRL-MI-127703 +

+ + diff --git a/GDT/1.2.ps.gz b/GDT/1.2.ps.gz new file mode 100644 index 0000000000000000000000000000000000000000..519ea97dbf1f90f1b5c4f7bef8fb197bff799724 GIT binary patch literal 153638 zcmV)?K!U#?iwFo>IH)!N12HZ#E^u=I?EQO>Eje;02>x%M;w}~q^kAoy^uvgtM?-tj zH$0pJ0o&Rhl^_(*Xbe)7w^TV@s;tS((qj|#-S;?YkDrlrKYk)^OX{9HRe8hANT-uV zqtVRx@PGIJ_|L!m#iwt+``gdIxcph5fA{Z}?>~S2<9FZxhY#QU!?%C<;lF)&{IkdT zHy@Vo{_xM=fBSd8{qe)}JU{)f*MFZMKK#|!-+cS)uYdPp{pZgge)-*Re*DL;zyEwK z;jh2_@pqs9!-xOv^{@YrUw`-efB5|SAHV(X_kVUR?waGiu`|MBmyh5hHR z|M2W@Es|JAqO{pL5{{`&L(@i2Y$x8MH$t8f1Jhu?kvr$77c=dZ8z`0$JG=->Un+KOKv z_7r}}fByRWumAq@kDtH4*5jK$eEa`?3gLhK@N2HfhwnfC{dfQH`5Vsr%kTd9{jaZe z`S91DfA!(JKmPcKKW>bCe*UvC{n_R6AHp0rhP-_KyKjI0pTGa^*T4MyKhSjqImk6-`z#~&<}e*FHApFjNY`Nvww-+%Yb=gK^O zV^@J6uW$Op_utW1*Gj)`#atf4hf>}*pMUc~i1p21{pC+r;-9vJuU>w-R`r`L)1UtG z!&krj>GMDR`nP}jOZw_pKmGE)89_*T4JC=g;5#1VZ?gk!$AbAlRhgKmPi= zKYaW7ho8Rse^%hbueL(+fBO2z|1y;>`bGKkU#~CCFE9TkaJOC2-)|Cs{reYd{rl@5 zzpubIzxlZ6SDF3G4_BClHGlW}AOHHF|M2-k+;VT&f2xf8&F_Ew{-1>Q za#cbfuCM&?m)EiM{Z%P_`08)JzIM*9K79TC_h0|>RnWh^K6h0Qzq*2d{PuU>EWiEw z`=7Rd3ikQ?jZf?EfBoI2&^}x%`Iq#08yA24!%tuR=DY8EDWqZlWnWL{aSjh@5ZYu#;vZIe)wqNH3cu4M{^`}34_lNDBMPNSv=Bhon`e{9{S$_T7U;N{@-~9O7pRT|+U;p^^ zmX&_@`S*YK<8MDa7SZ~<&yw-CpSSq}bAI5Q4^7U0x+=M!Xu8ZF2AQK|-v06f|GYgg z$a~rH(3W?)B*_2GKaFxvO*Ov#`|E@x8EMP-<0d0t)y^;Q@2~0CzS!PSI%WH>4#2lslWQL^{0-C!UqZ;nBgCoj~}$F3OgXR^mJ()XT451 zKmPjbU;p;=H=qCEIw@HmD`FeiHf(?KXOkAlg_m;vtg!yS{v=HQQ~CGrzrLCq-+%t$ zYJGg(So!+XSHJ%CwH13k@n1JZzR9aG=bC?O?lx#Px@XRo^2-lp=v<#>tw>H7Hli@N za`T#f8%Ue3y1qHQzkT>`zyIzZ zfB%!9TZv!&@-IG=4*$y!b?%~#cbo0-<98pv`kN0Ujb5rbmFjKKHggnR9gsv@UH{O1b(!%vL;N^<<_!{2O@ z@;CI;_UEtu=0j_+U0?qFci;T+ivHtoKY#zZXlJ2-8cWw+|0#S(o8J5O_g8!LAFd|( zHT~bee*428e)si1|1Y2aFAgk-S>YgWmhwbjiW!YRmBo>ow$Db z>FwErjga3YmV}*ae!Iuhux4~0||F55DQ%(i`{-;fy7Vw8F$WQ)w z1-Za~y1wg|__lw(*5^_uzCX~{Rg+##!(aZAm1Jq+m2xE*o2%;xo{aZR;`8)4U*@YX z`55Q$e03;`YI)-S{(bML{x4Lu+0_pG`f5K;)76Ij&7X&~hAHPc=Dehw*X_T&^OJPi zGH(x+hp+#8igSMW^zq^8!|R`K_V+dK8sImT$B~o@g|0>TJeVuWd*_a__;RiVF>!`@>Q7Q-+a_gO~y+D0tqgFO7npoo*HWD-geeXE0P zheAy^I5llxzDXfehL-7m%D;R;Cz^K1qujA}hvIr|>UyoG#7-ch6f~hcHxmttUqeZL znjm;Sp=4)%3I@dDHDn#bGLi@;1Dz9KtRl<)-xz1#$?ws!Tlh`ec_7WS@v9-#v zHl?Os)4t>PuV2t)q7t^QB2>#FDiuqovF$nXs9}@CB*Q*N=d&JTbUy1bM(490V{|_2 zF-Bje7@cU}TMWB;?=R-dcR!dTVl<0FY%r%w0V9uY*ikY5h4JC4{~liFtNx=YJbrMN z_5xK5Vij#aw(Y?{s#;iF#*(YPA{`c9fhCYf^Cx?HNb>IKm=bm<1f;dgbO!|sTDnYkP<}4xfAWHC+LMta18bhd(O;0{ zhO3Tb3q6i0JY)9-^_0Fw4=G+P!_K~q^&GLvb5d*zgwIygR76?} zF4#K8lELWttdCC*Q+g^Px@mfp7+n)9pPcW~nn5F+Ze=)&)4dP;cc>8> zTi1?nY?-z7kk8_@pP_w$ufe}X9{pUSFI^)Gew4p51_igj)DqtMcfX*b1=9`OqDoZi zJ{dtw7n^ZHW8znr2AzO1M3^Z$9*jvcG_=H=aprN^Whn*JSoHbL%>V*-!15T<>#|uh zfAQxeZj`NI_$3mxU^FTwhKD6>6zWMmr7x(bkiHhzFy3I0(%T;UxuBm5+ApXV$P*OS z7g$h)Od6|bC@B3pM5Ul)gB5I7(uxYlTP-xd(E&kysVsE#BZuXx_LAqB%2e9fA zTpvXjfE}yXO%>7l;j8?&AEu`d)0FD|lH&Fhb)2fXy6*j#s^7LbaaZmnZ|&XngO}1h zZ!Is)mnF@gUZ(u}DnK|8RL1bC932I_H5Gj;>(Z$erC66fA<~J)7 zS1+g~#VoDyl=YZGx$)w*9yv*IQb)*XYs>wV|6vPiInw&aykhA|9cx3^M*|9VogF%X zn0m_EO4dXjDXwix;?h<-Esd;S1#)SZ&txBpmOaS`)v{x*|AKqJs%BNkUHxzk-dLc6 zp8~bQfuU8*b5=c*6=-vURi6ny9K%`2)7-WfXyC)c!&CaSJUmVxpVEidKhN;1IjEUw ztO)z91Db`m_~&+mVKA&+pq2$<&t?-H0^aM>K-12O{rjBO)Ln(xnIoFahDZ~?Vx!TB z2pzDVp%vvnB93VD4H{v+UB;v~#wi-T_2IRl(TMmK8olRXq|ukCO>Npi#(xoJx4lef zWJG8obi%1LR3n|l#xboM0b2z-$GVoU2S(Vj*J89P@g zAJFJ*##=Oc&%;Qgvl(yEy1hYOGV6#*W7>)|8WHzstsG)6Q$x&c#_1%DoWx3w5e?-5 zXi`OT3cSsTuT~h|0t1(D;j0_r+Ey@ot^KZFM2$Gz((9l`ZI(3EOC_DxE`l||Y~bjc z#f5n;^V}JB1oB~FTX_DMCbfm#^SoMZE3jn5Wsq@1TOCart~E-wQav}cSFA12R-|qF zLYJ1f(Tw{>gXp)?#QG*ALYT{FP@<#Fo<(-LjO8EqQS0??c*(q)V% zX_Iby17a9tG(Qvr9gap?WS4P73nN7+rLuOm}s&Z%E zu~TJyGU;Ax{fPT$6FD7UTdFNX681m)z#J=x^-#m+8KTnP70lK@1Uw)uvKphKjw=QB*VHqbQdY^H=_YWmI2Ljx=f zu(`)*lo((;%m4#NOT3$E`mG>hv0^BKGcQYGlRr;~cLe-mi6SuNqV6Os0;qYR|^%46kx*iUT1ss@TN6C=tR zu%ONOOpD7joutgZpdi!Vy_DP)G@Gq0meN@Y-O3aiivAi3oyqibK|dGtFH}$hH6o<9 zQJR7t9v|Y>khwgpAOAuZQD&~d2JX@5Y{r(>1ezjs>l>jgi``A5_dJX=`VzGXflu;$ ziRRIls7?5p*{5j#_ha^iv^1ib*IfxTA_g<|9Bo#bXf#<=F6zjLQY$l*GxBStMuoNp|3lyA;q~EB7gl_^s%cx| zTWH{kv`y5tuiVlG+YAEXjsfjVPtz(a2Q<5d1&}7W2s6xGM`gAgH%?D#9?0@oID(QxGW8-VC(ca zfOV!gf(Z~{_fi|e^%xdMWXS()!wiTH+eeM7;5GGTzf4CQ8-1Q z`3wn9^%mE4fkg=Hm)w`-k}17@dX4ep6s8gzPAmyp`Krh2zg3)u|I+HSMpl!Z_PzDn zxtd}VtW?^t_b(65N&Y27-t+IjPC@zLcNOeuet5bXVvm>bQSJpuwYzw*Zv2;(kcx>m z$aOBryd+Z4l$m`6v(X*5Xnzn9-4=^i#9o$)$@cjuVqEA2&ell!nv1r&@w|PmWKs#69Jm&UiK-=Dc(ff-@YKh$0 z49^?EsG=QHh5bv{8PHRRkvO??&`EJ}r8TW#t42NRDw4Qi$jhJjKhq_bX>Ts}Q$3Dv zda+*RGFMlv>O=cyJF31x&w1kYkC%0-Gw@X#3dH3o#QTMmSQ%1PU8>r4t@Hn+%^5;= zQ_A!{h`vlQi}Cus1ridiW6wgDO#!`rS}rSHHPA!RK&7~-GhmukWM&OH%+*MVb-t{! zx%07*cVml+HThUq|M`k zESy~#TQ~6N1yEL#HJC*X5Nln5f`?dR=T#45tjdvvTR%0>94cSy%X#3!?!cE%U0%;u5bzc=_~tS*i;(NkhoM8&EuksEg+ylzV6zKwF!1QbS&Isz54>`2Zp$ zhLHOLPeSybhY_MLQ;b|9D8y-fGekEsMpha@H>4qyc(7d=T{eqNH9(?0RG_3|Ni5zV zCdef|TxH~C-K4Ud;{5f8q?d#Tmi$Gsvr(L;`sOX646Ygtnl?(@hU&G>)3zTcl?N+a z?e9&urPu1Q&GGu!>#T6C+7(V3dAK4^BCrj-292xEOBty3WlNaEX|EZBMnX9uRaX(= z%BFP;>!>e=!y0r{m3Q`VtxHCjO^I1yLL^oT58<{2yA7Z_Td>;zI^BZUMMfjSB=={jw{#H$~hmNWM))TVh@WArg2APU&`=$ zezElR9#)&S+P_jM;ex}a2Ad5dk81N5$C?FsB3p76hw*){k;g>pN4jhRSc7nL`LD)d>_4HG`Q%ghl;-()dZ|a; z<+<#hTlK$f{-4&X&JFEcb&oYOR0yxRQj z5u_tCkD&4|LFYK=*m~;UMrU7q>uq)r^yqab-loCWgBPD_9IP$h-ro3q-Q(7Eji^mAiN#dd&Lz8TXMb;k~ z&>()l%=sQen_==8xi-ZHo8kG?U;!|7HM80?(8%jdMqDd{!}7P=FU3pvAM;ppy(1kmJl=R6EQfF4Cch; zT28#0v(v|Ex)giXFPsb8YUfr5PkJPY`J*GwuvLtq(w&RLX0d~+mm;I8Vp2fvu_V1! zc=n{0Fd?WY;B zs9xjtZwv#B7)`E)Pk^zK=2#qJoA`jw@tUjcSJ!#3*rZr|&b4YqGPWJ);x3kRfO8U7 zZOqn#HSA=TYD^_6`V!}l+htO(AGWB(H`_-4^MqrtR%yM1LT57lT+sh41*JyFAkJwHeNSm0o4ca_^_%u{LI2$f>bt)c5jrO| z50{6_)qtL^Ds21qVxNH;%6-xZzn!fu_y7KWDH=3ib}VH@uXy!Y?j z1?|n|`0a18t7DCz7XK#Hzvc0jR}nb-(%b2jgSEet_^vAl@Ny7=c-%*%e zf+;0z3*0YoUf^%_yo&<_f5U>jTF{^J%sXI?Mk8=c!a@gz1JIj;47uq==;bO$Da*G9 zyu7r3W%I!G?-RcuNz=sVnpjd*a)qfJr2&brYplO1-x|Lvp0OJ9@a0p8>C^KR)8?-f z&GmPi-kL5CRd>lxc6$5Ixm6kAzZQ}VtX!7>tHYLGHa37c6YIvH5|FXj!X~;k^ISGx zz{JM;h33Rxjkfa3HqUKa5`<*<%gL!GG`o{aC7&C58q-M=Udqe+2V$i|Mc2}bO4lZn z>=`T7g!4|ad8`!f4~VSxBKy&_lYo1~&d_P{W1@ecIVli#=;XJ>xd+bEV{`uAZw+$4 zoAh>G*EbAb0IX@xHxMK^BP;LH*CCzu5}-5p=t!RE3-!M3q}zMotW?mxxZEcda=vid zd0=WfC+*SY8+d~U?xWL=SlrwDdJgCG1A9_IyS;y>RL~ddeJ$v{y)Rehw4HF!`(aF{ zo4q`H`k)UsvwPc>g~w3p!?L2yQ;F3wWS*LkUS}X$V)Z1}CSi|0GaYA%ZEB)&|9!|l zVXA>i*zO_pdgy4vzRcUywA6XceQGvswV6Xbk+dLvI~}pjT%L1vQ)1oO>-uuawm-8S zHZwnnah7wx}<>>2}x2wRKX?pQ%>D$EAG&N>w+Ugxu zt$-(l=H($g%*0bqgpFlE-v&1b*sNfxR$iFw=xn}`?`F0-W^MRrKAVJ=4jMf`!)HxX zEFFZS)>Isv2_KZwmzVqUCZ}nk=)Pou((oM&k$Xu8P3?9Gy5iVFz6_}Non+zrt?bUSOd_VFW%L})D=A{xhtigA82oa2vW+Z zzXtOXTnE%k+31PA$x_br9p)a@)B5 zOFc=iB+n8opSObQdJP+rLI36bZ*|9a)m6Mx)$&UoOJwq_t;|X$DZXx>oi3M4O7prc zo;}z-sp3F&f3yB2B&^+ESvZ=O@F}I|PnT!2qMS!*A04gjMUPFEp}2GMYVut>!0hA& z#gUhw#7SnKVuk)!Cx6@UDOk#N!#5k-X|-cQb) z{}Ov=EOx@uC7^yj^Sft520Ufg-9M(VOGR`&m5lV%%|Je0S8pDs&!C@Zt{IJYwf`hwb}g%I*>$~1PcJtcOOG@x_^X-2EG=wKwF(? z%kVWy>M0V1Fp@>-7@7u@xuK{DOF**PukK|=!C>#|-d>EF)xBObp@qeIFw3M=rwC=% z)GVwo#=8MU3lhruFh5+4!R6`k=_5TovVKhN>9&hpDB6BX!}`pFjNnFVa5=ZMVygpj z5#rp?P#bl5K!ae9j5u=oY4vU2W!Hd4sveNn z9W5iRRV0OG$F$}_#+G&fQAZ4ejK-u{p@v2cv0>McX5-d`3g%fOW~8;c*0a#KK%!-P z9!AG&aX3+F;#5JdY$y*!Quoc8qB5QG<%{O)Q8qybTj2*_fi5;q1Gya5QU>Htt}!Uwtbp0x0w?N zh;fi{hlX*p3ebR<&eJUDO(UXBl4of;V|$`9QTy>Wjk>_FW)zx8PCTTM)wVI^Tbc@3 z8q(a_W2ezZv>kTQ4IcUyAFxo?9(xlZ0}Wa{H#B(!jblr{EruP9MgvppWi+dUvCAU9 z$!V)k!7 zXE4*qVt$LkOgm20-3BxDk`>@}=M83?9>HQ(+8qWn`Im6UHyF(1h@r`Nhrx^!wpz?@ zF_>vD3fY=xr5l4|C9~A0n_U-YJg7!g>r6~p^&E`i~zd?37<;BlBC08N^DE7^|7LsXF*osk#?7wIe{a90j`T1j-njrH6kaR!4+n{^tuAl|QT3gpbW4i7Y z?Dh2#J!5CDP*9(sST~W6>1MBY0&|1j#1LsOq5OQb@Q_Z5h6XyK5Bh-i5(GQj2daX4vUIXq5l0flz+K zKA^pXurKdJx|>CA4+OWQ9LMxyRlfJ#AG(d-iL zrPF9?*?x~N>_^zvH#7TtoUpcCM8A0e*yp=9v@8AH?AYOH4~KBp(qwRWzJ<42PqN3S z6L`QHO??Y*;SC4qBnu4R;ebuFw^(w619XxFW4sysXIOBv#dMMd+M+$Y@%k-;#OWl= zsCMeyXvn4F7SCFnsKtsc&U?73pVIxQ(MK)b>L09Nu3Oe%X|)M^xEIzIB~(hYV@TYX zja|@1ixX-A$$o-)(rO=kJCX0XShQpM!4NpEd<2#ZCshq zuVwVE@(Aw+7S^bBjYoLc!{yo&%rIiX5uVk9K!V4QMw)Lion%4Nq8=wt2w@*T<^UbB zAW4hFgPnJ9jre>2TrO7L!`moOsxz$-vIeM!_+&U&2t!w8=|FZ1cS7=`m|}A`z0|g zq3Nijv_;dZJwod!amu)y6RM7aw6+tg!Ok^p5@)G9r!Awja&0r~Bk2rHY%HN$X@S$I z-UZl2^Gl+f%yAvX6#3khq!A{TP<`8poK}c!Cqf#9wxeodsc6@UHBBvD>?djxeJ55{ z$~TFdjw)%TqR|;zgGAp^8|hURK9)9(hKWfY)R}2_1qS2dI0;K)yCj1__NEgnyTtSx zCiZDE1?pU>No0vjmpJJIiz}RK#%a|s053b~v?G$$PEnec`?NAa=5?G{(v~(Y>AY8t z0VXIYgeRVwpI+Q{5#gFmbx`1hGbuqPmvBs9fvyuROtaJ86x&9jl&&&MdkOq*@e^;- z#ix2->&1;8Br9&S`WNfSJYsYID;XOUHblsi@L$N-vNK+g=?$NcN*gh}=J=~-xHEO3OVV5pY z&aMw=Cjo5ub96J<+)N8S8;88^w5iny5^P?om6#RUr@Luw`~aTA&%gjX-?&A-sR@0D zW^&i(F;w9lxY8P&jpU)#g&=sEhSH-t$MRU_;NgVRQDYrQ1kFn7k|vAQNa6O9*cd@J~KgW<7ilJAc}nHfT?A=Q;=%z zc0n%}51m?qTg=#g50AWI<+0!4t(kJ}iBcg~i9E>ZXgEODpr%-}`2>Yqc!vzaUV!>6 z8iSr4v7r##4bdOQfSf+)oyB%MBHbWr->*)z~7V?zWXe# z$Wd4FUfsNL1z*?B(pt$6!qiD2YNsd5#FVU=L_z8ooVN#K4Y~#U z#mYBF$xXEKLdMzWFhLIGm>GjH#A0h_|Nc)b2Z-8eDa!c{SD1 zY;kG`c^_YfxG;iwq!DM9U)5LxJRq*)g_N)YUBwe{F}mD1F;^K*(?EFzU#tNV0*84+ z`j{KKl3Qt$%fnPicv4eNtihXbae0_iS{@%J`Z!1aYM%RrCTK9GOomWT0?N!&Y88rz z;g-mhz!c(+utgc^lZ2E|DX8EoB?yI>#MvQ}8o511Bnt0Xpvqypq)f!D*>2`2vr>Xm zs0yP$hd83l_Sph{eAX|ciwWrzLU6w!WgStR+n2jcL&`Fu9J|J2$~>YRZO+`Igb@V` zdJl^ZH)r0%A}r`#ENVArzL`Z>(7RZK1>Is%AKU~A?5I}K9;Gkn;o%`Z<>m5_Khj_P zIZnl5J7N(veL1Ar<^xAWV-#Bli%Ky2*1zO3s$B!LrK3$F8UX{$iU=JoA%6v2fu?TX z(q+VtPZ(|PXwl@+h=Utf+`afZt#vel%2-A?OApb^5lu)nnmPLygIXTZ7Dp4qfm)*9 z)FHHBzbu}mrkr&9cMpZYyp{VepP2E%kBwR^Cd)_37!;5+9#F&iIM>nXA6lD&tx| zTal`AZ_%(o-Za1gA2gH7-{@$w(kvDml%^jh0T;o%30P=S2Vw4L$&3*rx@cw|tmAHcpIstL z-R0ZC{8aqyD5X|I?kc4-%V9pg4#c*+5v@0FX?NYMW5Y2EX+(6ilPsi$DBe64sn+Zf zjdn5)3Vc`|9-dyR%6%rjir#(J(>V%t-jnYb)!EX*idj?=anwYgjppG835ZDDB&Z=C zFdNwAF;e3>7}d5O05iUDnmLWDN77IqRp>w@No@d87X6UuKUMIh@Y^0S#F-frUd9}E0eV#QI zQ#UrxA~iPlGOzHgB49!0ff(tsrpK#6@bd9F@i!!#7h)m3eOOIKr&;4ZJda=rvJoQC zs~wzoVF;+gP@MMRWdv)bY0v`UQY@cY)bOIR27TVb$BBLF#8WPh$W?-mSweBz8i85R zW*BUQ$PD9chG90tSnFUG6GCKmW2}M00bFNj8(_1Iq#OqDXgAmIeOc5plRZ^KUjt1d19j9HNgFb|{+oOtdH!mp5lI~#u| z*&`A`N3hQ{3pdd9Vd~|X^|LKrGOxW);k}L{%$_S36M~^K9b?NL7QG1@TWsFLQj3vC z!yc9oG3Vr;Yhfy1R0o^uSp^%{?#;ei^v&9X$L1KBjz6}?#^#t|bk-hoe+K3l9ou7! z^#)jHj*+*=j;p_Gj@@XQv&Wn{w)W;2L4yzNv0LVt_S!9b?3Ou3C+)EZo-1B*(T z3e}>QPu*N0@6%s$3t&F_+0G+NL_W{v7i_hMi9ePPlrL6Fbd@`YZ4S)}(?i98@vyy# z$Gld?>2&AfA5B}$B-UYmZEtrH1574|T9Sqb$fE;HA8bRU(q0#H^5KfNVC<@&oZk;H zx?{FK#Qe=}7^SuZ~6)h{f zp*Y|ywb;05!Qt&GeRg91OUH09$yGywz~JMqoK6{6q4`5cmebe@9NL8Pc?3b<5UN|C zX(vK;-Je0zJoVUoT=%d^E?cgBF*z)@{WR!9@#$o_DdpAdd)3nDp{opTGLmpO%+j=A zf@Hr|HQiKH6x`J#lj~<~ugopbUd31t3B@!+0KmKb7Df=xEcRGojGa>w&OjUlT7*df zc)dM$d6?2C0+R*f;@(@$H!#>C;KGojsoi0{u7&OVZ<>^}@#Im)b|;>F2b}ITUYa(U znN?$#5OjDpR;C)u{P$)VX1RQRBjn9NK{n%1-HQe`u#o1Gl_f)1~S)uMygEHAOK-TdpHfg{pUbRnKAW zFsW$yJ}zm9DrRfbx@{8xO#SH%n+%Ivy=G~egf1Z%qW-4v%_h~-ZK>Ra3Epj#tuHKd z;_ZU67A4mHo$+rJ$JlPcfU$blZ?K8t#7~_*)vGJ8XmbXuue*jiZWj4F)nPTezcB~- zTVwQTtwq=G7t7zkUoQ?dIwDGZ#lJRTutx)|o$)WQl%L-(j}ohbylRjo+TU6J@@JBR z{H;M1`&)LH$0~p2Z}o&%IVz;{|B5vYJbKC}Y#%T2DohaXvwp{$Xz&~LJ8u6Za*b>o zz&MrJ%iQHN>s!TW@cZvs5KK%F8VR%a1SU@87(iX1TOk5s{)CT!m^3AXSBv+R9#;`t zK+NVL93j}m_vtM|7c1_=^9ZiuPlzZSMsVJRx61Xd`WcBDG=T1G!EOTxkrOT0Z2+BW zL0YuZQ(jTqC`{~IsoFu+192~eC!1|7T0}M0D$RY?`Xnd4R^=%>{%Gy#@^bYXBP*3A z@eX^mIq3X$d-O(FuW8+wIZPG(aae3He!GIGbwVbctq(6(h#Ny~T)9 zOs$_>W+`-(qfD2jO=+H39o5wYu~N%|fnNK$VNvhv+}f1X0FHc5W0I!jmy$`Qtwt7R zhXsc@WrMqCor}d-R&0${ZmwSIV!dW7bW>%9YHL)#hVgBrjBzBpU{owe?zNiOVo9;& zu~p=@8%k10(8|ai&I^{e@%2CyX{{y!rRj8!OPgpWax3?)~(}3b^-gqOi>&b6u5}HM@bC znQH~A;j!x98q%nLDb-`+LH-sqG{723ZmJ*`n0~qn(Y!oN^$rWWk&v5V5uw!zv$0#J zBA%$C4e@d_37>S~GllG=ucc-sXglGgxhX#$aXezHD?&4m3{IcKKb~112L7z6kwy zd`}ylFgooe_yC5@!5y7vT=G1y1W&gqddSkiBV?4|c@X#Md<8Rk-lQy>zdL%h*c%oY zoqR%;goZaVpX}4g^G1et%d;|o`XG|dG&(&(!h~@mD{qM-dXnAkVy}VrQrDHGX6CelsF;1ck+-h2xVDIA0>)>K$ke7y9MS<6lsOkz zP*?9P3^s!_ut~ds$cohpm)i1xV13ON^=lRThe@}^z@D>7<(W=u8Q#^bs9@ z)G6#?`4HnVizxr5V{8=B=_8P+WrM^bZewzQUtbyLhwI3(k;$Vcs;x5HO&^y=O;iq$hJ0CrI`kM0M-^7(t;ztHI`1<{&Ae^Jwhh$etX+D(w{}MCzaH zVP=0A!P#IEq^)(=f-$@Zw+NBA4=bxv$M(AWleC{q-RAmf1!;)NN(6W-WChzy-G^3y z8+aMPD8+8udW3<9aN381!uYi(ll=r50pYv@bG+~RipzQ(7q*czD&MsC#Ll0YeX$OZ z^7z%LFW<2$K0{^jQY~lxjzugHj%uii0hGTl$Z7I`i)SBj~_k|F!Tsf+zgg)7=v2VH&oefwdC;;}VwTDm}E> z%JE`K)nW~5!DXBAL<_A!Zh8tuUINL7K-g$vy*jJf?4hJa8$vX*YUh9Kxd65oII1>J zB&eWXwpMGsDwMTNpgv|L&e~=*6gM%F4OL4J4YW)OsU;#iDn`(?B^DJijjZr<*=CA4H zb@X?#ze{ZkRSiD;t*rLfK?S<^(;F%({1vY+*PT8S zZ>_3d5t@smUd3Tm$6Nhv@vBo-i(1D4OUH&@!ox%g1)iYhdHJ8E(__( z8~bOoxdY4i0{Dt1Rj{KrGB!7hg&bkR46@C|90kLdT7EQ~f^LKH87v@?C7dJ<#KKiI zC7SDkArj1Lad<_i9#;+!5LAgpY@xmB5QmLoQk4mm6}m4Dht7#5!a)lXs_?LaIfF;V zt7$yTTC1=C@vX&4z4I!IdRbnT(O`RySkK!~obs#PEL7FBntW9SZvU>zb%I!o6e&uT zqoQ#|v6skHWyvh!UQW)7th0GOv-7Zuh*x+um_{t581`xX2&KX^C|3WFUgomDO3o`0VTr~(BTE%ks=-X_n zlLcv3-8re#C)yJzE9*z|3`K>C8)#<~kurHo6*gM9$D)!J8=H$tVd9aqs4O-V(hzej zhJ1oZTrB21_m5CiMnNKZcAR{s6(ll9YC~Ci3c*0u9*H2GIt$#OkP5Z)Ra+6hP0=F@ zBH)(EQ(9qfDYTy{!MMo%=xD|Wu{AKkNQp>cW3wMM&DbGvC2&3PjznK|UF~m@DR^X|cuFSoi1v>mr06WRZ=9 zYR4C^>!m!su$)ZdC{^jnB9Nj5MeeIKuqm>M=BKN=p9dzgH;|oq490QYw2Y>)H;lbm z?2KYNah@x$(IEEbus4RiDeUdww93AmD(=n*?o8m^0Ej?$zhG$na+oHm_vx{v!;~jz z^Rl8vw8bXuDIF@_(ybVIG`1LfYU>S5zoee9O+zdzCRQ!io1JiH(bjP^fd&f_iE!qK z4*xCNYsI9XQ<~y}n0`LBs(NZ#VNJXe2`1A6f5<}Wxe7;oL^xD!KB;12Rs?D8 zJZ1TKsf)#M>p>G?Qh*Y}b`ip~S!E>SRNA}*ZI0~qC=-FX#rADkxiU4^<4Sg_U<*5F z-Bil8n*^Q#F0WmA5u+fBkV5_XGv0)V->3x*<37=WH!t_LvIPEJf;g(HrQ6zg3cq!P($RZ`7}Gv80dAp{}Z}K__>8FNX|y zGID~k{aZ}>PaGq&3h5-FzXR;5VIRUndU`IaQ?!2gD*yIFdiXG1;)id3^XL3k`r7u1 zm-Nb)qDy^u{}NxRu}CDe|2x~?tbc?3<<1g7qP1E6uA{#ThjQWXQvVXSAn0G7o2yu` z=6Sl()@glueR+)!ALB#4$B=g?CA94p(;?jNb#OE=54Bj``h~mF;Obh<3P;{bQiVgd ztrWPnh^nBd%%-)2D{mb+-NRTj-w3)YamR)2@|!FwRSKF)>=0!SGbzLJpI%g@oJOm z$YFR9EnFu~{qjzz(iM1oUh>0h+2#H~A0yxKxIMH=N4S0xxw2q4yh6H7uh!G-A|le0 z(sNqsIGJ|x$t)H1@9dtF$xSUTT2DQ{SC0MFgImoD?*1zUEzRFI7t~eN|81KnO`r5H zwMFliAWw7|x!0nyIqwx!|CaCFT*AZa)2B@UKR$#yaYP;0Nk_gZ2U8iQ08B+ppgB(E z+%G&xK@pxdTXh*gqY@b+B&PN?Cm?wuONo1=qy$ z`&8I5dln(}32Pz#rYDNZjb z%uA`U^%#T-a)2I|Peu_U6~+g!pv!Ye>FLu{d6KH#9XSA9&PCeWupLr5sgz z*~OqtI#X>c<&fD}$^o-))Eb%nt`5|Kj;ej81BKaV3RXGjHo&4Kepi$Iu78HTMz>Y~$P94S-H7}gnX7pB$hrhEoov=WnpX)j?6*QH5= zv%-Xk!S%fPF>H|>_VE)$KvK`g=-^qUGdZ{(`I_YgEZ#Ujdy0l*r3jw5P1D+!ab;E0 zi=D2))~CJ5ZdqiN%f^HnSC^NVuQtmo%&zKP$zKg>$o(}wfYU^+#22$w$}Dx^>aB96 z`)XcD%=J@LTPf9M(eEu-^rZhwNZ5-s?fu0F6d_mk{+@|Q5fIL_V7CDn6FJiY{mZ79 zZuixdm-Z(aFCnh!5oG|E0aeaOoI0a>7jOyicFw~81l*iLQAA=^eIGT3%k02tq*GA< z(_R8})U=xiqJjDq(KG%5z&_GHzazSD9f#@Jg2hIAWybDasx4jr>~(%oM7uj0#x; z@+|%oI{BA6PqG7J*52a{p(g&i&byFyE6+bT*N^9Y-I3`eavT!b^TiLO6V|`5Q_)Z| zNS;l~B32hcW5$-y?Q=uXc9p|;jr$la4$i$GzhXUR%lL`Jj7m8oNpKyuN!kb%8!Gofv9S+C}R8w!m( zo7`A8I2$|xj7CSwX$3Jd3%zW^mS-1*BY)O|Ijy~%bx%8UdeLWt3DOLokPf!=BNw0q zo&;vnXTi`>xbAs0J4fNwTiP*;ku_@q?{Cp(6aJZ=ax+r6?eJ9yqs=lWw zkPHeWg9348ZSQU-0y{XHfgmiZZlba-iTPx_7LnC$%~@tw=O(5M6S)rRtNP!>A7iaF zc`a|+*$T6lyB8)W8I)yqj<$;|$bQdSj0yvaM9dgLU+Uc~<}YJFg*Q%(4flhsI7&Y= zEUG-H@-I!}b}bqtiuF==_V-_u6B-+5{_sgdc8@9W9yE&+_|03ZRz}8R6)Iu42pf6H zI};5V!!$@(dpIL(;tVhyhgD=BB#d}aoFWv(rcwD1=00S-$wB0wi*CgqtWiQyR;{W! z?y^8+B!VdTISRhij&r2J*})m|7`#B}U{fACjVV{lAkc2!nCF5oA)wsT$fJ~o4;{)K zXO#9F*)v<;@KM?=XY6qA&6kM~xjk<8bph?X;J)$$+Qx5+7#NI3_Ss2(SUE!E7VJHy z#lcV`qhFSXkmAbzmT>SxnB?JZ27+!8AI|y}R3w&w;?In(`L3E*}%sp(}a-Lnva#E$-%BEb4 zq8_y>AvNs67$xpki@I|uuO@5_tg>?mQ=cHapj$XP7{mGH9J??9LWMRJ^8Rm@mPPCg zJnh2#wcQIrMGzx7z-tIhfFrBs($(%mcfrNKa9Xi$b7okQSChO?qC0uDFj1B+gYwD%k{oqiT2rr)!7V|Ja;w^OzH4eyGR6R%PE6DLkSu< znABR*Ypl>lXSi0AVylg$`bo+9?(NqsBt?FcZoucRk`5dH8$s_gBW0Tht~RUz1`xCW z6;_aw)`cM!*^(0;vv5duF_8@bmhEZ>lp0P?@i_(Ivsu+P<})9jX2n0pE}_EsfLky4 z7veQ@d{~Ko51xmqO;}t(m)9w!`P1{2Nl%lWK!AlLIh>)Z7>m*Ud{tp7IB=9VszoP? zk2iJ(@Yf_<$un;ddt}v{9pa{z`5}JW#P^Qmn$CB$Gb~kNcQ>;CG>K;|6H4h_!tJ%Q zRKUSe{4Wml%Z!^4UWuDnkRaC$I5^8ThJ zrc4rrVY`>D0WwptUeaI39e*oAXQdjvs0GV5nQAG^IIlO3v7x*8*!flnhHWgIBZg}o z@I`WnWm&w0EW)|x>IWlR<$_znyu9jk_4J^_p_$>)RYur0LE*XKxC1AIR}Y&7v49U{ z@Z{im2$$1Sq=-3jX%B`^H}4zkV*}kA>+=B&@n=bJ5u<(6)8ZeCIoOQw3lN?gIP1F{ zEa!?)r|-a{I5xtz)WZTZ+Q@bEujoHQB-s2e!#e0>s4#mKY@5oIOLe(sWm64d%cM8Q zmxoKN_th=+l^w#9Sl^lN)L;ix>R-zQj8rf0L%^HGR~3kq7%V10TFkK#wTRVcg%?h9IuAJ{Eu zf{_?=*dN~Dm*eImfhn1f9%QNxEa2*)t}5B=wOOP`%w)NSUv0tiN0yZ_27-%wRZO+M zIa_osgCnoIQev>#pdQb}AeD(Xht>sMT%rrh_)6gh0G%_;h5&UJqQh)co%s$vSRS_X z+YTN7S$Yg30;TK7qqniO$4V7TIK#~Pw__x1@Jw0m#;8y9+XemBx-rHZ1rcpAm3d@M z7%_^5JdF;+9ifInd_d<%{fY^X3&eD(2!>bCuUA_o$cv)Lsh4Q=DYz0S`+|{h1k$9K zz*vm0Y1NfRaKhjnWYd+6xFZaGIp}X5A(QLyF(yW+$Fzqj>~(*1*t8-Ru>0!I9}Nq) zG2OymjTyDsNSvjLuBS2Vo50N02@w;$Dm=!P0OzN0zZwI8&A=ayVj>pcSSWJwfTz1m zr2WLnjZ?X94D+DzN*+CjVdO?IDWJy`aGAB|m);jw0K zH-_N28$7d{tiG6-`r)R3#DJXT50?dB32X5-e=Uim#Gw7;yZAP4SIOW;*(Vm zYR0xU?d&t*!@;ZH2Bu-c`HiLu-*-mf03L6o-`_d23lBH2oY`H|!oxEHZ(z}QT6+VF zVB_-pSajFg^JEIe6p}hw#Vr7_Ls&Ac&BK@1tJKeKg5!@2$lze}P;5pe_L$|Y%l0J2 zh)1BYqZt@CIcM0LgY31-UNKy3R2zQ<4EPhEi=)_Va_$kiU2nfz05j33Yb60NQ;>Cy z$V`kCz#*1~N?!(n#7 zV#82n*S(5{E#R-}0#s#9Dkl2P_&+vz)oN9TnH2j@?~10_R`2Mw-v47G&2aOeYS2-ea8F3W8RoTyUo%&QMP zw`e$`#qohrftwRWupGcc8howUxo$RjdaOh;7tIQDX*DOyA3eAE;<2^C>_U)O6BR4V z;>w!_%N<6FPGCYNkF6D3cCgvRNKwEJ~OTjGmRJ0W@y;qx$v?Mu@%`PZ2BA&iSCoIPSU>KL0HMhj+b1ON?(=rEBZ6{H27HGgIJG!$ULa*M%Qcyfu& z<^|KuJD9|uCzog|@>WBsh9AtjBbjO(07tWiMOZb8$+kaKN-WDphwmvov89;iG=0kz zwi}9gO*`%VHpKLcWQ$*wLKEU|{rL8}aIV?Wtov3{u{3nZx8 zu(#*-%0tn!2AADyqwT77+oV_ovsiFNXp)5I)w`K8Nq|roiMz1mrZ*Wx!o~X8F}5w{ z!cd-$wz$O$@BHo0$kf7crw78eils<>K*QtH2{>+HpJHV*va0@^5!NXC7DiCj;cPHw z?#_Ac`(ufONO&`x^NulcCp`HR=J&-9n^@v_rSA>l*6~W;8^W!(j;h1_S_O68fvHcx zhf65&ksTh04%i&S8#wrYSm49!15C5}Vp?B>JUYVErSKRxgdL=NH|4fFh0Qau{Q`$} z-UySMR-8-Y4Yd0)Jx76--3|Y42ctWAx{Dod%01!jDMo|DBZD}5%!p)sKeCI(FqFWb zw>GOOss`=~c5%r10nCcwB#h_%Y={KNi)z6+Sup1o^XuuwZkexE66x8LvI@)tSLYgm zF^@f*7whBt`Q}+0K7f;hp=S_YDflkKEr&fE9qcH!p$hv9;f(M)gk=K*+RQvJB0LXa zlw&db(QG_A`j3Hbgk_7@!1gg%6@0#NRfH5|q6e$Z2y?uX;ppHX1=SCBNrHpdF>F(# zM4<^0hUXC+QMerpPCqP`*ZXyK^5Gy?;W}JG!EQ(}=Z8FsS$#3EbvleN8YB)E)Nf%N zI0l1&u`s}7L41VKQKIanzm4&#QT=dR(3=$5surwX-moAh+zYysfjp|1uJsq51Ge^9 zZCW{lsY_;M`OIN6U?u9zH$E`NLYmfEkfIe^5X*|Kh@}SGc8hR-l?BE8s@OG`2CT2x zykX4RrN>Co*Nw+7>ltiCtO~uZtW|G!*~R?M#Q|pF3?1`@n|fhaoM7+;9&xDq z@&}%8X5OLHV3t+Jq!okn2v!Jdt7?J|yi4`aO?N&b2`Xx5%D6kI-YXjrv+0+EFo6r& zcJ-JeAa|Web^P`T?8x3U73V$<$%! z6xDSdmB(jD+*at2r%=_Y!Qhl>?NgY68Vm#@RBuAD6@nXtQ@z-5hRwh$j!&&tCQj#M zM2*U^4z?+e>~g8Hn!nD_uIOG4XnWX(o(%Lf;r*iwt{uwA^Na z1bjJTPdNo$vW$38&+p zKi^`vI)A%$`z>w`tDb}gEv7dqv_b3WzE0kU`oVurjp@uHlLPNMvsvTZ$VjhcWIRWD zz2j4-J#6`PBUOF4eCl*b+p_BOBkwZE8CcNlDG^>~*#RMSk&uGr%2}_RpTGy2*h5+b zmR^_3Jr-jm2M=e@+C7-|5-cPTt*Wq{47bF{gpX#h_RA8rx1*e?!dRr#^H zh8&x-a1e%9Fu0bXTR%Y?GW|2ECN~Xoj z>t#*PpEkSZdAo!N11uPcG1>8O!mSiB-|pmn=Lx^1{H`@;{FZXe?h_euTw{Y(WcX7F z?8w*%NTh6)k))%26%k3`_tUX(3U15dW%d~nV94Vd>o zE&AqY>>0wz!EG!;HX5%wG$scxFnkP7FswNC1N-PQa4>T+kJXhKuG%pe%!-G!&+0ed zWi-w%CZ{vdK;6(#UJjGEMDd(VMoFFLD5jD_>816c$Z&pmGx(ZO3LZBxwVAjwv_0nEbcB;ozvR9Vnp!X*(qVNL43tzpkz?``3%!#(#P?sK1`7A+W z2B>i`7=Rl}k72$H0)qt@D!2i|`ZRwzMwRlaZ(}tUw8b8tfkj1&OvW$esZAWA?no1O zQK?1?vGkLuwK@V)wxBgnYGtCIbV#t2l^ZudfqIh*l0{%dq9o8BL@I`Df-TE4+}*U<|lOB@1eHxD@8YV}TjZR$35lNx?wQ8Q6G@wmArl2#vxA19y1y zf{N`7LDdD|Rkge>5LgIPcv@i{qCxfqYyV|X1ZL~2tM7X`7+Fi)yo;;uoEzO)^{VXX zH(`QEb$?LYve?*MCt1F**)cTZZW~i32xUuY*wLZh)jtP2eJb;Ye&Th3Vb~1GA-sZb z#%k>i8zuqPK+lPJrh&)P41F0E9r$Yd_Wjtrr!T}J@zrYeO;KTgoPqTVzs1B;b}jbn=4^$*i+*HBEV1SnKHK+RVJ_kYus|qkio$=Q$`Qn+FL- z5+wGRMd@dFg<)_Vz|n=Kk0;DW%L0 z4=I0oX|H$SSCF5~o2{l!)wcUPLck}-67PJ<8Ao%5w`gzt{r(0f?a4F0#X=*SJd?^;SaVoPXY@Id$wfXt*REtZ!GqEkh}fKyI9P)h!xPjz7NCmm*JMn za*T}+afPTec##m5+lyzlgO^1oot_Iu^BpjXjd-FICA7E=+OxdAbFGC=` zoSa`O{Ito)ARE?xN*Y=ackk)ZSpPWr?pxyVj2-W zbgsp;)8yS2TTCRoALZSjJ#v|^a)+xQ924RA%I0wL#eLhq<<8pVRva@%bvd{P8!oEW zs`e-?qVI}_q}p!|)i(~T3+uib;dY(fgnh+lB9H3uuQeH9A@KS^gK@N%7vV#YE611t z=P~kVeKps?UR`6UhZyZ9Y9%>5I71x<7#$@l28^0_7{xV8()APfX6<@AMRF4>@jhQn ze7r1k&|njbJP)y4*E}SGzz@qxPlniwzv9fK&)myl=!FX!$6mr6p%*%ez1$OexgB0P z8G0c^Z-!S6Fgl98+!K0nq4hTQa!=@mj$$wOgkC6KE(?S@t#x4EP9jw@Uk8;aku<7y zT4s4+t64%bqpd<5x0TlXM7f~CAz zbO~}XK}ao>vUw>CkzK$O6-rqGt-NeMhe=u3ZY5=tM+U!eDx-hHoyS1g*OIz@Ej-Ik zvP;mO*21=T*^W0{$wXblVM7r+>71%f)dQPRzB5!0_rGmKGJj928I_GA4E#q8l&Lz~7HjW;+=(GqwYEU>al4#is`+`4GKaHqeN$_7h>9N0 zx^YgV&X$u@2f^cfwc}h7j(-c&K^Quy7~F1vM)!jcaf1 zP_5Wubd=Z@6Q3Tg0~eKY+4#gJc3n?qr#t7}@87O|8M$Uz0>75TW?h5XwyTN?r?OzJ zAYChr-c8Gs4Ml(3n++I$(b%l8-+=??l!X0za5>0|Fu1R`zwg|_qR%tT`=!MagIG18 zyC^ThFw^bstBhM-_2*9^+P*p2aW86B=l6_z(v|&L0|w=!^K5!!apHeR-0k?R>+P)_ z3Ll19W$;x@%C9?oUIZ@`_xgx@bnM#uxkmp&M0=3e{px}at9kw`(W`_bYHQ0KL?T7o06!bvr(oy0N+tJ9q zJl=yr>)U|1!(_b4Q)a~U80{zeVu}g9fVQsAePEEXTw{`~j0wuaMe=4-l37og7Yu`$ zIqKvq9ENa0cn!8KOLW?54@9(XxB)Z3e*PkG6`E~*GbAjSfL);_Q}o*=@cl(+@;k+4 z=UT@OCP5ZR=;4g;q81T{7jl3@J=_(u3~HT8RVSemm-th8XswG6)qq7jg=rHeSXR?YVi|1D83@DZM<&HFnj^$Re`U z91oo_I*AKAL5zQnl}8xz1A0 z^y!n{by$@dn?JP(4kyMA#ysU*s66btLB4#qC_FjX{*^vjYAmDe4Bpqej$u&T29s$_ zoF=NKuPjInEUT#=CM~pj4F==wbqMPna~>x5Z8?NNxgo5dN@`%j@8G?_1mTt#&C2nx z&868^j6mgG*2_yuUM?bw6YCI_GX}^I%!CgQnOo!;Vd@iL>17x8U$O~S5GKe91%KTo z)Q`F>4i?eM3KJqrU7Vc-UImUhIheoWxNMiXGfYb|q{3_nRII`o;kkZOD{S}F&2kN7 zqSx&^Rap|YZ+6WcT;q5Mb0;qb6C&C6ce;~TMGI9cC0ylVX1Nf1t@J({r!~-{zR?f~+F#5(`+lWY|w#>k;QYg3GyY7CpP)9e#v zb1>$NE(H~#YOyk@Igy5|WEYf`9*a)g$xzgzBh0$L!OC2vVuVwdpRfiPj2gj6o1D-| zo3qDez@Bj|=(U~X(Tcq|>>x4%V~9dOz)l1ldW^8Wp^GBy$4Q#8@53~4|Ap|r*y=Xi zm6i3XSH?io98~y zwTan1*kED-n4zp)E<2ZGxnqT#YMg`ZXaTz!G6q=)-;ptULBOgEGtxoMgBSD}X(#9T zs$}!{EvMeb#>~Oz#4P7v5bHcO6L6>{3ycJv9>loD5D82Yijhb2PS!QFVuF(dtslP1 z-+o9>AEwLWhi`sk8@EMLJTIzxb#bP7i^(P`AMg+kZu8VepI{(u)5`YUP`$gi1aqfy zKe7$KvL`1NV*}z`WxidC{n*el((3#;*KcVP5Vz};;fvOF&M2mU?oa0-IeEFL@vs5W zM`Z1$X#2&EMv65VcWG%vYouDaVYKxutzUR>PNMvBN6XpE8;yv2G%2ZJYt=;~;)wsZ z`h6L}Pa`5Arw0$6?37#lmtH%|>uFd1?Vw^XBLD0{`;KN@n3gR=oXHu| zMa35v$@Fb5o|iMKH`-EZT4_=u)Oy=dGq<$s8;T@C1+gw#X zbq$Ou{*yBcq#)PD56k1H>+_Fs`8ZATwK;BhVk@d|{3ugb$v z>`}p4zjKB3Gu9>G0|ZqbZpt`0a>)@~`teAe7C$Rc%ASz3A9zKlALDFi3t{yKIjf6d zzl3Tth(3oJd4Dbh|GsN`M29Cls@iN7skai&(P?CfMwj#w=Y=?6iS>bU2xUT9rDc2$ zONUB*zQ-}l)!=2K^%>%;MajGkaD)>Oe6$jJyS1%GxU;ibd#c0NA)YlT*4j@k?(JJs z*oecuOO^q{gN`NOvLi=A2&id&c(=#l2-)mfD4Nz2dOWwdasYAKWzp2(^A66;o*I_K z4i9H=i)E#jCF@;y*|Q!|9y4FX11w45Hm*)~iRS|kC!HdUB@I3yzWFq)fraOSix|#4 z186ubd3;ICBT?AW&AK8mBTMwRgVSz?yG*TVA+NC0{R*2aZd-d@g$4@TTV9<4C$GV8 zDe#BD3jDcI;Kah|Ej`{R)m&Di{zE+s1u?CnvL(n5HLFVv%sgv?toiUa$XtA@d`=VN zsDY#BKZ`cKoYs?cYNy|(m;LHcz^Oy%s}wX&IHpg1K{ogJ zgmzliC7E%j)nk}Wj9u@8^ts;e>C|@)_I^@7$oKx{?sz&`f}6#rUHa_lGQ>&MV3GuU zPUrbsblEVV^M2d=i#S}q_mk0ADTouq1M|Hfm+0w|yh`vy=(;yYxa~R(=9Y%8Kj|A6 zJj8_i^l$L%s-f*1q=%k9mGjiGOrBpImui|kKGh2tiqluO z_-*x+O+$6l=*hp^0&QHbja;8X9iG>BPa7i^`&+tY`x`1v{%(V0ODi@+s(Hhp9C;q| zC(GuONzdiK4gkPobEmorNUB3n(j4iULijrxTPsRlSy==t(KIEvGzTHDt;IyPvIW{YUL?YnhBxrRemkl@=SKkfL)8vsOq%DrykRrxN^2vCvKxI_ldp*+mQ1=DMc##@c1~TxO{pHA7dh0?lRZW z++7pR`5@cI3|XHsW7Z%o&ONd70JUnt*B(dh;AQYA9Y1R$>aY|u9>B@*-umg1{X&m_ z$^JVAZ=><+M$o8n2GPmN99F9aS9}d`aAqH(rcEPxIt9^N9!>*@M#Ldxqh$zDPlSk* z%}#E<>!1Y3OZC+T*^o5|N)&Z)d3btA7g|1D=8q3S7rmdlj5y%;-$(a0L7 z-cHRebiS1189+z>={ari0CfMo)2j_hNxT+xiB~m{ixyzdHnK-sPhhU@zvAgGOD$qz zM|Ik)kjUQYHnN}U%OpLr!^$K;`Igh^HGP_&r}ZN}Y))qC!lpH*TK~P4_?YJBbctJq zyZYAur~>t3jKYgg;Zu4lFTury;FM&;Ip%Z|ROs+F%F?a#NI{MOFZ98Teg}+uADw*R z&G%7D*m+?ecWT~8CtrB;eYRKN_=Pv$M<-vnhwJ<7P{;8LZ@#bH1a+%{yYHiuFTD9a zY6&|ptmPjxxG~puUwHFgf+WJYOUHtmx>}6KRqYN^2_Wsl$|bV( zn-Xptz*dQL023m*47Q$wwQ`HCJL1vX4!&9GS(giCtC-a~tv$Q~s}U|{U*89lgXbX( zaW&Cvx|$b6aPZ%ZunCv!C||T3)*=$WD{;Uo))LtaT!s!lkbE2w6kyTi^6B9rJuVXr zBr{ou?ycHvG_t(gt+%BfHvY0EYK61il4<9{^+26USg^(C-MgQuA)dXSrn{HLAAW!x ziUzF*kX7);*G$e=keV94xq`x@9TuLiAcaE~?o~@5lLZoCuzedHbS?aHdAOt}C#i7y zeS(XiHn~*M%%>Tw5G)#BV&M}EW7%1!&&p5_2Cu~#EAX_|$WcANz7%h* zz!Blu;Dm5+a4D!>=u_Rj!q?TqE5Hl$`C5rdrjskMxrZc3U4yQ0UC&y|!>jh{TAr=_ zD*qZk_CA5`Z~s%!-@!LSc9{5~n07bWAPbtUX&D z#b7Bu2enp67T(G`S9=!Mm{OVLY^jC68ZHCX50gaCdrkr*eC0SYZnLL&3p=N;uet+PN2pMsuSVy8E)dMh?rQ{i$KE%-O+U z_gPV>M&km@&j4q;SGahoVzO-EHQnq6@Ky)i`8t~r=imfd;JCo08D$y++eh5T)!CmYAW;S>ZnlmrFu9aX;F+E4lDmW*OLkXc>ak+EckTY3^GJ62? z$iIAge5CXWdZ98Q)(0RsfOlT?$8b2BQITK?($PkqjC`HKY#xDVJGoFhY?RrB6T$|P z3&kLe6VrMdCOp9~*)`l(*iIi@TQjTG@w({HtERE@i+N6f5DO~(Qqch%0B*gqnXwAoTEKQ$zZP^_dcZ9ikUng?fT8)J?a_Il%E%w zw{m`&HTsl`8^iD#WPtIVH^niY^Lh?q!4vvxRHHIJhb-`)bY2X72>z2%^^}nAb~OPb4NU-Zr(c z&CpJz%ZrQgMZU>d zZHaU6zRpm!JI%_CZ1&1_m89BJwM=f}(s<5d0+pq3vVfT6Wy^q13 z`4n!#3H!_^Z}TxseXY@(R)=rBj-$&#;Dm5W)Awu#c3=(+FgN;5>^$$x=BZEcFUM6n zI4xtCI}H#AN3i_kI)*`Ik%T}rP{nPZ7&a}&2R_8EHLI;o*k;S>{W05Be>P4x7-PM_~c+2X)6bCIR9Wm2H&i z*`ol%-m;WsE8Ua{Uo#66>OL4L#jpbp$%JYfA!UrSCVEM!VFi6rRU@P> zKyfy-#(YF+k^1oALe94eG$}-KU)3V@urp*yFE-6`aB>^@Hcgtt%ljA7D#Wb=SU=1< zKLk++E4K?SoH0)gEFZHZXU(UH=KU>gqFdze3)$I*ZO!%&IbNh$S3=EUn#~^*XDrTL zRVj3PxZd_V9%^tKt?{UbQeMrUL83rC<4hx5E?1@dx;$OA?o%YZz|aLyU3GSFn4tOv zqu^(-5gr>n<;627H5uAd*bP4Arbbhf;X0W+JhgaJquhK^Gps(twDdTt-Z&#pHI{%G zQt)_g@YUl}{XQhM1bE>wp1g@cEun0@M;oNY7mv$64Dtksmm27O3ViiL@x=x zdVD&-Yh1liwR%nni5lkRw8+p~DHH}LBmfAytamZyjLaK2X&^qZ@Pe#$+q@J`(NdwB<7Nn%y zWy%A<^8=e^q_v@FgC?UaP3EM6-7^IE7<`F1tBDr{-=1p*q4N{vwTpX|v_5mVhe>{{ z8wJB&vptQkg)Ny{6O7bIfuUhhYFi&{5^rmnfAfS2F|?OM>o zboF+(k5oS5ejAR=1$Zu?POoL^2;KamXg6U?{BCl}Y5Rs2az!qJQ^{WW`_3?8{rJ*# zEYGKA(836jOLo2CiCT9KIG}M!vf{-$!f;!_ZV`NTIDS`k^__+U#nBk+@Wq>((^3{5 z9RXk1;Xx-L#Dmo+_zG>m*x~CTPAwsIcyCdUaN5hzwBROAM;XfXD+hQ2yn3Nc0WIDL z0dI+PLZEA^g|PSxyuZTm+2aYNn1Tzx=nDati!^jxG!s{Km|$OyPNOqEZTkC zoW$bQT^JwPwqVsKvU7g)^Ig(3CyWo_VNRFj@!?70POZHKO3o;HyP>a#7i`@ge_(=9 zH^YK~ahqaS7?aH!?bOgyLN|FTB~ecT+q0#lT?%|KQc32m3@G!MGT{LMA+7gN@$>%H;cHLDwqm9K>?dqs_PiiQ+AjZ^9l~cX@;OnvC2DJ2f~_< zmakCehGOENM+|$JIdV;uV#bhK*^LdRC4(c-+hPM#wD)7076y35;#16m>?A(6qB{)? zE^}G7AB{YFQ(+H}Y=V7iwVbLjD94kWo|{xE_H}}pw9%0&(E14B28A36EZ88Fdolt6GeM(V#&6zfxPMg-25*rHa%YuQ{8xOsK(Zp}yph@o)Yk;ryigmAtsU7M` zBT9Y6!-cinrPFe1t>aUB6xDJGb!z>KOFb$V2NdOSOQD`*2M?6_ZLnxSp@U3mK#?|f z6jGf5d%BsD2b3_RBrHfr^`w$=&k!b0dYhJmYxhQq4|f7V>v!}X^(6VR2nD1BunCmn zFY@SA;hkE}n^|XAv_m0BiXDq$%Odv!3hicUU1rp214yG{ykpWwOeAsggHSZxZ(<-7K_k&jAP%B~SE z2}!=wEL)xzAE2IIy=yULJ+h#47AsO$OSHUw_)I?2r9+L+2ZE;n?-v|8mUn0`fRpQt z*o02_(-SUIg7Q`IQa&jG<)*`D-}wWa4ueBsKc3d&xR=spxD`+9;%je>r?q(NK86;L zEYP=bh_6j}k|s#&&*CKghZ(#O#13BuIPGU(4cU0FDArZ%l}G)Oq8}F$An?wCj-x06 z-`18g*IGs-2MBcQY8_y!;9~dB6TZqmhG6kDbMc#x%(zc{$IS7x7iWXb+OOwmgo+9q z1kCYKbM-sNMZ%Lo0Q(JAs=4Y8k||*rtXOz-Y|j4 zG0;uh9u7BPa&6x`_UMm5BCH)tP8^Pp;2y$sXA5>4KzFua7p6N~u-gE-vjw{ipgUWz z+W@+=1-lKP+bxJEx&UX9+X>i~7MErc*?72ZDsEV_U0~SK z;?jm|a^4Q%dtyD*A6wemYNCn9wPUs%I>dT&{JEhmJG7A9C;F7wi(x~XA&F@;BC=>A zP1l`Pq(S8y&v+VS^cL5MCIyX5v)m2uv74ezLEVp0%>k|728%R8qUd2%UaGe_)fAtZ zO{O6yPXt*PuVS~Be_p}gu?OiS0c;s3U=Y5PXKoZO~f;{RMB35i6-#5o z7hPK)jOn4$%U|-mwuAwFE28Oxjb17g$#;S!cwTQ!YPZrmI)=q~UReFK67=2**3Dr| zr!GOPEo0ptowa@aL!PIaGmq)WYc64}bSucdd>$OHdC@N6pMLO9XbWkkkNeFVOqXEd9m=ygDRRa;>0Y4}w?ps6XU z9v^;ha75D$)Jog#jFUsXM$|E|-lWy;Wm?^IIC)}#gWK-IigIy|8&$lB?Z5NL3)Dti zM0EYv6jMjjl??x7M80;zXtRua|H($95fXj8AgT6277DG^N@dJ?Vrd_+tHv_&>{%!@ zFppR}%nc7~``S-sHZ&f8i`)u=#I^fMCymVc#9?_+hQZZQO;(wk~?oztKD5zsMV$*m(|Uf9zxq14ZAs?#~vgB zEMXZws0GV3raUir#KT7RuR$0#3Bk{Hszqwk_?A?N=FD#5ulLNrJY8@wg=%ckuVl3A zUF||$>TrEujre&dZG;EJrvdH)zD-bUj(6Z?h1Q)l-PFe&cw`(l!mX}FJ$}pqY6-c- zXpitFipRLym)q_#XmRMI)CftPMA}pOc$TahvW9y*)eb1@=2gq4| z9ogM}p($v#_m>hD%)UVsZ~)5q2>Qz_4xeiKVZ=ujT-|iK4!54QVMQy$#WPBkNn#~a>7kqp37}P^5$FkJjTUY%T0!MtBx9{W8Bu~^`N21 zf#4Dy5TjeUJ!82+3fgDE;&3}7h;gnG&GHi3M*3v|%EVxXVoVIw@bf~^1}mHR5OoKL=jM z^qtw_`!gL%C$)qgcjCOF5Fk!QuoM_AJa|78^)K=EUu!YDdW@NQ?a__jcEBsjU($T7s!4 zy)t)}$H-9pE~`3K9+O+e=N=<>O?4Pz$zvj(&WMb(^oe&E=am@n#0Gl0#vxz%9 zvpHFa<2RJ51?!vz@ReZ+y%gkZSB`sny|TDO{}yzzQo9ifo92%WO$SV%h9~JJmWCMn z*yBX@vISdvY&DDkElsrW#nNLgWZ*GU*>mbvq!-- zmf&&jq7ybU9b~x6My5fAdu`-q8|*1JVIK$GufGE~VT4aI^&8>XO*rO2w+;3_x1r** zuhA`*9CLt9vHH8=9jpr?dw6n+ z4O&PoVF$m-fp*{79t&E$YcaKi9Ts@kVZedrrglZZV_@LJrAt29!Hwgjg z{_d7!#(gKl2lX0!1-qYX=rZW1tDPYW(J;UXUZCMw?F%;Rg>z!T_&9P?i2dJOQn!JXl^#aNyeR6eWt+`;UmLQNf&da{OcCW{VbFpzB7Wq}D$^`pNg z*r;JdU3)6A1b3O^m^zQBb?ZS-4v<-u%FLeXB8y4fW2RC=cHwwgewL3}3N`iA?P|G7ULCok=K7Wzd|r^6W>(n^ znH8}P@dMqoHJ#Qp?x||PQCFnay`;FMsq~tX7tCBYn9iWk9d$C)jJ+b6Kh=bLn~=I7 zwZ0{_w=as-?5V3z!#)+yb%xgY+@peG(;u8P^^w!SR8-EkqoAjfg^8KGNai8cO%gH% z#ML|fLUuETMk}Ebk8_(k@%5HBB}WeEa5I}2fm zX&WN4Pj3nFSym0Vq5mqAqJ{|M0xs)g2^}s9%l8X|Z5WYQtUe`SQ7o#QqZng2R^vEntW+wo2fKfva;eH=9@Rb0R_lataIk*zH4z}0;aK%x5>2@( z+snOA3NPF>%Lem7yuh_C&o-eQI>C;QR(8uRg zWyaPBkDvgX>J*yH>4$6RRRl*6Zh{I9B2R=pe(MorLM}BRsG+axpj!HYQ~r4OwFb{? z^`Ik_Q@U0DQjlxxq;EJ}A6VErNZwRD5d{TWI%ooD-hs+_6BBe>7-^xT>ln6vfDoxn zXOwOmkquNT*Fg0~Km(amfuSmOi*!+i7KSD@*0|=Y)@ahVmlYWCjqP$Y?LO)H%k=P+ zo|cEptKPn7u^ro6CD*ByuM5^(bC9q?;7O$p&F%I41!DQ9dZ5KL%TjIL7bs9(9TeA{ z5z=l1lnxUp>_9%Afs%g=b%S9%LV9SLb=`Q;_2d~KlJ6d4=ken3B2ZQaS9A^>q1wLD z%BDT+&yc|3<#N?kkDp#HulR|Y1X?5aFf|FzN9{l8K7to52ssj_U3ebC(zS$0g~Jfu z{TMt~j_zzhsg1x_<)*jpQ{IjpgwnK3(Wd@wFPW_UX6hUK4}XOBGQ6ewt(s# znEwe8`8PQjs&!xskRbj1ziSJuBU^xaW8iJJ0QJVe+iZcdU#J)ZOI4fp1s36_IR7m%n03vWw*Ye%jERo#nfSHAqmA)g9Lc!%3_ zr`Du~ct%$m1X`08{Yw0dP%S!(O7vQ@?t;*NN&;0@)d6X0)3jzh)S9Y(aF17^s}nqB z^Qz(sANzq5%gGKYU&V2OB_4$Jm|Kw^VZ6=1UQf~)-FQZ466(e?@JMHtf)D|#(H1xM z;AS>Jmwz-h9JS2BwA19fTTDAmJ^}kVkH#&#H5pbo0PMoOx8eQP(>P(TllNgO=*%>~ z-(t66AdKNrKTV6l*XAZ$SIrhbzrtn)H?THkwac99;L`mN^!SzS(TH7z{7w*GCBv=G^$- z{h53#b0t7p`{OQKgb>;JHyr${)(Ih^weI0;@Z{jiYW$|5`<@~62bzG+G~6ZxyFdP_ zAmKYpZ#HP{q#^%q8MFpZ77k)_O*h6`C9?YA49j}OYGu|HrJ#(^NhB>coF`E5DD9U| z4&v;Aacql}hqXn<$Eil7yi3e1} zr-ayRrc@&6RY~zh`H2%W_~2k(@nJX&VFJA%L%vr9^AL6#iaHh5!|+*!{qarUI_

Upf{eppb+!={@uXs8vm(kc^n%yxxL(ennYbsCFrq_DLm#<>o#*RsjpJQUn%jv9Ls#CYJB#?Sodc9v3bV@dEZP+* z2-Ov>zGj-s-HU0dH-U)L7N+zN)4WV-9h3wCc4A#*-~f_`{eFny?_{Ul{+%z#tSWVS zCOtD!t~=sX^}AGqnJz0=XjH9c@h}PVmh~yNv6HHq6(LqP2C1K2%7iLv(~cwRXCjAn zsF_D((A+_zj)yA-hJpV3D@aWZ-&{f2bhg97Th%JvDU*?^VU1sBwUPi@H4d<<29#-U zTQamIS2ORWL6^tJr|WpMLAP&h&?w=I>CsNy=osB)dbE}YA!6_uH{Dwvgvf~&G_cIH zWp75d={U{&Y35Kig$R6QH;0-6>+K3Zu~}VK3|B)Yxd@_02|0eX+BYZj0(&%;3B)!k zv#$t*Xlxt8l%KU}F~7_P8%rY|f551C%CLPmTQ6*t)`;6^x0V(-Y_DxY>p?8zi#uLiXK}RRT-Z(snV_uEaS9p5USEC_m*VyRMJe`B5+xf z8n@Otdo*(gh! za9>NrenziI!{q}SVY@V@i?ju3=v4XkmwOIV)2`}R9nsL~;>uG%@5{4mhc+1voH84; z=$(0XvEFQH$?=2__2rg_tVF$N$WmcJ(3=Yn&vbcB^QV``(2h{OK?@Tg z8RV>W=vz2}Z=b3d>0F?(=>(2`T5J(CEw(`*i3+dqTGoup+m=ZUjkrdQwCTQBej>zB z<)t30RYvfhOdFVR9~q0{tbRHdrPe@@a80nyE5QZ2>>p+Dq)fN9 zV(lEYI(4D%@>rKBC8`rmd6g)DY&_T`gZKSROHn`dE4`Z?F79)@vR@qNZH&tdEW^5o z8(Upw*cN7%_>tgY9J0ICSZ|rNungSQg1yu3xL$%y83?1LHJg8p%IMCukXl88J!_m0 z9n77qB~(~M&xGS@)4a&$wCkT|UF5R;cBMGov>I*mSF_CfKA+JFdXvv6l>>O^D_eFr zu7{;cm#ZU|G7kmPxA4$hWu1dVj8)T?^}21CEP3YfSC(vSpH3aJ))CeX$w4*^nLL;v zgVC=7VbhL5wgzUiWD9YqUu4y0$9<@1Cag?gv#%w(3Ib$N{054~qvRQ95O07AJSV)= zNqXK^!OGj0$T!iK>*+*!tssQMz}0d9!KLJ*aMd_)6*?$o2>9qrvQR|$8tD4w%hOdB zpJX*l)(hvVX{sLIEVIqpnu}eQ*(ciuaJ`Fr8`cb!!M@Z>;mN=XO?BCAPP5$6tELw~ zrT|1alj-LD=Jy@YLv*@VZgdKr?3J4pq>lK;3c9nZTd!6V>hgTF$xDs})rE|bNM%MX z4+OU3OQw2QP3M2TV;-Plx>YSR>%0%d>T3$r@&Y)O4a&Nc((FFfsdilg-0YK^v&a0< zYTF+O~EsNj<|PCr(QhqZ)Xt9C4s>ATY{gDaLN8^W4KFAdWDzt4!D8 z0v6+f`>9OX+`hb)PDt;g2-eDF6C$^Zc^$)>ZRA4w zV%pXi&>_tGexo#iR&lS&yjq={ST}6Rp_01Sck3d)GOkDy@B7`X!fWp)(MB0I{_4&D z^%a{8Twd9!e)wqe66c@+5@KPgC)%r8Us+H#Xcb6gq(HUzSCxr%Tf<hP?b>!)f*w|I*5fWR9t!RE;>19Jr!x*1dI12acGU#D?kg>kF(VLD;Le$phiAc z)d)oy@y!Tv#SE*Qd8)Yde04O;9hQGKDWI0Xc>}M{B&iJU4ncg;PMg#~6iE^Fp&hd} zICI9~r2Mk=J?lDNM*@5u;2fb%qPoOekxmoQ+r`;I398mA!Bm(E;t>uT;HsZKX^UKw z6_4{V&I-dK{X!6ynf7b(dJv9n_HD=&Q%^2{zW!N@?q>sS6P98!Rq5QE+wBk@e&Ya4 z(;(ju5-s*p#r$ECb#EPXbz+q7r-h;@d6bnsE5ze0E*uRlNBY>cSgV^OC*&1oqg4*W z7nrh)b-sGkssQ6X@OgP|l}ZA{?|YjcjZ(Dscj1YnG(meK9K**;^xZB985Ik_+L1%J z4hjnP9blgL>iDkHS!NeNK2^XUl#wsImR%ArMQEz0mY!K+w#cGbMkS2Q^WkLICvn!TFFbI@&R(+XmaXCa}+m*s8BAn6=^IuHM6k zmA=aV=ZDAV57XuH;hW$5c_>OW=<-k*n|am0wdPs>S~~fRZ~yzn3lamCo1`mW+|+Gl zDm}iu#`rNlY!@m>)TRhLkI#ClkA02R!!h#;O9Q9O_UKZasazLz6sbBeuY4A=2-DQV zLE&sLA;Moy8^CogidjC3SznAW$rZ&Z2MzDn$_Itzgm7LOxH76}7WvquxR^Q|LC1R7 ztQfqn5I^ncZY$e45kBp}5jK{dd$EJ@y@bfk7f-tlhMA{(v79W;X4UY;W^)IZd+{=_ z*_5s_IInCJY~8aYfww(~B61TS=`1%|AtEp@HudnT@YKSjNOkd&%2}<2Y@@98ts+;3 zf`daHA}aU{^AX#+m{U2u@$0fYyr$RXVH>!K_yYgzds!IWeEOpOw1(W3#w%vpaZS4I zE|R~?@NL|Zo6p(vz8qg3nLzISZQirxbFxyiu;ca!z!hu5cE~hgFZNw}( z_2fR=wZ!J|!Tmigf9Y0+vwvwY+z$TkjE>2A`*=9E@99SM;= zpRUyxIosrK+bMgMzrXl%y)|=9Y7DMp6HtEbBO7$Ny17^TbJ}G3@ha1DfkXX%-`f3G zK8c#yb$JOdPs{wN)F%Y1tStP2oF1 z?Mk03ncl#TVae22It)uT4}4rQ8Ms&Jb0yR5N}ns4ZddwT$z-5Fup(~0`qOo!e&Uhp z%HR{TRj^f0Wc@l2FSZ8T?6KV;!`oWKmg4KIO3Ze{C$*-6nex>-tYrbDk(=2dIZZX` zDRBp>8eDHcVk@LoRoCGSb;l6B=V64%Cg#&kk;bpc5HgTiw~W|1V!P{}nbdn+x_|W3 zr_l8v4kjW~(fTrFNE_Y#?c2YXYxky?;_en%gvf~scK5kn4-M5BNVGnn2gR^DS%ECs z)Z>v3$;$l9J8bI?%~5Ucgy{Z9#t?m(V&u%xyNj_?dGD?l-O9+{Lt44_GOzgLs<{jz zOSVTYQ(V)NNg#gSME$ZXPHlOJ?47f|sYD~z8l&J1tmV9B4Zv(_tcGY{FZUu7e^fO# zM5Y9=q3m0q^7~#tRIm$J9tDt?#zCp8T(*ja+6v=y$zm#FNT1Ukb{djq=)&A zvWdC&)x7Y8o*=PuuM&tl5bo92$(JQ=HL?W}NVTqsb&-oA z{58LZ59M@M)#j^qi$ohZn6icj4lNAF1=I`B*%6U|ecP^cjKfnZ9)$gZJ%e2c(+Ylj z&s2G+pIWmn&p0jkhpnOsUS^f>W}k1G3kxO(TZB!94IxrX8Z&0%6%v7M!Kr@8oe*i^ z=-}D;@aUPUXYR@Ird z%(?P?Amx|HWz}mclqQn!5zVLu>86aAs{}qq3FGPA&DE%Fk~oem>B)xN zwKh@2I0xm*tP|3rJX`E*&jOJXiW>%xKxqKch-gu>*g;vZFR)u}H;7aV+B|8qrz{rC zGM8r7l92~(9iJs;DbkWu^lYwykAhmuc)oG&(Uj?q@Ymt*jknzN@i#gOz+GORmQPy% z?s5L`YB52u$t#msjQldOhw%Y3w!UD~P8Ec2?whdJ0_vN9@a5B#((>u~%BQDk@Yw~3 z6S5ZJ-yV#N8N=I($+IR*`B`6gC%kFsXV5hWhasG@X)b_iFTovx%HVVFz=Vh$BD){1 z`?9jd&~L$Y(Na8upc* zqZn+U@Z1=Ou;gJ?SDw8rbe(ixi<*0S)FH5%TT#Z3O}Dbmx)cbWd9Es3hAC^@nYXWZ;JpRyIz>>>&|Av$!7w$R+KL%jFHBAE*uvJVZ$Q{$*MV0V1LoTY7?@2 zS%4hF=2_0+sW?h5M582N%Tavc47HCOAyQPD4cT&o-!0P94LFRa$aECOSl1dFQD7Cn6lM+znNc%M|Bc1 zBg{_+cfoY_0g$f+C{|*GM22RJmSP;iTAmg1kdV>Wc%2Lh{NRt7CSWJTH>X|tH?Z& z+=B68GxNi@e(HCH5(nn057#5D8^{ zLnAAzkFqFV7VQ;qL`fKIj6`Pp1D)0(TJEGsi^|Is&kTWLPBb22A%p@I?!XCXh=4fn2i*UPibPPJK9r ztJLQz^;-~A!|rNF*_yyDxQ?u{qqIIZ+n`PeFJm|&T+G6S5b5CJ!rg>tD~M0u4&Y@B zCxllhs3H}P2g;D~S@!&8QWCYLaqT0R4(#R@9O z2#yG^V|ZGjph?z*s(`Vjv2cw42AoPktoO_8Q{7wjZf);;Yk4j>lCs8nY$uZ}O>Nqa z2N-PZHg5@`I_ghs9~TI`^$z-i8{EH0F?zh={q>?ZKP;TfcQAX@QvykkmM`G3PF<|) zO&q+mWYCTnZkJ82Md;k#5y&d;s|(Gq0B;GpRa)#D%{9sUf>~MBs#t3M4X*bIsDZcm z(21CnyD(VODO*7<0(mc-v7ntdW9_ik3E;Awe9(U33-ZfYZ#(%EUPFyME1lt5t2@JV z?Z^Gz4FAg(1iSD8?#(rx$mWN+&x8-VHKYVjfM6@KLV`1G=3rxlgDID zLpAu!28$r42^e_~OgG<#HNu)g3tZEDjB5dS1~?eJ21lpvfthDeeVf4(FgQ3EnJu2+ z8{u~q6jvxHWkXKDbAoRs)Pn#;v(8j)>y?dTfEQ5A3MY<(u;K~&D;rIU-&W**zSMm* zCB|7V&yh_uwR7?!nkFwG?pbV+x;MEzrN>zoSr?WlLG@wT!jUs9WuNB;l>NIF9O$6w zCdBO{{jh-wF29H&JYpglC({cmnFOlCHtvGPO%2iHe9?!h%C4}zb=G%DrAF0j_FPLa zTpO3OMTzgTPskf6xRrMH(B${wa!_UFHg2c(YReNWIIaq9o`DQ$Ud~*rQ&9^@syAq4 zfs_YFtET%fH3@6b>H0pH&OK0F$3hU!1^sF74-@W%e=g|XOhG%VnrxXlnA%MCuv^mS z@mLNGurf&^+8a!Wv@kwwQPo-XAMUZD8-aW_S%aCc>lmgkVK?h8TwY0{P7mv44e=ZD zpt0my&hVltv%J(Mi~|@xH>jrP88|`wy}&f@Bv@cR+Pq*BR|ng=ynOJXg_k^nLte~w z&I{b+)FzDJ1jp%Gm_`X`WgzSizSZ)t;$I99+ANT#8tkq!NoKJQI=9?}%MnR=m|s`l z>CYu*595U<29s?iwZk<6TZ1yfWp=uOesr*ZG)92h(^1yN3N{IqDL26u*QI36WjJADtc@>T$VZMZPFmbiA}I(~e}w zM4frTW)8M^9dM?(yk6Gy@~Ifd6ID-xN?qSRmGu-#}-ihQ37KAr9W@+Dcm^y1` zCLJk!mR#gSM@lY>hPu6_vb2VRO^@p#7La=td z9~x}O9zqcA9zlA6LzI{{_uftIsrCP*yLSq;aQ|NDeE?fm1p=M#z3t;kK3Z&UY&Z= z0HMTZpp=zp|p4JhWB!W5}M^Zl@x?DTD6{TtWF2fs)`M zJ0|A8E@otJBZ2_g21?BJ_9MRHVh-X&Z~r_67fj45qT*pqU|miw`8-T}3HI-eWhb5y z2m%S(oXQq)JOgf6NjCfgZX4nY=xPU+fxtsiZL9(Z{W*967(V~uJ`82*&({LC+?dT9 z>`FW{&vIP=v7xC`LdLNA*2(ofRY}5Xs8%RJ5I^l);M2jhlK=(b3$(>~)xj~9+L3g| zA|A*j3cCi~g9)=KZB*M5%K~+d3O$&=aakW8((9)e*_wbBr*;qC8F}c42mOuVfAJev zzWny^Vfx`f%CSiRmOz(O4-(Eo+ zdM{m4cjeY?!5lh)Z8Z#I9()*1WKgwUK!!(N@NxhRXU6~{c<1l#Tia|!!ZvPf5EI&sE1Z~k62 z_HUa{zEmHiWnbi1O=I!@R@RB|SLhh%AOEGjgnR$~ToC+qmweT~)68-t-z2SH+S?Z3 zL$zi&23*Y|ky^2{H-c9Dc^w_9%ofmObk->lJ(0vqqQ zD39j1fgI54-x73&%X(s`SUR>}EYBvP*{pi1y22P7p_v-cX!sHk=7%M2h>L%V!V|zj z)l;4JtC&}FVCK7t?2P^Zp7G)Sh9aH?)HD{2h)_}p^`i}9gy=Z=Jajg&eiNcEQw-d5 z^44-$Ot%kIXSFX<|GpD>WRh~J#*Q8`_sTFWa4J>o?)QTe?( zOon#ivB5i#QZ-cMTQl{3XFKju9qDO^)p6Njbd(rsYxp4SF-GUJ9%J-nidprhi+2`N znfRZm!F?1ZgK%68}q>z-DI>{4Y@uZtUGT0R=GmO!)9+eYl`LtbWiU z17cI!F`1{7sGp;JH?1RljpWPi+D? zR5b(edJNN)NH>VzUO=z2DcS{(8d`T_ug6uz0W+&~0ZL5SF4zr$OwFZf%}2oDG{G@^$7&MGHegOL!0bZ?>Rh z`b)jevv7Tu#b($OUMA`o1p-T z43U8htpdq@YME#&U1TKcNU8N|&(21vKiX)rFH4)Me3cg@TTw5|*&IsK_3NbCCpswi zklg%i7b=!uab?6Bbki%lX;CJmN&LvnLzSc#>o?qr{$8|uX0afWewFd^;CD;%HORcL z_x-Bvoo9^H8IoJBv{ffy^nAZqjs6EOh;IhxId$HFX)nQGy>cU{DUh5WP_%&T4__tb z5O7}v+;71>$5qwsYe6Oo(vC>T!mKbMV({+m0tn+-+L>*L!m-X_J zRu&9CK#)WhxStQNl_trA4}0E1UqC0dFzqE63_U-Ti24t?guEy`m2uU;QxR#EMT7~F zjl&8r12{0eNC!$MZfDB@aqyf5a4m+1!s}MNz_eX@r3nVtk=nsA+r~2w6C$|AJ9Cu4 z$weLWwP(TkVH*p4Ab~LxKFq>kMxmPQ-RJ#qFCy+yw7rXNPa|xm!K)JFctCv`1_SOA z^p`paOu^OiW5f#^)_|}!E#oB^-6s%dF2|UMfkrk+w-i}*M}sWb zTm(0YE1ZL>lei1dT(b1OT4x5=jS%wy_LE0_FCh|m@=#fD@Xj1D+vN+B=$G=kd1t~{ zU|vA@wZg0)$CJd~gfW6iIH12>Uwj_ID%ceXkT8bbR_%dlPkI$|U@b56db!nH;RtZD zo_V&tD`<6_&&<}9HpM9vR3Zw}<@Mnr1Lni};j8@LK0IDNOqa`tZ+`RVnoaeii}?h# z|AO4&fK07Fxi=8i5dSO)VJ~5?_(h+#!qmhp;l;ISeA2-yy`QE1^=0f`1dVN3v6>PQ zVEaNpih*7cM^GdXp@D0cV%oi=P_QI@5$mwMloe8vHElOZQ>00`%yZZO=^}S z9BmRrZP|--Iqddc?eUU9-ki5>-s>C>&I#0@445X znw{{IYgI=?KSA1TgFLde1}qv(L7Pb_=sphgql3P`St|pe!S4<# zy}yV`?avv#v0xgI;bd}df4fw=C-V`NdU8M+nugQ?*iJ`p2I>7p*p&SxDl}C7ZvyEH zRcJl7M(zOj?4!HE-eP+rNMEQzht|kl;MhKT1Nie0xedY)8Ce|XL3-nJ4fr@wwQe3i z+0Ly1QBs@cZD!L%TcXnu5+UC_$|u=Y&1Ms+23X;VJb1#`+B}Q2spwV@zdWm$A7pcK1-)K$z4y>=iG zjw!eC&IWk*7N-bw>QA6;dQC8)jc<@?FP*XRdpEuAgQstNBFFR*=tk3*_%R*G7&KXx zGF~aTT%Mjon%0+?KR$%*wsv#ZESHuX-HYk)%|}(7eEW6uFWoaE?oDHkYk>W1&mBkr z8CXl~HTFoldKsoWgm^TNlIiTx_~}Mc3-tJFl=%dOW{VpK+}vbxdyYa&7u;Rn zM4`7mCcCDfpxlw(M)|p*f9-<)MdfHv(Afde4_VDcGEP|)IQK@Sg4A$@v%eVjf% z%&_R!E$sRAzvOY6ns-ExPfZ<~1;wOwgo9>au)P8I_H4b*1}(sLNplycTN%#c^xgt~ zuF=mm`ng8`_p4FL_;cUIs%r0TI2j>6Kd$MaOagXw5>}_qA~wAOlYp+_YLJ>kT{7 zWK(>-e`J9-T2m`dEuq2l0iF=2mLND-c0(L|%YqYx$SYGYJh0&tZUS*O1-x6MD@N$? z$qUaIx2p=QUmXtRr{;Yg=S>yxg3y#VHRZ*oJhr^7M(yZqo)69Qt+_+7$)9}w&{TJ7 zvTHC(YSfrlTX~3vrug1EBcDI{{J{pUe_qef*}@t>uW>PcUSw_M=L6!f#<|wVEws*r z;4(nqehqGofc3*{IJxI*RtQFIplnmiEGMzSy@Fg|OjdkqaQA$#aLGR}U$8MeHh5}q z0jMSH;NccdE#V9g)9Ukn2BTKktN6EC`_>Q% z-4M}s_!{+9YO5`Ah^DCixxR2ee|pJwuWL#gcbl8~d)AHklExHy8}+0k%Hr23vZ!cR z(1NotQ$T!}0;1|Dt{`a7-4xffHq#9XgBfbXB&aq}&;mt@Y^ts{pi?7md9AR^twhzD zcs5pG1%VV*YYI{!_mqtK2qLp3;@5G4o1I;^J86_m&dhXIfNM6*@^EDf(klFQ3I zp@3N(T1q=OdJ5svJ^x}yNgXBk1$CsUql~p3VMei%sV8-m+)`RUve*zW z4JqZeAw8vD?dW`ng&I=SYe=k|mPb8h8)}|3wXxU`S%{I+vcKGx>Pa#NL5P=z!nanm z6kiZue1g&8wV}wNevIUAGEO}SDM@T3^99$2BA+#Jq`AoyJBr@QVQPw98Y9`o4n~<9 zO6(|lZNHIFr!83W5yigwrX+PEl+uLg1GpAKG3P`J!re$>nT3X8YDy{kYt8jH{|*WY z`s+2sDK+vo%KsDvrH+p{t>vRkXDC|G!+iD97M;LQ{VjeL%ZuB)R=r-rQWu!2b2eOl zbAt!KWph7fSQ>n8@Zj<4@qj!oO@_I_gU1(-%N8#sz-3a{x!*U9zVTSb#K0DlMRPsvH9q_fmv6p8J?r^W!eJvbIzzw}ySvE4*MQ&$v z3yU&$)aK_tu@9Q50-Pm#wFa21%RNa)h8aMPcZy?gSu&~03i7K(G~sauAHvnenHv{p zE5Vc1#3g}@CpM3O#C-(gQ540P!aYmI3h3INl6NWGfmBnG^fn3_iUJeXL_O(k zlsgKFhZH)KsZypPrQs2hvi1KR6ux@ZRE?Ht6scSsQb3MSpwY)f@k0^!8qP~LSbfR~ zQ92M!%B}LU!Qze@X1^`gY;#+Tlb;ZD`ok}4w|S(fE;=m)h~sVF#KkM#xWt~4dx|rY zW#M3HGWp+pn?j>Z(;Fz1LrYpaO1l=R5{w0fj?^r9YIN9@+)`%hNTH=v2}Yx#S)VQQ z7hwQI;>$ppBybFrx-N_*FF`MIw4!|`nNrNZld%EOmbG@jb-mb)M!^qVx{~qWa48w> ztk}unD5^X+6xdeNR2nIL$t^`}Fes=MZwT$fyLw7!3W{BY(ty%dNR==3BzW0!i(>vc z^(3g!e2bEmLOp4hA{ui@fejjiCBIvg)EAVFGVM}wXRsmZJZ0(&>K9Y*rR)|oUbx(+ z(AysSw^2~Pe1jC|wMHKMf$jo6506hVU9}rKfV$M*?ElIB?J_^0nl^=AE7Z^LRe;e3 z$jz1t9lbV0NR@|{q^d%PF)Spg!2%(l^4FpV^4iH!cx?k3G^QkXUYXrj49|}#xu=k0 z%@SPq9Z{CPptZ{c0Ro|zd8Yo3r%(naYbIr~Ezjnjf|0spyBMU8qRK);SxwcNSjihS z;=7vbC}L#_8Vm|+H2-vjnBn1~J}9(Ml1g@JGofVn&w)|U7M_~Nu%Irq*2jCm$y}o= ziIO;CT2e-v4SEJBCaF?(A{T~Vt%oHSP$r?Y9aFzZ1S@L1gpaYfJd0Rr;R{Z<>f8Ab z6ZlN*Q($LJLJ2JeH!yaYLOM^uf?}6xNSTKeXj&yXgCnx4bQ(4w*O(IV^Ya0}o+pj< zaTivNXq8Mh`a##sAvB%D2^mXnd`mVVMKm!|C&J=I!wnjvFo6ccC^UJD)b0Ec_~Am` zPL9r+N^POhS1zza>vC!J7A$6l8BG*6IHcP{6vBMv4%6HZ9cu;r}#e2gR|`T(X2 zhB-q)bCfc9Z~el#%2pDv*;0cwPzcYn~}2Ldx!T#X)>ie>GY|M~UNK z>^#j7(4ie9-cGX#aZn>?`!v5!+_6UciHB*;V5Fl&_ltc`VH)iwCMPL3)95Hsf30W@ z(uUnX^&)sPjgAtxpLv1DXl>*4;FY!PAfd5}+Ta*=*G5Bb=)G(BHh921jCP5Dd5F z!#+)FHPAU6XyZsjb;C`AA zHJ&v}9He#Yev&{(iHG5)2{bn`3h&ZBfP<-JdAmrqG(jQS)j&!N^$(1W$9w;cuvw;fJbMR;{nQ@rwc0*EQ z7-3Vt%-4{7#!xjbolJ9Ou?{a{&_T)+(aXc7-KZ?tpLHzRDE4NCezklQNwBD?Rf%dm zrxh}o1G&p4`Z2f%>(||2a*%Yp!I_FR8UpJ9087XU74g6?@+0+Wka{Am7Ln=WuZ88E z&U!OWnTyZZ$n|YrrehVMhJ;xFh~kQ#Hv0`0CkWgsr~=7>APY6|7P3P8D7W%n|A1z{ zzu7ofmhBHm{XdWvUW(0S6jFvfAe)@yM|3GJDZJMO5<8xz>Lp&|ds>74SbpL|s9las zWh!K0XaB(ve3?{@VkBpO)f5(=p|brwbk?d^CZ~E@{WL83{H1E*^z!)nHf0WvR@R}ug(js_`$NIVO$#y0lWM_CZ3W1#Nea9cYLxGSu;-XgX#O8p;*Cok{& znI;SDANV^SdqZsI%HepM{Y{71V27|_&c{%2^2M^Odn|;J?ul=O9cyEgGaH9^eRO%Q z+Y#kAp5IyBy)-rq570Cmx-Kr-9r$RFz*I`HA3PJ4zG?XJF7hQD6vL#cg2iFFT`YLX z{D5|eLBe0o8Q$_ay5-i178y?hsXcG2iI<>$!4d*K2R34raZU*N&v;S*pkXx{h=1ZH zZx3oxIq(XjtCq&IoFWGhs8BzKG$uTGQm1Yjh5VDIU$WY?AIYC}>n%`|Pje__eL|)~ zkZU2`p}9e8!Y6ZN%(!PCxcutQBYr{O5AS1z)`yhniOCuv+6{7WQ#v3A2uF9;{k7ck z9@<(?S}vG-iB2dD9lqr-KuY*``^XCY*qfuCxqWoH$@HO`x5)g3m;s`~L$znxUVgBg z*>L=f@$0}`d*|w6@SQ(=hn|ALx3qPie@+tGbNA^Xy!lNnX3NVKB}`v@VS9&G-&g+` z+i>aPrEiH3rClMohkAW4RVrV;*NS{tid-8~OXk5-n2&7aK@>gx-H^A)*zwP%-I=U- zAKym7PcR}xsm*HA8E+%NDR9)Ybu=)=O9_v&e)H9#)W&E=W#@_G!FycjDwpGPX;IrB zj%>e$e%|*XX!8rNt7>MT>CSgKnS6JtPn@3QALsC;=Wk#CB0=HFQFO3dG1Z}soGDWinLf{yWvEY;aFCm(ss1Ym zu+M47Sspsqd$%}!vvE-va)z-a9YOmz7+SVls!dL&P^dH+v9W%RyF<^M*nhZfudbCU>VJOW`oFEQMOeK`L8TkZ$~B)y&6v-vVTF+g!O*#M+JnOj;a_j- zZy!k;dD{=zlQiE+@mlQ9$N2{L&7#hahw-=3{}+bTb$R4vTU33)$KBL+YFXK>f_?xU ze~KT3t=!n_?Y6?O@xCY1VtYNqsI&3vzI3sp4{B?qZ1?{Zl|O@)3&h}_8T0ONIvf|b zJ?{)}afQ|fOY!@F1>ADFW|0F{PGqo)H$ZHv)!wSk1y+3pV+JviW6Ul<c zy@7jK{*Sl(K0a<3@Uk$@GghBn_&TTqbZq7WIbMt(efA~23CgSc2|Ns_IlbQN&hL_c zz*+DaNcQ#P8)4_ySYsvG@vMux(H5zJqAD50aZ4D6*M&hWGfK{z-fL^fUtqNNBDpgA^bh_bzCf0EtGS^~4l&c6pCiysV*bOwo7tLr8XumGN&_1S0nh5%J4VKN zgV8#B^pKUxnZR3ph9`O2?EElAMPygrkybi`qi+(zUSvSPUZkJGa<(?#B)6HwBl3@DcB)N+zqK<`&6r|EIM!uV= z0foUZaK*neImL@V_5gy?q~31R^MP*DB6V%hX_CQU`@)eLSjGeEl*j|i7DyO;I{t1_ zshvlMh0nI8RjD)>NL19LQKvuG=1%SE0il4@0pXL{_~)Zm2Td4F{RfiJ#aG4xz{m-m8s^cO*|> zt;yv0(Q%II*X4w_neu-sx)jt`%kxyFRwQK8qq}dvsny0Yv)m0RpB)$mUpu$w>^Wse zQFHKHj#oSoZOTDoxO{BC@s83p51)BRtxql#QI6biJEh1vaDV*W_R8QRWW`UG0&hpG zK-}M<5XWs^pKi&s^h5}S?)_T$tJYz^dHip?HS2_G=) zq+LCGLHFoC@j*v!mr%9#9@@(3PRwscUHQyIMJgh`T^{0>gQiAt$k6pTfBi={ap8T% ztUN2v5vNAS+E|~jaE5Lz#D@oLC!uaKs^uS&|DK;yQy&ZC+`5W<7Tv*N171HszQkUx zBJuQ&F*?Fi1s}$d+{-@nBbQ>X;)>|F-eL zz7yRY);`k%5zP;fp9Ja$Hs2QFW1f{-vnSn6@|C_4vW9S}=k&du7 z8=OVU#KooPkq=-UYgh1Y!_6rhUdnZK@9KDo{HZW`r^QuhQ5A*qWpN;56`bfxgpAbS zOL{`w4QwAcdxaMA~|0ru{eb@y5sT8)@Hb;2q;$ZS$Kds zWaEB?55%h9)$fiBYuzx#s-~Kh*x&{z*D~`6-3hZ8sW;j!l@!9C@Dd5t|emiOS3=UZx7}`6zvBtawMV=-fSLp%F z@;qPUU&Y7On4hqjPZo~p|4_9r@fbht(nZVmUecWZfBq50V8Dh&Nrsyd{5kdKX5Nd; zLrKi-Y*tj!r-#xvde?$2^`AoGkq@h<0!1!|UX z)*@j};QwIft{C9bj;!mhk1O^+ILak3@sTDPc=4Xs@$qSI^g% zmrGRxfu2}??M^2=mJVL{W39_?3&SMdd^xg37TUIwc~MZPYn8(q2PXBHOO>lNLRj!8 z$^4G?+)d3Zmt4AQdm9VQlgT{{EealZU|aTUv3oO77nccpy;S`k5jtmfk0e;IeOAZH z2&GaiOLL1}iz`U4El#1{(Usu%GhMh_wtzNqxHsnf5f3ki{e@dj=_1xJ7p~46F>mj} zGGbzv`I|}B8ullXU3!QM)I7+K?J)9JD-LjBITng|Otj|+tXmYG8nd4c}o*Ro- z;tcpPUpXr%#?pk5Ai7XH@5F_fb_+RIau1NQwXsuJRT9=?JM`w%le+`4o}$ zvL5sh!2?b7moM3JO~HIvNjqxtHCn#!^M7&=84T{0Jv^|I9$l^My=B;HEJCHmzg-C# zu#FRN_e|&yLGCkVmu6~4I#oq*LGjJaVoU8R>Se`{>(=_9SaVEREA+k;_~`W?c-nKdY_>xScy%@^|xtaRe=a`?=sOR$SDVfq#GGI))nC@Q(U1n8f6-5QbI6D{(U zA%kHjt|!K99Vs#VFgzp8!be+J2=pA(!KA$hy>7ib4eSY$^u^7Z`Tikk)()7#xY^~! z`#ii?{HeUf`o9mGBxlkq`d;{J1FP_ScD4ke5pJmkon}ug-Due}ku`Kf4E}lO2E&F7 zS`|gky9q*B^Opx$D|Lp?5oJLxZdWD4lTalalBAc^CbX zo|UoLvKNPMFf0=oC(bgJ=Z)v}aCw1Kv#S3gHgVbCwn4i0Dq|=(w(wb>U2Z-s&Z*H3 z<>{EUtMW?4pV1Ai*x!L`$FoQ+IhOFi2}q|u`9fhaX?~FCrGwCed|WP!n|?;28Hf6P z>B3mwopHAcE&G;17~`7AaO>-))-N@C$dT2>S;c$v$0mqC;%!ixQ_P`(pVi7^K=!6y zFTbr@loX>Hnj`Q6rU&)45OK#3iuCW4GGjVo68XdwAV(Wnp4q!i8h8D<@I_}~KQ;Ey z8s|I`q0B*K64|Wk;A=;5gf+i*5Hl_vRi7CC8%CV_Xg{A_*E2S=)|M%kK_Sw*EXfBZ z%aNk4FU&)3cuxE_j>M)(VzU+Xmfg>Y=GNDCw3;Hf)E> zpP&M>fXIRCS~0skw0xV1CP5MX%Op7sm*ha4L6b#_^&7cBKq+)GOn(Mm?GK4m#(Wd| zbe|FkBsUp3a*BQqAL(c)yx0ix%q%V#4!$fEYwvJLEB4L8{|K`Gt(xJp2HYN=h6f+u zl-nEbQ=uOEsWqH6uC>E0x0u8*W(>P4btNC0WCgY90cDI+G){(h1wG%wO5kigFt&~Db@MwuT!c=nx&iG-__bT`VWb0X>+LO`00I9nhM;g|LB4&#DOrF zojFV)>wN6V)=o4Y4v2qco%H}}I6xWnI}XPn&w83rPhC`+X_{GdMtq^apmL>mz=(J! zCpgGHaG#Q0e`~R|Y7r4ztCeKNv$yi>JGlkxBg@rJB(xj=72w>KV|e{%Vv26Q(mPp- zQ93;H=;4o-?l_d);${_MhWW5E;Xu+6R@5*6NGSR98egK1(m(jmhTU>;Dmd+k>~y(w zek*D=o!0U?n*Ng3&1vBLnkq&zG9asuH)hI8aOjgZQndDcQxDPe<0Tr{IToN)@X729 zN+!Fs-}x}!sEP7vi=ADmbnec~9e+jzF(EfiayosL^|wKqNz2P(bx&g%@SkJV6Vi=r z_TFSF+IsJicGf>)g`M-h?ay7RP>MPWxByhjyUuhGXz z#0V~~f6Rlw|10gtE5!c+KIW^N?hwj+7W^}wRWKC1NT$5ZW3LsuBy;g z7U;U>F7vT`##cu4TW_GdM35$Cmd{*ubp0s^jgf-`4{b}dsGG~6Nh>x?`-YE8S)vG^ z*V4EH$JGuHbj@bfvTMLnRgkGuoKSE-&uatCIX9A>Oj%(^uK(jU&f01GU}sjm0+*+v;-_OufN9>9tl+p>~?jG3vzY)qfilRP>|S#0?kI^viP65%FNk^%XqvlEW9A#3?`1eFCSIk9lrf zGXJOKGyYVj?h}aCPPo%0@elQhjUcIlx%(VK*N!Y9d#0{IF$Q_3hzj75BPLMOmD zZ36cH2vrHMts6`5FNex9055z@bUVghzfu~)WuL}vB0Q3$B)HxGaZ*&?KmI;$+j>k{ z#$l$h;o#c}wr0k)=g-RDVGqRNQO%& zG=Ip>wc#%pH;0toyK6I=c>4kBLV`zXBVK!Q;~baf!~dZKY2?jNh{NG;zUne zPoDk$56e)XO;078{OtJaIkS@TnD6k55uvjlN8yI$1$pXD1*^*64ot+bLssLIGpr-x z)OjRcflTWr1+~ei${VTh)Q_E7G-q~qq~L$FP1V2riBqv2On#JaOo!eyLIyt?A7fGT~Y@YEg2Hp|Jlkk$y%^^JIYDVrBh0NO*=_RZ7l>02BHGWC_* zb%p?I`N#t)>tfIp?@PcS4lNQc-eG$jBD}R3bm2i?=CqKIj60?ib!Fh3YH85eV=_FU zGaY()UHo5+c137Ao!c&wl$HcB&4QXKRjJh(Gj}rcI``@t)jbI`0^i~(iPVN%SS2t$ za7>YYzgM7S%71jcdCEU_^RIbe?51DWX!Q$KhZXRpSxgrv=C%Gp@~lFRn>wB;*UtCr zbDZp0|M+k{0}s;*6!;R;rx&W0z8_+V+o>KZ-HrmZtoS8m*_MXF2*d0!;{wOnLR(ek zbu^)j3B@1&MngP-OSg-9v5ncAcXP$4fct9dy&p!6>U0IEU5!}45XVoUXzXyI4b&x2 zKJ2J=50OVwt@N3=HKH&s>z6@!U2$}+LemtmW(@R1Q0oN7@V%Yd*v?f^%&Nm(b2w>*4@eH?V4*v$q8Hl` z0RJLh(t66~qgI{Nx&Etmx}M%?ciP@Zn!A$!g4fkIdQstbZ~nk|T+r^uKVs*+Cs;~M{c!9DYC8U^2%8!9iM4}k;eKpu z%4)gBW(`>aaDU`+G(!%JCbKD@}f(y2G#I*Rh|8$mi3pg3*?@x2jmF z`tExNYO=7MY`KOm)BdcCcdfkwH49f5-VUSwIhN=k0$;VMG)^HtO&5@Wt~!MZF%ppC zfH}=8y+lWY=us$zP7`Q~>7$<*cXsKW4<*-^92crRW{(%#8X=2It8iGV4uDIRSwXb7 z0ShUQ@ly^6vF%03X{fXPtgdF9P@o9lKvAWev^st4c*xW+^!)U4zuT_-|-Fr=jqWiZTOm$j5dOvfGAd*N15y zsb!G65M0R)ak2o6dr&r`|>dsGB86Zt`U5OSc{PZm*$NjPfl z1xC?|tkza4kr~4cF&uOA_aO6cwyUz1(GUEPL^b?H~ca%0-ehX?QbZ1+pX- z>6LHH-MBaF5C_QhUh*8WmIBA*t9);_uNzlMu7TmmxYEnuNnhLOHcped zahFXkTgpf}cFGb_`v$Xy#a}-)P>tzK7@A}oCkl0`J7uvo5V&!dqAIDuNY{wRS3JZ> z8tE89m5dX}OKobzew9$EO_D*ntKy0oZrCr-ip!p$ity8wC5f{;8#L9oGEIdek~Eb~SBVGG5GFC*^$Z z>&M#Bgm;-88N~|{kGQV!;!%t4Eo@bCM=hJ}`R9?o8%A2o?3hb~@DZJBK(*76ZsBA1 zK!cUz{Ho}Z+hf-*_cxVg{g%0lP&6LFPJgu``Azu3kk&N5|NSwRv_^eN*t*6UIL66~ zIf`iE$&tT+Ykz8`+ND;RrHLQ@xY1q3O_^Tidy1NrWi+C#lGhi)s6ziUSRrl{*`SXj zUNg8-am+60u2GMX?*PEneUfj7#sZfzQt?MeL~(G5PjAQ6=Dx`y#?^mgr1*9GB&$np z61FvBJaV-qc*WT2y!LrV@%2a|B}W(Fh1Pvwxv zC8bWSv$Z?zcDvX@@4 zUl~Iq4d#~=IY4LH(7gZb>ApI+uAIQC19ZDWD`$%sR4gF5dSHdGSHo@pu96U5HpVNO zSgE!mB_jEC)dZf6KQXxSvx~jO!eB_SfVf6`dwAt5LLtcu(d(tS$i49Lv9T)Xv2Omf z@~G-m{R&1gBC&oA=mncjwpJ@xG*5h2i<-nNwQ`Cha$}f=hSQq(HpK$ z3Vl{GSeV4s#APf#EqcXTb;@ej^w})8qb1M`zvL3y>6?BcJlz)|728)`<6!C_IW2dMmR_F+S22PT>-V(~^D zpzeQF1Ud}c(j0`qSs@7eL) zjEcBnsb0L1+CPn_WAa<$sHP?4WhBC{YlQC)rMbjlSLgjBBnV5ye?ctu8E9{PFvq-W zxrb%j$y>brO}P5YMoHXW`NDZW#JYDejt&xyQna8thmv*O1JF`@N%F|+jlGO^D7jSo zFA10lxg*X;eXl#ck%`9-pVqF=5$mC`$*0x(W)i*2>C?t~KV094Dd}0t8QuaFk#;`f zbgHvi)kKKX5C3AT%W2#Xc@}2IUULg$t8Je-!J@*D?}ZsqAH-JJB>z>|r1eUkVUKJn zwCYc8^PhWVv^q9IXBq9#@LGChJTqF;y%7f$2Uj|xmfiYEnYM(V=8UdkzCYbu&LxB9 zk-t)s{G0UgQ!^U1E~z04#+>+mNT*S-ay9npH*|tYota{61f|FSW=T)XsG{KSK05-3 z*falxq}a610G9}jR&6#dd%pz+g^VaOW-GtRIB8vUiw)FLYC^}mNcxRr4#M=;FuFOx z($In;3|ukJM8_zg8=NXey(J?pg^8PnfXCor3q`B+CXuNAT1C0*pQd!qRdXCWlS&m| zBVV0WSxZ9=@~pzcSYzrEq)(23vAx9KpU8nFjw2Gim!WJ}7T3%2#88&8d0Dg|J-97M zE!Q?v7@ivOw!eX1G6;T%zUr2MEZb@haS0N+RCB=u9RQDr5$i^|3FFCMS&}U5eLI?+ za(2U5Q^~0Wb|S>hykRY!+7fk)!in$rqgc|o75eIlF8u9tmUOufx#uZBYcbYom|I7C@?@}lvz2ZtQx40R1nMa*wT}uXKce$|{L?KP8 zo?74AF6Sqr6zCO&P}$hO0e8u&b{DkO<3TJo7ZN zEjRRZ-B8QYIuuverkeV|2fO;eR$jUZg46>a)9S5kEoJC0AEL zge6`_dAt9rRwLd1S7mV8lb|=hBDw)Zuh=D28~@ILq+?b&b9##x&gCaaH`dZXN=QV7 zof0YP5fzawOG{`UgWWl&jyh!B_c)6v2Xw>>Hs>lZC#X73<=hCl* zo(heM=O6Dhyh%_}NLH#L9L#U6Bl4m4-UQ+(olTJQ}|V|N;o zT0!zDPO2rO)E4k^R%Hq$KcFNu#@eEIPE_Eu1ZL1qeHZN2==)WxRf4Rd4*C)Gz3_HK zU%c7l5s!i|BbmCK}CGeAh2LhUyv@6LV ziN9+i1y8!VpNycxC8&+>O_nCOU%yRpUqGp5MoXo|!YA}z$W;iC8i5X2AyY;!{6(VJ zi(g2ecsGqY!xqrUN|jGnsQ=sa+ly(9dmS%XvT>o<(%6@UKS8!}`eZ-!kBf9r&G+C( zkGj9ew=yqL-b%x}x{8S>vL)$QOI4SgGhht1*?=6ek!M|<`3nlf&NEKM+(?_j_r?cM z2Yf9{2YM2Rx>J>HDpSujr|+i%Ki?y4BdYU$NPlBH0a@lbF^cFAQHBzC1Q?2y7(ivh zZ6$!Lhlr>Vv8dfj#`}6c8Dci5OjZ7+@Y zWJ|~%u^`#LEm`?Vlm>{`k_JnpgyjcG$q$UuPMA|-kbMnSwi>!-4S7kX;SPxf%ehk5 z!O1q(ivOf>{&CaxlfVD?NPkqEtOFL)8YmO3Nk6$0t};S4xo?+I(T*HJY^_RiWMz(q%`~$O0d*iqLpr{Pl}@Z4c+7av_oeEbD`)&@WFF_Oyf^n&Ii;#;IO zrCENRzgV!)*0w{s1e6xE%`Jh;muk;0;y}^`{;6c5qgxrJtVR@{$+a?t8IvYf0wfSv zbX^0@LX5*H+HI}c;_RF(qe62u4Ya-oK%$*(0{7nQ%`9lITQ=!>U0T&XR+wzi#r+VUUNGL8R2TKrhWO2=ZT3I|ycZ`H1DXbs9v2^eG4# zlY2&j@n~$RPyrOAUixw0B1|2GNlY0ScTkgE_vckfjS^Vo@Os%z&B=IPmCM&RW>*7<&_E`7PVts-cYM(g8o6l|US`-D z*x{X!L_ASG@PQLJ(h}IZ98%Mxu=WUmYqIk}9ly5#Q9TJoef*ws*C zCpa-o@is@vT+ZxRDGUsp^$Z0{4W`rH4u&5ibKT!}H!=A^;tJQw+L@S?L*Gd3Y3NUH zNjXJWmF?0N%H_CYz zU`ljxw(Cj>7Hc7I*BM{N33z|*{#(_<1QaNxq4jJos%Zado9&)PiHIi2#eNKp+iL7$ z2FeD?pLk@Ri174UM(iT5Xk~uP+rlR$H0m&oP>PA=3oWD7@c66{8x{l$A@kw!lp=jp zvHs9L<^>>v00i>Q&3<3m-QVFHNZXEQf?!&yQ~j!8*AjumNzxcKWGttnPgjJF=P@|= zeIi?vOmuZ{Fjin}1GXt;P;Z%mQ4&5Zkoyo8J{mECvY`e%)~KQZ&D?^LtTWAyNI z8l0x@iQR3w{^jJ^@&5&M_}^xCDLs!$w_UGSWvL*A;$^FS8)%-Stzap+2!TktGjn~& z|AZ3P+$gyTSCH@`fA3_m{MMgTLWfh-3j4Lwq#c7n=!|6%el@Y{AsP#(#Bg8PP&~vW z3Ljb`Nuxp}d+l8=li z2&&1sPa;yf;18^?yZdA0U;7Y3835nRr8UWH4qz;8uJIw<6Ac*=m;r;7Jf998ZRT~t9y zUB(VFT1E^?-~;J6O0;9eok`rN`Vxpw4HT|CpU4@;tz-{ywX6+}c+RN3kmPIo#yA}}c6%Q#=56cF`y+#g^8$>wK z^hK5yV{hLa)Wa<54A-`y0}nsb*bZ$lXJ(8 zqK4T%H#+WwM&5J(V4Uce{SL^=mkB@l=aRJ8);;Dx=;7;8IYk(N`m0NWdLb3{dGJmn zS6RzR8<%}WFctR1uf+=Xbi$l|P)z`&IEAMbb#$D_htbC+&!wgus;uc+YR?9J4v!Z9 zhZjsR$qJ*$nmhDQ)8v|gc3J^{bpHeqywvXPp-3+NKpc%%z%+_e_ zi%C~}28)*JA1yWBJ1dT*1Hey68!w)aJDr1PJe1)*Wgc&X*x@NR_A(X4!De_{VE)eY z>!bE;4qok^I6mZRZa(6VsV39dK&py*`4R4s7KMApJO@R*`IG>qr`S<;{@a2#P@*Or z0p|kfRDJWvyAhv#A2;d^`|EXwwROVCLD^ql&>Cl+aeDX=A6bt^sJJ# zHwa%bPkwU9H53O1*es@V+-OlxoYf)b25W^3q{K#|1;DJVaWHhgt*cVg-S@tlt}KN zX8yMvmTnjfSusnoIg3J===;RQ?{2&8%euWPQhhexW-km~Hxy6qPZzW6NyYwb^YUrB znp+8X6kL=GmK;=Ya75ylvnN82Oz|~YI^dJ(x>*+Bt&fr8f2jo9AsM$6jumP{zx!9@ zmKsj&;+>3@eOEw5*A&(ii=veWw#Mx) zwR=a%Xy!Q?5VhYU>CpBSoPHUz++KZ=nw0k ziPcjYA!;#tv4+w!uToaE{@R1rSTjyhLs(H@ta0VIF`f8K0uZ9qbJq_WsCue#t$g&b zCmrkN!i1EA{;bBoLVP$c?NbyE$(c=9VM51yDt#23$*wvem2>8T(J+%x`O|{&z!}Ha zHUhG`9&=>AtBY@w1?J zF$^E>UtLu!z=2HcrN_F2nwg)~DOtdXCmRv$LZ=KM#ouG2oSd^R+Dm;mP9SCJPZ6h% zYz?s7jMJoCiW~**5FUJOj)CEEAs9rg#SuL3J_F$fKJLJ( z;+p^+Hum?yq|I+%2+Ns8VafmPA#d>LPY*^0Qv+x&H~%IQmmlt|i*nXq?8H&mo;wM` zby~9Sxzb5iqkr)d?FtQ5q$-kz_=|D+r_U`3a+G?pBF=L$s934OQ@|Iqiz>MJQUBW` z55M8%qyHErcMEd@P+~Et|Lrbr3xtoBK;_AbEC0p$?e5|qg7(5Ln|JXhp?4q7fHP1~ z^)jxTWqucZf2sa1)u!wLW^tMq*ChtWg=-|L_7}e!d4tQqoKNIBJI>i zA&Q7S<*rexN^*~I2Z9Q2QliEw!uG2LpSrVy!20MRPSUDT=}}J!I@21wKdlZLX-i>4 zRXR|P>Xk9_@{ z1Wh3vjPMb$Rrpt3{i3kk0@5Y_Psum>wPawXJl>CPGw8m9r-%`*jzm;HWeB~6lCFG#n z|9+gaHLzqL=)&_dta%@D`oV_;HyS}36>4@H?p}ImRcb3ZDe7R zV+mnOMm|7HcUUTAVegP(%M<;!)Ns+XI}Ltq%N4#{{iJwog!#0uwp|dHDoSLK^C$WV z%2B_>HtyEIq5Kt{lyQ*{CqBpY;x*;#+5Kt5oB&?6(c^AjW?aKB6_GZiSl`*wuvZC0MY-&hszHmI5`tc zkGLO}PG5!o+c7 zO_k-3g<#0kZ;e%zLR7)aUeAfnKmdV=*#3ye8QLvHrk4Mq_62b6k7fH!BwEr?d~y;p zv#QlLVVC0 zaH-SX1lqL_MGtVtg&}S6j6RX;p?=$8Cjo`aHXZuBzf8}1I!aXa`n=I`&+``lioe#B zdUE-N^nvb z3OPAWyQF!**MNiz*07n%O%=Q19N>2qJ%%1uL(iqtIZ_U1UTT^`p2}FB8$gVZiAR_a zO_La!-Jh=}#FiH;m>v6mx@I`@vE1Q()Kv6C68;$L4i+G%tOyuTOdnXW{zlVlV(2q& z6qqcMy_|q5@q2MrAuvY{Q7%DB9RfQAC_Zr&2t*X~{`$Hg$bg5~ zclNJ|f4`t-GYnob9&bKc0PY~XT5v@G*}OqvQVqR{s(R|!$e(6Brv9o|_+jL&J5bzi z>1r$gsNs6i>eb!o`ZLuODf-Oc>*by&FZD+WyaV=9ObS0+7MLB{3VIb2O`U1S=U}dd zhfCh*doM9cj7|5;8s*I(G}?(8Ir`j}m#Tt+o}UeMq`WvzJ6#3GT9lXXCETsyQr}btzqn ziyb{4TKBWZu8yYDPcN0aamtC-!omE_1u=@k9>S+zu>dg=&)11xhyn4Wa*Fds{G>^i z*>u)v<^u zs8k^z=}|`SWt5H=-yrTzb|woAQG3w}9_dx$y3||!9*TbbS0l_zCb9d_U$Sw?!NGbO~F9azj5=8T`tzYXdYzp_lyFE z6_`5G>L|xSTobf8H|0nZ!<~gKUrFimuH@&WSJVj!HQmRJdroQ)WB;Cu>lFP>SvD$9 zlBSpGjX_V_pr@js2N1Ujsg`g-KS(0>so*$-rE|cG^xM&l@{?qM2nP=9fjwh1J$8@u zyW>m+?yTx}GOH$P2}HwZfsm{sxyM!W0soPx|2WnZPoa&qaoT21&2CKKO_!!)p$zM) z0XN)Bo<`qBo&BD)OH5_I180Gl^ZFO0o8fr*lcQX=99g+W?KlV)Bqq>Pj z+xYcIG?J*qiW-B7>2QKVWzQ@?>*Ur_4nNC|p<2y7c)Sp$) zZ3s5F_s`ix!S?(knw;7isysug@;71!x^G5MIdN=JFmcWkE-=@h{fgePMT-{wxDOndaXk-k97e@q`mX z$yPNiBEIR027v(^^3CDSnTae2FC|!B1>Y5n!xf= zg9KaX>UPuXSbZND4bL$xdGS^u=KSk($)J{W)lx6wghrAeWUiKrX6oEWtLb%H+HHK0ggUXBM{p5ujyDwLm+mb|0E^uS;?Xf(YHltB|_%fM(vK}U&8Edg&O zp5Rj(?2XK|U12@BCr###(8&V%moKQ8ruN6 zMVoWqdic7D`>E2qfIlPCvp>tacyY3#?&}ZFZ0dZAz*_6f>BOU;af-V2aUTq^RNsCs zTU2aC3(5hhQyX6DGW)k8+^#P>;aa{vh~=Dd4x21EO6pm}^ZW(x!FNBwklEiEE>nN- zUZDXnBV+=iLzkiGJqBuY+DUKrRfM7tvm7m%)VHnEd8zd|`3#q%?S3eWYC2o@!4lIL zwiz1}kH$F~Qloz?vtX@RNj*e1sygILKUn?NBu910rxnJoC9BjMk|`7W%eCXn#Hu9y zj?#84s%(VS`Zuogv99v7r-0-)GT@6%UTsf+louu(NtD7O3qJs(-(*;5FOMROX2(8a zo%HoLK9ll{Dm%+v{py}Z)>#%h_riBuMlc{G5}MEh-HApo~S8LBGCZbFM!ROzMbWKcFb|%(mb{BVYjB9V)-y46zdb zh#2SiFnzVd^?tu3FM&T;nw5})|Aj7^^743Rx%7&v{Blvb%9xld+H7L#|KSf#Y0$MC zT`hp0OSehw>ij32#EgFl(;#L~n;bSmubteOo-Vlacz)8>JlXPz)JJGeJ|_5g z8$P=+OY*E(M|9+E(fij&(WU~H`bLA7GTD+4S-I(NUZrfp>QPFD48U2#Xc z+df`4RH-#rJrAZTZ>KcJ9+JeiY*NCpNuK%=Vfi;-(#*9h9^LWL+9cveib;Pf-(>Zq z8L#NU8k))&E5rm&5Hy^>tVP4*x-Xn_B72o1?C!cG|4&n++m$28rVGn!>XrarzCXbl$zKw0`pW07bLhr z!cq+GRg3T;jZk%?ENj875@rXTQzx^k)yP2`M8qO- zA% zgHeRurNT@6%2LI%NvCpmbK{u*r6t_qRbt>PPY^1(NQ{b$uB4qeQ!lPGs5#Hooz2gt z?l5|*Sh$r^o1q;|ptZ?%rL_|a5hv%cIf?Q|UDdn1(_ zh_zL`5JC0zx55xR_e%IJI|jEkglL{gI_Q_r$QIK%90h1rMvw4LlpHsVB`9g#2mP}qOOYMQ4q-*ya~ z(jeH~Z?K&H+eS`T77xK54?&%yU1x@n+7Aa_@k<<~7uEx9%~vzlG34)o%!B#o8Dh&G z;)oeHvYCDeoU3k#n5O{(UNtauN(^(5pW*lvtA0gOeLB>aO3=cP(Jl%I`ojh(xflTy zMOdAWD<@II{~-lbTGME~LzE~MSNj{>gP%}ERAdJEsihOQGK#|Ky|YKPey=b2g#*OR zs`fkexf!w1l;>62l*TrqeMjNV+Gegx+~VAo36(bOlZ8CVs-jrw`qMtIW89yjMNLgl z2g1Rp!)GI=JnCZ)%4uyU{T)C&PSQ_>KTVV+;@6ebgq>ume~Mf3Il~7$N`4mM??-id z{Ol?Z>K9Mv|ET6auEU*eh&7Wz12i(rZ^LYHj_2pcX%4@!`ZZmMI zZ=jQMW>tGy7{nT+lOZ{ghCe5p}5}UkVxc>8m+{;r${d*ND=7$DY?$M2hR;yXlb|qS~3mo$T45(*B)>f+z2=0`t-_zf^+^DUl&~L>FV`o6#0eV^zp=wLr5lq_|85z z4A`QRmoZoh%cs`y`7;JbN@6ORYykwXO=w$NFq7+sB}EmF^YpRIoZTgvmBJYaP0jSR zGmsGLOAPTREh6B|7HRAo2AQam8fhWUrK}eQ7|oG)Z;V z|A}=LtEVA*A>U)p20qPlt;uWCQ{FmPJr+M~jKaKz# z_EV5C!Fbj?JJkxO;%S#)FjX0FGI6G6a2w4;VT3$?i$I=Xw~0{rGEi~NF@{5Q)WWZJh~XXa%^)IJyfp{YuaCu- zD2z%dVTKo&BHHA69wR(8nwwbJW{9VEgy``p?7$LE2fEa)wU|@;>5j z`k_S2O>@(f^(Pg$Zpj^-5+6?Az%~?j3EfCnSrlQKF)>7{ob#c1t6)1&mk@`vOc_-G z2sPKSY3#cVeY^12D#?_HaW6MGQ;k9(QIJDE`M0)WZ0cSXv{_N>XoF+a11_3M5u}qIN(~^O`%g|BtoO97>Nk&SLHBqIR_2noe*w*uu9btJY z4xJ3}>Ar-87ggQdL=K8T_-`u5&XcFnn>2GaNOQTgsvYr|2a2k)aR`4CpD#BFgv4?9 zwqcD}ac=!AYpuh?v1sX0ZTS0weQ3~Cd20J;|7_U2uWc0ye9<}0K!bg()E|rzR=RV) zle&JYD;h-pMok;ooDLjcZZWk?dDr=DSSrnE;p<>xBw;pzxB4J@GxEvMs<^5uYM6BJ zFphYXx9?y=dN>HDs*b_n6lqjT%k6M%0)RDX#eL1{g#OvqqJFj~N$JrbJTo{rxTUDO zE_Yy4gH|zFD%{~6CQxmJ$pYRxx=QLRwD*?_Q<2@Ilwk~GuxO?hDR5X??!2MoqZt!D z{%$f|;r3|QEUa$bwUSIbveZ?RBj+y$e%Q3z>pVClBO-xAMQ95H!n&#N%7s z(&DLJd#{R$9|tx309rNjg#Saaew3)lo=i?ioG1eq{L|B_~(baJ;1DH8_b zbth?NRxzmX4)XU1+V|h6sEk)Md%3dCV*-;M z{O#>XyuGvJ>VHIv(NBr%$MZSzRfs$t88&mW1M=@PuMIn+Ss9lkqNc_ga%*ff8F!h(_Za)YDW52{g7JJNEUn}M~0 zsS4Ci+oGd{>OnOaZpG)9fe&av>R6m{V%;o_^o0q&I0=jAQg?98+>7_pG&gyD?o0+x zK(K3__Ys}EkdylI!COtX8{uw7?sXYw+mer16{xlIaLubyRerE+?^g&EK1P#+!eQsK z(MXZV5c?d#{h^yWOB0yMMa+9_b7Su0ZFRxj!R+n6qu`M$vIEfj{Y$V7&y&x3?-!k~ zfg;*uAxoW6>`@Pb3Q3n>hcC^tSI9Ky9u+f@X}PrYHT6XYSjD4{n4UbhH$L61Ip^ES zhQ;U?u4|y=)yiaXLWO6gberAFr>Q*FF=zx+9T^JJ{ z!(cT^DvJ=b@vO$59eF`LiIA|eGq6>$A7>RZ8E~v10bGD45^R12@t&rkoF*o>M`Q;QPp@5jpC|-MMVTrIShe-&u3SF-=8g(^RmvpFihMTo)*6Y&wR5Vx+-{Ksqr7(zx&OP!vv0 zk>N}+{HG8qsm+KSz5c*&c(UEQ;>7lD!MsFQ^tPS?8-w2CkE_K zq#qen6stqiFlyF~%}kD)^8&3GvYjf#%Kl!gSrZ74KJ-T%ar||*nTsP&R|w1P;7|X1mgYd8G(lF*wbvdHv)FDx zubg_hd+YwZfbmoN@l)qf5iETH908uWGZptLXpCH6yhLDW%1teLcp<|`$(=3K3V{iN z9Zad@zbqk=d4)_tQ>0q$%wcsu>>R)Lo|ymL+!geyS0DDElyE5N_=7R& z;M}J!oQBT#k=XP6_6UDDKR+zKo{XfbZzrK{BKl!Dh@2ej2M{061(H5O6N+Etif7`e z%BYeMlt0qCa68p7ZOV~hjE;&}^~UXbF~BKMW4hBGa97u9AHzqX?V-1jix!SR(f{bd92(g9o9SLJOl{@4& zR~~5@TDzFalxfZWJ9Lk2@<#|*MXlCW)W~pEDN>^=t7unqn^pm+a+wJAL%&owxm#@E zKF*d;oGebLWw{)79T<81-yrLOl>b4{`G4#mY%;?;VSwR!-h}R~02C{d=L6Xea@st) zbZ2RRdr!^p1@OG%UDt=AjOijzC=`EHpdkkz=0~~gxI zsq=!|j~UQ3O~LV&>G8T^OCC3&G#YMwQ^6jif~PV?c`}jx=?0pv-+)C;Oa7JNGCd8FcB%Qm;iWZ`dZ2gk8aePQi5BjM$<5s-UNH3k`%|QN$Fs-Hc%{lk+IMjPp429D~7BwaqN60_1O$q`8p3j#_pEUTWbAvq24=8rY%hWD7D ziKHRj$M0J*#15q^cOGCBoX<75bMm51RW3N(=4U%soa)c~^IIe5TFT%$QYKJK!D0jkyo@jH z)&AEDA4auf=lA*0L&|-JvpJ3q zYU?6J@ibGhMt$R6!{+!vcGDpQD}^`4!5P45Up#Nsq=~u%Uohc()@3d1Z+Fa(~{K zA)mJ9LCxhWChVB7B`|N+w#bZJaeUrN0&<*CHl+-=m>H_wawIz}q6@w$byR{>%WooG zSkst!mAH|gwRj@gn`iwzsf$lcQ`9WTJKcaP>4}laF=ziq`QsU&0WVZu{7!*Xf0{F~ zvJd*VYd?gX7~p2 zv>S6kCit+Tg8PR4C$}MiFetk}+Ia%1JRh%GcrC(}amDlNE(Q9eHK`VpLtAS&2ILTf6-;yDN81%EOYfi*G$jZeLid8Aq!k`l*hzzhQ~L*fg5m~ z_0N!u7=1l+DuPu@&;&8vH<*X5k<05s)qYxp?L2@!xC*~zx&337UACi`r;c;<0gj-X z+$ zw|^b^hQp8dLFcl@*D>qaj{^he_%~V9e`TdKQc2&fQHnwUJ?3D19tY;5BVCi%Y>%xC z4BQ1MbK&uCZFMxvZgRmP-ca*`Wx0^~n`??29|Tc7zoU6u^&%Evv!ZbyxP1KnfkH0- z?UHQ7QTP zqz+iLK7tkmM02zPG1Bs}Z%(Tq{UuIcDm>3~Cu=$BT0OiK* z>jA_6-PXlRdg^cm%;X11P~75SiICW1;h^*fkZ1Km@J1noUT%V3 zaPNp6wk5oL3-aE5u(RF&y|2q2XfD&qbsBW{e^+}&?jDeg!P(|5hPt1-;7^-9L$es! z>hEdu686?Do=5OEHGK0DkASbm&qudr-5JeqSSKH^&C+Y|)t}(zQO@#2KE_xQ?-}uO zCKhpOHdlhD1qxE_vUL3W!0RE|MYu~-Xtq25(F;4<2Hwl!#79g>eD`q;exVbaH_!zS z&zDH_xRWhQ)5l)OpeUDoF~-RFuIL`gUC!axY|AN+--8Zn8PFw{sIu5(Ni?P+9&=%M zWU07wW)0rwk+Q~s!61ani5)s-`bfc9ih@d;pJ(~d@oxo8&Hl1zNxBqOXdEQULpW=R zgxH|of=>w5v4h7|_0dCMl5$hu-fh{kHtgBHt#O16hmguj_s7X6-G^bfjf=fhVxVina0D*!W9)y#YnHjgX3KaKXo*2WlTyJz|TZ@gAz1Z?Sy!f{^&KV1Q(?crj5jJ6cD={iFysIIfALxlc z&jG^LCpVMD8xF(PyoXx?=f%aWXVIIUW1XN)qH;D53h9kX*F2*O7b{uz00U-wp&wh^ zISjc@HB2Xm>B$!)g7)HEzO%EQQxljxi~?H>51BwP-3JgxKeEo}joS1_JzMJMhro7& zpYQu9z+M=HxHa)cUi_RTW&K3NH@-p8aqK>n%NDaK6@m<330b)2VWw8uQ@~sEv8sSl z6!QS>UB_eJOfrzfLEkv*=VTh! zngm%2DnNPu!dg2=$|;yr;* z>6@k?q?1%y!^)kPmLO1wb_7ctId2Z&i;ir`tO)8BJ zbXEh$1hb(c(p{?H*ZlV2&#`r?^x?;NbOD5;+LNhU2taVDJCT*YcC=3Af|MSRbQpHB zo_qsP;kVBFH8_s907P8wUURvwZYqm6?cAx)eJHtZqdlgiyj_a8RUof%1wZAir|`NS8@NJ{FHzIai-VyV%jv_8&j0(N^MEwv!xm4an6bTKbJFlEVk5MBSuH2dU*V!CqXe-`j1MkDxqKB8 zI>W#r9Mq$5;nNo61(`d*Dr78caVVoi@=D0+VBQ^`cn~Qa)htPnyTN@RIQT`Joe+F` zZ7SqTE9#caey^us0}{%6K9K*FpYSvip?X1lBad@M5QLhCr{%2#2HU3vR2kDOgaQ@g z)YAPa5X`EP;2f_i$2(H6P960&6uDv3iRIKq!oX^~6WUB8?ui3sK@fzP+z*j0D~-oa z);NviEns;pWx0A0di|HSqR)=4dHWkryp+Wqm$KoQW?~lC0{2KauXOVGLtg$v;Xw$x z8TwLLvEKroHypFHbYTl5=Wd$hp!aelbxYaYvMQ2?VzLqn-icb_sM7wSDPw9Y(tH$n zDn?8^zrvu`+&$0Se?XC@1w-#y!3M4DHmF2JP`5SJOLv<=v}g7`)ldlDwKj6PqySj{ z1ktQyp4P@)b_oW!;U6ZGJ$oiC?Omz@!QA`l^n}cp2Lr^^Q zn3>uK&PN*s`$(7MWKM!orse&R7C5V4EQ3=pgo-X9L#?J6e{F_4mvkxh-f0fNqM9`E7sr7n z+xdb>8+vh~%G+OWeH~=1iomgo<6c>v)zTj~hmf-)v9pP(Di3__&i@|;ny8;Lw4$+gqR8?~2i~kn9a_q>lY` z6ho#=fSr%v%h%wb1uP%7%q=Xq;YGOJ>83_^;?Ww60ut8Y2&2@EjgJXoYZLb+DhhS8 z9PaeFBf^RP7_}#z(`b66!eXeT$OL*xkN^8?{tG@)F0b`)jPn%ugZ{~aNlYqK_Ffb|TvDo7j z*SXNMaCA%_8cGa5GN+07uLri>rp%L)N~i~cZJ8Rg3pX&&aHyeGv@v5yCS;!FWo?SR zzcOaVl17SU(qU~C?mYjp#e7+ifpMwBbf;za>Lsrp#A1>+md$?uUhXcj`QJ=?dlE?7 zggjaM`RM-b$?BDW{?je_o8$kikG7M(Zx!`x{z_6#7up<2(yhitM8qVA_n}_coSUnY zK%j26z;eh4ud*LcEy}%4>hw_7hL0Qpu^(;&cc&JAzcaHYz@|r*bnsp#yH-SQ4{Q^^ zPRe~a{gi@L=6*-Phq0C#TS19=pmTeiuw^F;O=}+aFY_I(rl1i{Z`e9_hxJj{cTZ!b z{rE2ZPe<{|4xXEj`abf{Jhl^gnLMup>BY*rI-AJ2;;xHfY``LT-;h6UBtkl9!Fkmj zSW^ty`Aqc=XycB-&Vl2fU`_WM^V&mqsB6np^=Hx4xu*yXE3jF&>}j;UOBr&c+o7mA zZ>)Jrfn+=6+%1Dk4vMaXIr_Wwy9N%?6hr#OF@JRFy^2+$uV{{u3nKoHCFgl7pe(wi zYWsmV_Vh5l?6~V8TxL};##!u-SQvS6;Ki2)oDCNIVFHFnYpa0Hn!V(ln}kA3dtB@Q88 zMmc&(BR7*>uTQ7j$(w)LxAG_6Yb@4W&11i(Q1DkZ3JnU*SAJR0T|?A2yZy3)7bVY3 z+mNN$DoQiu`!y<5k7{w$l#9~%;QTMw(LJhO#pyV=p@>&~U)Jq}zCAbos6<+|p!#Tr zv(0+(HjR5#1AUoY%V{k5VsmxqDypnvFGhDHA6x3hXqVY;q;%w*#DW3o&PcT;{jNb#wQQ_h!fpBjlrjHohRVY zb)ULR*FkGB{T)pUP{%>Ca~~ZX%p-iObbU9-idhk0ExtwqO^0nM5j4Q(WY04?0yC&d z2eY_yhap3dj`dpT<%43pHMCK!I{sn*3UJ!p{*lS#AIusNjOCb zuRT?SjT$4dQhofXG*nWxXAp56^G_&3-_4#PRw7J(0LpHZhPMZ5(`;|twsMh>HjHpo zCv@TJXdz)4vFs4PhGlB~l< zUMyM2I4%}u;RY{Ir7#}i20C9uS$gRAKTi!KmQO_~uxh6JMXCRn8C-7F;QLP#nvkPI zw}YF{OXEJ5^pldIf|D31vJJxUW1th0Lj&6|Wk`usNf6#>@ z6YtvQEs!6wDUhvhU(Dy;aUkpS;5wT5b#gnmd$=roSDg6Y!>QB{@j^20bD<#P0l=ee zl{ajP>&2Z%vkNeD2RsUyjXm8y}_3A&)J=^&4tfk}ywN{d z>Q-JUQR2@g5)S2%knTt-4Y0&PC#T}*HY=`^H!3%p(n`*yNrVNB52(**53dsItV9Vs zcZ`9~qFbl)v%rS@T+`x824*bi2Mem*6(v>5J(Y?L~CaryS-Mxmhdt(c( zYKrgm8>hwh`U*1re$kmtz5RZ?{K?&J`6c{t&zflS<r6S{CX5i87YJC0_ zix*x-_l3T5wd>hu`|4M6fQ3sJ7F&<)uHa_?1+n>gujj4o?=No`t`@HCtMtzvC5pw8JlU@fv*&=k z2f2&u4|WVL%(J@NPOQxrvGx0)>(MP2JFhOCSH7C0WpCPQJ=#qc`XV38*3M3gq_5py zN&4+ttX$iV?Ug>q&dr@$Z|J(2+g5)wwK4*X#p1hLG}vYiPX}C}kKG-*wrQL%C$X_L zdm~@Av|qPgrmsJKTj#`NYw-bFSkBrIZ&{q*FE=-9 zF7DOc)_-4pJYM^_v$-A?8>I9$ytSG_c;8MaS4@3gR7`EtKYgWNzJ6Z*_2~+_-36P; z(#}Q%MCBlMl~Z4hP_O7{ecv2@vQ)M_jPNZe78g!`-3zT;&1CjM@%nV$*z}lpJ=6BY zmtXQ&3uI`zw(|HqU%A+KeCb_#-1#7|UVSV-U)H$+YYk{ffx(|<&;Y`b;$ z(=ubfR$jPF9`rSL_Y`_@JUwGvE=qM%^@Fg^z3zRgFjM^Q@;+R)pUGGx_!!&V-1L4f z2Y7vbye@vqz=k|8eje(xz0ALs8@pZmcsrd>FK18j(Jg)1)PFC+e{AT`eNLaL1hNP&+XnDx+g$5)s<|}FyFuz;McK-|)4#8s#J_ZEnO-dZ1p4;uyJ46+3igs%g}m%23e z@I|0pyt_l3ZM&;LTC|8;+A$+{9s8=fxxX*Ab$|@DS>K(#P@CL@a%Qg+M$x*ut*z5G z{iD}Cwy3+aK7Z+U>Reo{jlY)Z-<>}=zFfbs@NY9q?r@(13H<%Ak|S`q&myG0KL-=P-S_>1 zdcrrK17=rz2!c=evUZ1`UAST#{J%E`0sasYu7+(?_glR)Be>g^OEvyu1*_7*zKE>!VqPkY&Ug?@s7~F-v=tVJxB-Po_P#*o1BzD6MU0;5H0*7gMSW@q0NZ1 zE!f@mm|#fdf$JilxRQ)m)}<>jQ~qUMWPSSS*ulezEQ-H*2;6aezx1iBv)|!L?hb~v z>-#3)8jQv(-xS)!>pG^4@6p#+&xf#sllu35A7ckLf&ppF4NMaT{^Fp0*m-kH-sO#- zTtFts1v{#ZhhWhAB|kjH_i?fDrZP$f?qkfmb6Ar!GQ>z@}?&)2F)k#CNQ&-UdndQVUR-kckv3y5xWc2O*tEA13@ zmoLNzmTeovyBhLEH21ySfRII4Y-L~^8^0MFAD@W*b>vDvhS}+_&nSSrFo{;1-|>*_ z!X`p2RiAoyJTTM54UVma98eFa0ehYxf>DYKbgL(t}RAz4R?VDzlw)Z;n#5h^@hUT4`R)o zYwtR37FF1}ZYPF&#!S=^T)8R*yAq6njpNQfZr$7aqW7OvW6<%{1fSlwuwX#`P#heq z-jI!K_$04YTfMI$8wgr)3WInP_!k&du9`+HR zlf5G@)SlIZH{QM6>btFL+woH@kyCE|sk6Skrjuoy^Y6{=e7c#L>bUr$l)AvFs`2+4 zu<^8BVR*L3eL_W?9UH~TK@Dudi}D{aROsC8o{Jx@>v8#$&tpk|Bl*+%D)^nI%{jnRh@XGpTSU(+X%0pF#VwIiT>~q3#Y7Ca39w zVEeidvMv+=w-7_PE?OcM6pOD52maajM;r&M=AefJSi;Ba>)){hd?0C0gI9 z58=gnO3gM{pZ6fe9-#n00LRBP#c=aoMo3#N>-`@v^uN|yhLEn&T>QQODE^lRn*-9~ zOoHgrdusIq1_GQlbD4=tlsnWnWmm(TFwifVl}JZD>G0vUUD}cbr`Y}PuP;l6eT_J^2Vv;WuC*Z zFK20?=BGvh>uW+%9iUmk#z&)sHkg2E4A1H*2QveAL4 z;(cgn8+#woG^^NA_~~G4*Q9)DIxde>#0d~XexW4b#Y+`XTsmNp^8!zAQacW;c_F^K3p>TX<5d**-pUflF8ur|PlNGuF3ccJdKF3}*R z_9b@|x?qX;#7QUv(SIFhvR)Bi!B>I(+`v(IN^UZr3P8QzLB!J#kJYpg<6A)pG<2+Z zMs|};P3UsHTRS0b7hxFZ2zO6*JBRa{U4X{^Mi4H$1SyYmjBV^3pI{7$A{Naqd$eJP_5%e9p7GtLW zV`_MvB(7l;4QAPMPqGNBjmZx!ERzh|XCFylGt9<+ZVww_!9RPw{>bBJvvJqQ&^ zmW?$;UuJ|K%!!MS;Tx_trYDV7eofOg?@M60C|9RRXb*|fQC}B^MEnE>f;I9$V$HAN zlqFJ+;O7zY{t6F%Z;-lgvf3=w&CueL$=0VP?Ab-sIWA_1@UOXFTgQ@}r{e=&3Iow~ z6Mga}3!dizR({xr?Ud&&xo)=(Y{-fKiZkIz3%xK~g&8Y7k17%|YR&}x|aYiIE_-wZx-T?mu7P(60@_QHz*2w}ddhHcw-S6j3VWLNMV zj)`Cr|11*Utd<0|V?U5bIXIqBM14#|MZVwuYF6B}{6cu`*@!!OZM#s;QJvWoC_8>9c zNw_1VyYk=!4(T2s=60CZ55a`oeUiO)iB8L)G}o^nf`UI|ukKg#jUrE=s^bfGO=|`h zJMo+J#BcA-xZNTzI=shmk~kIqAyR!M3c36ZL7EBe4^W4aZAS z-{I>EfApFvVoEhhzckUR@$0*DW4nDsV5DX#{Q0FJzqxGEbSf5 zF~<@rfwbDQ7+1xn#GV9_+;tJQZk^%IK7KChSol5!U6(cKBoN=Y63)3+{$ajlky0Px z50qf0*ugg*nfI7NEzp>xmt-GN?P`Pk4cuI}Jz9RV`imU)1U4yRzUX;<=st)>U(fAy zONdx*QI~K_gB(>lByqWO(W%($GF+d%7rwFE;+NEs*>}tBc`#B?YCDu7L=ZWNFCresy(oeW_g{l`;!cXo^%lSl z;o-2WRp&z8#B=fL`YDS}5Q-Iabmp)V4x$(YxYv3#jk_ECDKd;V2OoDU1V3Mz^$*tW%@<+#l$b#4}KD;gx%FO+gbQIjsmHK z{dE*(kbw;kBub*FP;yP)JuY@Ki1NpCLQg9H-!VUN)@#x_pk;~;8WW20*BPqCHwhR; zPl&GA77oCt8YF=Z6c$9ADfz{f*SPFxS#AB!vFDFE??iHRZUQm>UKZl0pl3h&Xy?8F*L$C%$G*b4j|T>v^`8F7pnX!%{RQuj`+ zAEOT_xgY{!j)sZ)@8S3qePxI|*PQr$UU-@Qr2mm>1K1&wdw?|CmD}Q9NCh(aWG>Qt z0$&l7MX4A^tD9m6`PP^0=k7bMN4ev?q--mJT@YrdL|^=G$7lyBEU{$BJ zdy?Lwb$!br+ZOc;2uu&uCLQz^h_)aaF#!@QvM>a+r75jOt6+2y@F@ZA`9zdhj}CJ0 z+y;y{xtmB|S`ABjG4(YP_d7hyOh$H_VCUoF6LY;VdwtnIer|3a0}0dn@yro!x@*%& zop7h19(@l)7?RVgRab6Gahg`n1| zk;`E1Y6#21oOga-^dzpjAM;g;|LPkVjZ6Y!sT!1t=n_v_RBNh_9Ob9-~Ee@?gg9G;GX zMfy*|CKT-?oc?MeDpw5cJUx85`C2+xQQi_5;>;EVaxW$em)a0FlrybWh4(AWWS?Ch z1jdq$re~{Bo-sO6)P<$LB&xg=(Di5pH4{@o*XqeKlc%LqyQHPb{|8Y(uD^&JL6Y7W zENyhXkyv|eSqYpNMH87Tr_h^QWamK@4Pt0fYse>Q^5q1n#AD?W{v1Txbn@*yY$Duh zq?8Uh?q6wEq4waa_| z%}Z9Te@GuCCWH5Rc9v%&AnA-a83hhB37&)p{_TdqLA3VfTrCM&J+L@696;A37=s`q zu2Kpc5wd`O<^dFvgVEHzw-%6HK(nV#0e{`z(U$rk)YuZpVZZFl`^#`6C2?L2rWx}+ zej0|Ar{2tph<>EgegrVlaB73t4mOzsp%H=#BmxA8>nMVD8{ozt8xi+0`vQ`6Mlgg5 zSWh;}aynE<^1ge=c-sTmX7_^Z2=IIM4(c{u&>Qb#0tMA9C>x?)-F-&OJJysi^N+?3 zd!c=pdH|6(y(xMXz`6Jhp(U_wQ?b2S-R}`JRcvd^6Fx?oEKE2EgU;e*BOt@3N+2y^ zRk{=UkeL*FDgImwX+aQu`oh}xS%tl(0VBz_WL z79}5aq_$i(tHu2l$YsdpgeLv2#S^mlg@iuCF_5SrYMHR81>9*0BSpo1MxuPs229x~ zfV~G))(XTNJ@p|W_Sda&BP^!9B2oE@52eTr!Cno`oru>&@`D~lW0GaN_Q&@^UqjIc z`X5a5iJ!W?pL3KXLbqlhZ$w36E6VCbn2b!HQXAOw3p5+Qzat)vN%}-6M8<@PW+QV& zR){S0gG!p09wB>>eBbUjh-7Z>cg@vXJs9(cZelIj2_IGQQl)}F84eXlr0=CkH zC#r9@U}Z15Rdq#J1uYb-XkC$`aoC*@I(z>h>&<>*j*%oYqY)nSwqCM4 z7Uy!`T&-R*&!DVppI{q6v0tAJYH~~X(HPYW&GM?ixANAj2H5V?Shb;B9m2Qi-7|Jt;zKUc+JN#yNWH+Ff4L>=TOM1Jy!k>z8l=QT3txSBosI*);&~B-?$*GAxrp%)36Yjf z9wWo8hPYd@R8ew=1Q4Po;Dfgb2hWk=#XI!qX_Ef6>0oGUpf4CfV;3np#>I%u0N4lu zC{vb*HNd;W`TjuqUW4~oJf<#m0qkCEqPByU#xewKX{wb1Oz*K^-VUPJ-n2@*O z>nIK1swvij$nM#oE5~I9U1M^yG8o!9Zj6Qvg%<$4v_~9I7rijqXbnAN$RSRwIk$CqK%?x)Jx7 z7PKJk?LC-jVt@AcmqhMmE+3)0vi6NcZs@_@cxvwfBn6+z3Q0ePqKPeOH}#a#rq^%N zf`WyTJ_A!oHipGifUq6wl^?*}1GVE{6K6n>ow6G$0c=MkYuvHkQ+l81T%GDK#s1^u!}s?0@w(g$U$2Avp9QSAp5py z{ED%(4{fEF&F$R=tf9o~{9nQgz$gK{xw{7`d82LqvuzWH@1OJFgCATuee$2=Ng5Ji z2MUxB5g@~nU$h62ppDqt1CSNUkAg7@$eq+1Rm0`DFwhHvIIRBr7jL49^Ej|`Edlq$Xd@qqGZubLg1+d$rr9^~^HPTIz)ZIPEp2LuNN5?_&h3LVm3(d@?kVZ20> z$s&%_wZxi#fHa(}IFZK!uw%)=y?`@uA83WNaMleAA8FNvVZ-d^>ilDSv*d%hA$kvR z4UzDl@EogB5|BYFKrE~~F;;tY|+ zhyv!QiJEDFkU0VK&G~2WW@VYl?bY^z!Nb1m#VzUz$eb%R^kBt9 z0?HARKY(|WsDs&z>x|8~2xpOkX5Z}46Bv0hI8T)m1Y^DDV$0c!AY(_tKTXWCy)Y69?5I((sP*}5-RzSyH37H!41b}^9Z1=Oa;3_iwHI{}TTlWC(GLTio{}OYK<_@1! zjRRFEqgd?cw`@hjwJ2<_pP2{Z`a zkzEdojDgkx|#FQ^BLPDI4PF5qbRCH{TcDKXFIILm{@%D}R9Yq>Q}ZS&}ny*@_(}9WP%GX@IteKmW@= z{}Y+`yd&2Fz+)|8O2C6#&t!vlU{R6T6T(-GP$qC2LHP=FR1!?3We49?!%h=u{EhIW z(|aNw7B|$R#V_qXF~9*Y62>TEMF5%;v>W#iVYq>fpciY2(ovTfcrY<6GH490M?8wi`bPo4sm0_Qs5Vwa}MeSgY#=!qQz5~2c!&* zKjvkm^4^5CfKvjI;+Q92bvKi!1U3pI67v#{@3nLUf!OJBRDr?MIk+SPHY=hIrDhf= zJb+-Cp69qBp#ger?Kq5LgI3ED$zC{n{ z_J%wzMi@T@q$h^nVvlzO)arT>>?@dZAxQmc+iFKxT>=0#F%B%ms?ep{gtC{Yrg~Z` zHj+m;j`1Fx*hq{E)KFT)1|x3x#akmcLmFkjS6fxVXzN6J0u^^x4;ThyL?VYZX(5m) zVCZzi^yvgSqZB&;9+b`5_-u zH5tfXZg2h_#_~h{J@$*t(v>Fu4jy%548B7iLuf1V%V&?q-s#I1+&IKfy}0s!7k3x_ z?-eNB{O|XJkLSPo)xeF(-@AyXQx}j& zdPmLESkQ)AT$*4u-r*m*>)4K2Zp%!(3s>9>8yWM1PRAEU3QvMank*3_Z$%7Zp`vx9 zZPDK@ABgh}Z^j^y9Ylui6s3tU>B?F8`^Ks@0>83i(pXp%)=q<J`v_BG$iksKXi~OZ*V1SybWDs#3ei}jcCP?pGKB1`8Y%_%2vIJ4JWBF+P-t-cN? z<0JV7dHu3m!6ph6sIBB8gO|;i$A5mPnJwZR``z+M!jVu#Q$t-Rl!$Pl`~XLO12r$dztb3`>m~LZ1n8 z5e|mTY=Riq41NtglBMuqA4AuPBRCOe zFC*^h_>p>rvN91RQhWiUFJo4Y1-N$`@qi+sycHgb_>02vl)EykO;=ixrW~P`(%CDG1CkJROn{&E4fIKT&A-Y_z_S6DgLA>YmR^lTafdCkXLyTxYEvH0r{f0=W4&g-VZ}gwXEDO6@?H;Kr^wsO5rzm^`wo{+NS zeoypZQDjAz9t7&87!3=i=%!536=U@L1*)$ca@7P)(E^qC{a_69Zb=)Em z+0fXHvI19tnEpE}rf9N7bP;Lyz%wSIp<=oZob-WKLc))%dWfluvM>CLpRqvyjj|yq zr%dZi1}YVJHp)k1Lk9L!a=tA$8X~5W>hAza{X3Y^{P4opftURrf(z_Crw8;Ye!i$C zhEDzt$*~_^eEj|51(U{O3@ZbWoTg-P(AoEcRViH#K~9CD5bL-Gfwh=3aj9av4L^peNL5R3p_pV2t8jrd85 z$F5Lf)P)=x9s}i85lKngJja7OM&ThG+&`c%5T7(wA)t7h59wKQNWVkNBqW;oP3{cr zj_XDg8F*SeJ3fF=GRQn$5gu194j(=aLrU%$NElLBV##srF7X{gy?PpTH%DD!{)#15L&eV8hwlYPWw8x6N za^a7--=G&)6!n64H~6_@-0K4;1s|D!`KODXF^2%p>NUhH-{1PbAk6*Kxu5Qj_66AY zF+(~|7}Ckv#7#U~3JfWHRT)zISw2Yr^M&D0W!O$8T}CWJBu{dxvmGbJ`k&6h(R=ZJ z`>6qMR+90va~e+ovA5g+cj5YlGP%wkWa4k>*&qC(1CLiuhT$Y48g708*#%k@%SLfQ zSVDq`7;%fPs|`%vUiQAzr;wKE+uiEZ*BUVpoLsT{8Hkz-{q5-77m=0yD#S=&M?Q>x z;!l1!_sxyn{jbm@EFW`vp6sxx{r+L^ml+wCQ87}p%PeL2U}^AC4co)C>D+TeoBwmK zv(gOKHk$?;mwOO&)!u-NmW&WTa$bQIyTmMnnWXFj$W?(I6dDkao}wTduthLBAW0s` zJl73Gm%81sR|>xVLEP87hV5+EzT4ihnD2bOYnLxrHT*LN1Sh(}MiT%J(h_E-+|BPx7(|RUL$>q(fDkF#>X8n zmqEAJatfUm7u6Lya;shqO}&HIKu9-F)Gkx5zqOtpG6yT}zi)?DP#wiq5^Xe6?jj3{ z@!HWvZSj7u@v7M9wi7M?LOll75_O&Z=xGAF)S7JE%1YPi_8zGV;FO5^qnI=LG%vui&#&r}as3J7j zPmPPt&Iy5K4jWK8a70S{LHQs;#$$7TpcH?`GufRixDzy^OrIfFmrWLWc(no zzTWb#z8I+U+pw5M;nI+)p@`=~m%>``vQ?a)N8*x# zt^sMQD|rx{%;7~pjO%@Sh32kk#UZ!^^m{RU&|Cggg0+jEUsTYeAtFL$v;?}|1_Ev> zyaG1PUobKpfcU?(d)5?gNTb^PLIN3Nd_^JA1`eEz7bTd&_ESs7WR*w)iPmjN5P@7a zY3-4HOD7c&s5KfvF;HxI1PZgT?~~PW+|r(X6{STZcp%@i;5YXauheQmW1O&LVnfMSjt|Tl1%*3z0S}ruQ`Nx36FDarS zrWM;>U5Wb?hq^~$kvcbkVGA01BfSNK5f9@KQPKsPhci!60uVt0QS>6t1%e_Xjo}k~ zGa!!3`R&#H4$(e{&qxf7j_R7}Xi`F##1Wd09HJ48ybijmo7Q?}ljcbo37TQT@wvTczj(5E3~J0X z5}aT2)?Y{@9~88h$Yof$x(8lU-bKUdEtAZhp^W6y3J1Y5rl9!H%Z3n<66Gx@3MDWZ zCNw0vg7q@dG!ND>SKva>T8$jB6jw?*1}HO+0^PGNHbnk)qkK%A`^bJmvWbI_9cF*p zoO1+u*qq-FD7b9{3;e+`8oj~2C;FyN^koXt7|epyH5wE=qhVQSjIU#m1I-{oc-WZm zwHO9lgxlik$TUPlS%9pCpTY^ua>)|w1zLJScnWlB#za}37y@q=g6KCKGi-Cd1w;3J z1AYt??g9nofWvy6?S(+R3L$GT!y#xAzyX%RUV!iG9YT$jlaP2Q0pbwf3=2@uFdCz| zxW)#-AgwSyAl{iD`=GY}EaMcgzIgeW@^r8s0fTjZTn}VSd!`sX--soU7_0RDj3fp=1QF|&&H!+sU*d|5tsBH-(Zt0 ztwodW5K$xv`cIG{<~xPyy4`?ck$byJHFWIKEq?dok)-YY(4!QM!i-3G|2OEBGL!>y z$C)g4A=sGc-Zy=rpfy+?a>3nMh|=VoPZE(@c~-jeY=R;+U>nuT7|FGqV)M`mNdy8M zm51ObfN!@Bq3s(nKrj(k#GExyRT>7vPmZS6auhnHn^=p9(%2Cz&&gkf6T@ouojsTh zG8acqSoH1(8`Nk58?=O(#~?cR362MlM$ySwLKE?GB)^80);u&QKZ)E`UQv0v zz4=}w9&6)~`3bk7QQdEX$I-?pd4_qFJ7~9~d=Tvpf^PEtPF`Gt1&*d8q^E{DPOQ9$o;Qr5U60iQ(0U90ieg#dw_210?ZvDDjnD!67Z< z)FcC&=Z2FTSc5c@I1VVfizvNV7@z1l0%8sUGSTs8Vi6PrVahX$QPJ47OOmN}pgOVm zs{qY;usHt20bC zydokMQNEeTVLlnsH|#`iT;est1f}GX0f^oLK<{MaaugU7ky3{c!_5OG;Rr*w*O-fD zV`IC6ZMfLW#PzGW@vxbasYnP+KK8c@4u%4xD5v6k2cEI1-zf*c zBja?)q~Z7BA7{R_?dL=#(3#$HQf%c?tKh2k7uR~<*<6gOYmy6za|~{D%wTjb{ry8 zbLuM#bSO5FC{;}IrvM22@{|RAzJr)CA_&KILBnlAzTx1L@6(WsiWeN)tt(h(Ry##y zE-yS|if>0ih_f-Q$OpR*AfP!>2mGOvSCTFPh}EA-g@6+S!B#1~6Olv1kb7#QW=B0s zJ96d#amLK|5{{+e(9n>1VtsCsGOF8Ih~VqutQW;-nxA2pqO2E~0*ys7HUNF@cv}F0 z6SDRNWl0a;%Xr~}!~`CwaBMMzCh_blL=^D#pwr^8zFRF#8#@oN?$f(p7Hn zcbCmx@Zi!@m`j|naE7RR&LF>?1n^z#2zY{$i@a;r=l`Ef7KW|~A||%2EVG=z@m#D6iwX_M@Ih;ww79!Y?5Qx&_$}z0QQ3TH4Wr=rzVM4u`K3b|>czYB z#flR7p(FlXbNg<<=1ik(GSp8Ri>gda9r@!MPV^hT8QE*T(2zGM#Z-&bK81BKZvGJL z4qs3Y(Q*rYph$!;K^0JQ`+_-umSyRIf{QZ5Tm<#4Ly{0-8b-Y937TNCk1uXS0&Rq8aH>%%~FWI<`)T70^%jk4=cM9)8oNAZ&Bhj5a-34^_F=3_o~d?)5~lO zi>U$Cw4J;c_N=Z+?b%tAh^b7_JcHo8)t#ikgwx1CL{Jq$qRPmaMi7TJ5+Z@S(gd}U z2C~rnc5zmU_6mIDA&@NvTR`KAl4pveVwkj7WSor2oHai6Iizfrk(ik#tqh8w4oKR9 znF~enq!%b4Hrao&^%h0um`W8#{KG1B;9)Z-Voc+C7*|*v4X^!QgBf0?CRQigC`p>X3!slQnR-Biz39-68WKUKFR_6x6NOE6 zR=Mh|>;#1k+AI>pY}brp1zf_G52X8tv zmeqZPGKi8;cIMWqaJD&faa)Tiw?^q*hvfBfyk*C>0%U$h76oJk7|Xaeq*+R2rY;9- z7I?v$1jqZlvHLv9-Si4Q)6`RHlqM8pRMzx5s|73d1-kE}`UX?fE-*!{0?h61PHZVz z#R>S3*_}(9TB@j$G$a?4w_+8^DL;_S31mlBpn}tWH0_J1ZGs0c20|ZWj7W@5OzwjR zypl~!kyae$WQ(FiJEcX(#~+`m=BWgNTdbaGfj82-L>mh(~XXd6XPYW-sBL0{Ch)jWl{(T>{PTNZqo;^t7Z4krNpd&7J6dD_Phmd9|l- zMP89GX#|I517M)J3&;u-Po^X)lR1_2TsC2lGk|Hy$r42Pgef*41=2lrl}y8<6e{$z zUrLenH4P-^U?qPi(E*1N7aEzwGLZC-0}LSf29uVuJ*tscNY}1!BqB|74Gvg&LH3*L z^S=|d#>g--NGb#v_Z>`ED+>#IT9qMD(6o?x&=98aX=Zxoo|7d4cm5K6KhSD3_CggZ z2H;BYS|QhzkTnXq!dpXKUbeUGov8eP>66$j)WZ7V6@_V-jK>|W9d@R^j^Q6JLi`C} z+Vc|4XQJo6{9eng-hdp}<;IG-E62pN<#lxC7I`o$DPcVa?SJQ6u zx=q2;>gEUk_r>fAT)F(XFn=*i2>#wL&tm7R5_>eX z1P2=APhQ`wui6E0d;jo1xi`bSm~?uhk_!LnA71Rx2MkhAL2Kh8okMA9I4(Oa>Cy6S z+`#04{sbBX-=;tLwL~4%{!S7k+EiiOY=NQa<7UN8>l$J`FCJdzAmBI~jw^MFgcKDa z?5tsX1?CP0MB zV%sItO+$7A_z=ZkNKAe%jeTAZ?ZuHFKs<|LGDtLK*zHxHL7o}=pB@-oS}FIQ`@!}1Y9=EC3B@|@N|rBrtxbbx_r*! zySs%axmuCxWgLG1Z{5H5+ZyDi8TC7+mC3$Pvn;nPWNxgC)TSs!G0L7Sri+?C+0k;)s)Dn92*gsMbZqgm>7U~Amv*UwKO4gAlZ9_$Q=e&-=k^l#6rHY zQ^*&D=C}t01d{&ykZ0OBa-vcl0uJnho-8BC0T5y$FB6F!M_wj4X(UH+n0{x&^sCLN z5#HTj-37yVvSE-f-tXFmzly+{l?)=N=6NFiWwY2_bNZfz2aFGUO4D)8u?0BBaGF1i zl%x7?6J=%b=M={FhY0lJ$2?TcBoIZ2=zvE4>?KC7rfhR-{nj|7fz^7Dt4#0?ze8=> zXgc9!J$b?Q;y-hu&h0I@gmcT7Y!=c{OqJAr4!VM-=VpGYzG-6!|4m zmY158GBG`F942K?hR4`=5D8Ouqs zoW_2@UAh$I_U90KC1f1K!D7GyC()uqNEK3biCSC8F?F$ChoLwb3doH~QU+<<6p#Cr zXsElZZeNLdMC)JM+3;@A!W)u@03t5%1t2Bi2+2|b!p8!tK`Uj<;4So(TH?baaUPQp zwJF);WSo$0a|rGPdkSsy0Gq`MisHzYl(aBJgJP3>$Sv^_ z<9gt#?a7o*cRe^a+`b~4p}*aJhNOfxG16ga)X2iUwS@=xHm3kvAaBSg6d*bWvB2-6 zBnqMnq3OOy!XINJ6O$%|BT%5m>VV)&%qZDAlxMNU@NiQX*^X6bHt-2-C@?*y@>Pg` z;zEe?8*Co-$Y(TeqW!}*U)&?GF$L>VbSfdM|9i}r$APCBnTs=-i^$e@bzrJBxT~(1 z82FAW;R32yn$Z)k6p_#RJi8cqITzOz-FJQOFi??igW1jz2Qm4Lo&hxM2>df<4C5GNvxor0cxNf? zhns%a1c^B1#?>Q+p@;{Pqjzyd2IMH69;`JG+|>gb6i7@YcH0wU2+RnEvHy(p@q0#u z=tHokFjhARew)<-0<1QcUgTu-xX{^A&)+C+$p>*qk+93^5ga=NV~Vg5(-O!Fr>?<` ziMZ0|8{Br>f?pA@h1!deZ4oW|IXFTbZX8t#Im8KIGXogsvRUt%23*L%hzRaSW(*B5 z+Dw1qHK}DmD(qHlbQp!;5fW}f_jgONF>qHy|6unwxm1$~nx$@mGjefe-^ z9(4ue1eHgjXLEV35#eP<+t`fB6&~(kJ-0s>R(Odsw_(Hg_@#U}jo$)XDJh7Mhl%E4 z)0qagO_4xfQNU>%eFTGDVD7Pr;6LimuG3cxuTB3H6#Lg>cg%Z}g5X=U#1!E2N!wZvZk(1mZ)cFI4 z^Wfk#-7J$Fk%4l7A#?$_P3Y1|5Oj{lR7@wdIn+OoE3IAHN*t@!Y*sSu?RQ;YNK-#yNkj;{eV6<}K2sO&<}lmkhehb_#IzN3bNVvLDQ{&!1VM zvsr&;q`*;TAmjo^9zvt=13B9$RSmHMx-gortRk!xP}@h%fTQZmXeYCKP*7+Emiz~g zip5Tbf+vJ*Q!E}>+=*5AMpP0S?+TKYaZClxA%vO$QdI%vvIJ~T@RT+jC>@ZAqo^fD z_wv%?apoL@+HpL}fH*V?B{IfmN_I#jOS?2SVz;!84RYTK>obVjDwc3?@07{p;) z3J`3XJ}gW$YXGYn*E|oMItDBxCUF$myRjd01j^XHzg?z;VjZb!F|$0tp2cViK8q zJQ~T90c*X=R#9woxZ8K|d^%uau8cO|92Ye zUOdA&W}DHsEcLLPr*WGu{It?@K%0}TIf6UhMJ(|CJy`ZvDjy1v>Y~J zhj&V`qVOxjNr@kdIVa6#ZzoO}r_A*-PM~V6cx^BfMTdQ2mU44WhI)pB17kXPyWgFs z!!|KHml?Kc+`~4VMQoFC0b{m_GS{f!wL;ny_uYuKi$Vz?=G!3qL~$e6fpH>Z5Z$7d zX`-m|Vf6Kyzh`Z#4`c#Lt0fGKKr?+v3nmlG6goRp-r$rY1qnlEOcHPAEOiS4kKogE#Jy9Hn0(sR;o7>FX$cT ze1(({HqHUWI6+fAq)90O*miK(zfnwnG4Q_q7<6+Wk&hzKD5jz-aYGZhH~}Mf_z7`k0$IXvLaUD!}%=<^P0q>^#bL;ca#K|RXm*8Vcm&l8xgS*QSAOU zShpaifh;zRr`Xv;tVe>)G*YD?_}rMH#eM)OM)!^%uXF2S#&C+ z276MAWeRgJfmS}rTwflqvx4-rQ65%&gQJkeDOZc}Fr*?C2bcb2S7!z4``nYVT@{$$ z#=Ik&%Y!vo{XYrVD@dP}rp4MwS}ZprEs{x#l@IM3KUCukwo#d7AQT^GC=@C(6bcRD z$5W5_m_$WXJLX3K5DoEfo_$X?w#BeCX?XxcSe%{UM-(UbVzrTAwUwZ>z=W>Se5Vxa z>6Z}aL^;$E#o@nF?wCauh}(k60)=xxX~77*Uea(66yJ(Mp%D#iQZ$P?exWIvLka|& z(%~#k>5!(O1BU*l-7`vN5B~O?!j4iHe;Vpf9E}C+xG_L@gw`=~2GJsS_Z+GP3}Ppe z08+!gbn@k0kX(iW>4_pj zNY8R>U4+1d9r(VU5+GM3i5<;jLmftu$2uQ(#X2%(fENPMV*Z^jFmtZLX^3Y zd^1s#4w@?tyC#_YkYXHOHHwLmDkjvQM=_JQuHVs36Pj#AGFX}IJIdI_D*=Yj6SA1Y z!=G5Bf6ubUr9v!cL6$YK!b}*Xuj>=8j)QPDp1-A@#eg{}y&v#1!AD!ja5RiYpWdh$ z%1|PKb7^P;8w-EVI{{&E)qHZ)gN_!v|OoGMlvgOl7+O+oxbwvoiTL z*mQm}CimqB6SpId)dX)#Vjc~mLNHx~ruUZ{wh!uLPDPM`8CqHi6tP7cW3s?m-S5}| zX472VLL7Au-9luJ9CTlifG_v|Iry8U@zh%ku|w&^qn$%%4VFeuH%QNoBzk-Esl5s6 z5w_%!weebuB#Msb>j%oZz=N!K`%0am)pZ+8!kBVznj` z2|lj-yWRaIrA>iPzaod^pZIx#?OQ+|0}d$EdTib(<;^6uaq*+FD`07w4_hU`Pi=)%f&PBBUWzXG}?wKo#}n0E}PhR8NnQ)z0f z4IZhW0G2@3%q4?liuTJ|L^As*i!VujVBxZXxNRCkbxg@(8lkK+{@@Qbw8&V-tre^t z#remk=eB`E2Ao!9wN*Ujx#w{lRsEVHs$ZBmAtti^h$>TR1B83GNb0?BAidO5d|D*Q ztVmG{_F0YDXALxuo~1vnE{Li(!N!Dd0Q=mFfm-g8$vJ)dLuaNwH$o_i!lPrBz`K#9 z&lAj8wbTSYEw+ZP;*3f@s~Y@Ml60WTC;+|tM>{Q?`4+?QQT|=EKquvSm>`<BIq1!RX=yNr7w3-$j6s+!P)CWhPzR zh0LiG$hgzJ&OAeZgBk0Po44+XnSf+<@)R80>3rsfIhw3eylB{}g2chY83Ra*g!0|H zO|w|O;0~D&2K#-2Wa(F;fJ81NA~a?S>7YS2j=8=%md0q`g!@Y8TLXSI=kKo;Hy>Z_ z8;D(d*GNSC0jK0ykg5$n24A|Pupg{xy^Ina7!#!72$OEX;E8GN#IP94YLu3r(IGB$ zfG&)%6Ok=)WeefdRcS{C6eO4ioPqiT;*E!toSEcq*w3Co8&h)d|GQtm{h3xQ4zkFU zRAvbCYl>Y-dEqO^`NWP}$7@u<4ZF3(2Khv0-rkWl_XIthn9?zxTCrOJWzU*)`o+XIfAaBT+bcw zRjAKT9BOk#N-x&6ko%_+T9Iq|D(IU@C?J^tTtK705*SLOeoyDnZ=cjw=p`9TL-+d4 zhLBQ8&$aP81;Ic#Dm@4WU4kHiG`fY|G?HyR1cWp`tRwbeC0pn!2*+Cr!)44X)H=6viY{$jX6&>N>Jy|-lS*(e&7@%_OjiDY8tww?i>Z~E# zhJTa$IFk9(`hc(nEcoxjbW5txjk5_3+dGVqUS+I9 z|J#T65d9GioJ8KmHeur-c^6H(5av6~{y<*)$Wx^{lj5*r5+y0S^i=6EM;RfJsJ^gt zzWVluTR(ArhZfZkX~V|z6vN7YMLK7tPzXh}pQ1i9tyUaZtoH}nctPQ$qJ~dYqeO`IfV-^t7i<9k+ zAcqZuL%PE|gTa&oPKhL^1vgBHK{KX!2P6*Rv{JqevLr?L*`t4{(B=sqn^t28DnFJ9 z&OsMRtx9-!Qsx_vSJSG-_oi0^+DF;EfmftsmscU)^T^w|{i-C;!dwInhG)gXeGQ)sMelVDbd)d_~Dz zvFB@}N$0-fGt&~NW_boXp$~*OHVW?)NcRbuMrXm{Hy(Nl z@G^x=h~&%r_PV{ourZkTg_RJ(+qVqzFx4hS((!%UH(eD=p4w0k{eKSr<{cyIitXQC zL(G;tiP=Jg2AS7bagNwizh;dYOaE9zz3Pv)ccI}U*Kr=M9SM%fAtC4}k!W{HwOjT?L6)r}eo3PMbwV&Q-x^C@Ob+f;>XJD!l}6bM z&Jsf>ngIA3eu@n5dgLz~7lOGP^5Hb}ct#F{juIWU3qMS3JrIA=oV(g6YuqYMYkS&7ikK)fc zIXHgE3Nv2~)}|lph|No9BV|>&ITE{Sw0AUSaBSPmv7~rp>A;w33aQo5mO6-Djzn|{ zjc{zj*YWJ)ZZ>1$6$TMMv2c$_NJ=~m$c6iUF^p;qH@+j@JERwHL#j?zs3HH)0uGN= z$IUM-8I|Wf{Z;E0kmrx5lFt>xEc9;ULglV8X1wLF@~EQRfcp{|?_!8pj$%uZEnDgQ z_6i?aiqcqK7S9!@d82*jiV6)rC>mBO2uPP^WWt*wP+*lA(oL*2y59?lrVczLfnP^H zWt@_CoMjd{_xvDYH{g_Oz5?SlfhKsD5GQj zRyiUYSRn_|W(wtUiDv`$6u|`9`Q0Wq590Gmap;nJKnexVU^@?T=2iK-y5FmeUs1kpBai~U6YA)E(;q?maY(vs7%2#4*>RzSd8#m=7VY@nF^K0nTeq=z9!@Sb z)#@}v^kjgk6h|(fVyJ1jBC^`ImbGt+Ok$U%f=R!W6tiFeB*ViY267UJK!}l4q&EkMi3rZ&R@^sN zF!>u&-Skrikx71sgqt!gL&;|)Wei%e4GRo2Z528MBFUD-1y2Q5m*XUI3UAweUXT&S zReF6xVz_vK9YBua77_Sr458+kR#R^pwCh)3j7H=gN%NV)lr&{) zOWJ+Y@$bVPd}Kh_?j`P@j;9CCpF?kr^VR*0dbX4qhso9`iX1VM@D!!uj`u*0Rvdi48bY39BS9x641}Kc$8XhanW8*RK@?JT@|bm^En7_4(4=wt zccN@l|5GKg-60yAgN+5h8jLuPCR`xp`CIjgA=-o~Tt6A4O}>W2xM2k##xUYRj(Q8E zx8qGRd;1&E(sPQJmZ)Uhy>WaUHqvFnCIt={W`BY@82}J$^CvK(Nufw?DJzZh^TO96VohT%6j_SAwb&z7)GlwE>i3`5Se0msuPH4P@PS?Al7+-uq36&I>UQRPHo6i$bRO{*t*1vJOz2T8^7kuD!BnqJ@~D z5Nfd1gcDW|gB7TU@o(5*>Ewtt%17+{nL1+moy?A$KTI- zw)=kIa2REVp8Yp>NJhdJaeJ7dptZ0c4ay)Xk;F#!ki$(x#<WjDcxa8wKOOgAsR zh3Y;+ytE9!QOLZ!z=R~A2T7cEUdMc))X3O zRS$8l#CcK_N-*&h0D;_>4>7C%l5J%TC=_UctfJR>t+GkOPX1=w)YJotGsJh zXlOKY`DhAo16sDQ_zLhAz`-j`#+Fa=@FXImpI=DQJoF^MJ8Do(Si@7&l@8z%yPou) za*jAn5I_Lwz2v{3ZL?%L`os2qcLR~#p+D0SErAvf7Fpzxg?_dA|m?r?3^y46Fgm(kjofIS? zwu>;vV;jyIK~lCDy0cojI0zb$%EvTI=L@zK0sC6=*}5^N783XD6`CobO%EeJ&)cz3 zcSqJN2&XzqTx)SkS(nfhIH*xEgbOD;CpwuWl~E>4vQc%MYrPmEWfw~VQTCcoUkysQ zz$kL$q=FvbFu`zQ*ODZmVH#T8gwNZibou?)Z$iJ%ZOj59UEAV26J99y#m*rpNrDR4{^ zB)(bPj8xErs461&Nn?@w{$U_lmO8d4XjUW!5`^txzwlCiv~hQYWcY6Py^JbOh}V(LdS`oGYd^#x8!K6{NQI!j4gC* z6fyz~5G%4KUA9(}LAGZk@n9RVBnz54U1$^~^;CcpRGjt24WO7vSLAlR{k+$n4%U4d zqA1oxgNAn)%>bSGkcx-dZpp}w*KW!CthK&pm-jo8K1Wte#?)?jrovxn39M;WHXMSZ za0u8Vl!mZ(Xy5(<8axLqiJ*?oLLeF z@6}KPJ&wspr$p;2u`d32zPZ1KRJfFcBWx9-_Xo>Okp zshZ~nfsqCA8c5Zk36-jKB56^gx;Ty&1sgCn%W~dA7%h<5UVvb6lm=6T_WpClfSZ2owqNv|ZPuxaS>OUq1cajY0C^wAq3`#~M^YjaT z?Sf~jbtD8IzGVu8^_w5W{Ed|`c$lIIsUPKraQcb$@ z5OoDv{2Y}K4@uJ>Ow&%<#%K=QI9M8l0BWuH_zO9?=Lk?5ksb*?IRpRtfBoCP=Gnjf z>-gUgAh1+;C@GHCX|x<{TVW2z84!oAwMBd;$+xk%qGy@nAC8bU(e%xp!H01Uqz@&&kne~#G3K&!vVRDd_@q;e0!8?^^9mAb+|$h#<1l(xaM7Q%ueZwZUyDDX5I1W5v9dfNm^gr!mpMzLfTlDks<0E5JjboFZiL5b?$`i?xxNQYWww@d&=7|hBzTgG zo>>+XG2eIl|A(YCo6;rbYtUexL6|6LQ{V-NxDea4cVzW$@&l%3c-a{nZ)u*w2T4Kz zbRb6VavfdB^57ch38O(W4klnp#zD4UH{yodlHCPjHLEj`SS~>s4!1t1St|+(&Gz{2 zD_r91VYS>6ik&?jZm*~EPXd~#fbbeZf`-X88X13<5Qw$7v)M1yT)#P0a#4C z)8j&F* z+r&vCzwIZcOpCG_^;4$B6m{djhylRaD8o3x4?m2?9MzPh%&GE!IGM!cVy6UEP{np7!4**5`(GHUIbZ6#U8n#3I& zGZWJa6hDtEUdAMq>L)SXc!(~(k7{McBy@=^EqEksnlL@d(qfE(6Cbg(NG%9PDX%m3 z|2g_TZ0u+ z1iDC2h!~8u0WM19iP`{k*&|gAIpB4}3s&m=7eaA?Q`A7?aBJEC7Aj*+04 z$-E=6Jczax&Kx@s%juO{zV66=iKP3ywHQre?y`kfYiO|r=N!NsuJ}bss!MUQmA=3L zV0Rj!j$>RjCc}VJmj?F0c>kdOUnbSsSw5LkD_{oVqVSv1p404e6vb&&u^~k9mLuLE zXrDx&U1GKL&{zA&T4K)uj3b9f)O$pz*E&4n)U8lLrY36x@g>ka#6dFXG)nS8fE7q4Rm1lM}d#djJaW;30yQQhiD!Nzs`}{ z3@K0{H*P2alXWl*mQcbgBD;SwS*$bO`<&tN<8BOaGz_7UH{c>tZ6CX!j=H7P4^)xZf*baM$NwwoKY`rAbI zZBXs8 z8^q!0(9ZIPq`iKRHW#!Rz`XpDRJ&Iw#R0V?C0T+nZ72k}*{-ErH+w%N*=bSkTM>FsK4+4?T>CK9`G zjHV|n7(L;U9GY39`SrLW4tvG{I@wTTunrL`MUtcL@?e?9Du^@~b%5=|5k^i4pEiN8 zaY5H1%bur`VWfD^GxxV9tPEx~1lycegcw~F9Hj>Ul@S@Pt(2pLND9r(HvZEPg18Ce~uTTI!EXN0Io`+RA1eoNj|NU}%4@M9_{KM~l^V4sC`g!hQ z(OVLr52w>%hv{)eo)Nw%86nuk3`03#R#Jm`g%i-f6}Zv$n_rCzlfH}K+kk6I*m_El;{U5=L_vMh8hoUNinKl}Z_|JCd#bLk1=2)g@!_(s_ zCJgS`-k@_lj($1@7Zqx+Gf952itHND$4xMf9%X1tD6i>10y z(F7ghE6j3x zgI3w&Vp4H@(GLVyg_wGIJR>0M_FI04`gV==r zzZy&cCrXk(juJM|SmR6hN)rt5FSom0v#02M%RtVdQ%MrYR~TEnztQMD+CHxcDNFX6 z#U;if+(Z8Q2c$}oC=%J~z~spSn^YL(^5y+)bl^Cl3o3bGaT~LFmx;q?CjREUU1W;Z z(fOtG6Nh^dirEGK(4!8`tV}k}lp*N}4<$*v5<;r5EClLO{^3}Z~av6;WuWMSS} zkT!}Z=}1$}ETrf03QnT+Gg!xaZmluybTEw=0p)KOoD0PF2`gsJO^B6+HNT$yHj;*5 zr4_{7@q>5zA(EpjB%-xhhs<-FvLthu=?nCKh1kh{YM1v2IE}DdQX>6Kc0doep7U&j zVnuWX;qz;o_lZK}QtpZa^qc)w$C3qptUaGi%Z)lM6ZJRGO10BlN1i1?xym3+Dbf|8 z>Al1$m&pLQbF-kN5X@yyM1~R(+fht88|QK@0gNHw2|^Tvn;)&gr^Z|OytR9k^LUdJ zBfy|w4u+hgtgUneA#LzQS!;H{III})Fc3c^uOzrhX32B%S3qT^NEpLXsFqvt0310N ztaR)X(VAjC`->3oV6xVdHk|QdGEU#s5fT^TuVex-`lE>FkW}ok@@IHBu-W&78zkt? zN`PZ9XoQKt222D^vJrr$2pjOJ<>x$B5cy3w1{iFgpj2=L4VwLZEY3q`BGd})QYje9 zSb-B&G4NeuguQfh2jtBq>%HJi>*0WW)_Jgbf<|_4g@i!&l*!NK<#5u^OnDLnsE=D5 z3R)kb1 z*?y+9p}c#kQ^t;o3_ikxmZR9WK>i~W0Gk?#TXYX9oOKTxjz(471aTNI)w#tegf@;e zLevrCKWe>hkf`ze1+zTENi{u5?Zc)42h={p>1k@8f?}se&N!yKZ0ch?0g~<(6pQpx z)1zWIecTZTJ2c8sj0+a(5HZPr3EGriUHkzfj{O#v>6hQYblgDLOUM<=fiFc+$)U-J zDtjQjx6<}#6ME(vtSL-u7DVr?MoUL8_eYdAeISzC&4`gIs*jToJ&kv>xNg8tbX1NG-nUN;1y?A z0@_qz5#gwypK*TiD(Yw_KN>g|6q41@cjaTy(_j`FF?=H3LZBaqPRn?y~!}V7X^`?BsP4bM7Yhh&>!SdF2Rk}P~6OeIHX85 zxG@J*?!XEJB zL6I_=4>ah)H<+irhwYm35tJ^nqE-Y13`Q}X1vfh5R8EtuRB-6 zTMC2FXUL1F4so8KI>(_9!qXI*;bkxwU?(yb=h1Y$(+i9-;EcSF)fXFf=%*>+m_I zA2Ag;Atr~s_`p*f1AoeM>L!C+Q!>#t=;QD?_1X^~i;d{) z$H=(govIFBii9Rkzg7@{Ob#Sbspx#|7a5EhkfZuXVjKA=Dj6(&@?aYtmw-qS-`4=| zx(FF#0|*0{@k^}wCPW?co*~Y(AHa?$3(XrmsbxLmhrKiMVEqU55!actB4$}Jzk$hY zDZS@(!y)GE?UjiGuWmBF(RF(xHgb}~D0W!lj|8z0fEUorsIMncQfixf5GA`K(Q?irlU9IolLHZN9On84 zpKszSis8T3Y8!MA>^Z~(+-*5{VaY1lgxzc@zndm6M%%7w0E{^9$KQUp*&~%6*}ci zuMA&y7KLpfJ~T8Gb<8ohDZo9Y^C^x(r3|Twq@aEX#B3s8vOYHptz&~tfMI=rhxA(k z^eXY3`hhzbIWmgGi$-`ao+N&b(xgXrdFgYIkZGKU00x}$E(sR1m(XY8iQ6Vo)2|mx z3SO&1GSRwNTVl#D6QxcdTazd^tBu@=DI7N3Ye;Mi0k5&gOt_D-K)4Ufl1E~F$U^34 zNg@6;Lw8WxeBQKIjaqp1cyPC|2RB98O28Ax7_3k^OuojLwl*^38=1tki*_%rfW&Y} zwomr|Jl||T4}yiT=gq}?a}~uGBj07@k=Vk+ikLXcJE{!$0Wg=R&@qX3W{)e%D@=x> z9-dLEV@=Y~da{4hTsMkVXb(g~XLnu<{AW$pMp0HUqz5A5DKiipZk(1Ei~|v`Vc0i| zB2aRdMp#WaFL8 zm6?$w$2aCgb-JN(VtPa~Xm^Xp?2jysW>|C6m%LyWXH?yE)&HUs6j#d`j!sYO=O>6z)z`%vSH% zMD2P^jzXL*Co1(l91(rM>cKS)8;N~Sp} zL9PWvS61JySxF)z$;^!=Xv_ERT^RdAU%Jj0ydw!Rp`|dctUJGZxgjRW#7n?1Oq#)G zu|#+e2R%|Q!T#4Q<99pKixX!I*~$DuW1xbHcArHRI-27-Xsa{PNt?LNj5}>lHWbf9 z6!IB@t|Zqksj^qQEl7GZ&{4ngxEmFWeikcqWt>qcVHSSCZhiv1cA z$D$!jk}NSv_yB`rX&?m$qDUKemeCMy)CFac;kLW8lan5ynL(!5+ z8NskFhQ6*7#u)&6h-wWvve5IWKu64>5Up#RGz>tr2YKwxwr|mNSTEtU^3wZh7_za& zCM^8gF7iXXkLmdX^KcgC!4+eV0kOmX$WyTRg%VhRws<_rC0YC?>OS>eyW*GkP~X+E zmSbFLr3<& zcwKc804>v$H$8cgVW@wkfwSX@1{%l4ND4uXB2U7YQ-a~MRFeXYUnZt|ogEmUb&{gz zc86H+7g}T+V(iK&NtEO$8IIT_{7rxsxNs}xZtwVS2?+v+LxM z29Ow|mj)7|x&mjQ0ZVWQ`q+GCa&a|Cpqba`jk5RMvzma$KTdgOc5_@uo&t z5;_Da!!UmzcxNt!<$j6AP8HcS6FWqL9?JP;h3Gu*yzHa~1r&A}4{AJ|Qj1|!@6Lx5 z%3ER_f8^j21YgzPzxlKfODdQh-;D*4#G_9y-7y*Oi_LFSm!gD|L589M7#z)+s$wjz%Q8xUf@F%;M* zh_*f%vyMgs>;b#90~6s(2^t#w$TEmaOF^e>l+114`mvwbiXnukG$8_B=vt#mRQ;^5 zL0%e(k=n*wD(GfH>)Z{Cskv!`w}XE~_UwXY#QSDc#tlnAskQoh;P=NEf;l{fbl$&vqV9*h znJWG)N)wIw(#Zs-vjQta`x7P;=Ms)(QDJZn5X`VI(sYuA(>NSwAOb0>F=Wyhml+6! z%jsBRI#d{(Ia;$H@^XAp3^V@`0P6ngVg#)$&*5h>%|L1nKhsQdLk!UwIGoz-RW^pt zX0LMmZuY7e!&}*_5`Xho!!bnL$+5gSfA#9s1<*GB1Fx|8Jj?u*1Q0_H-gH`iz(@ZC z-?*UT3Rl^y;e{tK;3s><-yC1$@MxY*tFzGvBep?B{)+oDz8K^_C3Z_l$g-3nb}~XLyYEEj^yAtjy^AE)`<{>vtc0b0+2Uu#cLCyWo;q4Ca_-I ztXV+hSM7hns;$m%u2GyuMkPw}ODektBl$xXPmEckVAP|ms|7|+-z>-lf%zLDRIn98 zsC>^-@S6vWKnsk{#zq*S7yRVGFvK2wPqmoDn{@>%`Y<~ZR1@~M{pki{CXsA(0ooFb z;1ERUGA)|-j`9s-rzES(5?WFQC-7$pBN-AdihyhtWM`O5o=j`N!KaWIAsY_+5~OC! zv{gz1hVByY+q<1*$0RWUIt$*50O+#aUEW{gZXuTxU89c!+3j4T*j5Yy zuG{5H+b4AL`yig_sT7eVg<$`;6vYI!lZ@>ZN~=+m_*n>%B~W&ryE7qi(ByzEgf2H* zvG@`hwye~51b6x9pniaiMObKzi4{>gLqc-S-L&Zh+6f_YNo*67=PV%)p^hog@8`t> z#UU33$5P@Iqbxl`8z!d#kD4e_axk~hKIDNfXaL3RC7wAr$nQVw zW@mJ8?_o^uuO8wLNY;qKswZ3}HX!qUrfL8JBDYuDyF0S*XX`2=Sk)>IyxNITpcG#> zpSD+@NVb^FJ3oPEktEBm>)YFMQg!s`3;8Z_4S)&q)Zk}SZb=5O%(21N)nXr6ZU^m1 z6Dg@|x?=%MpLon_nYdP81r%W1-*2VZAB16L5!Y8wSF* zOW9YPata7!3P@|0<5kC0tV-o1U;^*pZVbLY`7D}{@lKY!GhVMA0omH883yG0&akOn8_69g|G!^uk+cTi``4i z9}Agd*^~Fy15up~v11{l9r(U@s9wL!#e@Y!1|gM)9~TZzTnN|#$Jc~+#RP@SY~Yp| z7WMHX1)TyQm3H%EUmwrN*2ZCL3Ayj9f=ni|xLKZuM`%eHS-&0J6DNXFZ78D;dAO0Jjx*+MfOptlB$tiX)FFJ=> zxwV?tu(3&_NE>?)a%|jTOy`F0LwA{9&)hJ_5LUk1AVR+e4o_6!>UE>S_7&v`a2$kE zKWXD^jybA0j*^88XdN<~W)4risx=YorEw()LZZoUoFB1OQRt(2w}XH0&^72Kx3W}HFQsKW1F{D0 z21m3887cHGCRdhBqj9nVj=X_JluIPvC9&tq24dNV4AU)`HQZk}Vjmxc##|;U&%N4D z`HcW}ZdKd8RpkRcA|Txfa&jv8Vx5Rpbl-ClOVT4?aYY`{y=lWc&2UWCpK;tMV<+hz>C7{MGi9$1rmG-!_m@DRXS?gbJ?_rA6~2 zzQj%R;ryn#lsq`P2J1K%563zejb%2FU}-r>oR6zwG2Kt_Ky|6n5}D-T1AMAfdyFbr zTpLN<6ME`4Cdff>r)C{(!@5z@j3yQenBB7=7lyU5ugFuTlj!Zu!+VOpD%N5X^4}P9 zh^!uE!*_#85&W152Hjn6B`8&hMRRq3&4dUUDx*f?+~$)Y_rp)^GL9*O%wrpTK_~?4 z$av?3Xw8y{M?hNahw?qymKkf6kqM)P5g?<^`0rLzKwFodcPR=~Kmb7hfgpJX;QsRb zJJQ*e1pL1rO!XTs`i6EN^RgBuroa$(*~i~61RzpGpBodcmK{!o2foAne8#t4r#uK9 zC~kGg54V=pU3EN!j%J=KCP>h<;EKNPI$e*?#8?lc9Xzh#OE8F~gQ_H~Jo(oA>HH=7c|z_*Brs`WB(e*| zxP+HAc8+;=(n>#-5NN*2gmx(r{la6NRGlbNtBUL_k^qNRaBM`wk=xHWklt$> zVc$%_I74L79|9lUdh%)Xz_IZAl5n)hkzkis;_;%j^A3&9;hN@|N9T9@uq!f-Iu7Gg za&|l(4~?}mr(dNL6salaHBC6LArPu=2<{)<^rX~qvY@(WbX;ZGyJn%kv}#Cr ztINNHkdi%~u~0O7H9{dN2vcn;g|HFeQKGd!Iv{Ej(L=2b{hdyIhjo98(R`YgBMQpz`d5fsrmh3c{z#WSq4 zx?mZ}zvieA9y6=b{BmhAjt6&p&Hfcsj^#9?7XQ(q1-or5|2K_`-e(gpnBgJPkdBx( zCmd)!D;Xz&suV+p9ZI2sCYW1_U8^0PdX636W7Q}`Sv2fQ?r~DWu23fLgq%d18YW{m zO2&Gwj_iq6PRNObP6e7Nb@I7P$j%QvS>SUDaTaWRkhsn(437&FGGCB16@!Ud7DfrN zwOFzB84rP8vfW`!q6Mb=02|juWJ9;vn%?IH=G8Gb9GIf&@(7_^S&mL8s z@9a646ox#%B$mL$0^O;(mzXrJ?kJi2rfV|bH$KH`6Z6GkMWln8Ctai^ja>}c)W!3z*_1Rye#65i45#j2N08Kl%}gZRPn0!yPrQ`aMBZsMW+Uq) zbCfkLI9Mm|H7_E#f5LJ2Ma>Z&(2ZVb?i~aM0=^cFAbp!CD+TGUNJW1^>JPe1omK0O zzEOB^hTbvIFKmgrMtf7jG|P8MsM=|UKrFDcBc3)=IythAp<&@bHQaXke*SMviTAvV zJ$W@KVO9CTL(~2V{^D0;T(dXz(06uO=~{ZxtP3 zu5kij(r)i}WT%YLp&T{aaF|{ZtK2O<5>2oXgR0f}yLWbrfV&Vp@&Ylx)6rarsu{(g zQJQ*)>4EwP5s+{=q&vvAqizQ+Nz?ZY8isGk(3hJc-;ezuvQE|+S$C%b_0$>W*816h zWGg<7G|J`|5IsW64zyB00$?_I`ev)E(TS6Ys|E@@G#ec^5S%ZZ&1KpTo)X;iYmCuu zLQ_LhQq;3hq%Q<-_j^fKxCVPvX?PRc;5I>u;McMDkb7ua8VL`IF}G-T*~uJ@Z*_7M zfa4AtjRFE9$cyV4da;28cB>FijSwVK2jXZn;seQj~>xg^}AlM}~YSSwN!$Vi?34MYp?lPqEWriLTnc=COHywYd5C zQd|K*vyXzm-4dOD*gk+`rlIN*UcLao)DPz`NW_F@+_zBi9YS8HyKe3_7_E|Syg*+c z68K#aa>Ly|Ij#B=6&J zDfaj`5(NRFie^BG1pPqpywfqhh&++}t9l1GXZ5Vv2!j$V9-=U^GVz!@fQ`Y`f`;Cp zJf-N78V<-gncQEuchL7`9M+Xpv!1HEYB(Fg`usIn>HB2>677wsr@5CN!(O2SD0oLe z5OJ~o{|HNY?Wg|w{yKOBh_TEE@Rguj4sTINqz|aN-l2&Qze6WqY)OdSZ$9y2Codx< zeRCPVd%az1#R@B$B?J@UckvnYO|UOvR41z8xJ|Rg9uw3~D+uHaaGiomfclyff{C7p z9CK_Wc=k{WQe798{S`Ch?fGxokId1%z73|Xe%etr^%m%Y*Gepg8*k8xU(5{H^-krq zY;ptnOpqQGnWy8l6-FYB4&6kJzXufU!S61%lv&G4FtB>X*5sbl+FpyJxeN?_>?I5q zQuNo5^<>>hlVqRjRusd8ZE#qR*9XNfI39qX`wv0}p#3i*SC=Z-kTN1V$L?Xz!@Yuy z(2~6t3!jZK28Ccp$kobcC^E?IzF}J%;Hn^hu%B!of$hcvX+${`_{+yIfVfrIUKYV~QpzK$R6jbwzM=Yd_RL!6o%EbI^VYRm8r-rEp+0}5X!ZQS}n4};AbCkxQ z2ih_Rkx-EzA|i6tp&oDI~wqTX^h2Rx{9*HIHazQ4d8wrtttd z^=kc!Y03RPXzx*H%0YV9I|$8y>I}4loJ3=`ak7V5Dh|&5e-8d;slyV0AokEEk}u9` zfk%`Sm*Mn?c9peIdJ@rMF$&?zQKIokv4I=rg3vB z2F{8!giA=7kvB2w;e$qrLsGkSPeQC$q%;F1KQ=NjCYXcR7=sdU!amVxx9h$X+>%mC z1QCt|5#>!1~oiD*)*WwG0 zqI&rtCAZ-4CzL?SFTN9UcX6dzog&Soagtq8Op`MbIKbio_-M~1EH^J@;Ct9jLqwU> zj}^Q7780S67!?%%3-lz2|9+3U@eLVBF%Q?qTUL|U02N{a3W|xvIySLb3W4S6K05q? zHl&oLq$!AfpwaJvyb4nu&>;`rg0t9X|79lDJV*nmC^i%Odq8=?_yUv_aC|0od`xC~ zSOF8;Iz-GSqenhBn3fAzN>V^{jUI<*9HQ%nJc?Cgieisz*^~_~ucjeeoqud`nUU=O zv{>K!N!;&=Fm65KUVPfNOX95DgbirOLJ!OjL#s(-M6dxSEeNQ{IU&t~6Sme2UP6H! z2)>8}I4yNDd?grk@hwk>_{4+}Ei_!sQe2)sa|7s|rYS`!*vs~wtw_L1HBd>mpoq7rf(qk@7KpGW>^?fG%s zCXpE@Q^UbS`;`QV`gF-uD&medx>Q`I$8D6LQ0BrN;_<}m4xeSSE&T?7FaB@>`4%Li zXKPQ*pmJ_Z8%RxIf=H{X#TtmFh)BYTFBv>Z^&Ow!YM*Xcb35;|$6?C6Z1`1)xu<(v zU2_Fyx5HnacHQwfwVaNM!O>rET&s%{I-`r1n1TqnYsraHyQ_cr<(@LvAQMpDyYs4n z*!dwDa&*tgylPlS^GHt8fjaZnzdYUVI_h(PzA(MW7lDp2QrnVZhI4yr9Y%xzQKgXZ zQc@{!CRmcez;ErpPSVX9kE7}FNg7dF4*gRiEfs{ zoZ=U~s9$N6Rx}XE*A9ST%>5F)L3YpCPeDGP60rZl*9Nyft*oMGhsx5OBdEO`50*f)Nykp~S*^UJ ztLfRdc@qu~#_kd2?TQsxU493)(-_hq)W_^SO$f!r89s6CV>JvKuCE7ZhG@;6bui%- zC|^knBmR+9j`-a*ui-cKQQl`sSix|R%z9g;nE|Tr2Yes;i9UGPPZN1u>Kwh|Yg}^+ zRr}S}A|!968C1e0#tv9!ynWm=ZSgfgBx~?^Uz{JfW7zd3XG`3t<-*$|T+ix`HqI z)k{yom@K5N0a_v-1bvpNyxMt-Xg3SQ!l>-vHfL;$f6wTIynVcdbASM2%XorE?d{v= z!?44@my2r~e+gxYQ}gh0Z@hEPL%L#YvC&9M=8;C2F91K=Xbr!L#eSj<^MROosiR=# zZDI84nj%*UzH?mbfY6*seUYsPqw++j_fBd(a7gx<#%GOJG{!j4s5bw(LotSj^q5ZW zkcV*4)KkBsXUhwjQb9t8R&&-+k8=Apx5g!K$_=?HBBE9JP_ub?nJ42Un#Uh&wM$g;wYNNt2bFs$4!v@RN3x^kS+Yp($W8^y!Y&IPvFnM zNn&{~PuXm$^P94{M%@#~^VhP*CYQ3t!o{>P@ek9}Mtwf%wDL)>Z)d0NIGxQoOi9nE z%CXP1ihlzAEa5YyaWM$3R8wggkZ(KsF8xEvkZECJZNY8nuJuEaCQDP}z}oJw*2=$lD{YVd1X^vkLUS&krkZ@gSC zZt1Kmt6Rb!SIsRQ4fX8y8pq77ux*!P060L$zfn8myB4KYAjMMlJ2c5nCYbRU*;d<;RL4O}6KHFDaF&u2P)d zsOHPKQD{W-i26%vGe>qsoy&&@V%;g;xO4^BD+hE?$AWl4cbgj&BgH}SW%bJMKJ$uL zk(@1;(PKi}r@gvf#nTKcBAQiC9Y_eCRWG z8VLB9&i{v#~8!V>9;Dd!&_ z`E1|blnMYvOo(Gzm!nrt0(RLtUmC4h5QB3%rqsB5k)}eDaRzE8$(oU`DZ{b?k>bIp zD`c@rg-hb3psN`E5rkZ=YL;5+G&b-P0w4Er0rpbSo)gzj;EUL?&n)d%Fc+{h^ltiI z@WWCdm0o&xchZNxn{Kr76;mHy^r%*n4!Fhp+q3sa%YOz{U<+rllr0-?DaQDJ*c>%l-B5w!rZCY{CfmxJ4LhN|oZwD|Ec`6X_|9gX{|sI!}r-h1E+*)R}I-|4=RhnntG z4Y;DPeyVNH8JlL-Tq+DD?dN>Szbj2Hq&zv?I$vbB{(4egh&E??!5-Ttd#tFkVk~FL zcl;Y3OKO&6HE?nz?Y`gJvc0z2dzMtx&S!k-#dIo8d@z#9YCg`LB0N2A|9O(l*m>Fc@QNwiOeedf?E6E z>5qy*Acm1fDwXU9EpJ{}1xRtbshNkI@xArc}s|r%> z<&#!>Sz`i+M**E7X-04zgtJ64uAco-&;kLIrOn(<7xaskGh={R3dE*L5zid$B>CH< z(ki+>z%9e7;U>;Hv-X;BtlCUQrXRlLvZLiZE^K^bL zFA}VZ$w+lIL_un#D5pKgldm)^il7D}x^bIuj>8Hnml>(#eot+RmsmaWE)i^j0iV;s zDZ=c3%8#B*?{tns(8Ndj6N=VWlI zt`X=HYs?f3E<_MQrZru(M)WZVE{!g^Pt}v2%5H)3(Z6^0BgzrErr#v5+l z3b8v5@@3U;2Cdhz-*Jc5k{`k7{k0qq8A7red;9A6C(zGojF23TLhSMQNFK_8P?&Ri z{P(F5fJ~irtv~EvaJsB5s>aC=3HP;SwBm+>n42J}TwC#^JVCh1bU~wMFJgjd-ggY+ z07nbe+3Qv(BgJ@>!}as9<`g!eHL8~>{u?*}N-e_78cWH)xupgCPC|Q6tk7~o-*2q= zB#1!t7{r)R&z>nw0Nxm}O}2VKy$P%A$AHSptUIlkmY)QxlS&`RoI?DfDPf@XO@5~U zZ2=W)^w;^;G?|La4UhqU>F7bQ+TPJ*!=i{rXNfe=ZiZR#Ij5PPaSZ|iE3VQd z`)t&BgBk`%`ZNE1$q79z-5VZrwKJ)}c{hoM`5t z`@E&GYgs8IWtNlR`wX;H2-kyPKN=0?W7l)WjO5pBC$`G4l_LC)!pum+txu14o?gI-|5~#EG;%6I%@jyNJg(x$|bK z$w<_Cb;AHUaqLkFpv#EM)C)bsgFWaRMJ_Kma(44Ox4ohG7;y0pY~1?p_1*McMpDsm zqO-u87vbYqQXAKV+|MiHfcJLGv|~q2(F|!3uX6eJLsnx}NVI!Cr#Q!uZ4@P9uPXf%3qs>4-KxG!3{{>n1 zJfY|?zV)?vka#E=H^)=AS~Eln;g!Wi{kE1>ZIZ&awdOA){V1penM%edqk0b5doTAE zhGA<#fPhlgPdtWcTzuNnz2tOjYzECrxysi`UXc#OGsef+2<$~7$1}=yBUgCMnRHoh zDZ}~_*3XRCjztx@-Fh)ZX`mNTm&K*tb;1J;P4?)tb;~xDVpl9M5{70kXHacu&P+?8 z97|~I1Jnp}9B9iSS5hs(tf*zEw7BtU(2}v+%kUYw@ieuB@9ij-R*z$80y6(dYo;GQ zC$=(I(i5q)!IgIGG)$C*bQpkW?7!J`aat>q-BPdVLyusGGPh5z6*Y1X3LXacx)2d> zEiyBlris+!Gm4j|y9qsNMuOBsR0+y^N3A7ihOvUR^=DRY)49k}=R`yHd> zYUaIuJ|b6)^2X!R?zD{E)|L20Ez^*RygM!UhYiN8M3xhp*jcXyP%dc4o=Tf?0@zb!Nwo&(h ze(V^DY+%Dj_+jcAhMPhA->lc!OT9i#ZWL*DagS`215|yw=1L(@>GZ5>qoWBTR+N%+ z>S(1huS0LOLVZp@k)r3X4}$;~x+{lj(AV~3R5hrH$BiMAJ+M=mh)-&=MVNF|nE-BC z@x+j^a|`)ixsv3FSU73SjF{E5AC20ek|i6pL?V_RaEWvTRc>41MY2kPJ5}*YkM2q+Z-G8o7CDI8{K4F#MGCNA}JKef@vsp z)h`-XX%-W)-?)}-hhB}TQ&NI%o@UpoP;#UvA(5W2=7Q^g!5STos-lf-FyLhFcSF?U z_B&C(dbT!EAGn77$cd!M!2=~kiqszOUykmj7B0S|ODx%?&@>BYuQg(67}w62iYAf# zZLg{lWov0t!@1q#&4ddUgkcNV2p;o;E%gX^Q>!)O}%k1m_RmWquXJ^|1E_ z+j=YIc0lrevtX-mpYTkyFR5-Fom?SJ7N9sA_P~0DqfHJYx@9ezTdF_0(L$WDy}Q>l zk=}cV0mZgp`+dtV`IOqS30(h|{JTV(1p24ThnwfMUnetqIP+Ecm1v|eZ2@CW7tJGV zRL$~-vKF>didZ<#xqEMRiw;@FaOx711_!rpQ;Y9;#)#pgXdqqU{dAaX5n6E;Y;q;b z#SureQ|c{1vUpvA^D_#kqZ+->yC@IywBAd*oHWa-WL!C45_e*4c<#`^rDAwy1FI5yy-q18SWb%M*Mu z8_sB8saz7#yQ6aGw@0@1HfF;g6%8omEQ~#1Pq;%x`*l}?77<;EpOan}4~Zf5lVVhW z(UVw!j+`g%aH-LmGgb{syHzD;W3&+3z!Vl|ze37Ma&Gq8yBbGyU^ORcsg?iia+yC% z8sQkfUXLSMP5cw+XPrm11`BcmcBXCSO>WpSp`GJEDO}mz;R(RQls8Ie`%zqT$-e~7 zLYGX)av;mm!i&|Hafyf+&!~#KqG{RulgJlbdX^<9!M_w*u<^>WU-{{lS3nuvDYR7; zD9Pk==`jLNr(vf-Fu^tVI^o?YzZA3tad{-M8B%gk5nO}nC zNkyW3sE7PBfgz((^gT;uRxSwMYk-wzQDq9eXPC*xCnHZio3nVoz!(xYf2&vdXgIcN1xmhl#a8$i`!?`J5-z<}um7X*1lc~d-XDg3k z4WF4WQe$+W7Av9JeooU5dy6{Xd#B2!6}?gyA1uqQs&RTURfX6QTTR>}*|iC8)G;bm z5i~v&k%DoWZ@IWd^4ELP5M83Hr=!!~O512r%m|>nw;G5@9`q**S$PKrh!7)AA$Kn}_MK>vT;MZb zXSl#A$WjJ1i~^UGsy$=Z!f<%eeSRazpwJ1z|4(paXC6A`Y2F~MpAGb2pDKh$JGEIp6Hq^s!8S&Sj|`#I*KE0wEO`pzznkorQ()O8g`V^RK&H< z={?4a0`aMb5F^ZWcv7l`Allg1ex|7X7xt%Z7%jG`>Oot8I#YFCO&xU9YhoD(k_xkDnm}<6pZSzen(-4XhPv$V_ zk9H_9Nf&z!obpdv2rCb|T`?|NpvjV4(keX0{l*d=GjcA9uy}-EX)BI1CWCqstW;)j z;nF3or!QR!ijKo@U|!(t>vYg!wD{sfF#@2^WMvvl#Qm*#-eZ5k) zNLZLID#dK~&nE;?NPdTpW1_Y*c#3K)h?*vj96tFitp1Q7@E4iE66)2Pq9=$b&#%I< ze8*^v6+2Q+>olIGnrS6&11}#{BeDv#Cpi3=c99$uA766qmQiz$1R6T+53jS^P}9N&GEH@#h#TaaX6tyd(FcVkNh!^mwxL*1j^pjGg5?M&+_v^0R*) z4bKoppr;Y;$JORBge``U6G;KpgtTBnVS#R32!)ZesSZ{Y;N^fYs=)fa&@j#>1NkN@ zaN!gVI#c+bQMbbvo6Y-dMO-HT?mVDs$U|jw9^BYxYPr#Fd*8LD zIi4&+bzU&E=KN9h<=?w1r@a(8oeFpQCLtuHYxT0^2b@hQwPr-e3>97ZI|ItRJZQ#? zi!yJlVn$Di;N+j(9}upt7U9)szG>SkV>dnao}u}J%L}4K+WD%j+8;9Y;E;(5!H#P< zWpDfhN?v-^dN@8LDsbp7I!Qz@2c}V~IMRs!=O782RwVUeR?+6FkLnYz?=afXz;sLT z3dvbyFV$v5a+hp*SC@hl0}{g2>ZY5Z*w9Lhok8(XsSKq_1r}`f&3PC# zXMF}hjAOLmqCS08EFP~2ddpDK6=&hB;4NxF$27eNVRtEZe3b5=Iz>}(E*}oy8B%cf zyB881??H*AiNaJMWX2h9LwucPu3Fx%+?2pm+pChaQ#l%YrL*cpW25o0nM&NG+m$Eo zoN_h|uFsOXdNo{7#FXmr-JHKEl@gk8si9k`df8N$iIRl&C(aMmqxw)TzL~b9dxOR3 zo}41Gf_zX&Dhk?-`=<}}Omak-)Q6vCmHN^M4uKcvrI4EA{NYDx)%bbCdTL&?lWBZ7 z8#Qj0LVuaiI)gjjunmK*uIPM=81lkEE={mxFS~2x$-9 zXgLY;b|N|g68T<7=RIpoD0hxg|9Ta_B>tA1!bfl7jg1tK3q(H6I1tjtKY@NW=e02P zk@3{}iyl!Q%G>zvV)G4eW-TR|7EBb2w$pkRjMi1A`U$DAZW8Gt&^WDgsvjbT(aY25 zAG+=@G!**-ap-I{RlBsZusE!E+{QB!Z^&wnvj1CIx^Sh;8`!~u-wxy5aA6kNA02+J z%AELxFBO@iEurN@gE(T1l7}6gDgbU4ErpS8<|vtYGeaHG+gs(~VXt964j*oo#qIS? zY{#NdUY#5I^=4jxSqFdL%s|o;=wZb4Ll?yRgs{g_4q^kAw z7hCEgHQ_qmNh$;V74M~Y&|mRxxtE?5zgrf<>r)G_kB@_hk-6d0ha*1UboO)-P7-DA za##`iY~s#V#@d_<1n(0rkaJg)dywUZfx2Y|DGP(`e&hmWXpm!=gjepzAAcf3z8~1; zTQLDmaL(z)D7Tj+!!uPRY$kY{L|>XRs?ILuf-kF!vgOO9A}EAjN@76~NeK4y6fi=$ zFB?AU{~**JD+xC{t8&U&*o3? zdJ|Y3I}~V$$fZ*^LPj*WjP|XgE_8^FO%gD&0+#?7ti9l+*MuA zE^8fC)6SjpP?FLSKI);SaJz@B2f3a=;Bt8ah`mZtHP*pBfH{w{>=)cSLztM}Jg=le z=YrD}Q4Dy-YDcxKFO&+ei+1VH1YJ0H)|P8}N|c7H>mB_(UGq6h$L$Wst*Rw;35l&w z5Z&ab|7F@qT>~y@7kCxrnMgeDqXqESa1|{fEg}M%Ot)u>iYI2=gmZ;O^%Z5) zbWCv8)6{3{5_4ke#}^n&?!nOVmoZbFRu8m0GaUgXq01`%Bo@!*MD&D^^aQv$p$Mjx zKCmHDla9>;0#?nCUZUs!>h;J`!MM*`PVj8nf2SXh^kaQXKQ~%n-Lr6R{bA_G1eAov zf`a>)fKv}>)lu!2X-34UGjXP~1ykBElJ7DqPQA)hgMA`mmbK%15mjApdEBpG zc6-vgH}XhU#r6Bs{&~OK_pfpJ5GORP<|q$hbF>9ax~Z4vsufN2>Uc`^&NJ=3&ol}* zCL_&C75Py$cm8XdYxHlGT@R`5y4T~sQvMxf=cU$E-dz9wRG(cG_*9_XY2hrZmdjZY zm!i}o*Fs05g~XS|7R1B#mfEe=$>ZnqqUPJj>UcX=$H^TU=x2M*t7+nQAq|<2nn7RE zTgkl-YYe*DGD*tfcDdxz>dTq@U2)<01=pd6-iG?=i7RtdczwB`%IjnUkoFe;-CW~C zJZ>?&#wXLoVkzm|#cfnd@eN*7Q+@WCl)W$Lo0=57$&4-DL0}PeaL6Km_}6;6T8_o& zbgdKKh#mD38{KA%F+SUh9$Cik|DZ{*gg>i!=4MkV7Fx#t|DXpdUC_wX5>{eXg;z)0 zD44-f7>_=xX0FAICOd#|xY5v8uVg5f`5Gw_o?wo%Fyg`CxOR@Z)Vy7d&1l_F4bp)Y zZ3p^+*I$7+37F_hDgYil3te{eX78X`;Oq3b+ zA!|G$QYFXt$difA`|RcvCYohxUYv%_%TDs0i8MqLL;9p`QqckQBA(4h!_F~oZuB^@ z1=@VQ63;y``vzC|hElk}MCQrjCPu~%N7a{3g6?OkKf|idGdrml$Cqf&%zjC_sTDx! zMWB$)(2&hQE%1mfmavW#NI4g3BPB?WsN?a#j}V78?nY{NBXq3X!1YLrR$##7Bcy>C z4RN>K76s=Kl!(_Uk4pl*Dbd6)iClFI{yubE=Zb*!I|Hg471O;hx!5Py^U7xwh2-V+ zF_ek&%qAc?+ctSXCIKwMWia+N6BChFeXSLq6|RdqyYcNBmrC)2x{Nc07F-OQF{J%z zE!?tU6jTAPn>snMxLVID28N6B_{-e zgxB2&nd}55A_cJF-X#8{BL?L}NOhSK-B@J0L*2a-zK0-$56&45Hv#JEdjGdg8D&X! z;`dd~ilCKnm^xPAEmUP3Ko2PrJ4#U#FcZYyJrF*!j%PcB&pQ!38bSuE)yB)S6nu^=dVte4f0)T?hBtbP)6rUzq!e%^_8^_! zm%Q)5G!1Jma7j(aE+}lDxla+Ka&jEpkI13p++xoF(cw zEG?D(-MHh7)*yQcA(G=MPOr2lq`D4o^|{pTDvBy}MeW{O#eyM@Lp)Bz*L` z7w3~{O2EX@pdN?gk>R)`07-pm=?MREcpRb{LPbI9`T5i`gmFle#GEVcdI(-;vI7Y? z(J*~{?3L6ge>!cC5@usC3upHm5i<&(FfYlIQWQQp&4BS1xan1IItU@mMFNzGWpljg z6>P%C#0i{0Jsev+2kbeBQ*wz}p%Xt7a~%p>A)hhrS=DwPUig@&zJm#&`Wj!D3F z@TX1Fax=%#>OLJb)oekY`s|>=XXAsW8J&^<=XW?pIPNUsW_y-!#R}!|<^H~lONwK| zpK@U)I0fj&JXMcZ+x>7afkeV+6idcv0K&k)EJA>sR-Xhc03EjjcIU8PS-oy*CoyBH zJs<~|r|TgSEoR02L>VVcq3{!}U_d*fs9dH~@yMK3EKvr~EW|;LmjI}M=y){YSo8Uo zrO(}+xX1GI^&RI}0NH%TvEb96DO+yMC$vyhkoPMCPy`OB1O6`6j4AbH5US7a(6M$- zv3vM`9;jgL!0VdZM}^vHhkq-^+TER$=snn3;RG}8m|iAymfEFf;GN!KceoQrKJDUw zQJa1^N%hbX3l4L^90zuaVXHWv0JXd}L6nN_@OV_8@eu{EDKna$@ucHPgbc6Ls@35d zOs`Bxbq(T!l~0ePjB~i5sHV)dRT3i32?DVYcp$0|Ye4Iflnm6?Q+N9A``?;wLmT$x z$Y~cBTQu!Qk@=%##*9l5GxAy(ASC${cqK3s-IC_}`@If2^X14P7?qkzI1_tj``KO* zifMK~SSwhC0+Bg^9G(PFkqCK?*x-lGU@MF^(I*U?)|pl>;k6#P*wGRM(_$r`=hTmQ zJgzx7AX%&o{pc(|{EellS(?!5y2A-54VQN+Clkf2Z0}IE{DSFnn7oY37VygI24x zN21)|>}YCOzwQP}T@kjCN0zZ2+qZ$RQE*8-=-#u|=mvz&!fXKGUQ!%Z)KdGX8Gx6E z&%t#kSnCkf?l4I>!ca8kn!NG*<`{{ zGr8nkN}~S#N~*PV;$Zr&drEnP_~dur{o#7ryCjgpl+dZwXsrumwO2flud%&@vLm2c zeP>h}=R!$~8T1C#v|Ptu7>0m_O$+Hx4jvSLM9H^nXyJ0y9Mf5I#9?!Gx)z}=*vz#< z*6Oo3#uX-bdFt*h6(Sw9a*o5Mk(`$Fu)&@0ko6l|B0vk{A-ur{<$N_v9+5S}qUBd% zen*cxlVRBIh%DKF4r|q#nT+ThA1rXfS#`0%xFzsvT-KEy(TuTFmZg9(MtS^*E6B6D zkB1kcIg#o<%bCQZx6aipzG%Bb{2y%gVS!yw2QgvH8@bv^f$(ZcyNC?~^ON*ra#~*; zi00u2+4itz*QC~P8r%zo&f*g8_<>vhK)*x4G?DEPHt$?%bD!ER6MxP3N$I5I`iw8! zo{g6g&kAwCux8_uP?2q0u_>5nx)AU`f$gti+v6u%Q|x0(sm`4{n>rz7T}`bhjvkji zC+%F;AOdezkkSqyWKPG+Lvo-|T0MZT_o@4{paecBek=Kbyt!VP<)Wi3k*T9x^=A>P z?e@VdQwPE_8?1bg1Z0DFY_HgIxQ-{ofHv1R5xhA#2D5Ig4`?>~AgN?=G@?05Lr#wCI;6ywGcu>`7~U%=BRc7c z^&|MMJdE(~z4#;h-Tou#9a+DH*E#n~voYyLy`{sLS+=i|xd_5?PGAA161g!aQBrEA zKqDgPdXkj^=2i8`_0^Nu&+Ix8)XCX+jbi}>FE7J!+YwBPJ+{zwh64`*pMD0}-xGY@ zZ(Pd-m---FqB&D_!}xee>~%_7<(su(9sH2UW7#KscgNIq?1t#gj^&kmcenZo5Tt!Y z93~9-BQ+4p{aY(=o3FsNtw36UcN8;L7j-qGRllI={G)44)4IN^eZ;h^;cByb9CicN zDMX+*;RS(^8zJMB4^A}F zoCg<(z8tnLg*z%LZ9|Xt0T9kaNy{Tn*S)Fa8#w<-w993hgZO;l^&yNDjF_Z(SdY|y z7<^RN=*hm`6>%;T|&WFXy%) zmVu$4xGkfVKA6dAwR-`qD0E~vXEaEq~TQYx;{d~q4#wq}Fnx-^5Vey|e$Z%|m zv=c6kixasN-Koh$#>s8ub!25`?BT*otHOy!qZ<#5#%Z)e^Z5c#S7#_X^c)bSAq?Pb zCbQkjZ0l~`#rIo<7P#dX2~AK;JxEb3;ZXTOfh`?Ue(&{JvS@^+ajovW2xMtD;e$P)b<*2aU#PDJ@h#6@Qg-^!4E?XO2X2 zy%L`I&FD&PB6Tz>X@wFvF+W8LD=>N3j(6vvOzRpn@rFTg+=Z$Deq(fBPDBVmJ7sZk zbC=?BqufSVQH1({U+}2>%KH~)ZbLLgtdU)}rrPvjCdKJfhA62V*V3URd7faDo;cZT z%GsI{OW%s?eWWN-D1;LQamb;PQ#Q~j{jr7F?z*Wem5UXeX`sodAY{*Q+D`L;>?Bt@ zZ)JOC;cCwD%6&`}ADJx}Oih;iP29FD$>W|aiuears3y$&f!Zvy>NW}}DT+ci_C~fn zp3vTu^oZ?MK?R@F%_&!VWY#wh@hxWkq;40Bpk}>gxZ&l+|G~@a^m1ihtY^hdyw7bm z9#_@@>p4Z;g4S-(iY#%ftueqXac|?Qu$1rQ>{NFZ33<4&0*6f2kU--8*j$-Y6aH$R zGv#7tW>a#B{M(b+tcVBJiwZihxJHk{&jUkT{H9Pd;JS?l%8C^8Abuj|$YNN@x5}m& zRlLpBcS-SMW(YBluG!%h)WIbs{d)Bc@3GjIZK0XPyWNdth=0R=`Zy_3iPckT>qd&~ zo3;7G2MsoFC@^=}YgW8`{|=ZQ8lJjr+u5>eZ?xH|na!KqqQ*Zdt}&+uK@LDV zqm&_)-EM8*kCv6(iMb)3c$tvym(~3VPQn#?10~Y9+hEU0oMMxrX`kRu=uCe;a?UPK zD-&aerd(3Po)6LT{kVHo#*R}GHs-}8jzv;j6d6YH=Ezut(RQ9?{yl79b}Yc(Mx}$& zVcgy{abHBy!5f1@8D+_MH}$t%qMYN3)~`<4;M{v+QLBsFrk*X_R_rglQC_H!JkYy> zd7=~4Y^rqr3w-OlSHw2z(qvM5(?TZ2Tg&2owX}!TGKlNZh4Ewe)?21Z0*eTxG}k5T zEizcIyaeBDd0i#xceOG!Ag%vKDz}N`tZ!<0d(?K8Tn>+RKQ-~*u4*neaj`Iu-?X7) zUdMaGiWiU%%;FC=il}_!?UEIRPQZ0^XA5|XWltfIIm{bd{k4#2cF6;RvYkTx5rjp^ zbkYu{9`o*U+BCEit%WRHT$oxbjgH5A>KrcA4n^bTn2YRIPmL%ZxL@2;6y3>%X;#y#3?LlN=^9PBv=A>N!g`3T*{^ z7%=sXLiC{@jZ%BW*(+%-P?R~N!)3UTt933<%gPCpQ|24*`UD<-KK6bt;-5f2=jZ9M z=P4<3vzFGpEK52zMLfLYzHpH`OXm~{Oe7F6{Jy$H5@*eSHMi|IEb45T%q}bwGZGC{ z$;$*n3v$=i1&uQsRWj`pNpj08T2`)>q%SVQ+y|g(8@c9%5Y(gOsp2DvXpl3)CVxH* zJB?q(Au4z^%(;8#nP-c5;K=aQJbueralRDFVo~^|FuAZ4LRl3`bUw~BiJ$qbxJVvL*1~=fB|jx9LYNY>Q20xml?J(%O`mQqC4tmY7m5xYRXvMtfV3*EAl~ zEs4W;!>wg}AfE1?{Gl%6)$Nk1G|Tm zxi*MhAq*B@OYH)~uMYSU$_f{*vqHtVuA%GBPd6%IKd;1fG!xg+x9YfUeVXZFX=MDU*cqIcdnyoS z??~UO4Gpi0v0F)4M|n=bb}+)6sNxH6>BCjLZi{RV=#0ejQ&O&sQNtPV1n05De5~ZD z#I@r=PHN{5CrPJJPxRIqcbm&D=ml8EhQsmXG5}IyM`kW^gu4K|KCq*{*804h9AXZB zV7kRYxJjOXHJ7hNv|)W=a0XZI*jv=zl-4vz=4z^|w#p)fQOVxXZ9UP7BdBq2(s^?p zjXY{n&cy(?S}6Uv1@W`Bu>1h`L@dUvu+;EPU;PN1ER894#NAz77?jlhPLM4}KU54N zPa{rS+0awbRC@#{ieqpNRL3xxEzkgWcgK#vyNs{du8!S1nA*!a`F%Rx>-GeR8Qg0j z&TX(|ERTvt>+Agv;D5YM@l){2w-fMl?U~iESagOq1ge~v6dWKANc`N9M1XLss zrQO5t6_)VdypD=!9Z%OVR+8sPhgGr)IAtUxkKYB2U3gE*C0yCu6_7U zJc-wuwDBWU@lR4G>a;y{<^y2 z=?{)3(kJ`)7+*l%^~?>tHXfik`JoN)Nd&fBoC!W2W%Y8pI1QpWfE^6dz!iFOh$uLU z$u7pN#p|Lr63E{_i#GGDrG)M!!c*nEnKb31jsGa4YVmnOS*%UjEGH|t%z#$n*j3ra zwSa4Xj#^C=`d_ZTr*?6i-I|6N*qro2xw4EG{vf;mnG$3$n^9w>(kFTqHLP4LL*B5O zI8c1UsyQ^=2l2dciZ#Rfil+UMv#aCKb6e^5pC_}Tpb#7qZnnqh34mbS@?h} zV?$Z{_)4^fn)Wldqh3lo*t^UbZ?U;<)!OuWD+O`PNjyCRO zv;!U|9IT9rjPL*bs1FBd-ZqQF0TIDA2cWA;q>JHL`dD{kSLZ1e~z208TxA8!L2`=z3@aJ`0 zig@e43*dEiNGUc>Hus~U;k(hUWohkN#O;1QG*{QWcca``K5)RB{$7Db~Y$4#Tp{m8kdjaQIdiYAosXHXN`vYcIN zLOwWh!vdlwTfzFMU*FGiCEj1n?fq3P?>BtPbC@*+f1#0lTBz)5%Lj?q*r=`aa1~u8 zmRdfY_Vzlhw(xhXHt7c6jc!C)Bs1C8?O8V_=mvdpsT;x1b-q98xk>n9%4N)XyWo2_ z<`*YjTFXsDYumEcL2u^s%dMp^xV0g9yGCO^YAik08XIN7*>LN7>;I$89kd60J9;py ztq1M$Vh;ko@Z)`{2iA6aAle>26H<_=orm$xlQ({K70A}-OX;YOzs{WwV$>6>t7BUoI0agWXJ3vl#1q|U2A5W~ zQ6=yQ81!hY7jt5iJ0Ix>1!8PQHHOt68lja4zQ>S?nvQwb5nq7t8e$ z{dah>qAzl1#R#D@i&$;hjZYHUoTu1)64HjX+E$y@am$pz4oMIE(?;P>+XR2glA8&S zb=oMD&mp5U2{(3%z)SDrif?8_qgN9>ve2E)=99UTsBO8N&@NbMWSZ4%z@wv;L~V{s zqOAF@pC`-ZTWU0EsB;KddMVXC&R)7K=h;gai<0Gn2%&(N&{aQEQLS_Fh@hyLWeoe1 z#Q&EAG5*HgO)cOD&XepRThPN*`xM|t_bdJKfzIQvmSb`8|03wYTt%cndvW z-qI}UU~l=+-Xd4%TYePOkM?7ezd^={Z^9d>+OE8T==|&r8+${%!jIE$>0F5m?8eo5 zX0+@^Sr$HdVAPQ__Y>mFS>@VeZUovxWp$-J^GbSrYi;3gXYpIKFB)i61~a5M?lI3l{A3UQs=ddo6+U9o0JW@9sGwD9)ZAWYn!X+18;LB|DV?^@1%r zzVv-}l#-HXo{Et(jn6Xg-{_Iv7OE22WmLD=m{%~{QP|xNTzu0CU@$}^0RwwR;-9^J zDsf!YRh6_)DQ=R(pr@M=990+?w)-P{@QGAK0lFJ?QYnPl7ATOYy2|1B;baiIK!6{Z zHObvx9gi=M7=3y@^{y6NQe4>Wg*_oL&FA2S`B{S> zt}H3mS}ipKhW=>}XDj*Gs9bHX$LK5UnrfOZx%Ax{%2DXL-5kr=qjM zB@HYYN23+z$pDOkqmdei(zKnmRk&k@CZbLD0j(4KDjD{5R0-ef2qZJP8vf{RllFkIo zq_uXm+F!@Zf5yvSUuXu6aJ696uUWo?P0H@^U?(b&ca~vF088K*B}KI1vfiiekh1we z8fg2E(>M|Yg_dVG8Zi!qChiXobJ;(41Esnk9|=rMvmh)m;}AcHC!{RHu~Rml#72vC zX<1o*1^yQ;I~py42KT7S`(Iy}qb3ywMoF!__MOaE&l_#yHq_)}h9wMjvsk3M;sQ0< zy#jnK5t2i;4H7|952*y}Zm~v#kvv-2VX4zciVvk0 z3DO(2QMIzf6=(|MLKF{yL+o9uBvq$4`{K5eNW35ad0Fz4l_=ZBt2_69)dH7EL_7kQ zq4HpqT4fSbDw#eZ7AH`v0~k|6y_`!4$4|00zO2}jU{&&ZDSRv^HrnB+3wxw`u3ahF z1h7|!5VG?rtVld8^WalIM;uEs`h%Bho0QjEgr)$zawO@YwX!4cE!6K((6+yE@Co2u zIV+E$Bp1OR#Q3~O#8Runfx>oxB0cTVy0qT<95)!6im;cq=5Tmn*%=d?72evWLWfYw z<+LGbp zayokwFSp-{=C(fHa1_*{C#Y;QR_@1pm{Y!6JI4XIa*@-&f{x zCbZii>2DX#(hJAvH_AcAqz+>Hq7YJ!M#}T?`*G7W;znE*Etpdk>Yi*w#gw}=`K=?2 zp-NTB0~NNs19CdX(LBkGliW z_rx>o@NzG80@sP#$WTrDWwZJgY!l1IGb%H0MYKqdJUm8L{CLcZv}1q7VJ7pu z%4{P?2V5I{)YfVqx`KoE3spDKS zyV*K;I#;)ADMF0Odr1qWTf1Mj!^h#I>qO54@I9fqqf+*2?@5Zx0Lld#f+VQ9AbBk` zrfSON#fhE$O}psNs8hHZox0{g#}_JN0GpyydD?kl zdOd9om(18vMrqn=9aE|=N1?5nK!?#fr5INh+5%w(f%e+jw+kDxn+a z>RS4oc7v873kL3o8|~O>ts3T~RM)DpRdInPO|hV-u)(v=&!J8M!tNYe@741e42}kq zMpOWV@GQqHkC?4twSY6?{?LK?g@bD&dVy<3jOHj*t5nJ^{rwB^M$SENf!)t1e;(PHp8e#ow8_@`EMd9d+hrMcsjwq>fcg2WbslAv^Dcl)Cqc3B|tZvR7>GjQM$>J`x?`pG$>&za02S#ve1hlGn zjJ-UK$YvY1$-qX$#Q>J|GU>-e?=FJx1OhE4)^Iee;%{Su(`bDm=$)}0ITaFHmd;L0 zTULymEnVI!Ea7jt+5;c)%ZV$RU>g@3&@&-q#f4!CA{johj~Bw);ZHS)U8IG1$rO@c zuT?FKn~65dcY7m&qDmq@i$K`;U?uZP3~m9xb#mAXu7#%~+9FZrDkbj4We~560t%!a z(d?+w67Z0me|I{nV)Pk-59THtN=YYJs85LFhpu=yNx~eB^F8gB)ZiWHVY}(I{pYbs;d1;ejKGbRkh)8MyXCK5JSCH zdkHnUnq{8uR?W(k??wf4W(D%yK5^gUKPB%?{?%|!ocyarGTfwE%=0Bd&Zcpw@#~3j zyalD62x2f;CesuYD^)?G&b~))>SEqp&RML>G$)d$R@Y^FJ|}ep)8Okli*?&-2IN;0 z)j~a+&%Qo+^?dn%HF;<&p-rr<+pnamg@%A4ex4h`a-?}KimF7y+5`JwoAx`+*HYCyBx$N077 znoNX*E~NO-w?{o>P32D^IF7q5^&o}I`j~tTL8X}`$yqIrzb{9w$V5n!xCtqxjnqbF zkj_pcP>23K$!!QpXmT49bW`fKq!Lw@(B}0G+;b^njJGCfPUREhF8jh>G$GNyR&OuZ z+a!ID27A#k+poG{pTwuUS5lbszz>Rd{;(0&9_Bb-IcgX3icrJpe5;m|ym%uV%BdjZ zu~TVaS{|RWN}CI)8v*^9e{(7&5t=3Hys^1f3&Wq;+4Uq5k;?h^B-JoRzHvjb;auP< zqZ?;?ad8K$5%n3Q*r#iWcXRfcdO4}`cb&Nd;DP(RAJ(a9;*{I^@-=O-Ag-&Y?w1od zsIg$#Yt%NmXr26;Xbo@6*2%BQR-Ok)*U+NKDqElNGDxKsoZrj_5+lJ z5~ZKTd1Y}!<3RkZ0%VLyHV$`66t#dA}N%sB~*OJt( zwIbdpeUwURPCvHaSuk@JG8OR(&V$J%U_3ko9M#+X7x17A&Vh+Id0hqDXN)!wLHw2k z=B;9$CaLY<2Z3O3cJ0k{DLlRz}gCu2{c7g-Sb`#cbGZ$kpzA9)(+G{;lXa0y#x2$Mb@i|MqV!SP1Q6=-hRUWfFDY^btMYINJ zKo8f-DZQQx?sA>hq#Z|1BuLldPAF+RRP>|1CQWBal@+n6xUwo2)?IKt@c8r(#dW;; zI_{!x0Df9Drqu{j_VNDZ=#~INx$1WKjIz?ftaOp_nZ>%8Fm==CJUU7F^5j7K=Y58z zBB8KC1eaHBGLJwj_yTfSUa;%y5f0188n_|?zYC-i)v_Q&Ww(R%B}E}O9OcAFP(tcp z?6s08jV;z^Q1;+DA;#?QI}amhCLhmSsgl<4EFnDI48BT0w0#L5L+ye5Cm(brv++^4 zs0|L?{l$uv=oq!B*F@^=Docsfv8TgtaYS}D>=2SZ|IXEB$b(2bP5GraaI3HdoB#tH zO-o!-e8rk?Q;|7KN(?tH2`R@Y@|#ylXHW%(i>x0{693s-oO!pxbdd4^z)9IezOp^N?gJUz4eP*vkx9JZ)dpR3%I{ z#OFvT)^k4cox*c27wg5MP^mXkdE#eulpnN;a8VR*NKncO`MS;%nzAZ>O}c){MQdK# zxoXM_7HaCywN`VEO3|zoS-cs`;y@$TG&$SKgss9PjG+pF5YtT8!6pVIh;JDvX zbb{UrO0;|D8NN}_6<_HDzQPgM*dq-LN5{+2ZxOLwNEKU-m5mfQWaqcTOupslvPn1~nAcI$z?)x=*F zM3a)za$GKAb6$uTOM$((gDe9{SqLB$?xgfgf99E}lWK`L*CtgYQDrqjH}lgQYUyu4 zzt+Y<*5EbSmM9Fl;nGdq(xcTw1R#n{SYLV;^fC?&%W)Wr1YmUMUyTH_l*w8>%XJS@dV0DE9R!^bf&I!X|j{J3pgUL7=r^@_{P zJVwp*@rAHXr|vN+VdQCNm;5yci;T*Ja$l%cQeJa;NnEJWa63Bk;aXyY!-Wu9?Myez zVFTkyu0`Vbeam?tC9|F>ExS%QkH_7vo&}FKN`T=ru1I^(@-K*AbEAqf+ff^7Zy1Hk z`2p_qS(6p^B*ux8T+w{4YW5x&;+PRHRoyPMEN5lGfNq@F29J6wtm!Jp^hq1uifr!D zhhDgAUx{HD)T5{Y(+i<4p&?&(d(D(&u!R5`QAePeFXI;bVMAb%6D;@Dlj6HKT9x9^ z`3Cfg#Zt&2;Y;JFZQ(G!9}X;|NgDy~aJTQR#sYLYg8(zJ#x<8JTdCc1q0B4~d&~VmG8k!*a<~%kdZZ3`LcQUHhJ93r44)L7?j!JUH zcpx>V?g!Em9E8j3Rw}6vT2F?S6?wRhhF9icx$J@@ZplgJD4~Cz9hnpXh380=Dr4l_ zTu`qu!cbRHWiR%QFX_2rbG%zAW=W=kJivtAgDMJA(~7y8p+F|(M=bn%r3YQ4Ey_4C z^4Ea0^G8eifeCB10l3^T{aP*Ai&0f62m$aMw3A-=%0@1X%Anaul6qEd)Zp^l7q!M=&!#5i2AZ|WzfM)2vqn!UuRlE@?WI6%&Nw^)LjV2G;=P|`Ktk$ zW-*IuSU3yGcunO~rlvH@g~svUKkiAUd?=$iwvnc{rPViiAB6=ZMO3U*ko@&Z9@7#x zjK28IXwEiMzOoojXloP%;X(LX=snt&$Z<`oQrj*pT2>b1mxy*`EiADJyA$+BofPey z(l)ao(T63tV9C1x*EW9T@csQ2{w^2vkN6n+yFwIw9Db|DHcK7<1o~NxVYq-@&L-?x zJ{%pY-P`Oy4y#8u3*0s-+wOwb0DwY>UWP#TSIg-(`61u9XOi7ha0sxhXatbjvx&m& zIhr3SAj0u_@<;yj{xg3O(ic!ej&WO1P00yr6IGM1Fy4{^glN^9M+HBYj-@`{Avy2L zLSw#-`*GBY)Do@pf`;Et+ajd0dE<&LESGcV2{oKSk_qeiHKxDQEu6(LBgI4 z19Nl_pJ|_*#Q8~&ujv@(G56HcaEHe=+cCOF6g<)ba4XG~BJ;>C>U4O(HS00IFHKo* z&U|=Za(ZjY0sV<{$F2A3pX&~8{p9;wqAje-tPnT8 z=#=s3Q_@M0gz8n24eKNIxjw>?hOg)&?9UQc3B(#AD5YtRg2(x)stC2s1@Bl%@%}A5 z#7B^H&I4zne(6Y`-eo~9nOYSo!=-$hCUdhLEl{Dc2MvLrD?buZ{wmrc+WzHNkN1!O z7j11<{yoQgC^?}eJKpK=!ovW#{MEBPIQZq8o|Sa6XNS7LH2;o6U7?$Hs5_YEhvIg= zEay1ZZ;PgcKcl>@mik*8Ka&?Sbkm>h#qDf1n_m|MrdyV?xkMm}Tl&X!d0W!A6mJNE z-IR~Y+lDS6tOoy9;z598&I>7e@_XAxhH`SpxCY0h0MA8_LLNKhd^opjVDqI=fh1QO z7GkCD#~=~3YVz?SWEyAnm;KFYtEiDRMe?czi2wJS)cXEnZmIh^y-(ceKV-#+bApSs zjjr`5;Oq}%V^DgP6>!_RBY=ky0n<%_7gEKuM%t@TlYE?$9%IeNII~hDAh<1O9@Dr; zEsr-NFpc0jhOzb|`2lRQU#vA9*mYGGD1YEsLb#F0&SSj6j^YWRxi}Ne*T~h#i`QYW zrYX_h7Zb9n;iKE~zc-FcV>~h{MZ-r=*ij-lXfU5i-IlE|@6UyK|96M3-$&!@iPH;Q zVPNLmHu`Jx3oH_z3zvu$T4@%J zLx3#LIghwZDNAgWw@j6T*Qh{CZiAO6t`f(b_fdxhWgiq=R1$mYSUoyTnP!aWlTk2mVfiC`TWK&8FTK!o^}gop5VT-r zsvL#_SWC(0e56w1q7y^el8Tc8Fbl5aq%vgY1c(^l&eeSy7n!_fr?TVnausYhoIY0J z7y$)al?sP*l|e1WIfzSVqq_6B2%$1^g{%~G)fyeW=^;6dn9zA5AgOCjNUbQd4d)i7 z7BW)sw4N(0r1d1+D#&(kfO-F6oOyR-wfDS$LCA_f@T_(oxb1eB-%PG(Iym%$k>^_> zg;dy$XuIcHn(@#D7VRJH8&HFupb$fYM;h<}lZHC9G1z)jvG-yVVnFQ*l8AHOSWDpni@g z$bro}s_(f#(a{oxrybP^T>QkUmc0Q7mKC={iK|kg71WW&9q@+lM!4u-;8$esg#G69 zWds*fnG9(#=IojzQ<~sBRvGl=S)G&yLD*)sK7|(y$e#FlZ)l&hBcNGIJis6RLYfj) z1Z4ElvCE!BO&^&qe$O(btq@v(SNHzv^lGNE7N^r= zj6!^5;t(_tfnGlyU_rEG|D0t8I$JrLI*u67Rr1gg)9I0pL*7s8EDa$|>z3C+O6dpL z0bAj{&?m?u>!6`Y*%$y9I}tSIo}ufw8dnC-skQlW3y8bxU+yqkkAt8>2J#J?HF_Rv zQ#{F0=w9F=BGu-?aFHLODV#1k${7$hgwbwFFVJmqU>|(hANcIfls(cv z30dJW1}x`fjy{RfyBj_sTd|myBsmKzxQgoOpeQ3?2qF@ek4U0NKr&j}a@^~k3M7Ky zm<-6d9}H(B9hh1sw98mLA3k}NvQmNytCtrBdLh5hPw*HcQjh(uP5cw+XFc|}I>Cyh zUQfj)B4}=KQFiiN+90A)Z-r6m_!5=%**zz#^rjD87G1NfH=^<+HUfrc!92`eI8vfZ z#ENr&5yB*hRES@F(mp*W$nG{_h~hqaTqW)U7#9rvGEM`OeHO1CD!stVMXYBTq4*F+ z)ow_geHGFgB__ET3cdEez`UL@Gb2 z6jWBeit|(hxvHWjKAfhhtnINov0<2XS{Mvai)YXLKc;MfwSj6j(H*@tMl?N|rc&&t z1=&pijX&LZ+iCLAm1Qm_%Vi<6abhDAR-)HQQPBzdzDXLm!?CFcU*lPis*RaaIX^Ji zCirw3giKR3TuS8zy%;%mXfIwis>rT3gQ~lr{cfE#)#uh!4{PdnI;ZvcN2jSbTFEWZ z=jSCx*+D=F8(2P~2W=`c*>+^CV`?XlkLw&K=mVnT4`l zGEo^Zz$!=0Qnde(x<%QG@}*Pamg{w3cb_{BnZ(Fb-SJwyzl|qjK|N4JJKimuT2n4 zNKJGO`h3Zlw8sw#ZNrYf4Z{(H$(e9jAS2^RW}_t;n?@N8@@Fcq_cELqZ562Q!~J1+ z7C9hNYDAq<1T^uRU@i7UF98L@0!w3XF!2Fd3~$Tg56t){(7 z_A%&-{o3SDVkVQv{eg8rsn*A3hw&U zqg917wpM)zZbNaX9z|a}D~;;FYBYbx2PKIg-0}_Pu~OKx;TX_9nk{}p92nw&W=`t; z;m{qK)DOuFfdlSwphlccj{Q*LfLeZxxxa4N@BAl{a!{~!fhL`kRZ9|e56*igWQbhm zu?ThDWb{;o_GiZ0&JY@Xwr$xoPsyHIYbG5TC$5$2NU>NVcBbKqjmmXGdu)Ek%iY06 zaReQ319obr8YZdm%4(s>4^p9(g|5e8#RuJkz|j^h%kD;k$=IRXmM8WAr&YAkRNG7J zkQ>!+h$20_y2Xb*z#}KXN)C@tQK2&zKvga9yLUpZ8((;9>2yCi`gr(u9#^kt{+@BJ zg=Bxko$_DDF1y=v^H5v@l+{S}5Ld(|kpZA$(E@bP-~G$<>FV1*e*a@s0G-b!d9H-* z@#QG|Wmy&g-fYA~qij76mcIc_G`O#$gC~TJ@NPp-XixQ!^|@xsE8e9_O-1pMk;Aw z@2tPT^)%(uc~~-5QSsnOoW{99XDIT<-s_-H}A;duJ5liwn(8i-`6=Z|tplcz79 zo{_AXZ@;%*m+`mXpk`Nrtihn!{InPo!G*A<8Of}$TyO#aQaeSUBfZTCM_97a@r!JF z07Jy<%|=v~RNDke{l}2u1R?*}-$=)Tql`F}0Ts8)rd^h=!P(I$jJxaDq^%UL-M#BB`dsyb&t3b9F$I7W>$i{f(5 zOlX!=1BFySNum_wV$oPGO{$Ht-8W}Wz-H_OY~vqNC}^FJIZrIRNFSW~oDrYcdvh>OY~)|=F$RBm zCCqJxkpT$NZj6>i+*(m~tp@NT+qzD)P!iCQ$=%%rH0A1ur}`AgHO3OLU3*%6Bt5PfdrlHVu3Z3b?k-r9n5pD%WOLo#-U?@2p|~p` z&P6IH^0_L3Gxxpt0LrjvPBJf8RcrAjOO}2Mu#02e<4uZm5YfV*cm!uR=l*O=l-?z_ z7v*LZ6HA#jK#eB?h*SFwr%tkBeqJ`CYMM6BUy?+Kl*uw?qL2u(!E^H?p;kgNtGa`D zX^Ym}3v$i(w?M`OYkso5wH&B=Mt*PAT6R3T4rm6MoB7!_Sj)R`F-TMw<$i zYRL$J(QYkC7u3e`=94TijGY>Ve1_j$k#BvA-FgcOl!3LSz}k{TV#S%1bTOkQx1-$g zEB}|0yP_|P0!{~Es<2cY*yr0Tl=nffMsm9oSE#P9law{-EDuWG>8-0T`@<*6=7;Nw zQjI$*iuXzjC0ksxY?3LzsN#gx&xoD&%E;ylO{GE-@1Zi&n8`u*OqC%^euxyOkXfUW zhV*gxIa)$GV_Gufq-^U#nEGPyt&zqLR_swKI8-AASg=0y#ei;ni4V3Da1MHd(9ii) z_D3aJ!o$D-dydX^#3zQ#_tRr-(hWh4(yFyExPN?kP6PnV3n2LWpc>R(S20kNfa|xU82QzRPmAIo8Ozt=0)_1ozc}l3AZ#0)ismj z4jiZN$bZ~(wvzZ(HRl5iLaDX|c&R*w<8U`V~Yrm?z$j z=V4c&8xkCfZjt0JWZzrR7JjIN7AYQnJwa1|$QuG;?+c*7Ld;AA$vC z_y}1i<@fMmSTAcq`F@)wW@Pl%I{dunEJIK%@N3UQ;VY9JPmaB!AG-)_JAZu59*xES zad>i9z@iafyLd#IiiynX7HllG5!iK9nGvV);6{oL;t7wVd=&?5MX^uw&-GWuNn#3| zCAs0!i!-EppQX1nR`ELFB-O{ehuQ9QG!(gfQy*5alK{L73vzj9+2t5?9x6bzzZLZV?A*Mp3Q6& zCJn~jO>Gg4QXm~^5J>lQu8;gdOy@sJB@*aT6tUM+Df6W&A*&qXerpSWB7b7W7DZ_Dy=zdZC>G{M?zU(CFC#pk5qp~XbHSRPfgODoU5#fe>hXefhyt>YoxmMihZHThdwabx{tzFBsMBO2?% z?2es<(o)0g=rm10P+LNp)W7laQe4;qz7A4~I4i<}R=izsM;utH28-IegxV``)x?kE zs;QboCa^=M9Xn*AIyDRQvpqv){gR+Rw$3;c4X8F@#Ghl_MLa44+1`*0|0bhs$EnUvDd#8VE zR~OC9rheR0Od#elPNVDUa^~*Q4t#jWXCw7sZ5M~CSh$N#oi8K>u&Uq_maiR6(I>~2 zZ#+ftA=TAY^;(~v71u8Vav)9IqOB&6;UcZ5Yzs3TZ2}%F6+u?`%Ns{ zk2ZL=FaD$WGyRDT{O^DMkN+P3`ZHJpQE&3-ljOu~VFb`6!|!jujl*HeQ-Np`@>;eC zN?pw3CesjjoRh#Mz>!3XSa6D1J#oqTcT+wy)M?-gff7g((v$Sxzr}$jp!%If#2?;B zCO337tIa}t8V{wBMo0vx^I8%NWGV~tupo}U>zLg3KdL|fM;XVG{t)yB1G)Bj@qhcn z>Qx`GDF4Q0{HV~|cCQDBe8O};qfAV=-MEAa0~T1thqZjKW>8)PsJ0qCtov+J+=~xs zhF=fnKxCv8BXn1Z5T2CcD4x4-C|jd*Ws+(`%w}o!RDBuN#US7d+S4gK%dZh{DPBXy z(Tp1VZ?K;lj-tCK+EZxC1OW*GIgsX&o8;njH;!Z$$u1M0rUItBUs5HuqhgmiLGVhl z@7a_g(!sf^l%?9)CF)=V)WjqF*_@kMn*V5GLFG4?0bYZ-rgD{V z77Xn+8Or+?t@HPFdJ;`NQe^8^prn$T7&nY%SS1e&Psun*`I-|m-elgyyRsw@df-Tm zH;QZ8B$r(jwcS&H+QunZ_$e4yvp^;0nVEkOM>Ywm7U6)7uFp$AhQzQrcuW3tQW^?x zTDOyriqX~Fv_a`k;BxppbYH}C1BpE*qhKGRI?gqt8MT$6yr87!1B(qQk2YrHklnS{ zRk5R)=wJEciwx}ZAhpi(0^FRt{wNgMYzD&=^eIydLlP+UELk3wS(#ZKz+B85pgPZ} zHY8(>o&PEx!IQW#PS8O(K}p1AoS;XGppfGvvsx`1B~ft*iCaqL&kOsIZwiVV^{Zj} zI>Rw`LYY4)0|x!eArv(0^=ul z5`5-5T@8W4Hn(WlQjJP(^+pkzrZzfo z<<4qb4QZ>wEbLvgc`aYWh5?8<b#$j)8gn6X@p*jsXgx8&YFz_vXNrfF;x&Xyph~ z>ouCBcCVPd^cmTStJh@q3R1&8K)K;=Zt}V`?h3PZ1^%%n%WU4i^aMnUv@&^tLl_UH;#d4{WXj z#7h;H|A{m;uWQE{LD%%vRBQCpU~vfO)Z53h*xx$ zKLy`2`QP?xDWSlkStMMHz7qUk%CKo)-iX)I-HF#?&w_gnq;EaO%XUCb}fZs|UNI zb{mTn21m)`w9#TB5aeF7>+>UC2r+)4V1MEQFzX`Wp=l=q?vI%H z7b&up1%~Et>08oAA(DLyf#qsQ@CBFpYMOS?Y=M^A`@=?<_KwU(LGpyAdgK;hJlczT zZiIOREPC|}5PDiDTx6ZwkVQ_e1#4mk_*<>3W#~y+l!;SL?vyVLrXGvOLH0sCxA#+6qlH0WnWW7=W1z1Jw^&0@$ECa{XI@fwg6HgVV)_8 zLOCzaiv(x0!jtmp4UiKI@Z?g%Il(h~o>sujth|rKIL&HgCqhncoxgE=_9~RZVy&Jl zOS01nK1~bZP?kGcWsUL|kbhyu*}1Qx63?c@v`P7nh&L^###~=6YvPy-S~42rw#frb ztf?OoE6f7VY||l*4WvnUeerqd%$0=*oLcs;S6AHp~s( z1b5+`|9zT3(3PF)^1utKYDl{3OZ@qf)fMF}X4YZ%1g~$!>ypY9&M?M8MC@)JE2^RWT8h<~VL1Bt^i zRNgHXuVA^!ppah_Ct_ZOt5kSut^p$GDR);%A(n!nqhaO(^$Q{3`1OjRv7mi~DU2V}^rcN!8T4(SpX>p?+a5N;h)`y@)*jrhYP^JNRyQ_qqg zY6ihE@Frha^e6>n9C-<9`y_ZFKOo2hKn`I6YR)y)Y&);jHmQW~YIA~POX|m5I7n}& zGmYACKIs$a3NuP<(p#MJr*(C=>i8p1msJl?$1aMc|6;D&*!ZeusQ#5Tdm{s?oG)gh!>wEv7qXhFqMEtGEsPE~6+Bg`Nl2sfYRxB9$WADHxba6*JZO>+ zYCD!|p{f^)UpfWB73DF+^1xx{rbbK&_GySWm*dTsI8GxSOI$uzl8L|7M*UN?UZn`jBxqtGn zWx-g_dd3>xfXwXDwp~sB13A+jKwhtS8k;A{<2<$`8;u=zQn*ZGC|@%x!B!|oN9F|3 zgiVUw0!ZVA5FsIwfx}hmqz~z|9ALT(2ecM-LBv*u6lZlhr@(?Yf#aX3tKx!XoDhq= z`DGyyazcjrwA}Tp+$TrnD)AtsurU1SpN;AiVM=x^%7J@psmm7&x3DC!GX*fa_tV)d ze9THfIO0ks9G-kM`i^wY!O3t|5f*_F?}5oz2J^p33y_}-cPa3GDg5?E9xVpQBGfvr zG+L;$)dQ|UsF0-sdhenK4d}t%`IU2Df^n%Qq+Hopwo*gS@6lXV5F_{ijzg(e4Y>am)fA+yJcz7F|68big z?4Lm2a#Kz3uZ}n61p{I+HXx#4T?YEO`2RE4_KgJDQNHasKQ{aw39>F`^JaoCoQr%G&P zJVhVoteY$>w80@4$Vk+3zQ1Z%X@c%VHJ1x1j)Y=Ky%bC8Q7j3S87oU2MjR&iVL{eZ zZ8ekVw;BFLCm8*%W z3x!(JNq9(@18))`xN|`t0rMpGHIOus(ngd(67Kl+NqE3`807ub=yH;c#I$0drxtmP zTV@3-nnzeeQ8k41iUSP9axElG%#!zf*vIjIyp{lsa3?V}5tDn3D7>)KAE{qpZUk>2 z5mTqJC`}IOASu$n=rmkGk)6J7IaO7dYK4 z_b93S9Z~vyM$-xU0bgXiHm;H;h=1#G5az}~D}o;5(FU(hDiGkok@1EB}|XmJ@6k><;z zOk;^82I@hq*D{90+HoGH{!Ov>>DGRKpQ2oevZ z1;hm_L`nl$f!)AFt$q|{RPB6H&?WQ3cL^utiGu3QtUv-W9kbgDA`ws5Q+6e!3z>MoWx)Us1(ukr0mJ)5 znNx>s*z2n1w`apHB|-8fx2(CL!42U?ms~c6b-Qe`r_vb4Ihu22SkJl5seD?4;Bj?Q zdRKAWI`U*TSvcVX@%i|2@17ce(oguKztn2@WHdh*qJ@+r~%cu^?V4C^OP-RZmUe@kkP z5l@SAiHqIH&ejup-ECGcyAzmA`R!f)b`{Z-fei1Q?y!wtwi^M4>lv;p6W5X2a}L19 zucoJtxTsg4L?(vzBx$$I(#<$GQ4-_EoW&7D6Mmk|+RIbqIPM2<8%UyVU2z^EcU9N* z*659WNfYIQcy+lvDx6e5Tf8rKwdPrN@4 z3e{%rxa{W)&cH;-O6?~s^}Se6Qo_Lv=zl|juV>VlI-rH_%@Ta$6z{Qu$zi)AJAdei zIw<37mpG=KCL1NJbGKd9%aO0GAy;!$_vm=d(x`X&>T$VD$u#TpVP=!a7{mFDK6IxS zVwpxQ7a;rM`Mc(2l51J+6(~)0pf$u988AgmenkOdoa?tjQkuCBR|-3fD4&X@IspRza;hnulD zOhu{LFnD4M;4wH8Vs+H28Clnbi{-dyupMEpR8>`0YeN3enb5fb`xMWd6$REwaM-1r zJr>`~le{4gDI1>bfBj(trQu0ST>0g(XiIo$onD4w4A+^iN1^y-c3o7;Hj{G|IF!r$ zLb1g~2i{zHF?Zq(W=!>AR1#2%`^tWRJH{PntFhToIXzc4Lbkkyiw>n}1#dU4Vgbyi zCiadoj?|PxdU8;COb*aOFfhX=6)GOyIQbU})8(^pa0Mf1I0hjmHicq1>nNAPF@qCS z>D2Tu1Ju4>y}uA2cQRwe?ekmao8KRd0)mLXdk+1HRuT`xqtqS2ra!Bz0`Gv-~^ zByXp4DIWc&E{87VW&w~+2u!jVWiBHPGMv8bg`sOy7AsgP?xDm`B_t}JSm0H19g8>biFV8VtmI5-`EDNBnTi7=~uZPcyloQ0CB_5Lal4^&B@2}2M zOc`B%ZpOj<+OSX)JC1|&M*ynneA}C{{e}f*@U`6RQiq~_P)|~Rr%TK7)M4sgkddG5r5W*lUo zRe-|bgI5nvoFpj~xsr<9ap-^@ZlS7~Uk(4X^Jk;h#`fylfBk{9+?S`_@JTDcvbIw% z=M&r>4fN($t-nldDrlKv-LZi;niKeC_1SJth@Q4?nIy@o-Ha7obr0u$)R2z%aZBh7 zJnq%q(VI%K9T|y%3p(}URT~%}Q?+0-AY6$cN_J9uMGi(z#|~z|nU;IzVoG{!TAfv0 z)clTYByuzZ-U=J(FD;b=cKUpkWzUK`!A2sGg!KaX^zgATI?MT1hT+r{O{o6}x zRGyfv(PcRsdmAgR$-q<7rA(gj6m()vj#hWi1+vNh4I885Pm{F~Wbk&R#y908QK(S{ zUX9%Oi~vxQ$=Je`0+^|ChY+&#gCK(uQ8aL{mi>_;!LbPjgwjd^^NC}u>?s;~j9QZl z^RqPd6h+>;T~N>EbHQ$dtdGZAd39;rXhqEU!YGrf@P(d|WYY6QQ-B(-4j7NeN{+|x z;`kD-m73L-NY2_pa`sSXs2N)M1jU|2+vT1gT`h3r9toFi7u2@Xtq(P{;h`Ykm3JRydgf6Bx3;TRhl5lgBZ`$O3w&C@b(8$}wgX?K91M@PI4^&EkvpMA z$EM!90g@WnS9?C+xpA&Y&%ouD0v~6nW))Y?`XLSsE6G-Og{$4viXUG~obsODwi}My z4q1C0tpvVA(zwX@PfK# zYPfMa4Lni08%W?DKe^$bZhaMYrbv2tIiAuE5EQ)@F0$=SqbqgZV75ppl*f3qFima{ zbow^E4AeqWGi(2p6aTI;Zp?~(=nx;o{Dww}P2fdHeb)gbxhU2Hlpe&=o4Vw5Z_TnD zxI$ugV{b$yQ4Gg}_sUz{LzaImi^40L^%=Q2Xobkr$$dt=v8d@V*Z%t z!7sL3!>dw+lRdE-PQ5az%NBLG!qTc(!=%_e zh1KHnn+KO1^wtX|TJ3EGPL;L}q%=V0Rk;6fF39ez5TSE)sFS=1Z3Sy@1#beRGe*ouANuUD+9W#NK{rN71g68)ci;aL zoL=`kYFNTu+iG`H8{c2UipzJ2g)t;LN0-Y%6`yo(Vmm490oAZ*kQCNpssartIgPSW(DU?BF%j2X%WXK?f_1# zW3-^6Y`-65YB><*VNw)4s(~5L-#{8Rc2ml<-{!kX_Rd+$mYQWTs@LbWc}L=tO{5PL zHN`kkRK`Dneim;9iVW(d(rM6`jj9~tcv~GXUF@h`Cta`;UC^p*3qSNo8gt*r4iZ<8 zTLMOW_%W$c2i;{jy`01iExhN!#`$m!p?gmx+YfB}u}V9@2^iPd zz*!L-{Ta&<5-}(jI^AYyKlnT10rCW0oT2gk)$il;{&W9yB0^|j-la6!qtM97VG&&Q zG~}nx#EX_IBI1hDRFc>CahQYpI-&C=#kxyuVmt>fTN0`dMwmF>*lDH};K5)L$7aiT zQgNW2ms|TAE~3~Q8iW_UxiF`QS9-Eo%r588 zlm)UQDf?w!3;%Els$R-`SL7O zsSH{jT(Xx3K-r9{g@5CF6%ilaQOXRU@x{(&y>5iLr%}=7FMI<7m z%yJ|w+U1gD^S^GitXh10kzL?Psqj2uRdU2%l4wy?Wqzw<6#P%)M6QcTMzDaUN{f{a zx@y`~ON>}?FNc2)|&0N{@{;I>&2NQt`ir7p)Ht6sf9A>EYl zXz3-Ipr~1J6(keneedrcJxQupBFQv;;{^ViQq!L`6%FE_bfP5wqlEP37B5KANi(Zc znieYsnHX*6ZS{s2C7MuLL?JKC;b_7(LhWi~!^sep{jo`iK-YVdf8i4r0X z<$>!?vqlwIqgtq^g3Rh>wAAZb2zu=d97#ArCC)TCs=nwKriNh=4Gj@aw*lz37o0Im0V^%IKF>rNyp7Xmz^A5>mkUzN$3+TB zTD9nt_Gn_F!DgCb(!^ z*$C*)FFjdE;u+GibFN(vDMnXO-nF3i+#=*y=b93uIy0>wC3A#mTevPbC#yJQpSG)t z8`)v|WHVGKDKXv%6Via?V%LOW6eV1zZMCR^JZa;f->cK%I`J>NpfzQovv_@QJUhpikEcT%1oM`;hA`%joi`^Y z%2*Gd2dT@+SOS(wW=s(?6(!A?BFRs zoD9CV-Xpxsb|ZNYI`jNlOURIzI#Bo!Kzs?`+R6LqT8V~A4UPRlH>2w0fs2x04yOGj zvOXsJlPzZIv#96?xzpkd+oEOL|hIv7;r_bHN zEb)m%Rpa~~lmKulOhHOg?W$?|cFbY1tmBF(W~_r(Mf{J|)*>*z4~s-@x`ze}fm+gq zR+fWS2Ir7{q=gZ?RJ0V2ptRKxB?pU*SG%OAH~G9Qc6f1ehV#@}g@*j-HzeQZOvt&y zS9I6$G?<4537sRLmFLy8{1`2!eB=sPoO&r3JUhkf>=fTJ<{Qz?E#dZ~bRabq=tS0i z()gR4|F>hDp8pmH>vu1-tJb;$4xK=BPI3lmri7sENI~da*)1>N=5R_-;#jx{gV;En zb5qMBM=*$x3Ak_Qc*40k^e@9{V3CKI14{MC@hO?d_Z|T;Ty7YwHdM|3o5})ZeH>;hz%s?6PU$_0#xY~Qn!U^ElT*; zZ!BOSkdL5<8F83vEIkUZM@CHhya}f`WK9omwmn1T6rQr-7KA)kzlX9!-}B^3+AZOl zm6thNwq@1Kn5bI-G~;#KyLG!<)DyjS9)4wN`s5-$OMYiCo>G0g&cYOyT=+{_T})$e zH;Qd(d~cMQKyQhb#hMS%Eh~Ji<00a>T5lCYa59wRHVKzPsm4!Rk#gguf)afR1OS#l zRjFB&x=jZ)4JSn}xO=EPutuSp-Ab;5bC3VRsfAt0u&mhky<%y>G6S(l^wTfLMC4lp z;ocaY@~VjfR;5?eyo*`B%4>b_CQ~k3!f(+d^9?;*FlMSMj*h?9FP6I$R_tBYgi8P& zo=ll9l!HADW}H=qjLWTx-Pv3obuq*JsIT`gA9tONyQaC}sTbr2VwJiiv)I;~m_M+V z1hYH9DZrJIM{qNJtf9vI`V|fPj`T52mE=t5*pBt*!R)q;yB`??LDv?iZU{mJ7SF`S zgl7)0Kjx?Btbny+u7G`Yk}xTttLgM1*WK}WOEH3qS1yiLhC$oNprGnjk)4{FOVj=8 z6udn&+?rwTaYA2)-RTD0*21)MMpcZtRv6sV>tF6I+r1u7|bEb28; z6JBvj?xWpF=()HCH8$7THZM+QyaT^-?3 zp$f< zhqFZ!Km3{DCaiBsHtEPSlI$ns0SRJA+Z50NX;eEQ2vI}kI?qpA&ZHYry@77Sy$E#O z&!mgxy_oRl+>0~?`YBCJxj-RH?F@g#1TK9l8-P%lr6yg zv&nKPof0BB5e@i0H1W%POUaG#cbZU>4)Q7GLPo=n)ESJyilMk zUx8bLMd=D`;Rh0sRMHzdV@q;NT1SFg!YB5O_hLD(iZlj{vkP)k&n_=V{(^@fjX?Y~ zsGzAUn?$~tUi{_ikKKI~dB^Eo>pb*%i8@xD12j$N!{z=i;G0n|GXs=7l*usE2~wXF zfthrhB#+;V%b9fkJMw=$okVb<|8#AdyUShCKr-{ zrY)B?kZn@VdBUnW=hO(w|S zSr_Az4qxj@r?V9F(}ie{+HwAvXXyt{AKV{Y&g&UJU|p&Ws9oYmP?uT+MrgX)nKu4k z{`k`$zm1KHdf}2kSEU&K6s}Av6H51M$Dp!@Y+Cv8XX{Tt#`>#NV#>$dGt2x&8C0dl z*^WrzSj@>%UoP~r?{sG&cp$q zU=bhft~@W=cf~@wTNlwdKzErLS7t~qI9G1v3sG*9!5BXdt~=Y<9TIypvl;n5R&BpM z45#g5cOs?=qG3OBz3^TLzAt1&s?o?#p&nF)_Mn`_9*~peX@59v=~5~IN}u}e|GeMn zjNx9)7!gAj?YN8l=v=DVrwbTlCe`F)-upC1w*;lQbz<$XKZ$?)pV7YU)xUy0%Jqt$ zLA`1SW}VKWx>|@$jQU=IjD39O*P_*Tkly)c#A70idPiTMStt4_TPL&}>BQn?DBtmk zg=~q+oM%{YgcyF-Y=Nz(rVwU|kmtkT?5Z31k}Rd=NivSI;P8gews3a1NtdFXHB2mTSMDr7Ds;*V zASw<9kI@Vy?-&+~>t`e528It4X#n+lQB+=etmiXvOguAJiuxsR;U&#kz8H5Op9b!J znKn3I%p2kR;m~=~{xL&KZ%nCSE5WamEG2OqaA)|L)EN@MJj;v4Ja9J-QDW(;z-;nG zXMQ62B8-m}xn7Tztvgoat9Y!$-7cY=R4Eh3Au_%YEj_RVqf|O=c)Xo;ucm0E;Z0G9 z8-c_#RiBJkQ|c|gFi{oXcM_2Y*B9NNDJogdIn!=ML!0NrqK#4f(xn78XmVpC~<#^*^N=QHi&ypldLv_#e{f8R#KGgnvT zi?O|b$>Z+o$J$+v@|Yh` zzUChPNC3_bvEg6oau{{_T#1(lQ3)GVy)pr)ZkVnEGj!W*+jltpG#6yawd8@x+WWszB@y zoh9zo8ams@CH zYMfZ)HRW2sXT7XU<`TvF=R&(=UFTF5wUcX_WreP+-B~zt$K2(4g-Wr-Qb#^ zCyNpX$ov;7%@a4+IkWsZG!U=zr@!L%1 z3wg!dJSapB{Gc!}>-gA6^hIiHBG_ zt4A;3TkQqkvVFDcja121SYR!JYI!BjN9ia(CN`xkDrw7dVQqQh!Q`&$6=$O?6VUI0kMeEeFz z^Nr0VSV+z_;#4M*Mm{Y=rdFOqZK1u*ZE4g!U{M{UQmwnsY9dJM_?EJ$#;xNoZ14Q2 z&%u86&*4VMb0|m&^PCz%xjLc%MA1o{5&BkKGnti*z7^A0wp505^4spJvS#p=B}q4Z zUK%QWs!WPpThni|Bm*gaS5|!TLZZHh77*}aPMz9eaVG9E)~++Ij5A}JnRK@4py#E5Tncp`lAsBJHu>%9>Sc|F z->y+TC0e-6o2OJ+G81w&1I_+V0=FgVE;TP+8ON5F=$H!Ur<7lk^L|I6+m4B)c$i|J zxoA~q3{LXGdDY4bft^k7Tu1re5i_s7l>m9$jA`?Fs)SefD^4r}t?xa!Awm6-L%Z*p zt^5ob5%Z7EQ~vnjDqNfV2baJ!67_ZQ>W`#?&Ggv=m7MAYWglz1&#N$Dj!x z&!CO!fQevK-f|{@K?=LS92DKFp0%U-;Wki=3wO$YYz}9;89Ee4SQP zg8e}y;u7R^GrnnAmweOD?AAh2c7`Y)G*cy!lEig^nR-D@uGJGyWMTw+qB=@-#RZ$$ zW7w^&TK6O4xvjVg(>QY!Xp@LP<cvI;`bYN zLcLE7`SzGd$2t(1B2%&Z1&oY6DAeKIT^(FL6%%uUms5{Pp?k(=1tqGc6+CbxGHfn1 zf^DMxlVO0+jE~U_h*}f@9x03tPgB&upjkXedUFeWiz*Y;oGB3zW)GrQby{ zqpDHit_xAjXgu?6h7kb%L;Xc{NhSfFN^PU0IBi>?woz`HZ`(|BPs`gv8ii}fGcS!| zQUR0={fLBr#g-gj4x81zHcDf@6U3o%309_D=rR-ns3Dev2jX zasS!Tu#R52klBekFfvuQc)h9Pn)ZdRfu4`RJmz&gTstmdy-@ zG6UXDx0yK>yN$drZbY2svsMK8J~hM6CF!GFu%Jpp){uaP;DNp+-RgMKVZeP^OXBtp z(vp&+Z`pVtWz4k%@F&WFlmU_Nc|`toW$)pS@dnP$!W*h6X3pNg-`g9&9@Klr6?+$7 z*_O1J{-?c)KgFvWM4RxcdUoMe{C#@WGnI`4HS+c)T4#vp$h-K`W>(5RdbRiQH+UcC zR>k)bE6(hE@c4UsU)CqywEBqIGV2q6ig!@Xl=Nv<_;&<;2Y;X5F+8SK!r#S-ol)Ra zo*v|5Qjx0#tH?jIw}a)-dr3i%G#2)ajr0(Ok>tQ?qJ;BhJGU%wH)BP5Yq7T=2W*PK zf$x3UNYb-HI7<1D7k2Q%FE*c*dus7_OdtilmxybRmgE0SRmfkI(?Rl`nZWBf^iQH- zYg(1%nr>Yy>Qb7a1!S}Zl=zals9HN=e>%hiqdSB+L?NEaQG2{UP5wvx*Ufgn|1=?* coTtC(KmYfziMRa2zx+1*f0Akw@uQ0h0D%SX*Z=?k literal 0 HcmV?d00001 diff --git a/GDT/1.3.html b/GDT/1.3.html new file mode 100644 index 000000000..bc3d908f7 --- /dev/null +++ b/GDT/1.3.html @@ -0,0 +1,3497 @@ + + + + GDT netCDF conventions for climate data, version 1.3 + +

GDT netCDF conventions for climate data, version 1.3

+ +

Jonathan Gregory1, Bob Drach2 and Simon Tett1
+(1) Hadley Centre, UK Met Office; +(2) PCMDI, LLNL

+ +

14th March 1999

+ +

+

+1  Purposes

+ +

+This standard defines a set of conventions adopted in order to promote the +interchange and sharing of files created with the netCDF Application +Programmer Interface (API). The standard is based upon version 2.4 of +netCDF. Documentation of the netCDF API may be found in the ``NetCDF +Users' Guide'', Version 2.4, February 1996, available from +http://www.unidata.ucar.edu/packages/netcdf/ +or via anonymous ftp at +ftp.unidata.ucar.edu. +The standard is named ``GDT'' from the names of the authors. + +

+This standard is intended for use with climate data, +and was designed with data generated by GCMs particularly in mind. +We recognise that there are limits to what a standard can practically cover; +we restrict ourselves to issues which we believe to be of common and +frequent concern in the design of climate metadata. +Although this is specifically a netCDF standard, +we feel that most of the ideas are of wider application. +Our main purpose is to propose a clear, adequate and flexible definition +of the metadata needed for climate data. +The metadata objects could be contained in file formats other than netCDF. +Interconversion of the metadata between files of different formats +will be facilitated if they are based on similar ideas. + +

+This standard is mostly additional to the conventions sponsored by COARDS +(ftp://ftp.unidata.ucar.edu/pub/netcdf/Conventions/COARDS). +In addition, all Unidata recommendations are supported here +unless noted to the contrary. +Comments indicate the places where there are differences +between the standards. +Comments given in emphasised type +and CDL examples given in slanted typewriter type +are not part of the standard. +Note that examples typically show only the details relevant to the +point under discussion, +and hence may be incomplete with respect to the provisions +of the complete standard. + +

+Successful transmission of data +depends upon the receiver of the file having +software that will correctly interpret it. For this reason, a +strategy that is as conservative as possible in the use of attributes +and encoding techniques will best promote portability of data. + +

+This standard also refers to the udunits standard +supported by Unidata. +The udunits package is available via anonymous +ftp at ftp.unidata.ucar.edu. +See section 11 for details of how the package +is used by this convention to define units for physical quantities. + +

+Useful comments and suggestions from Karl Taylor, +John Sheldon, Jan Polcher, Bryant McAvaney, Harvey Davies, +John Caron, Steve Hankin and contributors to the netCDF news group +have influenced the development of this standard. +We have made some changes to gain greater compatibility with +the NCAR CSM netCDF standard. + +

+ +This document makes occasional use of some mathematical symbols, +for instance ``s'', which should appear as +a lower-case sigma. Under X, this may require you to add the following +to your X resources:
+Netscape*documentFonts.charset*adobe-fontspecific: iso-8859-1 + + + +

+

+2  Filename

+NetCDF files should have the file name extension .nc. + +

+

+3  Data types

+ + + +

+The netCDF data types +char, short, long, float, and double +are all acceptable. +All numeric types are signed. +The byte data type, which is functionally identical to char, +is not recommended because its signedness is ambiguous in netCDF. +The COARDS convention deprecates char, rather than byte. + +

+NetCDF does not support a character string type, +so these have to be represented as char arrays. +In this standard, we refer to them as type ``string''. +A string array must be implemented as a +two-dimensional character data variable, +serving as a vector of fixed-length strings, +the second dimension of its CDL declaration +(leading dimension in terms of Fortran) +being recorded as a dimension in the netCDF file. + +

+

+4  Attributes

+ + + +

+This standard describes many attributes +(some mandatory, others optional), +but a file may also contain non-standard attributes. +Such attributes do not represent a violation of this standard. +Application programs should ignore attributes that they do not recognise +or which are irrelevant for their purposes. +Conventional attribute names should be used wherever applicable. +Non-standard names should be as meaningful as possible. +Before introducing an attribute, consideration should be given to +whether the information would be better represented as a variable. +In general, if a proposed attribute +requires ancillary data to describe it, +is multidimensional, requires any of the defined netCDF +dimensions to index its values, +or requires a significant amount of storage, +a variable should be used instead. +When this standard defines string attributes which make take +various prescribed values, the possible values are given in lower case. +However, applications programs should not be sensitive to case +in these attributes. +Several string attributes are defined by this standard to +contain ``blank-separated lists''. +Consecutive words in such a list are separated by one or more +adjacent spaces. The list may begin and end with any number of spaces. +See Appendix A for a list of attributes described by this standard. + +

+

+5  Global attributes

+ + + +

+The Unidata-standard attribute Conventions +is recommended to reference this standard, containing the +string "GDT 1.3". +This standard is registered with Unidata under the name ``GDT'' +in the directory ftp://ftp.unidata.ucar.edu/pub/netcdf/Conventions +and is available from +http://www-pcmdi.llnl.gov/drach/GDT_convention.html and + +http://www.met-office.gov.uk/sec5/CR_div/GDT_convention.html. + +

+The float attribute appendices is recommended to record +the version number of the appendices to this standard used by the +application which generated the file (see section 12). +This information could perhaps be recorded in the Conventions +attribute, but having a separate attribute +for it will allow applications to extract the information +without having to parse a string. +The string attribute quantity_table should be used to +record the URL of the quantity table (see section 12). +If this attribute is a null string, it is assumed that Appendix D was used, +at the version specified by appendices. + +

+The string attribute comment may be used to record any +extra information about the file. +Additional attributes to describe the file may be included as required. +For example, GCM output might include an attribute +to name the model integration. + +

+Although not mandatory, +the Unidata-standard attribute history is recommended to +record the evolution of the data contained within a netCDF +file. Applications which process netCDF data can append their +information to the history attribute. +The global history attribute is assumed to apply to all +data variables; +individual data variables may have their own history attributes +supplying additional information (see section 12). + +

+Use of the string attributes institution and +production is recommended. +The attribute +institution specifies who produced or supplied the data. +We prefer this name to ``center'' or ``centre'' because the two +possible spellings could cause confusion. +The attribute +production indicates how the data was produced. +If it was model-generated, production should name the model +and its version, as specifically as could be useful. +If it is observational, production should characterise it +e.g. ßurface observation" or "radiosonde". +The global institution and production attributes are taken to +apply to all data variables which do not have their own such attributes +(see section 12). + +

+The calendar attribute (see section 23) may +be recorded as a global attribute. +The global calendar +attribute is interpreted as a default for all time axes. + +

+

+6  Variable names

+ + + +

+Variable names should begin with a letter +and be composed of letters, +digits, and underscores. +Case is significant in netCDF names, but it is recommended that +names should not be distinguished purely by case +i.e. if case is disregarded, no two names should be the same. +It is also recommended that variable names should be obviously +meaningful, if possible, as this renders the file more effectively +self-describing. +However, nothing in this convention relies on the use +of particular names for variables. + +

+

+7  Data variables

+ + + +

+The netCDF variables which contain the +physical data are referred to as ``data variables'', +also referred to as ``primary variables'' by Unidata. +Apart from the general naming rules for variables (above, +section 6), +the names of data variables are not standardised by these +conventions (since files may in general contain multiple +data variables of the same physical quantity). + +

+

+8  Coordinate variables

+ + + +

+A one-dimensional netCDF variable +associated with a dimension of one or more data variables +is called a ``coordinate variable''. +A coordinate variable whose dimension name is identical to +its own name is referred to as a ``main coordinate variable'' +in this standard, +when it is necessary to distinguish it +from other types of coordinate variable +(sections 17, 18, +19 and 20). +Apart from the general naming rules for variables (above, +section 6), +the names of coordinate variables are not standardised by these +conventions (since files may in general contain multiple +coordinate variables of the same orientation). +The values in a main coordinate variable must be +strictly monotonic +(all values are different and either increasing or decreasing) +because this assumption is frequently made by software. + +

+

+9  Axes and dimensionality of a data variable

+ + + +

+A data variable may have any number of dimensions, including zero, +and the dimensions must all have different names. +COARDS strongly recommends limiting the number to four, +but we wish to allow greater flexibility. +The dimensions of the variable define the axes of the quantity it contains. +Dimensions other than those of space and time may be included. +Several examples can be found in this document. +Components of vector or tensor quantities could be contained +in a single data variable by giving the variable a dimension over components. +While there exist advantages for manipulating such a variable in memory, +we see no strong advantage in introducing this complexity into the netCDF +description, and do not recommend it. +Under certain circumstances, one may need more than one +dimension in a particular quantity +(see section 28 concerning multiple time axes). +For instance, a data variable containing a two-dimensional probability +density function might correlate the temperature at two different +vertical levels, +and hence would have temperature on both axes. + +

+If any or all of the +dimensions of a data variable have the interpretations +of ``date or time'' (T), ``height or depth'' (Z), +``latitude'' (Y), or ``longitude'' (X) then those +dimensions should appear in the relative order T, then Z, +then Y, then X in the CDL definition corresponding to the file. +In terms of Fortran, this means X is the first dimension +of the array. Non-spatiotemporal dimensions +should be placed to the left +of the spatiotemporal dimensions +i.e. as trailing dimensions in terms of Fortran. + +

+The reason for this convention is that these kinds of axes may +have special meanings to particular applications. +For instance, an application might want to plot a longitude-latitude map, +or integrate vertically, or extract a timeseries. +In the COARDS standard, the indication given by the order of dimensions +and information in the attributes of the coordinate variables +have to be used together to identify the required axes. +For compatibility with COARDS, we uphold all these conventions, +but we also introduce a new attribute to make the identification +straightforward and unambiguous, as follows. + +

+If the last four dimensions +do not have the interpretations TZYX +(in CDL order, omitting from the left if there are fewer than four dimensions), +an axis attribute should be attached to the data variable. +In other cases it is optional, but recommended. +This attribute is a char array of size equal to the dimensionality of the +data variable, +having one element for each dimension (in CDL order), +indicating the interpretation of that dimension. +The permitted characters are T Z Y and X, +with the meanings given above, +and - as a placeholder for a dimension which has none of these meanings. +Each permitted letter may appear no more than once in the array. +If a data variable has more than one dimension which could be given a certain +interpretation, +the axis attribute will therefore clarify which should be chosen. +Note that if there are multiple time axes (section 28), +of which only one is not collapsed, +this single ``climatological time'' axis would normally be the +designated T-axis. +If the axis attribute is included, +the dimensions may be put in any order, +but this should be avoided if possible +because applications not able to use this attribute +may not process the data correctly. + +

+ +


Axes for an ordinary time-mean longitude-latitude-height +variable: + +

+  dimensions:
+    lat=18;
+    lon=36;
+    pressure=15;
+    con_time=1;
+  variables:
+    float xwind(con_time,pressure,lat,lon); // order T Z Y X
+      xwind:axis="TZYX";
+    float lon(lon);
+    float lat(lat);
+    float pressure(pressure);
+    float con_time(con_time);
+
+See sections 14, 15, 16, +23 +for details of longitude, latitude, vertical and time axes. + +

+The coordinates of points within the data variable are +the simple ordered tuples formed by associating values from the +coordinate variables (section 8). +If a particular axis does not have a coordinate variable, +the coordinate values are assumed to be equal to their indices along +the axis, numbering from 0. + +

+Dimensions may be of any size, including unity. +When a single value of some physical quantity applies to all the values +in a data variable, +the recommended means of attaching this information to the variable +is by use of a singleton dimension (a dimension of size unity) +with a one-element coordinate variable. +The advantage of this method is that all the attributes of a coordinate +variable (quantity, components, boundaries, etc.) can be used to describe +the single-valued quantity. +Singleton dimensions also result from contractions, described +in section 22. + +

+ +


Longitude-latitude field of temperature +on a pressure level: This would use a +singleton pressure dimension to record the level, thus: + +

+  dimensions:
+    lon=96;
+    lat=72;
+    pressure=1; // single-valued coordinate variable
+  variables:
+    float temperature(pressure,lat,lon); // axes in order Z Y X
+      temperature:axis="ZYX";
+    float pressure(pressure);
+      pressure:long_name="pressure";
+      pressure:units="kPa";
+  data:
+    pressure=50.0; // Pressure level of 50 kPa = 500 mbar
+
+The units and long_name attributes +are described in section 12. + +

+ +


Surface air temperature: Surface meteorological +measurements are made at +a certain defined height e.g. 1.5 m, which can be shown thus: + +

+  variables:
+    float temperature(height,lat,lon);
+      temperature:axis="ZYX";
+      temperature:long_name="atmospheric temperature";
+      temperature:units="K";
+    float height(height);
+      height:long_name="height above the surface";
+      height:units="m";
+  data:
+    height=1.5;
+
+The explicit height should not be given if the surface of measurement +is included in the name of the quantity e.g. screen height. + +

+

+10  Coordinate systems

+ + + +

+If the axis attribute indicates X- and Y-axes, +and these are in degrees of longitude and latitude respectively, +these axes constitute a longitude-latitude grid mapped onto the +Earth's surface, and areas of the XY-boxes may be calculated +on this assumption. + +

+A coordinate system for the Earth's surface which is rectilinear +but +based on a polar axis other than the normal geographical axis is referred +to as a ``rotated grid''. +To describe rotated grids, a two-element float attribute +north_pole is attached to the data variable, +specifying the (longitude,latitude) coordinates of the rotated north pole. +If the attribute is absent and relevant, +it is assumed to have the value (0.,90.) i.e. the geographical north pole. + +

+In some systems, +the axes covering the Earth's surface +do not define a rectilinear grid. +We do not wish necessarily to exclude non-rectilinear systems. +For the moment, this standard is undefined for these systems, +and we invite comments from potential users on the appropriate definition. +The COARDS standard excludes non-rectilinear systems. +In principle, any coordinate system can be handled, +albeit clumsily, by replacing the relevant two or more axes by +by a single axis which indexes the points, +and providing associated coordinate variables to specify the coordinates, +point by point (see section 18). + +

+

+11  Units

+ + + +

+The udunits package includes a +file udunits.dat, which lists collections of unit names. The names +given in the most recent version of this file +and their plural forms will be regarded as acceptable +unit names for this standard, with a few modifications which will be +listed in Appendix C to this standard. +COARDS lists some modifications within the standard, +but we would prefer to put in place a means to allow future +modifications to be made easily. +Users of this standard should not define their own units, +because this would make their files less portable; +requests for new units should be directed to Unidata. + +

+The udunits package also defines a means for linear transformation +of units by a scale factor and an offset. +This convention is allowed when it is natural to express a unit in +such a form +e.g. density of sea-water in +kg m-3 +in excess of 1000 +kg m-3, +which can be specified to udunits as "kg m-3 @ 1000". +COARDS does not permit the use of this facility. +This facility should not be used as a means of data compression, +for which an alternative is provided +(see section 32). + +

+

+12  Physical quantity of a variable

+ + + +

+These conventions standardise three string attributes for +specifying +the physical quantity of data and coordinate variables. + +

+The units attribute is formatted as per the +recommendations in the Unidata udunits package +(see section 11), +with extensions for time +(see section 25). +Case is significant in the units. +This attribute is mandatory unless the quantity is dimensionless +(a pure number), +in which case the units may be given as a pure number. +There are a few defined dimensionless units, +such as percent, but +there is no need for a wide variety of dimensionless units for quantities +like sea-ice concentration, cloud fraction, probability and so on; +this descriptive information is the long_name rather +than the units. +A scale factor and/or offset may be specified +quantity e.g. sea-ice concentration in tenths +may be given as units="0.1f". +A dimensionless quantity with no scaling or offset +may have units="1.0f" or units=ünity". + +

+The long_name is a standard Unidata attribute +containing a descriptive name, +which should not specify the units. +This attribute is optional. + +

+The quantity attribute identifies the quantity by +a description chosen from a defined list, +optionally with additional information enclosed in parentheses +() +if sufficient detail cannot be given by a standardised description. +The purpose of defining a list +is to allow users of data from different sources to decide which +quantities are comparable. +Case is not significant in the quantity. + +

+We refer to the list of +possible quantities as the ``quantity table''. +The quantity table defines the quantity and a permissible unit +for each quantity. +Any legal and physically equivalent unit would be an acceptable +units attribute. +There are two options for a choice of quantity table. +One option is to use Appendix D of this standard, +which will be made available on the web. +In this case, the global quantity_table attribute should be +set to a null string. +Each quantity in Appendix D will be labelled with the version of the +appendices at which it was introduced, +enabling an application to deduce the complete set of +quantities which was available to the application which generated the file. +The other option is to produce a list of all +the possible quantity names, +giving for each one an acceptable unit and the name +of an equivalent quantity in Appendix D. +This list should be made available on the web, +and its URL recorded in the global quantity_table attribute. + +

+The use of standardised quantities is optional. +The presence of the global quantity_table attribute implies that +this option is being followed. +The quantity name may optionally be recorded in the long_name +attribute if preferred (to avoid having identical long_name +and quantity attributes). +Therefore if an application which wishes to make use of quantities finds +that the quantity attribute is absent, +it should obtain the quantity from the long_name attribute. +In the remainder of this document, quantity attributes +do not appear in the examples, but the long_name attributes could +serve to supply the quantity information. + +

+ +


Quantity attributes: + +

+    float tempt(pressure,lat,lon);
+      tempt:long_name="potential temperature";
+      tempt:quantity="atmospheric potential temperature "
+        "(after timestep)";
+      tempt:units="K";
+
+"potential temperature" is the description which might be used +as the title of a plot. +ätmospheric potential temperature" is the standard +quantity, and äfter timestep" is additional information, +which a generic application can disregard. + +

+Whether two physical quantities are different or the same is often +not a question with a well-defined answer. +Certainly if they are the same, they must have the same unit, but +various quantities with the same unit may have to be distinguished +e.g. atmospheric potential temperature and soil temperature. +In practice, +the most specific description applicable should be used. +We intend to expand Appendix D on an ongoing basis +in response to requests by users of this standard, +since we cannot foresee all the possibilities, +and we will err on the side of expansion, rather than restriction, +when it is unclear whether a new quantity is needed. + +

+The subgrid attribute (see section 21) +can be regarded as a modifier of the quantity; +it applies only to data variables, not coordinate variables. +The quantity and subgrid attributes together define +the physical dimensions of the quantity +(through information given in Appendix B and the quantity table), +and the units must be consistent with this. +The long_name might repeat information which is standardised +by the subgrid attribute. +For instance, the long_name could be +"maximum temperature". +The subgrid attribute will define precisely in what sense +the temperature is maximised. + +

+A data variable may have a history attribute, +supplying information about the derivation of the quantity that is not +possible to include either as a standardised form or in parentheses in +the quantity and subgrid attributes. +This attribute should be used as a last resort. +The global history attribute, if present, also applies to all +data variables (section 5). +A data variable may also have institution and production +attributes, indicating how the data was originally obtained +(see section 5). +These attributes take precedence over the corresponding global attributes. +The attributes history, institution and production +must not be relied upon to distinguish between data variables in the file, +and generic applications may ignore them. + +

+ +


Optional quantity information: +These optional attributes might be used thus to describe a gridded +observational precipitation climatology. + +

+    float precipitation(lat,lon);
+      precipitation:history="gridded using Thiessen polygon weighting";
+      precipitation:institution="Climatic Research Unit, "
+        "University of East Anglia, UK";
+      precipitation:long_name="rate of precipitation";
+      precipitation:production="surface station observation";
+      precipitation:units="mm day-1";
+
+This would be appropriate if the data variable was in a file with data from +other institutions or methods of production. + +

+The optional modulo attribute of a variable, if present, +records a number which can be added or subtracted +without altering the validity or physical significance of +the quantity. +It should be given in the same units as the variable. +This is most likely to be useful for longitude coordinate axes +(section 14), +with a modulo of 360, +and climatological +axes of seasonal or diurnal phase (sections 25 +and 28). + +

+We note that the Unidata-standard FORTRAN_format attribute +may be useful for both coordinate and data variables. + +

+In addition, +other model-dependent attributes may included to define +the quantity of a variable. +The Hadley Centre model will give each data variable integer stash +and submodel attributes, for example, which are codes identifying +GCM diagnostic output quantities. + +

+Variables may contain quantities which are discrete rather than +continuous, +meaning they can take only certain prescribed values. +This is more likely for coordinate variables than data variables. +For instance, a data variable containing the results of +a Fourier or spherical harmonic analysis might have a dimension +for harmonic number. +Section 25 describes some time variables which are discrete. + +

+

+13  Topology of an axis

+ + + +

+An axis with ``circular topology'' is one +which can be legitimately +transformed by shifting all the points one place along the axis, +moving the last point to the beginning, any number of times. +The main coordinate variable of an axis with circular topology is distinguished +by the presence of an attribute topology="circular". +A longitude axis which circles the whole globe is an example. +The value linear or the absence of this attribute +indicates an axis with ``linear topology''. +The topology is indicated only by the main coordinate variable, +but since it is the property of the axis it applies to any component, +associated or boundary +coordinate variables as well. + +

+When a circular axis is rotated, the main coordinate values must be +altered in order to remain monotonic. +Therefore the main coordinate variable of a circular axis +requires a modulo (section 12). + +

+Note that the topology and modulo attributes convey +different information. +For instance, +a longitude coordinate variable limited to values in the eastern +hemisphere between the Greenwich meridian and the date-line (e.g. 0E, 25E, +120E, 130E, 180E) does not have circular topology. (This might be from a model +of a limited area of the world.) When +making a contour map of a field with such +a longitude axis, one can interpolate anywhere within +the eastern hemisphere to draw the +contours, but it is not legitimate to interpolate over the western hemisphere +and draw the rest of the world, which is simply missing. The +implication of circular topology would be that one could put any longitude +at all on the left-hand side of the map. However, +this coordinate variable does have a modulo (of 360, +as required in section 14), and the points +can be labelled +in any way which is equivalent under the modulo to the coordinates +in the file. +The coordinate values 0,25,120,130,180 are thus equivalent to +-360,-335,-240,-230,-180. + +

+

+14  Longitude dimension

+ + + +

+Coordinate variables representing +longitudes must always explicitly +include the units attribute; there is no default value. +The units attribute will be a +string formatted as per the recommendations in the Unidata udunits +package. +The recommended unit of longitude is degrees_east (eastward +positive). Also acceptable are degree_east, degree_E, and +degrees_E. The unit degrees_west (westward positive) is not +recommended because it implies a negative conversion factor from +degrees_east. + +

+Longitude axes should have the attribute modulo=360, +indicating that they may be interpreted modulo 360. +Thus, for example, +-180, 180, and 540 are all valid representations of the International Dateline +and 0 and 360 are both valid representations of the Prime Meridian. +COARDS assumes that longitudes may always be treated in this way. +Since we have introduced the modulo attribute, +we require that it should be specified to indicate this. +A global longitude axis should have the attribute topology="circular". +Note that the presence of a modulo attribute +does not mean that the axis necessarily has circular topology +(section 13); +a longitude axis covering only part of the globe cannot have its points +rotated. +The sequence of numerical longitude values stored in the netCDF +file must be monotonic in a non-modulo sense +for a main coordinate variable of longitude. + +

+ +


Global longitude axis: + +

+    float lon(lon);
+      lon:long_name="longitude";
+      lon:modulo=360.0f;
+      lon:topology="circular";
+      lon:units="degrees_east";
+
+Note that if a quantity table is in use, +"longitude" is the name of the quantity in the standard list, +and could be recorded instead in a quantity attribute. + +

+

+15  Latitude dimension

+ + + +

+Coordinate variables representing +latitudes must always explicitly +include the units attribute; there is no default value. +The units attribute will be a +string formatted as per the recommendations in the Unidata udunits +package. +The recommended unit of latitude is degrees_north. Also acceptable +are degree_north, degree_N, and degrees_N. + +

+ +


Latitude axis: + +

+    float lat(lat);
+      lat:long_name="latitude";
+      lat:units="degrees_north";
+
+Note that if a quantity table is in use, +"latitude" is the name of the quantity in the standard list, +and could be recorded instead in a quantity attribute. + +

+

+16  Vertical (height or depth) dimension

+ + + +

+Whereas the two horizontal dimensions are usually longitude +and latitude, whose direction is well defined, +a variety of quantities may be used for the vertical axis, +if there is one. +The axis to be regarded as the vertical axis +must have both a long_name attribute (section 12) +and a positive attribute, +with one of the allowed values up or down, +to indicate the sense of the direction of positive +since this information may be useful for +applications which display the data. + +

+ +


Vertical pressure axis: + +

+  dimensions:
+    pressure=15;
+  variables:
+    float pressure(pressure);
+      pressure:long_name="pressure";
+      pressure:positive="down";
+      pressure:units="hPa";
+  data:
+    pressure=850, 700, 500, 300, 200, 150, 100, 50, 30, 20, 10;
+
+ +

+The COARDS standard requires the units of +the vertical axis to be selected from a defined list, +in order that this axis can be recognised by its units. +It gives special status to units of pressure, +for which the direction of positive is defined, +and makes the positive attribute mandatory +for vertical axes with other units. + +

+We have adopted a different approach for a number of reasons. +Firstly, to require units for the vertical axis +means defining dimensionless units for any +dimensionless quantity one might wish to use for the coordinate variable. +This is inconsistent with the treatment of a data variable; +the standard does not require that dimensionless units be invented +for dimensionless physical quantities in data variables. +Secondly, the vertical dimension of a data variable can be identified +from the axis attribute or the +the order of dimensions +(see section 9), +allowing any application which expects such a +dimension to find it without any further help. +Thirdly, the quantity attribute (if in use) is more informative +than the units. + +

+We are not entirely convinced that +the direction of positive should really be recorded as part of the +data structure. +It is mostly an issue for displaying the data, +and is to some extent a matter of personal preference. +If such special treatment is given to the vertical axis, +why is it not also recorded for other axes? +For instance, when latitude is shown on the horizontal axis of a plot, +is north on the left or the right? +This is the same kind of question, +but it strikes us as more a matter for a graphics application to consider. +Nonetheless, we have required the positive attribute for +compatibility with COARDS. + +

+For example, if an oceanographic netCDF file encodes the depth of the +surface as 0 and the depth of 1000 m as 1000 then the axis would +use attributes as follows: units="m", +long_name="depth below the surface", positive="down". +If, on the other hand, the depth of 1000 m were represented as +-1000, +we would have long_name="height above the surface", +positive=üp". + +

+

+17  Component variables

+ + + +

+A continuous physical variable may require +more than one number to specify it at each point. +We refer to these as ``components''. +The values of the components +are recorded in variables +referred to as ``component variables''. +The variable to which the components belong is called the ``head'' +variable of the components. +The names of the component variables are recorded as +a blank-separated list in a component string attribute +of the head variable. +The dimensions of a component variable must be identical with those +of its head variable. +OGDT restricted components to coordinate variables, +but the concept has here been generalised, +since any quantity that is used as a coordinate variable might also +be needed as a data variable. + +

+When a coordinate variable has components, +this standard requires that a main coordinate variable should nonetheless +be supplied which represents a combination of the components that can be +used to order the points on the axis. +As usual, this main coordinate variable must be monotonic, +but the components do not need to be monotonic. +The definition of the main coordinate in terms of its +components may be given in parentheses in the component attribute. +This information is not standardised and generic applications cannot be +expected to make use of it. + +

+ +


Hybrid vertical coordinate: +A vertical coordinate +h � p/p0 + s +is used in some atmospheric GCMs. +Atmospheric model levels are specified in terms of +(p,s) +pairs, +where p is pressure, +p0 +is a constant and +s +is fraction +of surface pressure (which is variable). +The +h +value is a linear combination of the two, +which cannot be uniquely decomposed back into +(p,s). +We would record this coordinate variable thus: + +

+    float eta(eta); // main coordinate variable
+      eta:component="pressure sigma "
+        "(eta=pressure/p0+sigma; p0=100 kPa)";
+    float pressure(eta);
+    float sigma(eta);
+
+A generic application would treat the component and main coordinates +as independent information. +The extra knowledge required to relate them would reside in any +specific application which needed it. +Hybrid vertical coordinates are the only obvious application of +component variables, but the convention could be used for other +similar purposes if they arise. + +

+

+18  Associated variables

+ + + +

+An axis of a data variable, or two or more axes in combination, +may have alternative sets of coordinate values. +These alternative sets are recorded in variables +referred to as ``associated'' variables, +having their own units, long_name +and other appropriate attributes to describe them. +The names of the associated variables are recorded as a blank-separated list +in an associate string attribute either of the data variable +or of the main coordinate variable of the axis concerned. +If the association is with the data variable, it applies only for that +data variable, +but if it is with the main coordinate variable, +it applies for any data variable which uses that main coordinate variable. +Association with the main coordinate variable thus may be more convenient, +but is less flexible. +Association with the data variable is the only option when several axes are +involved, and when there is no main coordinate variable. +The examples illustrate these points. + +

+The associate +attribute may alternatively and equivalently be named coordinates. +This possibility is included for compatibility with the CSM standard. +However, in the present standard it is deprecated because of possible confusion +with the normal definition of ``coordinate variable'' (section +8) and because the use of associated variables is wider +than just coordinate variables in the usual sense. + +

+A variable may be associated with more than one data variable +or coordinate variable. +If an associated variable itself has an associate attribute, +variables named by this attribute are also regarded as being associated. + +

+An associated variable must have dimensions which are all dimensions +of any data variable with which it is associated; +the associated variable can be regarded as a function +of the indices along these axes. +The values of an associated variable do not have to be monotonic. + +

+A generic application is not required to make any +use of associated variables. +Associated variables are not indicated in the axis attribute of the +data variable (section 9). +However, to improve legibility of the CDL file, +it is recommended that when +variables named by the associate attribute +of a data variable have interpretations that would be +denoted by T Z Y or X in the axis attribute, +they are listed in that order, with any others appearing in front. + +

+ +


Vertical axis: Many associated variables +will be one-dimensional, +giving alternative sets of values for a single axis. +One example is a vertical axis where one wishes to store both the +physical coordinate and the ordinal model level number: + +

+  dimensions:
+    lat=90;
+    sigma=19;
+  variables:
+    float xwind(sigma,lat); // 2D data variable
+      xwind:axis="ZY";
+    float lat(lat);
+      lat:long_name="latitude";
+      lat:units="degrees_north";
+    float sigma(sigma); // physical height coordinate
+      sigma:associate="model_level";
+      sigma:long_name="sigma";
+      sigma:positive="down";
+    int model_level(sigma); // model level number at each height
+      model_level:long_name="model level number";
+      model_level:positive="up";
+
+As the association of model_level is with sigma, +any data variable with a sigma-axis +will have the association with model_level; +it is not a property of xwind particularly. + +

+ +


Trajectory: The +value of a quantity along a one-dimensional +trajectory. +In such a case, we might have a coordinate variable containing +time of travel and +associated coordinate variables +giving the latitude and longitude of each point: + +

+  dimensions:
+    day=10; // 10 sample times along a trajectory
+  variables:
+    float hice(day); // sea-ice thickness measured as the floe drifts
+      hice:associate="lat lon";
+      hice:axis="T";
+      hice:units="m";
+    float day(day); // time since the beginning of the journey
+      day:long_name="time";
+      day:units="day";
+    float lon(day); // longitude at each time
+      lon:long_name="longitude";
+      lon:units="degrees_east";
+    float lat(day); // latitude at each time
+      lat:long_name="latitude";
+      lat:units="degrees_north";
+
+The main coordinate variable (day) must be +monotonic, but the associated coordinates are not necessarily. +An important application +rather similar to this one +is described in section 19. +Note that lat and lon cannot be +indicated as the X and Y coordinates in the axis +attribute (section 9). +This is reasonable because, even though they have the interpretations +of latitude and longitude, they are not independent dimensions in +the sense normally expected by an application which might look for +such axes. + +

+Since lon and lat are associated with the data variable +hice, other variables with a day-axis will not share these +associations. +If they are wanted, they would have to be indicated on those data +variables as well. +This approach allows the possibility that the same day +variable might occur in combination with various different sets of +associated coordinate variables. +For instance, there might be more than one trajectory, with the +same day coordinates but different longitude-latitude positions. + +

+ +


Transformed coordinates: Associated variables +of more than one dimension can be used to +describe alternative coordinate systems. +For example, vertical profiles of atmospheric humidity might +be available on a regular longitude-latitude grid, +but we might also wish to give the national grid coordinates of +each point. +The national grid x- and y-coordinates are each functions of +both latitude and longitude; +the x-coordinate does not correspond specifically to longitude, +nor the y- to latitude. +The appropriate representation is therefore: + +

+  dimensions:
+    lon=10;
+    lat=20;
+    pressure=15;
+  variables:
+    float humidity(pressure,lat,lon);
+      humidity:associate="y x";
+    float pressure(pressure);
+      pressure:long_name="pressure";
+      pressure:positive="down";
+      pressure:units="kPa";
+    float lon(lon); // 1D main coordinate variable
+      lon:long_name="longitude";
+      lon:modulo=360.0f;
+      lon:units="degrees_east";
+    float lat(lat);
+      lat:long_name="latitude";
+      lat:units="degrees_north";
+    float x(lat,lon); // 2D associated coordinate variable
+      x:long_name="UK national grid eastings";
+    float y(lat,lon);
+      y:long_name="UK national grid northings";
+
+This tells us that humidity[*][10][5] is the vertical profile +of humidity at the point with latitude lat[10] and +longitude lon[5], +which is at national grid x-coordinate x[10][5] and y-coordinate +y[10][5]. +Because the associated variables are multidimensional, they do not +correspond one-to-one with the axes, +so the association must be with the data variable rather than the +main coordinate variables. + +

+ +


No main coordinate variables: +A related situation is when the 2D grid is staggered or transformed +geometrically in some way (other than rotation-see section 10), +so that it is not possible or easy to give +one-dimensional coordinate variables for the axes. +In this case, there would be no main coordinate variables, +and the default of plain indexes would apply. +The physical coordinates are functions of the 2D gridpoint indices, +and would be given in associated variables of the data variable +just as above: + +

+  dimensions:
+    x=90;
+    y=45;
+  variables:
+    float orog(y,x); // 2D variable on a horizontal grid
+      orog:associate="lat lon";
+      orog:axis="--";
+      orog:long_name="height of the surface above sea-level";
+      orog:units="m";
+    float lon(y,x); // 2D coordinate variable on the same grid
+      lon:long_name="longitude";
+      lon:modulo=360.0f;
+      lon:units="degrees_east";
+    float lat(y,x);
+      lat:long_name="latitude";
+      lat:units="degrees_north";
+
+The lat and lon variables are not indicated as +the X and Y coordinates in the axis variable +(section 9). +An application which referred to latitude and longitude coordinates +would not generally expect them to be two-dimensional. +If it can handle this situation, it should identify these axes by +their long_name and units. + +

+ +


3D associated coordinates: +These could be used if one wished to describe +a field with alternative 3D coordinate systems, +for instance on both a regular Cartesian grid, +and in cylindrical or spherical coordinates. +The values of the alternative coordinates would be given on the Cartesian +grid. In the spherical case, for instance + +

+    float temperature(z,y,x); // 3D variable on a Cartesian grid
+      temperature:associate="radius theta phi";
+    float radius(z,y,x);
+    float theta(z,y,x);
+    float phi(z,y,x);
+
+ +

+A particular technical application of a one-dimensional +associated coordinate is to deal with the limitation of netCDF to a +single unlimited dimension. +If several data variables have unlimited axes of different lengths or +physical significance, +they can all share a nominal unlimited dimension, +and each have associated variables specifying the meaning of the axis. + +

+ +


More than one unlimited axis: +Consider a file which contains data variables with unlimited +axes measuring elapsed time with different sampling frequency, +and hence of different lengths. + +

+  dimensions:
+    time_counter=UNLIMITED;
+  variables:
+    float sw(time_counter); // sampled every 3 hours
+      sw:associate="time_3h";
+      sw:axis="T";
+      sw:long_name="vertical component of "
+        "shortwave radiative flux density";
+      sw:units="W m-2";
+    float latent(time_counter); // sampled every 30 minutes
+      latent:associate="time_30min";
+      latent:axis="T";
+      latent:long_name="latent heat flux density";
+      latent:units="W m-2";
+    float time_3h(time_counter)";
+      time_3h:long_name="elapsed time";
+      time_3h:units="h";
+    float time_30min(time_counter);
+      time_30min:long_name="elapsed time";
+      time_30min:units="min";
+
+ +

+

+19  Bundles

+ + + +

+If several data arrays containing the same physical quantity +have one or more identical axes, +but are distinguished by the values of other singleton coordinate variables, +it may be convenient to store them in the same data variable. +The common axes of the separate arrays become axes of the combined variable. +One or more additional axes are introduced to ``bundle up'' the separate +arrays. +Such an axis does not correspond to a continuous physical coordinate. +It acts simply as an index of the bundled-up arrays. + +

+The singleton values of the separate arrays are recorded in +associated coordinate variables for the bundling dimension. +They should not be interpreted as continuous coordinates. + +

+ +


Timeseries: +The Hadley Centre GCM +can generate timeseries of the values of quantities +at individual points. +Typically, timeseries from many different points are produced of the same +quantity at the same sampling times. +It is natural to contain this information +in a data variable with two dimensions. +One dimension is the common time axis, specifying the sampling times, +which are the same for all the points sampled. +The other dimension is not a continuous physical coordinate; +it is simply being used to ``bundle up'' the timeseries, +the points being +irregularly scattered in a space of two or more dimensions. +Thus: + +

+  dimensions:
+    points=15; // measurement locations
+    times=20; // sampling times
+  variables:
+    float snowdepth(times,points);
+      snowdepth:associate="sitename lat lon";
+      snowdepth:axis="T-";
+    float lon(points); // longitude of sites
+      lon:long_name="longitude";
+      lon:modulo=360.0f;
+      lon:units="degrees_east";
+    float lat(points); // latitude of sites
+      lat:long_name="latitude";
+      lat:units="degrees_north";
+    char sitename(points,StringMaxLength); // string array of sitenames
+    double times(times); // times of measurement
+
+See section 23 concerning the time coordinates. +This same form could be used for observed timeseries from stations. +The bundling axis (points) is simply an index. +The long_name and units of the associated coordinates +identify their meanings. + +

+ +


Vertical profiles: +A similar application is that of vertical +profiles at sets of points; +for example, scattered vertical temperature profiles through the ocean, +or data from various radiosonde stations. + +

+  dimensions:
+    station=10; // measurement locations
+    pressure=11; // pressure levels
+  variables:
+    float humidity(pressure,station);
+      humidity:associate="lat lon";
+      humidity:axis="Z-";
+    int station(station); // station numbers
+    float lon(station); // longitude of stations
+    float lat(station); // latitude of stations
+    float pressure(pressure)
+
+ +

+ +


Several parcel trajectories: +Consider a set of Lagrangian parcel trajectories (for instance, ocean +drifters). +Various parameters are evaluated at fixed times from the start of each +trajectory. +The trajectories are identified by their place of origin, +and the position is a function of time and trajectory identity. +The position information is therefore stored in multidimensional +associated coordinate variables. + +

+  dimensions:
+    parcel=15; // number of trajectories
+    times=20;
+    max_len_parcel_name=64; // max length of trajectory name
+  variables:
+    float temperature(parcel,times);
+      temperature:associate="parcel_name lat lon";
+      temperature:axis="-T";
+    float salinity(parcel,times);
+      salinity:associate="parcel_name lat lon";
+      salinity:axis="-T";
+    float times(times);
+      times:units="days";
+    char parcel_name(parcel,max_len_parcel_name);
+    float lon(parcel,times);
+    float lat(parcel,times);
+
+The associations have to be made on the data variables in this case, +because parcel_name does not have a main coordinate variable, +while lon and lat are multidimensional. + +

+This section raises the question of how best to store a single timeseries, +or a single vertical profile. +Following the scheme of this section, +it could be contained in a two-dimensional data variable with the bundling +axis being of size unity. +The associated information such as latitude or longitude would then be +stored in singleton coordinate variables, +all associated with the same dimension. +Alternatively, these values could be recorded as +separate singleton dimensions (following section 9). +We have no recommendation for this. +Either scheme could be appropriate; +which is more natural perhaps depends on how the data was extracted from +the continuous axes. + +

+

+20  Boundary variables

+ + + +

+Along a dimension, +the values might relate to points (at the coordinate values) +or to contiguous or non-contiguous cells. +The boundaries of the cells should be defined +as well as the point coordinate values. +The convention is to define an additional two-dimensional +``boundary variable'' with +a right-hand dimension (leading dimension in Fortran terms) of size two. +The values for which this dimension has index 0 +(numbering from 0 i.e. in C notation) supply the boundaries +with the smaller main coordinate values, +and those with index 1 the large values, +where ``smaller'' and ``larger'' refer simply to numerical comparison, +not to a physical direction. +Supplying upper and lower boundaries separately allows for the possibility +that the cells might not be contiguous; they might even overlap. +If a lower boundary value is equal to the valid_min for the coordinate +variable (section 29), the cell has no lower boundary. +If an upper boundary value is equal to the valid_max, +the cell has no upper boundary. +The name of the boundary variable is recorded in a string attribute +bounds of the main coordinate variable. +We recommend that the it should be named +by the coordinate dimension with the prefix bounds_. +The boundary variable should not have a units attribute; +its units are the same as those of the main coordinate variable. + +

+ +


Boundaries for a one-dimensional latitude coordinate variable: + +

+    float lat(lat);
+      lat:bounds="bounds_lat";
+    float bounds_lat(lat,2);
+
+In C notation, +lat[0] gives the coordinate of the first point, +bounds_lat[0][0] its lower boundary, bounds_lat[0][1] +its upper boundary. +In Fortran notation, the declarations are lat(lat) +and bounds_lat(2,lat), +and the relevant elements are lat(1), bounds_lat(1,1), +bounds_lat(2,1). + +

+ +


Albedo as a function of wavelength and snow cover: +Characteristic values of albedo are given for various wavelength +bands, dependent also on snowdepth. + +

+  dimensions:
+    lambda=4; // number of shortwave frequency bands
+    snowdepth=10; // number of snowdepth categories
+  variables:
+    float albedo(lambda,snowdepth); // no units for albedo
+      albedo:axis="--";
+      albedo:long_name="surface albedo";
+    float lambda(lambda);
+      lambda:bounds="bounds_lambda";
+      lambda:long_name="wavelength";
+      lambda:units="nm";
+    float bounds_lambda(lambda,2);
+    float snowdepth(snowdepth);
+      snowdepth:bounds="bounds_snowdepth";
+      snowdepth:long_name="mass per unit area of lying snow";
+      snowdepth:units="kg m-2";
+      snowdepth:valid_max=1e9;
+    float bounds_snowdepth(snowdepth,2);
+  data:
+    lambda=250, 385, 570, 795;
+    bounds_lambda=175,320, 320,450, 450,690, 690,900;
+    snowdepth=0.05, 0.15, 0.35, 0.75, 1.25, 1.75, ..., 450.0, 1000.0;
+    bounds_snowdepth=0.0,0.1, 0.1,0.2, 0.2,0.5, 0.5,1.0,
+                     1.0,1.5, 1.5,2.0, ..., 400.0,500.0, 500.0,1e9;
+
+A first index of 0, for instance, gives albedo values for +the wavelength range 175-320 nm. +The deepest snowdepth class has no upper bound; any value above 500 +falls into this class. + +

+In some cases such as the previous example the boundary +coordinates are well defined, +but the gridpoint coordinates are arbitrary. +In such circumstances, this standard recommends the mid-point of the boundaries +be used as the gridpoint. +Two advantages of this choice are: firstly, comparison of a gridpoint with a +boundary will always decide which cell the point belongs in; +secondly, it will probably be an appropriate choice for plotting and calculations +involving gridpoints such as differentiation. +However, as shown by the last snowdepth cell of the above example, +which is unbounded upwards, the mid-point is not always a sensible choice. + +

+ +


Probability density function of precipitation amounts: + +

+  dimensions:
+    ppn=10;
+  variables:
+    float pdf(ppn,lat,lon);
+      pdf:axis="-YX";
+      pdf:long_name="probability density of "
+        "depth of water-equivalent precipitation";
+      pdf:units="mm-1";
+    float ppn(ppn);
+      ppn:units="mm";
+      ppn:long_name="depth of water-equivalent precipitation";
+      ppn:bounds="bounds_ppn";
+    float bounds_ppn(ppn,2);
+  data:
+    bounds_ppn=0.0,0.1, 0.1,0.2, 0.2,0.35, 0.35,0.5, 0.5,1.0, ...;
+
+pdf[3][10][12] gives the probability density of precipitation +amounts between 0.35 and 0.5 mm falling at the location +lat[10] lon[12]. + +

+Boundary variables +are recommended if the main coordinate values +are not evenly spaced, +or if the dimension has a size of unity. +If the coordinates are evenly spaced, +and boundaries are not specified, +generic applications may assume that the main coordinates +lie at the centres of their cells. +Boundary variables may be supplied for component and associated coordinate +variables as well as for main coordinate variables. +Their elements are ordered so as to correspond +to the corresponding main boundary variables. +Hence they will not necessarily be monotonic, +and index 0 and 1 (numbering from 0) of the dimension of size two +will not necessarily contain the smaller and larger values. + +

+ +


Boundary values for a hybrid vertical coordinate: +The atmospheric column is here divided into three cells in the vertical; +from the surface to +s = 0.7; +from there to 20 kPa, and finally to the top of the atmosphere, +using the hybrid vertical coordinate introduced in an example +in section 17. + +

+  dimensions:
+    eta=3;
+  variables:
+    float(eta);
+      eta:long_name="pressure-sigma hybrid";
+      eta:component="pressure sigma";
+      eta:bounds="bounds_eta";
+      eta:positive="down";
+    float bounds_eta(eta,2);
+    float pressure(eta);
+      pressure:units="kPa";
+      pressure:long_name="pressure";
+      pressure:bounds="bounds_pressure";
+    float bounds_pressure(eta,2);
+    float sigma(eta);
+      sigma:long_name="sigma";
+      sigma:bounds="bounds_sigma";
+    float bounds_sigma(eta,2);
+  data:
+    eta=0.75, 0.45, 0.05;
+    bounds_eta=0.7,1.0, 0.3,0.7, 0.0,0.3;
+    pressure=0.0, 10.0, 5.0;  // does not need to be monotonic
+    bounds_pressure=0.0,0.0, 20.0,0.0, 0.0,20.0; // note order
+    sigma=0.75, 0.35, 0.0;
+    bounds_sigma=0.7,1.0, 0.1,0.7, 0.0,0.1;
+
+bounds_pressure[1][0] exceeds bounds_pressure[1][1] +because they are ordered to +correspond to bounds_eta. + +

+Boundary variables may be given for associated multidimensional +coordinate variables (section 18). +Each dimension of the main variable requires an extra dimension of size 2 +in the boundary variable. +These extra dimensions are placed on the right +(left in Fortran terms) of the coordinate dimensions, +and in the same order as the coordinate dimensions. + +

+ +


Boundaries for a two-dimensional latitude coordinate variable: + +

+    float lat(y,x);
+      lat:bounds="bounds_lat";
+    float bounds_lat(y,x,2,2);
+
+so bounds[4][5][0][0] contains the latitude of +the lower left (smaller x and y) +corner of gridbox [4][5], bounds[4][5][0][1] the lower right corner, +bounds[4][5][1][0] upper left and bounds[4][5][1][1] upper right. +In Fortran, the indices of the boxes would be (1,1,6,5), (2,1,6,5), +(1,2,6,5), (2,2,6,5) respectively. + +

+

+21  Representation of subgrid variation

+ + + +

+Since a data variable usually +represents a physical quantity which varies +continuously along the axes, +in reality there will generally be variation of the quantity between +adjacent gridpoints. +The data variable can give only one value for each cell, +despite this subgrid variation. +For many purposes, this can be taken as a ``representative'' value, +and it is not necessary to define precisely how it relates to +the subgrid variation. + +

+To be explicit about how each data value reflects subgrid variation +along a particular axis, +use the subgrid attribute of the data variable. +The most important application of this attribute is to +contracted or collapsed axes, described in section 22. +This is a string attribute comprising a list of blank-separated words. +In this list, "name: method" +indicates that subgrid variation along the axis with the dimension whose +``name'' is given +is represented by the specified ``method''. +The method, which may be several words, should be one of the +permitted values detailed in Appendix B, which include +mean, maximum, minimum, mid-range, +standard deviation, variance, mode, +median, cell, point. +Case and punctuation are not significant in the method. +Like Appendix D, Appendix B will be expanded on request by +users of this standard. +Some methods imply a change of units of the data variable, +and this also is specified by Appendix B. +In the above list, this is true for variance. +The method point indicates that the data values apply exactly +at the coordinate values, and do not at all represent the variation between +adjacent gridpoints along the axis concerned. +The method cell indicates that each value +should be regarded as a property of the whole cell +along the axis concerned e.g. a sum or integral. +The method can be differently specified for the various dimensions. +It must be remembered that the method applies only to the axis indicated. +If a precipitation value in a longitude-latitude gridbox is given the +method maximum for these axes, for instance, it means that it is +the maximum within these spatial cells, and does not imply that +it is also the maximum in time. + +

+The absence of any specification means that generic applications may +regard the data values as representative in whatever way suits their needs. +For quantities calculated at gridpoints by numerical models, +this kind of vagueness is unavoidable. +If a model provides a longitude-latitude field of temperature at gridpoints, +an application used to draw a contour plot of the field will generally +assume that the temperatures apply at points, and will use some interpolation +scheme to compute values between them. +An application which calculates the mean of the field, however, will +probably assume that the temperatures are gridbox means, +and average them by weighting each with its area. +Both of these approaches are valid. +The finite-difference scheme by definition does not have any information +about subgrid variation, +and may itself treat the values in both ways; +it might calculate gradients between them, regarding them as points, +or enforce conservation properties, regarding them as means. +It would be unusual to regard the values as extrema, however, unless +this was explicitly indicated. + +

+Except when the data are point values, +it may also be unclear what the coordinates of the gridpoints should be. +For example, +what time coordinate should be assigned to a time mean value? +In such circumstances, if the boundaries of the cell are well defined, +this standard recommends that the gridpoints should be +defined as the mid-point between the boundaries +(see section 20). + +

+ +


Subgrid time variation in timeseries: +Consider 12-hourly timeseries of pressure, temperature +and precipitation from a number of stations, +where pressure is measured instantaneously, temperature extremes over +the preceding period are recorded by maximum and minimum thermometers, +and precipitation is accumulated in a rain gauge. +For a period of 48 hours from 6 a.m. on 19th April 1998, +the data is structured as follows: + +

+  dimensions:
+    instanttime=5; // 5 instantaneous measurements at 12-hour intervals
+    periodtime=4; // 4 intervening 12-hour periods
+    station=10;
+  variables:
+    float pressure(station,instanttime);
+      pressure:axis="-T";
+      pressure:long_name="pressure";
+      pressure:subgrid="instanttime: point";
+      pressure:units="kPa";
+    float maxtemp(station,periodtime);
+      maxtemp:axis="-T";
+      maxtemp:long_name="temperature";
+      maxtemp:subgrid="periodtime: maximum";
+      maxtemp:units="K";
+    float ppn(station,periodtime);
+      ppn:axis="-T";
+      ppn:long_name="depth of water-equivalent precipitation";
+      ppn:subgrid="periodtime: cell";
+      ppn:units="mm";
+    double instanttime(instanttime);
+      instanttime:long_name="time";
+      instanttime:units="h since 1998-19-4 6:0:0";
+    double periodtime(periodtime);
+      periodtime:bounds="bounds_periodtime";
+      periodtime:long_name="time";
+      periodtime:units="h since 1998-19-4 6:0:0";
+    double bounds_periodtime(periodtime,2);
+  data:
+    instanttime=0., 12., 24., 36., 48.;
+    periodtime=6., 18., 30., 42.;
+    bounds_periodtime=0.,12., 12.,24., 24.,36., 36.,48.;
+
+It is not appropriate give a subgrid method for the station axis, +since this is a bundling axis (section 19) and not +a continuous physical coordinate. +The instantaneous and period measurements have to have different time +axes both because of their different dimension, +and because they do not coincide. +If the pressure measurements were made at times half-way between the +others (noon and midnight), the time axes could be shared. +Since the precipitation is given as an amount, it is a sum over +the interval of time by definition. +It could instead have been expressed as a rate in mm h-1, for +instance, in which case its subgrid method would be mean +rather than cell. + +

+ +


Thickness (geopotential difference): The +``thickness'' is the difference +in geopotential height between two pressure surfaces in the +atmosphere. This quantity is by definition one which relates +to the whole extent of its cell in the vertical dimension. + +

+  variables:
+    float thickness(pressure,lat,lon);
+      thickness:long_name="thickness";
+      thickness:subgrid="pressure: cell";
+      thickness:units="m2 s-2";
+    float pressure(pressure);
+      pressure:bounds="bounds_pressure";
+      pressure:long_name="pressure";
+      pressure:units="hPa";
+    float bounds_pressure(pressure,2);
+
+Here, bounds_pressure[0][0] and bounds_pressure[0][1] +will be the upper and lower pressure bounds of the thickness field +thickness[0][*][*]. + +

+If more than one subgrid method is to be indicated, +they should be arranged in the order they were applied. +The left-most operation is assumed to have been applied first. +Suppose a quantity varies in both longitude and time +(dimensions lon and time) within each +gridbox. +Values which represent the time-average of the zonal maximum +are labelled subgrid="lon: maximum time: mean", +i.e. find the largest value at each instant of time over all longitudes, +then average these maxima over time; +values of the zonal maximum of time-averages are +labelled subgrid="time: mean lon: maximum". +If the methods could have been applied in any order without affecting +the outcome, they may be put in any order in the subgrid attribute. + +

+If a data value is representative of variation over a +combination of axes, +a single method should be prefixed by the names of all the dimensions involved, +whose order is immaterial. +Dimensions should be grouped in this way only if there is an essential difference +from treating them individually. +For instance, the subgrid standard deviation of +topographic height within a longitude-latitude gridbox +would have subgrid="lat: lon: standard deviation". +This is not the same as +subgrid="lon: standard deviation lat: standard deviation", +which would mean finding the standard deviation along each parallel of +latitude within the zonal extent of the gridbox, +and then the standard deviation of these values over latitude. + +

+To indicate more precisely how the subgrid method was +applied, +extra information may be included in parentheses () after the +identification of the method. +This information is not standardised and may be ignored by a generic +application. +A mean over latitude, for instance, may be area-weighted. +This could be indicated as "lat: mean (area-weighted)". + +

+The subgrid attribute cannot be used +to show how a value reflects variation over a coordinate which does +not have a dimension in the data variable. +This should be done in the long_name instead. +It is generally more informative and precise to introduce a singleton dimension +specifically for this purpose, however. +For example, we could describe a quantity in its long_name as +being simply a temporal variance, +but it would be more informative to record it as a subgrid +method, by giving the variable +a singleton time dimension, which could also +be used to should the range of times +it covers and the time-interval of the data from which the variance was +calculated. See also section 22. + +

+

+22  Contracted dimensions

+ + + +

+A contracted axis is one which is formed by aggregating +the values of an axis with a larger dimension into a smaller number of groups. +In the commonest case, the dimension is collapsed completely to +a singleton dimension (i.e. a size of unity, section 9), +where all data points share the entire collapsed axis. +The collapsed dimension +indicates the relationship of the data variable which is +being described to another variable of higher dimensionality. +The boundaries of the cells along a contracted axis +will be the outside boundaries of the groups of cells along the +uncontracted axis, +or the outside coordinates if boundaries were not given. +The main coordinate values of a contracted axis will be values representative +of the coordinate ranges spanned by the groups. +A collapsed dimension has a single representative main +coordinate value and boundary coordinate values supplying the +complete range of the uncollapsed axis. +These boundaries will be +the extreme boundary coordinate values of the uncollapsed axis, +or the extreme main coordinate values if boundaries were not supplied. +A very important application of collapsed axes is to indicate +climatological time. +This is discussed in section 28. + +

+The subgrid attribute +(section 21) of the data variable +with contracted axes +can be used to indicate how the data values +of the variable with uncontracted axes +were aggregated to reduce the dimensions. +The new subgrid information will be appended to the existing attribute, +if any, +indicating the name of the newly contracted dimension. +Any existing references to the uncontracted dimension in the +subgrid attribute should be modified to refer to the contracted +dimension, since the uncontracted dimension will no longer be a dimension +of the data variable. + +

+As explained in section 21, this attribute will indicate +that the data value is the mean, maximum, minimum, etc. +The allowed subgrid ``methods'' +are listed in Appendix B, which will be expanded as need arises. +As foreseen at the moment, the idea is limited to operations which give +a single value representative of each contracted group of values, +without reference to any external constants. +For example, the number which exceeds 20% of the values in the group, +or equivalently the 20th percentile, is a single number +representing the group, but the +procedure of finding it is not treated as a subgrid method +because it requires the constant 0.2 to define it. +Instead, the relationship of this new variable to the old should be shown +by changing its long_name to indicate that it is a percentile value, +and giving it a new singleton percentage axis with value 20, +or cumulative probability with value 0.2. +This kind of transformation is analogous to reducing a variable on three +spatial dimensions (say) to two by extracting its values on a specified +surface. +The contraction or collapse is a special case, because, +in general, the percentile axis need not have a size of unity; +it might be a new multi-valued axis (in cumulative probability) +replacing the old one (in some spatial dimension, for instance). +This is like regridding a vertical axis of height onto pressure. +Having said all this, however, we note that median is in fact a named +instance of this operation-extraction of the 50th percentile-but we +allow it on the grounds that it is a common method +for choosing a single representative value. + +

+Singleton axes are not necessarily the result of collapsing an axis. +In section 9, we recommend singleton axes as the means +of attaching characteristic single physical values to a data variable, +for instance the height or pressure of the surface on which a variable +is supplied. +If no subgrid method is specified, the application knows only +that the single value characterises the data in some way. +All information in the subgrid attribute is entirely optional. +For instance, a time-mean quantity should generally have a singleton time +dimension +to indicate the range of times to which it applies, but +it is not mandatory to indicate in the subgrid attribute that it is +a mean over time. + +

+On the coordinate variable of a contracted axis, +the optional old_interval attribute +specifies the typical spacing between +two adjacent coordinates of the uncontracted axis, where +``typical'' is not well defined. +The old_interval attribute should be given in the same units as the +coordinates. +Further information may be given by the +the optional old_spacing attribute, +which may have value uniform, indicating that the coordinates were +evenly spaced with the old_interval specified (if any) and the cells +contiguous, +or variable, if they were not evenly spaced but still contiguous, +or disjoint, which means there may have been gaps between them. +The coordinates of the uncontracted axis may be explicitly recorded +in separate variables; +if so, the main uncontracted coordinate variable +should be named by the attribute expand of the +main contracted coordinate variable. + +

+ +


Area-averaging a longitude-latitude field to one of +lower resolution: The original resolution was 1 degree, and the field +has been averaged into 10-degree boxes. + +

+  dimensions:
+    con_lat=18; // contracted dimension
+    con_lon=36;
+    lat=180; // original uncontracted dimension
+    lon=360;
+  variables:
+    float sst(con_lat,con_lon);
+      sst:axis="YX";
+      sst:long_name="sea surface temperature";
+      sst:subgrid="con_lat: mean con_lon: mean";
+      sst:units="degC";
+    float con_lat(con_lat); // contracted latitude axis
+      con_lat:bounds="bounds_con_lat";
+      con_lat:expand="lat";
+      con_lat:old_interval=1.0f; // original resolution in latitude
+      con_lat:long_name="latitude";
+      con_lat:units="degree_north";
+    float bounds_con_lat(con_lat,2);
+    float lat(lat); // original uncontracted latitude axis
+      lat:bounds="bounds_lat";
+    float bounds_lat(lat,2);
+  data:
+    con_lat=-85, -75, -65, ...;
+    bounds_con_lat=-90.,-80., -80.,-70., -70.,-60., ..., 80.,90.;
+    lat=-89.5, -88.5, -87.5, ...;
+    bounds_lat=-90,-89, -89,-88, -88,-87, ..., 89,90;
+
+Instead of an area-average, the contracted field might instead have +represented the subgrid spatial variation of SST. +In that case, subgrid="con_lat: con_lon: standard deviation". + +

+ +


Mean over time and longitude: Here, the time-mean +zonal-mean humidity is given as a function of latitude and height. +The means have been formed over the complete time and longitude +intervals of the original data, so these dimensions are collapsed. + +

+  dimensions:
+    con_lon=1; // collapsed longitude dimension
+    con_time=1; // collapsed time dimension
+    lon=72;
+    sigma=6;
+  variables:
+    float humidity(con_time,sigma,lat,con_lon);
+      humidity:axis="TZYX";
+      humidity:long_name="specific humidity";
+      humidity:subgrid="con_time: mean con_lon: mean";
+    double con_time(con_time);
+      con_time:bounds="bounds_con_time";
+      con_time:old_interval=0.125; // originally at intervals of 3 h
+      con_time:units="days as %Y%m%d.%f";
+    float bounds_con_time(con_time,2);
+    float con_lon(con_lon);
+      con_lon:bounds="bounds_con_lon";
+      con_lon:long_name="longitude";
+      con_lon:modulo=360f;
+      con_lon:topology="circular";
+      con_lon:units="degree_east";
+    float bounds_con_lon(con_lon,2);
+    float sigma(sigma);
+      sigma:bounds="bounds_sigma";
+      sigma:long_name="sigma";
+    float bounds_sigma(sigma,2);
+  data:
+    con_time=19960901.0;
+    bounds_con_time=19960301.0, 19970301.0;
+    con_lon=180;
+    bounds_con_lon=0, 360;
+    sigma=0.99, 0.96, 0.92, 0.8, 0.5, 0.1;
+    bounds_sigma=0.98,1.00, 0.94,0.98, 0.86,0.94,
+                 0.65,0.86, 0.30,0.65, 0.05,0.30;
+
+This is a mean over the complete +range of longitude from 1 March 1996 to 1 March 1997 +(see section 25 concerning the time coordinate). +The longitude axis indicates circular topology because +this was the case before it was collapsed; after collapse, the topology +is not really meaningful. +If the humidity was subsequently meaned over the depth of the atmosphere +as well, +subgrid would be suffixed with con_sigma: mean, +and con_sigma would have bounds 0.05 and 1.00. + +

+If the same axis is contracted repeatedly, +the methods may all be +recorded in the subgrid attribute of the data variable, +but only the most recent old_interval +and old_spacing will be shown on the contracted coordinate variable. +But if the axis before contraction is retained in the file (identified by +an expand attribute), and was itself the result of a contraction, +it can record the previous old_interval and old_spacing. + +

+Repeated operations of some methods can be regarded as equivalent to a +single operation. +For instance, meaning longitude cells of 1 degree width to 5 degrees, +and then from 5 to 45 degrees, gives the same result as meaning in one step +from 1 degree to 45 degrees (apart from complications with missing data). +Similarly, meaning a time axis from days into months, then into seasons, +and finally into years could be represented as a single operation of meaning +from days to years. +In that case, the subgrid, +old_interval and old_spacing attributes need not be modified +for successive operations. +The choice of whether to take this approach +is left to the application. + +

+

+23  Time variables and intervals

+ + + +

+A ``time variable'' is one which represents date and time, +which we will refer to hereafter just as ``time''. +An ``interval of time'' is the difference between two times. + +

+It would be possible to describe +time in terms of six components (year, month, day, hour, minute, second) +in a netCDF file, +using six component variables of various data types. +However, it is more efficient and for many purposes more convenient +to represent a time as a single number, +giving the elapsed interval since a certain reference time, +which may be either implicit or explicit. +We refer to conversion from the components of a time into a single +number as ``encoding'', and the reverse as ``decoding''. +Encoding and decoding are complicated because +year and month are units with lengths that depend on the date +and the calendar in use, +so special provisions are needed for time axes. + +

+A ``calendar'' defines the set of valid dates +(year-month-day combinations). +The standard calendar is the Gregorian +(the calendar of udunits), +but climate models do not always use this. +For instance, in the calendar of +the Hadley Centre GCM, all months have 30 days. +The elapsed interval in units of fixed length +(days, hours, minutes, seconds) +between two times will not necessarily be the same +in two different calendars, +because there may be different numbers of valid dates between them. +For example, +the interval between 1 February 1996 and 1 March 1996 is one month, +and equals +29 days in the standard calendar, but 30 days in the Hadley Centre +model calendar, +since 30 February is a valid date in the latter. +Therefore the encoding of a time into an elapsed interval will +depend on the calendar, +and it is necessary to know the calendar when converting. +This standard permits the use of the standard calendar +(below, section 26) +and of other calendars (section 27). +The calendar attribute, +described in the following sections, +indicates the calendar in use. +If a time coordinate variable has no calendar attribute, +the global calendar attribute (section 5), +if present, applies to it. + +

+This standard permits two different methods, +distinguished by their units, +of encoding a time into a number. +These methods, referred to as ``relative time'' +and ``absolute time'', +are described in the following sections +(24 and 25). +Relative time is a more familiar method, +but absolute time offers important advantages. + +

+Time variables may have +an attribute time_format, to specify +a format for printing the date and time, +according to the conventions of the +Unix (TM) date command. + +

+Time coordinate variables must always explicitly include +the units attribute; there is no default value. + +

+

+24  Relative time

+ + + +

+A time encoded as a relative time gives the elapsed interval since +a specified reference time; +units takes the form "time-unit since +reference-time", +as per the recommendations of +the Unidata udunits package (but see below concerning +the time-unit) +e.g. a unit of seconds since 1992-10-8 15:15:42.5 +indicates seconds since 8 October 1992 at 3 hours, 15 minutes and +42.5 seconds in the afternoon, in Universal Coordinated Time +(time zones can also be handled). +In order to decode the values on a relative time axis, +the application will in general need to know the calendar; +the encoded time values are meaningless without this knowledge. +Furthermore, a given date may result in different time values +when encoded in two different calendars +with the same units. +For instance, 1996-2-1 15:00:00 is 62.625 days since +1995-12-1 0:0:0 in the standard calendar, +and 60.625 days since 1995-12-1 0:0:0 in the 360-day calendar. + +

+The file udunits.dat defines +second, minute, hour and day as units of time. +Units of months and years are disallowed by Appendix C +of this standard, +because they are not well-defined; +since udunits defines a year as a ``tropical year'' +of 31556925.97 s (674.03 s less than 365 days) +and a month as exactly a twelfth of a year, +use of these units will probably not give the expected results. +For example, 1 month since 1995-4-1 0:0:0 +is treated by udunits as +30.4368 days since 1995-4-1 0:0:0, which is approximately +1995-5-1 10:29, +not 1995-5-1 0:0:0. +Also, 1 year since 1995-4-1 0:0:0 is about 1996-3-31 5:49, +not 1996-4-1 0:0:0. +The udunits unit common_year (exactly 365 days) +is permitted, but not recommended. + +

+ +


A relative time axis for instantaneous measurements of a +quantity: Measurements are made +at noon on 2nd-5th June 1996. + +

+  dimensions:
+    time=4;
+  variables:
+    double time(time);
+      time:long_name="time";
+      time:units="days since 1996-1-1 0:0:0";
+  data:
+    time=1.5, 2.5, 3.5, 4.5;
+
+ +

+ +


A relative time axis for monthly means: +Means are calculated for February, March +and April of 1990. + +

+  dimensions:
+    time=3;
+  variables:
+    double time(time);
+      time:bounds="bounds_time";
+      time:long_name="time";
+      time:units="days since 1990-1-1 0:0:0";
+    double bounds_time(time,2);
+  data:
+    time=45.0, 74.5, 105.0;
+    bounds_time=31.0,59.0, 59.0,90.0, 90.0,120.0;
+
+In this example, the main time coordinates are merely representative +values, being the mid-points of their respective months. +Decoded, they are 1990-2-15 0:0:0, 1990-3-16 12:0:0 +and 1990-4-16 0:0:0. + +

+

+25  Absolute time

+ + + +

+This method of encoding time refers to the separate components +of time, rather than to a single unit of fixed length. +It offers two advantages. +Firstly, the encoded times are meaningful and can be decoded +into components of time without knowledge of the +calendar, although to calculate intervals between them +this knowledge is still required. +Secondly, ``partial'' times can be encoded, +which omit the year, or the ``seasonal phase'' (time of year, +time within the seasonal cycle), or the ``diurnal phase'' +(time of day, time within the diurnal cycle). +By contrast, relative times can only be ``complete'' times, +which include information about all three of these. + +

+The units attribute of absolute time takes the form +"time-unit as time-string", +The possibilities with the recommended data types +and their meanings are as follows: + +

+ + + + + + + + + + + + + + +
Format Data type Interpretation
second as %S.%f float Diurnal phase
minute as %M.%f float Diurnal phase
hour as %H.%f float Diurnal phase
day as %Y%m%d.%f double Time
day as %Y%m%d int Year and seasonal phase
day as %m%d.%f double Seasonal phase and diurnal phase
day as %m%d int Seasonal phase
day as .%f float Diurnal phase
calendar_month as %Y%m.%f double Year and seasonal phase
calendar_month as %m.%f float Seasonal phase
calendar_year as %Y.%f double Year and seasonal phase
calendar_year as %Y int Year
calendar_year as .%f float Seasonal phase
+
+Standard abbreviations and plural forms of the unit names are acceptable, +as usual. +The time-units calendar_year and calendar_month +are units of time defined by this standard (Appendix C). + +

+The time-string +codes show how the year, month, day within month +and time within day are encoded into a single number, +after the fashion of the Unix (TM) date +and printf commands, thus: + +

+ + + + + + + + + +
Format letter Interpretation
%Y Year (including century)
%m Two-digit month (01=January)
%d Two-digit day within month
%H Hours since midnight
%M Minutes since midnight
%S Seconds since midnight
%f Floating-point fraction of the specified time-unit
. Position of decimal point
+
+Since an encoded time is an ordinary number, leading zeroes in the +integer part may be omitted. +It is not mandatory to use the suggested data types, which are recommended +on grounds of precision. +If an integer data type is used for an absolute time variable whose format +includes a fraction %f, +a fraction of zero is assumed. +If a floating-point data type is used for a format not including %f, +any fraction is ignored. + +

+In absolute time, +3 p.m. on 5th April 1998 is encoded with value 19980405.625 +and units="day as %Y%m%d.%f". +The advantage of this method of encoding a complete time is that it +can be done without knowledge of the calendar, +whereas if we encoded in relative time units of +days since 1900-1-1, the value would be 35888.625 in the standard +calendar, and 35374.625 in the 360-day calendar. +We also know, without reference to the calendar, that the value +19980605.625, with the same units, is a time exactly two calendar months +later, and 19970405.625 is exactly one calendar year earlier. +But to calculate these intervals in other time units-days, hours, etc.-we +still need to know the calendar. + +

+The only complete form of absolute time +is "day as %Y%m%d.%f". +Note in particular that the forms "calendar_month as %Y%m.%f" +and "calendar_year as %Y.%f" are partial times which imply +no information about the diurnal phase. +This is a very important point. +For instance, 1998.25 calendar_year as %Y.%f means no more than +``a quarter of the way through 1998 as regards the seasonal cycle''. +This meaning is the same in the standard and 360-day calendars. +Because this representation carries no information about diurnal phase, +it is not permitted to decode it to 1998-4-2 3:0:0 +(i.e. 91.25 days from the start of the year) +in the standard calendar or +1998-4-1 0:0:0 in the 360-day calendar. +Similarly, 199804.3 calendar_month as %Y%m.%f means +``30% of the way through April 1998 as regards the seasonal cycle''. +Examples below show the use of such partial times. + +

+Note also +that the only form of partial time which is composed of the +seasonal and diurnal phases is "day as %m%d.%f"; +there is no method of encoding the seasonal phase as a fraction +of a calendar year or month in combination with the diurnal phase. +Should this be required, the application could construct it as a +two-component time variable. +This exclusion seems reasonable because data which resolves both +the seasonal and diurnal cycles must belong to a known calendar +(it will exhibit a certain number of days in a year, for instance) +and so its seasonal cycle can be labelled by month and day. +The calendar-independent representations of the seasonal cycle, +shown in examples below, +are more useful when portions of the seasonal cycle have been averaged, +in which case the seasonal and diurnal cycles, if both present, +will be on separate axes. + +

+The forms of partial time which do not include .%f +in the time-string are discrete variables, rather than continuous. +The interval of time in days or years +covered by a time axis in one of these forms +is reckoned by including the both ends of the axis in the count, +or equivalently by adding one to the difference of the ends. +For instance, an axis with units="calendar_year as %Y" +that runs from 1930 to 1939 covers ten years, not nine, +since both ends are included. +This is a partial time indicating only the year, +not the seasonal phase. +Contrast this with an axis including the seasonal phase, +having units="calendar_year as %Y.%f", +and end points of 1930.0 and 1939.0. +This axis spans nine years, from the beginning of 1930 to the +beginning of 1939, and does not include 1939 itself. +Examples below illustrate this point further. + +

+ +


An absolute +time axis for instantaneous measurements of a quantity: +Measurements are made at noon on 2nd-5th June 1996. + +

+  dimensions:
+    time=4;
+  variables:
+    double time(time);
+      time:long_name="time";
+      time:units="days as %Y%m%d.%f";
+  data:
+    time=19960602.5, 19960603.5, 19960604.5, 19960605.5;
+
+ +

+ +


An absolute time axis for monthly means, encoded in days: + +

+  dimensions:
+    time=3;
+  variables:
+    double time(time);
+      time:bounds="bounds_time";
+      time:long_name="time";
+      time:units="days as %Y%m%d.%f";
+    double bounds_time(time,2);
+  data:
+    time=19900215.0, 19900316.5, 19900416.0;
+    bounds_time=19900201.0,19900301.0, 19900301.0,19900401.0,
+                19900401.0,19900501.0;
+
+As in the relative time version of this example, +the main time coordinates are the mid-points of their respective months. +Although they are encoded straightforwardly, +their values depend on the calendar. +If one was comparing means over these months from data sources which +used different calendars, that might be inconvenient, +and could be avoided as in the next example. + +

+ +


An absolute time axis for monthly means, encoded in months: + +

+  dimensions:
+    time=3;
+  variables:
+    double time(time);
+      time:bounds="bounds_time";
+      time:long_name="year and seasonal phase";
+      time:units="calendar_months as %Y%m.%f";
+    double bounds_time(time,2);
+  data:
+    time=199002.5, 199003.5, 199004.5;
+    bounds_time=199002.0,199003.0, 199003.0,199004.0, 199004.0,199005.0;
+
+This method shows directly that the main coordinates are half-way +through their months. + +

+ +


A partial time defining just the year: +An axis of this kind could be used to record +the number of occurrences of a particular kind of event: + +

+  dimensions:
+    year=3;
+  variables:
+    int year(year);
+      year:long_name="year";
+      year:units="calendar_year as %Y";
+    int count(year);
+  data:
+    year=1991,1992,1993,1994,1995;
+    count=0,2,1,0,1;
+
+As discussed above, this axis spans five years. +No boundaries are supplied, because each element of count +applies to just a single year. +Hence, both the upper and the lower year boundaries would be +equal to their years. + +

+ +


Year and seasonal phase defined in calendar years: +By contrast to the last example, if it was appropriate to indicate that +each count applied to the whole of the continuous period of time of its +respective year, this could be done thus: + +

+  variables:
+    double year(year);
+      year:bounds="bounds_year";
+      year:long_name="year and seasonal phase";
+      year:units="calendar_year as %Y.%f";
+    double bounds_year(year,2);
+    int count(year);
+  data:
+    year=1991.5, 1992.5, 1993.5, 1994.5, 1995.5;
+    bounds_year=1991.0,1992.0, 1992.0,1993.0, 1993.0,1994.0,
+                1994.0,1995.0, 1995.0,1996.0;
+    count=0,2,1,0,1;
+
+The use of floating-point years allows us conveniently to represent +exactly the beginning and ending of a year and a point half-way through. +In the standard calendar, of course, the interval from 1992.0 to 1993.0 +is longer in relative time than all the other years. +But for some purposes, it might be more useful to record that each interval +is a calendar year. +This could be especially helpful when comparing data from different +calendars. + +

+ +


Seasonal phase as a function of year: +Here we show the date within the year of a particular event, +such as the highest daily +maximum temperature, or the onset of the monsoon, as a partial time +within its year. + +

+  dimensions:
+    year=5;
+  variables:
+    int year(year);
+      year:long_name="year";
+      year:units="calendar_year as %Y";
+    int date(year);
+      date:long_name="seasonal phase";
+      date:units="day as %m%d";
+  data:
+    year=2011, 2013, 2027, 2028, 2051;
+    date=629, 627, 626, 703, 710;
+
+The event concerned occurred on 29th June 2011, 27th June 2013, +26th June 2027, 3rd July 2028 and 10th July 2051. +Clearly the date variable could have been encoded as a complete time, +perhaps in relative time units, +but this would have included redundant year information. + +

+A time variable which indicates seasonal phase but not year has a +modulo of one year. If it spans the entire seasonal cycle, it also has +circular topology. Similarly, a time variable indicating diurnal phase +but not seasonal phase has a modulo of one day, and has circular +topology if it spans the entire diurnal cycle. +These kinds of time coordinate are particular useful for +representing climatological time, in conjunction with other contracted +time axes. See section 28. + +

+ +


Average seasonal cycle expressed in months: +Data for solar radiation as 3-monthly averages. + +

+  dimensions:
+    time=4;
+    lat=72;
+    lon=96;
+  variables:
+    float sol(time,lat,lon);
+      sol:axis="TYX";
+      sol:long_name="vertical component of "
+        "solar radiative flux density";
+      sol:units="W m-2";
+    float time(time); 
+      time:bounds="bounds_time";
+      time:long_name="seasonal phase";
+      time:modulo=12.0f;
+      time:topology="circular";
+      time:units="calendar_month as %m.%f";
+    float bounds_time(time,2);
+  data:
+    time=10.5, 13.5, 16.5, 19.5;
+    bounds_time=9.0,12.0, 12.0,15.0, 15.0,18.0, 18.0,21.0;
+
+The first time point applies from the beginning of month 9 to the +beginning of month 12, i.e. September to November inclusive. +A representative main coordinate is given of half-way through October. +The second point runs to the beginning of month 15, which is equivalent +to 3 i.e. March under modulo 12, and thus covers December to February. +The use of the modulo allows the main coordinate to be specified as +monotonic, as is generally required. Because the axis is also circular, +it would be permissible to rotate the values in order to begin with a +different season. + +

+ +


Average seasonal cycle expressed in years: The time +coordinate above could equally well be given in calendar years, thus: + +

+  dimensions:
+    double time(time); 
+      time:bounds="bounds_time";
+      time:long_name="seasonal phase";
+      time:modulo=1.0;
+      time:topology="circular";
+      time:units="calendar_year as .%f";
+    float bounds_time(time,2);
+  data:
+    time=0.7917, 1.0417, 1.2917, 1.5417;
+    bounds_time=0.6667,0.9167, 0.9167,1.1667,
+                1.1667,1.4167, 1.4167,1.6667;
+
+Here, the periods have been constructed as exactly quarters of a +year, beginning two-thirds of the way through the year. +In the 360-day calendar, this is identical to the last example, of +periods of three months starting at the beginning of September, +but in the standard calendar it is slightly different, +since a quarter of a year is not exactly three calendar months. + +

+

+26  Gregorian calendar

+ + + +

+This standard recommends that Gregorian times be given in +units of +days as %Y%m%d.%f with data type double +(section 25), +unless compatibility is essential with applications that cannot process +absolute times. +In that case, Gregorian times may have units of time +formatted as per the recommendations of the Unidata udunits package, +which specify a unit and a reference time, i.e. a relative time +(section 24). +The recommended unit is days, with data type double. + +

+Intervals between two times in the standard +Gregorian calendar can be calculated by the Unidata udunits package. +Udunits implements the mixed Gregorian/Julian calendar system, +as followed in England, in which +dates prior to 1582-10-15 are assumed to use the Julian calendar. +Other software cannot be relied upon to handle the change of calendar +in the same way, so for robustness it is recommended that the reference +date be later than 1582. +If earlier dates must be used, it should be noted that +udunits treats 0 AD as identical to 1 AD. + +

+Data type double gives a precision of about +16 decimal digits, +which means that it can resolve tenths of a second +for years of up to O(1 million) in relative times. +The precision of absolute times is an order of magnitude worse, +since a year looks like 10 000 days, rather than 365. +The larger the year, the worse the absolute precision. +If very large years are needed and the precision is not sufficient, +the reference year will have to be modified to keep the interval +small enough. + +

+If there is no calendar attribute applying to +a time variable, +the values are assumed to be in the normal Gregorian calendar. +This can be made explicit by setting calendar to standard +or gregorian. + +

+

+27  Non-Gregorian calendars

+ + + +

+It is recommended that +times in other calendars should be encoded in units of +days as %Y%m%d.%f with data type double +(section 25). +Relative times are permitted, the recommended units being +days since 1-1-1 (midnight on 1 January of year 1), +with data type double. +Since the Unidata udunits package can process only +the standard calendar, +an extension will be required to process relative times +for other calendars. + +

+Apart from the Gregorian, +calendars recognised by this standard are +julian for the Julian calendar (in which all years divisible +by four are leap years), +noleap for a calendar with 365 days in every year, +and 360 when each month has 30 days in every year. +If any other calendar is used, a suitable description should appear +in the calendar attribute, +but generic applications cannot be expected to be able to encode and +decode relative times or calculate intervals in the calendar concerned. + +

+

+28  Multiple time axes and climatological time

+ + + +

+There is no bar on a data variable having more than one +dimension in +a particular quantity, so long as the dimensions have different names. +A particular use of this is to decompose time +into multiple partial time dimensions (section 25), +of which one or may be collapsed (section 22). +This gives a method of indicating disjoint intervals of time +belonging to corresponding parts of the seasonal or diurnal cycles. +When a variable has two or three time axes, +the first interval of time which they cover is assumed to begin +at the earliest boundary values of all the axes. +If there is an uncollapsed axis in combination with collapsed axes, +it is a ``climatological time'' axis. +There may be more than one-see below for an example. + +

+COARDS recommends use of year 0 to indicate climatological time. +We do not favour this convention. +Firstly, it does not provide any way of recording which years were used +to make the climatology. +Secondly, udunits treats year 0 and year 1 as identical +(which is reasonable because year 0 does not exist-there is +no year between 1 AD and 1 BC). + +

+ +


A mean of a corresponding months in a number of years: +A longitude-latitude precipitation field with time axes +to indicate the mean over the months of January in 1961 to 1990 inclusive: + +

+  dimensions:
+    con_year=1;
+    year=30;
+    month=1;
+  variables:
+    float precipitation(con_year,month,lat,lon);
+      precipitation:axis="-TYX";
+      precipitation:subgrid="month: mean con_year: mean";
+    int con_year(con_year);
+      con_year:bounds="bounds_con_year";
+      con_year:expand="year";
+      con_year:long_name="year";
+      con_year:old_interval=1;
+      con_year:units="calendar_year as %Y";
+    int bounds_con_year(con_year,2);
+    int year(year);
+    float month(month);
+      month:bounds="bounds_month";
+      month:long_name="seasonal phase";
+      month:units="calendar_month as %m.%f";
+    float bounds_month(month,2);
+  data:
+    con_year=1975;
+    bounds_con_year=1961, 1990;
+    year=1961, 1962, 1963, ..., 1990;
+    month=1.5;
+    bounds_month=1.0, 2.0;
+
+The representative year is not likely to be particularly useful in +this case; the important information is the boundaries, which indicate +the range of years used to form the climatological mean. +These years are also given explicitly, and optionally, for reference. +Since the con_year axis has a discrete form of time (%.f does +not appear in the time-string-see section 25), +we have to include both ends of the axis in working out how many years are +involved in the mean: +bounds_con_year tells us that years from 1961 to 1990 were used, +which comes to 30 years, counting both 1961 and 1990. + +

+Suppose we wish to include in the same data variable the mean over Decembers +from 1960 to 1989. +To do this, we give the month variable modulo=12.0f, +which previously was not needed, change the dimension to month=2 and +the data to + +

+    month=0.5, 1.5;
+    bounds_month=0.0,1.0, 1.0,2.0;
+
+According to the convention, the earliest time indicated by the combined axes +is the lower boundary of all of them, which is the start of the 0th month of +1961, equivalent to the start of the 12th month of 1960 under modulo 12. +Alternatively, we could have + +
+    bounds_con_year=1960,1989;
+    month=12.5, 13.5;
+    bounds_month=12.0,13.0, 13.0,14.0;
+
+This is exactly equivalent. The last January used ends at the start of the 14th +month of 1989, which is the start of February 1990, as before. + +

+There is no standardised way to indicate missing months from the mean, say +December 1974 in this case, +except to include con_year:old_spacing="disjoint". +The information could be included as a note in the subgrid attribute, thus +"con_year: mean (December 1974 missing)". + +

+ +


Climatological seasonal means for several decades: +This is an extension of the previous case, +and of the +example of an average seasonal cycle in section 25. +Here, the axes are set up +to indicate climatological means for two of the seasons +in three successive decades. + +

+  dimensions:
+    decade=3;
+    season=2;
+  variables:
+    float precipitation(decade,season,lat,lon);
+      precipitation:axis="-TYX";
+      precipitation:subgrid="season: mean decade: mean";
+    int decade(decade);
+      decade:bounds="bounds_decade";
+      decade:old_interval=1;
+      decade:units="calendar_year as %Y";
+    int bounds_decade(decade,2);
+    int season(season);
+      season:bounds="bounds_season";
+      season:calendar="standard";
+      season:modulo=1200;
+      season:units="day as %m%d";
+    int bounds_season(season,2);
+  data:
+    decade=1966, 1976, 1986;
+    bounds_decade=1961,1970, 1971,1980, 1981,1990;
+    season=115, 415;
+    bounds_season=1,228, 301,531;
+
+Here, precipitation[0][0][*][*] is the data for December-February +(i.e. 1 December to 28 February inclusive) +of the decade 1960-1970 (first December in 1960, last February in 1970), +while [2][1][*][*] is March-May 1981-1990. +The choice has been made to give the seasonal phase in months and days, +rather than months alone; +hence the modulo is 1200 rather than 12. +Under modulo 1200, midnight on 1 December can be expressed equivalently +as 1 or 1201. +If 1201 were specified, it would mean that the first interval of time +began on 1 December 1961 (rather than 1960), +taking the combination of the lower boundaries of both time axes; +the value 1 is a year earlier. +The drawback of this %m%d scheme +is that it is awkward or impossible to give accurate +representative dates for the middle of the periods, especially since +February has variable length. +The absolute time format %m.%f for seasonal phase +is better from this point of view. +The season axis is not shown as having circular topology because no +information is implied about the other two seasons. + +

+ +


Average early June maximum temperatures for several years: +In this example, the dimensions indicate that +maximum daily temperatures (between 9 a.m. on the day +of record and 9 a.m. of the previous day) were recorded for 1-10 June, +and an average maximum found for these ten days in +each of the years 1980-1984. + +

+  dimensions:
+    year=5;
+    con_season=1;
+    con_day=1;
+  variables:
+    float temperature(year,con_season,con_day);
+      temperature:axis="T--";
+      temperature:subgrid="con_day: maximum con_season: mean";
+    int year(year);
+      year:long_name="year";
+      year:units="calendar_year as %Y";
+    int con_season(con_season);
+      con_season:bounds="bounds_con_season";
+      con_season:long_name="seasonal phase";
+      con_season:old_interval=1;
+      con_season:units="day as %m%d";
+    int bounds_con_season(con_season,2);
+    float con_day(con_day);
+      con_day:bounds="bounds_con_day";
+      con_day:long_name="diurnal phase";
+      con_day:modulo=24.0f;
+      con_day:units="hour as %H.%f";
+    float bounds_con_day(con_day,2);
+  data:
+    year=1980, 1981, 1982, 1983, 1984;
+    con_season=605;
+    bounds_con_season=601, 610;
+    con_day=-3.0;
+    bounds_con_day=-15.0, 9.0;
+
+The diurnal phase of -15 h means 15 hours before the beginning +of the day in question, i.e. 9 a.m. on the previous day. +No bounds are given for the year, because it is a discrete quantity, +and there is no further information which could be added. +But if the five years were averaged together, this would collapse the year +axis, and the extreme years of 1980 and 1984 would be recorded as +the boundaries of the collapsed axis. +If, say, 1981 were not used in forming the average, +the collapsed axis would have attribute old_spacing="disjoint". + +

+ +


Daily values as an average of subdaily values: +Instantaneous pressure measurements are made at intervals of 3 hours +(first measurement at midnight) +throughout the days 6 May to 9 June 1937, +and daily means formed from midnight to midnight. + +

+  dimensions:
+    con_subday=1;
+    day=35;
+  variables:
+    float pressure(day,con_subday);
+      pressure:axis="T-";
+      pressure:subgrid="con_subday: point con_subday: mean";
+    float con_subday(con_subday);
+      con_subday:bounds="bounds_con_subday";
+      con_subday:long_name="diurnal phase";
+      con_subday:old_interval=0.125f;
+      con_subday:old_spacing="uniform";
+      con_subday:units="days as .%f";
+    float bounds_con_subday(con_subday,2);
+    int day(day);
+      day:long_name="year and seasonal phase";
+      day:units="days as %Y%m%d";
+  data:
+    con_subday=0.5;
+    bounds_con_subday=0.0, 0.875;
+    day=19370506, 19370507, ..., 19370608, 19370609;
+
+Note that the con_subday axis is shown with two +subgrid methods, referring to subgrid variation before +and after its collapse. +The only point here in having separate axes for day and diurnal phase +is to show when the first and last instantaneous measurements were made +in each day. +If this is not important to record, the two axes could be merged together +thus: + +
+  dimensions:
+    day=35;
+  variables:
+    float pressure(day);
+      pressure:subgrid="day: point day: mean";
+    float day(day);
+      con_subday:bounds="bounds_day";
+      con_subday:old_interval=0.125f;
+      con_subday:long_name="time";
+      con_subday:old_spacing="uniform";
+      con_subday:units="days as %Y%m%d.%f";
+    float bounds_day(day,2);
+  data:
+    day=19370506.5, 19370507.5, ..., 19370608.5, 19370609.5;
+    bounds_day=19370506.0,19370507.0, 19370507.0,19370508.0, ...,
+               19370608.0,19370609.0, 19370609.0,19370610.0;
+
+If the 35 days were then averaged together, +the date axis would collapse with bounds of 19370506.0 and 19370610.0. +The subgrid attribute would not need modification since +it is already shown as a mean over the day axis. + +

+ +


Average diurnal cycle: +The following axes are appropriate for +the average diurnal cycle of +precipitation rate in July 1970-1979 as a function of latitude: + +

+  dimensions:
+    con_year=1;
+    con_month=1;
+    hour=8;
+    lat=45;
+    con_lon=1;
+  variables:
+    float ppnrate(con_year,con_month,hour,lat,con_lon);
+      ppnrate:axis="--TYX";
+      ppnrate:subgrid="con_lon: mean con_month: mean "
+        "con_year: mean";
+      ppnrate:units="kg m-2 s-1";
+    int con_year(con_year);
+      con_year:bounds="bounds_con_year";
+      con_year:old_interval=1;
+      con_year:units="calendar_year as %Y";
+    int bounds_con_year(con_year,2);
+    float con_month(con_month);
+      con_month:bounds="bounds_con_month";
+      con_month:units="calendar_month as %m.%f";
+    float bounds_con_month(con_month,2);
+    float hour(hour);
+      hour:bounds="bounds_hour";
+      hour:modulo=24.0f;
+      hour:topology="circular";
+      hour:units="hour as %H.%f";
+    float bounds_hour(bounds_hour,2);
+  data:
+    con_year=1975;
+    bounds_con_year=1970, 1979;
+    con_month=7.5;
+    bounds_con_month=7.0, 8.0;
+    hour=1.5, 4.5, 7.5, 10.5, 13.5, 16.5, 19.5, 22.5;
+    bounds_hour=0.0,3.0, 3.0,6.0, 6.0,9.0, 9.0,12.0,
+                12.0,15.0, 15.0,18.0, 18.0,21.0, 21.0,24.0;
+
+ +

+

+29  Invalid values in a data variable

+ + + +

+Invalid values are any which fall +outside the valid range or equal the fill value, as indicated by the +Unidata-standard attributes described here. An invalid value indicates bad data +i.e. a software problem, which is a different circumstance from unknown or +missing data (see section 30). +Invalid values are not permitted in a coordinate variable, +but the attributes which define the valid range may be used in boundary +variables (section 20) to indicate unbounded cells. + +

+The attribute valid_min is a scalar specifying the +minimum valid value for a variable. The attribute valid_max specifies +the maximum valid value, while valid_range is a vector of two numbers +specifying the minimum and maximum valid values, in that order, equivalent to +specifying values for both valid_min and valid_max +attributes. Any of these attributes define the valid range. The attribute +valid_range must not be defined if either valid_min or +valid_max is defined. Generic applications should treat values outside the +valid range as invalid. The type of each valid_range, valid_min +and valid_max attribute should match the type of its variable. The +Unidata special treatment of byte type is not included here as we do not +recommend use of that type (see section 3). + +

+A scalar attribute with the name _FillValue and of the +same type as its variable is used as the fill value for the variable. The +netCDF package defines a default fill value for each type of variable, so it is +not necessary to define your own _FillValue attribute if the default is +suitable. The purpose of the fill value is to save the applications programmer +the work of prefilling the data and also to eliminate the duplicate writes that +result from netCDF filling in undefined data with its default fill value, +only to be immediately overwritten by the programmer's preferred value. This +value is considered to be a special value that indicates undefined data, and is +returned when reading values that were not written. The _FillValue +should be outside the range specified by valid_range (if used) for a +variable. +In cases where the data variable is packed using the +scale_factor and add_offset attributes (section 32), the +_FillValue attribute applies the numbers as packed, so they must be +checked against it before unpacking. + +

+If none of valid_min, valid_max or valid_range +is defined then generic applications should define a valid range by using the +fill value (whether defined explicitly or by default); if the fill value is +positive then it defines a valid maximum, otherwise it defines a valid +minimum. For integer types, there should be a difference of 1 between the fill +value and this valid minimum or maximum. For floating point types, the valid +extreme should have a magnitude which is half the magnitude of the fill value. +We recommend a factor of two, rather than a difference of one bit, because +it is easier for applications programmers. There is no special treatment for +byte as we do not recommend that type (see section 3). + +

+

+30  Missing values in a data variable

+ + + +

+Missing values are not permitted in a coordinate variable, so this +section applies only to data variables. The missing_value attribute +indicates a value that is used for data that are unknown or ``missing''. This +attribute is not be treated in any special way by the netCDF API, unlike the +_FillValue attribute (section 29). +The missing_value should be outside the valid range +(section 29), so that generic applications will treat it +appropriately. +The netCDF data type of +the missing_value attribute should match the netCDF data type of the +data variable that it describes. In cases where the data variable is packed via +the scale_factor and add_offset attributes (section 32), +the missing_value attribute +matches the type of and should be compared with the data +after unpacking. +This standard is unlike COARDS in giving +a particular interpretation to the +distinction between missing_value and _FillValue. + +

+

+31  Compression by gathering

+ + + +

+To save space in the netCDF file, +it may be desirable to eliminate points +from data arrays which are invariably missing. +Such a compression can operate over one or more adjacent axes, +and is accomplished with reference to a list of +the points to be stored. +The list is constructed by considering a mask array which has just the +axes to be compressed, +and mapping this array onto one dimension without reordering. +The list is the set of indices +in this one-dimensional mask of the required points. +In the compressed array, +the axes to be compressed are all replaced by a single axis, +whose dimension is the number of wanted points. +The wanted points appear along this dimension in the same order they +appear in the uncompressed array, with the unwanted points skipped over. +Compression and uncompression are executed by looping over the list. + +

+The list is stored as the +coordinate variable for the compressed axis of the data array. +Thus, the list variable and its dimension have the same name. +The list variable has a string attribute compress, +containing a blank-separated list of the dimensions which were affected +by the compression in the order of the CDL declaration of the +uncompressed array. +The presence of this attribute identifies the list variable as such. +The list, the original dimensions and coordinate variables +(including component, associated and boundary variables), +and the compressed data variables with all the attributes +of the uncompressed variables +are written to the archived netCDF file. +The uncompressed data variables can be reconstituted +exactly as they were +using this information, +except that their original variable names are not known. + +

+ +


Horizontal compression of a three-dimensional array: +We eliminate sea points at all depths +in a longitude-latitude-depth array of soil temperatures. +In this case, only the longitude and latitude axes would be affected +by the compression. +We construct a list landpoint(landpoint) +containing the indices of land points. + +

+  dimensions:
+    lat=73;
+    lon=96;
+    landpoint=2381;
+    depth=4;
+  variables:
+    long landpoint(landpoint);
+      landpoint:compress="lat lon";
+    float landsoilt(depth,landpoint);
+      landsoilt:axis="Z-";
+      landsoilt:long_name="soil temperature";
+      landsoilt:units="K";
+    float depth(depth);
+    float lat(lat);
+    float lon(lon);
+  data:
+    landpoint=363, 364, 365, ...;
+
+Since landpoint[0]=363, for instance, +we know that landsoilt[*][0] +maps on to point 363 of the original data with dimensions (lat,lon). +This corresponds to indices [3][75]. + +

+ +


Compression of a three-dimensional field: +We compress a longitude-latitude-depth field of ocean +salinity by eliminating points below the sea-floor. +In this case, all three dimensions are affected by the compression, +since there are successively fewer active ocean points at increasing +depths. + +

+  variables:
+    float salinity(oceanpoint);
+      salinity:axis="-";
+    long oceanpoint(oceanpoint);
+      oceanpoint:compress="depth lat lon";
+    float depth(depth);
+    float lat(lat);
+    float lon(lon);
+
+This information implies that the salinity field should be +uncompressed to an array with dimensions (depth,lat,lon). + +

+

+32  Compression using a scale and offset

+ + + +

+This standard +endorses the use of the optional Unidata-standard attributes +scale_factor and add_offset +for data and coordinate variables. +These attributes +can be used to provide simple number compression (packing), +to store low-resolution +floating-point data as small integers in a netCDF file. +After the data values of the variable have been read in, +they are to be multiplied by the scale_offset, +and have add_offset added to them. +If both scale_factor and add_offset attributes are present, the +data are scaled before the offset is added. +When scaled data are +written, the application should first subtract the offset and then divide by +the scale factor. +This procedure is concerned only with storage. +It does not affect the unit of the quantity. +For instance, a pressure variable with values in the range +900.0-1100.0 Pa +could be converted to short integers in the range +�20000 +by subtracting 1000 and dividing by 0.005 i.e. multiplying by 200. +The units of the compressed variable are still recorded as pascals. + +

+This standard is more restrictive than the netCDF Users' +Guide with respect to the use of the scale_factor and add_offset +attributes; ambiguities and precision problems related to data type conversions +are resolved by these restrictions. If the scale_factor and +add_offset attributes are of the same data type as the associated variable no +restrictions apply; the unpacked data is assumed to be of the same data type as +the packed data. However, if the scale_factor and add_offset +attributes are of a different data type from the variable +(containing the packed data) then in files adhering to this standard the +variable may only be of type short or long. We +exclude byte on grounds discussed in section 3. The +attributes scale_factor and add_offset (which must match in data +type) must be of type float or double. The data type of the +attributes should match the intended type of the unpacked data. (It is not +advised to unpack a long into a float as there is a potential +precision loss.) Users should note that Unidata may provide a built-in +means of packing data in netCDF files in future. + +

+

+A  Attributes

+ +

+ +

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
Attribute T Use Section(s) Description
add_offset N CD 29 32 Additive offset for packing data
appendices S G 5 + Version number of these appendices
associate S CD + 18 19 Identifies variables + containing alternative sets of coordinates
axis S D + 9 16 18 Identifies spatiotemporal dimensions
bounds N C + 20 22 28 + Identifies a bounday variable
calendar S GD 5 23 26 + 27 Calendar used for encoding time axes
comment S G 5 + Additional information about the file
component S CD + 17 20 Identifies variables containing + components of a variable
compress S D 31 Records dimensions which have been compressed by gathering
Conventions S G + 5 Identifies the netCDF standard
coordinates S CD + 18 Synonym for associate
expand S C + 22 28 Coordinates + before contraction
_FillValue N D + 29 Indicator of invalid data
FORTRAN_format S CD + 12 Format for a variable
history S GD 5 12 + Evolution of the data in the file
institution S GD 5 12 + Who made or supplied the data
long_name S CD 12 Long + description of a physical quantity
modulo N CD + 12 14 25 Arithmetic + modulo of a variable
north_pole N D + 10 Long.,lat. of rotated + North Pole
old_interval N C 22 28 The typical separation between points on an axis before contraction
old_spacing S C + 22 28 Indicates the spacing + of points along an axis before contraction
positive S C 16 Direction of positive for a vertical axis
production S GD 5 12 + How the data was produced
quantity S CD 12 Standardised + description of a physical quantity
quantity_table S G 5 12 + URL of the quantity table
scale_factor N CD 29 32 Multiplicative factor for packing data
subgrid S D + 21 22 28 Records how the data values represent subgrid variation
topology S C + 13 14 25 Topology + of an axis (circular or not)
time_format S CD + 23 Format for printing a time and date
units S CD 12 14 15 + 23-27 Units of a physical quantity
valid_max N CD + 20 29 Largest valid value of a + variable
valid_min N CD + 20 29 Smallest valid value of a + variable
valid_range N CD + 29 Smallest and largest valid values + of a variable
+
+ +

+T is S for string, N for numeric. + +

+Use is composed of G for global, +C for coordinate variables (including multidimensional +coordinate variables), D for data variables. + +

+

+B  Methods of representing subgrid variation

+ +

+See section 21. + +

+ +

+ + + + + + + + + + + +
Method Units Description
cell u Value is a property of the whole cell (e.g. an integral)
maximum u Maximum
median u Median
mid-range u Average of maximum and minimum
minimum u Minimum
mean u Mean (average)
mode u Mode (most common)
point u Value applies at gridpoint
standard deviation u Standard deviation
variance u2 +Variance
+
+ +

+Units: u means the units of the quantity whose +subgrid variation is represented by this method. + +

+

+C  Modifications to udunits.dat

+ +

+See section 11. + +

+The unit unity is defined as a dimensionless constant +equal to one. + +

+The unit degrees is not permitted, because +it creates ambiguities when attempting to +differentiate longitude and latitude coordinate variables. +This unit does not appear in the current version +of the file. + +

+The units calendar_month and calendar_year +are units of time, but cannot be converted into each other +or any other units of time, +except that multiples of 12 calendar months equal integral numbers +of calendar years. The units year and month are not allowed, +because they can cause confusion. + +

+

+D  Long names for quantities

+ +

+See section 12. +This Appendix is not yet available. +As well as existing as part of this standard, +it will be made available on the web. + +

+ +

+ + + + + + + + + + +
Version long_name units
1.0 depth below the surface m
1.0 height above the surface m
1.0 latitude degree_north
1.0 longitude degree_east
1.0 pressure Pa
1.0 soil temperature K
1.0 specific humidity unity
1.0 temperature K
1.0 time s
+
+ +

+ +Version: The version of the appendices at which this +quantity was introduced.
+long_name: Case, spaces and punctuation are not +significant in the long_name. + +

+ +

+

+

+Jonathan Gregory | jmgregory@meto.gov.uk +
+
+Robert Drach | drach@llnl.gov +
+

+LLNL Disclaimers +

+UCRL-MI-127703 + + +

+ +

File translated from TEX by TTH, version 1.96.
On 18 Mar 1999, 09:20. + diff --git a/GDT/1.3.pdf b/GDT/1.3.pdf new file mode 100644 index 0000000000000000000000000000000000000000..84687af3ed6f67d320eae2a342b5730b424593d5 GIT binary patch literal 287970 zcma&MLy#~))DN{h$g2jvT^_+Eg`no-mZPgywW*Rwl_ik0Z;eMJK<<_oh z^?y)%dpou4j)g&U9oO0HJau1oxaWhzTysFnhyUs>y93ePdsB67omb`_O4j81M+^za zZyMr)%z7h%48UGUB8E(U@4Md#6kDQHGAd}uB2_dhY+4lwOq7~b?!V=_gTsO$AH1w* z``YDwIQHtM8fK1Q=EW3Is+e?34SQxkELvQ|v8s$(GjF{nbM)WeIVOU!tE*jgK*B?% z1qPU;&fG_&-9@$;f);5-Xf$$n71&Wx^4pX@&c(>Fp6qsVCno!6tsiuZGgzQ7YMJud4 zzR75ZY>zer;6)v=d@yez=YsiR%`}##(dxk6AI|UKU_ES-?1Kjb30rDC)v!Hku3%`z z0@HX)pfFyz|9Hm*n!pF>1K^nJvo?*k-3-!+OgsWSN4N_;UMOBo&r);+P(0kwcC-@ODr z9XNs*1G)9$xi#k92s|Bl;ZWhd60F9B{RAeGv-2=ijumH!;|1+{aUAq;cp!ZaUZI7C zgV(ShvJWa2vE+F=$=|}ryZuCzDIEA&*pnjI;rc`)zeuV^VfDd1#`QgJZ#P!X&>`?r z(FXY`UQS~>pzKph*;`!f#cfZ$Ad+B6(p+czGO7gFZ_O4}@3dPK(?GbRsko&gOr8-W zks;cE;+7JFoy+_6bb0ys!i?7qz~INFwii~8S0&Z=SsOO951~@jSG%iAGoqhacE%FV zBT4~9JuDb%nlNxOhA%ER;(?Rh?nbg(w2=^*L=eXkRzF-RnZtznCpt*vK`0%e!$tEF zTqX4~MKa%|Jk)1Fq!1srFN+b8vXHBe=y|pcL6#>x$vV{I6iMdRUoPN!aXHx77So{R zRl+PQ&nzWRA8ABTO2Fh_GYhL@YZL}YOC1?x4mZS>RDvCWwgS@)G3DAp&*4b;iG*oy ztTzn9%=R(+xgLd1qjo$q%I0Yr*mE0m^iV)1c{NkV?d$pBc(^RDv#e}dZB1i~K>gxn z)q`1`P(lBZ=$kjOs`lSwd0~vuuJ|O(e?~So$$kU!v^zRsl4<_<@-&eDA|M(hHRYK9 zYl5HxLKWZ>90k{OjZPveykh(yh5wC;OK!S?$!{@pXsg7zzZ2)qNOOmmDJp@-&Njs0 zy?tDeNrg|F`;p4m-x}kYsmP0*=Yzf%Di^M=`#N4UM`Cje8jJ_~Gfv65aCb{>#J&F@ z^j|%J=gISs_Y1XLyMxbz^`wi*GL+m3eD0~lqEXk|KOf>oB_;4w zCfuMXYHnkIh_NhfE9#7Ny8*}iETVlyEvjVOTDP)3%VzMRfT)RZqzj2gy?u_r2K2{a zMfJkSg@`nG6zUg+Dn$&L?7eCDZN|;=uUfPCR+3=dh)j`T9*d$>dtLe2bl?i_JV+Ij zVpX^xY@Fw!z`}+I_P>P8WlWKw_;!jygEnr!%!>|GK3z6-W8;|<@IOel#QGUmkyG+q zjkleRzNiPqgeD8h#j%*#m+lJIAEQa8p$TfZW;~+DBj^;@3Z&sC4j_@9|2P4vmh>Hk zEvcBHh5~yAH{~0{!SfzDU>B?7B(A2(?pV1PgghY;Tb3zd=h*z*Xt@QwCbb60u-Y1f zR03vZ;||-UFhzkdGp`R-4V2+{p&bE`L{O^95tBEc&s}0m%Qn%oE4Wa0Fgz?)i2&&j zFbomI6nC)IO)kDnJHJ@y2xo}^R&s3wh<0Qj^$K(nqEWsUFHpIACIxoa^B_JBWe$vr>-wnQ$xT(ILCm9Ax zl4vWKWqN04J|flJcTMjaYGbDa-fipoORYbbRX7mQnSWqP2i^)xIoFO}!Nw!lXUHQ7 z*Kk8Ea11ekn!ymt9om3o{=HX6V;tlB!}^ct_KD_VxO?rn+rM_!TlcdCdduB z86umZ&n1V3zStQJIrHGU0BjV4(k;LcZPV$Q=goD9)&%J}h|eP|y}>|$ZAHWPu=!gV zM?Q_yu%$myS=5|+0&1zU)gVKTz|-1IWL5UynlQq%J#NtX{3Lo%T@WW<*4_kgFc!Zh zk4MaXgiywc8pl%adopTS3(`e4$TPW8V$S#a+e4N{9b3E^lqe_QXC08T;#CS4o}V3QG?=a;qU|3)_t-6+bk`+l3uBOJIUP=>qjDeL9?on9=2BFkOp7Mn;lA6ktCZ zeoNYJF{QeM>Nj~{zElRC3HNh{b8(rztUkzbqG8xxG8 z0BBYKmW@?nGgP`@tX+K83hC4`^gB`oc*(dfnD#XSI+aNyK8%C$-WZf#7BruayTpRu zG96dSLIL4`NQ=hsCQp3&H5#JDh3iNTdD|0AcI4-}V?cIP#t2A%0*}v1jn;wj=e^3} z?>Gd=5)G!M3ehb4d53&+*wESjdu{SYK~}nz#0qx2W*xyTC*D^O%o+N?_%xt zA)8SlXOZ+UZBSEa)-4J*KW(5hmmtEas$GSGDZI`(j{iHWEv{_*mOIm#{+BER*Z=1% z9hI;BF%6Y;*$h^z@3%HUm%?Yguyte}%ON++L2WK@_U|K4ai-7wgbLCw7J@~<#h|K` z4dG02q(9<`GrF1xv}!C zvFPUNbKO4{HSP#>>R%?HCHgpzRP=WqAiS>!2r}H2ka|KQ|1QsGX^Z!QaimM$-^d*`U@-H-0BTI zCjFn=`%yoF4wwOysh!FHkEQ>){l7@d$ogN{mW7S^e`VWR%;~t}jz>QI#f)T_m{law ztH$=tycsgVCcrlCjbS;sCXwx#sU%{>-Q#|F)M^PyUMw5piBY2Hc(48o{QSLLf4Qe_ z4mfGb(cR@oI&^nWx?u|WyS%LK{UEvCnC{+*rURB66O@Zhzk2SGun4am$P{%tiaJUA{nW`P4fP77r%sVlW`BrSIv~g0BK?Tf3H*qS;Kb zG+dmYpNB_uEMA_!T(DfG|A6MQ$1D^As8E`_BVLq78HGs!BPb?1<06IIHwPNiM~(KX z>HrFEWV%-%u_R7eWC1Zhw+j)pz?UWT&M*o5m@|JXENxCTu0Wo9kpSZnmE*xYLIW@( zu^P9ce(aUluDgs#8;Q^aPfQa~M59PZ!6U5FF_Dbv8M`f{0sX*OiNv~-ga%wbl`mtvZ3@paWH^|x$eFpg7V|D@1!DOn&vRJ-ZIYgkBDP=8vf2DGb! ztAC)wE=3HfBN9$Tbyt>{D1t4+NkE-MLoSxM-hsWVpc&j3%WJas)rIulrO+@HASux9 z=5~tf1qo$Pxc?0Si_g#LGVv^>k^6jt5{Fiy@4_Z3b!^uhV*gpgJx;$)Q5upQDYnw= zx*$Ugac2Tm*iYacZ?}&T0kLi2#1wFsNebx;Xjm392+p{Ia`r+`K1GTi(msO6)zGZ z*Uzp`0cd(YL#cvCi6Dv0a1_km1tKV`lj*diLSOwO64Jxo_KZOfTrToOm$JCBHRH0y zg&P4Vl5Mj(HKc`f>uNe_NHj#Lgua)?6wS0R-(vcd9mbx6B6ZGdYyp%xhnJq@f#U*3 znqmtn0$td7k?-~hiT&2gPa>Mfhv?EV_0=2o)yxx}vaaOXx}=y_sD!V{x=5@wv4CGY9%ukGG(j-M`P($IJNvBZgvaEiB^681i zFni6)ktGmz-#x)kAUv@^Kq?bV`%Npf(9n3&8(bTlOxPh{QG=JI!~#^`w3orEQh5{S z&~;iNKvnPUR1C$2s$r$NVLqc*QvurW$)3h9p>!$20q*o7&EhdEyGKtW4vJ6`@4F!^fro9_!Daf^g2isbX^PRy;!)zBOHYk zgFD(WW}LuP&Gsf=V4$wP?RlHc**J?RU`K=RG!~B>pDF!2EX%!@J3irHRh$#%7tb|c zX)&8}18r?wxY0(FIAaO0jH~hM)=;UThW?U4*9&=JSI2JMa)6~Em_7KY18FZuxXPK1 z#I8jM&jIHY6@Y*;udA01N<(bKr5Ni%AkbiC?aV^~_+Ts4P;7G05d2+1ftp(duyxCW z8&)dpMM5gxm$eiM07uYKMz}=;{tX}ozMDodEHA>u8%p5tz&RH~ddISF`2RIa5 zY9tQTMolQ+qbU=IXkpl`mx^hZK&(`Lq=%v^IHN&Ft96I*ep2`zC@P8Y2XR*sUON#j zStm>{?~Y6>Y*Uc1`?!9&`!)rdML@9%32^g(hS5ZtrzLXoZil56(=Wpoh{(nQca z&*r;;iRq&)bduS<9>OzCmPm0cm_8Vd=m@ksWe22?ST!}uQ?d*?WlSvDIa2_dTorJh zwhT8V#EauA9FLfGHuq+CFK*lwe=^>L`T-;LYc!^L+i2vunx+HsNGDu$&UuUy5PelK zAO*<c?@4bvmjtH7=@K+-u)#Sdpoh=Vpyb?UkKN&l??3CP0h^7A zdY+a^bq&m5=$7On6uiXk=+p=UwMFo3u-KQC=IP_B(tL?zpYg;JX40 zE*4_h-PmKDsDYX+d*gsrJ~ZH%d+*7+1`DOHxa{IQ?qcXwxD7wL+J)?=cCA+LGbMF1_k5)ZeKN= zTK|-{@oQMmtFxYK?aMkL{x&UV6sM`s3!I0xy%?rVD4h2tc|X82Wk)}4Su{tVFzKGk$*PaI|DeBA@bz5=_CNuW(h^m!3x{!D$jU42cnQsxfZJ%YHCK3#8r z<@56hEJm0jhCvki_K#$y7&Lie?%y~OW^NNi&$bIg@BkCj0BLZS<_j?7#$t)rKy0%G zn1BaoGxdRwYG9h5Q2XEdgL?z6P=O&6e(}tcf&{Pr;v^g0acqkHwoUY-7}EMEd9$wJ z-`^jw-z=TXU0KYiW41n!Gfg=JJwHepLb`+pUJ8?C10zW9sXOkhTYNTmw+uh+~+wu zef9H=E=H;T=7W9vYQ%eMYhikd%DV;%?0);K<@febya0Xnk`ncsN`f6QnfvYclm7jE zetzEle3z2`x2lTz-vZC>xosQwo_*^3?>IP#%X52pHoqN;4GfzVGT54WBIzE>LkIWn z&y#k;q~YVZ>_IlmCJ*wJrST8NsO{?6KKs9}&o~jC1{AQ!{0$4r^Sa9Xs4n{KE&jK1 zDG|*Ive;jeMmZ#vnW!H3LdahZm27U*FGCF7sN4Lgzq-x)Y284h@8SD*T|lM#o8R#K zy3;kwi)(_-5yb;?t}kULYZL<6NVO8GS`V=tG@;=$qFTA+XEn5ppS% z%}<>b&K=WsF*Yd23x`L*OOm%OqxBrQhS~|t1@@2Z1)*yL0(jWGHmbX)d0Vwm376A7LkJtcgSsbG4T<=Up(nR7v2r`u$==N0N_y2)521y z8$?bI$8RV+fY+)WQ(njDfOF$c;c&;3%fgCiw(IC0qy73osA@->;JO4!^_vI`C=?n} zMJ$F?`lxvHb-XXO6}d-vnp7n{&HL=PlMUbjJ8R&2OsW%c&^^p@QSX_#ro%)49y?DM ze-HFoXP}cfp5#k$V;OWE&4w`&7aL!U+kWk2!ZkmeV(RRkQDPZ>sCy9vEnO z)r`IgylE?MaBM2Ubj@cP=PtW42`7Y<8c4rr zdfxoD>r5(r_8OY|z`4u#S%UC65hry!O*Ah`%cDpN+`dE>F+{|zK$_PyH*q#g6loujKiHl^Y7m*Mkp0DTf|<%wX1>?_zd^B5I`0B={H@G z>gzD%Z#`z6Ee#R+^{i|sV%z#1REeBNML35Jg&EpT$=J=lljxC%vHDC*jPf3=oeM)CB0Pq;F z2)k>l1PV4$5tlLvomFP23vs3~Q26SeQ26lhY4V0|h360SVD6U5Coj&Wg3iz>@BDL8 zq^XGjv#AS@IY&lN`u$r2!xRznyQXB#3n|HmkXrw@Cxq%6_tk{wS#l$hkFP(BskRCA z;0#Hl{i;Pv->tDV;(s=*6(rob)DU1k+d-WOFW>P0M08;6e=}WxkCi<+rW81iB-N(x zZ(0XaPQv%OxH%^j!OFvi^-n8aQ6S-OF<{kj89T(GFPDet_SG+b;1h^K9CBTkpASMX zCH7nBw?(|$Jq*L*$4BQ~A)_~UQ_2Zmvd3TwN9Bi1YDA{uBI-4rfm{Vy=C>0*ceYGi)Yy4j81#FaeZM zP=+p~Gww-f8JKM@in=C!C_a^41DGcwoB@APh#=l* zfN!{sm?b!OpX3TX7v14Q0tzo@1TKdTHVi&E_mVYqyb|X8Xsebiz80E7MFp-35{4g} z4SkQz9W5x7Jc(J42F#jS@<%|D`oaJr`%X=W{}=N~$fMw8Ovr)whR5Ilh=6V`Qzt4= z9e98dIo%bEV(vgXsrn4k)c24I2Dn0&h&PL)r^|QKCq^G+;^c*N5)okRKx-j>6g_aG zERlt;v%)+k#Lz(lq(^N@-lmVN_$kPhH6&^zj7s;F83sOA9uT2(BZe~gkZx1xqCn}s z`CXgt!UejF)%gQYiUKBAP-PP@$}mUj7=k!TV~H`a#?nhiI<(=@Plc=m3Y;ms*k??@ zggJI-ln9R$GHyzn0c&E_2(7^3VV426N+r9VxgRaFqwe(FScEW{Q-Cg(v| zP1+smxal6lLJiem$3O&8LxaBvM>HJ0fEcpt@|B{3RVDZ)9O_8FR6*Tkcj!Y%eMLW# zvGZvEyn#v!j^H-Xe!a=BQb6WaOeoT~c&23Nkx_yJ6{Ol0%&wLsO3ePdB|JVIRkV1C z`lqo-pG~Qd8b)k?-&MG4&P`xb*kb%B(c)e$jzOWm;sMh0mI_R;zoJr1&+=I}`_y!a z_0UsYVk|+U(o->3=Y`lOdn8H?go9#AT~WLmGFs+`aM|vCF;Y@5N_Y%zA1CBMh&&!$ zfOqr`0E!qn?)%2A$r*N`Fhwp>2qOWtj8MHwtuFH^KpGj~0Z~;#N*W~0F<&U;Gh+-% znf|N7dGMv!r&v?M*;v9UAkP$Im$BR>qiOajPV+h5;+hvk1C{--Eo+(J>r(%)16^aV zhGIYHSZ|`N;kb!u?v{r+7{xY$7GJKkBgT9VNUc+#U>N9EV4Mu@t+rF8f)2< zZ6s^)v!XYiUVF9y)?S+Q6Ll&*WSwpQnZ3$^lEBT4RQqaF`Dn{r!dO0<* zHbgy>x@Y?W5=77SYDS?~r@~I48v5i4`ofwBB{Iq_pbbHNX6LtEWVRUT90N3LR|E!4 zQ7)vX_VOF+a^0+mxeY7IBY=%N+kjF10>VTI>P>!>QGMkh5ong)yEf_a1Wu;}%Put( z#k@T1oZY7vdUymzx^x$q_5kkcK!0~?Sxvbch*fYzZU(*+TuZ=4+d&+Aso^o;NZS&(#KmClpD~F6{D$F>D)1c43v=NoGNSx z@Uf@vG)`c+;B5mRIba^F0o>gY`ldNz0M+?Y^8l#dr&|Sp=n(Cqg~RS|*IGF}O3mFv zYyrvGB?Fo%6Xa^@B8YQeN3I3@RUpdu;CX{3tRJ#fb%%>f|A1`F6jw)+Ju#;c%+n?I zH%%GtZw5;W?aq`pVQ@?71!06tSS^s1P`)e*75S}W!5L+y5>B7NdX2759l7H5XBWUs zm%8Q7Hhh2&DlJV_05%H5PY?utC|vee|KzkhUkU+~jBkH+fq;yp;>ctIVsT34v8LN| zL!>-RtV9v6)?-dvkLX>MDWTm|(CRQ@f!3q#4I0HW`B0C%|9H)9_3y%{kn35y!ZXRW zUtrX7)d}pR8Iv17YU7hajg7D}D-$1J`%BOD<<0KZgqi$SV;7z&VwD%++?qY4<2eln zkDn8VRKQn-M=$CY1L>3!w5XmRKF@^;5o6-cHR@txtElVtcI$MrdY~LaMEYFs8qqDt z;4m%ox1Um4>0=s7W*f#Ef30=GlWJ&Q_VOv^uRLS_G=nClZn1xuqOL4DF@+ihut*0s z3D}({fAXp?qt_7qV4*4a2SSkaDY?=aiMvkY&5Wl4CJ+sbmI$nM;>rV=IeCKk=*x}! ztxgpH9(51JUg{wnVYia_c}oW?NN|P)bsN!C`o7X02&y^0DXnKdT46`vC$=)4Ev#gi z6QoC?0;RP_Xn$}k>E~4k5fGQb)QyZV`^-t!kO8=;pb%RcPEx6l=}j5^DS>%qFiXO6 zcp`;=Vy?F(7Gl3gKJ?3)GEuZigJx~+Z+KX}{~y`x-4T(@ zk<2b{&VAHL9YqrZ)lzr(*`2ssU_c3k6?8W{j8EgTXxL{nL8|ZEwBJ%kJn;Ceodb}~OG%;|37Z<>&HPB6$N&}3 zKBesQz!E#O)oz;vcpIy~UyQOUb1F7l>E<0rKeXXGkC+-l7X7UG#Iz|!X(uPKoA+)e z@m1m+i9GOj#vh(rhD~ga*aFMM0*NzN&cl*(IKop047W4e+@WE6b!&{(-@zfbE$C#n zF`>ga*rw2#JbM2o$7SDQ-KT>~MSAogs48U3wEKy@y%lrZRGcG*@ zMVhNQyUPw%S+rvl)aQd2=~66#80OIZHZtIYmBrdE_*_LwF?A z5mf)cRNtW(5?~SrDXHqqFWj4pfg#?jQEBe2Ar zT574K1Vc_hoO}L3<-?Cqc1{TTxkMTa0qC}{sSgDT{z$cv5w73e3ULR=nFpcW3TRYD-$#4jxWGJ2*H3BbB;SO(P&LxVZo}syJY^1s49U6wa;|o=!k`&UmCZ4QUu-Omg$|b%LbhxG|s|~t|)53{h@h#&95gLZ!@v57<h=0?if!i68`V2xe)`0nVDhZBU%^e zGC15%>LG~}Xg6I>7~ewoi%VjFyXOl_V^8WX`f^$Q+0WweH2~!1BaW`pKo?-MlCkSU%eb(ij{O z1Lv-bz2fo$+wNnJ)z$zBQOm$vn0UO-f*EZJ#vZ7ZXwW^D@F2C4LM@nW_p{9fS;dsx z;0M38-Bg_TDe6WyUie7F5~OChf-UJXgwu!BAIwI4 zaQWIAqwLVybJkUXnff2!ruibHM)gLMT(27=)zq7$1gM~E^Q4%me1mC<6Fo(+e;pcI zIA;kslamIGpEi>ygAsOlfrMX?ABvJ9;lR0fx2|@r7$6u}=U+t8aLC!O9g1^0LH6bZ zq+X=a2$6&Uo~(Ho7cap8=-Q*dp-sF(+tWN~OGD@SB$#N~V-8=*q9}eUE8%Y;nN+R~q0`0fAH} z1S10ZPD2D&E|grpPY~~j&4(*gi8(0RHX;=4*Ptzk)Hhn6QeUB*LA8QV9K)guMk@d< z8>~{R6oQB?(mwIIB`5S6JRFAd>|!B_$|(zP2T~g$f&ju|X~8SN%PHar-6MIq`#Z)s zL4Ub+xgX-{A#>bnC|Ls$A!r6BYMrv^n}nt2>-eUAlP^p+*U%6>O$SZ-!6j!tnY5)jRvhqhva-Q;?84JbYvb`y;jFGv=RXtU>^ZNXD>`p zW59Q-+73BuL6Y{AFM%5Q$2w;;s%TMx{Plyg za*6~(vrSNLRYNsNLyPT6zn6Ka0m?$;0#h(5EhiB(%&51+0kTMyrWvJzy^t(`P4SCt z8k496;Ouu-Fy>0!FNmr1Sdwy)GH^LRopvU;)##+=Uqf*|({U{D&mM`-;ly#LhA^<1 zgqo~{AY61{A~HczA%x3FQLc5%r_kEtil}1Mala4!cd>RRK^sTV+bD_|MNuKt5&{8kAdw4#W|DOh-i{-e%b%=VKBz_K|ASpsR;CdKMS3YM3xTg~Ilpgi@}97zSd5wxm4G zJGOx#Db^vq({B2;WY%GCGC0qYJ?X^= zLrZ8?@o+E4z7j4H*cksQeGM(PhSErYU8ot^R~ge3=aJhXeqJjeedIQRIFN?4aN@^a z?nn;l&D&#Us~V=#qfLboKq_4#f=}86N(7QnbU(^sY6H5J(x7F2G~dK(;G7R?YM|O* z@p5A^mO(8$-D{RUsoN20@jm6ED_zV^i-JhwQ}PNc5Edy$;S@4C;0UETq4e-;R zbTA^q=aCF0MSds1MN`$cC=?Yl1=h%)l;wFup23IV6+5C8GgJX4>r(lSKr*+PlHv%Q z87yr(>^^>|hN>s4Tf4ysXAwZN@xm(;Hb5_hF?N&&lN+(Et2y%79I#FkU z63gV0nd6^FhSVatCegxE)TjR_No@A9)v~(XG_}~_H~4@8Za-f@gG_=*FDBoXz+0c~K?JV_-AH`)B;I;k&h*~l-7fRYxa7sc^21B-KH+9YZ<)|F`o~2o zE4cXmQ&3?s-Pqf7>;ybiu@e}OE#-c*cWdOD5sMQlrwD3riA47Q_+#*r8W1!YC3}kf z2o-lc2*FmSkkWL8don9e?vP0hS4_OG53iP>Tm$%QOwPy=*j6k$KcWOiv z&_44ssHUn4OOj@XxHQt+9SKqqmqt}s$+Gjc0|5}1p$fgJ6YK49u*a4jt`Ti4n2jnG zY3y03%E7HbF>)+g!LQH?$F^h1JJj)$@Wm)=fO7GO6^i~L4Imy60ZmHdQn30`hZ)k4 zX$`&@t`!~#oK-F1$Ydt7Uog z!O$?Yy*kLQ9H_5(=63$)HJuxZnzgLDJH2rSw}~X4kGVnm&b`wvu7N>#o7yY<9`aoiSva| zrFiW?Qs{+Qra9PG^U27)de>!!P;nSHn7;z+nB$%qgKN~=@~|EEFzy*5i%d-p9C{lr zCFaYH19z9=+lDlymJjcHlPDb9h9aqgzHd0(6D5x#(d9C+)%CIZ;O+!M`Zp6R^NLup zZzm*A(*o!NO5dxvWj5mUgiSJVP%US(%6vx@0xTot#$%QCSkAb%t>cVkG3c0tx;_TK zHAg2aE47Zk%l#|l%MvbuNB5dJbkVH>;Em586@{#@7EOpg&Bn$HS?6DRAzkXGwuw572` z7T-R~+uKo6=&$VSp5e5)>~-EXY8VSpP-|3){q4@R0cHx$Lunaxb*Mh)gJ~-KKc=3|LCIc=4C5}TrSv^XBf+Q8Yy)O+VT86o!P0^yVzb`3 zMfR6Hft;0(o|p+)3HlsUh;?WS%*;L5H}!&(s%?5QuI0}`z$?m1f;1tWrQuyLSqmM$lueq{800uwrfe7GBb zGMLd*o*dO@XX3;TS|j^kBaQZ<&bHNnhC^sFb4!nrf&P@1A1pG?FV~NO{g9lB2#TDEDC_j)ix`VT}>{?Io22JDOrH*=?^$^Rd!|0 zXQE*)u%2XkVpXkoOVCE6jpxqxV%h=|2hGQyON3v@?S*8juzc|SDp^W^@i7FKGo1$} z%>T)jcHrxm)IbFoVk5L52T>w<7=%(dL1Hor5cK#kpd+Qn}qIneW1&U=b$4}!r0k$dpz=S!D@40jO%zfQMH-E9U?lD}l@zF1eRkJ)cSSmfzeU+>c%QNS z@y$vTHCN(uA&d4STi>_P!s`k#jZnv9-mgZEJ74+D?ZSjX;x=%oIryn>Vfx`mP{Npj zpPgmz#`^cN6Q85@Re#Mk&?zefM$qFCFT3=D)zgf2nsc zF)^|J&x=(JP3L1aB)|81h1qE(RkpALS@}=GNx7;f&kdWg%3PbItsqhjqfinQq$kDi z^LBcL6xn$^c%%Ta=EaL+VPO4zy`Fz;wdz{cT=U7Rx4%;3tyE7|s%f23(=M*vy1z)c zVBB%zgL>_C`@hZXxxJp~7WaHL{@uURRC>CXGNzi3N72UnACG6d_xu=!JyK4k$3Ho~ z);*nNmbn!?^689qfNRLwWE}&@>uT?SB7T2p~-5;mz9<8C04xv)lR<9B@yHu*Aw3~H&x<9u^*P!w0 z4g9)6HA<=k(P^Ui^tS|+s!HOJ**4 z6)hCQZb1rrLRn(DuO+3$zMX4^&DIg4EfdLr@UdgW_r1-%_@CFwtO>I=q7{!aj65f= zVDU@)UeTTRW_4KacZw5! z{y6pNMNB8BJeYVh3mM&OarWhRzSj=J8Ri$5+y&wW%$_q}`fXr-b>LBd4oOKl(wV+q zws}|s$^Jh-pP&yX^_u1uratFm!+UTCz3E#t#2Y#Z0UPgtLuoEKw<+yzTPNx4<2=r$ z!PbgtozCz;c7F7^i9oNgy`ni4JHPl9+4G>g0wTU2`3H|w(bW@txLan+9(bX!rjC(= zZZ_>Q_qhxq79)Rko(;UuZ0|Bpf;SxUiXZJQV&KcJhgwM`yL|2!o<3IPs)#!;tva6`_stW@(#et+^yGs(G$Zd5@dVbm~0 zU6~cgPnjVgw7aosaP=4*jM!#|$MDey)9DcNc6r*E54_$UrzG4Nq|6 z_aLt0%^tq_5A%|V03f$u`b@F`F=0$Yf@^}5(t{<_6~b4N!TIR9AO}cenvfK4hcY=Q>ITC#~V2BPd#=twN4%}`nk|E}RQBq#*WiIi2N$SrE0w3@D| zdF-4U*y{bJXgQ)d+cOULV0R8}!so=B*AyAFxC-gvA3Q`?-)Sll^8kqpY`v^Sws0pn zye_j7-GFtsU|YiY&HcO5pp`ZRARBb@xf7bu2G}EUDmV<-K}`ku@NRh;W)`{&hq*w% zxv$-VIRm-^kw8in5jZxyxnU}>;}{Ym=?IByA=SG zBIrVoB&AZ5P-ON>)>@Dfc(C3aVH*$XY@##d9(ysuH4BeisIoq3^dLr29@v)a3vC5W z)GEJFS`DNc9;>-{WDgHOYGjZhICFhF;pnY?i{ShLXF{c-vP@73qU-tUTB9*^tGcon zvT!sEZ7*TNi$gsSrjzii5zrO}jq$xDUmC;)4vFrR4C>@mlG4vh2SBW6f1QEf5hu7g zQy2jo7cvj7mCZMAY)7E@OX8C!ZPo_DF zOI8eWyBR|eVweZHGcX9cHc)uQvc-8;L{}j8UwDCFEx|?ZGdjpmj2g_Ls6lB6keIr8 zrx0dQs-+nOaH$*Nq|`gAo8jl3QOnC7yj|z{G{des05F(OzTpfDgNCME)x31171~Y- z>7+Q58`>SqWnoVoO5|deF@=VJaE=VFJ4%P*AP$d!ildOba|(%xsO1g+s@oOs?;&VB zK+Jj|+H`n@&Koh`g@6LxjK>j!u;TQJJPq6BbpWT@e^ILde~{&fb!C9@D6g5m(hg+G zIv_&j&bf5LgdyYKnof=_2KkG_DVTvlWl0lNM929yM9Dr@7GG8B0B%o8St5lS)F2{| ztW<`I^0DZTW3Pdvrlj73uRF*Sk4*ax7l3rdxFMRa(dmUy(m+T49WrzP7lfh**%W2Y z?Wv>R&?pWd8bsNFTc(Qc|G2`s-l_BdeImy#8 z64NX&ArC_O?c;SL7AcqAn|SW+7qViDZL~m;N>DljC^_6~^t|v9VXcCWQnlq~~A+zTgF;)GOhXa17Iv!Oy0#2Rn@? z#&j2c5X#eoY-h^_{JrcDtTU3=qlYV-41B&E)wR($q^u`efWlU-SGFRsfco>%31VtaZ>vm*0x~=2Vt+bK|BYfI1CzM={H(=;h*B^HHo9@Q81~Z z(>JB!qb!ARHda3d1~4?7V|I-LZ-TGQ#pZbjJS<1X(>N8yp0Jw~oCANQ?Lh_*ofYwg zHb;!Ti32xp>)V5zayHh3ttj)l{i~o?_WnLYSWP14@n1O8#E}V*YQ* zG7HQ93_!J7n(@bDh<^F{gg3~bH5PA@NP+=1y>JNx0m$AlJRZLTi4};AD3zp@nDhPa zy0gO5k9G$=758^nS6qEmSKHs+_J4LOVwD{3a`t!e*(&x+E2ENDZ13T-^or_yXuE|x zHtu5A6&IyD_tVn~JS(K_^|GuEwUl$EGF)xA>S4;Hy1xo6;2Za}Y_cypu;iNxX({tM zobB7Wa>c^0gowR5BK>OL^w z+$c+J>R$t^(X%x1rsK|+OdI)Y3Y8m+=_05$iw=P30#cAN=i;iPy}E<{x_r;x9=UY7 zH?N+n-E$Kk2B6jOo^@~}7MddB@2kH^dq#!4e6g}e!pc87>A87`7Ztrvl=fIMMHIEj z4G~we(KHh)E)y+qP}nwr$(CZQHi$*tSu}`s=&X?zL{`FlRY< zGjm7mcp~eWDz;$JnO)jpE0OFcRJ_}%x=4O4sT7V&1lD)C!O62M1AN?D__c(o#8bBu z-=QAZeSvzgCO`X90S2_uzj$%I6NfHwoUYv`@)KE?H?l6=0fB`M)beJWR%*I%jSlDe#>gzvclj@TTDTu&KjQoK$s?r%VN{|uXak)Qr@k;W0L z8IwEB3vga#F1RPZ!Kh#;w^qUG-xCCTcQwlnCY-3d7-R1O5IFCm^peYf;#wNmrKm15 zMxjOu`0aNes^OMEd6=baAQq?<80meaNgsl-nm^g2bC89gpmE35NEox2ani8;-nG;1jz;ghPYj zY5IZMHa3Mr?U!?@E#OSVa4i`!mB`q)V#!df3xBG>2z8!8H$OT4XcnB}sUybMNFK7Z z=2vptq0We;$(A8r%9y6j+K{jWFkzDGgpi^S+L;oWC=|^eZ}#!(E0_)v4ODN9MA!xw za<&cbU{P8xkTC$PEV*&%*f}3j(&J*LmP-*}cfqCqE$v9bv*FT~$}nMxsT|-AAJd^W zPjqjn21UrA+*P$)fCjC9yIWfQci!t?ETUvC2bKU*!SFym!iBe`+rygn4=t|{4+AT! zAeifV&~hwPY+~_pA-olso69vD%7btX7(5=RX7J>6HNzU~tnfPH!r>RdEL5hH9w@+2 zqk7$3eK64*-ULsah%zD6!|=M85ST$6H9f(N02p-gaLW?m2kU&P^*$8_v!{l667@}9 z{!;g?$n>RVo}m}n9WKX0q@?vD89*8Msd{h>BjzVbT>hr`R1RI`quo{^cc@ zGapzWE*P#IL4^l|gfK&>_gSxs9GitJg7vprt13XqFBA7bm}Pz$^K8s%*@zF99Nc-=$oovAdb@AI-D;V?ARNCpzSmlU{fpEuK~ zN&1?e2RofQVXnS>Oqs>BsoTd@Y#T41!HC>%G+%}VV5HyfVGF`(zGqpO9Li~8Ft_Pz4bM15kD(H!3&Zl(*we{vbs@Is6)0-rPfwr+aC4qP z)Zj%o%5P-MGP2=O1GHCcP5=c(7ok{Di-=lNlt&YW;oyvC3SmYgYBKt>4`aNGH)W^KHcYk2fWR83o_IqkHnJ1G zu~MK3NhwQh%2@3n%tBFw8)r;p*P7Hr*UW|j;4$>M0K7?w` zHS30RD_aP?-DS)0_=rpgXs-YAS=SXXOy?=LGE=| zt1b0O11GY5-kkEXh%{kHWzKq#I8KIrlj!h10twoKIEa>3)7ot+nR8}0dr8ZxFi7bY zM+zv>gS^Vj0e6rsySUEPjMsnQw0dxzzDn zo7nUo1IQuRQ;QmHSUyfj5sGU+65zF0kBs*=;=gef2$ z36&mLye2~~u(iG00RAt2Ldz@)U~@bl6}?NDKB`x5BDA`Y6`Q>L(`ZNCkjELavoEpF zr@vsEUyZ$^ms4I7Hj&!K)UZ-(M;}ZEk^fA=nQ9UVCD%8rWVXN=$d$tgoQ!3RGe8nU z9aUwhP|L|Pw7qD7*RDJSN$EAc0*Rd9b>>;))Qb}Axf>PCEMqns@RcE??G0GUvyb}P zPfTyEBcSJU0s0t}CxzsE1De`XVLVAcCUXAmF^u}bXFZH?JdX|j}<4Qsd$Uxfp$51A+ z|0KuWKqVHQEejdgtdIkTic3||4YNBR8#j2u57l!ErUGjg z#7JwT%vuLbvGIbVnd=er8H!i^4es zdCcS_in~iHi#HhsHidAjP=u!fp7c-3&Ir@_^%nKMT1S$N0QjMag3;8w`7}LWW>z#E zuyBTw3ln-WH3cIs*x~>}z6Ip2gA1=Y(aZ4Oh>J||DM^)~qZT`vWR3GLE0*<>26g{E zi3F4!!?eI}zczt-j3v3gHN%umDbHjEPi9(sXx{y?FRA`nljl$nXBnw+>62wS{M5R< z`teSzVZl>EuD!y54G|FAZ8z8JbI%9QlD&TG0v&nqJFD8l$XG01^qmpriV?1PJxBRqidN9luR=;sEAA&*RR?F< z^c1eKso8P8Yp;Hu#9;R8O#fmz2GV#kUI=ix0{OW^`hO49S2aN{9i5*|;T`$jTSbRw z(5`i;>Sj!s4TgGX`YsU06Z^vmr;dbS>?TsAs##n#-sMYixf1@$w;HLbT=XyVHf?kQ z@N|kt6v!XSDqD9rW7d+Yl*EUpC;fqiZLS>V7e{!oNj%aF$;gE)Mr;Sbb7jvLtSnNF z#Fi+P2pB*{?sGrbFiL3%wJG+$?EIcKifYe@-eT*B8wO))ae}%uxoNkP*SWU zp%YcVIoV%B=an?l3+|^ZFx*HXVQLM0`_rki+E+lglA{ns5rpoPOP~Ct9@~|(CTXrQ z|Bb5Kh2S`&!9EXRnX_7C5CCvXFdN5OS=WhuQH0d0`Y$bn{>De6%~V~1Ogk(e>gT&L z_vqlGrTa4*?HNyv{6%(i2BD1--0Ty%^A)de3S`e|Vi5Sc_Y#~qrhw!SVHucL+K-Ry zEa9=La>`?P-=fe6gIL~QbzG@&E0KJhdhdrw50>{4nh?VjyRO-$mcMrp$xST)PC8%A z5Feu}S@(b=AooTb$46Me%88s7(fTci=*5ogq3 z+!@$6J0)XS{9wtX2D;8XG>vNjxuO-;`T+j7K)8H>R<_LAKzki!8pDy<{4J_63BSv-iq1 zlSeXmM6m5|xaVA-RR3j&WB*Ufo0*A?<$qPNOR=WouG(Dr^$lmHfOV=-BqNt;naWIq-y8m9>g9V~X1IQi1o%y46Wv8s(L?9GAEmIvZ`Oxf zRZQg>s?Jw$9VxWR>WRZ+SKF?y*2KdeKVx^01 zZl)eqA3iwzrKaos+&bF9cW0ETRPB*EByox)o@rUC-Z{8k+gqs$n9NtBV#?I-U058R z60r=+U#a%r`Sy7QMznOZKr&Bk^^+@w&FkFMxU#TA*0xa2!+x~Yoq9GuW^=&A4c|H& zUQpdy0?@z)cj3mdCL)3h^W_WCBV{0~pg~(lqc~kPUEz6ptE|0x0TiHrQ$`c`8X=YO z$uSS(G{Io2hp~MOp22^F?BIn;bGqY98rjB5(@GBUzAvD5w^c8Q*{`r_kOf3E@aY^` zSuaccBUvE51H`6|vx0_ka+o?i`jd@O6=+Ge(xQ58%$I7>FuM+TZY8ivP~Z3H<;g+biq3uVc8MP-Hr}XOSwxD6tm_l-9#>)&MJmz)nDt2t+qX@}sOYr%ocPGAz%JI=hIN8PvFAq7 zTS?|-kj=h`Cs^v%Spf^6_ab1_pze*S_MOK9_c1|l1OM(X;i~ayY)o8^RM@pYAEsDt z(^-&fftoEC@LguKg~9~j3N{C4*iZu+U=OwvuK?E7sGT6W-Iv7tWJ_=K3<-cncKqeH zmv{?G=JTTeq1DXA0bNB;2%d(;J=QB^`PC3Yc$NFItNJ3Mxk0+dv75MaR)^34R-Z z_NVrN{0>7*3|C7G#=QS@zkdqPWVXu%);tHD#-KnI&6KWh)}dJEc$Vs!z&Epd);_%q z{VOlmydd9%@+K5t@6lU~xtQo`2{a2)jJpKb8UgOJ>63^BBgE+&@_Oz{_G?hPO2U(u z{99HmNXzu7vBa zMeJ6ZF5f`dYfrS+CIP}__ooj>%QxaS0EXMvSuegS#P70acIXRF!^cHE}_^P;^Iq}k@Y9nl@Toc`t-+d+%GksnaQC=%Pc4%yH!0?~#N7I-QwW=M9EY7;HTV{fTR+=-~HBQ-Deh_RUlCLNSt}%RzAt z3&Qbf66<$UeIP(UUE{zMdRq6m&o#Zd!D=m3iOCCynu%@#AqEPn2wR!S0y6`Ht%vj( ze1N|>$^!LCvRVZ$!@7^O9Xw|b_~Z-!lCNU)sK-z-!Np%LW1%>()%yJ@6cSdt_EAYj zwFGM%ne5-v8-fYvIl8~4*I*7EHi^9B*9$|kpiD;O2G<-y0HTxR$)(}S{0{c{{+a#< zpr+S_jBhO_eG~Lp)~ouDbyo5sn7)E`!ypBD3Ci)uEG_KAaPQQ?Z*(nvCzXmPfM7yQ z0(Fi<2J98k0n_K6iutka1heLM?rO%XFnM`~uEVEvnyd=w#eZG%7+(3gl)PYcKH5I^ zPoZ+!@qig<_%bpM8E6=BBA4Fn3>?LOvJmnNPXHZoOI97@Jl!bn+{UR#j-xUcBBQrc zFYMO5fUK5Q{yb1FFL&sHGu07t;3QZ9qVLc|$H_6@5%4^gA){ z25UjEG9H15rZ@$jKp*p}BU7mACupw1kx6VB+y+16K)Dt0Zs98(<`r(K+W}U~380)E zhU1(G*;@@ME-B1Zo1v#5s`W@vi^U5oVFVMNNT()`c4CA$#00$W`s&+OcWurf$N&vf z*s)rYv4D!d5y3r_V!4kJlDr?{VmoFGI=>{0I>+Olg~;}5_^juj$lEaZ=0#tjg%FT! z!Ja+Se=5Ndm5#Tei|<}vA|%l>+?clfH9uQSq2It%7l$&^E_0? zOK;|yc&^1APXB4~y{$5`l&`tSQQor@y6VVUOlLZu#=Sy^bBXz<=eu_Tj4PqQbs_Z- zr4~7a9Kf}1<7jC~{mc;6B1mxdO7BXk$ij4nG2=^yo^jTOgeD*hXgD7%Mzrf)gK`4K zPetKD50<4Olr#X6-2O(Iaa!Q;?TBm-tB_WcKta@%z zE8gZNB}RXBPDOTR6x0n~xt%21s?Y^g&Lli!E>k*6kn)<-;~Pu_I(OoI+TSY)Qz7T4 z{>7#4QEO=LK>hgFlp4{za0A}{MEvh3x_ZJCX4HlB4e}x zvij+e;Z_imuuSjd-a>71v2L=)pSFXPsgHf;q01WMLqUmYLlYafLn3%J#SMxiD{1>1 zAPUM9g(+KsrUXtQkT1>Sg7q{l$s@#4{ShBVq%`u4N4=rvL;5oCr1{|U7XWl1X19}v z63%MlJ+{n5qE7>qoo{iY0^guDtWEpnOY*qsik-M>*qaJ|NcK`4YWSg}^XKT@hEH3= z4eM>}(MRkXXps?v(TosPu|Q;1Ppr*$*eC>s>*haL~u6VxeAM{}Mwe8-?U z-iJD-)SA*l8B4&n* z@yl!eI9=@=H+$QR&C0i2Q`Fd0VO^T4MlvWwb+^Q>{VOU}WP+vLX?}Q%fi^s@A?TS? z$1JHi4Y>2FGpQdzfg@V9iv@~(i;BK5z0_k{$P+jD9mL z7?Czc9g@wmJo{O3Hx!tkH-5EBRI|2z-HYHB7Qwf&!|RhUICw-QANnJ0NO z>}9>?sN_g><@mMlTFO8~QkWK=Hg55Jw}eiiG~D=2Qt-zHAojK%-uKnxH9fv+J)}^d zM6$(m+kU9veu-qyMDq|lbCV}6TfbM?-0j5#oOiwL{p!z)@RsiGE6jQ-#Shewdj)k zUrO`sI!W!k7OU*xUo741_Samz8Ks|~=hNNEtQO_9w4=ECo%kKSI}xLK7^0s__MNA! zg%#qsg5sYWU)`s*(mosEA2=C;?Cb$GQxl-AiU99u29Kp5L+sKi?sj9<$NLLHNS@%M ziKTMo@THe!O(j64b>3e%al#a@g_+-)z`IJM56x(ot^0ZhU$`NU!^UApN>NYm-Da~} zw3myVoJqG2){~Zj^8KYvm~k}I6kg=}m3wtNVd*c7J_>4`H)${~Dgn0hDg`-pk3$+T z-F^42B>1KGaLZe2kDmbk%7f8rH``#2`(YwCA1@59PZl^k;AQ)?gf@eZuY{4jbJQA+ zDP%uD4IpEI_!uVp1k{GOwqfm~CNs`~1aIS58+nNbiQEPUzM!+g&BSuo{H*07Z6eo;^9F+9oRa}jYT+*u4$#i z3yG24+rPr+s!g%HENHM0a^1*`f4iJJ^~Gnk1Uoq=W`*bVnKSV3eI#$6#oPQ-QRKvX z92({+APnKl+YIq`A^AZ(_7OXY>M!#ok=MQ;zgh2MnwIAeCon-iOQ)dX{hS-l(GFlJ zQ@@R>pq0bZEayxed0x!;)q48=i3!A9@6pr5eSz$$Z-#kZ-IJ_Zk4cMc$B$4mQvS@m z&c1L-1?(Mt#>+z|3}8+WWED{wJXoHsIC@UGBV~LriJkk|Dt8CwFJk672|D3NKlT6g-p|q#FZ`Uc7}->)*0k* zgZh2FmXFKd`H9SR!^gYp-wA0zvK@#GN~&boC$R(_6Kq~zcI*2?vysvS^}j*+{L0ik zi@jS0=BsAs-~Kqi;*w%T^QxkJ2tpJI@q;V2Fk%q$%u`cf1VgDQAlzzMyXj`h@I@k< zA8?34=Q|42IMnOBIcm|s`mHf3vuC=cOlbpV9>FJ(xr43)#@}Fatp*Lj?*llh5a4OG z05i;D5XcIl)h*KoT3S}oN>I;6Hnvj24l?oB)BtsIGYmX?lO23RA}L1Mx~)hgL^Hd1 zf6>AM85)N~jRop6eVoMjQ3N?|1qp(1Gch z(#d$(XB8sL40+BXuBq_wZbiL&LSaz*T!&=<9LC^-h_AH(nYUp^JQ#Gb1}G*5?Q?DL zc1ZZ^bjL{fqRPAlnKcl>phQtV!Ro}w18A3NC1;TsydOYk-!lHf7`3fo7}dO&T_JmJprKsU!MWALV9pMn&TjHz( z9a}&{3ups!HyR!LlOcZ42c_^p9Og~cw^gZt>5Vn2D>r})(O_x+5G5}Y4=)3Qp3gD# zwtUwy2hR}PkckVS2@INb2A1YG;wpAhjkUa~THzJPJ(0&2E-C!IX9H z0zB2=a#0j2Oo~YAja6T!5$cZmp_Q@|^1^bIkVnAWmL_tAKBo+`HTGa71zvc10fF&s z+@AKi%<{PHgB4$J72GYalH{DQN{c^wF# zUynyM74xJ~A(Ot6HvnDhh%<0s*4<$m>~9K7<3G>!vvL#F$?1s4_Nq^;0}+P!!hFYT{!TMw~_^3`3F!| z+6p?=&$x+pl z*^^*sEQA&YoUmPuBtG{0Z=0+uUt0<^=oGg~E%lFs6aqwl1=Lq6jWKWAI3tOGRpKmS zpYn^gAyZcsp9Jxu&!V0wc5(O0HQ(RlY7cb7vh{-eW5_c{!+na#={2hWU}Z~puZ$vG zi1?@Lu}ad_$S7yl)FMX06r5DGw6_old=qulZ7iEQ9(OwT1SPmRNZ~p%;apuM&a)v7 z&m$e`GQS%ngYMu%d#qwqV+x6$-KMUj<>uh8moMtrFch5PQ2Wxc^OAQ#>TSU6#;7zh zOGvg)D!GM(depYD9cT_W+u0^c(^U&N%!2q{jiV5va zKxcHBbAruaRQC%13EP<>?NAMPm)9L>Q{V@%+rXi)^4@BCl|%sGVYz)YKATJ%dh`(%oo zled;0RD@Cg(8ag1Hr?Z2!n5B&rWjB}91t0wS|-TT6)s;3>*;5&Nf0JuBV=>PH+Uc| z?kE_`iNNHEsXxnUmjLNF;J5iv2R3gX)*i&NiOZjQd?a(s>RSSJg$)5&n8T%09q1v) z7t9piGiboz@_&9}`~h0c1P?&*cNsXPwkcxBj4DhrJPq-D1h`Q%YpLNJz~;K$9O`sb z667WWJ#-V4JcMSY$De{UbEVE5l+1vVRRlS5Hire7KJ7u(TkiEfnK?+)`|Eal4470Y zf~uP7kG}yjx9X^T-OoZCJx;P(e}+tpjii+iR1}4DC;n>XdkAy`gr8Ef^~1Q{s2&UP z#(X3$R#ES5yR4+tP;;Q3yhm%-gSmn0z#WV!b9{ua?D&O~9GC~-uJY@EG#Q5Rr~PR%n5fT4Z{ zPQ8bPTI2rWw_79H{g=P<|Jn}2%=y2%6}6hu&f9EAeOLd;M=VuI2t4l*)8ONA?(C_v zd^NnJl_;#jmKxK9rb)ybzkdBRxtK^s}c5=(zJRH@LGcuF~;a@;sjlc$c)WOV6$x zRph&JK*xixoifdv*o)m^rB)dvl*e~E*O`1=1*%4u0D&ZqprAmY_#-_9<5gHn_rW$`^2&o z(6T9TCZ$lX;hKFsVBW~8%^v zO_^!nuHOd2&8~Q<+>0~pDmk<)L26yX+$@KB=hhx|-^< zoe|(6bbGo*Mdm@_2~LMQd+=;L42Xz+3oMOx&aLJH8NMe5d9x~r5oA?uEaPb9Vh59x z=vORR(K1zw>EdQVJrFt#O(!~7zW%eB*1`*YAW0?hExE7DosWtEc(c>~agNFcFyWpD z?rPvVFhxyl$3tAH^3x!3g!!rR8->V`H%QR$o)uPDOPs(PA)Y_@WkG|2%;B~9 za`ii@*X{9jO-`7IRqE8Uzc=dpr&rLT*CjAK_JYZG>SH9FK5F!O`9Ie- z;QMB@>?~UpETQzUZZWKCo6f`6d9QFz4a|Ps7nOxn;g6w2#Y@9*>Z@<63%kl!lAMTJ zGDzIMyKc41sMxX}?e@(HD}_n}e@37O*OTXgmyt8+Qo+S=FNejMDJ@#JQ|wpnN+t8bL%t}X$;8Qt$jYLB z&z;(3-2e1m&hyO6WxV|AF}W7z!FDqc%$V>_A%;vSWFIAruQCQToI1?K+H?ODHkNM1 z_0Y=BXE(JqWS{$<1jCG7V%yWmqB~E@%-2E5CcMQsD;u7H>*`0mc>IuHQ9zVVK*B{G zqq671I)+iGd+AA8nYaJFVdIcf;7}=6)E2nW_2)ZI8VluoRJ_jdxogicS**D`sD;H6 zzFHZYv3#W>@BAw{?}c{dzFwh3%ojv8X9`^|pODwC8rCkZY4fIowKb$L{(ZzbMYA{N zz<1`t=avUKa)n%M2vu#PaQiWGM@lvp%y&E&Y!oMg%4n42q?{zv>WMn#m82nXGT}xe zR_NClkelY)r7f&Ic3(=A&g-8-y-^$ z(Qrvp_Z|s+dw66&D~PHG(q?SgZV^{)H5r1#e1!b*12>pUBoFg5XThVc5%;wRWrWPI za$-M)S_##=c{o@0EwScr?{$@}DoaYMYMb9vytB$l{qOe+FFm@i|N1AErrdh4O!^A5 z+RJjG*9X?DRYr~h^oPrBL#5i8Svp#UmhrCY8w`wy%=v3YOL_MUg78HqxI%#0f~&i7 zcPFn+oVS-&e)N-%q30aoB^@u7S)p1q%V*6C>2Px~izbTdy%E?>V#Wby7_e3^AD4zd z`@_1*vL^l90@(CS;?#B}u$}m_WfY|-lJz$WvB~SiSw#OO|a~l-Ap?qIDVK zxQ~`({2|6zw-MYs4>FG+Khih?m<*Stuf^SFSMT@V%h9q+9+np(vNom2i6$4sQ2BaD zUkr^cpUP}~t@`AVm_a&BL~5B?xt2_xc`*DUFfub1w9#yztQ84MVp`-zzM%F7!g6bw zyJ<+)MxKRHNG7}4mXK%R(V`&z?+h8K)1>_lNBbvXOozN5zRC84@!BwD)oLa2DND-*l6%AU=31$pU8+{A%-mDwh% ze*k1YT61E62-EexdXP#TZ0pvbz1PtRW9O(24E5da72EuGJ@ar^{C_C)FJt6H| zr{0^7@n!4j^?EYh{T`Uij753U^n~_$Vua2LqCi{dmb?rPm!cKEzXeh$k^In(WT=tT zY2l|3;;d^KWriT^Jj(Z1;lh7lR_-YiVsSZxCs{@(F}{`c^+Zx>VZH7uCAgK8Wn@S) zi5IN5hp~BVrLs;bf0rUuj*|EgQ%aST8G*57LRT)7g1-Z#qNL0Bja*mY(RpC{wI`1D zq|To|AQbbIUr!EbceWak@;H8>h$${}q3h|cYtx|BtfRb*xc22)Co0IIg|0?l3^ z$-5HgMN!>n?yCUz^SRyTm78&}LOW2+t{U#$L(eNl5P^|1RB;9>&N9YW^#K)2QM=gX z11X<+@>MgdJgp)a*~;yYU~0)OL(~@e?2}>5mz{8vS*#pYu}pdy>a0xG?^jv;oZ;j2 zQ8h>!i2}VC>7Mj!Z`+2W060Kj61B*WDWw&Q|(t_ zs$c%ifiTxWr4OeKJOk7k^#;*ep%SPb&-Tz^btbOMa{MSbvrzhjgCDMXfKllMRFg$? zhw(cT`!1%T3K{E{<54>HI48SRXc)cIjAH2oqNRBRBzO6A1BR7TsKk#NE~PjRnSM2% zAXY5VajXevRq(T2`a3UJtYAq+k;m&6paWID@`OVmey`x=8K9pBl=|^V371L)cy{S1 z@TY8b{PK(=T=Y?TiQhNQ0yb1J_p~8l}Yw-kAPlNOMmq` zyZJM4!RdrH5vk_CX+NGlX|xfHs;*CkiN< z@Em%uvn~1zl$my!dl?;;UKboKq# z*Ff;XkqYr$Vq@yQHhtlUF3=9;`BbBibxI_;J9v+j#>QUJ-`0ncUbs_pe(VM*X zaWvB4*kak{+SDg9)$$%$S;Q#@>CS(>SAJwY+pTg=&7Yo2sS_fe z7X}8Tfs1bwYwPac-Sh1mK1RA(mOGQoux^%ylHknT;LW@9`(fS0HAIku0ZhKri+Ju? zGSu2{F?+qZ)pzGt5YxA`M%0jLC!yUr$pbi5K#3IWM zT7g=*zLt%s<2kL6a66lGJ3gIeLIs&@a|p=3ry&s-X_jaF+LUQxgk1LL7|n_0#SnJ# z1%mG4dbScwf^4;VFGSHt8NH&dLdXgl!8xyhEOuPc!LXkcxPjZA#0r!zn zFh%k68`QML-^t`qPe3eH%md3$n7*A}&S#~rn0nUCh zvFD&4u!<)gX_K>0y98Y)=3#Mm0-pO{TMQ0HXtbIiO(ZBXyb6z^WfoytGZ|#EpGZsWGBZP?CRVB5at;>J`!oM*!_)kq?{xX~IeUPA4F(a{ zPa4j~`-pI686sAPd{D=pH4u5D%ET`!`=>(^GPwTb4kCay!}s$Wp$O!g9o){OVk+kbFZp3{lvjfG%wZyGDSp_mH#&fZaMHYAFZ- z0;mFIWhmi3IWrNd%BoW5w!3*xr&f|&Wh&xWgt&tfQa4S=!@dSw>HXS>a&++R8xta2 zATIY(pnBfTXS!>znD_=$yv4}fb?GWI=3quQF)o&GLPdb1qu7;j2YAUwv{ANqwg_UO z;QJas1lZ~)%!-&gQeD%>og6BUXBxplEvCapqDd?EsH4S z6@cnOs_~~EK%s&fBld|oZB$#;Grw8>eMv^pimPHDe8Zy9#L);F3XGvy19uVZsJWj4 z6DL#wBRHc;_?Kx3aB=TZU|MoWBVWhwoO|Hs(*z(efEQsjHRK{4o1?VXr*p{U8m?DJ zgOZ5^Amz%5N4zZ2K5~QYp4g`@(>gnDaoy~^fy{1{H0)mEgOtN((I&93#V?p}2I$>+ z7@oU%M{biEpy`)fG*G~QBOEV;!VMDia0>M0;Fm+F2Vg9*!1K{I(3@fV*bA<({qf5r zub@Jp<=kzo`l`HiOPV;9{G9D(RSPp7O6Zyj_=2PuHfL491B~em*1bq!#}I-#vjDh2 zSH+-0gz~Uk9a)D;C#y}PSsQ7*2*^oYdG^1M1|Y`wtJQky=zR_z%VQP|G*%$aDG(YB zR|S_h8YsKaqkM5mah50CgKIjM)`$ph(nUc_vckG!{W^~4hl18i_QU8;vWb^klUoTv zy8qZ_WrPG#e_39Wo>g;H*f%DMPLB-=#Up z33?F=KMOqlN1IEUUs{nN5-i2Dp}fna+6cjAki&!8!(IpE6(+3@7ocCP?AS?F_tl8_ z@XteiFxjgJA`#f&jx1iHYG@LT!qDO8;{w#tpn41v#EVJ%v+sS|z{h;^Bo0d| zoH2e85eny*Vu@5XiWW0!)T;O%<&%V~=s=y5>+?SI&K2JOxE{s@sBo`zK)OfrwH&KI zZlNtjP*z@Oe7R*8i1S4R5E3Ai1)uQ{|0!Iyw11?h=VV!oR()SAzC{Cr2MPY2OgVjM z3M}qt*$yYdm_QAWZS)%<4d1~*@S3dC_{ZtP9WXn+toRD91JC(i<3>2)46XZQ7nQ~J z^D=Qd*(T*v;m>#! zkywRcM5alL9zXwDtMQr0+^En2|4)+?;xEX*v)A*FSqxnjw3F%R*rQ)`NGu~$ux#nr zqy5of`&{9YJ3L8u-gs`S)E_j*qh&wKJ5U>|v%fLkI0)Cw;|4=x!pn9uU2a0D0Eq6R z*WPWg-`v;)K0q%I-r-UzS2k@25oNE;@aAZ7S6R? z9T7|+heHmrr?pUoBM03N5#`JiS&Xf<`K6)R`Wa|`?XMMT_IbU7?{99oLYR|G1uT3T zI5z5vuU@7}w96QuezbPvQ6-hCpIN-c)LQ0YqjqPR)L90}&5w{TLFn#OY$kQs;W+rav(_lX$HbH;5};u{4n3W2ah#R6UW4K+Ojz zs6{!NKny&AZVuhhYql;>@n*%X=VLXu5*K(@XR>b4Y$xZmr&7%tp};D?gPor);)#;s z(D2>a=Iie6-SgZfi8QjYuTbxEyxOP;Sm9q6TUCjV@OeuA%C2w6qM!iXJF!Fn+*`^8 zQ=~!PHlEG@qY{l9v6jJJ=Bj*`Jmatvh8Iq^(F9!-CgGL%anXNd}P z_Qp%8hS{kUL^wb>PN3=5AkYKN2C( z;Ce$ZT@^^Mv~?5=DK%kbDH)Z^7(pCqf3Gl`befjKC4;W23T_m7fPR8SHd%#;upNDF7lT3f>Nhws zIl7G5lSk4il7&A7UweM<4)LHm6qT5)J-AIwL*=6!nUYX^m)v2lAB0HU+UO=I8r>b) zk+{g)bhjA>h7@e2)(4lbBX(VAXyym^1Aw%8hjP6*QBVOkVOhA=njt*nQ;TQG~CV09&M+(hY{r&FJwd(@NBj24M~i|_!F?FWb33}5ZQ z2^Wtyy8&Zmld$w#gjCHDc;VF?y6~!|igq5-O*A!;RpViBptZRduhPBqA%inJ;mmWA zyMmW}0~F$68Gy&2GY#KEc3detmc9}Q`^(X;Ry22!huK}G*(V@bPaty41bqI3Lf7ie5O}%Z5)AYXfCs z;~HNkN5=w$zNWMhghG)mP>rdFV>?o~no0YD1491hb8nA>;LZyK|e z5AJJcq<;)5SY#_%RF&OB%HUE9~T+%AAZD^v~!qdMID2O8b`)5Huy8oKazK0&z5@i6b{K)<7kt0b(glSG+ zhJVJA<<3bKLF{5(K74b_EnLt_2!1su1G34Zz_p7TGAYcBNxsnfEG4-D!Unm7?ua6v z-E@J5j!c5z6Fvrr)s#A_M+=ySwFgYW7ETNMiQP{z5-RH|tGF5?Imhe8Q|UHd)}5>2 zSCmWV=kezbg{>H+CJ>&d*%h`iRM~a(T-B4ZIxan`&?+$WmgV@lc9ese#9iaqO%h%Z zSU~4rIRmYfO9IJc1wyfpRWvZO)y-eDsA7^_D|J7TOcv zilKbC+l}#I7!*T7oi&%b^W?PzDv5fS$oKz8**isN)^*Xkv2ELS-q^P7ij7L%*tRRS zZB}emY}>3@yY{)=-^G7gTUYCPwK@A5eLTG@*eqR7C2RGlq1^^H%slc?y*Su(`*}V# z{t)g*XBdTQ6}9mJhQBOLf(fk}x~7bZnC{=aG0FdQVZ6_kTdV1QmqY7MUAt&d;){vP z4U~xGN9F6R)fT6pC2S5LBLd^$lKoh@(^ejwakcNwo;=UcYA?;@~pIMyWY)#?H{4)~zsn4i3*-tNfbFU~I z+x@-lxIdXk7FOhW5!+5K{a?eI3HyeFC*X0={nxb=NRhM6jq+W9vyQkgZ=CV^MtgP} zF+VZ=ki&~iNc8qh*wZ0{9K-I9uCd^O?P%8xi-N+nWoK8?q?u~%m_65<+bh(M=vNp_ zO6Q6P-rO_9-!dONJ*oah^7%_kYFFY`QqEKzW!hPG1;KT9b@qQws0!`aH4&)@Cd*gF zbHC9*2S4|~FFsF+#z1Oa790v4^B{mZ=bZYMMWkal0y*`wtVuN3gswadoi)@NbU`Sr0Z`Vx3@m5-3_ z5!ZAN4Oo061isW!k7sk1(6@jEh_pGHVaplzMH;MP(LY2HicVlmPWx=KQ3)}5cD+6+B=g>?#x zl+DSC?*gm7QJ7NBVFP>ySe@xozAheh+6c*a45W{t?Je3fGD zRTwAFZiUqh1A$RZ4);@x(i(Pq^CiXjW;+N9N}=G9=8e!MFEO!?Pvaqme>diL4X=fx z^Wmm6!H)-kvL~p$b_^B|=h7}>ItARrl=JTg*olG77=oYhT_B}d{`{LCzC{59HH&nh zzVdL+2s>n)Rabk*&q75@NKhhMEVNVS^3~xQXcmx)Ra!zjA8rqbgyF+3k!%HuXi2GM zkjxt?&=(uNb^}(64R-FPnDq%b<@`ilPwi098W=d@_+hnP{D>HJv(SzOdL%*f%2_|o za&oTP`V{U$MG7$2zd=Y#$b(6TIW~>|D8R{D3_^*c=a@x2FkTK4nj5zxkdvggtD*8l z3osUl6a4k8PHcI#UDz;NOO#i#TG9|QMksNiB5m$&NXlf9kzyZkr={|RTA|{wXa*vO z6_?J4xkUQN!pHo$h)VU%Sg3$hHml6?gDPn#DTS@3YaRJr zJ?9%yZ;f=v60v*GP+LuWl=+t0{|y@(9BX2Pv}yX9jhl9hC=GhO+i)(^Q@it!&YgXj zAx@vLI=s{&u*e0$@jhJ92 z1T$-{oS?RVm9rx>N5m1lu^(LAc5%6y$b{7=RgT8cRHtkI*jb+H87x-~n>YC@+JpW`gq|g@sfp=^FW|hA|8V4 z5ml{(A0oi*mk@kUgVkvM-m_@kGKv~dEVvLQ#3i%jp}nQ6bix!#$5x}WQLFG`8m|_J z!my8P!!J36Tp}aBKUXB_PLNR#7awGT;qv~1@Yt@&0?6;L+FNb#`UB;X7?7KMemwD7 zCEL)@=ksc}f0`%>XPcWo_0$nnb>bOjj0HRSW!n3K9o?3LZ^TuE)Qq*3bcB<4xsi(8 z3ZS>>hT2ODHYMeP;5+0wO45a?|oR>}jAuuS$^E z6t~wXqm9BY{iu7z^14>-?)z|UY04qyo(zkB{Z`brqI5tf!gRhgH@uYa$uVx<9jOpX z)F^;8nbl+d0b*=rz<_!{?x1d@i6c)C&*wgJ5S3HGlv;=pd14d06g^ck6~d^Asb^14 zQDIj&c`5a8>RY#~wF>LUt=i5LdTb6>*2ocA4k1TUl31}M0V%oE5LGvines!-(Ke&A zY(K;NWtLXc>0Q$(wwfKqeN2|eR)=6bDJF8!#|<)IAgJ;DXxXGt*?cYZi+wE4peN?e zeF`#kz#90tLCBY06v?O!L!rj^ad`N`IEx?R;mhMeraxS$znZohvL z&58t8@M1*B6|u~pHEzG=?$pwV`oExKUB3}W)dr3_7g|Y8`HeCu^NKdC-Fr2n4mm$E?!N>lmxFvPrjCB9J9DUg ztK1>w`%pmJ!LANwt60SQ~`?q>HVRZc|wKN-jIimT;|qI(fd32kVCrDvIK zgQ;CWLY5@@G-xeQ<$r$W+{jwYD_6|b38hWb_Zq%O*%$wbGP%s%vd`>NT~foRo0-;@ zN#V?(LC&T3L>=WEIq}J0`!cCdv^StuR=yaf6ta3)$mIIy(}fU#DLV*07XvZo%(U$H z$150D1M#5Ql6N&4a!SOhO~974%em7?`8nt?d4HR>bhBkh!@Wlo<~}3GWzC%H^!Ec6 znwa3J?$hUrc}MysNvJnio}cYgSj|JTlXKImWGx%%vUCPgyaABR`a>{63DB#{7l9Mq z#iw{X&;)E|3V$bi+Z|l*@e)jVu6!Yp>SOmtsA|IoR_gJ)@-FwEupeqn7|^^DGQ?V6 zfFt+#|Jixr`*QgjOw;?JOJ>*^e7Ie(a^vX#{^;Jb^D;esdp_45NxzpIcwImC9#V1g zh1I#_w7#ZSH=dC3_0iG0wRSSI{)N7|9sa3x9eK%yH)ci<$Blq>d1e3DonEm(|0m0W ztiZzztGidA8O17(W&{|l)n)`9Ioju`sG2|SpH1iNFY`$}Vbs6%A6oRkf}q(s|7S-(K*)lTDb)S350pk?{*F|0Q7l`DC+9_Cy7bkGjR=M@($dYuDu z=bPpkwZSJI1!xU8_+v(ykZ)$gT>{W^72#3QgP+M<#NaxfpV#IAcqs#w4%3u#BHPol zka|6S5z2&xqd0|FI&)mQEs2}q^$$EtiefOjy5d3tuM@JtDIcJ(v3E@F+sN9e0o#(_ zxD|gViDvqQz7KJRNHU-h?aAuh+WGJ-3Lvsuq*Cu7Hh-mJ6zzpMnC|I$J!&&{4Ukcu zh>!Al_($zf-Loe!uBvC{IevLQ3|>UQ4$f*h%iEgQV_i)`xzZ4p_z{$=@RP3#W__K` zQw13#BqwHx4s?{?7x5FU2@Yhc9H&laU<;-Aju(|RvLDsXHz}-*D9Vil=24y*ZTjf- z1gK(MIm66jPbzzJ;YbET_oDL09VH*5zT!h;`$yL$wp$RViXZlRbmtuUC17N6On%~Y zfBAdzY4{;A^|y#>Ul2Mn+dX8Yvi|006w1gR@kPPbej z$PS`NF&&VUOE++Q64P*BT02qN-=!0P7gDA@q{iVHds%qj;C~hthRNLF6nygC$hpz? zwg@yTPPTeJ++_(RsfNN0QvA#CdUvAC^_w&_F`@(}d=0R}<>k%N;=lkQsl(bpZJWpf z3V$G`JX-LqHi+Qf#|GH2P$Jg;OdM1n(hs98l}_;DTvEWHLaUT!$Pk`XE^nn6=;ttJ zW@r-60%Y}Y#EPUQ6hR9Bv0{ZWG(+GR=jCoGZn!65=2cJW!2sv?G;t7{MV@3lSe{xU zosWYiQ=*&MWB;n3bNIuYWg3%ASWRVMPJeh@ox`#|MXnv8{>NDeF&pIW?Q+{WKw}bib^UYFb$>5D}E}Z|&R!3${b* z_UBlSa%*9_nmr4Kv67lAJxLV9;3#u){*-lYQT4_fj%8UD@< z2pjp^r^*K^cRzt<7l@5&m1;Qltpk&W~c<=pb?`!`{yWS*7+i1$nboWt=l zJgx*IPEY8NMO6R9MbxzF0Aic;184uKcZ#8qyY7Y~irA;e8}$A@B4=sH zw^1|Adni`p@;?&Xr$l_g{qxS-*o5tZ$eF;K6GMeS;^Gmz-Qp2@dlqVDASKK%*U~H; zjJ@L!wEUNn0m2nc5JIb^i2GuB=(;GkBYu+MN>F{YE|gT_w?*$VyFjEOW0 z#x7^GRMv=mP;Sx8My$(#E8nbvixs_)n<;af+BE|su6q+tUX3vC1DHnkcyBbAojpDd zH8uPI1pc+Qdw#UIP*gW7ogKDskquwVA<;qxMedoRu1L>qhNL20GirUt&cOch8>8b0-pQ)>MhR_xWUU2+8Q#aY@)+m!R{Va%XmUw45XL5qA4#xGa1MQ-c2=j{-F z-X|twt6yq9;Q+@1LJH~jeabs%r1cjjzU(99T*PJ2>91`khu6pw@z@_-`M}s~)lp+8 z)ygb7){E~` zoVnR1tBS+!LT>V82+&H`oR3I)2rG{ge!EFmB+gzey-;-Xm%Eb7t)1!CypmF(@J7xR ziXSB1qRLe$Y?Dk#gGlJKH=`R2B3zjqv5>n%p{L|tA*Qxb0=qbm;d>5>4kz-vxnWWW z*$IEod4rdkC97t&{Kr#q`aQpL>B)tZ@48R|Y`#fSKT^f8o*}<=3~8i+crr#SzG`T1 z5-7EX+?-gM5(*xQ&T^p&a+Q zvK>P>9u533`Nxwy@ zfKpNYG}aePbxK1qPIG#dfAvB%WPZ7F-bII=>L`Rii_a;*L#~fh(zl-Q0GeJy;%DIJ^VPUu0Lj`@|h=yiGFAU3d)-MWQzTU;V}PA34;l?A%=ctNZ?2>!1GOLg~4wV{uW1lS6&1FJ|yLEo7X< zqVbYeS~Kn!Z=Tk~LQHh8mhktn1f(!fHtsP-8B(}d03N1;#qJjd!@KhKO-Hnqi>*1l zwWseXq!)l`=GG_{cb5g(0$wGXQMz&(7w;Qo;YJn=jckH)K@>U(@+Ie#@ItzNFJZ&e zY?&5UcmvYVKI_{)l69a?W~HSNIflIyHjVHFiyfrLx~o2eIDVEK>#?lQIWR$8`IR+u z+=IZ&KI-6y2PcKZow4v!qiT;5V`&%AbTrPN_<7)6I@+k z0WiIR>Qy6lwL@-h5yuOj5&A@LQHdUsiCCFEN z#4`m}y^x`Z2e43XjdrW!9{_kPYr%&iei}AOC?YP^oi`O!Ps#j&*{Vzarzb*R@Oqi8 zcGhDNZH$C1l`7!J^;u%5QvBHh2E+)KNaHE?V<~r#9g(=%;Ur_e@J&LCty8IBI70A|*9kK?(D{hlaJS`@{kToHl_+z6hE?whQw(_y{kL zp_uM|@$FkIe+gf75ZNU-SfM4^iOb31aq;UeAAZ*_7&5uQj#k>g<1 z<><)rRTQk93Jhk(<8B-_+6A=n8oKMGshr8ou`g4?25S2x!CX0J(nAS8;+iF;kn zk>$x$iUkTo$DW#R=S@PZ1A_xnG+w5dR(V$j_+d9&)SpN6)A{)nR6N7-k~SDRH&S zYJbsz`K8w(q^LbJ>MAlAqHPpvVEy({w_UffvcixcN~IZdgmY3rEU@-bmU9C=e^YNG zavg3Vo0p>XBJG9i>2`%g6!UF@5MBxYNHv^_<|j+xwoP@OzwI2hfK`G0A|%PPb_Ddl zqQl&EIJe9x^)qQZ(J9EFl4jThVe`6{;!d!%M|{9C(F)^q4B0gvldE_?Bim}okf5Gr zy>WYqNhg_D#U#G6la$DU^#101$2BkjxIZI(w>;2dfG>+(Me%I7)ZvNV{JZ0GNVVwj&|I$VH>iup&F;+n4PcK{g8@LcDHS*A4N&Q&Frw|W2WrPPwd z@fBB%ASR}2X@QMt?mD@==KFeD`w@=%< zc7{^okSn&=8^fSV6duL!Fi5i+eScQpMRQ zXbnIXNbvFnnb=mJA9DnmuX)tyap)g`ck_ zpqRn)NTJP_HJ%6X9P{Dj`>^k;4iMSJltJUYh*kSxxLc-}Mwy!@G1KWH*pXFg{^^n$ z61h?T)s**AwgIgjM4V={EW|5vN_s~Z@f)BR32|%4RkP`rYBzd=<%IEBun%8&OH58& ze7dw%xW5tS`l!)rV1&*k{VSGIUm_!&s-p=`jhPnBw4t`6KgSB4$}oH6HrLlO&IHF> zCR2xks!B((0lRd#uOfBl#O)7ne-EcVqn7h^SG68h&*T}&MlP4AW}!I91rJlI2$%S0 zx1k78khkjlJhAIQfZhj*SDLQB;Sstn2BU42f3sj}0!c^X1V8Oh6(A{KM)ohsVzvJP z%d@G|C52=)YkCz8vwn<7$(6)kL`>Oan><>YcTtztXX(DYpYM$k zlFnvuYk=o5(NB_GLoXGiXulCV3q40zbJ%*4sJ(Ejja}KT2{qR|H}9k<~pS-OUcNm42%TkR6^pKfb)DNvYTPp$2!}{(XYevv;8(y z=h__QWQ-od@J*-MuqhHLQQwETMRYe^o5En{SX&|x`$){oGX(8DIR*xP=OG%5OWFAT zmvT|tu9i@^Y5~b9TpLDb$iL@hy*IRq6elheaWuV&Cs&kkEL^mMiz(U+eerM|z!Q!e zfMhd<35Y~AX`T;=7hN)gAgl;bcGhq=or;*dUqd$CR!{TlfFt2|tFEc`!Eh-(ZYn(? zJf*l!RMNDH-sR1UfKa1eddL~jo>B7vhk4VLUSikLE!dUBhSP4zw>MgdoO7wg*bo~; zi07^be}}VxY0D{}VpnJV%kJ8tmWyWvo>QTc@addVcxP^i6EEq2R-pFNPEEV;`~xN3%Uw4<%PwLLZlr z9P%y#&X6!p@}_6#aw)^R?Npl*X(B@NVcThYWgteOipPT7H=;k3jy(=93uMO=3o_yX zhZuoao=)?a(wU3%qw0uGK7a^;Qg#B5|G8T~EJIIDrLgNR3a{0A85lTM1$s>L+TM$c zlh%*vG{0PZ+E`rG+C&M$+|UpTdpN-6q_?g~u$mAla$nqjx-44epPfMxtJa3fH@D#o z267d4O?k7^?dzDJr;8hylAyO`uy;Qg&&a`#zeP_@9;bc z0ar%IVyUmmPC@BK0J6WQqKxm#BPTB;`KuPWUUPSoWS@}21x5^{fRiAT8Ouz<(jXJW z(};XtCh-KNa7Yu3iY-!ZOnz#WLRi(0EEA1gymt$vkg`h0GUNxkf&A&1J9TqW5liMO zQ_P>%t`hbo-M)huuxLd!S1WO#kx6n#1pZL2u9i{7?~#2<1g>3~eUUqmSwc(I*$9_2 z7(zS~8)tTbmzgI||MguV)g8$}j9(S&E)w#DSane0nWe}qhXOl$(PNXzCKjFxLDO_` zaR@Fk%KEdmI0fz853CLq>|?ni9VPg1Vw&XxD! zidoooz`>{U7({R$K$dVya(feLRad~}nQ8;!BFdzt>-CC1VLO9T8wc8jwn#`O@^9k! z-9}s~K`)y5hsI4&Q()7-ztamVUC3}JN^V3z@)roeX!41qKw{_f=|PED&`-P9o9`NQ z?qhTgonH!5pRgeyo;btWYdzo>5^(pqn=ivc8eaZAr*nOG(Rp*_ubhhS-Bn(;3aP~n z*2U{L6o}v(G3eoG#dQ;)MQ!!>uypR+WwmdBQ<_FKrjd_-2cp<+h=1bkT)&-21EGBPah`F$aX zY|HhBue{w5A%Qig#gD-`b4K~nq@<0JVf2>5X1MO z7c2lTly`zF8+dzA+NgUMxQM;hdMQq=sVQI+D?FJd)l<{0a;?Ocia;7;0>;#xs{>j& zmh;l96p_*YJ@7NQd}0w{&0+usdj~<-rA4H9c6R65G~u_$$0#F?;Y!Vz$S9lmwzf3a zNugNP%C#95Ki%{gp#9WLFBllyt}jj=Qh3J)(tt+S7!(j}e>@;fPj0r zmq2=EZVkP`g%kR)z!nEDt`k=g$J{^=?=&C9czDDAw)lfd$skBg7lE?4KD$HMB>{-E@h&s{rNI z#sn^-L*IJdx}0(}CX>{;M_-%2C|5+pkCE;1?;ixX8$a8DNQbY3)5 z)antXao4Yg%&jj}=e?3piIUw#diSfs@N_7YjY`Yo8o@g5MKmtmiwzl}Un%Ws*_3$t zi0|#imtd0v1M}UJDIN0sxwNVDn_`PPhh3~qRklBOZMt;*T?*9ZypOM-AmU}2L#z!l zhUMe-N(q4gH8Qbue#o0EmXxUno&rf(Az-xwaNV+InZ!(uqvRj6%>W#xX-@B%s@2TKRyecdaf3R8YnT14 z#it%aY^Vpt(nZsvtd}FT2}tlPWFcSKn_7BMZ%+05)ijP^G8gt93;xS*Cix`(=`83= z=Yl*{J+2DfSweA07?$m(rEzp(i8dWtk#LO;1rx(G#kgxPJAB%W1{ylm;|R{~_!z&% z4VuZmR~MrmqC(zmMMZ~gtK7c5p5#p5#NYWj{?`FU1!KlY%U@7LC2K)Z70uQz$>7^S zg8nbCv<45-IPn;|As1H%eE}1)7R!AywZN;)NaF3^g)kHh$xOd40^~M+F2oAOmlYP3 z!bImP5J4?4CbK4JBb95AQxmEXl>I=WFaU?ubI0(#Xw7P14B~+r>cJ~(%=$CwP@YdM zp&t9{*l83v5=`Wu_oq2rryXtEtRR>fA~#NuKq$3yS5O=nKD=7FA=;ydMt5Wt}B1<$C(pTRYJ0*w6v5tQ^l|lP0C>-H?a# zu%KH+`OUpBXTO6EM_8;jnphFCUiwO6I?c|BhB`d1==8sWH-E2y49MwJb;}<4=FE=#XS?Ajb7U6d0lHPk7~aT*aW@o#?kM8k6bJ*FaO&0Fw+t{jzCmavQI7)15mY z{4T-|InMo}l{7F`UsiY^;T)Dr*948+Bs0XCviqWh?n{OY_RiK)G#>|y{ph9+iNoKPsr2Wf9st5|F7iW z{ttwS74Yxv|B5i-j5};{{k^WkIV=X{XKRTqh7hEXaG~*{HEkA7Cdrp-4i8O>9J`yZ zyU~+oG+RSjpN#PF@xkBWC9a)e5~Mt$iyB~?++SLnuACZ_w$t*=4JA|3)LJ^1b5iq| zujSUshI}_GqPHjRjglwuGM{SDIvJ6z>cTRf>w0XfGSX5ZAnQ;_3A64rC;K|qvzBjL zn6I4VZlTumxWezD%k6s1pFYa!l+)O%)m z8zsC?jU3b1E%o#6eS;T0u>0x^{5Xwq5E+8x7Rg^3kHQ!iy#&<|uO~3A{fvNb%OTs6 zr@+D;0>4fEkziJ&#^G=5+(r`!HiGsrpW_j#jH~wdCpnfMG-be-v9e(I4f@Q*h z0#P}i4l$`Tih76Zq?D0Fil97lwSzPzpo854%w377gR(4y{OA zXKOBsoz)~N49gWLpAhbOth(2iHt5-xG2xN$skoG`;-1t67@T2Gzv>^^{+F*vIok7& zH?+1o9DLSo*~!>@#xgx~$;x$tq)6x0@V0&aQ}6Fa1{NvvlSjplWhLlZNiI}lTI6y5 zEksTtBU@5B}d-;8hTxzCz!+Gpuw$^vZp zYYGe@GjnlIZwRiwf~5Fk%E8O0d7f+WtVnAte&@q4(T)=b)3u&Dci89^MfgkQ(K3}O zYkFBrnP5DapzEU?zBWR?kh<+c74ixAKjrU1J|dcENu-Z z7@S58#oztV8$qTp9xUit5a)~)qXo-WVm8dMkVohR?XB?guvDeC%cR0k9G&K#FRAx#E46_{KoA zmp#{%`M~+ri3n;;`a|hxlf|(%RGcb=CbqktNTwvo)o16s61XUk+VORoaACjjXksaI zNe=nzmo1OycA9Kg+zW@FaUaS~*7StxZEY>!DnDo$6pSx8WbRn8oOo;x6F4W5y$Zal z*76#dnb1!A&J-F5D>46~NMu6i)$?6~fGVC}6Irn_*=h|1u3c#q;YfFa#(=|;qSFVa z=%hfd-^23FHPpUD#qZsfx9mFgXbEGu3Gt`Jjr+Fg$VxnHiyaDHn;hk!du;3)n)OQn z%3!qhi5rGRHDpvEOp+{pu$%$;U-J1m-_C8U25L8{KbRu-K1*hxcNYWg)9|66(%PlL zU59e9wdQXj$p{@UhbKvx<%s zY(g=R6-tPaWaoiuU8>cysK+{U@>Bi#NuRX)iARqxML`nWqxMMtCP?jhP`hg^$_jH) zksN8|#-{!JkbUiz1xJPLEm&0gfQnC+w^99zqtIt6d6Pftbe$t|cX%GzV%F&U%*vH& zL}P<>x#AMhsS2ttkT!aYtV|do{fpzJykZV|Ba$^pf`=W4AQme4>|pnPuwWG1C*05Q zU*0FO#E&RXVLPe}csMvNpjaBJ1CGk4ArXR^x}A!0KkBYJt$q{)2=9n$v8%CJMq;f` z1vEGQm4(@e+a-qIC5Bkx+|`^yuhWAG0iY3+Vg@jwT20c&uQ13T|I3qMIg19&@ncqG zK7k`hq_SJjG1=IP`IkHgqqs0tTcy@Fxq)g58UTj|?`6sFvcliWM zbSU*&|3-oI+p-SEC~js!n@Cq&-)a^m#%Zz*`(!rjv0hR{^~K2Zl)|u{kXct%H;4Dr z=3V1p1;%o-W(WJ(?Trdjd(W>fgmApO$mk^bj3(OgL*UDh6%0gCi5*9j8P1@HqtY;~ zfA~#7_CE%|b`Y>E(UG26FmEzmS;#{)7zrUxRxHzoD@R=aqh((^#5~iC4;@RQe?lVk zM)n($P;|8LKW^v$%02x5(b81&?{+?reERCaC}{~2wmCVk5;svV3_Lq6Cz#CDa)mVZo+9YByWi(SM@10%#4mQPY@>sng_luwqpT;lCF!cI z!Cl*;9V8<-D_Az34D}PTf0iuIw+^=h!5ssDB#QWZ`PFW}xCui#vwHYM(vc;6$^%Kp zBi2`RG%yJr&k;U@s>CKKcyEol0Q&V=5?z>|NFcIM7(`RJ(yhLG=v#M7$rVnK^R{Oc zVZNz9l^i+_P1T%a<}hR8Rjw#{EbOpAv4G<0Bk1kq!BVOUX2y$l)Y{05{yE8goZ$#g zRiS^=N`#sp*>M^ivj}yE3gL}WSuruKr|>of)n>XMg{w@Z4c&)9e?#VS&O`JjRnd^& z;<*1XOG!A*rvfUssGZ(tj1~Y3n1$J#6d@aXUm1|yO;vF!McYZRXD(q(VAp$?|8d;K zDV_((d%qCr55-h81O{#BP-S+N$EYF(T(A%XntvF_wBDaB=Dz6#$n6c>*elW^5d`js zE;<0E=Glv@8c$}+4w-+w;;?>*#3nc0p)@lEgJhbeTXYn{zkr~V7Cz@_U@)Mn#R!Dum_R8o^bJF z1J--*K_}#4c*M+3cl^s9pzVIMEDr>u89Zr8GH4}*xM?Dv6+3;E@?jVfeQ8}MqSId| zyzU-$Yrya~_{7}1oz)MTWY~BMn5-CUv@Z9Ee2*)XKz(s-a+Y`v>)Q>k#K5jtnY7w5>Hlfxsx)U*O#(F+A)KZe2LQ+oSpLU)hlaK2puMK7I#S({WA3=ax z`uP3b{SJCR?0KK(IBCWoyb7e6#3EKj1s>p_T$b{vyLi_A zQj18pIKzt|Y|>9iX8ncGmokc?CO`fGk!U1Z(?6>RAmLIMI3!}ynBf)ru``z5dkE?w z)sd*K3M7vykodAa_nv~fq8j-~W)z$zn{5gWdHaY}jJX;t9Y|%}#wmkXR9$I)ydOED zm@G!w&kort3lbEf2blRmvSPM5aPb<^8qzSwCqPD{Ado(>=$2Z&{ z20L+KnOpTYzp|MDhJ2jPdO~qef`R^^%sPqt+VOeL)jO{kiU@zxX`=$xue$4*(e zyoS8#jAr?ya|*$+p?!X|{1v;YnL7rC6*LQvsLnouJLaCLt|1u0rUtN%Qlv)(vpUm= zH?jiyLEiGP!U0j5!zTS`_yz>{g_H$*({>Mmo@CWjHYs!F3~utLpzGDDg*HZKTgFZd zBGVl`FI(t?2~G_x6Bvgn`F!r~zp(toJt!$Kbi|rX3$Q!odav1H)Yw;Y+y+aL*I(5` zdGQCm;SVt4w*?kZfpP1#GEgX9hYOgG&J0?Mz@9hImA?mG8*|-3TPP45o5zqd1~OVRha< z2mdAn=|c^t8Da~WTEHT(>XTrYKqbn>x)Yy6>{q(Z0)Ga{$1_XCU@X#n6IaqGOW5|( zod?k+=WmB0S;Ib}M6o>o9QhcQNWBc)T@;c*0)n10*2Vhc-l&wZuAfuD| z6WwnvEeVkz6FjV5ok}p5a^xenj)$dRXE0cOSSblO9^34F1=1UD-vINl^~JKfcdLUJ zDm|)or8=P>L7f3%+lPZDV$9GX)?Vk1GvqOb*mG~Bf3Bw(kty>8@7j2lkhSNXr)L@f zm+s(|0~@1AhW!$*K-UBLV=~Xrq0GacNRdSQUZCIHz!O^ zmWgfppDNE=Ly4nxB_WLdTR;V`wyo^h@Ig}IogX`LzmT52xiR6GRhPJwc5YSNJ+|^X zM=zD9bb!#ZJBv~0*_d4%U1T%wcbEl!I$NvJm73tft<|Nj-}mQe@zie^bU=!-3wj+ z8&mo%;F8((3WvL{73FS*Rl8{fuTceeKuNPCFA(@h&u-z=Q=~=XM5>BcROWoVdcwD) z9*KbStPUUD-Ccd;`uV;7ZSx-*Ho$ocMVZ^0JgXi1V3EFQlz$Py+WD>SVZ?8{Zd=m1 z_d)Oy2=oC&NCN|jnFfKl77HwYi_8IaZ#_4XksuD}A}D=_4})##{@aiq;VgH_6i1Ws zS5|cI?~-9s?LMgB7IXa3#;q{qL}K0GYeAZRx)^%L?f7rezxG)(&2?p>1_Yvv6Vpcv z=%jzto3ys>k;aG#ArEbl()o~X8VCsx>;vw5piJgB0;$*4&-suy(7#5A3{gi1_*w** zt<0dUC!kS*6IZ{_qZ@tXm!;PG{WT`3nHq`8d`GG7!#WHJT-g( zQ$yeTG`j)|FW!QlB8DoOK+LTw-nV#J37wu0_=S{~<}{C)`TI=kSC7jVt3U6>aBuJI z`ou#dOZbinl0M>Wa}sxPKtPBg%IzVA?XA-8(lwraSei#zQ!Bl*&q=G(RVxd)rAa`* z0vz^RFCHLuG}j)#nT_g|%Foc-;HOVA*VqDTcW>@~lh(p+c{ouhq|QgU7@27!({*4&LL&T_ z>hsa6@+|%Pn9BGUxvn;J4n1~vrCrt_l_VtmHVqBFFb1WfoAidH7PK;M$se7!?BZ}n z+1^ktQt%TiV*tprwJHqE+HIUwbLGCdwc|`LcXBT_XTpph>W^f9quO7CGS}nhWYZ)@ zXS%+d4=`;0BkVGB=y@M8M?km0)9gWv0KyB4{Ia9*(OYIT! z2H?*)zUz6V;bey8v|X^2Pl&7Snc$u@VDMish0Q--?%t;)U7q<7y!)@UfRW*zs;Om2jp%~=7?_O!k1^lA zGc)J@J8bM^I2PMjG4$8V@)e8KLuXl)_!?R{s+;Qak@loW8`Ki4i}K0*GB&@QIDVeC z;X!Bc(0q+Lo<0#h5*WLcDGcl{$%6FBGN*jpvj-BrQ6$=k*V2zzI#zdi37;- z&eMax%hrML<;E^V2%4z`WwkJc3?|nDfn3c;J9AT@BY$%Au4@VQQcIfFT)>atdBuON zDK@Y$rh&5YNYTW;j?$SuHfZweZJy8bz%HL5_>k{KGyLf34K#)vMiJESTpq-S29*?8 zH2fDIeT&!qQbaq)EiZB$v=paJ^PzRFi0nd|h`4H6uT)!W=cJF*Z-(WFDqVuV2sv;T zNt#%s8;>#`pODFdD8wvPR6`!{ zCF8g9^!jysz%C-t)*Fhz#&4vfAbkWu-#Q|pk_oG{T}@G zryLYW!+r)9KYbba-phh23&0Zyz*2C9;VzyVEj^E6UxKF6YaY_Hs4 zDrm>vV0lGRXn1I|%f0;oBtzeDjrEJi$lTvWWk$Eh#FoAqcRyvF;@@}_A$g(ft$|{tS>qRDY=xV%I=LA4~F3Z^;_@{^hOCZWXRum>PD$X!yzod8s%uW*E z;cq1O%UOU3dCT4$VyyxD3773)h9Me3IeejauOywr$(C zZQHhO+qP|X>GnL|d6=I#NzP8za+cc}nY?dJA{n&ncet07FG@`$f4>57YfZM%EhB{j zPi(9CsXGi=%^AM*4|fXPw7C54JjuKre6bF^qKL?AcxT=V=M-78sp`Fyz}9vB1_v_q zH8L1sqz+XKQcunx98?y4xo>@}sI7A0Fu>EcAg&tTT8)kQ=M^s9*RFTF&JHVb)4UD9 zkYUK_!k;U{2VbKSz2%`G!iF+~+L2>+f&D%f7(So`a5y6caIjDiJg$-jB1L=X51)Y~ zmz4qKelJ5q^TbUHUywjAIvoWy0A(TX(qGhE?G;vzq+WXWVLl=h&#Weq!Gu1%TV$zK z`=}D+)bcdUMmD;p#JMxJ0&Pale@J7BuLlhfT!He2G#g*-qCI?u3h%k3urLU`q!CA_ zK0GzqdUFqOiOH>4{bb@b7BnGiQiW5Wo4`A6vtgyRi4I}sgzwf%rg#|b8N;wGvMuii zD>_9~gTy93cc8Ja!kNKFVYkr4w~nDjTrb3wNGeb7{c}LSpXWh!jOVRCk#7_2k5i_658fVecRihV&lNIwK>4zH+J6?=i! zrimh{;F`Iy_TdRZR$Oem*5sFyrjg4M!Cm=PjQ%6Bn2`F^&V&$(z0yJ9aqi;=0w=Xd z>D#}*@BZ8-`uqksYJo5CJX>C_ONQ4!jUL%4X)X??9xHRz)r{7wy&u++hc)!fDen%$ z+S|dE2j{bT5xwVGWa0TCYNx{=0$n5PXS$PjUHUlId=Ybt@VbbJ9x(_Vo%6OO7}w%j zu2K6TtZD=K;nhg&Fp53z`kp2HXva=B-U}!{Qq!Q~zfzby_^t9+CSr=K>Axe9d_gE*E;qRCY6CViI1RUb>-0+GzEnX%j zuwv`tn9)xTuaPb5P8vG6mH`C6`J5c$d!smGedYZTdu|M5ao*3*5700#4Q~!cGoYKmp?1X6&}-x(1f{-gUfmyr})skhoxo7v5Ib! zwM~S6KdD(dj$SQvNtKq9#wsOqTutg06=z#VMj5uX;$XkCt5+@!Tj(B%Ydl;GnwV!F zxKX5g%57@T-(X^A7H=Xcs~1%2?G$hyD+o9o2b>popj_msjo?9fCRRD_=)N=$E=gr9 z>`!xeX5HryWb;3OW03M>+IPYS!&Pqr27335aIq7La6Va{-aN%T50>J;vT32Ku-_Te znK1$Mo#{^5A)XRY>`LE>R}$%e&jn{o^t+Gl)7%#yhGzTT2iv-h(7ur-M`sL4`uje@ z;GLT(5+hYT4(i8~jD@N??a$kvw@~?&k1sqgVPy*sQ>U#}rXJAL+?|ECRnPf06Etr!|U~J?@6Ny1lNq z8+9{F1KSnqE88X#R7;67QxqA3lS1$*yx6gQ1nh2v;GWc@k47hHB@HWIUDJSz_84+G z(Iro(-o@-A#T(@z=fKs85bpZ|NubhC3sO+aYhR@W$O+M2Mj6BiLy4&fn4t+?RS8Ww zU71G3o`;eFs$srAt^}(*xESUUnO!kwcfN_d2|;k#<}RpoNjPls_6hU>zf@HAEqi5U zVqQig1r|xGmTb!zFn6jW{(YEbW^raP246Zuq_E@m5w1Q_YOLYI<#2A`zRssFl{uY!()`H^bI>R82cAyr=wLlZaOunW8+d zfz{UV28?`_xQB~gsA1So1Nypal}MSlu>!(NmtWXTmObYM#vaDj18N&lr(jRxwL>iP zq^$o!Fr2?jL$H_K$a0H$K&@Ft)U#2 z7iHL>C<+H68T3YfqE+?Jcf*$Ki3YCM7h*Ht6Bhe>2Zj@45C@G30BXletF-bct zx^^6`BX<82q2O48!7fQIiAb`~W1)u^xk8S2Ps`+@$(9u?o-Z5F@G2MKrpT;WGCyFzwM9lb;4zCa52-Y{jv)WTg&shh4Q-fBZ@+*lHoY@7NeK}4reyOF8%TqO z*RND+zNWb4SG4HPVF`$fJ>w0-)l3EDZF z|4DLrxCLhRkV^Il3&<2sQiz1Z?qKk%QAmZwMi)H@K<@zvv7&aA4}Krq7OTcoSbCIa&=1Q67aeu@1#Y0 zwS@p!KpHP7bZ{NDXkfIeBouBGnNw3tHW`_iuz9%>Y&6#flqvlR@Dq>YYPkGH9XP!3 zV0jpkNw;tqKiOh`!pEtaZgNKdb11IjP$Ht3iubvcGt&7k0TAi;@H(lhE%(sGm~ zLPAI+Nq+=+ykALs_E5k{^ry{s3b3Q2h+dVa@Cv!CHPBdG`3IUE;m>3COU4ta#Yd8v zHCJKP)*W&izm>%S_rZarQcLMnoQQhsn?P4BHRn_IH&KAGI&DE`*aKTz4Yp) z{)M#Iqj*!ynq1^^?&I?1-mErk6>un?VCN0C#qULb(j#j;OQ*!^g@M?Sh$I76zpR#_ z#ZK~rH3yhJH+@wDh2^o}Pnrij%{hjy|5B^|ODw{~^1rk2ORy$vum8hcHRw))S&t-E z)N+J4l`-{)bd0jgz&4L?Ws;MMLu-rq^0q8WZT;Gn0%GQG6i1ubtXj^)^L-zpqwI}1 zs+V3mK0UY~m9u^1hAq+Hf+NPXPn1ZV?rH6Wziy{c9kB}#LZxim97E0d z;vNAFe@>18*UG?uz>eWDoen_cfCnIKO>8=IBmzK6NViKfJH$+L5NMK~N~i`e?68ew zV~(B{;KRdKTX|UE^(!GnAR_h2@lxadw>*feD1c;ZjX^G`=Jz~_PT1#cd>hnge zV>@e_uFb+`n*X*<2fr!5-OY&tpHroedth9HiU`T-A8W#e6J(1CJY{|)s#%w!fL<`^ z*f%e|w2ylx5&tH)Zg~mx!pQtiVA8{89UGQY|lW*On- zE(1JD28HE>+h#RQbzzES(F}=-9g=hf#olm~(p_U{ON1{mD6}Z1f$?wEdLe~rmB$*k zZ?A5v$1ixS5Q4;{(}x4Ds+R}eR=ccWBJdxV4bi`O=(T}iC+%|~@Hzl!6{b7C`%Q=^ zaYr&pk1Ld^pdl>$Ha1w+gfseRk?(wIp{R*IG`9L&+KaqL0=LqgU>Yy(o}X*Rj$N15v>?n(G zYBJU)-dsC6EjZG{Wd5An_Twj%H(u^RXJtkJ#Y=&C6}GeRN!5PltvEI^j;X?d8jhtF zVr^>?=-T}3O+ku+LGneNA<)vF5J^3ctW}p^@L6i zZRIln*Ahm|BXqV7sCPEPD5sR)w_d&Zd{>o2R+o&@hu5EZ+iS0W$5(JMH9HOd<SKJ1}5RkZm5wpEt8W8%Q3 zm~{9Y%$(*!#IM0M{00l{t>XW+5}av%f6)43S}OKFHu(?Eie;(p*-FN`3=+iMSTx5F^` zMy`9C5WiLlv=1GV-eo+y^NF%kB|(CQL0zbVupBOhf|(-MR>RH$3a(+JpbxkJ1IXhN z1EdRj-|=3Iotge~QY3awL%<*$CvHN52{Aj-m2Z~WaE2=stBo|v8!0rvyLI7xfacGO z2Nf6@Nl$^UGE6U4!h7L=zKp>2i9Tfy=b5kd4h=NlwnPuT$Fqt9NI!M_CzIW$8R z4j*ivpRQV$3);A)yUT7!`R3W<><5QNEPj%Ewt9Al3a*3GqH zKKA#Ym>!lN$yhoY&yZf;fzG#o7*cuGHi z%C!Cq;X-zWIW^WZuY8twbvcR6LIncUs(}2qdg#ItMwq;^nXdsQiimv33CuiWF}aZS zJGQN?9LMV@K)rIB4?>I-t$DqA~D^A&5mzT(4dv?fc1dJ+jw@8*WLtFZ7oU%$8P@{rV}F zBj<&`&MF#let5!22Hj0k90P4fEY9Xw7xQy=N) z5;FO{1k8Xv$DD$|2lYGDIvCQiVN;7oWmVl)SI1t(M<9#oqr!h3v&yu?fm^?Mv*1t` zJqq-J=k>FxA;6fymC!X1kwjPYcg|R7s6X?MLqiPY^;IS1#}3vPYw;Cgevk*TS2K zoJ8y&kzr31IiyGc$fizIT6at{NB9{BBer-V<_(E4_$Lah45Iy7v;i$FOBNWNU5EjmCK5I!nhU zn32$ua5p;1rFP?cx!l*zywtNxhIm{@Cz}|++ph;Kutf3rgRV^cB@q{++*z!A?Ot}& zBX%!!q}g%rRG$^~(O7LgM^r)ja?dIb@ZWogmFyx@%i%qaQKKV_cukf<9~;m=RkK7^pmS|EQfoBId0mvf13>jKlBPO(9&Dd^&@;jw<+ED+W@*bg-naz3i^p#8 zehP$qA{=KJ5CK3+@iM2Y?@@pA&TKD+6B<>{F;0Z;@EAplm#;ktPk_2O+pR0)PmRMo z5#C2f7DL!NvNl*S$jLWnhYEG3<&_*8h8_PGfxy#hs8-JK(z&%~PeKDE#ktBpW1Fg$ zML%T=T0Dc9U+WonT^RRW;e%Mrbm}yod2lvbA4xZ9^RFeGZlkj)iLj*2;|uEjj8gg6 zw|X4@G#A_qU#)*2lS;4>&6I43H|!!jAbK*`9pw#MPTit;Of`9l3FVum#%%83H4Gz>Xt zPBS}v)&Hclwu4PMaV-^1u~!t@O-ID>``eDdC`^dwKLJv*pb-0zGjK3hm%=0GJu(&1 z-rGy8MXAf7!W`%5oD!y>?ArxYG}Nt5$**RhgAIX*?Ne3uv^uU}B1K@6U1c6#2a7@gKX2qG*zjIWZU-fgvPiRp z_rJRRQ1oW20M~VD4n81W+L~c-YW3Xfzcxj^M`n}XwbOtp3XR53h@=xJ#Yif9t>n=< zJdEOu+0umcAUZ0;>KOKh6u2f6r(XEd)NY3yOu7cccaoo2TW}F|ODD=@T=znc;7Qm1 zq4Zxgl0^%XkqeV=-q^&5(<6lSoSTiPvf_$=bq7mapPftGJ5gwOASJO#s^*R$*J_~q zV~1#d+|`Exg%D1p5YNcy|HhNTU>gii@WeajS7U}!QQ|7y0@Wl-;$LGG%lE(h5$G}d z2-6TI#q6GpjEkZ(xgVA-Pr=#de?7U9Y|i7eNG4ZJ{@QgQhXZr*+J5Q5oOCH1q}pY? zu#$a;e4BSFf*G1~%x7e^b1S^V*>fN%YY~FZfHWrlQY-Y`|P(_He_^vO@T8&8s~0M*p7k20j~^(NAvH{DK0NFvNXwNHh|zq+JiX#fqDE}x2>csY@Wc2|gg_$EZ*u+x z7C0*}l9`u5-2NQ)pJ@&LIIs`qzs!)V`1JU8hL%v=+)#AVCbnkI=J*WEEdTSZU5WaC z=xs2)b9E4&k^$>SX4YlrCILG7u5{4WwoFk0np6$+zB1HoH+_1F#b4q%twBA?j3SE1 zWW9@fuX((`h^AC~ZcXfPs-HuYG285LY7$6Yx=6LJAD360YAMxRE6Y2#uY2ElRDAT@ zv3MN{%_O)gZko%Qz*j!rZdbzs-9KL$tLbQ_h%MHCW5rWYbo>(Yb(uuWHNX2QGJlLo zTDW;48kyF-p0cz&6lH8_V`*t)OAi}WUmOHDCg>``yja?O?!4Si*E2kDk(ZL#DIVk} z+4?z>V_vFwCg5~_PDgO7_N4298~xZWA!)Tu4szU8cfMnq`Z~0PhZ|F5e7)J=xH!KjTnS zuYo{)rIHoY6&jd;(TgUz25nScR<(^_M)#o=sE0?l{*Yi`NF#XBVfz~$`qFCY%%{u* ztTy?BI2_OE3M2~9!WH~2*A#ixR`bCr5z|OrJHg1U^n}jqz948ZeF-9X;LT=PlFtnV z^wv{H??cn4!mO8FiL;K$GpRT-gIi0#ncechdog}(-4z4&4ZIUtTH{C;xq5pzhT1MV z-U8@8{r&wJtNZpddBqNup$9skURlyMe`ruW?)anictab0O}o9RliYa+;J|nP_X3~K zJTB1WLoK#YpopUV&&O2&_#?4`jiNH2_0tf!#g3TL7C1~P$|a>Ll*g<{^xNindJQZ z=C4)auE!L3066xm8ma4P==%a$eNbj;ljQQ}ggRz|a0fxJfJ%EITcQv&BMBudl{eFC zeod`z=Sl2v$4QePjA};}U}f5OWYDV~2Y1#14B0B%_AVCPAuVKU=#mn~MD}M4Rb1~u zn`YWz4}|rL)5OFoyIefU#DPa_+aZ=vWSl6o3CU?>R`uO5I-r^DO*d+Pm%VLeahFa{ z5oS2wtdTVaTd}p<2*KrFL23^n6i*y{gna~ZO0x8&m$V^#5pI-L)8_gEkYy@_zPY{+x%#X(A*_PMw2x&l&PylJX{6WB)rU1y6vi0b_xKS{S}OC4zff%SD%=BH7)T$H)!h9{=%INWq#*il z_kvb%IsI|@C;kDqiBzTIB~1DE$jC{9fsol;^?!npafBniae<^Z|AHAmZ&f#Api!os z0@V&R;YkT#3*yZ7i%Z@{O1d4qOQ~>4Z`)*Rp2Dv3$5l~w;lHzdNH+|e3G$#J_lcO* zEIv1lKr+tBv`&oFrD=$S?Nk>IIhq0(Op(*PpMyta@y9jrN-EHCpc~uL0}QOAVt|+H zpc^9g)!cza*p=98Yb(dh(6(yRHVE%_6Nnc438}b@sHsK{(ed7k(FM>Q&l6p8_4^Zs?&5G>?1`grLniG&)vcR?@~6GJfpg*HX( zwCrrvFKW;agFM_y*v;c<7AetF&urnoe`BtSSX_3j;M(L5<2o-yLaIN)u(YO@h}|uu z@Sw_Ld7&P@;;&ky7LQF~How>KDS)O-&4mFm1=wFbqa_24$yxfiTKj>g7mmR6`FMK* z``YOb{&~Ig7jXXPUR5w76ZJE-$LbWwG$^rOVkQw02(Qlb-ShhkPmOkd`(GY^j{pDo zGjgyn|IcBUhKAktIHK=b9mabE6m_W2{pJX-aT4_*H%WawdR>D#Yy~DEx2Z&I?D1S# z?J3FT6+R$HL)_(UXXj;AF5aD;?{Crs{J#bo`)cU>NI=9ReRzXF=7?T$=Af6|8*0?a z<;zS9Zpdxu4Z_B(pVJ=sQfNVm`P~aKwmO-P=9%nRK_t@qX?<7ycUog#Oi{i!9hK@U zvqJh6#yhDZFY1NukGJ2`RZL*Sc|P1^<$S+_M0QxhN|=$;(}j~WxjvFHoJ=lGZbUOj zI=u{%`AS$En_u6D!&ngYqq%+^^>noc)uqxi=Fh9}luCKhG}Gm#p&S;hOwMmAGpF33 z^2KiyGijt?R;ga9s80~{oC>UJ9MF3m(;x2mj(VJ-7jln#&yIQfIUQvHmdjhr(eaY# z$sv?%m!rUDn2!V-SnqQPRrr@Kv7beU?Rcwy9%3SsDNnUC%C&;YZWX{hz??I!?NnkZ=Y86=PRb~R?4iy5Tx(!qmk;(`mip>HqPQ2d({gvCl~x|n z#ltfF&#(+(VJ_V6uI}?oH4uPIgbEwAv8No(zI@RKxg>0ecSY|OZ1IlL`}amcdUdrp zRSzlPdD?m@z;pGz(e20Xr?{c(mfdJ1$O&CX@9Hd!!2}Kdx?KL_L$=mI<)~##G(j|P zAoHZl``Ueo>is%T6FLzQJvHoqp-@FFO#gNSYArOfsXugJ^pdz2DPk7S}2CK^_edpI17(oYS5gre@mi^+t~$Sadv>(f@{ z00Apu`CV#nW|IxBq8@lyEivJ=;8H)+h=mdq%$efuW4{ln5;a3?&)8fK@m!e-8unF5 zz&C8AO&=Y*9E*6K(05(R99UVFnia_XN(zn_e~~S{0Cg~_*}m0*T`WQAxK;Deey3oA zImKw*o*>`ckhQ*T&SkZ0J~ROU)>g}Oe}9Z5uGMAN+zVE0ApJA?rNz|ru*W%~Hr*#n zWIu994$@T@GSY28G-1EK-|T-4$8CR{9Zs+-ijXAWm>cgM{i}(4a6aC+$bUk%sb=e3 zObqqdjw$MY%L#?(UQ)0!)?g|3r`JE4!$R^I#)KX8X1&g#F3OrV;xn*JKj_esnpzd4uTarB@^aqw4n~S)~F+INhLPt*haR^BDP~taFS;G+KvCw5uKu@>M3?(J`D!(A-KUbn5apRFrv2{^UWq^B8U;~KGDuU{g zI;fu}E+})7v*^q`1evUl8>%(olCzRN7-J!|uP9s&`X8{)R$tUZF*)h<@KECvs}6Ya zCSn0GeLK0>!osW7v>U_%=S0vrp7ktEBR$UEz|28n&CUb3!~U|$_P|^{J0_f9y-2== zXdk~p<{bjTIR-O*U3SR}7c;^Mes=J-2u1`0WpXj{5EN{6Xi|zvvb(Nz&?YCKL}WEc z^J@&Jhg_5zI)8+Hc_RQV)iS105F;Lh0*A3Y3i$7?&?68Bzzd*Y`h%u?y=^3+Bi&Fo zR_0fm=&&3g=Cc0Tl}LoTXcA9yc4%o%_clGDV}j!&j`uwcaq+ifHBEh9-!0o?uqr%j zF81kJ_G8z8qD6~xm%l9KkD#I>;;aFM_C%Y4kiRL(4oIh#p!TL;W`=fX)~qm^nxc#_ zKJ0zz>qU7rRR_z1EC29dB*e`R2YQ2*7HIKD54N~__-N@sp}Z-hhCZ>3|-PM!g-Oxww@XdgGY zz!0ce8_IfybUBslFi-srSp063?k5RMCsk&o*-g%koHSB{)_h`h?T!5^oE*=kgtNpT z@Y@NC@J6b*1w5W(+{!mbYiv?{aR(qt)(R6SlQ)c*!}UMmhXJQheErJkYf>;`EAU-` zrBL#st>ra@m36N)`eyrwxMGv9_R-pa>}?3w{uKkbWP5Gb4o@}>QYs3!QzZTS*cU=j zz$qpMjU(AHZ?)5o9B_J(qU-kp0MWYl5a=BlB&epxU>P|JKsP{txmxy_L-LGumK=Yw zC(#sN6h`M$srPMxnD+TBm@8i3Ap~X^k{MoV1a0Gxhf^DS$_|(0ade~ z{86`hv-^?naw#+z8by81c=L%)kn-(V3$#SEqLHcP4(Wp3;2Ni8mHorZg_{`3 zIC0!BhfM--+wqzoROGMf`0PwLgExJVDPHS3z)TgtZQizLU*>~HK0L0^HD7y)^|J!y!~>FNu3KEhdWQvhuiJW3WRqBGKzq%5uWoK}S>z`w1H7fgaTI zYLN$Fr^>Ee#8)V%^CP@0NV*M58qU-?!<@)Y?`_j!aY>)J-}w<`q{5Z310dLx)Eu=- zPeB+rNSAmoDe?e2gkr^|$myPug6hOF+`ceN(=t%(8zk2vLy>HVji|aBd6t)hlu5)S z0H$a`q+HWQl_KAR>h;DhTOJef+?u82TstnX;*@m8Xm)2zW01T zB?64>wYI1Fb+yvA$Z4flvd6$vgJ1aT(Yk~{p->EP{6-rO9L)tw*dyg;ryQjPN+jh- zigAAaBAIc9Zpp0Cc~({1;fMZtVZGu@QG9Uu%5mCuFqJU`&&!Prw!kMaS0Os=|I#PH zeZQYXCL>qYz*C}@($Edpz2dppFnk7-$6ZQGc=+Ch&r`I`pH1zJxbMQ{R47$Lj27g5 zc%t3@HtwZh2mkRB;fV%4U^Dvi&tNA&PNaVdj z+4{H@8B?)MX4?`vT$FG4MO4j%?|v6c^m~m0Zs!Z7kfmo44OO#%JUhhszRm#i`njL( zRnW&!tQ=E1B*3d_+QGXWdNPmw^bwq|k9(ROB{H7^sE zZQbhe@;nm~Hy``S6h;U&Mm*D=UoLL#`idt=B?>90mQOECl*(lkj1{ztv&uZ5K3f^e z23FtVDkNLGZrv@f&F_+IoJ{7|AC~To>~C=CJJIxxY;Iy(aj^SHx9l9yy2eGDNEoO; z@=W;O2IrivM!i&Q>)uQnefZNi;au4``)ok}kT5!Mtn033l^zJ#h&ug_E%mWJoH8sV zta3a#cKRugv^w4JaE}7UA6ar?!`J`}J>Uax9exx&{Zmpyz5&#o&cOButFF9W)Wb>(2**gMJF{R#4h1qV;_)Fi^P6G>?lRbg-v%U}DHwv)M=b~3_roXbkkR6 zqkqr9Xv4?e7=-KEg=|j!=E`NV0T+A0%i%u;*c`YCFchseO=N|Is^dSXuxTDe(9@_4 zuDsn|wFlCH1>jnL2?;9_+ukTmuvl^)7tU<_**6|Lyq>Ww-V&>3W20XjNl*-4x}I6x z@xa!$!NgRMBtFM=F&ABm&gT%qWNN126b74A zXiAEASiCEo;a=f)tb>4s)-Ljd#7VrDO-r zpj=Nt`YHrG>BNxV?JmbO`&znlw9AewTWT{w3-FntgBW8Hr7ax01{C=BuuTLWpbfMu zznMSG-S8CJ{$^mDA+&763<`C8

LX*z!Q*#7aVm4icj$4Fuq0bzUzI)NnHs+-au$ z$2*Ezo4XtSY^kDyvvqJ;OYt%YnBo%Z+qsWi{ze8n%)5Y|Kj?mK3|B(VHkH{9f7+Q* zanZcjcG$FMK%w4Q7@lmst)Xo!MMgi0s%ZPMZ%|ZTh!+F%F7=iYwj0LDEm$-M=juG+SQ9w@7CaZ^j$iaXe{0|^^NrC?m{p&{L(wpxO#k>8 zr{;R8y$G|bROnon^pD|_g1BzDHOP|?3s7?zuuYzW@vzxV=#EFa`9MW#9?Y`^IE>0g zokaoziAb`EyD6%P3g-yB=*(F;WfT;?l&v#6-<_9Q-A%<5`5bpHrWnMk*|A4kIQ|ke$El;^G=$KZz#sw0(j$+W@a1D} z(8dL}wf;V=A`&sL57e+-7~IMeAvt(}r{q#dwScW;~5P}O1LM3~t>yw)sKba=X%$9dT!<0(W0Q%z4bm49U7mlBq|a0(pd{$^S& zU!(#7m>jjR+z3-+O;_| zSX}((2WG0JIwq_2i~=E>*zZSjkNgL{94NiV!-jwEtqV)Q7ojqDsR=7t_i*(-=<(+5 zcve5xLxddXAF!uksafE2ZY?^dSAa%Kk#KoeVa#Q*JS4;1G%DBVF8Hw9ZFd4xhhpc) z0D%&ZMjl80@|`yJ1z|Vbs3ht18CJk$}%t@ z>-(4Z2dDO)mlpu}2$1WCeW{3!cK~3GGWpRuHV4X)EM}rPiTbyaKeD}86XV1vGY$Xf zQh*2XwjimdXNV`d#vf=PddDLz}FEBm?OxC;*Z7#Zu^NQrNEh6h<%`6yqB9JpmdvkDD) zPrLb2lYFyrsKE^MF@{eJG9nqm1{S12(`Q|shw*MtkJX3T1zzcMoucpVPymch#H=sa zwd8A&$LZuf*%9g8n`{YVZ11qCc~2wp@-V9VI~ll-s*-BParY%$!ed`tuZKiU9Z~%a zK$faIs!=e=pUbpWF;PXtyGY$eviQP=HiGiN|D2!PXO}6ICYHeGii3Gfej66-8-$0%16Z|P|{WR&{0ZzRV-`3xc7iM5KG%a-?}B!krv$dCrx zZrasoE$){PSIEifSCjot-b+#KQw6z#lizBL-=B^e$f!%8ilH7C5CU690&ec`<=*;k z6R__R|4}K-XSFCG_}saUb?f^@w6jbH`Ujh!WJ4F^wp}dGXYSnia(&C|s#_k@4Q+x- zWFFtLuL5GEnw6lZ&{2qEjy1%2WS0C$dL}I<8p2-8kO++@9qKw5>s(=^{D8*WYh<6C z%z7WK$siUxo5iGwe!%y}vW04tKk(VL2wO`Dn%SLHysa4smmb=Mged= zJL-6F0D3d^A|s)nt93YXY2DFG8n(+V=q(a0E2--NeBEIvQp5rf4wT`VuEY+0f`aHq zfN0;on{9-wVBr4D0N$n)^c#F9m_l4%;9S3U#uOmTp;t<`od zOL3;sH|!nrQTrjEabO8$Y^Eh)ft5+& zt&-|(&MG{VRc4yeg&6!-ot`~it&145kFrLYWLM^pt@LWop?ZB7z}5LP2&QLfs6nG$ zsRi^1V~6Of{SJ#HcUrpFu>}d>v=roys+3p~1)FP`yl^aON$ogw?RZ`C;0}Hz9tUPe zA>XN`w8G<{4E3yo`G8#f;e<$JeNKMYyrmes-Q=iOD^yxESYZwPBT~m8I%bf8`ON)r z#;!G21BV(K0Lf3;;CV1(j`f;eJTM@Ga~bZJ`I-pPgFfBS>QX7J5&vit784XUDBmQ2 z&@*UrhLQ6;?xoyRTZW-uAuS}AF>KQT{#qwALDT`Y@DBsjR-*VmKE0a3w7953i9*Vr z^E^4Xgl6Glh_t$PvTB8LAb0F;3Y6b051dO>vh!SWvpzz+LT*Hn%hG336K{Io_<|B7 z#qh*G-8*7W&f|V-YG{@6$hsL$T8;V;ysN}m(Ou_k5{oxk3W+e7TadccIzjQ!hTLVI zG?M!*ea(x#Ulw8Ys4Hk2yaNiOj{5-$8hFwZut`XN&_vIx4NQ%9Io2#}Y8cUwI^$i| zQT3L)I(sf}=&=Z72SE^;3^a%`ydut4%B+BIR?FY0R+N|yXcR@L>n};U-a=yT8L7QL zm4ab{`@L|?M-V2&QNUdz;K-5VxLwX&9~|7gIJHB^B=#vFBk)s3a0&PErjiUHAK6ddIRPjPx86oNJ2jiND=BcRKJj+%-BjK z>aSPOaxFWVQk99~YAa`vxFyS)3qe^=A17iNqeI;|=_*utuKG7`tz8g*bfkE8s$tM{ zWuin5tCKBIQ7JWMFdpI}|I`o^J?$`Lr`dr-M%iX3x`W3@btn?`Zu@qCA7ust8q8^( zH=QV@xm|p?Af|ABYx7a--aIZ+1i8I)r0_wvE^J70xTTD{V_lFzMfR;MAI(1_PZZC> zU-J^L9+K6H*(FDR9l@}WnOr6#1smEXggty~dw`zFhZA(k{^-wO>@2Xckd*P)k?trp zF2iRQHcB%7^z?raQ92QRtM)~(QCY7^3Z*13JhH7sbD=nIlbbIpoIE1{h>_P!ufMFp z??}GJu5YCE!!v`xG>k{iVHz^}i08Z|@F63RbtvZiT9W)cxXq;0!No-=DrBNe*}vYo zE~w~zzyg`2<*gtL?z{(Xy?lJmK_X@0Vo*I__Yl_k@&{baCB`dD*mhlm@Y^C*VAnFN zD%AS!LXA^AWVQjJjR}|sJ4k3*q`cEv?(6CT`$R!cFS>Xn5L~rK9~QQMa0&5tMS$a= z?J3f)AAwgTZ8^kC-NPs0d^a?mxqYmJ_<9}HCAH?zkjRhe=}Uj>H6`TzeW2T%xwR2x`Q$G zAJBZS9>F^?FqJ5=NW|n|SgH9u+Ihh^t|swFpw&bxl$ImOMwX3VFBA5qi4_~%F_j6TCM*;4}AW4f_E?ik^53wNID^ zTi^D+kE>8%=I3@Yp$WqTbwtqcPTpKCm(!dT z=Rvtp$RzHZO09o)ka;t^kn9XGiPVdQNpsCMI^4OMX4$4{IsArQzS>~rd^W8k9u&=a zX|t-gTZxF>{N^*x<30Xj)RBgSyXnB%alwyrCJU_X(LP7LH0!=;OP)F!x+*!SB^FO$ z(%Iq{eR`G$d6xQ$$Qgy($}a=WA}Wg`K@GXZOS<~pn2^z1V~Dz24KC0pYGTA2TULhq z_LWQ3LQ1plSRN86OclRLTub2=HPj@v9ZW|H)lrUIXVc^clpT*JH0)w!2+s)n`LA%T z5{&CmY+PLg$cV??%80~67Oa&XKvo-0KyHOZrX-2f(VzEX5{n35si;;#xMDoiU9M)E zwkO2x(Sfv$C{f9oo&kzWrOo2sU)F6WKN+v)_1r}-N(+TQjMs+$loa&^u`gASRK`%6 z0x}8aU@v*x)Pw^_Bs4W3<~}lmjw@Qt9E!>&F-C z)rC~7&h2$Ho?F**E0vt2`YUsrl<|X&Eiky$*cawXX{V2P^I$`hA!|nm5G?m5F(s2v zGgOJ}24gSnFf`gOXKU*3cA&%&PMgg)2UX40HB}NDDEn=6Q40v9mt^a!&3~%W$oCZcO zY_g6+Cf8z;{EbpD8=?m-2%kWx5e5VWe(h~eyGWEnJuC8a591)GIoh!qi-&;Y+ZJr;XM5paSqhayrN3YmmZc+h=Fonj z9>25S)uQhPcqw(l9yY+ThS+e)B+G9c>5QZpNO2AlAbwEQx#==K1-5Ke{oomy)Z1ajLi#1+tp=q5wi`^zmV4eIg2#z|0b=x)M(WdeJwPC1YH`Ev8)*n%)$=qhQ zE%OA1jbgR&Wj~ewly&)GOgVjSuZgm5u}A7H=~~zI^eY%9dY>A;9`DgK{&@WySr0rE zd19|6LQb&Ih0Ra^^v}7$rpsttx@_P~!3-DO0|(DtEAA3{acX%dB%HQ`-(hY!q~7kv zihn8RB>(JhpT9(vCCYO2_4D}5P zGRx+Ar%LZ~s5!#FQg7Qho|Crf)qV!5v5gRlf39lRVi+x5bsEjn^hPZv=z?YbVh3u6 zNi>L?!oPjvU#&qvr0r8@1cn{yE~^x~3hp5{=0;qEgm^6ZK0$Hy#E0%W)@(*0?#jCA zU`Mydezqxd*!HE&T79KVT6n_zmyXNd}U2B~XNZxpI*>aI6 zUU*ag?4)b$YW4ciqu&Dr)I4C|F|(*~FUOK!B#@P2=dTj`!7Ce-MDSq_uVQ zcjC$fd>;z)5{=K>j_{fq%utzO&VQ8tL<7Qm`<8OjzA^hDk_6FS#+6e-8yJ4SOk#>S ziXuA;ZU|OjMocInDaJ^2!L%P|@lyrJu3J?9rUMdY3}~b6L+EE^q~c!^bYOEi-IZ@> zTwgMtn>&sk(e59;FOB`gbnrcQVP}3qI;3O1F#Hmw&&$2}l}6+xxWQ8cXi~zb{hQ$C zDtrCEQ=u?{Wt%uaRb(o!2 zi^b`R480McOqsn<68#OE?1+z+l<2^`8ni}w-pA3TaaPq*ofETRrE&Gd~@}W(8n1A{(SKqcQY{a zD6?Y33g?!dwKK09H)@ZBszo>sKxe;rJ;mhl6~n|a-M03$vg>%=d8Z#xc~-%Gv5KL^+oPoE$}y+6>p=EHnTR5$fJ z5Y2p8tH--1fj;#nDZZ30&cx}S9}~7>YI9|i$fse#PzMM4_r<~K_gf(ZMpPv6!N3qN zd1Ls9MZAdwiW#9hc@W>AjuPMNY4P30SeMOU(*3vB-`mVe#WUHoW)W=7l=@1prDZoi zo-d5yfz%1{$>env8K1m|I4iLnj_hS0Ni(g{B4R|RXtNcp7v+oFpplFM;^4+Qtd%6( zu>)htrM@3bvW7N8qY!J#6*~lw&3=kccZ0DrB%V0<%cy_f(0IPTH-2qZlX2nwkN4_S z_Ukrot6WWNT^*cu^KL^+1uQ4uHEJKg33<(xJx4W%$nQ7HX3)@+rT}p|v1&TK=GTpx~1uK6Qs%n4F zesaBTwz_SU{#1M&MwZb`F3zE~-!cv6oxXy+{kh4!Vr#KzSx>1{qjI6uNj)hPKO!_$ z=)TMofkE!%e}<(1_m2ix3p~j^v3nF_o)MTN+qP}nwr$(CZ6`anZQHi(>?Hqg+}?K@?;WGM)v3m+RW;|D zk9Qq(OiZyLS!Kh|p zYN}6w9f{yF@ zvlOLB#VfDa$hcP&I3cCxfhqYPvl0wvYO*!hs>G;z+!~cQ*2PjE7`gV^Z-v0GCLVbI zb>!48uDIl>owQIlFG7e7P!5Kv1loLX-8bo06H-H^p_wymuz3yxK$HavRHA((EBvE@TbHLXeMf~s!gm5BQM_)A zWo3dQ8qu4o{k4DaU_M{o&1e-MK|}{d5m4l%TCg_oNEzVMsFQmP1e&KPq2b7!ftyro zHAu7u5j|EZG)EG^f7vTiXiSGR`ixdBEYh>IpEMqSRfE9eE;W}X^{Q^Do5!-`sadw^ z>L*0xV6t~?z+CY!`KfK}8-#2YCGfZF&zC1^bT*QhO6uzEjx^Wz8tPTM@bEP}wlAM{ z(Dt#av>-cDoV#H%AE4KgZ{rmvJmD6`gvQWv00Y!F`%7di$^G)HjP2>p2rWJknvdRIoGif+X-40{}K;?Lwy?UhyVlM{5Yt+XP&T%wt zqbE`jqyo4>tXVSda@vgoBbq48r~iP&r=&8ZV(teMf<`Ol>B28H8$>hYP9)EFDN)kU zPdF!C>T_7rN$-Ed0(DP~>SXT)qcBa9Wa|%CF%qPsA`v0)zH$Dlw+f7&cmh@bOPOr% z`rEW=D6QZ3mqT*2(bDFnNugeuS8WI(BiJbOUK*=nn}y{(78cP>>#Z0 z?k45BlYo`*)i_(UJZ*inuf#xXXEV)_fKN9d z=yO|pQU;e|8sB25(n{Nk-#}OfP;pL>=jkx(OzI*IRK2bUAweg^$l(_@N%20JFC-nj z>EhVnW~jR!=YC$_E--EUTA!}T!NmgwnSP^c_;$+|K}GMuT&MQWCj*KD>U0s1Hhaur z=fnP9Df#C~LHhK|Cb_hC9Kh7#mJh(m8i`TC^RF(L-`+us5uH1)lRS2|>FA4l8IC1z zEO+KUe7OY@I^7?VS4^<2#YPUDQ)lr{BdMGi#1Sm$f)GA^WfwVIDQCMWMcL*fg*NXW2#*71L_;TUv5eW=DAK{67^Ro_RIY!2%V0ETe6cy)kcR>=5V=v&u0vi_N;De z!?E+GONwQ~vUjZA%P9!zrbc+mm7*#JBiZ3WJD-#<^RnninfY$}M7gr7@uHYfthX@P zf}_7ya$tHzE7YH~Q|NdAKBD+uXxafcYSOAA7QZJtlKQg^u>-zP8M?aprHJ zw0-}#Nnm97FHxA`|3}EZmbCK`8`8gQlM0`tDjGV$fVxS_%o&m?jwAN#>uoEqa9#nK>CMa==*j&aW$Y*>4B`r?xk)JqpxR%PgB)_-vkyV^6G+j@|9G zZ2eYQ_Sdc+#*^1&_oIb9LDIFYhjT6z^d`97n?(<6-@Q=1);}E=YZ{retzEn>7W%c| zz4!eC&cdNBKiiY*;MH7qSQUSV`FCB3V^iW*r|92aWZ%|1ckJJO*N3xvm=YN=^DF&} z9$2Drv+?ALo{YBe-`|F}E(~~};g*~D9cxm0PVFt# zwDark`#1tCeGs$V2PqsbUK;j(@4S26U3v8CPxk&bF9#lBOZ>Hg|aQzxh3T@5gNYBS1BQr z{M<@ms?9an54jWXGZ6YmM|Jk=)Dzs-1SEqc&H;FX8RL+T-OSrR+jnfMc(dVQ+b%yZZ1?e+29mahJ7%#fy(7*JYnctz zmW4y66Odd|#p?dHUmKP{q5VYU?p2EP;r7KbAvm;b%d$lj zdCPQSwbf6Sw?0HIaRHM4f!B^^_ou2E(-Tp3kf;(edfN${3y0+;*La}Zm>C!E7+3&# z5GhJ*PS=WC%6yl|Wr%hD0J$Xt;4Iq?b5eH8$908+f)nQqulBp=IdQI*KkfEvefRNf>{ zPOUTIb7~kSqv8ui!KeQkVk6KoxBtYRd<{7Jy3VX6jo&`Njv7j{S}{ zD^zJIau}9;PLzx8@8>Bqdeslfk~_1n8OG$|duW^)xQY>?v581eta#-b+LpeX=pY6R zIW&2W<;8lwBaq5iq$>abA<$h_?;k@-P}fq+F)>UiTZNnL_uIR>ju?c6Fh_HFLFleU z$@-zIry{KphA4tu8p{RG?DqU!u0@eR)baLwcUQ~`5Cbw-t^n5 zt2$5Q4LIkfz_*U76kw?nAhYB-($jI%z) z?%O2)m{15cQ_nJ>1`xsId7(GqfR!k(1zY>I)VpE66GVvDeFf<@odQmNt8#2ha$=hk zlH8&Z&B4zv+klA(bj%>Tw-0Fa%yXbEO$D7W^n5syAWxiD?D@cSRG?vRM?OZFL))jH z643_`FXPmd1eBd5P&p=S^F*6}t~aF9(Kj=2+sOdM0ifg^SdTiaAy!euvgbEDZxJK- z-VoBD4mTsSpGu%qTB?#MIV85B**VMsQbR?ps7Z)1o--{OB)aF}=%v+QzeHL=I|bQR z3OGT{aZI#IR0m2Sn1l2s`sVbZIK}~uWVupqZU6}FG_|IF&hb7UNsD`fyeE)eu=-`MQop z&|x~Xsxd(MVQ(A2pYJr{wz4-*g)^ZFo^_IY_RA3F4hjS;O>0^?6MGx<4eDxBS+DZ2 zBd&}3TTxa&&i!cX>qXHi9cTKH(BW5D;$(8b!@m8%m*Wkn*bEdrB%fnS`YW8^3K|fH z1A$uP#zY_mXS%26T2I0McpDCIdpN)yutHV6U+j!=+d;}}*FH#RPKkffEL~)@kYV>N zZccq3DUK;(njfT?fJj!+EdWzLoeyXuQhrCplUoi(fgo-^o-t(Y@UoONV36<~Aoq)# zP!{?WQVI_8j4&$>kk&#N2!bTU(UNNhjRc`som%8Y3;DNnCpJe8^gZkVcY`QHtUDdJ zbENTrbb5LtAreBf>DDWg-3|OlAL!yazB31Jab$-?qS|MbYRAKNna*_zwqNBucL_X=P za0uQ>v1*;+;|SUu;*BxuCZGz4t7@8QC>dS>LiF=tF;QMc%pM*mrlgs zeGv;KlY^5KMuSH<46zQZlaK-NTaHK|ol#t4a>BS~%P{g&3un5KL(t%}5j-siy^TVd zpJ*4tHK>dhpQ?d=ZBhw^N>a=~E`$n!lVN3^y@7_@9sN^yrCMCl@t~bj4Mly8$tQt| z+{U!!lYI&7)S+tYO&sM95xPK|PW?wYp+&&{e47(AEKX#*yH?nrdnVd17ofJ8H=saEhs5~!o8-nB|?Y@T_-pHzV6s?kxGQPE*g9w z4e0ma;&4Gb{~s=n&Ek zo8pDWCOJCGj3JgJIxx$kZQzzf(X!O@77?4k?~U%-mtJsF!k9sTaq-bFKGwYQaQ1Y8 zK>`{)({L!HNWUa-xiKSoZ-Vc%nbPYIwz^s<#N~y7@w{RI$xPiWg0PDP7zzMj3TQTA z0NTknO^w>8_gV|2lb{k%0Ze|O#P_SAW@r6kuO?{H{5jM`uQrvJ*xXptxw|A9oK6Ye zeX{5bKW1+-9i#AB2%DG;I0RfL&$9?aY1jDtE^Kj_?jBqh5k!@RAgRT@B|i#86dlIW zqND2qRdK{8PNq}fDgzMeGVwH$CcF7BS)t|}BwwjkLsoELH({_|M>@k0n%^#2PA%eay&P+eOIF{1? z1q!q00;Kjqm$S*+M}bx-TXuh{@1PMIJaKC&N6Zu?=Dj^d z1u9oWlCOcYH>=H^XO=@}&aeFYXb%9jeR?(*!OEw4qh((i3y&SDl7WH#?e<{b0 z(V`@kI(xpn^-H}{&EqC?ZtDD_-bbR${ru%xHARDY`@zY}#`E4`e`QtW;ZyEw`S`WDQQ`%0j9;3ZJv5x`i>$`(aV)t6`#fN-w*~jziuu**-=4tH(nUtfxnH{4 zxBCt8@}&Gyqe3*bEYw$2!Y>}(-x4{0QOBIV!*)+5Tb7@G#Oe>E`Rc;#R4L0SE zVc8PhM%#RHcomInm5EkJW!J!*)j?YGauq@;W1{)i+Kt%)PSK_L<$^B>DGUsXpJZ!)Zl{yVSt!T1kT+%}zU!g<&N z4cF6)L- z^KeJAwq=JR*Uh;q5iuhcfq?by;|QYU`CvPP*Wl^H`b!SS)?HSX{fge<@j?CwX!%+; zaQYz^lU)O&FD^|{=qmB@y56)FxtxT~*{&@__{vek%kIUfjwSb6TKV0)j7{G8$_Qr7 z#I&idJ-f2*8ZpQ4)Ux-mvxj#iB3j3~n}a1g9UX$*W<@!5TfgSX37$b7a4dvuuhptH zo=K|IT3?2PQRWJ!;;(`td4-4_*}2dxM@Sj;VD2~9e0s#;$!Fw2jc`Ja=K zSrS`*DN-%PyraOBk&XnH&I6`(ZYWLj^Ls^N=)Q}m(*)>@+l}t^)6&Yych zGAEj5j+gMc;}lRz2PQL1NJZ`?S(G*_n6VY@&;vCghSCtpmWh_(Hr4()6OD1J%o-7X zj@1b>C&;CazmFMrA|$mCC^I%E5-|)#1jkNus>K3clk~RZ;p-r~lZ3B?J~h~r;ZqNl z@nd#QVOY$~5|(nnF*IPx6atVplPNny+Ip7Iw4ke zi<7cL7nxzb6vZwISxIcQi=ZCrVk!hlrdCu?sG_h`s2v1aVa5Us0KgA^J+89Okg2uj>k70V0?))WpNWU!u9v zzjF5z#lvgtiJ&t6+&>TdB5DQGcR%y;cbDzMdBDU7b9lw0P_e0OUDm?`H7w1ca{ebmx?&EllUKrPjCokQ&jQ^yVW}WiB$N@A`TN+}QtkWYIduE+l%$7Ou;jwnw6$z({9#R`oGV&l@#V zw&luij(9XY46O4CwI^nP#;h%Hjwri14u+N1`FM@4uQR19nbrgHw;TVvqudtr(SZ`0 z?MT*-v@L3GV}RoXE{}n*)rq6g|VtFMqha07jRJPy?EkUcq8j z#H#s#cG+9Q`Jr&}vA!S3FYG;Uhp(Ay0fNPDKYCxfGXQ0CVZ^!upq<)RQAX$L%$myH zsZ)LEVCYu`p^!vmCKxvxI&&7}HT$coJooN>Kb-_`?Grfb%_ygUGb5V9<6ZH-r&Hq; ztX9ZjiK;~WRI*uZo2kIvajXC>ZbqkV@sArjjN};bMHS9mPre@9#@)|8fVtj|0$}+_Tj5EYL4p z{adM)ydG%*f06ARz#Q%BB9zsODYU7z`G89?a~xO=XaJ7eO-Kxn8TxNq5LgTW!GO<- zZg~ZdY_9g*No7#d7J)(3G*LeSooS?Fhs=Ay*>nwbct{WNa1Oj3=9tcT{-d>jdJtjv z-iKuiKSP&mRsFIz2r*1A{NMM7F+iNnix0C8B9Ka&Guxj?#YJW?ld=c6kjU9dGZcwu zV;-iXhCi1KRb+mnO}ul5%5N`yIsDZIk4J^@_=?xjrXXTXv3XMF=QW3XgYMsXwRzo+ z%rhs1*F8P3xf_&%49Ci>trdwg{`-pX;0O*}NPQlq0uRpKf2I0Te3oFphwORd4DR+8 z{_fmZgO6b6C*SS$sL3{}ZSaU|oP3QwV)(K^o+aHB za;JIeqS2&5_@pn|j0&^zDSQ26XGU@s95xCEH#ZR+*zG8e@wVzl9Di||oV3AfKKVm$ z%s8A|ixBbqdlTFAX$DNG1NlhKDPPGg$B5}&6=z1n{fz=?kQ+~c8Z#&vENf8X<%R=faZZZKKjnq{lr!qlzu*`=g0o z%cZZA2gc_N^ihYM_aPL;IoJL+JOSJi@sz#m`^c^($fuwE+2x_;{6SY&vqWAeK6O$I zQQki2zZmDk0ehHtI_it90BVBA;XovrJCnqpufym)#Bo8{H1#0@9kwu9YS!9Jv_{^O zUh;?hP3$Q&K4m6uX9ws=$Ra!cfwgi{NL#7sP95E`{MmZ38Ywgv@IKX>5J%DJWEMmV z>3t8Qab465D-9towTJ?W?3fvZHTW0I4?cwaznUkG|7e~#**O1a#a3H8zL*2euTCGa zMFlC5Oi#a~x*2|nd+QRfvQ$^~X2`+|uOdWc9-#t|iMOTCZ>|BBF;X}R<=<4r(bLCQ zoc;Uz(b3KxA7qOG;(^S>$dMc3gD4865OIUYHjN)_yH+j65!L%cWo49bHmy35l3?~- z(}{gN|In4q0ET4B;X9XZH3+ei!_(^#oqZZLrAGbSrf!SQ1I4-l5PVgQXF{rVc3%42 zX(F8te|3|s+7{C0B;U?Pxa@1vs%3WOv1^O}!ikEh8l5Lj^UOR#+9$A16-Pjuo2G5D zyJaM+;JFMEr7*B=wla;xj6U!=6icav1I5#7$z|igFShAposF?wRJD(tyEIzD300Zo zed!cTM@S{yc1y&=7ZRYha5jIs6leh0;^s4_^}m?H$4<-WHPdO}VSINKiuCSTj%6eo z^Vm8m%h*d&s9_1|in&*XjIZYHjA&MkaA-^(!AeJiCVQqesIud*gUMdj628UdBj6J7 z-`E0=7hZ67j^@+Tx5h;hl#NqPZcpR%5372)z-?JAVt5@8%fM3JyZtz+VvG?f6zd91 zhAQ9SSBEAnjI?_-V?B?>f+jRVYf`2F`>0LRuUo3N4%W>I@o34N;3 z;R^$&Wq{a2QF{b?uZVdDQsjX#^=v)N->RjuZ@^SRc8W!|UuR1!ME=RDr}9-9RkYR! zJp;Rzel}GT+SNLL?Q~k42HuwODObmJ0!W^aY+KbkKnc~Xkmcs9%lctBh$JEl*ig^a z+soQOvZU%2uKfB!8Mt5&*xK`JqBO#HZXV<~;^AZ@^>pABarQV+M22(H( zxQ=%>8xrmvG<1MjLkMTp3y36{0CByTVrvf%$j^#ZQ+d596hDG`B zPFN*bpQXQepQ}6l!$V{@%T}pfo)}Ftk%+i;EYc?)H+41%OQ#gKl9a7ov{6GZ__F(m zPrLkicfTFS3WQt+&>{(9=v9OyA(|3jes0(H&W(WM`+FiK4oBQ`9s3gcAzZW_bZPMZ zv7YrrLS*7n0X$*F%L3xPz*r2pBq3yrLZc2u>0s0I>wo8EFK~F&5iN$y8fJv6#duhL zlJ#SpNrHd_m8OG0rV;P=02AoV9Of?5xvXXDoWcYZDL8rqVe?ZI2X{|k35j7~WJrmg zJA^9=^uof$;E$$KjTl^m@|qdb-{6n^xDWB23odYn1Dj~6^7UVL9g219xH-)HFUO4T zWq8rzJ%&tR%(W&HM4l)hci2%^OW!;z#nY83`{JZO{1|!?=(4P(c1>3Q6kEF{B_I># zOi*wf)Gy|UN+AC8zSxJj3fP)1V@>Y=Jhg{86I`ge6zV8Hhz;rC>UJf&W?kN^#Kg9+YrR%oVEHvq{lo3bvU!)` zqcV6zc|I2-LLT_b^dUZ+3M@vS2~A8jxOo(7);ZO#=sPq*Kp!MeG&XJf3V-Fm+^Xnf zj*%EX!DJ#u9TzNv`qs6~Y8JQv!8hnw=C*0QTJkXb-A8n$crz2S=fOSA+(hJwUFWe9 zXMCo$61hRFJnLVOz5&9toodT^xxCi$%~-T3FXl*QED3R9jgKiL5h6u5!LhG*?mt&` zan#Ag-mzN$cP8$W+ZJ3A%$TJ?mK!`&lOUa7jDP?7#^nmKMNLvlVPNC3gT$zeu{Mn{ zDd&hJ!OPzLLE^{Gt^LOc#eXZo%s9uL7;$bn>zk=xOVsIdhPh2{$o)tUIBCu>MmB@ zU?7sAQbv9ZEIrRvZm!oer?-U~0>6C?nkr!%E95Kz6(fMyyaUJ|rE~3Cn+e@9OB=VN z_bgEr*^fiwfzS#+yvunUNf{cXG|J`5kh?)!na?+_%<_ZbDG^X zAiITnI0_6I%5$-El-o1bpMyb79unfNIWG!1aI*`xd@UncW-9o}I0(MSNN<7kxyga! zlE~km(EKdzio_Tm?-cx}5mUC6ZW})N`DqBx-!=noH8Ol&qm4EU(TEJ05d?j!dE>$Z zn?{V<$EPajH?0DctWYFgUlAeSJDPzA2uyc#rF4d_8HKdAE#(~J^AV`1Ra;+^nt6}g z5E1wv-7$uj2@yXFMjrtjhAl*RY-T9KAkmmsc2FgCG%xn67lBklXz$|yEln&$_;&^s zg=#~SB0LWpT{D^}&mD(lHUHWC-wdL#N7OOx0SZ${ZP+BX!aHZI#a}r?5Si57F3lxq zV$g0&-Wl%LTaosKDt(Z(ONYXc*hP`~%vyLJKU-lf_!sLn4SER{%;IoeIE1|}<;5oG zP_mt{k#E0{9qgLV4c8jJEP%5WGXGU)To{zAQ9$SD80FXaeA)FU!k)=TlWM#)Ow#L~ zY!u$91Vz^^(A}EExRZYyHQj2wG;YEB=nm=WO$66%wb1$fHr43Ar!2 zez^wEV%2(l;Hv#tn`(u#a~rYnG-Jgyb7N+PtS|!EE`JMmA>XY5RFK*s57JA9((C zSfDal&#zVdgh|S7t1`TWTB2}1)9->3HQY(wl%zBH(pOU-L-4PZm%-%js+OORPCct; zgY>YN(7E~Q0Ndul4PKJ(lRBs@L6aA&Au0bUSk2L#uREuQ-pK^j`9r?UOBCuk4IY@0*#s}mQ(zItZ=iL; zi}n2XUZ-p^q@3gg8*od9nSYy>D zaG4j8_#KHM!7jdaiPMLXm2%fL>0=Qp$qcF=k0t;T+2x=p*fsj#|28hf{~w(`Bjf+= z^qE(ERCbe9A87)YG4IX{Q3`=(4=?gMe= z^$ouNOA3RzR9AIC~l}2 zCjXh)C=c)mNIa9}5 zeqitl)%acKDa$kI+vlauU=^EFpy8yjhghODF3V%sRs8n*m9Tbhdu0>N3EC}0zG=@M z7b?rUa?0}*uzafjFC#zf$#ZYxWT@k@M@u)|F~i1)q0kwfopr^Z-mil$B5+L3b+6rr zr_a5$vn!%*KA$!+^cxk7Y{m5w4XSqDGzujInTL!uU%zCP)fs{( z{*@WzaZyn6-Z<|}k*_k;hy;iuWleNXZ{l<}8f`tma1-g&?st+HwYnU#NwUll%Dn~8 zmB#|#mk1xF8KgoqT^jc6p)4*FvVV2JDYO8<=?PcEXhJrZvSo?8rwYgbF>%S+TbdaW zq?!^cYNp(b19clm6xQ}=k;SxhgbtTQ6*;_iPBDrGy7L%9h6gZx!rZ<4K6E&ivE=u- zD~35Tm6Rya^w=JH%(#VkM*NMYW(UCQ=_tlhhtU54X>GZIeN@HC=QPP`D0VF1Y!uYCdj-~rtXL23A#gUxw+;` z<-w?ry|p(#yQ{xQ`H7Chu$|S4L7b`h8Zc3%;U0lr6j)K1fSnhMHc@0sfJD(> zVA+!}V+aa?t1Z($_+O8tOx^ffTwJiHim3%q6rSJ}A@KZrfMY$2nqmeGt0S%Ur&%Ag zjA6hr0)5DoGjo`vP7(GbV5x?DE=+8@uCy!}HZ z4JyduVcU7NrQunbFw+sHCXyy~QZOE*+|OM3MBff17UnndIP?As>iM+au?HEnVD?^) zUFXQM&gF=;*)$=5NO6VqHm69351eBHuT6!GW2V&;6Xmd!mZQ_9hE3vm?kL&dS^g)9 zo%^OLOjoGCz~~-h2I;)*?k)2aRfRR*Sny-r)ar$V?avIkcBg3#v8Xzbb2<3&rqS9B zjJHunKG1+x$O=znvKx)M*j2a>5-i$eY zg|Cnv$nyl?^0{L6SE5K`=w1!$0x6gC!&n|8hG5;NUJU>l2N6f;L9)0NBM`K6*N<0m z@$zesbIwbj2++I^psa0tR>sbDet7-SDTjhZF%JDo^T)us;Iwe_7>0mF2w!IH{Era9 zE;vLWXFkJY`QCt(JD_cFyj=yoVXROUUuWNyf8pxbB70T=Hu94N>yo!;JnMl#BF{Ng zBi2K);01q+na)bmuybd4=_zac{^5jW?bPF=J!{TfHmfQUgGgpQ+Mk1nBGT;M_I#)fgS7ax}sGpnJYj$06``c#R-)J-UeL8jm&70VwYGp~{P; zOeDwVDd-QY=*I$#)2~2GBO+7+|4|h;_jL-2wL3UYVB+e)K-$OeJB|=Mp_K!ahw!-A zDYl2?#!lE1gvyILh4~_G!_ydo1g`er69Ul}p9)e1aW89_yDK2AL~pTRc+lX&U2vv4 z=R_@-jI#ono6j0TJa24^f{bQ9-(SRF;-^uI7ECj)?~h$c`m9a0Bcq2r70^RkV|+!u z&T}fd=W_7}ZU>I2H=BOx8#AStCo~&cq1MbCWPa}G4Hl=(5I74VFBug&6EN?d9h_<^ z{sNvjSYlC0)ZJwrIiS!?(`wVP5+2Ohm%wBB;Glvq8a(yZ$&*>&X-lb?n*vP{I(Y7? znk~O9EZNo|=aqV80*<4!!G9H6N9GTtWfs7wV=lcs(bc3l6OS+V8q5bHyA$D7o*X44O~qoE zA(PV<;VrVNfjLI25|NKCchrM8qK{!>= z(Nou9Ok&m_)+#{YLg15ciUNd5)sI`jjgRfdcoqIpfE@GUczPU+KN2_{2t^{zsN1xm=}+1N&zzVp_I#zE<&nfU)3sPMg=(9Pl-N1O;z>#mgA`f zk-3AigqZ~sI|2CJ^+M-GzqX@gP;pgc5(yED)ymCK@3$EXM)5|X4omNbaEb)!z~1gI zWl&@8NHXUP?J+@nJr3PIqi-M2n5ou*viNH#P`QaBJ;7?8%9nf$dxW9e^Kv_0mY!}@ zE*&^v{6kSQG00<&QFLSQjaPRZMxI2rK(@YHo$gJZv#PGgiKI`O`rb1p0!)6@iFi!4 z?`GCFpRtoVosAuvPp8p_*u0S%gmvDQzVLS{Ul8i46U%S?00@}{(}~bB1YrFkZgG|n z01gEk{u?2qX<1-yeYZ|EgTgY)OYG6~M zYF^D|4@;kPw5-o#iq(arg;e*R9IAlz20&~K7s2>fa;fRbnPPgejpsZ_{!)+TuTe5e z(wg)J1Y&am02=Gc0Wlh`7Kj04qVEK9`(tg{4N!1*LQm2IggRa2AIr?5&x&Ti#3gm2 zEboL$al-w;RckYiEy>XsGa8$X^4gOnxF*g$Hi~57CxPZ}ZQ%$&&^G z_`o=B-EG48@dMn^JWRv#IX|$sv(7w4(t4scQBlTna~J=Tv`?OCr1ie~*#zppLhjDz|BK@R2@V(A+dwSIZ=EJ{)Z9>A{BMt@&c)BrI>E-vti(4fFN zt;(W)*IczX5Rm*mIvE;1w_LOZ+0c-w&G+c;YOUO}LYDe^=FN@CBM1n;4@NaW?%L(= z$s{G#gh!emXjR@63XKG@B%?;C@C;`z2Yu&3GdZ9-|3!-;SuZ-^Wnc$Jn!j*jnWiBB zLt44$uPC*}meQb$8>L7ThV1)$GbfjRga}Y$iSca<8b1C zJ{~KS6T^9N@Adu#lmbInDg#|^xUlWs?_1ZAV&gHD?3x?!s}bJac>LJWj_>1K_8)hAk=A{l>5>?<>>VO$ z;n^h7tjUB@qlO7|{NV0OJm?;s2??+o$`^97O#Bb|DSoy-S`U=NAJq6E4mk_-+IDKH zuV(|*%DXy8KiR2`>B-rw){ef}mCC~86nuNXK1%z?d$Tk4IwDBANlKAgEmwCpQI9`7 z_zrVTG=`ACPqcvEq3tTe=}?$b?$2HDjM!sQJd$YCh~bJz3z~Q7ifHR6RHu#wc-d9w zTCvP9J4xUtP9wlKGxTMImg+0bWz{xa^l~M%FG^;CjWe{EEjvX;gZbMVPD~j7cwl#N zJh{DA_$d~+LLffl1Qe}|skUZ7>|UAP2SgHFi>pR35oi6F^&Ml z802w|bFCRR5fHcn_byP}T1O(0MRl-=a#q32Xrl|Fvpr)z=0QklBZ6IPB^BvOpq&ae zoEG*#YZGqstYwsye{d-^k+cdHwrPf25%`Qa`ZU{#gvQNU)`1W#X7l(|l&cy6}MOUE@s$J}e@5U?z0<>^ip*hItK@+#wZkonUR znM(#ZjxB(aOp=T!pwAQbx$qpi3L4rE9xg^X}ULzZ(%(`3H_~CcoJ4hWd zQV5aUSq&EMXz-WVSIug+qFi*ibKTCM1yL1Hy367-<(!-%9|6o9}p1wqP z0ST-C8sv;KiYq0e+%+qI+$$pWn*60^obi$N;bzQbNg-2&D+7ayRg*PU6>K`|vFbf! zw!EUgFQiYFM(a9ec*u9i_!~1+7D7t*KIN~`GouIWuFJ62udVFMgRpGtG>S2o~Ov{-j{ zYJ>%e=ae__;rl0X-AW|uJuL|7ktA+-k0^COf0T8CC6h%m15Os{Sn+Oj1~s-g9Yb{O zSNeWeMtYnVj;2vp@}6}jSNofqv))$S+wOeBv{7MUKeL8)t5yMTO5dVc1DLv*zxme? z{{#FRKJFyB+~`WdgfS;O^y($6H3A}C@4K<%6I)ZYSmD14)PKmxj4T{X|Er8VqoEkT z#RlVdR|f%wat|13uIXgs-X4y+K7V6KKPBA&7fPo#i-$uQebaTi7mq?Y*4uegmtrQdUOlU;fw2K{DN|6NYDNH&qpw5f8T*VpSsz~6GEmQt!QBp=0(diTIgoo>h}c~FCdzoa2waAyyRKzaZvL&0gLiAPwkzs@1GemYmNh9@FW z2XnOUftX@X)9|i~Wdc^D^AULSLZp;LEH@gS%!am+QWC0Z>Iv_7RglEsk!)AYXeOTW zSssHejW2L=lP+H^7AH}3?#Csv`YYE?J*_1Hkn?%8eX)JMi;x}{qT6b9PKhWUfSAP^ zV?Vj{76Wc}d9`)yTb-B-0Bp6^DCeH>FpLpC3TJnohDU4M>v#aM?O;;>_ZFm}j;G?!od|%er&s?0pi(9W%$i z+|2Ehrch<^w_ahbaZX);&hnkWVKMU_!{G#f$bEJEedR#f)amkVIB@@ms|SqWBM+$S zTOO2ccC`c>`o@K@9GX>EpjQ*-bC^;o_x%!f=omykSviX&O9;y<^i-VB_h$MnoLpI& z#}=C$C!(}xZyWRaZ!xW^$iM%LpOpcK4JPLlo`Ol+zs1f!X$jn`zFJ-pq#D&cgut%D z`#SdqHo&T#JqRo1$QT!b`R}^f1vrm~KUCN>-L2qnu`LEmNHG;?$EF|%Vu<{2bORNZ z5Us|}i8}5K?w1ra);oA?@t;C}ddqCMx7XJ1$9X_ z7VKHilr%$7Vmhs6ln%mxb3qG=AwSD%2)4za&Cb*9&iA`{KqeO`iMU|RqrA+yU=mbR z&I4R+Zm;r=$EIUV*k6uT^#uwH}!badd-KB$^<(5_bEA)4~Q11aoJ6 zHSKzkZaY9cpTJ*)b!1SU@3u7$Q2qqtJ5m%485GiSON)qf9>RlOSD;~6sh~Zu;c7p( zk!{x^F?k&~{MjL%;rVn>`WO|1MCD~uc(_4+eJMPO;)Cm_=d;OUmw**-v40ho{vyB@ ze7poqgJR?+P)oSXMWezgE5b^!s&7&z0v^KZ7AwI&;J)O*(cHul?z>~Ya)8@|sMyKP zKX++m=9Pz6-v!!^BeRbeTR2}&r#ol)^Y0$7*T}njBGD?Ld?cn&_C{#?ASOHrWL=Bdfh}IHP3_(7(JUL4 zCc7q;R9oJ48f9nc94{I=S}7GWp}LrBB_c5}VrREArP=6Y`q)WQeTj)M9G1(rO&1wK69sW!^#?lJSBgk~X7t-Ac-{D3q;)9K(M zce}_mr_SS~BDLn(Os>XRW2C45QA-M%eO`DM;`>;17}+D3)Faqh=Vr=};ky+`_GxuT zVQcV71t6!>a2-_#=jsfWeQ8D8Q90I8$t*E}vnpV93R(AI)WL*Jy_zkH#7k$31WVYT zS)RN{D~pz^v8?Xh7i`foZ414&no6DN$=RWnmUzXXz9OWPZiXz_@whh8BEyl!zs6v; zZ>#y_ zwhL*XEAcT4=tHXvx=EkS#O8ODBtm{)~bj{San~t3;$Vc7<=lgh08uud!39n4F zE?E_Z%K~`*2H;aB3F@?$iwG$Jsn;IF*>y+DPE=q{*M9_z^Ka-Zb$^MVH zd4J)85G1hZ7^i0gxjVKj*)TX_#3RdtnBH~WjmN{d8#!1l7)Dlvv6xqf4mw;`UizQc z9`jYF`Ok8Q8LcVGjvRNfYTDLUShqTG$6Z`VF`){5EeXZ!{B$Br zvEff>33QssV{g_GM=0#E^$M=$TGPSYlDb;h{of6j^ivBWLh-}0>UM8~Si*HLas;5X zOYVpRR2!hh3l7+`EC87k6DKI8)AKcDDof-vN?bNND~9nxIXj(_EDBEnrq)qq8%Y zvv7@VKG{0N0HM@OYyf75Al(I@VFmM+HpNP-_c79TB*_QVM8rWY@0#7hR|8li&D}o4 zwgs({ zUZC&lqU@4g@lz%y)6kvzlf4V{N%$Lkmw>cZy|iV33-(sp9S2iT*x8I-xOKi|_~bY; zGIx%}`y;_0pr97n`hJmBaW_)s4$A+^YVh-xoYL6-*lBz~TngjhWJ>raQNdS|gX2nO z8?25|yo!0#ns9zNtS%d}q?_lHp04879jP99xz1jS0S?WUjQ)h8tagw(YtDn1Gg%mAi8A;(R@onT-Pnl})91m^(~JLo zml3HZlvKrJcJKq^MoYg_CWIS7+hi?D@TRTH|g(8D;*fS zML)V$$HsjyH^@}Fjj6h$1%2TW_r{B`2If;_f@zIKybkf<@DeZlKdd1M4!= zMw#NkEW)`0SLsvQnkG7SUnBlt7^JYk)l}@l`Bfv2jDc%T7b1_$33o0a%efYgIhI9P zQ*%FYAl~F&jMHqRHstb*-P59g%yE19nglN13Q|V(i}j7%laxrd5Un1FMPAB)+@&g_ObNRKr(F4ojnl+IS-7a_0sYpc z?cim_S_+RXzg$m+xR#OiC~>r?;R-{Z=hssiT zA{>tTu(hOk{icq}L3mAy6~JDVSjpvTI+gGCx->iJSYh=%d$QjA=y$NYpI6kM?^M!z z=Df}6be3>!@AUGl{m`?P{KAnNS%K?+>SuOTOaGp$X8qM*h#kG#R+zbMDdJ|EVa;=Z(PEq7!}U&RNJ=!zTYQhrC?1x&mgCepnrs%tFo05#&LnQZv_10|L6Uf3X? zKR{CeP?A*UJsopZDL*S_k{?c~12>XR7voS`K(WLij+RD{Klfp_%~zrz51c^~8?+B& znnr=E4Y{!iH#)&^XWcW$T9)2&zWA^Mh_1gZ#D@pAQI4@Zw*OJd0W%#F@QW?Ox72Oz zA|`Nr921+R>z;Ybx!%%z9Z?^eX3LgsLkDAvftC zF{o3X$r>8_fqHrLbxlBYw9xA6Tv{}!a^eTETo|SW_w0wrzJ~tQZ?O0wkL!!TCY@1j2ZPBt^cm9&j(zWx<*JlT1?Ow zUMMz@XajO)jV=zt9MkX^M8xWEpl!ya^|(pFBz<5T>ie1JAO-C_0qCvI>#bXw8|u&# z4W`@An`AAWD}ddygFt_#8L*>UC)6r-Gz#RCH zuU1BYPLpvld+!Ubgb#r#g5gVz)kyvcDi5V713g`|A&8c|Bg#Xey9vw>jvSpR$-pos znn>SDAA%hSy$JYYSS$DU+A}&63`IRK**^~VY|&5 zFSk#+RkPy&>|&ftx=6_A!4$A7-oN|LRwi}lWDaC_;*jvu2V`{Z!y>wdG7qbG>BjQk z(o>;Rq7e>x0JM3=e@KsfjKeBE>EiyJwdQkceawmIz6pAkEP%LjmqBm>*2NqydE&?p z9}`6%SYoCdaDbX>G*?!V<50VscRl=kz*}#&vR( zAyTxV9GyF8@B{RX5J6?$4C{sKPmt69Z2iTO=fb5Vr|`VWD9uaP6xYeWJNJlmL5b}X zgE^+T8?+*Jyc=l0Z|GXe#mdpo<79n!Bk|IxF$V9I27JQCx%60d8QnMOa2`4R7wH&V zRcRV^kI7+%ue+Z(W-?OpftdT(8M&-FIkLVOR-H6e-VxEGlr}cF=0Z~mWJc>`K>b{K ziQ#zV`PV+b3x9h~u_QmyHGZDPUL5BJh+Wd-{!Ef-OJ~L`wkPwE6vuYo=+U~@zd-64 zx!qMtun9Oi6B6YnIU$&8F{owSbjeym|GN;|y(RZbBte7VG!d^lug!^#HAQTdv`}fC z9b1jLc+LDppM-OYk}*gd|Nd?fT$BMw_Z3>}n~@jsiW1NFqp#jEk|s+2J>U@XZOQ(y zMnG`%f=vV~Od`Qf^h_muo>F?@&p`GSI#dC__Dvh@RUhYsX~e-rR^7{%6Kt0K4|1 z_FM4$-5wfPjBPmmOQf-_!fk9IHq(0S=q`OT)nC1;V19X+61UJqr2z^7Kr0YKntKUJ z(4|Pr8eoKHEtcL5s*Q5vO+(noYBASyg6!n=iMWp>LkqRA1e zVnOgQ5hu{fP|u~CM(vfg*}&?)57Z`j1r!E=%B0e2`mM%A`#PDScZ1Mw_l^}zjw?l&oK}6YB&s0WCSTDFdd_ z)^dyhjmbtC^vQRI2!s>t>^4Nr<177k8QDm8vU|C(l+;vQ0Z??xk4ek&G~l?P$r_)b zp2@|K53azUMl{mcbfDYZ5gIU_AhP`q#pz(v+ z*L!++KePU$Z5Q38z6>ex)xCylQVbdK|8T4AmEa%GP0JO+SdQ`^zZ8C_4!0sZ1ke}~ z%qA`6O@7A`%S#|G-+D@1GS&TP9A}4(-rG|s66)q>CJ#i?59w}+D6KnRUtcG3Vf*sF z|0RCd{zLpQ|L;?X81>741M2@vKN{Ba17NCOl1Mfcp}fbptO@}&foAeNMM@!>(`=Yh zMQQPLG1T~dO-D5xsbm@kpyChibNewpt-jLz-0KwB9Z)OR%i7(zv1N3OGvbPBI6CQV zf3eB0TjGvNg7dYRe%G#C=~zvgHp?>NT2F2;tE{_L>fP=F7QT_i4hmX@L>FYaAca{J zq|a{pHp{YU=yUzFm|n(;vD;E=uYr8$qa^O&#mKLR_SW#Fr-nJ7Q^K9;YM#T{mQmsH zRjp7$!^lZDx-h3n(SC)Fyf|)L3x7;2e<&XtC{WA=7S&EL{qbmg!MBajGe2;mttN~q%5=wVyN zrM|%7Hd|w*loYCdXtybe)s?4^8k8Q^ReYHY6pVmVso=?QUuXR!O^tvuN}16X+c~#x zG*PiftKGLduV7a+Kr~Z2#I;#o`v$yTbUNdb2uSCmd>{q3698Zn3 zTqvQq3~#@};EMFg-=uaN_cqWittq0BLb|bDG$1HVS%FgeKp~|x=w7V=Nvlh|h?{eO z9!(Q8tO5C)ST|krZ}^cVtw&(uDTPP)GZF00O&8deaDysL``=?V#3Q9j`6Pe=Oi&K* zuvlnWwv^Jm{ooJ=_F9E#dI`BATsRa`j*XqhPXSne6xd~r-e0=Fu#J? zrjl8bF-VPVKr*Z_N(0Y%WN=Zrdf^TqP=&};&a`^*P>>+-BQ9PKWukW z65yPRfKGZzC}*U8ab*4wNW?%}8*SCeh?ob_j(V7M?q)y#HHtKAkD8q7Bgd!-r^VnU z=A!=$q}5ve_fP%?0HD}q3bh^*YAAiq1qql*zJLJX1mkht#;&?D9k9X@&IEXvBKALF*w!@I9RJo~!zmXTLXPb{NahisL3$<@!;z<{x?eZr zp8!?6_h2KSGy;a7h&Ti8ogqbH{ln&1zWP^{Hv!r4?05HOq+8XsI=fU1*gNbb5x{XP zNu4mDN^Bb6T1k?q@8B)i6U~H}O@*Y;Yp)qt972j3(B}S0D7$Et6tH!kruA>cwKvOz z={ITY|^F|n_RaQveg9{1aD8-H|eJpIfe1j3;lK=ULTe#LeYH_CBEM#EM< zFs_5ZhukqD{mtWTqMfg{z(>#f`Q|HyEJ`^2on6%LTl3C6o)8lZ#!=*+ba3!i9oBuA zAzV-JpQhc(`uU2sj*l4AE9WK?$!_RlqJ9Pn};&!Qr%ZrzVi?3ZA{6t_TTtR zH#iQN^j2MKQwgGVxw7t?g>*CpxAU&UCcfbD0omh6YvgCM-`K|?J|S)yMJ-g0lp$qg z)Tw(t=NZ$uaUA?)tzyeh*af-3F>%(_V39kLENUT#IN3#?6RvdD@8q1?%tdN2lHaTs z@~xr-4JTPQJ#A5P}8pw}8Kc&^`^m2g4$cmAuClRYS{5z-$xLS30$ zZN2=|`*G?l+sW?0#(Sx**4iNblO4r4|cX(a=Ygy+q6NL@#PqCptY=&jk`oe5WUiSSv zH1S-iX|@x@E~q352S^Sq3-x@MbX#4-(TTYB++H?VwYGLvQC@I!E$(fTG0t;Zy?@^H zw(9MEf80G<7RvZDvTcB^j8K1r`dA4ij={ zh0M{(y(MaX__!$T&erW{(66$TpeW?Kn>_o7j!I2(iFexC}iPOB{ZRAay>j zscmRHz1cL}RqK+_hzbtc)fcncIJoSLS!R$9=R8g*VmYM^s?&%r@p0lE0K`fzhpns< zLIeaZ!pfxMN)v$&g^tnGgFC(dUE!$I#K;w&IkK(`F*8 zKkXS{AZ+csU=TtmXpkMU7p#_nRa}iVCUEVXZlH$20s#x0PypoE#mdBjz#QeaNrL#7T*DiQrZAH3)+;eHyuPM zoLs4s-Dh?z#fc1sHBkVix6tseK=~iAo4T*Z9vwJDY`ByOSQwh+ zl3`MCod=yenbdAy&8b1k*8C?)+f1H2=PItpdn77ntbm%J0S%70lc=QZLTNx^`^V8EdOe&;F|8zscSXGGLv8 z_z>{HQRxMQCz-BoDi6YDPyZK^sS^pTE6X4nhdpO8)f-x9XPSQPD$P33^!YFLrQ5nL zpc$YtpW*QzyxD@zCF6wSM*7tb9p)mj%#N6`@gKNE_Sf_}9x(FK?0LaHhjHEX zJ-k63J#vF1^_FVpmO!rr-t!RozWzrk#L}5f{`WLoQxWe+*Y4DhSiopj32Id)F`tlP z9Px;Oiv5L{dN^#IPK1EQ01?cWEZzilrAj8ZPK6=0wiRb zgs~S?GtIi>4#$%x7hrn&Y0ICEV+JZ)tojVd8Aj94SYv%$!q%f@1e?QUo6pfwr;Oft z9e))cr}=)F$NF2SejB&5P11A7HhK8&$!j80YvaJ7fO;eFYuyKt*hU2yH%g@9BxfK@ z^C48#bfU`Qd?^~onX5Rr_uD93=jq|~-aP~1`5xa%{wgL=!+A&VH1>}X?{<52%Z_}75DQ;W%pHGKY&vt(X$$P>uSJ|9q;q*4?j-V0lTjrV zX|qCsjH`@Aw-*wa?85F8^TM#hpH z0a{M74`jc0+s}*Qc{sYjTD)KTnKZ!lz%4Az?i*U7*l7LlPa>Nu`0Wo^o>7msAom^L zt72&X`^N*w$ND1zLy-Rbzr7k|ed{3HaQrXuwgv}5AHO_>0g3-5asI==V_;+ZpIJRQ z>K1mJtQbE#|5D%-1n@1Qk_EA7KXnFFD?s0=fK&?w|5vY4LA;1aK9;-dC7*gssli47 zQTxh^r}*>yu{?V_U`<$@`v;Y}#x08ZAq5IuqQE;LIO@{d;qbVHS;gwI!l&`e@!Nn9 zxIklaEUsW}KD~Non#@}Y#n5vNYkOsxY;W)x=eOIos)uGO8)}%OvtVW0u^Q-jo5j@X z*1D9d?xImXaJg1}zG7~>A`kf86$M?IMb@}$9=>H;ySkR|+lA$FAi3k(?loqbO|9w< zR!ap-=RSIxW3z|T?Rx-0%jHzNkc)2f)_H5iCne~d+xK|e10P3B=SWYW)Z)gaK0!&6 zCI5pZ2W7U-F*xCr*^7{Z79U+;f4dUh z@^N~PxTtB7eO!-(?y`&SA_gWn9X5~H32^H9P-Jal_1zD0%LHKtm~B(G{pP8cwQ}mf zec01DHF}ZxuZvHjuyHDtK5`+5cIhGKunU>4AjqE-;<%zr*dr1E>~^TYC5K?a8s#wo z02HYi1#4Uiy-19arP)M(k#B>{Vx+$s9I=o$Aq()yK-ar&pD}NNeti48*^HRb_|ELA z(0|}ePF%o)l9CnswK_ZY$L+vi*ArQeX<(-pJ536>oi?93Jd7EqTAyQP9giNnYTy{_ z{wOT+yRp=T)T0-%lkJ3{vDYuRoN0N<*$)ao$HuLJuZ@Wn$9wSx10l$%19S-+@j+v< z*o<;FhH}P82Oz1kPj?A2X>qgIqha{dBCFd0(SQLa#P8heLo=nVG2aETV>UP;d$%Xh z9Wjo7QOA(AgXrM1sLZ_FJt1;&Op&C#BB!z$i;*sLmdKlkD{HxCLc(uA?jy|c{X?VF zgAk1&mC}P*rDS_{g>3<7ptO7U6_IkRA$9cS%yk0u+{6=s7-^_&=O1(MW61$TYDRk$ zaX&}|O_|2mP@%Ez2S+V+4SQVSO1n5-Z)G=(=m-z3M+X#%gW&~|&J0ADavMy$1DQ8# zkIG~FslMBLl6{kaH{uoPT0BnNz~eHlrV#HPdjk>yqBj(}dJv*L+K*z)I|e^;h!-UF zdU?gEiK^)fkPr%ZrN=vJD^`-9#mB1?&Gp^79ilDb;S$~Zdbw4^R}ZTiBGYS0oUG5> zFZm=e30Z*g*@FsPg=QFR)Z-htwLRqcbhazI+${_~9#v_+RuPok9V8qbGkxdC>v3U$ zgh*ucms2$cBMvp&rwekN+qG;wCO8oqyt{Z$YwM7n~DrYPsXY-zrecQdra!z>KXVQjh9+1RFCeV@~wV~v+NfB?~rA35KF?V!J7IuG^* zFXRV1hi61@wH8-C*`_IkftqJDC0G(sfTEoBg7yM1;Xf74zti**WJ#v)CUgiRz#$(L zgR0H@eBB0o%qkJ`?f>0uq5YVtE~_xb0Th!G@7oe0c^07@75=PsbIO2AsVzXp?4|CT zpNt~Jr3($NWm-Ehz2-bjcZFrctStk{NNG8ziIlqFL~yd3M(m2+i?`!h<2u)=DOS73 z^HpRxrB`(sebq+#a6r)J&VYYLSO<6Kf{OaMLh`W+)D zc4t_4MFC_lv&6F*hSV4Chb=|4#K~Mi<;yrgD01Q)-&ZEGP4AC2;`trIRh<9G?ibHC zlb@qPor@ax(=ZI!4*_6lBG-V}#M0zkSiZ+Od~$cQ%qYaJ`n7_+%~}bS69P5o}u+cDwplz z0%uEN^hzqJP6R4Ur_{1S%o^pZCbEcxrp=Uw9=lBgS{mnVd$YqX1N`YB3p;MKweOL0 zI>Qt^i2VBB8fUmJmDFq<>hyAwaVciUgTf$uRog*W11(;p*8Ewxg*@%33n z5`bFP2frUhRKHg#RJ3Y~I`wlmcxGFc%z8&CIlV^&5DrMek}ni{UtD9~%o1u|>4c}{ z+tu11K$?9zg8$k>|HB7jU}FED%ScnyY@!ZX5q|#a8K9wc)hUop)@nVNSV-OGg)Dgg zaT_fdk~~1#)aW&~`}HwnZZIO>If5d@iV=N$aA(Hip)=$ACPNm3NCgQvrDluWgMYGfCt!pv8vh^MQ1BB&7zYI{qOACNkyKB zi(1R+`YTuGuJ)h{zyy04AghqHL;1Awc5VdYLin4kf)NWY0+^BlUUSS>vSX$ zRm;##l}7sh{U6e7AEB26x1EBTKFG^9oKbFu|TUx zKfC1`Fiw@sOn1d#>7g5^O?9KJ9dJYM;{iTPL)SX>upLp`ibG?6lEM~aINnZx<6;Ti z7z2GWcbqGObVc_YY>#M16o6E=o10}o1-4DMOkpm+@R;dcCULbx<9y)cmwmB?CBu__iiM=Y>WNK+9r@3N$jnXd+Ue8}2 z{0O78g1GNto89SEuyQzI0=J7b3!6+zt!8`oGKAPn;c+7}9>5>W-aIY-^7Y!u^-Gxc zEyP%9q*VL!F#=;x}t~?KLrgu$+CRW zI%k2J(ojVyn8af;Tx@sbC+?VFrm(Sd-$Z;oYLCpvCFo*gjF%^+1-8n7d@>dhdQf(# zk~tl=pE!XF=5&d%5~TV&7C0vIHBn&S=s}XvN!rDsfk+DorI$ZdO^{Hc5<%TX@k4)s zpp{kFiuR{)nJs8G7-koUHgB-3`NJ2lC?lPqBeP;lOJ%vW86d#K!-@jY1K+td`6~#9Xom-8zy68gWG#^WGM|N${QPa=q{M#2`_{Pubh^5yWKfXTk}_ z+z&ZA2i~ral60yINwx*LJ@(`0#J;-T+dj_0qw>zAKxiO@kN|@vWtF9)k}2n|KZzBg zC}os{1u`mHY883QfBq4i$$h?bRn(3+c1|R#zw2 z&O)w)53l0m2)$RSU#ijlmj2;Vfo7dtN|+Ua%9bYdOxU0>UV-Xb-0(9@dv8pCr$DY8 z#n_3Gse`EmF&pWhvV}N4zf!m$aRRfZ;#OzY>r!K93+O>A?0+)9IX><6q=}+4@I6G; z5H>6e=w&_7-Kj5Zly)CdvkPkNH>-Hnnp|dFOnG=$Uz#@QNZfKWx$~14mo;LvB%l*u zzt(K*Q}7l`#w*FGbQj&IZ7d}~;!n;>s&ZA5sB%_Epl)s?^kh6OW>MS;4ApR5Hsu%xhaSXU~(6_ zId&Dj*fKeE%A!)x!HwCGBz9O%3=RgkrHL3|--hG&(H8}lIHah&QZ-KKv_WJl?Ud0g zb|)%TE1`bSr~q~IYjc+MGK(#_9@?a-Y~*Hq5XinLR19lM_V&`2a2PcMa$zd_-2C_! zVrRw**RFe2s;@;Q_U-0GCv_HAH3>eAGhKi{PafOc_)02-0}tX*>fJ;HC%a|H7f9cq zjhDyg(pUJ=5H4N}w%~JjekO$Fak4SZ@R#GST%PM(ccBVD;e8?>4c+8lUY=X~mWS z47uA>`O=M!=Ma3YH2lZ_BunPG;nE>&=Wsri2(0Nu2GX92Nr+tV@&gR2!WCzo^FmdT zAUu{OVUQl8Zi3^1Od#zPxLp2_+yYEGB76Zgm)3ygm$>+3Q;tc2<;h7xLx^i11Wj{1 z!6M1Bu%v6P=F%*Y-niL>k zrbnz@^7xuG{A*b-r4qFU?8T$oul6+UNuGiSHvBc&AFlaRy_UQp3C25QMz{sHpH%{i z>3*|+EMJsW{wYGBL0=Sdf;;DWWmU<$NI>j0#=Ind+{T_W)xf(C!#c($e5QN&2ALk? zn_aJz!*JL!TrkhNCid}Hlp;6Rkp~69jqdQDE8Y27Pi3tuf^+Y;ug9ISX{3fr+wpM; zHCNk7@~wA|BLLv%W;0xgp|FxNV>?^zuIe`J%HsiYVhj9EsSQ|T^AAiwe6#=QT!96_ z7ASi=?^x#(p-S8Tx48Q)h9St=rtYitaoO_q{MKUW8Z*pcr~+RGUl=xZD`!eiW~%G; z&^9$-DeAIjRO9H-D0^KT@xNASrvEVFnCSoKn0d9jhTXq94&T>W zbLu*OtaMZN$R!ZpbQ#9+6B>0pe|8U#Y&Z!8VnwPU}Z0XHFHFjbr>-mHFn@T_2ydn`q{F z2lDdyn;|#9ogdfN*=bqv&}>&hrfMTiU1i+)>&uHtr@ySdj)UYj0#>W3FUt70mrg1D zJ5Rq0abY&^Au28zJ^|a0Ct8|2ksQ?;b703W78z<98*c(gTsG&w-+n*wsnCd}6Tb0< z9ncx@CONYDoGU7?YjoFKYW@$#R^smwlk)Q&#Xr8zfPD!*e(tbA5lEBQW*t@g`nq-Q zRkL{AcYY#X|A=m}Mwgsk7e+ZblOL*Nshv@TN zs8b+}JWW?~NZ!^J^_X$f#^#dkDAQBIkDCthlCX&eRCu0ilA(Al0QzS{QsGIGLr(AC zRrG^UyC}9-MoN+Sw^`;9x{?s%ro_}q1LiQ}T>Tu2NYbLf_~hpB8P;xseFylJ)kyQL zHdO^p)D73N8bLac+fF-*y=7;)3-C0cTAT>1BA$9D22Kpq;Qr}I(o_!m0fHH?g}QucncaWmemwQk}&B=YQu%Ol|YD_%{r0!429=S zmzf7YTOQ5%xyRRdVD^m^@XT=@NsZUCPN+T==4I}H`FuSDR5oi-80@>H0nszlP6mB$ zBEH_ zZFrd<;?+*i%@y}!xs+Iag#*9!=9>kc5Z2>wc_Dx`(~m4X|G0;!EDZ_{MeL%yLkuf) zG6;W4$ijUTF|ky(6%K_@7jUBB;&$Rt1N+z+d&t`1!RNZYtEnsXc9=Y%E zlFcI`>uv?Wu+9bG1ebZF8U~y9JYG)eZBK^d{S9^hWx&Wl7>tnA!Rp!Po37xVmj`0g zL+X&_2H}L4F;pEjAEyU_jM@A;JKw)W``CB+(SCv1;MJH7WAy z?3V^aZ0I|D9Nb?{;y`rZ0qp5w!}s6CfswP|wt30)U**{vW6V4LxaJc8fIFD_3cGLZ zv8X-*aJ8LO)8@Tva+946^r7fnz^w~lhsbKy7!r|+yDp=-5T(zJcj^})PiLb!NmJleMnsOH47$f}db9_SgwK?I zBIUf&SqI01rRB0%$uqczcN@Y=4kkzeF9rU&8);P!T9C&uOja^CxW7!0U?h~L!@=vX z5MUy%*g`sU9^b`&Ta8#GJEbo)@+Pi-*9#qssXjIv@?Q$~|L zsG_X920PyPirMaW(ufHYrY9Ip)1NUH=ExPD@OVpd{pDuOs zpf|&;9dVfP|2Bs!&d)$8ot2vWD&jFnB(-;(9ca3y+@4(=MT-4@NQK3q2Ug{+{TQZ6jZp25gDECuQ)GZ%6;Wy!F)Om4tX1Cr&G|0 z^AM)N4gy$c+#$ZIbQsZ*SNBTz0$F|MTcIL&8iVD$MJv*j9qY1SWlY*G9z|yy2E!Dl zY%BKAU%^pF_Vt4=OOkn?%FaKwbRV1CMT<}AwMXG=^nBI-Y{jh6~ii}6$ zvmzQlo4c55$xL=;%pz zx&gJ)`^^zA&$9wg?qS#C$E&>z&vX*N_tsj71`G0m*>+YmUAT0 zD(}2@TE7j4Y!FfnI__`u9jy@Qd^y;RM;)-lJ9mdWCHng8*Xps& zv&o%BTBJ9L3slDT(QR({>6vM2zLzDw=8EAa_R+@7>U*Rcg2^F*wQ6EfKjA;%kQ??( z|0Ska|9>&X#QeWoe$`l$v4?H47Zr{JxhV>o>u3lXp{Sn-L^sz(xs^_ddsINX#1%kQ`dw91{ zE;}HDyj_2O-r8!}fZQ&X)Zpvo*bVkn(LWFI{^8SoyFWiE^|sN-m`48X2!zQY4U0wKM>F>8>_GC7no>^m&_hgk-=4HP$ zT z75o>5>y8gYXWHET+CL)Jao!(Gx)UiUk8KN{lJ&y5lWre80FHr!|JRYC%PHk=W62%C zJ|kQ(S~**$(mtE4VDV08pl4V6ZWKK(xxS?J?>Q^2F^#bVx-|&o)HGg2h2+wE#9Djn zAvMZto3A$)v5WOa#)iI9bPdqFKylu|SU|0V%X_Vmluf(J#}Zu+$uo|wAAsRn7QJO8 z3LJ`&7G;4UNCe9sXAtE5BkkC3vQ}><5qeS=B}JZRlc5422~-F92@1(lGZxKS$+%jl z@ZExJ&q|dX(yn|GgJ>Hv4FIFX4yn;cswdTvom(A#0<_I9%55 zK?Q%<@uN{v89>NHxH>(-iYmV)=vLa@DLFH|W)PP-Q7gfNM!Z^O|JAwLl&(UibZs^$ zvGMn2wjiV=90hw?i{UdI!UT1;72wN!?8l35geL-*l^&t4MOL|;VAlfgquNkdDd|B9 zP`yNQ%*)Y$fGL*u*lPJIM@8x{+0!MEN@eEaFH4%zU(z6JF7bKhhI!*jY+IfoGS}x& zt|Au%!~j^#XUMUj2(ypmY;JnjYz!|{P?NjFwwfKsGz-!v6k$UxZ)n5#P-5zk7DlPa zP4}^ry=^dOd*&VyJ~mru!h6W`+OY1bIj9q7m58Ew)*qKYnk3-Y)orqT5iQ+0pjC6E zHv^vuFe00)Leu?^>>|YBz`Y?_;%(koBM~w7Fr5TBESm+uYm~tIwo#wF%RbuK@n%e` zjWpH>OD(*0gv03zmCs?=dI+^m3hB9hYY6xS0APlAb__4(oUj?v4c7F$q28i$POu_F zeO6KV*lyTdZVf&l z7-)%>M;-@JjAlIqSn>K=@~}wjmfTJT=!0O~4Q3IB5h@;p4LTX;n8c3W1L3_$Kq`XaAV%q_6kF$k{%po3?ng+8Wpm(w=P&^ST}F% zEq5D33M7Cx0jlk}0$?zYgCN8gM5EZSO9Dj2FYdM>_LO+${vT!k6eL;?whh8l7V~fqg6BfE%nRya zJn(R4xM6^6T=dspr)?GO$g_=`F*SRO55GD}N&cVHL;#Zp5x4-Z94(j>EYHc{h|t9}$- z66=%7-ehx~>?pdEteqz7cL%~Pry@|jEn^)a-=g-HM?|9U8$s|s0IvvH!-dum7T0m; z<8~NfBnYg%RW4*ESBf@DCQxR%O)XDB^>HZ8DCkyTwNB#oWC~5IvRNisa4VqS96q+E zQO_BgsYf^mRc?U+6IYBxy%1!nh3$B$nnJZqKgin6j z^947`4ba#5-G3C#iLk$rv4h-nElrpQnFT0UAg2^nk5Bf|UHhs*QdoqOO$3Py5DcZy zO}#?nFJO&|j})y(?sJh-O9aeDILQEdKUU>#Fd{?hP4)a_^W&B14s$yKm$a*+!7^`l z()=*9&o<|@F^}JcLa5wg4@v`pGSf{K+W=MSB|4f({S-K0z)qi4j_B0|>6h8oTl+Bs zDULsU6*&6n$pv$VC5=E%rilX|JQbTB%2Pn*G^#3;oa4xyYWmTgV22Ctvur==LC#CZ z!AEKq4}(JGSVp%(kJ1C449{?Qum6q8Qf?qD8bW}dxVIvsF-cb7(&V;0tfdH`keYue zwo6$gCm&sa$z$cNBf;p~?m}E`An`+o&SGRtDqphzLqO&=N;y`~3XrzI8;I`_+NGU` zB83DvsVrxWMubR4?G>(2AB!L~#)q<3W*RnL7M&Mj27zE#9(BijftjZS1H&`6DxVJt zP`DUA3~`K2Qmvm!8c(1RpH#31UkIisfZKNAgBMTUFZ3!tQBObx_f_MMjCp^zRG96t zKnO@UwSPWUnB{b9KK>G_4^IO373-0QXsHAcdWPYW9?}&aV7-g~acuMztwn?phJtM| zuVV(sJjovEQLCaMq$HheM5@s|wywE0%r*M_C_rD;G@+AtENvD+6@Kbj6k;_e)YR2E z(C)M~wv%mH!pG%DF%GrH)0wNa;kCgqsj_~xH+ZBgpl2&(bIGO0n(y0Srm3ZZQhG0g^h@hgNq6;*>pq(CUzGB>oWN*?*9nt|aV*;u!91iJ2V-!#`B|%T}AXOJ9 zeFcVaLJN4PHulN2brt}mZo(@I#t*xfY;gWcirc5r-Ca8MgaJuA5V3Ocg6@2?;COpjUB^d&k#Pktb5bQ1+#N%!vD z;|*|szqH_~IjF1oA}E>jz%$!!@6v9{{s39cpuviq2_#wZANMd=q}JLGog~D!yMeu@ zZ=Rh($N7cv8}G-gd~slXuOR5N;YSSM+t+F7WNnE9O*5A6}_vB%^p| zz306cnDc^yScjg|u!;ripF9Hf)E$E+WyU&|V(1R%$Dw{z0AdI*nR^0OLV3}}e1U?2 zY}yu_n;gpx)ix-i0hPh);M5d?XnuD{3TdTu|ENbPHV_%g%Mv_Q|E7%uO3Io3I(zJ} z8?t15Yv;Qcqb-T12!H5={P#xfo7|rk3R3K2Uw%Xx?T6tE%9-bAI`B}bRwV^|6a`~`Ec@<;)P`?kz=AHZ^`bM$SMJBkxg)-Z|(ZeUS}A9P6Z zM?C^T(uSzpE$SYD@F^>iNW>+6XqAdX!L_G~A#@n7k&y99d|WlDKOmLK6a+Jl2w{?0 zCdyEPKk&6S6(#?b=W+Z$^E?JtHrD^GHp?*jxa|S!@@}1yx7N z$Gj$}^t%;r`*cb~4x3#H*?Cu*U2{{jJ(jAUO$Yn_x#rkS1Lyi$y*AWy9dK^6#rDE$ zX%XDmymm>eC0LAy4l9lk&opAwYGn0GaGu3Q#Fr2nq$kZ|=HlJu)T4tulLq30sk@2J zH(@O@!XyjJ_nFLhzVtE~8&zP!Vy)4i1y65374YLJb8;-U*sxHw{rVnm7il=wl}Zk^`DcKXgboBl@s1JhL+_>xtu5d9&G?>HHq=qE zZO7j2kI+4Fqf^IzTGAPdp~gzQNzNQ%>>nJH(Y|W4jI6V6YU>=Bl(pZ0i%qm^% z?;Mh0VYewvokF)4mnd%44x{4g69!Q<;aDNRE6l#UoVg&WfhE5MErXCSbEc9Wmx3uM z0F3Va>73?Gzs2KTroAn&YAd$yM=-c%sJz){=L_E*4pSv-8m`I6S-{+1VKN9WV7o=* z3v^`*w$k?u!7%Fn{S|(rv#uvEs3H)%G`f!ar-Zw-`a~TReDH<#qkDnAp2FyUs#TZ; zt(Unbt_n>E8&cxB>Z!t;c8i@}`W6#>QqH5t)-70ge#lZcW6KQi(ZOMsoP=DRAfL-! zL)&@{Mw|LRUX^-Ekj5VjI0AjC(h0&TeD zHK)MsAaoyZEKVN=6(K<<8sU}JcX)GQxb5mIT9m3OH3je~WR!?Du`afix7ZJ-+5M)2 z1h+j{bHGCuUjf3PpH>{BIy7X^!t)WIKZ+bq3HR*(e`Fy*GS<(-0wlVBMUf zEoNEzHnuo)>C2u)c!pFOLK=dlvpT^#ut8CLl?5#SyREgzWHSIzSSNy4+H_+@?$UF` za6;6IU`Yz2;=qxEvixAMYyh&^>U*}3uAEAv@B^9j*+S0;JH>XSRi`t43Xi`!VbjK< zk4ig$dUG`bX64G=yI&V~F7MGPb|0$VP6|fth=U_K(#c$2AM$L#Bz67w1oOr#%dIH# zg+Pz@)!&8!uep;^W?SRV=C^ulRCcs&h6;dH>WhZO8Ft+}To(d!cD2kgRTxv4F%r7= zqT&92e_Rg2GvF9fp?zBVti35DKAiuycNwlQ&W2_}wX(<7omxJ|oVo2{#r(st z$!joT_{PN;@OjzdF@MZrJ!7wOeXln{WXD#(TXS+JPPWiC9xm?)R&vOJ(P1=RQ5j3+7*xhJt0bVGJ8?=FSO{r^`m@iw z5vtJV>fadn4>otvmupqppgrcVipTk1E)PGoG7U2`M%rRbmyC6Tlz?O7zaDNq&Eoa@ z+jNqcvOr3K6sD@EMFn~+Po}O05-Ly}^0H_tD6B~j4iWN5VPj@)(2^-R2NGvsfBKr+ z!VE(DBA4ejkM9Sr`WZSFSz!46j0>I;^1q-&{vrbY?T&zqP>3-JG?(RLq1KX;7hu8m z7z-~LvxqWK#B!HzO;h1rZ%M1dab|Jiv3akQA#F1WkqPlNoVKCl4}O;M#Z5t8HS?J} zYeiuhnVy{qN`!MES&O81hAJ67!5SQ)I1w zF#fYvYUGDLV)kbOsW?(C-T#}+uwF7yo%XHP0?j#=Y00W4^L zH|mw09nTpzcC|ep1q=M)j@)}60)8$t)!}85B zVp(_>6YtA=`RJr$NcsakWtAnEIT&A7)J@e=zGry<90G4@WUw_uM~}truHNdM)_PhW zACrV8X9w`|PO(+?++HcO!@$!7g( zVxDlyNr#UbIG-oyJ8C;zjoX){jntA2n|DZp4?U&(>`hzSBDhWB*k2>@;^15j;^Bs- zO5_vwS_*l$J;Xn2IO*>APqD%-b|oHUTv@csqDc)d$+FZz!M^n5Z@}{>Y2nE9`roq-pvF zgd}&#)<~y~)^w*>yecQp00aU*B4=JTzo3|h3xQz~og;2!L*gG>m~-iJ)!;-T98Sdz zCy@KqVtPBvZF>g|a~~u7mxV#PA2)MNt^%huJwbSp73QczE)qVoyo8ChT4=Qw!gO8M zBFhV)#Sa_!(VlzKVZPm5c3+V@lcK14bX#5|Jxsdb-qnq&m}9SmUu9V%)wwLrR|F z5D%(08x}9EuE(XEKtB=&*lhec?Z5U&ne$h<7tM-?sMS2fPmnO?MI|koAsR9rm87^U zaUxNmJ>GN}Y_Kl^<6q5OcuM@b5jVSaq1B7QqD)9gI%g7Ais(0Z@)B8eNV2Hs%%W+1 z51E)jfog*B^h?MZvPRWh(UKh+xVdTQyi%NA;TYO$csHB40;RcrIpU*U^@A^zkpu@^i5`%zH60%c^3FkbAHz!Ip z$bKCGCp$DIX}T)>m>~;YDi?AGUCJ!-h_z^zEdC9>2dRM%2|V(Te@PD zdasg6C64OWN^QG0f7eRugzJ)>)PS0dps9z^2FoZ3FNjX3&c4=_pyhBD+sF8s)EF5f zqpQr~CzHk}{N3vTLFTPty`H<$U@L0#1xtbQj@^1qNuRH~wUf|0J7RL;t}8snRF{03 z(DdgIFoe1VaMQgv$jAb{XxeJ6LLnGvvUIwLF%Ufvlk($=pCikIY-z6zDL_SDmEg7B zkq7yZjkr`DN;)=mmpKykolg2&vq-loOE&vx_s^zE=;$@KUSPn%b5Vl)S?L7ReIP46FLDO4TF-IBkD3 zT@#oi z4h%mJw1M-rX9;|bSz?7Pf*(LS!De1ybj_c;`|ESUE<9~w5|rDL6< zfVM2C_v6x)eGlhsY1fUM>GS z>)i;OZ-gxG366ZkLqPc^CX$b{vq`1lBtnP!xrOYimRnpqZ|T#E_R6hhVd_nMn1BjM zFcWaJdJzkLIV$fqPb!`kak0oaqcJ073!(WH>8}gB6(^|$-gVh07h-2i)UZ4 z(7sVab+Dkp9G!Q|OuhDA!gGx73q4&+sq5CWXo`JzSw8J4J6Lb0rct{zGu&%L9b)K}Mj7>tZkCw?#7YdkfKqfawf zosodnH(a*f84YH_8c(ARHN8F?{?rf}bl@4_$>s4Baf?4fjLMzEPEc{k4?G_!{IoNB zK>IVrC2cSdrd5n(H%NcVxJCQ_`cbQ<6EBEz>DGIQRV7Xq-5=?bG_0P8MwYNid{ZEi zR~Y||0jLe7EMIu~(bvPHJbCTu-NmrwtLvE=uG2llXweR#AStRGXxZ*^6~g0F#Oat= zsV`|=4>%h>T2byE45-i5{7&Fnwz*3L9+jH$>JOkOpK?=ZDZJs(%l+wKkRBnG5qoaO z6NI6ar-|TG(WSjAq(;c(z(GAmL|su$qI0cFm)9uq!b1J8(soQGoUcJ#1 zr~A&OQPf^Z_9>v{{%>1#vT)#Z)-yaREzZcaY!oRj6gbnY}@;0^YMPH3Q@ktg0xX%H5qqO!@l+D@N!UjbFKSzlDrmt zD`&lHC!8@Lg{ujR7`;f;3R@%` zB-~c>rghk2HAPG*l0#eQD>+=XuUivVfZ0#-h7nSk94$9!t@K441d_dj5Ax0TCP5n_ zY(=Bj49p8@i9DQ%A!P&jPd?Ou+1&B>x3Y6^x9ZA@9i8-d`Yjst(i^nUP8Z4NumH|i zNI7F?dyH{c46y+XTt_2nXz%vTWX-4Xb;+=N^znHjYFX7Me5f^& ztw2WGp%uDVg^egG`bF2QfoWFNM}jGf>BGr?aC@4RCc{3QES-#{@4{B=zBWee1h%+k z$ZVmCY&NLp}2NiAzD={zv@e40GoixHmEn@x7S&^lQUfPRy@SkjXj=D zSR)$oQ7NR55y%j&6IV#}B=@g-Q7>G^f~7cuK`n6@0~Epb{5gZ9^IDXM_61mXQzv=% z$b+B8i(s5`A2WE5>61<)2j?j{YD4LskMhf89t0vh9W+h~R&Jq$98nc6sOR#31)q<# zZzwcsK|V}k#)a*&!Ox9j^6elrocP5wjLKlZtHn7^H5EIK$wL-dXoK}u-)`^gZ${B6 z4oiL>T*%36u_0XOXz#SQC}Hv*W(#eL^Fa}XirqXwvm3ie9=nmQR@qFIT5V9v7TH35 z<+fTm$_o1!RJ44=RCnsXvJf~=$fFlIXeY7U#y<$N^&xX7a>dDJ?ifzzRIo5cv2X*MUvUKG@r}G z=(WqNAcQ^}K18&{bnzkKtZv$6yqjq8AFd}Oe)Cq26&L#?6)&5378s?J228;s2y+@Y zMiGsyFt{(xzl%7?%zJR?rFSEs;>&(Om*+HoB15345K9||fF3$mnNGF^@Pq;bQt#N- zT)fF_#OH(CCnMj*Ne_sDG#XZSjU`?+_GK6!qQ$Q;pI*2dJV$d5&4b#>P<8F2c7mpK z_Wktx1x`@E4nSQczrg1KK72f$`o!xOkP)RE{Qyomee!vD!+_YFKKTN?At0BOKItEH z*UPgdk(cAE&b<}ib4MnqHDpR%?TPltLh5j#oOROb0(_;JT;Ve3U%|mPy`gYfi?on4 z=IAuz)-YgBOLT$qS3>9|3-sL0+NZgwH&6Er>JUY4HPB7N=|k_=eVA$TlVQ-DNsR%j zzZoWAGQrhOcSjJW#7SMT7if%NN|L{`mcWP<(cn-&mUCo^pbq_SfCyaM;zLo-?1`&9 ze?592ZAd04rtP*(Pf)e9>CFDZ%rMN8Y_O&a1cC#qAVDx3arVZ8VW@NkaQ4}^e11c_ z#LYy-J4@}%koU#?f>KVOJ>lgHuuDmw{Q30^AdXHzeqx%bNj@i+{wX}E6!A@lD<`n# zvj^y4S_`F3YvxSo8xv3quDk~J82dqcVHiOZoonLd*3QmOx8K#@L}dX~vafB5?1=|Q zM-1&H^-`sKgx3!1tY(#GEbzY(Bf;S3(Lwpjz2PC8oZw2J2`VFkE;H%o6Gi(lYO7$E z8{^o2;2?Bdr|~sAFNEl2TzsLX3%eMSfGq}JXL~#;lzY22og79b5J$ZwOKd;1u2kl3M@AhCrc0-WT<0*uaCfKL3m0*sz}aWeya$6i#M zu33FJ@#0j@1S7?-OB&@Kol;U=Xf0_Bnx>cvE1ZE~!?WEhG4-&OgB|krqSJK#WLG^q zl7ZKx1(VtbRzCnUu&Qfm^+wRAG9p4Rr9%jf)53wfFe8>EUhm~`zUZII$>iGNKf1=H zls;m;n*!-R>>a(Xo$O`lQ!Bt~eLKzATsC4AJf*@mx4k+EiCE+*P8*Sft}3pz_Xsue zsln!7CF}RLFr=fu-1BF*FeELpg1h`5Sw}25q>1ALgnG{uMfU0g!7kg$>pOYpqZw9o z)`Jo4ZwbM#hU>5W{h4N1p(Xh{ajY5NLsZEXnGGIKVNzp+QH9>xlEuRsbLxHk1eSN?H7q9aR71=DKtr zXa2pvPY10Bd${M#@SsahN!Rg-cTkmdZ|{6hto)1mYi9cin%fE}~?tp0r%POc`{+T^$5{9ZvgXWFT|H`|W;4 zoRAwlNTKR@6CR^36Y5WZ#5u#T!5qB@lQjxvo}$$de@q9?Cuy*!%}}LDc!UISXo50{ zNUH|}o?{25+fBN<*Nrr3nRnY`q8^oQ46p?(l!6bq*_atY)#P_n-4r*cO>#*Fk{?47 z2}|-f8klLs*);OqoT@))w(0Gs$x38_o-@bhD_5n5a8_U|)rPY)_GSNL_^6F(|KsPI za(GfAmzbZMvXRL_J}`pj3YzfP+EwSI5N465t9j#U(<(`7Uwl^CTW=-bl$|x(aVa{eCbr9f+9!MM~S&_a2)0R${qpW_0I6u z-ClAyC{@j}#3NlGEXyLG1cXA^wi!474Z;;oqnB%I+|z>{kzz4krZ1as<}|CNWk9!7 zrZ0HWJ#WPVnLL4D2FOJ;lfTxSOMQ364*8kx#xLVJVdU9}HI9*)&UwwlgX08MvVouJ z>B$5WG+hhZoiQqlZZzAZqE9T6Obc{YA?1E6|l!gn%OE<(9cESM7m7ZmuG+*nm3?c zv4NT@mHb90HGOhKh2QAZ(kG9o!)G_E2z3Ij9p1?c`@+sEm)Co_23B9BND>NSa2i)Zmvt%Ot78QKi16qi#qzzAtD}jMGjx?7irNmd0 zMKDGgHPz?3ls;gaclMt(WTs)O_ltHaPJNyW&1$U`Oo{U}kW0siyy7bMlEP}*JNJuJ z^sMPS-4i=$3J}Nm!Tslg-ov>3rsM`lTVg4S*v{N%!vq?dTZto)f3(F_%NUaRP$-n& z`Kp@^c5J4ON>V+{D8qBSU;AZMv&$1Fv{FO@JNzOhG!unhJ%4sBy|Z^PUVTYCs4Y+M zI^WTiki+?{dO-#p&6MK8(`>7T?sa>5B8sRgcRx*nq()M|rw$s;J4dSBmK`}yHnCH~H2c(% z)6P0sm$VA<)4I_!B%Ki9%mXl_6T;r}+b&PE|2<{vBzpn3%Q1%T&MRDn3c_&tM^tV% zct|7N!n(l6_S>6$(-G}0&YNoRF*WHe7_Bg>jBlWMi5~2#Wa0acD51dY16GbK>3iGX zFcaIl@z;|by)u!a+-0EH7kL)EhY3GRX6^I4jIMV3^~cNjSzLR&-@nb(H-BadXlU40 zG5grnSgZUghh&;(2=((kKoPXJ zaXG+8vL)?HpNG~DW?M|rgmufJ?X=NX!QlMHS6A=r_Fswbe?b!1e(!Sre+7z8)WX`? z#1WrP)Y`z=MA*d0&e#Nsmlw*(+0n$n2Fm?kwZ^#PFWb8NgUY?Af-o4M9-)0$15W8k z$TCaY^q4=YZitc=GqK^ISI($VJgP`6S~jna$GKP6f9307Yu0asx(^6(%KY{f8v+TU zU!Hh>mjSvb_$8BOBsc`Vf@>+$CVH?cM6|BXx5cI~HpWt=IiQ}l_PM2%hvzDjyOB?q zl^br6RIanDhII#*50Da$UNn)RK%o%8Rg{?&^|4X0F8;Pog{u+7H#}ayXm$JR&i&;J zZZLzqFCZLIiXA(%AnjgXZQJJ!6%-@d2Z|77lGvSvY0|#|LV<9%6M}c=i=|$SI3fIY z0hd0L5Q^P3%B!$9fmo1m!~$Pm7bJsapM5I;V!aqZ-BEZbW==vF!Qf{1r{=z<8-$cW z2#MJjOCM?{0UP+I=9jU8$8%=n4i+tvR^y*b94m6@moQ5r0*S0<6(AQsTP&<@SQ(s6 z8R@a2Hn89x`Gk0Fk#pDFAj`-L^kw!po}NZ_!qa+NlTFsf$@-kAO#rE)WP|^~)GVDJ zXmG3F&Irk2!J@;Ilia*q!303XtK0|Nh3!U+PfM!XU{6UYiL!{qVx?jkasG4L&=5Zo7_J_YBw(Wg|n&6ctY*wbwy2c#knV zs(p1)Yd*Sgslq-*4R3hPMZ6h0t{@#lm*S>HWLO@2N-ccB$?|P#Fk&FPgsl+z9s)sl zTuxNI6V4@!me3|ii|PdD^a{eGz*}czPk&KHYh#qJy-EAKd@8mm3oNMSH!8KACl}YxoJ;eL2uHCg_MFce)d)3Cwmf}T2RVDmb*Sj;CF(U>M;zMTq zxG~UOf23R0Lq#T;6UilWy3i&l?P?_F_z}Xx7BmSDZj|<$nWUMPP8&Psh}BC^TIth! zknRFh5`qVo>R)t1u;zjmoTAsJ0P0zK&;uWiRXx*iU6*)_oG>yCLeK2mMQ(TozLy0n zNwlJy^sZ;(RAEPV!-WlQHuNyFW_yeuk=2=sEx|q=NSD`Ne-mu0ha z9l6N2-FkNdWLsfF6`cjuUQF(s{;l z@hGoAjnyb?c}h)!Vgy3CQOr0~-jOS(s&i@fqF0@mXbQ2HG-4o0Qb&*!Ow0RjO8>_u z*{o!!Adg{2R38jc+0pp10aBI0c+*$q8#p5zsFOaUPP(q}t+su%(K zC2QqNKcnFS>C8JGbH&+89)m9bcRQ|PSvLWGOTxB^)Osf%+;8T z?&1t1Qa{%2^%C>BG}jAV{K|TLEmHUcl8tuA``oWTT{=AoYUX|RfmZaZQ&YM~pZc1bj=?OWi0|NxK1haLSkfW*a6~cAolW6L zb|i6OW9s3|hBCBrh5$E7y!8cB;mN$78X=$ZCOnH(_P10iO^r&z0WLK>nwAwV2xC6M z+qI@Fz;%B*WWBIeUZP*kP1zsrwR+6xihAH?E1PSNPV|@36{Xy(GAj*FG_y=@DqkUph!yrbh zT)}0dFkR0M>!R%6tVp8z=sa*^EHX0K>e}|9X+#)*)SK9f6ykBW%aj+87$P24LPGO( z6|8)CJf}(wb1JGxb%@t;zf(HuB)mss8LJUi=u(+Z)XK;Y` zQnAlO_uv=!><4h(e?&@F#{U5CG5mjslxZ@JnXD+GFW-Nt4N6q!TTf}vMhmVAvfUdb z8slTUt1+;+{BNeeUU&r4_^F+22A~i|J!5A+fbKndH9tSWj}TMB^=Ar}Srds96AP6W z%#+1(_jeLWMuJEE@S;kMB}n3+edv?N|~q!!y4Gyma2y9F3Y%Qv8Ck5pz?e}1`@Dm4%@vWnRDz7zOCj}F&A z1J_1Co6|LQy34XhHE&(G1Riu-n5x`$y3VAM`u5uS&Bgwbl`5*WiZeeMJr_*4?i?;| zmT1}3^RPlKd72;EBuqTQ`8ulHEtJ)ztUq!FDYitnt<7&5tosv)Gd{Q z-RN%fJc{Nv@g;OAOX-f6Jz2dS?O_lAE1bx&Sgp9paO6+n-p5_x@O?OfcR^(8iu}OWR?S6sCqv4`m|W zU`<*$2NU~77RhS;Js))5Z zbsuY&Ca<6An&~}bNu4++xUnNz9&zu?$Lh0ry~Az1g7#*`i>JN&Oewd*I17= z;5Ml4WUFkiV@A)vZuCj`b(0wqWP6!lZsaU6n5Y$U)mQ3KhH%zo3V2)#*=b{kJ|)D6 zrPv07bXTR(%~z#*f;f_Ms{MRNOJcROHmxTmSI3>IJq)d1H4LR*G!rt5m@Yx7&pU9s zy7rrIdnF!gSu8Y4Y%$ead(Eqdi#Fj;fJgP|ynJ@jvHzyPmvvgo;r}J`AbRiRp2nGZ&^Hj@*tfaj zChbIK{_!S7go*7$R0JnEDDq)uiah1NMW>@+eBi>B<8yKf=*5=gbIlz`#(+8-6Po)l zWKR5{e}OF(q~OB;1a%Ud!d$~ZO+YH`_mXfBTJA0eLWwuSSYB3wonkmxVCNq%VJ{(| z6$x94*fWwC1mh~m5{iSGII9HvJA!^;L-L#5`~ue-0+sXj_u$Sk^vMl_#0wP>Vfu)P zH7Y|u_W;<5;UZ9Bp}Uj^17||hkQb_H8wiOtwOYj+sOQO?=ef8BN(l&u>P3+u(52K- z5D)A6p#SCr(Lt;5LmvN0BLKBRPtF6KC7@qy;l#Lj!SF z=g;7E1%6LuXo@FJP_ECb=hugw8rvDrtDi^i46lw)uS9C6#|sh>M%uDSnGC(G_-1|9 zJx4|`Jh3w((#PI4^iS|{05KFS@jpW1=P*A(!+vQLU-5z@5pikyYchs$kb|}OLUkBV z%E?m;%2cjPH}_m^YH*~Apq2@08x>#A0O?os*drYrF}tsMOAY@l9%3~U3rD${??RSQ z5BJ+) zC0>NeB+Qp&VTri181Ry5PI{Y_pmm~;i@SDo6q5#sD6(9g zXFO!^g-EmH8XlV;-e_3^T)Ws8enK%1KN~Ly>8gKD-+|U-M`*c@DfejVZxQ* zw$fsDqBiof2gM0|4RN&YUCjc*)2Ie_-iSD!7^&cALoP2!YjQO2uhT~Ux^mv2aq&F$B^^=lFqNS5fq0 ztF0=k%A!`9Nf>_;5Sa*F{xvNO(hl;{~V`;Qhnqg+(on#MN}EmT(G!j}pUc zbn!Mjx>)(Dq_2X?$in2F%xPFVjtdY%BX>JUU#PaxL{rRT;*LRIEj#Lpl^fV+;sFz? zMucnxnvxApb12>&M5GB5Rc4uM%%9HcUt3d9dlcFNIlz!ocRW*w@Yr8utBk;3)(o5K@_sNwbK6-uSxRpsZv8Qu$H^ z6;ITfhOutNfV#Rnu*qRnSB;K_4k}HjhGv@Yzc4AI;*&(3QZS+;Wp;WLwf4!vn!h?xs zxTQ^GPU-?nG`Ta)_ME7a=5J1;4M)8_xus$E$@*31rA-6M?)9+!L3$8{mo-v;ejm{W z>mseow9nqNFU^g7i0fU_{m+iX53Sk{AaqzG-G4;LUuq6LzMY{Z6gT((gg^#-Mh15F z|7kJe|7HLTEdSH|g+Tw$BR&HwE5rXE{QUoZiWc((DvzzX0tyA(AJ{2%AbWtjxwVC9 z{#y}+d!vGg7ynpIXn_j*8CJ=jPe%6Zr&+N`{Cujx>J~oJ*Z_ z^>mF_07IyPe^u51sHm={sHm(Ul$0)1YFf8`hY-pc`etZU;9fm`>fszg0yTDw4)GV6 z`GG+^{43L_eG_o{NBc%c$3_N#b@g?(-=lN=>G1e)j;)NqYQC%4uOGqxVh5=TxqA^W>5p2qXFQl)TI0=*FY~X;aGt`MKJOp8i0Q?F`0n=@NX?l z9NC3r1pPWkj$y#KBd=+u2VUuHYj0jo_s#gXUIMUSc=#t0t$o`Wz4gMP1OC{;`6s78 z_3nLNeW~G7efV(%Nz33|VhcR$sy?FvI8kZQ^Ghh^tS>JnBLXzAeM}-4U+mu=!#jd& ztOrVO!+fi7Lf}&r00DF}e6sPODVb4;{Z0@xs*PTDpdYA*6;ntrCI)aWuDb3@=mEW_ z2rZ5vnVt7*(9f5>e`{RsYagHB>00ZV=|3WaQ`3lL26jh>U@3?{t<#whdvQ}hmr%RL z1_pb_`=I{W0Qz!DQoDMPOWoN0KZ*_CB)$i`CkLmy5cGb9fX|JL;Jsh?uWkOD{Qzoa zY<0P_e1CqT5Y#mQO;t;30FdiKl6eMlPwfor-gaKEKC`R20X?bh4o0E$-!2|@{J`vx zO^@~ME}lj|^~7M3;??96luTdQj(#XbL~ytPd8M;10Z3V8RRO4|xaj;d|AOwm^g1L7 zsQ%Oh9mUl~ceexkzx*C?yg%k9ZvF58H~r`$aQ6L-C_1}mYSa5a&*oh;+Bc;2_|$*< z;lB8_fBt0N_ZEEi!oB}gi!V&CJnMP>>xKRHd&K9W(yH~->XNC+yx2e`XYaZS{Q6;8 z2KwplD1d7*{WVZOfb!e`Mf8H+_-UD+5LuW&wl1+gvb21AY5SyEeYKjmuDd3WI5M~k0y`#zK4??K!5>7N2!0C{JAh$opdOa0GR_}@ft7u^1)KmK;q z{eG7__n$maelXjTtUtgzQ`0|yKZ=@P4rV{l9X~|ddY<5HCw{?RO!cqe-b6NUdSiE@ z{}v|akv?ZVcZ^kkyNI^_0o*zL_f^=z=IH0mvGZ>H!|L!e*PDEEOSYBSIQUbxGvV@C z=nH-P`kb=15jg$B931qtqj>UjwYbp<>HPlN?=$?rDm!ZJZ=pw6El(v!^g$&IVm z(Y-6;Z(+YqBmbI|nO#5U=W2E}_%J{At^}74aPN!1AAB=s)1P?r>)ZPtU-^%w7qG71 zTx&np1?8C!SW2sYTgi4-?|7fJqwp}#Aen?!)SPMeVVbL9uG`Cz2%hv;GQMer{}v^S zjqWuZt{uJj1EfkS$6>aedk`gXFT}KV67srHbxC}lIV`P$qs-+L&)gq<7$udn*Dv_h zb4D<7D7ld=&ANQ5^FZL08J4e#e@}CiTq!Wg9?kQ+uQ(@yMSjkwmtr0H;K{*_j z5}v9clH%lSVh+BTp^&J=w$dTv*i_a_kZ?XMDpM9?!V)(6bQnLTES3Ft3l@BCqYrSD zSaQ#>37M@AJgcTq6IgG@Q-~ikR3TbE==bSOBCIg@{M^G={ZdNq7Onj4_Zc%Kaw>Im z;m45Q+OQ%&t~c#0B`0uswQeMJPSoXnFayn_fgY2W6RE@SR$ke)teo3050*s{66-Dx zV3M+)gttnp5-P=NXVXeTIWm}9eHl0wL4gHt#f%E+cu4u zK{GXdtN$;;&M8J1CRo#B+qP}nwr$(CZQHgz^NnrWwz+?@ITxFpWbdlGQuo!JC(pZk zR$NG{_h00r3rdD3Vj;9PF-v+MQ>ep9+LqJsEeE|bYyD#F2wpP&`5st|=S~Ff zQJ8&7yEw8H>P`;&#W8<}X4(iX$Xr~MeBE*D8|B;>M%w#9YfIA5^Xn|}HB+jbm-W2A zHnCwn&wc+!BHe%c(ZjCtL9iA^@D+im=Jx04wP_N5AjpnZEa!1bgQ@&FA~v>7xYZ z?EF;p{6wyxRB-+8ef3^H?wF~{OOa6DQ8-NLk5qG)Jl1pRTn`<{2c(8c)eETzkLxjJ ze+oF$@m268LoJ_`rNtMC8Q=5%q>cP1a_JGK8PWy|=ffwdy>JcHK5+YOsJ`Z@R}kR8 zM0O4sd1N{In=5Z1jc&3syU&^2U3{DojTGv9))CiCE2Yn>rPEfm6D6E%s|2}6?d+%T zl^wh2S+vwRER9BF>WSOnh;Kp6-0X~GraQStR(mPhr_W$trID)yi``nY%^HN2>DaNDhHl`6hnPxvU zZxw?r*1Ua}y0rNIW9oq|TM8WCcg2;TnZL1UOS;jdpogzTe8mit(;l&)$$qQveNyNt zyDAXFv^QnOk?V+>r{j~k8~*UDQN|#0PE3CiiG~c z8AZ6I!AMCo|F9C;G(c@?K_YKt^@W|FvVI=>g52=XwLGCkjogP8Bk4UZe!=l_;pJfx ziDQmY-{ftoz29|?rkZSCqSHMFBX3`SUCh`8!?aPY_gHw0XMhF*<6$hpbFH_pjPN3+ z2?wyTsk#f`b2+sSi*64R)t%*@*lAe7QF$oPi<9AoR(;1f9_p%Q@V;|g1pi%a-@nT2Tel~R<3Y1FNIi5ntb|6>KT!S8EB%SB`Y z-#I(d19YtIl5j1Px2-QWP1LWrB#+yR+&mGwaCo!zg5q{6g{xn3sHLyC0;PmRm}x)d zq+C-a&h_u0vS!wMfjP?JQjk1OYsHw6SPA04w0C!O*NPl4D@?lqecddgoIdKD`;-lQc8LXQ7(O$NUMV^DRP|B&3N$ey0Q9oA$x6 z80{A5L(*_n5g?4{xz`KVTtGP5P97f-1zAu&8>j1F zWANkv`(JW-u|3m0{%FuC2uO!K^g)6uV$bdr7QC;{3G8wT8t^ptWM}g!*EzyI?oQtG zj&<0-*`rdBh=~Vwuq~Dh{9gRY(I9_zHRX%%5AP z1k$*z2V=PxR;tPhecVn?hoKg@x}Y0(BM8;`uBq`sc^Y$+6i`8uKUWN~C|} zzH+k>CQ-Sl?l!#)xs_sE*T6WPx~)XQMPmM6Z7o!wKsVxIc$YwS2#*bzCsXNn6A@u` z0<+djvx*!`w3<+&B;e{c5hvIO<7|HA0yK6+@Hx z$NI8|zSQ;4Swi$U@|A8=Bfr+%;E%F7q|xq zLn)MOUY^sRwB0HCiCiSVfU=)fdOg1MBNc@ND&ID@M6z_sV$?4CV*6dky6xdXNnS#+ z9nm(I9Z-heN5c9M@Y`Qw1#~a5w`tUER&&3kj-I>(Q)IMWmy( z(t|la5JA%6`A+(36Q#{%Rtl$z&VaS3HpxX6rC56%mL^vg+49^G;!I|ARWh2XrhVi# zJ4{=rRnx*+@sdiKXDL+dL=AcWJuVtzRo5&*?I;-5Ct5?=VB<5Glj?ytexSju-|_90 zI-UZzeO2*CaWmjlDuIgM{MXA7CZgDoA-C>?`LXc#8NY|88qHqhVfj;I+>l2SxHyLm@ z>F)ASb0E|FC6`%#XSEXiiTFvrWbNa|5(L3236WAcUQnST6(ewq-N%x-PjaGz8qQEf z+%SI#ox+V!o*pnAX*{QEA=pPr`cX4lC|8ym^$4p^3(pNd9dXlD&$Ux31H3R=G&E zO8zDw{c@*47EEno{x+69b+_zVx_;J?oa=UHb>G_ruFRRm2W!$tnKwVHSNuLKIj~Tr z*JuA?e%=kdDY`p@wK$enL-=^36STbw6=F`mxaK!(_u^LOjM#V_c=a9dY)ys}qR;cq z9Z0pSeASq>c#fsgc$x2dy(un)z?1|yZb{-K&rh@h@-Lkr!queYW069RqhGj90a{jW zjK-cW69eCLo<^%9gdEo|&%=%dim@Tpw3P_*V}8bp=_oc*3ep$4kQOsaWXQ$QW61k# zO}LUFfp=zs71qo&XPLi|#Sk7NI>9B;n2|ve1ikLP*CTUcpW-NmA*ZtCTlm6(^*z)Q z50ta^+F&pQnrf%XLL`mJ2BU>ETsh^D(Tv&YVla?B0TaVhYxk z>Zv8To-rfO(%b&GvnE}?%cEI{a*ARzE`LDzO<9)~UvnO^6{n5jzc}aYOoOn`ec0D- zia^A6$rm*s;*i9eEy~!|p)#Ku-%8xXKH+btH=v)dFD9=o^4%yY;Sf9T%)R&I4DqYt z*r>B#Nq6a6f3FgWT|>6;3Kl$uEHqo>h;zd=pFYwLAqeM;>5>0EPt1?1)%-3)xBBKr zork8_Wli>JS_?8MLqY_W2Be*;qB4ZB`tSQ!UM;)ov`JZ!CgY=tXyy>J)#VA^VJh*N z+-vwH2)T<|r2<1XNmT-b6c2qrQprB~@?ns&g8on%sIS7t`kXAG1 zRGaHn*Uj+4n~*^;#|}h72CW6ene?KzSJ6M38>0Z!|UhTM(X2<&OWj^-J5qTz5F{h=xFnaGnT`@W7 z_eRvjHiP%W`w-a%$z+fb5n`|AyWyuT|B?gw63hf8!yL<3o6mR8_A0RW*K5CZ zpNhi=QkGpuEtLHfa36faJ+fhtK(y)Pi@74MmT@EMuCeM0?Oqrx09OZgOK_1bl0)2- zK1DWCea3h7ZA%JwxJXx%YL|Rh?;UE3PpJDuLf=F8iv3zQS1IpH$E)&;;ZkhMf5P2L z#L!|3Nt}1PzQJs#MZGSaF-5RJm{ui)xMm@s&U#s}eE(d&M;UgIhSp;?Q7HO^hy{Zd zm6Qs9Km`RlX(XPHH%}AEu($sF2|*G6>W{ct43c$xr;u621{P|+;hrS86n?E!a`s?k z_#l#8M_tr~S!wLrIg+=@vgTQ&eTKKl){=w{P>?VDdV7OFjrOe z*;&=vLx(z1nuZq)K|Ws>$2TDc9|41GMl^SU82-s*Fi2p}Q%t6N_2mKbm&2h|a1tjf z++~`XPiU#;tFQ4c8!=*X2xg<7m|)Q4OhL}N?@)qxjr$wuk&OSe+DOgg7q%G20L#bd zNRMz-r2+oS+9i4A6o)i~A`4WtRIo&5|FzQ5%s=85&!8s^uT2AJ&t(28vyc zfq#h6lSA6(_6F$B7-Fc zsIX0Z1e8l|PKb4^W6^!Xu#uHKhGNIPx-UJ;J^a&{$rZ09nx;hSQv` z-ql7JI3Bj>m^$Cy!-ZU3Y!4~i#n)z$C-#}Mf2IONWCu}q^~aOTTU7-X@~B_0%iQCB zL9HT7F2h0EGXAW{V+U<73w+7-kic&XGR&X8I!XdWV_)e_J z9l9n^`$h0Jo8f9ty^Ncqb+EEEWIR=lt65z16>71yb5<~Y?fs{35%ibSgWP44ku#t?DF0XP#j#j+AgPpCl9u z2|~kMIqyZE_eWo?L}+s%ko7rZRDaQ9qYC29c2*l)w)zP-CmaDmo<3d(%b(Key3`*BheR_kYp zc9Xj1ER5~_qIp0zy})XWtt2}j#{CS#TJk7eyDq%BKJ;f*JLUmz3=;^*Kw3`NIZwJv z2?Ji%jL5nEs{^Nx3~ybg{w=Q!al-joV1m}XaQ27@qdm<765R3X>XfKZbYRGrWmuu; zzZF7irB4c1T4)~8l4o?{IAZJV*D29m$Pl#zYXKC?g_G`yk-J-AIT@Gxe_0i2Did4# z(TdR{l2(dsD#8CUxAk4^z2(*@8mt12D+jcBZI784dVYM!m>K6aGf@u1u*CA~sOH-} z8~+21C&#{T$NiaZ;PSdI8GTA6e}2p%JcUGe!6^UWVG1Vb6!7j1DSLw35D!ob&vw>2 z)y;RErdBH#d=GDczUnloL^%BeUmWL_ZQ1D)h(Do0pJ{a>*W|atpnjV5;+4#7*B(Vk zbxdAVNSv_a_>nM|+^=!J93+l^T8A-xbo;U_B^_1Qyc>ryRSX6#od?DxQCE~<@*L)H z8t54~j|vJ&of5W)<)cc*_38`|hREy_jh>jlgh!*NXa(Zxxk0Yb;eYxiW?k84aQx&d0eeCd6#8ZE6@m@g!Ot` zG1%V&XQ(qKknoV}eEPyV0l5|jN;Y>g`>msyFRSdjcewci8-B_c?aY+3akb8#5xwJ( zX@ij%v5hWNZ7enyQFj!>09TDupq->(%&AyNsB(KEZ56tL3iD{)toPgH4S@@EbL=dX z)`JB3_kXk4=(aP)dT-zobXcl3;lUDLotv||ulycI5(>*AbUz3!;;TNF3(8_6Cj=O6 zP^=$XH~}@I*A|Qh2ITnCKRcTqFp4b5?$aFu4dL3sWYaWKj6+ls{%(3r0l2T~A8I?9 ztyJ|<9^Jov>%aP#OU_8`N`mDy3LUysQ;4DjNt<`yXAjpQ#Z}a)>$|cPz^)iY2R%zt zHak@)p%_mOJ_lTSp3yQ0* z#+$)iCKLBBR;0uFJr7J>#`Z1yt#QMk;9V4!B^)qMk^R z>gM2zLgMPQdqdM`B$BVM+*u_~@)K z{Nw6U7T}3JW>(x7iF1q6wFdlp!zpc`nupi*!=rGgX#>7Tx<7Z1gGbx1$2a-76IVD; zQV5N1dTR3gD|bHr7lIJqxQrXlS;S=+Vn%@P+Z!%s>gA1%F;0}qD%Nq1_6>(5fM1!V zt6#t3pI^o^MEHcFTboW{Vl|X*4<}B=*yJA$H4Q;y%=TLaJsSo4P8*e+RVwh85@DvY zo-JwIq0(MXJIoQAldtH?w+<(p2lO_rA$k8EJ@LHI4U4T?7do=p>DPba%)~HCmY7=pBD>&T34*Qd}#? zSoax0Jt1{zc2wpG7wM*Ywu^7^9k z8EI>!-Z)4~%EP_(I)Rb!>rK(Z#sY%5hKOO27U|Y)ijxJqV9}!DufU?!B;pf-*ImA; zYs9}9f#;^=idxfsK)bw`jAU&{=6=K4!4kBdoY{SUuHf$G2Ixe_@`y?)-zVA;PU!T0 zK^h8fD>^!N^v?yr;uNmIp)=d39Cy0%aLI^KB`o+rw^d3rx675gF@2P^B%k_yscDR) zQdPhQR8#nzNtvr%MJJDuJB_(tgu%N0(c!>h!`AnZ+6OKJ8|ww;$Ex0b5G!Z)cZzEMCz9NTWhc)769^U+-O_xt-ksJ}Z%#*5FJOZmU39voA9T}cl^jAzYfLkFx__}9X1Su{gx`U(&BuL1 zap`-OPDkwah-@>`&4C{j}0 zvhsVlGTJVAZSGi!1Ias#+^I9;N2sn4P+n;NnROB-!9m&~v=E#ywLwK+gS)I>t)`TV zR28s6gFUV+A=f^@3}G?G1ZASyI834tuU6#PsVT~Fn;O1@OYEMw@DIZBq=y%oOoR7V z`iut}-xj+O$<6b$*@s}?liYsV{t9}to2zov9r>j&ydY`ldPTHPg6JVs6N>-rDW*_Z zrqG>T{hFU*5TvMUw{oVJHQeAuup)#!yVSRc{+L39hlATGQCn5h@>wa&7jDP^T1l3j zQcIfQ&zOmq=T;~bDaX%HQ>lWjwQ;76%gEG#v_n6WqWnkE<*K{USEKeYX%I9FUlQqa zV~lebouK$XtPEGY&EeKqA-1n&{EXNYa5LkgjUT_lg))w&IvUOlg)~c6A?`T zTE^jneQi(GF&$<7*pf%Lp)w6`luOd-x_(Bk?&?MK&wi|O`1#%VZE-Au(&)0I0a@$I zJ@EY$p&NuUcSLf&p4|Fu@jT53-SToTbzm1yZJg_e>XKu-BMrSmsRZyj-<#aT5SGFI zspp@dOGURysg*7lwgyAY-ogTDQli?LhV0`kVamBXT;jLMe z26~p6rpitqJ#9_`Yu{0*z52bxrVw=p?ih`X40LoQI&F)kTjrk zDZc0h2Pxe~TL{;sOKoiahopU%cUht8V=iY}knNaNPz$wg+=5rQsH3-EuRNfr@8Xp3 z>{3OFlxDqr!b#XiuD3l}d-?G*sX_@L9&l<&in=juWUC ze$9An!x@MI#Pk}U>}9P6A}vQ{-OlIpDyivftDy-;guZq2Q=|2qLz#GxKGXAFxN{}i z5j&IHBEu8Y_2i2MEQ|NVrJ@PpJtB*{wn8}tFH1}(XC<5L-{-!yH@>zmWa7$OMpe!| z>RF{Ln!L`LdZIAt}Y9QdDOnb+xg41_#Wj||ItwmXdoUwhJY-<_im!-sT9 z0l#*wmWL}Hjwc-ZVGZBPZ?Y>B7$9mIC4 zX-r(FD}iqOV=Ft=`^vL&it!T~ti$jM_jqS2oIjKRPxK5@t9giP#ow%e`(Z(tlqOoN>)(1xgPILvo(R=bQ-YO*La`(% z8rI9jZn1d}k$I^T5F9wINOen=u{fcH$wNu|CTeD;uQOluu%wKOZdl^X{OX00ziqemqE;Z1U~z4T!Uu58r&1H^MsO9du!Mk%OO zGmqmonCS3NTmxxeewGm2{uS1t-@ZOabH};Lk8|L8cfeWRz7t*QbF7+(axOLLSEKE) zkhsI8t~IS|imfs=(@+($L#MB-G@A<6_7^7nlA-zhx=U8%Yj56|W?=iFv+m&P=r*Rj z7(Oej?kwf~d+RQX-9YJUpMn$d!pm~kV1B=*11D_6WPe5_f zhjCWPH4^%LEYYnVn-Ltjf*iD+7vo%u#Y0QR=6>CaC(N!E4zy~B_02U;Yxop3BIIaE z!#q+@@aH59May;NPSBxcSH)4&uL~z+hhtCG6N>UopV0SVv`w}%2V950W@I*H+bI0g zavY5mHkSE3Cpii3wIn(6p)>7-vWw~K=f;#1GL0DdYkk&V95v70s2OS#ʹFKy`&Xa}|MYs#p(tU%mO|*q@x`)zM?w_Gw*5a8+BRpburMgEBvPF++b`7BLa$N? z6%6pkqq;tPybdt4R}=;gj!d%P~Bsu4zi- zbKMWT4(&t%L~ru8Dayvg=DfuDOy?*~F!nAw@7S*#%V}{sR}fSB6PR`FZ-WR3C)$e2 zxumzi=Pg|@f1Zs3`*m`G28et#=hKrpO^d@&naR4@wII0>QXT(CVcq(dSB^6Ku?TMZ&f?*Y*uYahp3ca%WvQsq5Bg6{7za)p52%r&E^ z84L=#P?Y3w9^1^GB5aW#5x#pEr%Mn?f!`bdF7wqLPfDCgYL}X`2xE-#+(6bU^q@52qisz7#E6M zZg$xaxFy>>^ZJ_xoUA|?C#tkl!GVBRaMwj|lg4(AI_sFgX8umd>QBo7c>{81HV3Uv zrUpcAt6J&l%H}1*C@?>T6BgM>PAcZsbzSE2;Gh zLQZ7a#LYUYpdaDh>pomA-<_CmOy)pag7fVVRf^v9el!HLuO^$h$kTRoG$oiLtS7;V zzvWMp0avq4hi;3OP0SDymbiNyw;ayO<14{#l#WaogoWZV<)q{(c}<74(615jDF?c#NyTt)IjMj-A7cJerUd7h!24A#ZQQVOZy~P zeWB;&V-DDYKvwneU;-$D4)41N>B7mJe3tGrWD1{AjQDxp1D@6I=U&#oSVJ+EJw5(^v_sv=}3+UM_e8Y{8yf6d@wRDc&7B1?$FXU`4n$nUa z6a3<&-Qy0X36(j?YhN}?KR=bTtp%m!P2I&D{pJAPRaSke$%=A?)gFLH7`D8uPil#^wprS2c}02eR~6RYT=4 z6>vvZc2{0)e!?fLU7Cr_zp{XJ)u7^2{e!9iqFP%~GChdX#?C<7TRB*8tIX&tqp2G^5+RnE*LMlXU6ii%P zMUTCy_Eu2sb7zpn7xZ-dtinWfJUN?4*Wd_kn2|fJ?mLd_nDYQrJ`o`r|$~i3t$U|7zPcZuduJ1%y zM;8_2lYMQbAv`f3L%3I_Gt1%a=yjspD1oVs?)f_>G&)BqwPEy84`g9ToCM7^^T_m2 zp-MAq286WdLrHkvT*0X7jGvA6CH_jkW7R!;7@H3ILZ{~s6)7l;cPhiQh)LMg8G#rc>8k73g z40#$i$OWGpV**1rAtw!4d}-nEJ$|3>-6_Qk8eTaF7&`aE(75w?dIBKld*>2PF#f{I zNNXQNANLz^b7E1~XS#x3J=D=aW<&-C`L4_-ZJ%n|LA!g6b7P0&>w6RUBCM7 z?XJ5#>_x7$Rs2k;kO>o}`XWgxFzHkMr zeheEGV~WjtbOS-EXQZ zpcj(u2fQzh$r-*jNuIIncs249P{i^ zpcZY`ksc@^7yXRqDEH^IWIr|qbE8Siaw`WDsx zjb8-95z^OXN#2oXq8O660_NDFMqJmUG%vd;w$bAaa-B;(ow4r}!$MUfxVp^QG6JVm zou9j_vXQvA`6Pz9 znHk7~xb=hw;N#Mx{}EEvJ<3dTSpwpXCbGceec9al#gb8r7zth3j)I;ACa1YVdj|Dp zKb|P7-Gf%(rp9JumZ`3z!oCN7ZS&n@S(x0+W|geo|Do&M_acDwcQnjJClOAr@Bn#Z zjF}8@2QhfvVNm+fz&sdb&$B*Z(`VaqL3n=PKJnJjxtEpUF0PBgzmeo`%Fs=Cz542j z?WvKE!O{5$ZM@P?qdeIa(lG8j{jTX(ifj5q;R#=YH)_aZvq&gVHsh_>bIh7%xJDq?p9ZV59yQ1P! zdL-;mB^IO&xn>uYqD6@f>`yt#U=XzpP)-MH`Tal!rTF{Z@>yW?9B~%xx5x#p2e&cc z(8Mq`SZo6wCAzbCFEDPrRG5NQCN&xI^In3yQB9|CNUk=l$$z}2T`G@Od-B@3A*N?n z2Y#b+@HVgMY^JLPF5!nIHfnw3E}Q=ut*vWHC$GWbEmuR8q9Xcyj}d%mO~UF%n+Pc|Io&zPm4Q}nW!PB{R?8JP6pd7cRK~9>!;!_;-m7g^tZLPDbIY8 z-?uZNkODal&F5+_%ij?(g(6f|GYK|)14kcE@Jub6&oq_@$F6KW*#(z#b$30>P`Njj zn%q>hK91F;rm628-_JW}v9(VWL@`U7 zDAvJ9IVP_%15c~wXNo3!fH*J{vtX$L30Ke_!x<^g*Ha{hBfL{Z!+XazyQ!GIv*Tj) zq#j=kf(MNFG0KuG3BDJ{q%gB8**W<6p`**Y zvF80Icf7^VUK3A?eT(t+>a;k-q{=ND2yrQVWE~zorS~v=ce9$TX%SE+-x1Yo?Wwu6 zYADS#5^?YpCI(issW$x$#bvA}K?H`CmW|w;+Y@;IYUsljI!3}@3hDfYbRRC9PIH|H z%A56MYH>h<8H$@uX;^(1h%M)o&yk@<$T(kF@PK6EN=go2_ik@kXB^FBBRLCO`%G@x zo3q*}$6&MHez2!``YYd?W^Z}d=2biUz2s?UYh$>9^$~($Jzr&2IH;1Jznxf8o@5WK z?;=mkGzlKJxG>_U5(vC2SZ>F{3%n@muy2ACRxj>)xvF8w9!{{XdlFp-@icCcR+Wiw8m51|B{NQ z=%%s_QTzG^(%bFvSKFB&e2@HS?4GQdr4*Zsa)M--;uC~6glh%%-pbYq%yF^n<=G+A z6Fg|O*<#YZuF!gvM1Y7LlR24TJ1n6+OBD4Rf>1u^jNZ&FSfKRr=d|7YfgWBxbxQg= zMzR-IeVWP<%?#0xu7g{G={TTpIzMH&oZ0=IesyqI&o8@4Vu8|s|LxS?8}c|*jv-gP z2Jpl++4O%jSFd~IwcS+F2^)4xH97iOqEcG7C!Qo*_2ex&iWps$nDA{!UfSHy*sd_d z3@X`sO)a%+{epj6tY{d2WMoK}r8qati$ikUyZjE6_${T!I~!%lvoq_vnyakK6Y?h| z;iQEmLuz%qH6t#L6x}ePwDT1$ORLK2b=HVc;_8~n;Hnl_Ye9a##8i&&3D)C?_l`1L zoOXsdA1k0kZ_My2gm{4xt1_ADf8V~Bjt=zqn-Et!8`wD@S!)vl1ak9%#Ey*_p(16K z`YI!QeLv(`t4#9~d9T_m_qG6ysPF$OiU_W6?|!>5;P~CDGGfWSCKf6Eix<)ADvAZ& zlx$8qEJsm>QK-V!WXZ>4-*&PQ)Ks(2_|;+CCJoT`4ebuY*Qv!}U3xRv`RN)YQD?)3 zyK)VqUC1{v(2E)?2w{jqBVwk(*PR^+Q1t;)LVaS3`d4%KhFD$mkta0H1(%8J_Cy{G;|fr-dr zPL8A9TJim)9sI&4g7spjb;{M6le`Qiz6MO#!vXsDUw-dn()4W0aHOTGzu=9>0l?Ew zO4E%P^**Sm8`?p)JAqML28fW7&p=aMw)}{5R2(%(6M(2s;n7bG5OI>fXhvFTNjfwk(A}O>42rthm zQF~Y@Mncq?qa7JA!<5u1BNx4%(*krMKQB{JLyVfnyfe5-T0WBAy!2+kV$?l!xtEo7 z+;=PW_UF5G+`eNQ10)q6r;2x{Y-`5uB)qzee94(M7_5$K*`0k}B9*WE5S*k)uzkdp&h>iw{+~@44^7|_^LENkJznIh7EPuvACX&c@%WR&&$5yaTsQ7%~lvn z=KT=}nzcLjRTPkS`w&(Y@=IB+efr>yVvP}Nmg518SzF5tDd6TcdfJ^U%u0xQWJbZ{ zZWDr|!SdbpUU`TYRi4FNpZ;uT_UDElqb&5%SXp2{5Ps{cWI?aF((0h~$rlYvOms-h zW`@=?@lk3|8f4~mj^YY4ypGs$2}%E!@N%ZNDGITWWFTV-PXP3d-q};FfW>NH3g3cXQV+`{~J{34#AsZ}iK@yk2f8=3ekIzJPHsWfQ~ zz#=>)CCH7mw$qe@%kE7e*aS%Tz*$iq^LQ=l&VGv_0kt!9$c9oyX~6qWK&x$vCgS*s z=GR&t{$GuQKx~+3kI;k$fG~)t;M*6_Ac8CUf10&T0_;>VZM%4Kl%Md|FV{k$9Ot36 zJ#Qevcv$X;iuQ_ab5|wiR0|)ub(r|3`@DcPktmWaJ0rxM)gHk&oVAx<|G?K z8N3#ZE<+XO>*-R+W2RS7<>!ezPzzxG!&I78Osy-OMD%Lk$$NU|t(Nh4v!Gtm-2HGnwP_2L_^DVXl3e2V0Gp>+&Q<&~?+ z)#D2qED8x@aK7Cwox#V;sJ>repcH;UMdX8k!_i<`9{p9;kyx2Fr}pHk7|Ohb8DFHK zlmu9DTUDY0FZ9})U=2bcIS&DnN7XOek-_TQfcZ+}G^tXvS|NtQ+={VcmtphorsmUZ zM*c+c3n&P3m5%@r#xFYBlI{t3OGA0sUu&zzUV5iC;4vX)&+yt=B7^YgVyV3qc4my5 z)KJpJxd-SO*(pt-D8lS@F3zcKq0D+fF8KN)+&h$WKwX|A%pSTWnGMFfsb~}wRHK)w z9+LO$b%bYPSGXW`x2pW0uR~p>_sxSZaO9w0sk~*UzLNxnva9T!hIJIA*bGdI7={ED zq;Z_KxMKo1n9*C3W2B4!@utMCYb6o2Qrz6Wzq*5TQ4piF2gQrN>{>n?Ihoq_TgwG_ zc|4T{deizkJ+u4|q)Zeh7(ulibNg-oEu!wDBDkM;Xo0&6^zM z5;Jo+vO%RnwNn~D)EO+9UEJ)8pS|?VT=Ave#3(SsH}!9rc)<$WG1~vo6hxT~Be{Tr z%3$H|Gig2yuct(FH-s|dDq~}v6g(QpiQytw5lEZC7oTWBv)X2J*ZQ=5#CmWdbp$&^ zfg=XEDNMPPUg z;J%BjgyFeD)hQ}PhzS#R6F}7?p^uJ9ji3(xw1RsP`nK!K|0EMb|C;L77%^Je8a%mq z%L3mQx2RaHo zZzN|IdI(rUMnq+cu#E&cx9F2bDRW>%7LaL6%3NU*xb^dbmDn=QnCR{@K=d=7N${iE zB4mj379!$MK9{*^<_ZXjr~c_!aIkHTV;SP!XbS|tF23Y zXZ*G5mz+I?NbM8{L!X9Easjn+zB1ToF5;C0@j2_wm`EXbJ+9oN>lvG~&_h?|(n+zL zn)w~)O}b8H2Joc3A9b5HnZ2aFO1VctN#7<)d?lSf`5BBNjp%kQ) z^yLF!vvn-jvqf6MUN6x8otR9@*=*3J4Ix2PAU(S;vCy7)7$O;-1Fo-_A&nP~P-Bc}0*IZBMQE;N$dY^R476wO z1(+xnGNleDKou3h#}HjfqBd{$+@2w)1KNeHgmIhTmTu7R&P zJ)r5=vw~+L9;AdFb^jjDc+o`N)H_q7_#VW7{b_nydD%*I^|_7S4}YDw44k_81D{cH z(D>gJ3mpHeVu6kGe?$uZDHhmR*xCQ*`2Q#t*xA|suc`F^uULp_H+Cu2UJAqBa;slw z<8GUg8*5u(HIw)HnA@` zd$&Yhb^dlqZEOUS;L-%b$OMeB(TTa%fq?-a0|T4=VB?c$chg z$FTtT=p@{`>Dq3`jtBYd+;I;nO zkMlxPe4@wF#omuZXXH<_G;D^2j>E3mgZks@<%;$CK#=O z4Yi@6iK)rWHSn#P-;mjxICwVEz3}%IMO$O5Q^Wh)(3sg;_@J)@lWTLK`d?0TP7F!N z-_46<(1$TIqZ3F269Yp-Gc!=kHGo(hSlP3`ZPB%v_4C`|-l;sU9$p(;8^Dn~rI?=C zs93(g0T0e|PAov&tDKk~-k;99evqLF2u2nr2hek@%yf<6ugM<^Hs)V6fA_%dG-|=f z{YFR(Kv=)u@0a9R^BjzA4eicv$Di-Q^C8k<)8x?dAL`GZsh}XXdH^^rGI>B+WGdjm z`0y-{fw?K5$B#XkCvUUA`eVEXq_r_Nz*#@kFHhwk%FXwkT*3QWfdbGkdwR``IXE^T z#V<<_8!&b7;2Hhs-=6KC`}m)F`(NtGpZLGOy)2Pj8rr|tg7^5TzbWCl$?3(LJ-&S2 z>*IG$cqX=dWy2qT-IbVM*3lrgw6H&S*&)R{XDW~ySw8OUw8YT3z!QBpW^hGjbz)&p zW^v#pf7jXmYbyVl+|ta}#GcsT4gI;SISR0_zvfp?8!)u|IppxU*KhBMbLX^QcYQ=_ zVC;YWVR4bEum$G!^=DvDo?H3DQ1=GkIda0pYyH?*0`lgzNAH+G?<(h@3~i0!A6Hfz z8i3D0{+Yj`HUM6O{KIes=%2zk0A2@wi)SVC(LaP`b8_FpGyq-`{Lh)npYXn?DIdc5 zoTh&WeIv*S>`FyRZHV z<>iC_kgm?M0%K3kTEEAJh91*LCNJf`_1{G8-{HS&7S7?Be9=BU^}XQA{v1AdD318w zF%7=qzpWRp@n`cPWqt$mLBAzO9z7k)?Apr2R{!qbTYi{-pGcSvG5JL~|6La4{MRnh`E~5e5AFH|=3nak0p=I&{EPi#HUyJ< z%_qY7^TY7dJM(|PyW=1xRwo9|I03~5z6E&73TPXn0@Oj@?1dyldk?&t*QRLIMYK5s z2IVKrE5|-=dbr%gdFD~cC`c7*|pw$IM%AL?L8yV@LxeQ~$oDOd2wm&Yl| z&3G;4;%4n<*2c~<&|#s3>)d{vQ!-jy*S(o3$BleQU{&Qw$<`A{V$WF|43rTG5j1ji zVE7jZ(gI!5st%zBJw59xk-(wzB%d>K5RJ$CbBS+iw1cNXSqRAX^#lF(j*{}Rit&aS<= zo1}%|7w*$1tz-50YBzWlV4Xy}uWkX9`r*e48Ewx_-9<2-oFBB~bJa^OD+a%q5PRcg z*zoP|1u`XO27LydfHGARPBm6v&EYe+P7uwGRspIdG} zE7HlPDEnvO-2_qZ?hibekUL9^OCgVfxuESiV~?g8Uluc@8KgoldGcAp{Z(MzR?P6i zQq}CM8zC)H(%L+@2NH*kF!4_=0l-b(6%d=dk$iAt^mL&@>w9>>b~r8mfc22kOg0|2 zd#cO$jD6!dj7KbghW;!kPt21%;8WyDNP(xpo-vo?1)hP9fWT{M=}or05IGTU!=>fP zHRe>W0kEoOOp-PW1j{gm)KcI2Prt0$+w9%tc5#eC9MJp-fri&oj*@ueHT`>fp|TY{ zNAc`{;SWOVMemf2`O*croAH{U&w1tvrUyC`Ci)sKYh>gk9fxzI(>5g+EF#(m-6$5nIiDS&WsSvfs zYfxrdbb?lHSDGUF9O{Tu!3o~p#%ZPl(Y0c8wi}Xsa)AV#c<;0(^e&mRvu3I^=?nPzcLlg4;bg-)r`e3#< z$9X%-9Dj!g_$di?^Gfnl+k^d+;3OB_@?ADxcGlz5+ctd?arhd+QUj`K@tK1eVefi@ zpx3leE)Zp&+8Z;w`Oqtt2rTdgRx`Gmk!7hY+x539ereAo#V<6^c4;6kD7L(kcCmzc zJoqNhipw#}VL`qOz09Ii`{fJwYD?E_BH*4JnonH{9Qb7TP?48PSr$s5u@c71(<~PZ znKN$$N9J3AkD-mr$7lTn@o6(RhXQA-Pm)N#m2J)vfPT0|9SB5%C*!)HRGgsyX)z?0 z)#$#K=zf=F-NJHuU>*+{)9u9`00@|{Ce(X`2u9wbI);7-rPTi*cAcr{&zlL{w~Ub1 zz5^EQhp1H0C0aB`1(sgfxCOLPHJFh|KlR5>mr=Av7}uX9Kqw=Q`z1PrpGK%!f%%%{ zFQhe{B+MwUa}ao_=B62IMG1h^B)(WjnQ;^wCQ^PJKX8N>0?SSESlcXy{30Jk?z9Rd zpCf@^0z#R~aNEHU2N$S8)+{mST0;E3`-$rc>6Fkxu5i&F%}7&QzlD;UNiCP4>@r`r zrd+@9_~fK6F?458K{w!&LM_$H!(9Pu>WJaOz^HP6J0yqxFm+LBc=h)2v`%`sXPdR7 zq)csi9Un)5$Fr&*W&qGdcO3NXyiM%{pf_Z5`hxmo0bN7gU|vSpRvnI-DbjVSG*!>d z)iggS%tk3XSH!lNa2E}b{6m;HP8H%u73EY0)k{@g>e4RDl6(QZ#A2x>%H=7vq|>I! zRAx2uJ-OtRF1Ua*VEB^c#2YibI{Uzg=7fk(G>91tj&G7xO!4GGa9VD~0%8~6zsQWKtF&%mh<@vN`>rYFpOg7~VTb zKjKV?+hM=HPqCE4Tgnju^(Li^t~QyCCbCgVrqeiqPv7l1Gp3sWI^V%Rxj;hzENu~m zAWpCgbBamsY?DEa+XfqvNsNpP6@5t;ag(HZ?~kTJqh7G?P+hiuJZ5=rQ3Q)yy%#L& zbnV2owO;Mk27cn6Aj|7RJlP=a(?vX)&Ch1oIx4yFD9%&Ho=UmId@FjQFmI3%h`H(c z*OnITLQdSHVFmT$MX6@TFYIUqn=V76sI_wQ7A|&Wa&c%l2{ZHl@X_AzX{^9pxd~ow zq2p&Nz%S9xJT&X0Yyahel!41ltR=(v)`J3>vq3PwbO9@-U`dYr$>_$XNL+7WFT%&` zpx&JLVB=L~qhH;VX5b_Q)kEuC$WfEw@dP_FZmlO2o{pEVlLYF06gtk4c&i>O5p-Ol zaxC)Kt&`{`=>XewuVS_pm8c+yoj3w~u5%lGz&4V4t%s5P7YYY#yx%w6!&P};xRrYm#Bdb*tFv)@&%Hyh>><-A=hCJ;B;H~WlfKE9lGuCedHf|V;* zdCS=E8Tl{>h4@cVaZI9-IQ$FPvYwql(fshGn+nE`Qw^#rk8#iTJ)ciRjv}+XgWs;t zSxvdHz$3s6c9B4E(hKmCy+sI(CpUi8q`;5?5Z$_llS)AAPfWTvi$dDIVErBIOh%;^ zZB3nV_wt@ z5P}y~RShqc>L`o{mu=UdPH#o$ay(md%v1N4@*SE!ca_MVx2d6XF-a!znlBZn`jcw3iNoVqe6A9h{Q!cG z5WY6GvdKSziHBz_d_F-W>H12uukY6S7w{1U3ir zrE12#Rl_)QQsd>zC5oW5ac=}wuzX~E4C&cNJm6$`N-|Un*-B+uUCm$wEx8?6^BZSf z#uE8U8->XO8geBx@zPuPu;CW8WYzfJu|;uJJ?s_43D`% zf#7%tuiCtgY+c7^yPAUegC^{pqmGpS&V+71Ege1M$`$Gs-EI2U)l%!Ss`@rM`cJLW z&MUms6gL@rKT|1?)}m7kN#q|ktC?uPwqV3O0)-%@Dy+s{9eL-uEORP!3$%zp+3jrf zOh3#WKz248B}B|Gf(#d8x9O@C6Em``p6isAwq(Vp8Nxa!=_4^dC^9P@)u08rq9at_ zI2*qe0Psh{)}Ai1@>z&Yj_)QZ@^f@S|HFaG4sDht5Ijht@D|d&M-jB`>d9iF-gQ{P z-LskrE_|XXqpm0^d{_g%RX7VNu4iZwQ_9o|SMY+x4V{k)gxiI_|{R@<96($rd}r22@*_i-Mn)K4(0Jf=bk!ivGpR9W>RZ2VOVfz zFRox%&Bpe|WADPu+02dLQt&J;vobN;y%9&9D3BI={ffjbsto8(ZHy*EG|+CsZj96? z(?YiY%y!;vIoGtzQ11{jGYD@dqEMdciwAR`hDt02Q|i%ANg1y>(MwkYu7q-qfg0-q zha#7mbnJ0uDnM5khh(QZ4i1yJ`*I(L$zhE+PVQjpiAvaNm*nvLzH@%j4(aD7 zKq+aTyGQ09gRjyw;D2V$q-+O?N};KaBj}IR9ROQmi`JE~L47EpdxHAdM}ZJq=N6p; z4Xsm4GGzSWSP>`yZBo`W?SeQ_B3yQ&ur01F~4*T@Yq^TGDKrdq??pg()>==vy z1Kfg^#QfPeYu5TBui47OivE{H_vVY^iy73wN#AjZwFWg4GNU{(ky&S!sUrKlxRh9{ z1b|fk)ytnmvcMYe{RxV4E$}Ef5j))Eu7%CG48`H_2=1@a2Mu_z{G#Eu`H2xk0C_FfP=(gq*6 z<#~G}Osv|Yp(kTc=QKZx?0O>XkprD&K1Rj0C*T%^W~g*ero0I6RJKZo-M>z3Ca>V+ zZ>Sq4pHXn)L^ND(s94ooXZC4RVz~%+{u5d2((=gfy7E>g$^q4Rq6gqx=&p~J=l{oH zf#%84=P6NK?fj9MZS~nkUz)L_p-P=tyj!)lums}F>3~`NcV^s(Z9sI;IVGzv)tUjX zY7fb=n7K&hU3ug!ef?ue4v(z$iqMM8?1c@5r7UoaRF_a2qNIH!NzCO-x$L)C+^%Tp zA%SH)hajRWR;3-5vZv(a$<*a4Qh7HPUFhZ|R3);GW5o9Rh)hiY3PPMvhsMTyc zERYwHwaGkrI(j1!xt%9RXPP6?#R5Iu{6J)4GcU%Xr%K1^l!9z}LbeSUw6Zjc3NTsu zZdG{lVNn?LQ@KV5@t)3#P7Up1nW!+jk2GtvY1JgSBT1whB=;lnR*psqLNlU0#`D(Q zOevQ3JMAv~e0gfxm?!rtJY@}_^XeYzcf;iB)5-jpOa8d*1MSOEC#4>CM2#9P9))5* z*&<0WA)`H!BNeo>YR9B)qp?Zo%Aq%8EF5ahD(mY3#2(~&A1nje43=J10TNsPCGgEx zN7~3uQO#F}v9rfE3aA%7SXP;bWmkAgb7_@AsBDC&K2)$xUj%`b4EM0Ia$HLK;T7wM zDMho?N0g02I^>VsuhQre+18MA;Ew-;JS2AUPVLR`3j793dMC<>Oe zCGXSc?Py@v7h+UeB(x8K)6gLprK+@g5Jl-w*POikrGZ{C;h4L%WgmV!SxP~yZWUT8 zdp1gRCbP*{Y}U!__PaB0I8e{g=4gde&o|R{-d}xH`|-K(aiiIR>rvG9`M4%kUFoJN z0$5WvNac!K@*k}odfuPu6zb`Deij&M#LTZoE)kwIWnWHcdjMgYTlUa?26zYFn)G?{ zggQW?wPP#joed@qvqpI#C9`l@eFWxDdoj$Ndqjog^gIu}SoIp@LoH zfMY)DHDgKAp0qUv{VY+|c}H(RCBWTNZd1PbVCr==n}0?Efd%Qh23vL!SP`!@sOnob zvnmc4`lFZt00B$snmtPvUBJ;5LpR_S{@sPwWbU+jn8MGZpVUkFCdyMYxax3H< z`O9nNhgTTIRuPHBCQph zp5^>EQY>%WU|ORL8c+3F2V?Urv;P$D0l55+87W__E60*HhKchGe#ShH{!NMv_3mfv ziAmZcJOphxp6{}S{M^{jE(E^ZSvy_h;EX^}RG*(AbhIZyF{}E~i^9KkjeFJ&XMB&y z7hJn+zr^xa>Mk`S!!jnih4NpoO#k$l!*!f$pW#mF{Y>E^7dE5x{_I!LC9EwUD{LbjpCd zO6(~=M=eVMeQ(}=3XD+7DPP1}4E8DS!gsbQZV-@e&I>UzBKLyKW{A)>Ve1AV+~R^M z_ShBYRD6e{FEE$_4~^3m1AbDaMdYNX_o2caW?;Z^USDQ{nxNL;==b4h) zS*eC4R=ctZI&v07o1w-qfVDmD=V89Eii3_Lr0}h_9>Is12TM2$C;VX{TBG?uqDA2r zZY-Nm4JY!oATH@^ztfvS?}#dz5owiumyv~EN9dv3B~a&dn}Qlp6H<8io!R{WxFg8)SL6dAs`!DOEbl!iTfDSef`U1o=hYT6Vdxj z(2Wl=|D8GkxSs2b!*m>)-MbXsnKE`><}01^P4WQtt{jVPp}%Kt7P()TH|TFZA-{XL zH`aXx8y-kEO6P>1g^`{=1d10?>MZnZCWX>q`XA|uKuU^w7>s|Wz=-t7$6$DViyxE9 zFR|W5Q0W#pUBb~)Dbb5mYTnb&*!C^7*#ZHUorN$UQ{6e*s^exAE07lGlTSTEAJa|# z%5@A_$Y%DXwDoHND)rA1)OdVC#;Hv0+YmysOUutDeHJH^CNm;LR4`v8ggJfoJ&=)* zb95~K&iH%}OdS9C{SPGg^n$VBt_wN!y>9~FK8)|_#a0Pp4O+a2CWtmy30@qSB4K|3 zUf>^XU2)O;B$q)#Nn1%A-kv?OS90QQCY#9SjuJQU5#S3fr zCNH={N~NOyue+ofsi>enZ1L4LVRESry|%%hAC859Ot+6HKR*2e3|;YjJ-vjbR0ekN zwV!}XRsgCuj%N83@7qPJra$=N8m}6m5-;TdIAa4i#Z_C34z5(c+?FTTGsKT1W;^@u zT-9YG1THjhL;lY^a_eP}1tjc~s);xK;<)$-ExgH4DU)*@;6I$iOQ-cr&Agtl!AL?0 zObA#kY^er=#@^UGY*g_ymm66@rtmqIE2Luf!N?oo@@c8r@L;AkV2PWn?PQ!18leEv zXokgy(DB?>s2CJuYl>D8og;_SY9aq^x25~4&|=`y-_h=O0*e?TlX&ufwn)U7xU++*bLlYZu1<3thQug3vbLP!)z1g5QO5z3w(pX>x1;ZZ1LH%oR0MIe(9zeZQp zdQ8L>K10suh zbbbBcs+($2m1vep5h1iH>crdeiwZs4xi}v~yQP!f@6y8W&rTIC^}Lty-}8IOrmNNo z?x4iTL2R(cJ@sv3mPvg>s31zZkU3ROd+QYy+9{MhCTL>P3>^nP`K+C1Nnk)C?z^QU zFvBhllf+fk2`_UE*`CCEnne?sayMo8I6+=j(WiXz1$BE!u9+AjNrFadb|d9eHC?rL zup3k>ZnSf~9G|zN#%w4W{tG~^4Sr0W@Zp{EdWCWTS2X6w2|QwDx1m>6uYxUnxg`H2n8<4 z6T}cP!Y#MBu^yb)ez^7YvcGK+tR8Mm)UnHXL}$OP#@6VhaC4KLVxv&IjzJd7D_qbI zcruf@+uiCfNR14P*!cpq#?~k+BRbEZayk;R3C!M_C~RL3-V}mKa*zJ};sVMJ*|JhB zE1eQAf;yTp6a9KHca-ZpPA-Dr-p)J=_Ln8B!$4mc*uus&lwct`kh{6H8bOpqVZxY& zL2!`2?Seg|p(Ne&as|}{T#&Wwd1x_8#swDY>+~WPDWGtH1b5~& zk{2_qkNcfY9r#*N7*zlzYDyX+1+Cf2Rj=Ab>V)*x7;sqP;L+_ z)+6=Yz(b%BNyQ4VNTg#`6|S3%T8(Lqg+6F;jxteJ8Kp8uXn`2t2+qg$n4yvC}zXzHGMXFFEm7^84F0kDyU6_NX>N*`I6ZI z5mG^T@v%1JF%wF;O{&!aMPAA(09VG+IOjdV&JbJHYZjK_0&LD+;c1rMYwV(pHbR5= zm_q)&E=hijBXBR}j*b@2HjdW@_t6EFyYzyN#V{YfDbG3_Shh=9YW>Ip4yB>ahEDf= zZuvpNyqcY~w7c50>zFFL=cnB+Hs(z~KYMpBLI6C%bY#ymdsXji5&}k(Dc_7^cbx`h z2af+>hn_>dx8Moe;O#EE!$w>m0Y*UEs2$U6kq`4~ZVO7mfgi6y8)((62zCQg4c7nO37hln>pyhuZjaX;5<3zfp;(Kc zG{H#BGcO|BT@~V#2kA%|N<%nZdVH&@y+lkzd8&Az{wirb%W z{)YF2bh@L6D@e!EtZ1!uT-t!(Mtp&@xULD@*{u>sK?CojL&)!A;7f4uH%j%**Q#ZNqeyZ zYdeWy8d9Giu1=~8R;GkrY~zN3l-tLn9lu;st=sGg6SqC!#vL?^P}v@2QdLp%v0=Re z)SmCbK+R0l&ZEPrkrzndnz+Haf&L93=Rv9ta-9)b;U+De=rDz06rEP!E2Bm1%^5<$ zXr_@_C2PF1Lki)`nII3h>XrO$p%X`Od0?J=0ax2uoxg4rDO%hSB|mWYadKBo!pZ8h zr^gjdqHCnZ+2f`ePJg=HsuM^{l^7QWbkLZ_M zp=U+;^+>Qs8KyPokT03cog7SkS3+Tzx(a5_s>u8~FfoQZl~Xfh&&ETb^D&!eMJDk4 zfhv%VTJ;aY3N~sl+JNcQBKGb+)s*g?1^dPDYgi!O*2_OJ&^CSw^n1fvn$VanK}ND{ z(7T$eqkOXGrr_+&wFtX@^MCWq63@K*77$UrT$YeTWGD+QP2L}jJ0<9@2TB3VBvII71YJd7J3YW^_+ekdzfNb#G!*Cjk0qGYs^n=FEey^} zcG2jWXaNV1ujP_;&Hi0;~Yj6@d1W?B8SlIPMp~00sYA8bU#&0XPQ$gYp zGaZqSdx^<`0RoK=&q&{`hLVq;NaCal>}-k zTd-+<;p#dtZa9J+q&De|T$VV1k+)YIP;Ts=AAMC2#SR9iG9cx*!ll5oLy=8FH)QiY zaIlpfD)z`x!XZuyu3D`r6tI?OPij3`xHplc4q0X?zMj~F1q#Pdrm)+M{?W{U1fTq_ zOvsOdW`__Mh1NQhyCOu2o2pt}c3JHyw)kJfcDF6xIsy82_-jx6nSKbkI2ihL1MF1n z1s}G{FzUzISj|~ME>RIWxI~nutI%@;{v~`Gw&-TkM?vo<%IQYbbXq#U8{mF5vw5qGJ(U8}w2`ZEF9tWk7%}ODy+qJ}UceIdR6=q8XCpMVDwOz*zDhPffo#(CBwh zJG!_4QI4Q^(3G!~^Lgr1W~8Cv^qR~CYPt>>rkyTb4gG(e#xE622`TUHd}vqwBr)&_>;G8>MBby@gaG2lW}H$C&aDm*auwUfV7QeuVGRxGrq& zlst_Cx#D)C&XZN}=jrb_Zt5l97M#?pKZ481{-r&L+g19i3(MV6%m5Dq@i3{0C&MmU zV3d6uKHRP$_KRRaAnn=t*5S>Rv*2+Dj_ZeFhy#+1c(MnTJ0xs!Md%)QH+w2MWIQ!RfKFw&Bw8TRJ|ZJENAVY$I6S& zTK`((5RxYIS@v`E9@{i(ahdt_VHYd57a|Pd)uZca;*uWkg(pfchk{$*x6AXUZk7$d z73WgPO7`w6hA9)E1jN8;H!hjjl85kR+!oMdEQ5ukXBaAfs&vI$KBm8%*M?f>cYVjQ zg}6&WDXFdg0nMg>uS>yr^Hc9Kgd5&PoISo}ovZDHlsmUAnK*01B6Iv?%L1TalYid^ z80vfKK9*$D|1PihjQ;~6B6zz-D>VosASt#80soBoncgXGDJ^JtxcO(zfca#=mDteO0J1g zyV&1!N9*>=kStK0Nk3nPBDV?cmaqs)J1rFB_}Fq;Mwh!N)ETIJPP=|VnIZ6n@jDX6 zvvv`K74zaXSqEch6rtY;9K?&8Bf{}ZpAMFA^Y)GxpQKG*`s%_Gab1x6?kRT4#yXpe z{_w(zBE{~bv43|~S4M>F)uSLa`#(MH1bePYsn_wAF3d#)s zTxO4<1a)VB310|~25`OcxWhM=&+l(J)U7UKnGOSlf&HflaA$(XrRgvrIRVpU;7r4) z)P(GpmvTmgm6ur9j~yx0S-pg1RtaMA!BsZb!$FV!=&_f}7UMY|SS*hb7AsUpF!+;V z?{7Fo&2COXoie&Fs<3Mj$EzXnbd!p9%K9)e1ZM0;SNpRYUjnjLymaw(a4ugvB|l0x za|kG0N%51xK}6hH_W9DF)deY#z`ufWp4q`vq=oG`;>Ha_6a)(WM9fmy{7Ij4z&w?IlqZPJzOGX$tQMCXhaH)GFq~nB9_c%0~48yr`ZdX<#pqx3?nh-?!ZT>LM zalsl#KAtx{u|+Gd&UTz5MawgZ!_LyseXi_v|wj zq_sc{^AAL7FfOr>Q9F%4k7vHbS^^kUUEhP~Tr^Uy>SCU|B$K#;y>qSt7Pr{nt$(%j z&3`HI(r{W9NyTl6Z}wU0*{8x;qRKH|FyUi&Ef=j?!?ME=b;=9%6#4%(q9GEdwf#MO;TrXmk{-_JeFs#mrIi^uR{`7PcYR?vIzrHwM!T! z9TKqf*>m&@`l3OE&L=U(%xGAj8QUxc#S}Mo)@M6O|L!{0r*4J+cmyfor);c|B4sKa zl&;WE0H>t|0@rW*3|}`@bawC~^C058ccvn8`x%-MrOs}XCB(7mAD0(iVdd$OfVitk z!|8Rm7|zb2h0L710Zb%55Oq#jqkLHR-}nSJQpq4eFCK6_aH6cZ91c;1aH-bxxj(v1 zknFPPg%4!NsLBjq*~l~|O^9!E&#RL@gie}0J&+Xg{(cmet^`)RcF)bzN2gfVVZalb z)ya$rSORniY(dvq!Wbl}xKMVFZa*yJdGCFPoKHxK`z;=W;Lw^iJcZ$hVz^o2gAd02 zb1?-CYLA9q^$|=8mK`b|n>IE@E74(;gcsx+pe6rqao|+HkJ>le5Sr7=$q#z(386`! zE@sB>OT4(L?)A8e0oUs?_;Wk#36bN_TSKg^8m`hdp2{q()-)Zqs%00M{TC*|#&}rx zKO7>U)khiEoa}is5BStkQb6@&2@Y0SW-OswHQ^H_QXyC`oerid@2tft12Ss`zB$_KH%Rtqf|03sY|u z^Zln!T((%0?wTI|REOG8$QlChNR;`g6vd9_ZvXD$ruKO}uR&&t&-flOs?5gs!=!$cc4bCeob}yhC70N|293U{FGCyKwr)boI6(klm)kJY1d(< zw8YoqImJap;9#du`#6mGATwynz$F`!;MV+Ycbmk|q zOZ5u*z7XP0*q9z#G&2f~c2$38-QM)i9nFV?!M#=>r0}GiGQ3Q_l?pHNB>4c52f7b6 zH0!QCQ+yDljGG2;@JyuTHG?$K#@_9sUvo3a{=#Fh&}NRHZ`+%R^ePzPO)Pcnfx0HS zQcc*(?T;dES~{HQgcv|XC@1g7AnH7IWf(U?_cW;2Q|x{(vcTNU6v7xBe_y?wo|r!8 z>eLt>)wAvKxJa358O2mcI_7GAm!6U~n+4m1erG>u^(9cBL>jRcUx_-}em4P1%Mf>H zp|)L^wq6?{O(&;Bk1ocMnq*FK#6vp+p(SrRt#d+6uh(XYI6{C67k^RZZ94kDGinDy zizODgmDDKyGK?kN6!le~%9m}IyuksjDx4AoQjE}={Wg=znEEW&u1sHky*k&O=o*itcf&Vs=H+pQ2j?JI2ZJ95Dh-#h$9#`bhzaaQ zbph0^8U)`lCMg8~a-KsOD1W?2TA;}^6p1hMVWb&jcJrs63gy!fQk_47?{hmI7>jU83 zHMJD-KC);c>NgPgR`^{@Cc5M?M)+sqG-(J>&+kO^Ur{2HYuF*Z;y&f}>W|@9XXaC|Cbe6qA=Y zhN`|~KkkXfdwKPDNz9>zTW3#&4n7BgYp(j@$`!VTg>e>7!b5DKx2z1=)V0N~G8@a|oe#{zLT=KX!n!>ACSYNM+xHQjmU%sibrkHASQAxO zFU&Q=5h|>w*C|y=VwxbpI+B1!A9*Rh3gwI<^fmp9an`~%yCjdC$xROolzWA+iD?Tu z!n=4eZoghNM&Sd@_N`V*7v9~$6pWO<{+x?d>N>Rj+T(!L??Q-waD=VX`ozAPy?tb4 zAl-rsPI4A&Wozm>dMCm6K{_A(N0_Ah#dFAP)GD{-C~;#~Jrp zA(vzhP9^-o0ikdsv`wTU#Sj7iu5pe%5&9#Q)D^m0X+P})P-?nb-S0M#Pl#eTU7&)o zpexIg`p9q;0u;D<)MBw0T&bv_4n2 zFxtU{I@tvzZo4ym@3seBPqnvr3$MaJskNS1}_o^%w=8jro*n?JeQ)bmU;{9 z7WC!IIOcMaX*BcU>+>SOWwq90$SIr zTc5TuxaT__x#DamYiH1#<(=NE6B@0(^Gcn)jeIpCy7x^1Ln;Fn@(C?$&EaJL_ga9i zKA0f)D@Q}Y;&Vvg!zMqF3ZPEy^N({a5T(}gGxH<#=uDJ*l_o-Ox2;5Q?oNf@Pdoka z8GM4rdb(w*Cy|t?xR6=&G$BM&^dWBz*8TOZv6M3MFd!gClsASI5Nq zZjFL~m!4Mnbk(T|oy*d%7{(~B#^E0-i6sz?52DOr<>`lyuG22YH_Y}DLJ3pYIUAqZ zjjO8lBD!Zd^BKbq3!hd(ADgDLMJpHRr2r&58@|J&urH=x z^??Jt+{^2_p4{z}vhh1Df9Rh%N>ql#R+h zN&o%Mf+QZQao<0!=NGQOb|MM~gMq7U4UA7-`*kw;AsqE>FGYJmp4EEDOQfl&hs@a2 z)Z7;Y;HsQd6||y9+c%5mo`moj5z<_sBGD6e`+&bm%9s>0v0?ZNN%O>&DD^F4>ei@t z+FRW#sWJ5FMjDEbXZF75-Tm!ye5f*yvAZ#e6SeL)%7VwE_h|xm&9R8OcPHWAhb3*tv3G5?(0==|t1}ZAI_VQ#$su6cXyv zwC;XBI2)h|EzmVSJdft4LG*LS9&>^PYFVOeM3^h8j)L!8PLJsO&h!KODJe@>%o3{G zUo(@|(kE{$u!AlDIS_`o@X^fUx)DfF5fGWn=*3Aqno)0V!vBUxb5r{Z+knV#30cC^ zy1f;0+(J+JCavsg{q(QV4(n}^;~qjkgdmg34S%axyLD52NKG*?V;BtWs~JZ)+{XX6v36aaz8AN;yPk%wzCx{jIozq?BZXw_ zijj!Xf9n`!`j!329o_WHNjhEMMc6}5h|bmgQ>)9?_;LMY*7;#e$B_f_#eI1NW|M{+ zl?SoRPAF{`Cn1wC;EWA=ARMk*DAbgiT}gyWJ=U?(zZ-GcM3l!`G=Kif!MAtTW8}mvydubhP1?rbfur^^s?Pg|);Y|l;dm73lzumXF?CeqYl@D*m}pxF zHqBBC)Wt?Iq=^&7m{dQ;bz|p*q*We|V&^%L6HR#7 z)TrVHJ8j4ur>m)tx1k-j_?V5?>@gosM%#*TiwW7&6xNH0hM`;3CPOjdHLQ|REZoOF zum4k6MK+oJSp86wfiBxX&~b&|UpGlxGYfU7SSGVVqphOWc5)fecXKRS0lf=B=8_Dg zTpVRFbLD#rN{iX8u3i(xAT=Cb#9xzr+u6#)w-5imHbE0NqU4jVXTf9le?MeHzjH4w z+Ru6cK^e)iQ;ZEaGF>nc)H9jkh6%K%IDbgXsUu1EZ33-ToMdGz2%Ex+sVXutvaq>x z(KS@gaN$lRo(}z+N_UUs57R98Z>@QWl(2vr3p_1%*JwEum*iFS@t<>w*bsvk@=WZF z5TZYNd_umIL4JmH_VR)b8||rOCKygA)gie~Seu$t`;T+qq1EOhs$`E+APp%M%5eRD z%KLEgwXrJRf5ISvXtp#m_v z@ZV3MfgTzqg|a()n-yUa4u0+Ik_EIbJe4Ui6-$!Z>}DCZIn^*qx$~LOAxMIsd?v0s z?=!RaTpI=ngx}=Bk{ytjyvgtDfhCtOX>}2~_ZS=Z`MDS*mr5Sr}`U`AlIO;)5DxvaQ;)@z3s>o0o$ymsHLIsaIWtQGOnUiW8o!5EuPR z96@47Ym+)$G9U zCG5Vlv!oLm>560?H%rJ+3LgDr0J1i(uW-G*|GB7vw*=tT%q%sG?KJS#G-P%kw z613om!tEu#)DrWr;ILGw`?(;-NjUFh_n>N{S3 z*>M8L643TKNCEa(m(y5^GS*>H++2IZwvn(QdN99(;bka@35b91X98TkqS%=*QD$v( zX#~_&SeVXU?5+J82m(KXQqv50SB+yiBlPAwd6R8NQ!1nJ=#R1sys`~N%a??Oyr&&z zus+@kAy+;%rT`pKK81K8hL<;f*RqQW;x&4hQ>-Vk?8*^}DubICL0l9;*(I#Ghx|!0 zU3VQ}TELi=hV6PnA2+;AE_39D(bGlq0OhTmtV?wKtR|!95B0Yua74Fg{K9;oAw{`6 zJAVtq^j$}+KrDZQE~oWo?_)$8OFnMX$T#Si5nhZdLF zVB4MFUw;TzlVoP2tc`K;_=1mOXH)X5 zc?dwBl=r+Zj<^!Bir{YZUp;pb5G3Hw>-8b9vm8@xd2Ta1wnp!PLacIx~FkYHp@{);OYY6>2hi!BGNVPmp)DZ0Wd{ zh(DbJz)50z#24((#$yh1{iI_fkNp1OF4cS?N-7*2!TdE4=go^#thZfSl)aZk>q2rm zqMn{fs&fp)MJ*)9f zDyclHR^b8xncD-a7$`8{gh|4kUY|fUG>H?;eVifoTMmKkBO@apd?UlfI{$N>0wL44DP;Z36*06K#L$NLixgTaIl3vcHW)Kdd^a0$qZ z;^v>gg#L=E{YCXdzJECacyD|ENxiu}--#fc=jrFiz)nwKLc9bD+yJy9h%*RCE9PJy zMC^kA6y)!T2I9!vTa}-{1Z2v(+M)fn;ODIehOB&)u<% zTdDa69_!1==~Dp3K6UYVgj27BdEDvp_`$!CqMQQ0e2g`Q2ne?JZ2@#}K4AzD(%I^x ze)MrlK%DKR<6;0J9HOEkBA|l#>j3D>V|(T9fuDPE2K{k-_zsQZeS2pc%mQc`W9##W z(_(xC9RY)P1_45-v)AX%`EmVj4IUf;`xCH1__MCV0gZgPxst%N{!EF@?x7w+G8~Cc z#=-Z0eLdBYLl+MUFq9Xq^Hl~MuLJz$$RLL%(_!>~+LdHI zK|O2|^u7P)z5J3N|H+^BHGS)){Pv+zc5-t4J~aJ`zW==k$RUW=^{V7u)X?CoYHUa zzYTH<ez21Os9{NUp|PFd>`=@Hu(v;8>&9*n05e`6loVMl9M{?!TtS^@+A z?DF{%8UW0%pq_^W;qWaY0lnW7uAzXueizvSz_X)-3WNg0^L_zZ8`N_2fzy(J?F-h{ zBj2-sgC8IR-uol|84!z(9(~g*OQ2W~jzI$dKp%>W^%J`OB?za(M{ocgy#6JCN;-ZS z{i^@4+1A#MgZ)9kZ*23w=BYzsTtl`FWaI^GNCa~%hcy3GCloyIZ{nnQ+h>0+rRP~S z&m%!E)G=7R(CX4tI|dc+n5wYd1%7yEd=dnce#3 z`%2z-6c^?V*a=NAOQn_|(mULUFO-#EIQWRosmToMl zDrZ*636Yw**Ox|_xBiy*Q98?wCN`a-V0ZtH@-T@q=jh5d9zl`s7G>(Y*qZ0GHNgTgBGJuZF>ML)aw7i># zNKk9jJQx@%5nXI5R@29s96beVcwJvli)$d;KZfUNd6f~--9iM8HR^8ZvA4!xv5}2A z*r>DT@u@kh@` zD*$Tsl%#vynQqeAcrud&Yl^&6HZ5Q147KoBakc7oQ?x*upAnnoF4!c}@caS)2yG2J zop)=*6dtU_n|4-}9gW8DX*s95_Nv~VQG4?V+T;Q~1h4o&sy)s&(_$Q^Scb%>)4-!N zKM!QH;8~rZFSN1=d|t;MkO*Q+!$Z+9hJbPGD*^85`{7va@^@m@=-d&MXvhYXU0o&Z z7!jqQMBv5$PHHzA2aN25)ZKTH+d|MX4`JBF#ZjUL(xL z{9|fznnY2!P!8ps8-&7Y_pOOC6KxQShn^L#euHJ~DnzCAB6Q#|78&7yyQpX-ddbA` zf5+NxV&e(ejT}yQMOEx^m@OZv(N=Z`H=wEr{B0d|vFfWtld|5-^*Wx7czq4zF;TMb zgFx?R};&Et!Bz(b=%k#BC2uqo49hUG5|$3Z5J- z=K>*jjbfA&TiB>IrtdFygCJ$hxx;Rj^+WCIs%fa9{=J0!o#0L=hL2Q}!y?AHAj+Ik z*&*D$mIae7fztP1eFl-!BlObi0JQh&nfZ#xmjs zk10peyUQ2h&d~;v87Hg!*Rh%uL>^jkioy|iQ_w8uoBsz|Rl<__n5GhelV*bmh*iC znxO}`u_oN5!P345O~(|OHoSSt1&ki4>ubl#NmzL(Fg947<`^RW*04!OPrLFFc>rX2 zBX}({$P_iBmY~Bx)=H&ZBDom7vc~EN>GK5_Cq`NSQ%k*%f+^<}6@bIORDD(N&cGwO z>O9!Or%Oi%5K?0-|~{+TbUFql!KcWf4EaXhg_=m(i?`>k zw5R9bSZ-^qDmGE8q9Q?|0+_098$finfEru14jW75};cu=ch}?fR82 zdIR+5i%&n*wSLt?PPn{Bk?tSM0N)T?rq`Z*JpZQ9Of$}-67nD?=4Yb6{QL7%lm%#V zPLpWG^kpq*U6ooz#Mo!;`8&|Nyq+<0=35f#$Se%(axQy0-hRku0Vw>&L02_zGTU|2 z>_4v>AfuyiwW7w{8>bi>q8WNoEli3##9S~Rd2VelWH$pNBFh&VxsVVaj}f}x{V~T# zizW{zi7m6?TXL!qz|P--7h@S1rNvkqQPp9mJBzx6k+YBe z?4_Qr3^B&fmb}@xAxvMYYL5%grOVmZaBouC{BDYu$Ws{mlTxgxzC@ibg{70_2wL#98}XMf%T8=x3*mo>=#mm0hV!YPKc-@t#px6^Fv<%L z*mE7qk4NN}?Ltu+usJe4MI(^%q;u+TxXFokD=W)awcB&(vfqXLYv3xByA<(p)9L9l zMtLuoj-NtecWkABU@z$Jj=&WDU~IQ3(dT)*L2){*haK>G^Q+||Z&<~<)*(pE`i-O5 z;8_`+C?B;|QRwia?CK`04euyn=i;Q5H z&Nj^kS&<0yX;7nRVJ&9q(9)Gxm~T|+Gl_qL@sSu)q5CIJ$J36S*2@Y1 zTAFj~`c-R(7nM1WVMxtiobsHXz^9i`nwJ6cyMAY$W5j~kCTnrX^*lZQ0hZWxbNS|a5ez%3la zee8Nah1_r=3j*H83wu4mY@txRI>=ukWK1CuFsz=Yc6bD~j=y<%N+wyU1cgkM+E-=u z?(4%VeQz3k$XN76FG%78tjMK+_n?2Zd?`rOW0%jk3K5hF&^%KUgmUiGO$D~JR}s1OEvww#{A?MeSY+3bR( z+s*YZ2_ywt<>RJAUWSZ=b(5g<++I~0vgr4bHC9HHA|bZ~+n75m3j#Fw6^%k=%hRhY zXlGN9dZWV2)6t3G!9(cB+oyS@34oim#p%dq2XUC?FO2NPee-dCxI^g#t?n8E)p1n0 z+h|9I;?0iFv~9s)He2zegS{PM`S!F=8YhS9C>?^zo-KFhd!S>Vm1rsx{b9`^VbUMa z1551W^++9jKOkT08Ak~eA9j}$b$Y8=bu5dn;`dqQBW0OfO%imsIP{@wH$EqI!m3Hz|`Sh zN^p>cvzXc)PiJY&d)d8g-obK^50f>My3~_&w;jC;D>hX4wH^P9^X1yhP2*~JA}e&FcqrqPZsSpxqR zU{QYSjN(j8$mvWOp>W-4x!vw_^`T+u`yiLc{Ge^uELOH;`|U7_gsF;vCG?6%ihXp{Ljy z(A_5HVu_nF;;4_aU%X!dC^XnCk7eAlmu%Q2GXY|=dOwqpG~6ctOv`}kgY`xMMDUKG?==w3hh9Rr_YmaaFDoffXB~nKGBq**2d_0 zo&pXE6?psxg9=X+3=*Km`uE))$l)35G&%@)_NpNSYdH+{!1-&PY@qNt9H`TMUSOl8 z-$km?4>t_=2$n1Tq&XKuSHwKY(>WhG}YMw=?p+yJQ zroL)6+pS+^nL=j?jijP?w=3lz1}yqCFqjQKER7?Q>!p-$Nwb>XtS40=(PD-bbXg1T!j9K)7hc-KXkQAqhzd+NF%Ba&um| zBO3|jwJNEFDv&|oN8k-2#^u32DH1&7z_g!;tIv^R-xl>GC@8YWs~U*cvS|&S12}L zgL--98gu8mJkC_Cu-AH{aOhgmkKJJL4^x)YCtW6ZBiwfI+YPf~_s;dJC|9&3?BMc*L^9jj+J4Th-`geuU3aTpHJy-TLs15C6eEM@W(Fn)W#v$88*Uqn@!?5 zGCM8U7t=8v_$eZV)}P?fF$JF%%>(Bch$W>QRKZ+AiWz9I2%0fcmGP_qqvMT?UDQm| z3!w)y*Mdv@5ETwP?$x>@JTHdO6ah6W-W6=Hm4d#Zm>jWnCp(7}K zX^FwVL`I2|eqesCspTqJnj|W{aHYe76wKibdKe>Gx49D0-aJin)AJg;+TUuTS$ag; zE0tR!G-$Cs*{WTl=xe9L3>3yVx89m#fGkv3^qo${Yb~J1ON%;8kBKn|kw%sW<3#Eu zG;6W%*KnmIy5`Mq4+Eo$m-K@Y^`bM}LgjCNkZu5MR^yHCx?-l}i z2d;$A#TH=58P#|LJ*AwRC0@85@QgjaxCmKE9F6me${zi1zdEieliM7aqLTc3owIu2 zQ5vv&kLU^mZl_OWR~f?R&4M&>IfuOwr)1lFY#IRDFdN-#KL0ae8WUjmnI(`p7$s3q znYHgcb0lzcEPA{|FO+nRYwoiNx7fC!wx{s5FmID4vuK>>0oUK{83iT939ILBok7Z5 zAL~cZp(ka2H$f3Z>^y379a|ZrvA{-C3k#=3-+GpCDiQhxBmXTa^%yFhJm}r&HpQ-Y z=pM$4KFOPi0m3jzQ;F()vWYX@@{N0w>2i7h-?hq)>{~@^{D(HiX+jb|A*Mn?W2t_S z9EYA0_lUUpL$ihi0pG@>Z&eC?mif!>Il+XrxCv7{p-`#0`YT|g8}PIV-HG{8T&LOA zLdh?;ozXjh9>W8!-OVjCi@%8&`I=OZEdEK!UOe{B5vqimks}>ugL3s?YbS{*s->ZL z$bEU7j+LSwsb6XftsHv?LWIxvh(8_}_c(AkHi!khOgzrBE@Gv|G7q^cqCt1DJPt>G zm2@Y5(~pOw;(Y5jX}DJoKOgMLhpk%AgF&Q~nc`#@iwq00&4OCfQ;t?@5-4g+{NVT@ zYnhER+2ZW5jgTk(2RHQ=y&H~h*YX|oPI(^d>Fy0pkq*xl!)sN$N*qX!Zq9?`%Y|bC zmO0*1#d@Z`ej^0S)|oR?n4L)%s8`qfgZnH*962)_O8f$XKBIH_68QtJa^GuXjnWwV z8(MftaQxkAq7QSfu0tjd`aEPz4XQ*#>6f`Hx0$zJdXiT>#n;>(ppMzO#p-2y@rP(IYRP;YwfDq1qzW z%Y`fs_B%n*Wo0xRJvuRsW+x(1+uLKx2E;M}?AwI^o)MMHxEn@LtmC6hnsXalMa=^>ga5Ba@kT5RIKFEVa_nV>&@8n7?{H` z{a1`q;!HJ2>+DF!aU@toTHe6u$KKUdL>W|7kQ(aIouOW9bj@)EQ#@`?df$e?i%tiY z3B1L-O3GU`NVm7vf1OEwS8;y)bUK@oq=e82ng<2g!>hUrf-c`^-dq4cV(-suo3{P*re`50TU}Mw9Z;J4Mek=Q`Mz8%wu?nrh*`Wqiar2HzQ9>KzX@ z`ef_==lY)^L*}75yR;;5iP6x-N^{9vdzk z=h{a9If%A1x1svPYnji6rIvdoVbf7skc00Hr|o3)xH&97CZK+?f|7mvXG>G!#|@1) zZt3v^f4oxcE^&-;aAbOedY_t=@JzZ666CCEuG0aHr)MesK?eQ`xke4%TZ)?#!E~)v z$M{XpYF3Ux7h9Yc%$0PZ4ooM@=?DP*84`k6&_Nmo61#mdb#lGk+#W=^ z2==xls$~oD${@@#v#e<#~LFx@X_Lw0BTRjk9PS3rM^r6T)^4gP8 zfuiGo7U#D0bS`m{a8mtUL9#ikPxt~p5^FY@?wW9vj{QkJK7{h5Emdb>u;!jZN}d`6jPJK^K+FOd&l#QAmVYGI+KX}&CeC|#!QO5{y`r#H@bzV}TN zEk_|(f3I4{q!1s`PR?VRk4xHGrsl@)CRY`-raE=Zh(aa*t;`AE;F0jb^4jgA|J=fA znVBvAx4e>AAA-bOTDIO+8s`;6Q??1_xg($QP5QRxOx10!Sm+8w)>Wng^BX&`hBLC% zWs4+>M}u%lOV0{BLtir7DnGlz0)ggqXRzz*nNxHEU1LJA*OOQwfZSXSG`@B#^Bch5 zx4jKQ^4s3I%Dkex8g|*xjzCufFIsPf(KZS;GEV7|oWnLlPDQwclO+2)q6@i~rMa(^ zKDEUC1jWwHLXGN@{qL-Q*!QEOuGiJxxUEuzCKu~z7p%keBK8>jj_!jdVWnqe?De_0 z!`@?XruaDjXkT!o#-d+7Eaoz>R;b#Xsfej}H@%TZn)r9Pg;=>m6w>vWFz9{F@6gD& z!8quSZWx`0AGVB41IIOrT#l|785&vev3%ERnM4Lmc-mhuA}&?buDHB5?9H!Fc^@22 zEt5GSPJ`kr8>51ujvlMjK2_idaTO-B2OOOHrAqp|pS&hUCqq2;B2EZnN<=p@zd-Af0S4GN} z==T*PD1>em zacT7#bnsQbZN4%I;}#c7Sp^#mPv(f?I?zTGj}cYCSBEjkR*|(U^n4j%NI!M1h>Lcl zAf2&z!U)Ue?Zn2@_cF)^WXd&pN+YXa|0vOIf63`vQ1Iqga6tQJ!+WY480c>E@`_zl9AE)DUg`wJX~I{| zO=nSFSM=|uzmDQJ+UVYH*rwkcuE!3i;YY#ZdmA5fFSVz4``V_F)q2Se07> z!l6^_glTg#sA{y=8_@e>4Jr7mI5o#gn*YFHgUjbnksY?Ln7j7nv(tIbnlc+?tT08x z9;r|ngGU*N0%hfTb5z_u=s~+;jWNHN3mD1J)p9NtTA>+E2+Rn`7JkVKqo+{YdI}3? z7C37)a&7b?=>EcCXlkSsT_9m;1sW>2YO#_aMT(dYMMz{fCrHw(fpLG{MY|J+@L2JbqQ#91sBV))ktzwN}+T z6yBUC1JHcE?q0WRWHud)>9R_Nuz_IKf-e@@l6C2E{+6Y{Dr@R;m3N{53B}8x;2xBp z(LhmZ5R>f8nN`l5Z~SSk!wt`l4o$b0^<@8^pcv=RTF69s%U)8*oV~f+Sa+!0(Fi$i z?pwU1_^MB5b&wMCEPZQ*4H*vL^Fjp9F!`1{AtKEM2z3hw|NL0SgGz2MSQn;KS(?Vt zd#X&@vIgsnnf0+657yJd26-Z8_&n6VCf^X7OUCL_b4EITUXEg=i)lyakwfgG#XH!W zwrz!ehU%0Na6#}x61nBR3dO{WI?hW29caHt!xVSNlbfE_WLKF`VyoYge`Zo7bgQNs z5@`y4484n2hwXwL7TiB3$Z44mZh;)lRsfja+f*8ue?x0bzwnnBO;ePvL|`!Sa2o~_goeXh(KEQVDZZsYE(rm@Ke~d z4$Jv$3{?K@=MtwgB?8@CTjByfIju!~)sgq6adY|EXeWda?QvE`f1qS4y*~$QqGI>G zf+dy^{}P%tj0jQZI}Tz#FLuUc=xLMuL0M zB9~-K4fj&=r%7aFu1|XKX=s$eJU5lHfxct6LNpIb|C{`*5)6dqw_VB`Cy^Z60F%UG zN@7u^eU>GtSDK{qZ2M$$5%N^J+O(PWj)lQNI-dTBJAnqT&_K=zi%N}xq7d>;_r=nT zMGzk&?xufH!+}qCjlDNs6Vx%K0VETn@l1te7XTjn)bUlCh-q-+fV9l|d*^>GVvPM? zKnv6V0$P|k82=B#!bHHv#>Db}s{b2kVP|D$BKZH`{{ztC49>EBvr11m7``it*^{PN zBy!jDrq6(Z$sD~9A*V}JC_*F!5g36a&>fuJz6FtSqDK50Tu*@6a)z^1+?-q%U>j*u<&rOu+R{Ij*giE3hYA*cE8b| zz$YJ~{rb;QK~$)pqFN3%dhqjtcwYcl4+DV&4n%Syu-HUYG{CU1km=7+guo=gQlGyZ zT0sw^Jjh?z8$i1fqE3Khc?LIx_Nk6Iek}tYA~7KW?Hd6!-@C6xVCvM170;cAB`ig^~Bq z!TpMS{(3(Zf(Zy_*Dc_Y-y7G!9E13H-);L)AzkgiSc4;L(!)4#FRvlyRX;a@s(`;m zPC@zpiG_uQi%f?C;p_-V``2`j)IFN{=*Mk{&(bkD1T5>&2cXU)aDb2@okI0~{`&a! z_#ojzpF%=Ge{BbONB|%}odP+E0_5S?C~)ucuGLtFpW>r*_>iywn~{bM;6RbTpFdyD z_|Wv|AUJ66_-}h(s0|B^j7;@t@5~QSMls!2WM`j`x6n8B+iOqO>66U-ZyO|<9}U*u>XBzTgF@4&mTh&Z$W*3A8vjAi>Lu)eq8?|!0%pdYCM1ON?5yq z5AWYrWeku}L_UmL(9aiE0Uy>pKzI})%E|RFIIZ9FX`cpI&>*201paxQ87M`t&Wzti z{j#I4zYmvpef+mgpnk^FT;ip@S@0BnWQ3#?;D7`P2ZFy2lUgDOSWvyTa{$OU7^8v+ zMy!B=u>bmlP(UpN$^M!csj)xN`q)L6BWRw&Jxb%oTlVuRb8lqzIOuW1V^gl{ z;|8kBE^&#)>ePPC=cSedY{h5{zc3w6qHO3I7GRQaa~>FUYx15O6EY)Y24zi{wr7}z z2v2|2*iZRT%GHfm+VVN;N(H|d^S%%He>0A&_l$dwkW|_`(2pp^!gqs?$O46WuJS7e zFE^l6ct(7lBfr1(^SOKQVI7!@+gggumy9(5gB~;G9{=cnJFl6LmfrR9|le+Pvj0vp&mQ zj=V}$MG9%lGMt2~aAu@f7zmY0$vZH_lWA1dr1nySXjavjQ>jG;{asx^5y|SU)`5%7 z0Do1_nWrZ|WX>?39mh?@jgAF$3yspXim~Gp_(N6K*67^qxG58VUxe7po_R~KxpES; z+!t?`mjf~mF+R?R&Pg_6vo-;hrPG@!xGnxtt+DO+8+m~V>lYYpR1pd?gWu$r;+moo z*qR)E-rXHKLwi&ex9P~^sYzwqPYQe9un~H<_%3P|H+35~Rz@7*Cy z^8D(xMi#|gL|YWhGe^<{`fdvavz<3phTU{)XAZ$~pCd;%RiDWgK(j42U5*`xGSMis;t<#kB8f{M0s9=DJ^_K zV+Zrw%2p3Qk%}j=6&n{;vH7Ceu@Uq?)Yf@n>Ks7^Ca$7L>0J5Q$44+c8cgh4!&2{4 zLm^(>aYuF9M^~paKPgpZ8C1;T{DNq$U0?UA9JZ`cQA*S>)>eSfQtpkXFHoT6bz{m- zJgGsADYKeMm$xZj>#8%#_!DE@J!Wf`F;gA+rSIe4*bDQYAW_xtA2(vFwTTFjC(pLlP-o;ntwP4tE!tt@wYrc|yC>{WP@<@O3mUC*a_EWWBw?rr@Ox^dXu;uz{d?ryejqF?PCZ0da8^L*a~xnS{f2^xq~u~A`h$^@jD4q)ys ze(2jlBu1N?Drt`4;56Xx%S~H_ELo)8Y+baA5&HB(lIy~N#B}~5YaeOfUd*Nn zJd|_(#9#vPL4<#i9$|PEeCnmd>~JVj`liSJ*o${=P3p$&P5t0d=*|zHSd%SsIW79* zac?wkyL*LF9j(!5d^ARQIwH^Rx1Q?bbm@s;UK{C`YoTk>*WbK%_g1KwSEBKEd8%d4 z%AHDkOyLf%39BN2P8i#vdP}DA%f73RC~ej}BMhA7aGMR4>m2N0Fu@P=GXG9HO&h)? zp2vmmcMOvjwRxK}?#=R`%I(4eac*QKP#Tk=`yI@S@5HJ!I;m}dX<+To$B>h?*K-9*0yJ6ky#s0vLMSVSk_;pIPLl+VnX&t$LJ z-nstV+T&rAYDxW1P5caMlG`MwU8qcN`cD$B9?I&3SrK(I(WpW3 zlu>yeQ9V^;;YM09U&%T70ZqQa(DArGS2Ns=1DmAAet-o+c^wSNL2qAp2bLv6q(rgF zVCvjm1(*88o}FAb=q!xzISi+AL_9}JiB_AQm?a6M{=!~YGPrlDw#lC7ugF;)EESqx z!;@f*6LVK>fjP42^?9v=E6N43@<9=>hWPLDW%KLqA@2~vL z{tVX&tcDgxao!2%hjFGBpN^@6{~>yk3rNrtNU5v76slr z2`Nu7N`54do2-3NJdD|b7TxbVBnWf*!vc4w`=adE=mBkbAYSK6s-#A80qxA`5i&Lj zI#=)&PBa-kV!=*Jb4>P>`H3GBpQwE;6;U}uur zLko2k`3V(3$@dEHs{7T>GZCBwM%sqBOEtAe5sz!OI0b3g{ugYBA43vL?;R8S$gE%` zIlq>ida%42r0Y;dCEvND88}o7d@AQ7OMcd+c4jQ5{NWV&gX&DanitA9d!F<5QNeSHMR$7 z3={+kfBh6}nSQQ7V6xZb{CZ1wGH9?B>7z8sQ0~_PXjVW)a-bG4IAJ$OvZl^5wmpeA zDCXqgQi~JNg1!sr4**Q+%0$-xC%f3wOxXYt>#Q4OI4KSm?r3j+l9mp&Q#+;g zZ+k$NlDJP}-!3xk@0sjwLyBCL1r=+q>?J1JJH`|#qO#K{dSq?wHfn|!gAVIx88AjD zSdrP(Bb|sU8u_O=!$(8A`(jwaRLFFTY(bn1E^W-ABs#?H4hMK4E`~Y+@1=MHli#YZXdo^(By+)6e!H0+0I!SV_=;N$bvV;8qlevmK1HsQ z0EJ-!W3N>d%~aP}3*SZQM#D>3D2s|b!xo(^Tjt*wX3ie~jQM_MRo^aOw`VthA@&NF zSm{f8S)~}@+Ng$wzl@h9;z&xLXHM>lrPefg%zE~71<7`-+} zNa#&0HJBK4$IEdK-=Xqq!a%ug*fRh0f1rg2-FA@VAHhD?JSvLDHTp>Bu|f%w$}aXB z)oWOK;hs4%;XsUCn5;Om1Fl;8zISUEj1g$A|8oY1C;J~|s}{hZ@&zj_kyVMn(ieQ+ zSa)ZFqAQqid9)v`EUBUF+^Z2X?^bopnnYALZ<~FuR$(tds}IVBrzHvjVl8TtH1gw3 z>WOH}b~M=(2Wtl_YFGOt{6xi2)XNenE%$3eTPsLe{&PSyu9;+zGG=pHV0yD$(d3Ci zOdQ0n?Ke34E4c?W)u7g*|(pcTg9k7uuE zGrkrY`MH6N%5}znFmHXF7T_JOBYAa`NfpUU6xBWYs+(E)+o7EzuLTWpRR^-#xAJfi zfoX#iRUEDJE~AZ`bj(&FWjA&Ym#wwxoMe8BLbKp%*diT;6%o`G9&&D@>1QnqWSK{l z47uBT18iQ5IXQ_1W4DGNCH6MTB5^<&Igfhx1%tfR9Rl-_SdP*YL)fcwqIMziR98Wc zp-xKshYuyhsqhw!m8c`G0HF=?>%uxUXX?YvRXt+w>N@;}8EXhQQG{cg45qv7)*Ajq zZQ;EO6>oB{B6znCe#tC#bMEI4j$;STN7=61eo|F0B1L*G z6~zh^aPqa8H{8A4f8E-TGFm}f76S^XA`_~a=>$7~3$IbY(>6Ww1v#p3MRG0gHpUn1 z=!Z{fNOF7;;DS*X`Qe!_5em%xut6%i`($2l-nYo){zFUTXW0&zhhzLV^snTq;fms8cilPk};y7ym4pzKI8ORtS~l0F-ayK z6I+AyG(`K%lxl(O9YocwqS6z@w@mMmFfd7EOol8t4Rl-ebUduO@^NhM`)+Tn(1ato zy8ZaQbl;C#fXC!H7k5=+ilv6uDz%4~$X4dq#0oUG&hM9*(qtX)aD1FII#rQYF_NU6 zNKw*%-A@SGrPF8_>&--8J|g~HV;Qvk`hTXL#Y_3WEAM*JM70Hn&=(n6d3(+^^;W6H z%E>nO?U!FLdsrqdjVygIJshBpFVW9+)9?2KCKJW z_7^Yh2`^zz7|!0;>llJbZCm*61d#xd%38yXMjZ?xY;ms{k0|hfx(>#)Z`b>aV5fRu zw5pA@s(Qz{u^27ycHrc^Wb<_&4DWXR(cPDFISdzlrtSfAja+YRPb1eZK5iPjdIOrW z$Ag|bhRk{a))dU{-~cCD`@^i$oZiM7=kBswZId#gQx9769Y0nJ8vlWJ(1tRmW9~YF zhZ_v5o5Dt=Qr{5*{R>grn-PtUFTE}*xl7M4&rX-2#pbRhdBW|RP{Tn>EXYUo`W%^~ ziU`m?05hF(!5_X`NYk>RrDF3?x+>0Igo+6l&tni)Cu>eN%@b>f6Vu@JM!||=84E98XsO~Dld0%Uy;KRcxJN2SU1acN0AUKeT!()noUFm%cJ$CzCxRWTiA-U zgk5?U_aFIgZfz0Jd_!k?)ehs&%kB0U9)7c7(B9do{HrUv-$U7av>w(Q9V#|Q$0)=! zOUtzQ8Nqr4Qd#75L}p+p{7!GjOOSdV&r8kfIPPrxU4A1Me0G&^8r%-YnO*0?jjK0p zfjOyb-%o`|!|K%US_U#W{moWr_X|OIWi|RKcNn}~9!hC3j3l-7lMfZg^n4)yOvi$P zb}7xwFRKRas)()#j7!Dd#>>&P*^TV`95~aa8o!eElXPR~Luws3y&sf zxYkkfV-=GLOH=84lu>A}_T5Lzu)xA?s3F-b!Q8{E27;|KUfXCY{jYS zKw*!BPK28axbGZe?bnf9<6;vtcIsljc?(kVqRj&`#L3jOde&!>6cnEKW9n^e)u(J6 zN-eqWn>N38AkD2t`9*XHkDYlv5i~BHlIW=0k`L3B+U+ihYe`@C0}(&z428#^P(YNVGiN`Z=E))`f zBV)O88K%<6kD`N+wQRPzBeD|mGM{HR{qFM!+`R!(i<65<56MNkNZG?YA#*-0Bh(xZ zIZ416RY@r|-9P1sY42=psjT%_R$@=1SB(Rq*0I5ZvlzUP26m?J&BB6aVpiAdfh3k- zufx+N!MxtL_gYKij>8vO)L|S@9_&vQL_M}+7YQj+@MiSRAtGG^`-LGxAus0oYza*u z($t(_5aw>x@9h4uUyzcrEN_;FIAwLi>k8QQcoW=J!H~z0hHdBZGsj7~%}O^MQ9RLL zpkN;Vb|pdAJOglJpR@WwA)Y&v zvG3f_5bsUW6Rjmk6rXmI_1!LqZ!Jr@g$XB{pQ|D3uEw8?GGH0H8x{N&OG2YOImu*H z5k4qL|JZx}gng{YyyT2^FOddbBqojj>0DYw z<~7$`?b?l}5G{$RDIfI|<0FTrWSIkohLdaqx>`HZ{*oyjkNFEF^r1L0jhtB3t(xGq z7i2mpcTj3ovor2FvpzL7`Sh0P;fkQ2*v4v1#Gkd$?49Rjpw$e?>UGoGaF5O!vvd@vC|LtPg9X<{bXAGBgFS5Z1cxW>~67oouGP5fm0$ z^I0%i_ZL6#)|noEvMm`~uHlUVDel3~J&fZR(L7G`-nS;=yQ8pliJb$M0LmnGwsez;wJBn1DO^P$U2oS}h*RpeeZkISG7lLYf z&a0Igwz@6>0(5uXCT_gFs$34R)H-NFJG%hk->xXN->yv_yijS=63bm)3b^nH&q2LY ztK*mcPua$3FkT28!yG1FS|*jYdwbMgA3MTcm3)-^TTZ@M_Th6sVswvtGME#+NZX+C zq|TQDExW;=y_Ck*O;X%{=Y)+&xB+_oy;v`pmGdUeKC%Mw>dgN1pFI%D&%6TY1m6Ah zEdkpkT7TT!mE2AGKsLM>b`?A0mOItq@ghh|2se6Rj*9YNZ*}uU8M_mRKQYp4j6$;O zW5cCa)O=c*-TIhk=IU$P!J{y24f5Bzk=JEJrEP$ zPzW*vKxy~=b9O)YJ>X3m|0@e-`5##@6D#ZgB{yK?WMyLeUtus411AT||6>>&-3qFp z9o0%3+10f~(%#k8rA5FIqh;>~e~ZNf;?}m0u)V{}0n)a%;kB3KbNi=yr>efdTN-^~ zB`8>0JxX9`X$OSPzt*v4N41p@~RAu}qs|c#C}9R=$j0^2`0F*P_cH4US0VtV?4FYG0*>))PGRYEBc zfJk6v2ktDCkI>|HZ{N()W)Bwfk(UQxiHrpR4G;hS0Y*T%zZd;q?f?;6prfUUksUz6 z$k_sD3yNrBWCKvMH?agdd;Av#HNSqrIcKAPqgh&C=NdpbB&X zI=TW)0ly0dC>q%U|E`P?fdZgzVd?a@T+QCh+0Doi2mlFeEKPuRP9P5#J5!(|02Cac zCL<3}asb->Z7lz{0X^WK-T+t_S^h)!kN0mPOS`|EjZ94JZ5@p4JS^?Z0cMsqK!B2@ zJfpL_Gd;k_&h)pTk&Tl*$lu7-$kN8h7-aBQ=0*TX5fy+DsK9^9b24$XbZ~ZJbh5Ph zT_V$O8c>%d>`cY%ZEb;e&Q1uw^Aopp1e$=_?!ol;WUcM&-R!*n0cMtVre?p3Fm-WY z(y+7q-!OO|W6$o&4bOHK!{af+h2!VwKU}|aN3@`?oTiPM~VF!tUW`ASQ z@EtAP0lLhf-D3eT|Ni{vmmX-nOzrJ#JpP#fHD4wLZBcbidAh$V{*O*n)ZQK7#lXf6 zU|{870q}5h1GqSt0Y3lDqHJXOPZrF7TxIOc>;XJ~CkyJ*|0L}C&-|(XxeXe?e{(6? zgO(Nup#HPtADKCrO+bHG{+~DfFPHzn1^>c8~=AGMLKrH#iw z2B4jFaR%+bf<5R8?Ebf@7Vz(zRREe=y4e12t&Fn~=nh2e%x(VDMoTA2OLw5DvZb?$ z#ox>Fw_M}*k=a<<0hR5YEPp>$01Pb5%>PRVx-1iG(38Unw2^|CZR={_+34 z15Bp>f}mak|1f~E`g;NYW55b33A7bPP8NS=^V{c>3+R^q(16-t_6LFt%>D)0e?!ZE zaf0w4$iem-y8g+6<+tDuImd5vdl$!lMFSa_{|kcn|76Dw>a&H1g9Xs;Ult(Qzv454 z%C!0y1a;W@Ul7#Bf0qX8vF*P~1NH9@GY81k4s=NW(t~QZ|C0>J+5W%Opkf^Us6k{7 zpx2|F4bbdQ&)Hf2kvjg@DA+++JAj^O_J7954r+vhjf>O2_(76SfA#|u?UReWGtku7 z=D$LK9@hU+{C5;Ko_{p|EoS*W#D9%|1=P_$VL3ofPC#4B|K2{1-!?$ke+_~IMB)T` z?fjtu)$}I_=<&_uY~cv}*Fu6C*2(4)tG;Kob{7&@7z)dP#vU;eYU7|8W3;?m!cSrFnZ3zF@1`;P(4! z5kfbHy>WpViao6~8V0W=#}1cA6xd{%s;s~b$6Jx);U2Wr9SQ0i;bqbnuj7__*s#__ zm9{7EXM=dvsl7IY#YvojiNa%%x_)AKA_jHgeeah~-kJf{kS*X{vJ|Aam?De2;PtI&!*Miq z4yR*m%QlI{=8)QRPf_*W@#spQl463F8Xu1d5=-WD)JCgxf^9Kb0t|ch#j5OvyUZ0= zi?I{EQiA21k_}fKhbkh^UiV{8o#!DeHFm8wZJ4SBc5)r5T`xx%F?dcLB}#a~p84Q& zH17JxL6fx~Ygi4SV*K?n5%Om$IJI;jTwJv^H25zf3ai z2fFshoJ?);^}&-V8X5ux4y#*}bHowp2tVW0WfQyH^0W{_6GXpH&dSJbVQWxE95m3Y z-8)94$*ekT>vu!A5AXlEgH@_g`X*9Uk;|QhUI;G@8yoY`1QE71=7J+Nxy38KjuEyc z8y@pY9h!fwRnc=j)_HxdSHQr`&sb`%CW?xrXp@Ag^F=96{(L4chJ#R_r@j~KmA6=~ zl@{9jOcEh5_|2;~0QAk9) zB|g$HAHIjqu=0&_Gt^E#1diZa&El%V0SQZhnr1b)PwXJ?*YE44M%Sjy7JUm|et!J6 zFmtSKr=vM3kD=Hh<{SgQkbx;f=U@mAK}3W+Qh^r!QN9zW*RwSIY-&?@$RVa(We~Ex zVDq^0B%f$M-*izBhV1aWQL*GY1zhVZin|sxMtwocVwCEuSF>w(*Ev5cOu+G+$$VnW zR{NZ&@QaitbcKJoDeBDiMR~@;|2;p8%zGt+J*Tz8w!R_BJ9C%n-knn%<4vm-I%k;B z`?th|yo)E~99!iw8h5J4omfg%PA@|Tvv~{D2;94gXEfMoOQ;v~OJaqxiH~iHu;te% zmn1bpni=1waowjgwSi|QgbtQpf5;(qr=uRFa;^3VsiaZjWU@o_>GaLgq`nr?5k%Ok z^ig5*aBeNiaKYpoaqm%yXF*K8Bh_Xuxct$a@j&-;!gNhf^e6>u8d+#C!MxbWqw!qi zYweW&Zd0itOzaI0P6h|#k!3PnP*kTZw0jj2Z+F4|>csg6AJu~i^T9nH->S(Y=OY`9 zM{G~BQa!3LN1-PnS~Jp(d+)fp+bQaNiCL)|tMZhrrY2$o>MQTfyG- z9oa8Ybs{-5U&hnQ!gTxKVcN>EGKe)IdIesfJTNwmGP-|Yl59>Z2F%Fk=cBW zR#+;ZVl~H9I$UAI3Q_OlO2m6?bel-@qY-L>EE0p;$W75S#^Tc7w$A)gawE%EZi0-U z_&tgC8hI&Nis<{-!p$Y`%!Juql5mZ}HED;cqsxep`v0(-s+=(7N!fOB)T7 zIdLSUI-9N$o%5i?l#`HR+1sf0-jMgTk zLwm&dEfIKosRrgpbN1DL`H?D`8p~xRb4LIhfE(sNl*XD$@KR@l74ck5f%KLp%*Bo< z{04^olbt$&sGU&&{0uNgYWQW^g36(MdWX(E*>z0YC$v`R^C-I1JR7<=F>h+1eq(Gy z-b_v$EoBS)rNg4*wtMvR76%ZLBsW&y0yWcG_)^VaG-i)f%tl`yX4Gf7k>L}MaO>0t z@NFGlxTfv(D{O9)cchy_q+_B97SDz1{KkvCa zTM7t4^^OVa>Mghovj{Q6fu|lbxkNj&a9oR)8 zuhMG+)2QK{aL~a_&P>c6HW>CWdvi>EnJfZIuqe{!(IN7(4@Ps<(%A4K@qXsaI5H|p z_}BdFd@!j`uHO8ooUpS}OS{6((Vw(KqXf5xV;y90D=4a|Nwj~tCjZ#y@?ojx0k};yYG_TXj?_q9S)%u)7nPGk zb&)~0n_8mcL`0U1&50KpNGp>6bG2X;30mj1bV$OfUU4eQw%%m2^pcTr11yFx1UJQQBJGdBpDTEP`vO;w)iqcnC%{o6DoyV z`!GK!j?>Ae4(wE6_H$C^lAmNg#qlG0VElaVfON^oag)K>^ptjiU)sHv_eexFmD|){eIL=HwmK!vzqC%;U-Jlx2$M( zc0wlOO2xMmDuVP|NV%0Bw!IGY%d4ZBfHEbE30aZB^o+W8LH%M39@m_iI1gGvSSlG? z=jNG8UQ@;Q=c$1ON%Av<`$nqnxxHi|&!%J@s)*o8nSS9J4YZfJFace=Jo~x0YL)%s zrkK`?y#xuy` zkr$KW{)GT0TyhaFWpLnY|FZzzfw zura3`o&kPKtw23YG3GGEoT62dfai!quF+7pc*I|`aqZ%1BQ&4^f0)G|C3g_Z%osQ( zMxz_0ZdQnapKD)Me$xD$pyZJEW6;?|%&>JXcbhpgrmxn&lAELag9vp$)lot?i);oW z4H7{WX3~tn4?7_VhnScxqfhAhksqrbep1nl7_cu9clyYE$g84jFr63FH(A&G%sGI; zETJ>tOh5F3g^5p7Zg~p{^aCi7)>X{paM+=LCA8!g<#?st9|%r$-?ZxxCheAZf)H1J z6Ku@|l4Hr-teogK+bw-JLi|uFh<%@zZ~cC=(oLw_p?9nPZ7i|(Wy`rC8;5x1JDC9s zpnay#)Unp9=-Cu0VTV3(TKGzjZm}`GBT{Sy)wAWs4&4jth*|YSqq0lxGKz?vaxP%` zh>*1UBf&sL-0JRRK8w6)b=d^Qd)eJu@P>=;*RQ>o&BPA4xXjpN04%U^?GRZE^msWg z9eMG0Mg_^g%=Tqiz7Yw{S~>Sp4RiA{QAf#qk;=VRpqm;#5&yEDn+{xFzyHMDCO52N`#`I1sX0wO0ssL=!s4`u!y z3rTDxbugWatCqoVaETnMszSka1NJBJPaN;M5K5=MsBxr9PHnP2^Y?b96JffEZs@zg z0d~~Erb{uY_Ote+cf|%FCrS3KC_DUXZKd!K;Wb2q$la(ABw_9faWY-{2pzvFbjeM{ z)qm>Dqdg!^_geKeZ3ZS#(<+JQnDq$gKN$q zaH^&6J~m1X?H*9y_qO~+!jGi7d)b*A7tKd8=hwVUjQEe>q=dAHuCk0G(k`-qIJ{wq zb!h*0dOqNdl2~2+-{Y4g)G=9yY*j7!;lK7cvKiqp4<7fEk(g~K3?4$aMXF^{YGHe6 z63&BF!mJp)A(x_T_7yiUrKW3{AOs-qu_P{XAeaF z%toWtm-HbZeJ_NhOt>G@?9Tx6eD0eQB3e23HhY$-pbYnAv7|B zEr)FN0-wh*A^4ae+7kW<&CsQfo2-_DoS@DG>f3{cM*6eT4?bU?m2arc==Y|A!I47* za6eE+Qy!2>&09NRp}}o5p6y+WULteo(Cgl*P7M%22Ypzrgj7iuam>e5W@mNwK+37C_TrA0lT?Np0q_QpCCb04tpDza3J z@|P1<6i$me#!9<0tLocbdkA7(uP*GdoOjIQnt~3cAj-#dkDMA~FXH-q%01F_8th*a zHk8j{?K{W&+L0~p=J2kJ&+4*T1nRK@e%w$^^<(1)D3lt;X*b19+!wK$cuf>EXNj75m#ONS3H)%TuAMqOIQ`pDj zUmiERMrMUU{@=1h*n}i*V;EJ@azs(UH=y?dQlU*IG`y9rzR|U45R=zCP6;f9MZ{GS z+Qq%pA8;6sCKu3eZZXYX=sWUp2odo7SiC?LRLN3YkkYY)4JzNb%ZfKo4RsLk3Y{@P z9ZF}Ol^zqM9;tWH>X<3=^v+Ztt>4j{xg2(g#;eOjHT|;4I3G0;@mZ->|TL$Cm?F|==LZS?yX-Ti|lx5 z32pR3bytbJZhBK+v=gm~`Hk?RudTJ2>o+m(wI73+!J){=Jn9ZQ!m;}3L8AkZ#lg9j zTn-m3Gql^Tnf!d5?mfn`Gft^je0W$QPCJ;t z?>a4|h4;bSxj;wk|H3_}b{Z^@LiPy%ScofjB@32I6F|A~-Buxjf)HvaceD#O<0F(v zSNG)T6_hHSBgChT!mgPRLdRsZXPDxWPnI?amu?$+Z?1<-s=^==Trf z3;xv^QS(o9dAQ`majg562I(cD*`BUY=Y~XjVn>5|CUl4X0aH<}!S&FzK|dZED)$r! znOoXJ@A4Ktjt#xS*bk7dT*qA7wrV1j&=;xewqL^(EX^AQpHuZa#w>R(HH{0<Q z8#GMQQ~7`GQKnhGo-F?2NIQeR#jX|Nut1)RNTlIK%n;T1Z7NVSg76xU6BSDEX*oAN zAHm2_Y1I4EPK0S1bl&>bINK2cWF!7LmhL2|G26O3rX=gNu2j#Z6Izx33-`=(r&(>*t}6e86Q1ghf_A+PNS!Pf|armZs{ zGGF_m;efNz^oP-XbN8eo1@n^Gi`Ml_#vgBX4r^Uy*6#L)A({{&^Hea_F|M=0#cO7N zUO!piJEi1X2sUOIhQ#L>MVBa7Q~O#`C|8Cy4-E^VA~Z;8*(A#cObJ1xXNT=JTs(Ze zg<6X*q3n*daJd1a>>v>V4<$#73XGnmc#PpO`KXNaPKx^7I#U03>BGxSL+I8C^zip? ztmJ7lFwYPuqNtoyx_61Y{)Ej~%gsThn%mNm=?OQ>_zlr4Qp-7a((aMlA5hzm)N9A4 z%Z9<7;uv}z)S_U@y2a{hb3gGS1VZxgxC2dK^i*fOJjH4t!&-$rlv>s18TaWn)!!RF z8@^iv1i%ki$7flyiew7zQ4$5|*RCDnt9nien4w1;rO@1)TZybCQn`m2iz=J)SM3z( zPsh;DkfHYGxz6p`E+Wd3#n;gCSJp78)j|yQKnN&|^7v>I>#nG>Z9R&x(kTX+9>3dm z1(W22m$*sLJ!eGC8=uipR16d+B;M&cAhp25jw)th4SYE!jz~Tk?|kx9{7db3p8}*(sCz!>y z+r+Nfwap+%h({yUJHQ@_Ir3gsUlXJ6&7P>BK?xM>g2f+VZI5PaH4sALaC?!&+lpx{ zHt{v7IKL9^Wb0N&>THx^pj!=O)O&w`RADC_+85);)&Wi|p5fZ>?mdc!I9<!{ZRNNoqE`Ek$b=fcA9Q=-d5oq((Ax|5i&18`t}K3Xu3TFvq?Zg0tGk;DY(t~ zMJ~X{jP=sZC|=JZY3Kp~Vy2i8$gcd^1RBi~`3 zOQ}3{DhPln+i6u>lhK3ajh3IOo>K!uF&@WHA&$T53m`F@X){Om!2pYM70u7w;wHl` z%yipc-s&vEo(7?M{DTIZ8-ym7>Ivj77qFbc7-DaPqHeFu>$G5rxYFfw*Y~Pm;o|(6 zfGJf+IV}(P8W5(8k|~r@_cmwX5B>D_N=tA+RQ1z|Pq56Ai9gCNV^_A$T^OguN6Qlm zm22EL_}?5*mSOqS-VY*vk=s$W7;*{xq&RS6Se1PjkwNPsLlnX*ny*5CeHK|lI0T!= zu~4a>ESsX9#&Bcukg5;7EU{d0aWAK}5pasxMe1-gifF!yxgi~%T4bR@tE8!bjQW1n z8G21kh&F}lY3$;G!vy`^bGm0ccm}#a($=6CF33$xOG2`zr0LwD4t;Q)+epCGP%l^r zfBS0kixZIz)0dysbRSG>AMw~Oe@fQE!*#X0Ze?P-Lh>8ool}x`=1N0O%@VT@Y<~@I zF6;Sel#j*Z^`20Zu-94I?gc9IiNT-no$~Qe*_r{?`z_y!hX9ox>gkP#LiO;Ic8)gs;_#`x1i_SPW4z9*M(xqBufxi$vU8NRldT(ES?p^n(s7N8aXk zsf_ElcnkWFtenJ9WiZwllkP#i!luq8m!2({Tgt2@<5FyGce^*ql)y3hG&0mV{G4xT z$FeRi?qU94jn_k&gayafvkIDC2xOLU-r=85+KUllOZo9eAO)HWb_ayzZ#?B^2ZaendlXV#Qhy76s?*?Q1 zY&lnM>g4Dg3JHMc+5ImvC(bgo+4} z&$T=)*l*O<(UnF|<_08AF7cDs4(|7^e=^Y-v~+kua_*W;wM&tUdgG2tJ(!42eXWnz zK4dgWcdPYZUerQm57&Zd3Lu6;m6U^aztG`FF`E`Ah@9cPn=6P+eHms4Hw-bg8lTVBB!6dwgH*5%PY)UFvDS9s81% ze1vhwAQE~}?wl~onU0w(7f&pOladKU2uw6y3Hi}3FKMyrWIoYjdN=z?YCqULp&A-z zqWpYh;8cAT9ae?;GLhXw#*?0&akQq3fE;r#hHc7Nj0bT|rD7?e>>-y;lMBZ@&F)5Jhkmb%_W3%>))fuF%Ho>YDx%c+3K*G8@<{yLv_ zMIgtbJo*Rm2vRzz_B;-j%(Pb#hi&;p56PzJ)`};$Q6gz)L>3t!!TqLm3>0!?Ro3g7zF`8f6s-|C#ERL1t;wM zfJmi4u`Ep`p^T%@kTbFJEX}&qjcI`L$P8+FMWKQ?jv@3*YEP8_f~J}jt99;F zqCL&{Vq4zwpbERbPVS!j{qvwDvi(Ixg9Y7+E(6vm6dCFl8zD6_Eg%}Xg++M;54U-1 zmPtcDmCrcw9@eqSroX8T9koC z_<_NT9TU$tPS(LiB8a9OB94~}rKpwOvJW!RQ2LNMUX3;)yx_8iR({`%KgpMX9dA8? z+c|<0O0@(5dJx>Zp6(5PpdC(oRoTSLyvgt)+d^om zy23!yGhCBuMCn0*-jHwAO9; zBJFwiD37%6yvQ>us_r8%e(>V%lWNZ2d!d|xQygo`y@*q?Gox}3E=6r(V@wu!?Z~N& z8ECmy69l8$NKgwXnwvitn@U25G2W-DsV+P|OIY}5wZCX-kEQI@a+9EO^C?PJLpTQq zounLxqco5S1L!^lbk#>~?cJD%L0LGLVZB`aTFhU&R$4B}3&zquPnk2K2>nb1T}K5S zKB#fk9n=N&im`<}kY0AEI$SfKBN=gRoLrUR>PaM7W+az5ey^xxxOXpxX-uID`vo`B z0(B{XSqW84Nds*D#7L$|TeS%C+5TL4CV!xd1;R!$;8bU@rjTRt2i=@4KjVtokNgp`dRfwtL8{JAeWdhBk zbg!krb4^Vu-H}^9GFPU{Z8mKb=T@tGh$7Lci(+$P-L-dhXtloZkLudJV&5;)aTX`7 z*sWCIRozrkNHmjz8hpvFCa&?(7VYE-Q=Cn6P@uF}O0gYVt>mIEU!SQRLw!~Q4wt!M z(}7ZXFL`Sih`i?vv~)1=AK0yiq)H~ z(zDMiLz9JkSqh%3=jzDd#A5B&2MNGew+iWXYHBtL*Y#%KEG$upxXsSMhKz_b!J;|1 z(^bfjCPSf6o6s42k2@+}t;CuJsTNAaOZ&&H|y3GaJg#y;B| zU}>l8yvFiJ4BG&)UWr-F2R#Kt?ir5$hxdLbmG8z?a`tJZnZ}2nDJ(vx{E{g?tT07# zzmqrAQqaOD#@2t(cSd;fE9U6HLAHhe_UWUdeGkI9f>PEhU@RO-x=sFzh7cIqn3RRK{ zHn~omSNY&4KS;l zLECPbQszjo7QQ}kT@ zBNj|po9`?&TGiB$mNoT7{moHE8<$C>ltgswXn%pZb)`eO_G5LU%2#ldv3Yn~v%bKq z`0b)Mx~6$ge7tpfeV{Pj4MqG;-$)06Kwc5JoV+0QsYI6)JmP0oCE7=puu*+_!kF|^DB4=a4Qj}uID;-KEI2hQqlx8>7 zlYAZvKSVa=E^W~@r1o{y5x?~9!zr$YQ~n-r8&;wGs&tXBI9DFlFd0B8tYU2`rQM9Y z`^ZR2AyW<>wY1I{slw+Sco)A!=T^Ol_7EEt0yf;>6%yCC(Ophk6XW~PreCMMCdKfu zxL_wXKN!cm3}j-hKSBK?ne^IEF{pMqh{Z`pcUDkZ^^mZ}DayTqZuP$DiY3A0l_O!Ju&FmWEaOU!*}*YOluJ z&ykPUnmL5^#@|JyNHu`PbHGRW0r=Jx(mU1|nT|($eGydWM(_k9MaPg@oVP26R<$f> zZ+hP|E*NUDlafKtGQ~@)Pe?>)*n|*S=x!p1I1r!``41UC_M-;Xw6T%{WRGE6Ui-Q2 zMU`GLZ=w-&X}xX3c3-n=IELvzhK#zqLXOlj-oubi4hdk@LlhCPi_+UE`n2ngJmh(3 zO^E+2CpNA|G-ycq1UU7ivCn{RWO9-&xb}I41=99Y3tdGSui2Vq`4f-rHbzmAe9+bL zdZm*hF+1s$Ag(&J!1qn7F^8|mT)Q2zUX61y{o8%rSkR? zGX&Z1D9cs;jY2`rd;SU}7l)WRog6iveVPRJNn>%Tm2Z+?I0#${sNOP5r$^-pRE=En zsrIuYyjdoh&javE`}TZh1-~iZ&&;&?c*SmGR{70M>D+0QB}d)l1^dKW@s2ha^H{kJ z()Ff4-YhnL3PrBZtUVchzP~&+0t!pSHNLPyy{D6V4=ytxb;dW#nSAW7oRJ1YntI?j z_kNuFzSG7vSVIF{uoV2)=$U34{9=%qujS?1aUjyi;q)YJOKGOuErd*!q}9qM-D>eS z*4+r`^i&&bx3prq)Qum%ojrrM>fO!G=B!_`MgYCoV9-cCEa7hIYWy-q9ft3DE&*~3Af?VP4YmOled z#H2StmR48U1v$mcJ!&^WwOENkg1U~x`qh1x&uLl`2@`3uVp}G+u7wIpjOtHW^5A+W zO>%bwL)8fR57>sj!2)kT&yM3^4%|}ulXSbtX?8G{2@My>sN?bzPE2*oiH)s};&yNL z;(!-R7kt9!Kd^dbVPO%G&?d)e1%{dSGm({R!)#N(hCqol5+&%W&bbwRR%7>RW488Q z^wZLPkkg_5t_MBVM9PSOWwww|73{#`!QG>YE$V;zs-B?$z?+COpdfey^b(=5q$8_NR$C z1)MmcQW^d4ds#KIFR!m&>rDmu27S9cMlpbBT%@Phh3%Ec8iK zF2Y#ATBRO?A9xFOa2nSoaBe?!f8}Qyfp0AY8@M`ni49C1a0%_A9ArDJ zOztiB=qnj4D48=cHV<=Uy89B>bKK>MMj%ipId`b(W{KDUxeEy`7&oOJTS7O6D&xO;yrHR8X0Vb{v*R71TQS0#aTqkT+1A^`~ zMXa7%Nolw;lcWuk{PN%25TLd(8ln7So8mS|#AmT+J=XyU7E)Y@y2C#eA$`nj7RDm` z&I>vG&mDN0HavzGTnY_`7Q!J(%g=;Rm+rf{_1FS=R;mWKV>8i2b!Jrz>0>9;IY(3*0Xjw3K!saG}#-jXHDS#OtjDT&@6U;bA>L4S*SqOo#^3KsT72)PD;*W zcWFoxF5U32Yc;1ywm2<8s*0EbhWC7WGY5B!xFo(>E zlQWP1jTL_M{Or68=^;X~Nzt$p)k*;pkrnpYW07o$d;y_De~+2u?o{HQkJk7qNW9mX z@q-)fw{ThA3I1$|i+f>jQwPzdvPyhwH;+_(IKE~-glWE2X1npJNr{1KNWO8r7*PjEBt0qhHDyWTjTcR9u5&NZk>jG_ zh5pbqZF4O_u#rf2ynDI`kIxTKg-+L2Iw&gbI8~{+72b0g(KC`##)!MP_1qbe!EYOr zIj}p*q$OD%a4tk@J@uO#O3LYiIDsD~%OwpYmDFH$0mBbBv2OLTO3M)JGkW`6VN@TH zhZ18~9RJ*yn%%m~-LZ5$Lf>0Q$?1PqT4Q!7=sucl_2h_}S-OnPaS zc!8A2jRPurR#G<*xgPEZwtjW3UV{%=J!z`wCv6`)8gZ!v+|?I*v}4rus7d%W0=jaN z9ML>j109G6+G6ert!adJRHRO;QcGS5svA&(=rhW(QD_JIt!Ia-P_{V8>O@-4Xno0g z5M%d5_0a5%)uBuZ*6b)oo&^UQ2et$C`7S^IcTgaqOG^Z=Fv5ZB; zLi^iX`_MZ&B*K2W6Fs;!v}N9b+53^A3gPsOO87d_nY;D9UVrUM->CM_K5YQv1YJHTm6TN_>9mBHFX+HLP>zRywsC zV;YlRC#k>4Q)6_VMVB7fE&WL5O=laA-8Y6wlm$H{AeHE)AV<(%3AK3 zs#2g;gpu#Zw)WC7k1vd;P6l6D65N95e5?YIN8AGvV4gjR;mnzHZD0ph&D%6L+`pX zA*8XF$|oA}O(Qq+;qg&mX+&tpcC6N^ zF4-KqZBsoJT={)YANwBccu8A{kTy{wF%zM&8pO`kUI2%=jr{2t@&VP?23yn~Voyco zNE4fBTO)c5Lw>{M-0gZw*R63`y{UK1uC3_HY3NT<(QKnKD2jv*mhRnWSur=_594fj zjA2w+t2t-zR0(M`jbB%AG0uf2^Ah2q`AUf-Bf1gx?lpF%je~xE4O@QjV|)m3=O?VU z?#B<%jhMoZy~Bt3{?ej)axFxgNQtd+JH;Q3ynR!(LVN{@aS`BS;lZWt*t+Prie);s z+Zk$2Zy=dm0Cys?MdMIg9Q5s{drGm^aks+2ytg-x#n6|pnF1nB$ujtjEi%UxT96|5 zGIY};*)eZ9wCM4q4Pjtt4OGL6ctm$v;4+`s$LG3JGD6D$!>@L?d>9JJC=-m?OfyU{ zWHWJ*&j~MU<=KW#!p&G8??fvSM>U)`ve;HF>9lr4kcceku zHr0kmsc@KIPKcK@*zTfW4K$nW-p$)ur}y8j6+$#U{EFW*KKY`-rX4yq?OffaJC2_- z2tH}kNzvp=zDxp1w0Xxl`6&Fpb#ym1x?NAd-UB~;B~UxfNW;C{owM2SA|xe^o{)ua zDAZmVD)p%0%6CD*_QNg|&SK&&%%+4?^hXmO$DtcIqfVGqOc5?E6sBp@#r2H12@=}? zd{_;UDy2AyE_u|2vd8suR(uiP^i6QbxRK!Jo@jhkmvOZXDLptirR}cckFu|dhN~^B zNz5b#^4Y(7R7yLRZsG2qhZ&`nXEp)?(U!<+J-nXZ@iT5&qbn+YNVD5O(!BH&Ov8Xu zmh1U~valDy>pf?<8CHq)Q(X2{X2j9)b{8zi&+OnQBdSiKea-aZPNz0oEUKhp5Eu@v zT{KkvX@TsjjCpvr!G=OQk81tL{!WI?rBf5}cNL=_T(6qSDTQ8#m8tW;d-9FMX~&=n z9vB<}RXPK0hqZa3B7<)IZ)3wxx(vQ(Deir}4w^%z0h5nT?uScOEQsjZI=H8|2!a?^ zq_No)#852JruUM)OpzRXP(641%Cezn|5Hdb_NMdOmjBbCY=HVR-_HATFnR6?Rjky? zpx|CD>NMwfQ4|{Z&2JU*J9NtZ3o5C;HIN?nJHKYYrFiZOF`u zEaV1ag5CCXKR-6;vbtQ9$jd2L0nQf9qDvA{u037-`k5aQ-9*2d1adlD4b&V2c>yN8XZr z)%c3x)UPSmG7>)`rS7n?5zUhZvel8lq;GGW!6kH3H=dB6l(L6)@WsNENcRtJ6Hr2M z;N?mC_-%Qa0}at}S`dh7H5-|~=KL5Os<2wNGd8&YM#~MoLX?Wk*IyFuU7vPkNHx*O zz3b`2aElK;DO2=rT|+SPed3Ec9_-A{5uL^SNO_bxc@OU!cP=Ky!opibPrN(aT6%La zqWM6x@+!x!K@kHTBzQ;s2GmDik*&rx8o$gLxl@58m}jr6&;__q13lT`4*7+Y^GFVl z+5)4r)6~T22c!iVE1;#XWfm5k<RhG*b|-wnq^twtc^puR^(^HI0%UD?t6X!lUdD>39SyhAH1X#a!-7>$d4o0*r$)QiV?oQ=V?Re5gc`+2@|GWp}Z zI2dWD?Az>5$!eF@9gHNO?cM5m2Pa%*J8$gBWnXU#+2N6=xF6XnRe>w#&AmR+TLU>S zDQIM=DZt$#j!Hp8N{%nl#WB@vXtC_>uy+Pec;fPqQ6FQ(rn(eFs?6R^aiqtzLtl%& zFYoZahzSY6GSj7D`NXwc2RJcj*Tz#gK;NP;juRM7vLKIGXI4aOtv5LdKjU(m(X(&U@If z{|ReV_z^Y^YN2>V6xg&GXjX8Np|%omEl%(5aq5+T|NbI`sysOO^O%Q#PU`V64{48# z*8$%NEVrX`DCUQ`G_$FvXDi=GGba_+pv1c3Clb2bi0Xw+ zueRa1j}MSI5?h5NNfImuC9wxn$ocJs?k11dG5jXl0Yf~9-_UO$b;~NlAlBhz3a@1A zK6&W8iTk3AokNr;z?Nmxwr!g)ZQHhO+qP}nwr$(amp1EFuT|YO_=6tCFk%pcd-mPO zR*b+#TZR(s-mg6EVewYKyK-DgztUGoaIc1*APplv&9L}^Xpt44ZLpMW(M83MAx#-I zE91#8dinLu(SC|q`AVi)!T82dK3zvR@)~^sEMu2}Lb|R|?_GngjG$kg^W}7_gj9qP zwV%NCnWFhngayLIUYDmrUw)RLK>X+gsFPN2+{$Va16J{)Y=MGT3T+9mXP?^w#jct@A z^=Topn$6MY$XoWZ9iCHLF!ER0w1GxF4;xNW4V8l6Tyii6o^9V8k|hvIDARz%@AZjs z0G`pn3$e_hXuWPCc{0aVN?~L|ZJ0fp&?>^=ieu)A3cfo2wIN|tWhh5VI21i6q&n$z zfsV6U0>Jp?KRi>#+~2-fICxhF-2kp6=f@TRj)av;c1>-Ar?D4MH0e6_)=P&=sTWvY z6z$+y+*Z#^v6|=R*wXPo~?!Nktp~u{? zn)-LV_Y@j4jX_+L z8wglfA=G7TSd~2nh!R_{K2Hvretk3%urpu~HVM8Ul0W^FdMlG{QC-t9CP4)}Qo<>> z@oLBSke|Kgvsy9onzjiLakV9o+yDkOAwoETIn3pcD~I*znFFY#IpUA-MZ(JHoa+jD zFx0N;8Hz+2l+PjjtCOo?YE7gz)Kh@lnV#2{!4q8T=v8G)k4TOOeK7q~#?P6xUMPgd zqpB`T9x5ikzS^?YYLBta%K66K<;b0v)S}&DMS#^K-V8ZUkkn{dvheo?g*^ij#62msGW^X`D( zUV9KQg(98UX2kxaEZo|(iNyXQaq|_Sz|QOpbSglI2m!dNM0u12FJOX_>OTDwjDy+a znH8C4QO$+oLs_Yd?HVxXWl);WL>!MIlT@5(WQzY%)(Yt;@I^GtoB2GZQ2h}jo! zuq(yF$caBS7VK{04ud+-=pZ9w4%SCWHwuo;b2LniO%chVmU*Jr4x??VHIWFaKgu7^ zTSwyx=WC|pWF_<#m{U7KHF=UDzG|)D1s;U~B46ZKzB6Ry6x%yzV?%O|Bb3cM@LfIe z$8l<~#`6R)wcg zhosUG=C-pF%`WSs+k00y6PL}Uit2z(n;#;o3Z}DOoj~y8aM((S5HoMh^5IEJ<=lv_ zf+?A&Y$7u+-fNhc_sK3lWc36aaoFGC^~=#5Jd3gB?E3g^La*}nbz#XXXMLQ{olD0^ zBOujz_&IQ{K?6&Tc*jn@i{NCcpL4F16F%ig*Z2&!1ar$NAXV(%?Z~$60vT~!3|g3s zynUE+NifGy|?*Aqhm?*abE@Hm`ZMU$^ z-O>d2-uh-kJt~hBjE4f9^2PI&nmf%(G6DrXX_c^|zS(+3g`ZgCklL%WRM6r#$>=e-ISBQRH0= zJ<0*Paj12$Sb@fU@SG#_-^kJ?jlu(?sUe0~d)(cJJgSfn`B!{SyJC1N9`cZ_M=R+z zBPbo5f`JW}5$@n#2KqMwMS*l$AJ_6c28jXguKFKQnXg-ewpMQ+$U|O>O;!6R(@20FFF&HF zOWTrYVv->OtL>@&k)l?QsyQwoj-)TP(qL?aaTZK5DRj@v?_c^Ij7>FUHK){~!x>08 zI+#+*_9$*_$VWa#a;*$6rRX~o>@%g#mGFyXvci3Ql=K46;lPQ_OQtQ3x8z~MX-v2? zH-zKA-)Tyz^Kcx0dI(74a0SC*&mV1b3Kr0fayog+&bovhJCp#pIO8Uw#@@!kPuo=D zhm>a|`DrvS$Bdey&T9-}OlJ*p>y*Elsg{i*DBz4i-yBpG2lZgi{kKnvhev58R!U-{ zn9n{VuV0!WMt?Aef|f*XMaAS(1uKjcu88%pMwR2qR-)H-v8Xhe=Rmj?(E4_O06M}* zbvL%~H75j>BhHD)J0edh2|{IZlFVNnVr637OK7FS2k1B#L9X+OCf<5&)JdrH#2a&Q z6XAFzN+GWcIKdJ3y22qmfh_e#+d@FL9_xLa!z$!c9!71NLz71jyf&!6-HLdNPg~4? z8aAC=&6ZT(^YR1I_9?X4GdC2IKRQGX5Sp^k@n$k@Sv6Hywgn6nP^~-)%VT4<#CS5I z(P&O;=^`C*%$c_o#3myz`fFz@44xfWf@!5((ukm@0T&1;xZ_T3#ZI@zNMK10JA#I_ zNHBu#h7W6x;Nb94z)H>9l{yvM#x{a4Il9}uN#Pq_OWwsnSonlUkQ7(&jr{hdq#t>fz&iy`O5iRkngb5c8B*7+tj%}N0NJUTOsaB{mlh$7HJr<5nN3&D*B zGE74lzq0zZ*ISZsd0cxtw(c^#-B-oO|IW;6J_M8boGh@yne?B_qkF$sRV^GMy4(~l zDB@HtOghZLLc(PE-9mLOqwHh8TdTW5z!X6RRU--T`E~Rc z7%G@mWA~V$=mzpbk>FOVn~cu&#$&SlwCAvjZ%`uIw40M5J*O@^1%_*O>iQgUnO-yx zQ)+Wr)5bBePz7m5&tiW@kKQLDuvhklt7v%Q8=n7;Ak?bHFCjXk zw+zfPHcZ$|ow1L$-(-P@>wHe&eu||;<>=Dp#foJhNm89zny6&(6w#{)r!+;&(XM51 zYjZbp(^WOFXbMlLP-g96yac2}ucaZL93OX&F#(kWF_5o~-nR%h=Wr{n+QCJ#5_aslJ%VRw5ZK=#)7$8uGmX?ZoQF}TmOc}$I#Fl4w+ z6S~R!4VvY#O++`f*kCuxY*F7~q-Nnf^%_-fY3DxtP!)q*SGXiJ2Hp@`Ix_8G9GRz~ zMBOFd!23#f=Ch0a1Qn`9o%s_tkRik+8RnQ>by%e8$HDe_|;W0QW_ z_LZk-AduiCY@6$s9QWhfocGl^o?_X z4=GBCL~RnZ^E1pt+l!0P2qZuj>~gZr97)prk=-xh8|3N~3772h3?2&w3DnUa?s~OWlW3bOUAT&=!fWC3khI8p_+*tDA+kNz*M&>gCLO|gy zLbic1DG-92OM5Rh0m>96+P((z+pIlw_X`dH3^U@1+}d%6c7eB8GG}=W`XQ(*Y~tpMCHPWl~yuPR` zDSfC>sF)!mxumQgGq{MLP&M4^dnR`u#$IsK;I-U%{^>j5J zHAe1l9c&sRc6;o7_xk5AVf!zD(JQuaNh5oFS`8#D6c)8v{VvRr5252@<`eqHA5a!J zto&R7h9MNoxyzY|!-4#~#hSRO6^~-6&X?vqznvTiwoFe zF+DP&Mk8K&>}u0aaol`BKZpbDj3wNWRy$T;XI>ZuL)u|T#;&Hol);>X#lCe%rugvG zG9Qy#Sq|c_skTKPUli;oIYDn^F80*l$J@IUaWT=7@f2RkEIUgKxTGF$l`yV4R6-F6 z0#HNN)(sH7J+2?Mm)T2ct$yZKe_R)dfZb>W6zlc6iafj~CJT>bF`}~sNVQg9t1ayU zw>berJz6U^TV}%;qIL=_YE~uzZ%X_p;pCtHAe?0U!@>4HhDjy@jz7%*_w>KZ2N@aI zn3?}SeUmY5pz_%lNOaMlZV*V;_4^F!_HJ%6_Fx$LruJ@Y0FZb%fkO5WNZNs2plST` z`*?I4J560@Uw$h+tSZcVOE3Sl+RkfiPs^8@)<~%gPVqD_mxH4da3~<+73Pi)fd2fp zt_1PWIk~#n*M{&PQqi(?(2kB^L-~QfO!1CD86Dgc#WUKu*UI|x{)JUM{z=&V?~?IJ zatIJ0_J9FEzl^~gkU+&V+x@8ksILA*Fi?)7WodyN?wkQNHrlzd?{mcd3xSaRlmz`_ zA7NblKtPT_TIm9SRHk%)2*#Hp&`e(dXwyM4jt}4U2r)uzw6$?SIM}y$b~a`RX0I8@ ztrW;q{khgh!15s-!8kesX#QI~S)M7t@25H%D+@p03dr?KxfZC=(VoQ-03ZkCRSybr z`E`5%)dIo+mq&ee%tPfbkL5lo0|7bEpsY z3?ftmsH*RUz%Qeku`@Chf~pVJ@?~RgdJE(Az~;mZyp@%~hxwJzfyu8b2(162ysvk{ zp~W76tDU`(Gw{+Ub^L~VdXZ|9pQ^|J0RrOUAo`n=$FdG-eDC}>7stAocf4AF4nUUxVVt}l8h`_c9}n#hPT!D~7pMP^ zyyFiVpUVE#UjQcnbski}2T-kmT;KVgoSAKY0CaM-_`26V)ZdtS4i3OneQ0!lj&Up) zkzewUBn;zca$n0{~aa!-xH{FmI{^xp+NI3H3COjeClKI$JW za!SZY0QV*bC*bvV_W%GuJY4|%IDr1wzM}H1fZs`=x40^xWhgL&H|C8ptq-|DTfgVL z`??K%pf5KXm33qb0f13I_8+iu=w|ni!!JM64}7{`KagMCli$(%Um5Y$mE9kB)+79n z-?&W}pw`!~sr`tiS~|HewE!L&z$ZVMXMVrgYN&CfYp8d%hhpaIOvY?@2U%lPkrn0i8x##`lL=N{4K;0a9q6f9KIGcX|3PC7Le+d8q!bbMO&n6x$ z5cz)=eruD49`x=gna=g%q5BP<>Bph>0e+!*OU!@i-<;8o@yFj$FhA(uP%vNQ$mCk( zyx#D?p!yH)gay|791pxFS19k8_F-4;vMk&MZe&$`t4`%IIHE+LfXOWo8TDRrT#UVe(pk7NY zzbNJ>G+9LXy?WR`ExOagHWN3G*IB2K&$##V+?UqVmOU_w<58p}tIz+$8+P=2`lEg1>BW51 zqFp<4vMnGfW+^Th;2yxoeZQ#X$fl9tKHJsKfYlb^6Ha|Az_|3!Y~!Pwkh$Q_mA|#+ zs$79A^0DtodszZbQEhYnsp3!1bSX!_p<2PYe17X$0Op97Lo|Z`>!^OH#_b9XqCGU_ znsLtFgTAh;e+aB!8bp<$B} z)5W+GH$wFQan-Elgq$1dXcK`(&bYj(@^M9}*1PH&3zYHjHVm#V42iyr3Xg-di?~*} zW{_DB#uhxGL(leX`a57JEHaqfp@f0cCCMd~)M z+*>~1xDUQ3EQVg}d>Nx++?B0^YRue2 zsy3ezfShhxnbRxpH;Og*Uwm`=t-_Bh)!MRgUU-T)<8_dS(BUsl=J!zXteK5|qC4X<$ z`vA=ygLF^DSXjxgo!h`C<-Qble|PT_=$5?QzoNek8ymg_ly$B~11R|QootCRN?wLW+eg-qE1scBR zqX&Ig*ItG$8?p3BvSo@=D$nY=E9LqKWk=+>ypiKcqmT=t7o)*%VxI_Bz_})qNN`%% z=%$lAQAvSmd2Gdjh)&s|{t5I|1)DjO!!AW&N1cpubSOzzfK_<%=}o=dM-n&O4$up_ zm&G>{?OVMqTrp`JparyYtoAG%QLU@ZsIP_DD9m7|HD@EE9_E}{L~Se3TM$U5nb zP5C04s^99THGe1YPcZ&(}tR%9VbRdHzQ80(uvQ-LZi8IRmc+Q(wTqG{-$1y?>$Y{+RZ$ zZNKz4+$_mgymioQs`8W(b=SQix$-a8Li3wQg_aQUMB0J0p1>Kgc6JcH1EXqx*GBPd zVp#8lsrppO&>}|0q;9^}A#SsCh21GXJ@@v%G`#!)N=t)6SB0M@Szhg5OG&hEKe4rRZ`-d}HE)S>1@@Jg{e- z_0yEhKWdxorf>yWTvbmbCRh^5JdtyLvzdnY$l$icjv}KvRIj-zTaWJTTwJWCIAund zISz;EmG0|0Tlv-Bd!o{@aNDaFZ-#HQfUjKUBfCz9SvYYh1!3u1GmrXiQqE4EtrKe= z$J*P37`RcNiL&5M`Y<>cTqClBkBh9aGpio>ASu6 zHcjtm{tnGBx}TU981o^ka5RzTS_CF+q2~caM0dB5%Giotxc80UisL;poRM<0zYSW+ zEy|_f+PdRm(5kqcxR^cRi*_-R8anj!vE|2c32)evUMm(MNU-k(>&Ygv`rvtIPU70Y zqmU&bsP_VEN-&j3LlF#jPz)ghtmXJ+Gy4Y+O_$akVvxZFinopxA z-dG_ThVrkYKRPNb^}YmCKomU)@TLIG{Dpa6fJ^g)^RyUn#BS3iC6NhYhR4%^qlJsa zQht|Ssv&Sw%bVEgrCE&{T;%2bisTkXCkVZFP8aH9ihe0la(9XyQEKbDC6)*^6Fak` z#tEVE@XYxY9L`u7(;*yb#I1`u8~hRLn!`3T*{o!nqccv6lNX{tuo8xq4tk&?9w^S| z4b}G_rg_P}I9{1L!8^xW{w`X6((Ef2iCewg)!zjBKB+foT6TK9D-Wo35Ys7MJ*P!# zVV69u?7WU}M^Eygh(&mzE>MDOT#KBr414!mf1+sZG=!f{V8svOR*mVGQRVr)vh66x z;j>T4MBEoN1pv!VRkE+NGH{VZgomDfuvAL;udgv4V9Vs+UYYd5D|HL<;=a|0BcQ+3tl8pXos4zPKV&b|0Pe~Ts^>?QZ?e@Qf#t!SxrO7`aFN^uYH7& zNw5uZ*j-%9b0GlN)*4Z>WyE2Ftb0}}9P}-wY`5r^8#0?>uH7@Z)wlGae;C!|Daz+= zxF1fs$j3{=s0qMi7QB8&zjFP!f88szAd&-&@=aHGDcYmzPKHzz1HH|TqS=f!iEmR9 z_T*Xc*@wYj5!YCy3TsO*;rF;8Yt|^N?X3_8k^=g3MGAyf@unfi^#x9)*m4gY#K=nq zj`FuwWvM$65#N>@x9w9PJ5=pZQ*+uuMJF-e{ImI<;8!Tp#mx5;tlzc$mPDlM6Z)%0EFhr*NfD{dj`QGB@J%30YM7F(kAZumjSJDftVZ|zbgrA{7VAx$ssWY29&X$QAE007M*#uSNeuk<}^GRK_|@5 zmeH|dstI`1GiD<0?fjO^1A9=ld1{ukdek;^cIlsSbO`hxiAkbR_E4+qBu=Wmaaa|x zpy20^xa@by@Azc;-I8sR7k8i`L&9@|wlFOxv$mLEHo}V8^Xc4HEQ zT)_yKv{4Ga($MQ$50hq?LK*>Ecdp%FN4?z&I>vn?_FRm`zb4}(ygW8AkL^aVC@X=u zf2M{{=`WS@(day=@!tUurJ{2lTQ^nok|%^kZW z5q>YVK#?^0i_FhF-{(c7o7QvGN5D>nHuJT{!*nD_8P%oQqL9XlO3B13 zaObn8pTiJma|9jD4`}eI78qIAl3Y_~-nWIzIdg&R@s3MlBwphrZ_y8Q98%tdP-{h0 zBoVyNR&`xF5F@g%vyX`?U&DoW!qCN&kzmJ8!paex3%MC6&NF8o0~fx)KsF8rikh@k zxGOX0TtI|T>s@oXl#|G%2yoI6eep_52MsON<~1$Hzf>~4-=aRa+Z*bmJ%s6nyYWo< znNw+wA^j#gs+(E`8bew(x9MMn8I@;5r|iPkrk=K*xC7XKI%GTUp;hi4kx0HsH@z*- z+ply<(OQBc)K$wuvU~xkH^cL6NZd)oE{*!^4Tfm|sToJ&7b0TiHap6VMHp@hQl;xZ zE#t@A%$!@}s5!-8_~-NhwHD_(o&Fjn*vL4=P=wb%^E6}6hiTPb-*P4F-Xk9@H&v(aR4MRYIdqIM_H%`r%14YrZZ;SI2&3T zh#&yZsW3;9$CnIW^3_#*7=STN$_uY)O+rDxR*+QR>b^A_-dS|PX=YrZifL$Cqw&0e ziK62AMsfBl+8N5>WcPkF(~zeu2^r_*jhEeLtl*UwV*AG7RqzT+M2AhWtNq4%N}YQP z_e=xEbu1uO7kK~qns$`^U?z~N{)+i%DsjWxLx6CR56W9{8oJ!A>^CUwH@RU3lJ#~b zHNQ-le?i@!Q@5BiD@Es*ssJ3-awiprIb`Z6Jk7CpN zXqw&MTwWVO@)G93zKVva8wj6tueNH3B~521b{#;X(lP({W<0G29Up%6d5F+9?DnY>IK_& zdW~zgj8k5?-Oc1a$;O4aNra77w*Pgf^EBVmjRjB_Z4QgOn5B|Nq^atnHA7SHb|0LE zLZT27huqdA7k~!aJ)rau!#Cl+}WnR`%jqb))QR?|RyVr-UvS zRIkpDb8+KX_f_tf4t2qmgQN;gwyz2zsJiK&;osQ!I0Z z!@6A)iJQ1(EY-)llf1}8H5$oG@Db}4fg_=)n6;b7fP3oK;0dFAmA-|i#@@T(Du{1P8;FQ&uejbH_=8!7~Do0`p2S>XQ;&FnP7|rhq=v7lBT_Qzk zAlPWn^%|CE0My7|qG0kZhX?@a0}Ndgl~6EtwIE=%t`S5S|Qro>T6t628nWc-M1X?jn)L~EV+`m}y2Ww(jzeF#;=dfsLZuyH+ho2F>l>!35BAlp992%K4 zUl!tlm797_yJ4U&iQZg`xAXcJIX+O}e?SVSl_kTPKHaET!r;R-BSB(G%HoB4Dwkm# zHY47H^+F4DFN2Xv%Eb*fwtPfkT>t{;7Rc&Ti_)3@9OBIsxy}3_X6eE~+a$rjJ)j;x zCl%bLrl0DOnw zV|%g#{km1j6o^$2Gr)b@Kn_s4MbjKM6IuX$w*OX8rTKM`Q}f`5G^AQ<)=0Jhj^Vvh ztRzI%J6m9cl4TlK2fOyJ7NwLVhtfv_2~z5HL(x%g%jXoGD_$V3M6LR_AGD&$TU&K6mDr(66q)Q{p;0XI^smD_Y=ygyihtF)1wa=5xP1S+P}*JP3TLB48)dZ| zj+sA|ki?jeKe?^-n>H)}GyX5Npos#7(&-Mi+S^wC{oUj)I; zq_XFwNs-7%7w4Yfu9oBIAFm%xU^+b%!mld6+z+Gc2;0fmW=`h=L&wSwZ7<}=?`N0? zB0YyF)b_wvY6-)T5ES0ac^w8nOX0QN-HOqb9xgPE!v1477AY>^!xY;6%(D)e;*QL^ zAi>jfk@3}x`Qc#>m(M!msN)!z3JlTF>PHF(a)p3V;JKv=0@ZYBgN9p#S-}D2RRH}O zfnU%aUID?Vaka7r!x9yau|_QKU8OC?E@c$h227k`7e2SWNe+}7kb-MmCjI^$OZczRC2-7OKd=D5+*RWb@?92iJVZzd$C1q`PY{a=*3jT&#Nc z8EVI4DPq_bvSuKop_JfHgvs<__F*MMF|3{$vSea1P1EGMg?cqd%RHTQ$0g8&H>7R0 zq;ta_^7Q60xAdb`B+*&4SROZL?3LuDZq?SL%u?@E!kQs5bi|49(_&ah6;$pg$Al%T zHR@IKHGa!TqxmU+@2ljVS#Iph^me7hq{jXdxN-kwZcWf{b%2N7!W`cA6|zL`#7kT% z$fe;IfII3tU@8AVL$gB9p6UpS{^e*+uC54IT)L5|z~BOD!hVg6$Hp!rPlqf!hv=D& zbm5d&r&0X3f3#CIZ!?%Qj;YzAS-eUdpv}WBJk_eJxCd0K+WDV&U3oSIo7z7}Zb3+zM5O>yM>Ux`1xUOtqp< zaz}kGiO4|VxiH=v33+^*!?fMlthucj9kxd)Q}uZ-w;0tPg{NXsT;OvBIHDRz&r3oD z9=?Ku>|^H;&Q=6-aS4+AYe%GCAUi*r-A$8KPw&dcr)T?Klw#i~VLUepLE!2*wA&7kVL^F{WA?ks523is@Hn z2#(M4mq`zqy`$0&Ilfd~SY7rh$n%|mdx7SrdMyhh!8dGt+xv#AwKR?~?5G)vUZEKU zuNpLe2+;X%VL-4(H5rJSsC9iDB2)}n_nvp`z4=Lj{yJNCdwmCl4+O?6o_GAR-(Cla z`5b)~*2B29H)7C0{;c;bJ)3Ex1{E4LJVeg4dGRDC1)+;wkmt_sX;$=+n~qy4*lfVw7&8iov_sfgW%D=`c{r4wQD8X&o=C4cr$W{_ z$g+go-J~nr7A}eeu^4kYVKuR=*4&v$5d;TIx`A6)kQ$#^WNmg62spx@D<2j6No$TSJr4_?v9CJyO_FW5{^1??@e5(o` zH?C)mhl;#Xlcr>%1-ZHb1dTEY5StRr@y6!m=VJEleD zLJZ1&2td!MY(~l47^dheX=Z2|aO=K>`h*S!Jji798X0~!cO;M#a%(I8A}TgQ_Z5ng z79jN27NLHI^s&l&8!)njP@!s1l4MXgc)6d6#LJ?*=#cJ98By-$HqDC`wNQ+%k2K; zM{rx>&Mm$K7xW&4OTrtp{`KL(8`^#EN%Rh?2-n3@25;`A#Yw2IfH+&=uH7trPXs(ck)78o6)nqo=1Lo ziwL=F!`M7O9!_Mw>07`biCR>?U<`TtF`N3mvea)IjTuc0>Vfixz>tQJnLY`HAf8`RU2bNyL^`-TTkl)aA`>_VLSjDkOy8pU3$A?cVoUW*s( zBpxAy^(+U^R&P|&u&KvWclsLNEcpAr?^v?x1|)^+z#82@7uckK)PO|^zuf9*h_~U9 zV^qpvaVV22w42>uNLu%6BT{hR8ReCl3Gx{EEDt`d5pv z%OQ5s~Rj3V|zvq|h|%%?j~P-d8# z?mcn!E#kq8{t$d#CM1IN+|n`B_#z|3xI6>fLT+W(FyQN(=cmf4X-^Yym9bLL33X{1 zMe8pTVq&gi*P!i$?(fBr-Ql2g4dU|Rm&sB|ZkN}X;>Zv$Mz9pUTx2|_M@`WKt)r{!rgfwQ%NDjw3G4s&DfglR3&9w%_gI)tk zZ+8*y%(X3>!D$rWDvS@@V1%JFU~Bl~7z0kIVe8wLQBH8MS;u%_nDV{X#~yFWu63OD z?fm@%B(Q%r^a46dLx`{Q8kGY@`9p7#GP6ui1HR`gurguD%dpZ9y(=aP*CVP-#AfqK zpUL@=94Vicxq?;V0mdu-m^6d1Jzrfs>4W-HF)sEe<$L@|cvOfy?i8HFQa^4nJjvoG z3_wF43m1LRVoAo?`t23;qG&o!(1GC^w-nfmR>I{v<7yvj3df!4i&42(B(TX^UFqy`*2R5Se?a)qJ$R=xTadzhy@?(@>oa#T zy1L7C`zuI=m#L_J>BKoj6oy`+7|b_hOjzK>zc#Kq-u>}_Ik^HFcD?ed7Mr1x@;HB_ z0art9t6>atYLu{T0S=*F-*z~)7XQ`bf@G!b;wQ>)QM2T2!bMZx0`(oKq9&`-M~}xo zK8US}nPa0qN8xbDM|#A_ImCE_Ei{LpgiR66D-UW+nOd8VQ8~F{N%%ZF4&kGeFJ;cMq85hs~|#a7&-Y#Q=>Hu2m@}S(vt_P zzht6~OY)F9UQcXI`;LZ)|aSE$W&FL*wT*jVPzh0%y%QURo8B}3J81Kp)}cb0dWOKH2J~Tsj5RFhf6%(5t08`=f+SD>@7w zglaz?Oce`Lo)IALcr|mHodHx6{n+?5m34U5N7}#oAevbgwT4V4abh}yIpIFSj+noWp;MPd=QKhqIRfOa1!P z;u$}s{5icRr0_7e1aRA#SWg{g**ZJ0J`vr748voZ z{pv1{yLU$N!MHfuav@4(9ZqDZT7Yame&C!?S9vv&c+#f|DKZF_KH}K8?E^;XV#M)e%@&$cFmjwn#{Qy zNO)BYw#<>pt*6SL8orfmoJ>c*N#Bur4_R#4&s>G4y;qQSe@Sfhv!B^Taa(jzQo(8@ zh6+F#(y|8d+!SXCP==&dLW4sR3quc?XBd@=oPjBwRN+00&T+GJk1ve z#@*ZqLFVmo0Sc2#|yrYV=6e4ddy5cyBavE1`5(I3t7gmECMzz>2xSq<1%vc~LL zYiF}|ImcC_YqlkP3^pQ7YUy#$Krd=-;A|pnVq|A*0>#S<<>c&WVqgR1z8T{Ts=U1O zK}Sl7o5)14BUI+^%tmwp01N}e%mA}Pv_ncrDjswk#zI89vl!#=PeJ>P|IBsv)#r4p z`C6-S#pcClcm29^`)I97rUR_&q1Dc+jtM990Q3$*0tlt~aR>kiNXQEaDB!KCDg=Q9 z^%Hd1gcXEni@+}Z;J1N_ppSqQLKzrO!A8ptlYavT3J?OouON}1UhEIRFVD|^s26J= zOu~mtpAP`s4hs_f^dklZX8}&KD3X}Bx3rW5e|$X(F1F5hdJO0;7!L>FQ;&^t z0N4upWr>~->j?I18x5a_s(%jP@Cl6F|2ICD00hGyC?5fvVHy^7AFc_M7@)if5X>?Y z02iI%It8wGT%?o{gOf?-($UUtpCYY%e4TVDiBe0HLz4ugj@PfL}Y_*4q~0ybkW&uS7tn z5ouZ#@a7sOlyK+HZ>0=Ez~4r`?PliJt6G+!oqYOrZW9iuspWGeu%-*T4-w|@5>8d= zYYQ0k;1|3JoEVS@2nfg!i63Ac72u`OVgEN}dwvV_YwHbLbmjK;3djv`y+9M-Q|P+C zoX=a2Egl@hU${N!+vm6KAdkzN06`zRK>(B^05D0*A zdGI$+!&xA{uL|K!t{Q6^4le#piRLo(N2PvjZ|Uq!3nSa#?{qP=2m>0-{wr+%6o?4u zPdWVEZ`Ns_(2w2qPs+)!z};`H*aj}tx6R3$$q!#Off>Z|ryW;h-HkZPYs)?bX#daF zGv!;Hc1!<1nbX;CYbBuo1?>3Hb-{11aCjEeFtl|Qw7@3&SHUFy_IX-=J_3e1`VH9o zb3TB301(ixiBQGlDCDzwd%@e=7KmuXX`UF>fc1WB-!dr$5flI+0)d_Pa<~E)qCY(X zsgPFvnjIMM+q=F!2^EOHPgRrK=vd#+&A&Y_JUytDz-WIqtBKKJo0W1jk=S`+f6D3Kkq8g-$x9U z+w-lu2Nxa~_owbq$1PvBTHnzD*>st6+{SD&n=Zg zERZ!5uyCqhAS<__CO5DtWryCT_;&UK{s~q*5+q= zgf2ds4RtY64)gMNM#|+vBh^fmr0l4S#QKn?IJ6bD?q%O_R*kOva}t3Y=X}E!Vbz_n zmm`-88{0y~tqOdXrNG9eOKj@z1Y_N2Zv18L<)Yxuyn>T!KdQF=p8!A7CU*q+)&6vW zcTz1ET)-5?ktCeMJIUpfHQaW|j-~1R&jmIdEBk`c8`B=?HAMz&gmlrczhzqi7m1Q3 zHbp`kZ?%fP?f=EtIRl9TEM0hQ+rDetws-H^wr$(CZQHhO+cw@!UezDG;14F4sZJ*A z>eJ^tZ%<7p63zCvqUcPe`9cQsWyL@g_2?s17I5FNw@NwiHd$L0RP&11GiLbhcw;{9 zRT;}8IJ`p@xSK?OW^BB@GLiyKM2=5W?ho!kQ!zc*WH%95&-f+1ue<v>nqRmTI624_c8sGhw$&bs2{T4pLt6@coINx z1qkooK_FNOLRJN3j)SL7GPBoCl;r2<;QQD0H-_GUe1!n!uTymf7HHcIxmoEXp};DV zIT`K2zfzMNdtUN95I<0-^LotXUI3^uvrCX=EU-VJx1Te3Vui!8j{8@6`6$#Sbdt2= zdO$w6h0luaMtH)M5M1Z+o0t}nw^SrO- zIv&o)Y)I!@kG;?bm<^j=fA|c%up^46#VO832rf8hC-C6qBJ~o|h2zs(7QySUa%q4? z5_-b(Yr&Wm4zJQq^7Q9YXG&x6P!M~MS)i1!a9mS{5wOa4>Ac5dI(;RO=XK_qbOddr z87Hdw#L&we4igTm-~4o(mj)+S|AbA6CPGpD;+(B_Jl1=g2Sl}GG9uL~xT{KS<_rSx zSl}Yiat((!?^07OQMt9~>c2HOHI0uA=z@!ybLrT~$~@eV&E-Pd zfMH*5Y?ER&xJ*@;3KkVe>cL=1Y?AK9BpgrPvDLOsJ8AS9QAzAU5X}Y9ElITV{7SiV zsH-gIsdtC&mE5Y|4i&X_&{rjlM%`g_*TzdN2SzUa%(wkLOHMMltQ6?TVnFH_hI77& z>o@Fezn9fYt~xr%J`b#_MX8iUdTeGQL>bk)_0sq!jQ;pQ+TiI}63&u_mqMZ&VEn8% zdO`ybEdK#k(iAn zQLG6|ZRLw+Pu#%&g=n;tGCj`LcH31b{cUsRk%UJ!48ovrWn5dO*+!}HE9))M;d!E1 z+mTkmftNWY-~;|ty_Qfsq>L^Yrvwdd$=&$7ol3}aleMc~$BV27hg<2wvLw(tH^d5W zF^ zO1zqKa`wKh z+Jk+=t{pU=&kCxgsrV;QZ{7W}I+>Tb?4R3k_Z^~qYxBllXTi&3wD!T}63Mq2TURr; zt9+0cd6d3OqpsSy2+d@J#l~887i2Yus6d*hMB2S?LS-`|=)XII2yyZ_)@zeMq1{ZN4FxBqdgZyC> z+pYxb$!A~d)Hq2G3x%fc%sC7C78xA8U3)N<%fWW^9YmYOq$~Xji7zJ^x}*eCT}OAB z{=qPv7Cr)-JI~bh0IAnWnn1h6qFssMI#d+}v(tT;GEB3yRNmRmhJ^+KLfGCA|e9S;On zo^R@B&pS3AGC018`dF1ye=yLm@4(Z8Jfp)&9~GCkF7U>9J)FP8KK00ffnhmTU-y}e zGuLHGo~+Mo)S*$sC`AH)1Gy``p<}?%lB{z5J|}7ng|xmswV;7X)3xmH;~34};DNcW zQ689eRsVW)QMyRC_b6*tL1QH4&%cJK7i!*<;D6VUA=`>#wClZ5=e+p~pJa#k)&Z;C z@-{6>O49}}CGC`THf8t9@06rlM&VQ|^(<{*paLuBUq;FV!bsD)eCIN92n!@!1B)9wQT^)S8tPmKW#n-rWJ(_k zr{7-M!zgN%wA0knrE)zMh%4^XWBx2cQZxuxukjwzR?&u1+Fk|mZvH941Q%bAm>T=SbLd!>x?9Sq z-w60im*TWC4;O;WcXogv-4I+g6^F$jrxwM5oVs;*A6-_R+}^K5rUHe??D-*Ezph#@ zW>p_7Q1BQPqs?Q+aq4Atd4YImgmRWf$^m!C+g+M4%e2Wfkto?Nf(PN;iE5$kkzeIv zTGoM8D~bQX4S7wyNL@AJ z#4SEXJI!U!%NXRkI0v6_Q(D6DH+1}U`eTVzP}VB3+VR5O{WUmS3?g=#*Oy-E#`O-n zUT=&DeO(K4AXVP>EY{j%&6kshDV$DTg1I~1^4Bo2WDzQN2f1ZszjL#K?Dq!`rm?L_ zVwhK5=2%IH(=>|RW&zjPJ-V{Cqbbu! z^}gBR^A;cBra{%W!D4Tk*=W!~h5pq}-LQ=u*}QC`E+b!on56sAF?AKs>yAO^ZJS;3 z#-ntOhspQzy)l`h;wMG!!PQb=v@XDAwK)R*{#GfZryCXfxZJL0cuk&i#}XykEsN7# zo2Ud!lq#21W*ZETul#8n!!s3~%95L-OuO%}z|$nH3kG-mPhI0|yke)d=DBq8wCRkR z8zTo(`5BLjr>>*!B}2~wl3Z+^%)E5zE-i1hFOdSsd^b&cu5JYdR!BE_H47YQEj)up zVGVUF98GSHA?7%CWcT9&G{9e8P9;VBhC_6LRdni}+c%H)*fJj|($IWq?&R_jQS-~y ztchrkxhk4Lu#9FY_6AshN753)HN2l?qV z%RdAk;B(^uE0w|Th z((I3gcncUS8%~y69dq%C6WUwy%Ou~wP<(2OWHUmk7+Nm+Pnizd#TsXO22KB{@>@a1 zKuIn2|Kmyq-+(WbVp-2lR)BBwjcjM)gX=xy8n^#Gflv?{I%)vM=1B|u(8VB6K(5$0 zWXu`X2;uUDh8l~=qc%GFKjXHzg&V%(ampu&4|D2Sq7HcSsjevvBgf`EyL}(&aR+$i z#XOLh6?HA)Svcv+eM`H&%-!9gunClabgM5R>1DYs&3rd$hVu6Gu{pn<8T|d}=uT}O4tA##{MITYfAvQ*SVqmw1+O)@e(H4=oz z%$T4p;DR_g^02LX{7^@(^TN>-P10H!#Y7?+F+EEx9U0bvd3qqE8uP9W7w->PqX-yJ z`ZZ=AdP`d1sl^pan<8_=!8slzU=lS%(WZCT^t1eWQTKB{Z&uM1qLf*+>Ms@7f)l{I z_MJ%nq(sow-34CJKw4z+B zarEKX6uVTs4i+sS$l8XP5c1U}Ky-*KX%f*6If*FtJ82rj@!J4pHW~(km z%nakcuETw+m&b#m&H47#Bb-3fig*`>^@m$9$EE7|7HHgcU&`wH7BB<)@gu#W>+kZ_R+_6ALukyLN<6V$x0%qsm_lLM96b_PQz&?ep{+Su>ww3FgNBne<}4su0t*(WT<9q>YE6Ea|~XS`C;w$?#vR zS0J6SEnYLh$7OQtbSr}~0=kEul@l2>d>3O6-aKTBQ!%%?ocfobKiJWW2Rdk&E3#t6 z>e|MeF;QcOGLLe>Ec)MPau2s>#Ht{{;Ay=Sd>)-u@yg}yr9;m~tL;fX^N9q$;~}Pb z8~$XfAMx6e=$ACFgQzly`&~-6mD_OM%v)<>_U5BtHqrL^TZZgpF~(Ka-QSi>s*POQ z5UbT%&lgvFF+m&X%R|}ynZmg<<}G$J%CCz%pnSSRq%3z*JjUP(jN$3o(`gN$_dNG= zaS(_vuoyK4s*>&i+uT$d>K$H0_PR?yhA2~hY16-Y;_8ph5I|?P4T;-CR4f%6s?7s= zm4n43Qx+^7QK(g?xDqqq5GPa_ifA#M4UD6EXHd?E{U&ze}j?$lD#s zCS+a3nC}_Q;|k2LKU?FgL0&6i6$%yY6lR_G@+UgUuIpa!@mZQeh#G61|9zXWpKMj? z)ggd15V_X{bCR>~S)@yMg4WpW$eAr;XQF-;-@(`JMHYQM$&>s%fk|N9u_^L$!{BMl z6GPVy^>x}B05)^WuA5bnl7#g3MG(xfC}WS@=LhAdx$zLwHC^;};BxPc{I~|ss7#LS z*Aw5nG_Owz8b1feEH!9Oti=OpvhDL2rQUT%b=xW$_2kWV;F6~enp^s1EVH+R3gTx- z$um`~DHb=*Q_vDWA^wQev43j6p?FopD`m8*_Eo=itq=7Kl|yOh(HPx~4MD%;pY%7%mDGB$~88$#9taIvaOW6O%`+P_##^^Sp3Oc=KD|dS#@aRlZ!2Bij;5_2~5$ zx(hIpkfViy9TQp!Uh@g%9m(J4KO*dX0x(Bq4yqU7ypNPu*LwXD_p+ckaJ zJX8-#4K_~i_};#d9Xu&q{9Y~*hKByuGSNYc{gkJiIl2nMu`L@GwsaGMFB{eP?2gC4Du9-i90%KGHV> ziZ4GV+Khtq8B6##qRRZ?Yp~b=y9ibas4mdgpkmCXjoy*&k;4VayjVmPpo8EBZnCwlx`C?Ym;?owemSJ* ziJz5VI(#1Rv16{(uRTc;t9?LB*@s|2MDSB&-qhq}g%Mq=e3c)jUHDvW(+{e!Hf=dr zT9t=*knV~#+YQ={T64xPqROqLUrVJ4!%#05Vl{-vUG zy64;9&-kPY!JS=$g5Fa9pij6-mzk{r6SmX|mH4Z4Hu~_WWD3fvy|wvFXjDHhN41bZ z6VkkUD>O3Vb13!Zm9GiC&oSOA#wux%mtCRo9M-5py{{2>Kyjdb2AZc2R;U??t%*4l z4v{waV&jUFnNRN9-p?iMs$rqUSD^n`np+pvt<1Q_r3pTPXEAFHjLU=V_Pr@;j@{MD zC33~)Q#A`n)gX{`gADVmE0dQm$EzQjFBUO$$uqgOURh4Q?U^C6tZt$dm^Fi?fXc|7 zGx$|mw9)QsXH%W^h}#Xg43=@dJ+S+wiY$W|>=oKv*(pA>-!dh4HD8*s7R%^bG(bj> zn$2x)H{Fapx*fBo(3=$#J(!H$tf|LX6imEwzEKA!BCQM-j=qaGbBIUEOkY0JD^qIR zA6h0r_LkU0U652S@cKJ_He)X>Wtr}p5Q;=NJ!&WNZh4Sl+0+^wcqEhcA2$Ym*=0uy zX;Z}Mk-+7{22m0x&w60RK`$%$@4@0`tFy)s%+jGfkvOpVG_Tf5d-?Gz^gmm|e#}$F zYA|4Qnvoa|%uYdNNA`Dd@l_5WqgUxsWuCO=pBr`C6AWRDA8%E4bzP~}fK=yDqb@S) zT3ZvOHJv6!|7hK19WkRI)-aRlLcPVO_kV38r65w4w%*wU*Q4`D(NXh#YUzoTj6Thy zDczZ>-hlCk)N@!Kd!J?oWC#ARSoC^5WmEIqO@qMc^Q%VWFBm9XJR6wTw z1d=*Iy1b9j3_@^^5&z-~C;YMfAA%+a{r?a&8Ch64{wrZJ68vN6u(161`hOKo26kp9 z=Km))C(6ZGDSN$v<_2Ml5F7diu77*`kCoox%?+GCza2CQ|J~lr&CS2Zz^?7>Qx}ID zPre;J7`>F43gzY%F_Dt05pn|yTbP8VR!4m!V}pI*Fp9GA%F4zD05!F>^nWL=tZa!! z+ls%JMO-O^fXr0NldFf%YAB1JAH7Z*=M0Edxvmv>tYZT}Z5^;m%d2Y3i<%mM6%B3Y zm!Zk^05}YIn_31CN*X}jlPiu1DLJh>ur#r_Js6sO@%IZlk0Bql?%~0K>FXK}p&pb& zJpPe<{vWWEiHX}dpq5)Wk&i%_9UPQ2EbJc z9omm_4D;v+j1lxpnUx>W4Em#$os@*quOqnlk&lwalL_554;cagZA~r1?<%Wbw#KD| zfDC}0k6%UCS~TP8OG0Yw22K!D|!T<{F3YziFh zjPy(jtFQ9-mnw!`;h!t%;VWxv0TC7RwtlzKd#&HuUkaT&GpVr_umnd6qgbc z-(-V}i@qv-)AKW^1jOH1RJEWdO%pf=K)Z&zx;sC3v?B-r_cS&7-(=zKMZl-lgl9U2 zHaJ~da8GbQ%r=y`+G0$IFYxhA%rh{cHum-aukK%^FE&8~W6(5>j5dHMfs+&$@!rC~ zVS+|qdN-&nY90U&>ODdu(7LbZ$6H^pJr1Vk*2eoU>z`MFe`dbRiHWj>U(Kgp3k3yF zct18`Y-qpu=xp4+;R)EigJYn#A6Ek7ykp<6uYF2GQxo{T?@+onh2O#TYra7K$6TI# z(0|k%5zjk-Kmg|6Q|;W~#9@%H$md_ri(i*_U)WDwg&)3?Z(qj*r{?*H+Zm{1Y!TDZbh)0HP{rDZjz#hu0DDm0 zfF>`p;=Z^v=wyECM|AUiVc6XAy@>FB>R)UJp!8B-k!|V#QwhC@*na9CYzE-;lV75H zFb=(l=zi)yY(S@f-=h28QGP|Z_h2S^5!nMTY(Zx7Ua;)}(w2RR?RcYiA+qEi|J5yi z{HtsF_vFp_OQd7xvb_s&1^$I@@7E5H-EU}cZuQ;Cqld(wx{J5BHqwhog~0(RQzfuh zI(h|9_2PT{OFh*KMUR>2i%7t#`GE_#1vckv zy!Rn>dj_@BwM)C#G3C?g*PgE4@r`&$*W6vXh!3ZS{K!X;mNaY|vFBKK=Trx<{;dy~ zqW%NeUsd(vw@0h_3)rvxI}W#nw7HAmwaoAvGX?3=3`{qL?_uZt(mU>#ef=v1OFOdP ziiI-;_(i|*y?Sm}_apYPYdQ!wa}3}B^pDS<=I~9^*!1fSgtK?TzShS18;?wF|Ah;- zMI)x_i5^#ekFVSiQ0(iJhPqX^i^$%u%pPb(^XrMm{@Vc9{ylFhYlmQZC-p#wN^fY2 zo!2;w@82Kfy9lm3H${sko5u%-y@iPG`pJjW$DrjqOU8b>iw`$L`XcMwg6d^rSy)Q9=q>9&q zK)o8M;uiw*DNJ%7zqy^E+LR25hfIZnEXkR0msbAKO@u$e0W>`XD#fPn1>-Z33d}Ie zd(PC5povQU+F`&t z4j1NKfn&*273MwAHS%E&(g!68NOp^|ZCU|fwPpy&e!M+Vp``>nJysG7?k3=n$%)X* zk-S_b%;d&j>>X{Pv|eX@OGGRXP-}64*;JRlbc3c;G>Y6suZLW;cNBIqIUfdHb1njl zjV}%k)x3(2w9UIYd^i3mg6&IQ$gk{?Idko{Atrd$bdh@IY0&JHj6a@ z_`;i*_dY*2PXd1@bu(@7B{5EE_0U}38pMOdMuE;J^O)&8-GOP2=azb#<`t$APc_Y` zIF@eVg)^t>p3E8Blnnxjg4oTx1Vx-5ogq!lZ;LqXUGl??fkq0N(vnO_XtaJSwru-K z#gvDWL+Cs^CFIF^={m@b1`RX+-pU`5T8veKLZkqF!#gd2;alQ+#z+8OGk4DEnBxIJ zn~Obo)XXYNB(uyc+!*cf#c-pb^Kb(Z8MFr8`D z=jCP!nI20mQ}R6CgfR@?9t?}nd@40jAq6nRuLkSGk;EI1SYI`LM5B+r z*LtdKCD;P<;=dz{%wu05(I~Q37vT{uXMxlh=3C;z%F`^8e_Dh|C);kv%q{#)JDpX? zb|ZCaQ8ckN!WuF_& zdI%-cWA9MaB+^{RE#C?|DV=(BQdWp69Lo?fVTzvry;|-mhq=sL`^Io|)jordoFJl% z1{)(X56HMOsRlPU{?4Vi`P0tlqYFGC4`m+a4i(gUUKQBBBqMyt5}#nY+(F9r^^pw# z_mh%*ApaYUz^MZhmQ*sW-+cY%AkmrX_iB9IQ)@!3Y~Ids-3czBLHg&g^$(mz(7r@= zCow4OTr#<{cS_W4Ap>&bWg$Hh@=`jAPaZ;~N)x6hdnbnVq z*?GI2VYLNBY%j z`b}{B(gkVE3!|06pZAm)_qGdt4p@N)2|2nj4L-Zw`MMiqSSr$F`~Jr>@cW|)AqEFF zeblK;r&)h^(ALTFB_T#5>K9S3&O7v5i+yDVCGKI#V%M$2m9g}Dhsfn;IB5cnIhn&{ z8H1?+1K4q^?9=rNixl162|YE*V6yBnIF>O}k#NN?U=(v=a|G#v|}8BYT|Y4+Keh zs<%6k<;%cJf&nxUkuuxV5+qWI?$Y7=@J99ZAT01-tOMwCV*L`YqchqWz{Od7pB*m* z<4De3I$}F;L5cVcI-eH=d+_I6P=$s#>MoEJzHIUckp=0^6B)iJetxH7)dn|hu$*I= zjejmbmy2)>6<)&fQwC)g=Er@nG4ze!vz&W^?UxsRLBd3?l6}g|iW93Axh0cSiekr< zo}5D1e+;yhcgspmWX;hl&mxV3T$GVPIgP$Y4$I=Oko!OnrFQ`lZe8GTl-!}gF2vj= zr|?oy=VN`L7LJal@I&~{#WUSu4V*GN%BaT?f4VpCq@`sWyx9u)t1>=}I(>Iuv%FQ? zO-R*-H?b++_6>==xL|D+cUjIVPp1jH1ST{LY-)EGD4e-t4h(ASQQ9#9I(;_^a?Hz3 z1N%Ho5hXCXNv6g&fMKjM4Z>#ZcVNxQ~2e;J6W&CSQbhN4^AvpYGM_aBkK}B{AjZ;kIrCneawqjZ zl}J4}i5L0M`X;8Km3&sFT5z6J4=uJz&#yi#b?K<#UdT8Y{z ztufGI$gic=Zm0D^>vZ*c)qtaL%Y_4=W2@`WkuEG4Xe56lY+T7db`|I$LtT@fkZRBW zS|DK1pH!n68AT3W4{A%fr|~KM8KC6&$WSC~XIqE)W(FwW5Q$zNR?YCLkclYHJhd^j zi1~vFyEQk*>;1)s&6KrLmM zb;rm2OO{CIU8`!E^(){6Zl7lRm(UwdZ>+ZufW;uY-E&3ndF!mW^v5+nd5EQUPFn?8rP;?U>;CSzsi`kC zUQ1M%!H<~^HsQy9lVUX5#94h?2WGaw&z~U|@!3VCGzq?{GN@7SA;Wr#1K8^pK2yW+ zopC8*ryezvu}PBZ)FOCq;NhIU1rb8iiG*|m6d2y+l2myWgpIeP+5A;=1Iq9I5`KB+<>bBS9a4>=a(C;J&H`wkwnQbb!gPk0?pJo-gd%ZQz?uDWuDk8#-QIZ;PJ-eSk&1ek z8CDAr00Rx7h}s?{O!POD67POusPb@&J3LD^qc;KE@DvF=D%p~#`8Rpci_bhdz5jB5 z0y&hvYsA~&&M(EWDxj&b(9km(0ZAtFk&pAFcCB_`HZjCf;Pm%OTvdXdK!Dnfl=_lO zErmDF#;a!J8Vl#n1qXrxI*4Dss|na%)N9k!YNdh?4iN-)+!?o`v7M|scEkTGv-r!* zA;vita3bLrpX9Xvg@O6YeFlRXY?Yf1T}fxbp)qwKL(HMH<~Z9iK7R0`{Nd;*KF4v9 zuYd!VGj-y>wuRqe}-BFp2 z#w!Ig_jX+qQVF{#OO~i~!#lIeD_%`9iSwh%6B4tF;OcoAY`Up>-s;VALIiQs&*ZY> zwmjiUKp`H?y8smy{bjgZa8>Th;MBnPl!2inIpc|Lwt(@Jl$Ckxca#kg1XSv!6m?#E< zPlJe+OIt6dMr+wus?9^ycGv0FqMfcaBX^|d7&?LM4e z?s-rOmn~-WUvTDfj?Uk&<=QJ51f0vOtp|dhs9x}2X2XFD1M0+;`A|Pwn$z$+oQcEK zFm?fCB}V4f^8D8eIu17`7FIe#1CdXId|YZLYXGR$hJ9qeZ0S%Fxt=~C8Jy1}$>xfb z`0N$9{)Ed%?0WK5Y>0=#>Q<>p@muqi$Ujii@+YfrzG_u}&i&vX48gTmUB8qJ6S0E` zv6W-!WUty!t!p)W1iiByTSf~d)GR~cw0usVt#cp@(7mZfHFIS@16J=TwbFagN5A6+ z6pV6&`5gNlQAS@zskU+21#T2dNgK(jdCDZ>!jR8I$jEf2g($-K`=VhLg!Y~^w6iTw z*ejbt!-?-D@K+M0AUE4-_T9Yg$252Ey-cY4C888s7icm~@vje^Mc!)kLxdi-MRf(-r`T({Lk_ z;Zwi2qigxj?!rw^1OH+Uqm!hW5Jm$CWf*N`Ltt6b79;W1_e=T zTc`pSaJ6N~`uJxsbXptm_8wLEyV~0#Q0DO``o;>k_J)$>9GTGHqvs6a!dvNb`aT;P zbcTS6>fY(R^M>KuvVfU;h2?rwA0Z(skiaq)bnt!UJ*r^t5#aJ~2UxavD|Kej zrBSZw0MD{<=zB31yPo}X)io>|bUqOv2MW%Di4U~AwBQV@HYHDkOoJFY{RQF+G@a2* z^l4b{!XOsU^yi_#K*unMwXsejIf`L0Loc&7gTIYTKc1@hiQ5wf<1FLfebUj!U1+(B zD`p8(IrQ%eEAMLWHwW_FN`aJTqQY{b8}yMC23u%EOn>}A-FMWY+43hhOi*iDBZ;k& zV#{8oBk1rd24w!L0rFSN>5KZlPa|N4+)K(iw>7Q&@8-Du)?^lAhmtuG`AL3QyHl|tZNcJ!zo&*nix-wh*jTdTt=E(NQpl}dknERp3bs5iNV(_L!sA9ZkQ|X>L0sY=a3c@= zp)6gY*afa`Ffd;nH(mtrdSS^KbrCXy%FrAf1N)mxkaDcbeIeK1vrN?jFtR46$+@ZRB2$Z<^)lI9C#4ch14p&m!DN!$< zchkw+m8aj6+9yCR2sx}tpLb-<+(!*SPR(9^5ZAOXuFpO8NI!07GOkAnYxs6XR~y?O zkQ-C}*4gnAol?g8T|3+f96hn!llD|wO5|1lP>_M~Hs1c3527(x#Or&tlL>ms+Y0{F z2Ct5&V}8vJBi%04U*d19=)^Y~eDHW1r>>(YSz2cMZY6C+g3(9#lHz0A_A4S7d79Ba zV@f)7<~GQUaT;=hiG5NTrCU>ShQECTD&pLC)*956NqM$E{#%^EoS!@cz`%v6ft;=I{KLbd z$yL#%o7@vK=~F*UFONnRZ_b&f!8w6tlRb*SqE(MSSkR>b&nlZcSwcY=MrM|m(_53| zJ=0Yd<;dfeIgpbQ@!1hqj{|?qnh)yp>!OLMwOCCw=`KrR18eq3PYSs8oUrXBX3ln*j^0=oGZT}6)O}djjo6} zv=U@>iwk@*j#n=tvN>$~-~)(frCrPn9?X^^Cn>JM#6bd)D3bDXp;8}eC9pm72LKJG zByUnoLEKidvcg3^h_(G+a6_eR0geA~Ef$;N+>r6|rcYPPl#n=uASWZqJM#8_EF95c zh{R{Skz+H$$a^D(ZX{D_$@211@qHq0A7BAkK5j`;!%F24jnq6k&KUeEvt^>SbaX({ zw`uQm`?wb+L_p`#b84^)cL9B^jS~gk;=xDE$Y|%kp%N&b;b2a!SltrqvzBJU2{mmx z;)i!9mOinzbHm<#x1#U`P92@+WstK!Fb^}WEK1K_LmwX-=LPaY8(IT2r_w}(XXWGM z9akIf$NcNMTjL@k*VWRS^YIp#_1(c!E42xv%+=zVW8HsN?3NBo(LDsV;?udTs1I$U z9@u*b`37TT&R{l9^g?qE5#0SZ>YB|E9RE2;O^4x{C9aS%?CA-h56@ccj4eA;8>$gj zJ%=EKp8cg^)m*$j2Wque^X5%J9u4qcsghDtGA_0aBR45ybXRV>8i;yecfF?}1U*)%gNW!TXwwPU}L)8!)b9MXHQ(lRcJ|!X|G2b#5X_h>S}RNwXE(eP#=ll zIMK}NN+T~q9tVIX{~&IOrO0gX{e#P{jDplA2pVXW0(5gd$Dvad((AUd0<3yv46=W2 zdhh6$O??Xh?vhiTy8T-|8yW&(pOG;;a=+k9qu4QuI3qfZQ+$l$@|D4_Y%D3t9~Iyx zol2{RO}GWf^V=<7T0Hf{PSpID?!0v!r4(s%7MJ6H68~K*h`lJ0ZtS`#X$P;ECdG4H zB~vPSh=3Zcb|v}-Ndig{71|mI2~|w(rGs_~Sdh&f-jiPL^>-rO;0{ehiC(nMCO_L0 z`Y;X!pE`wH3ADk`@-$QC#0O#$)Vz9Nc{jU`Dymyowfc{w!H-DQfv1ym9l!cXph8Pk zlg`Ki3(}RJrvWV0@pC6B)rE;rF_Olu5Q+_Jvd^=;0|O5ToiKS&1)KHL{aF-zC%tZA&EpBAZ$@L41Y}-B z4vLitg^cn=Pd(?F?%*XL%5-#2wsc|wRiP0*X$Cua6Nw~3I~4FWW^+E_ZYvI#FFc0X z*!&-_?d3%oT3~CmOVC7`t!|`07>n112GZ>K^cT5%(%a6q!=kZ3Vll;waD95mOs~kj z<<^7jEijjD+!$GlxSe>M5wX&EcJ5Fa*t%14-H&5fQ*(c>Q%lR)`A$pEQD;;1tQ)ux z7SY5~wvAa_6#5#lk@)IbC>t0pk7PsaYG$QN%YW=odrs%2Aq)h>a>Wa79Rk!D$h@pK zmN-Supe#ZsJ_n#UxHMH&)q>BOH9t=NqD@52wG#zwA~3J3DXKwa-5SZ$QCu48{hApp zbgv2;TXeW;{jQ9avNV@GQl>V%+H4-p=`X{r(}g%{sw(cLJ=`dK5-V+dFI zlE!rklv+}Gg(Up-YG)ELGD=|gl835i!(q7VfIH{@wsCLag>@xu5@Er;<8owmX}%OQ6rc$Wk8R`4eXzJs|d1 zbK*480blVS--C7@qM+o1{hBPV26t*fgf&(+8}B*(v+>k!{{6eql+piaUk;z)EhHMY zre{YBN9;uhuxUg%D%DJf(Hf*n-8O7X9W+0e47I5|A#;L9n;n}_zl?&TaXDNXiMatU zW-7r>G-yi2ZX7WBkbQjvEtgb=#LxZEDU`x58X|@rnyU93f$&TVBDlrFLgA@vKgXZ5 zgIVmfxru`(qfweHkLijBKfY`H>wyV69bls9Xz955siQLCvDKckFDddj8dRyur5uNO47IK$1mzt7?I#oQ#)QETD~N~tDFb_U^PWV^D6OCDmU2knEe zJ=kR5KIwY9cbY`gctJ;Wp??5*jm=~l&X-7>)Z%PovdKks=jCW0H^T0og$W5@;$l;; zl7(Z)<_R&ErMzy7&p9!_D&%oMouF1$nnZLu5Gvd_PtMjqg!KA3FoG~Z1<&?)YHqEvE zz0*Px;LZ^C{008jM1)?jLkrr_X_CnE=+ci={m~+DLg%#OzJtXAs%3iLe1~(lqyUWp z4&v?)|7f@cX(ND9LWSa$f1Sk^l^S|MiC2AD&ieamUxm?bao%}LVxoGUEMYM~D|&TY zei(BQ7aw|(xdcPIwblI?8tG9q;)gkpdC`vw@PbX6f1>xP)1Cwc-q zbXN+$x4W~Et5nEfXI0_F0U)oO=+e_HpQQOMK7q>)ZZ?Oe{8>&{pbabEkP+O2=_A#~22WIP)x9bwjGYfk3YYj-#@|T}X}6Y+PWrwHfWGj`JxsAi}Iu%8|uo=TGaoRAy3Kj zlg&#jJt9y}pJii4TWj099xCp&poV%^=G3aR03obB6vRl9QGsN$DQqc=#10auTo5wd zcyC|`^X1N_IKX+)*lUY0m}fn2b^+3L5XCqW4J)O*0|5HFwpl1W96#1}hF9kftGMt@i<_9)GeI*i;&eJVf2XclYuW^e30aV! zwBG$n-cAF$Y1q=lp?q;t$)+Ie^PRv!#A`CQ>KY#6{9@v41%8Aa$30Yoa4hls4|%nx zIZre}t@{dIP_JG)RdH2{{?M3{{DC0FqTP>^F3Qs z8)7<+CDy?$2DnARhCX)w$Y6R?bO9cjW{1=0Ox-t~|leAlZWOOkgFllnCQJ&0s5KXZ%801(p(t8a-)rvg$B zd*ZvBQc?;4pOTmV%-q%;I1~WzhT-Rs1qDr9{+;@(e-h#8ygTlx)H1?c7whr6U@%fAl*>@G{DMM3Byj#}118f{lX2=NcWFKgT+jr06 zGp&d*aZfVEzf&@ABGa3~usRAE*^<{WHWw$e&PDU26q5ymRVbOW5;Mhd$gFL&BWDT0 zva~037QALjzHI4=lfw|9qi7=M=*fTx>+Bo@ zzCKIVAwnfKlaYU>snf8^rj2+vUs~EHBy$hctqfRu~ah!B(l#T_aLlcNLoH z@oh5-j(7N^YC=-=!qc2R0;<(g%qtDCgm-ZcolWyeL+Mk?^2`{tx4 ziiyO;>nXk327DrB10#M{PHd>&R)YprBB<9R2-8uBmsDFK0c{NI#Y2f}eT>EdeyGiF zvGpn5Lv#JljEGLZW`STpFDU_$ukLsRF(NRk^TKh9^!`%wGU=Ec!o7}ln*C!#NpQY! zmeYhHMariDtzPB@6GJGD?Xv!5gGX6krCz;!oDoz55PmRKDlnp8LII)*XEN0i!l@A- zM2-$K>>Q`DI&6JvY&%BUys~-0*5TuqXp|ezC-#ArYTEABj@9#yXn`k&O^E8P zQm+2a478jLRSUF7H#j$)Pd>2htD4*Iz1*~QfjVT^r`N@oVLtbFm1bnuU#p3HO)EVQ19z$RVP5< zcKsh*awDtDdRn5A*o(ebH%VFh0v(mno$?hzX>k{WPkFKgUkI-Q4kL@321VTR3`}1~ z4mu}yDI0Znoyntwowvrk`Q!PntI*o#q+e54b`u?HyRZNj6navi3la?do^NsuyeutB zK;^|2c#@YLkAD{O=9GF**yx2%jI3dY0O=QUo&g<=zb84v@5K*RN^!@$e9z9-yUYjO z18ScBqI+4!V2n(+iboI45Z$m%fqu^=+L23S3nWZ8HD2m2`)nm^r+ZlsrO5_{ ztqiqqm#SuIcE{ODfn(bfd(&yykZmv|53hjv{Oiy&TLVGE-k1 z>T|Ham69gQbqVAq@5LeBu#kxsU0Kb+#;PBIIOI3*Iyzh84`%6{M68OG-Ux&B3r9`nA+L!1P$}{BTp|W~anOnWrZRcf$Np9F zzH5sw>J2rkaI5ksi(-pq#RfdSXz%%H@u($^|4 zzR*F}0Um`?sb$0L4i|y{RN2o*HgN#(W+!<0^+B!x&c@KR%YjIR^Tzf(6zaeRHGhw7 zjX2(Igf?cCgrx#VCd#&@6rIR< zar%*r^W$)vjXYMO3;%6i%MV6R5>PfB4V86juVweH*QpVD1pg!0^F@&{YrNJiH!aHc zE?;-Wr*?p2`+GiMirmjzl3gDZ5%P^eE_M56Wl2v;IgyQ}QFMsQVrOjnIQW$*W!D-M zDBrBuf4DEwsDEi_q$*{MXDQrHE{%RAHdl1MlWFIZArmbgTlsKSV^1;!wVtQTRpIu> z)Y$^LQu^>DUkW>$UckEng{$hfzRH%+3!P=%7XKwKJleJ`xyd_@wY<2h19_fUi1`AH zyEf`p`bd3A>=(k~#t?3J`r8op5U&5ly|?*Z-gxrmEhQ{ai)mf{ViRyfj*oELUAtO1 zK1+vVfvtlpgSH6-Uzu~uK4V|sGX69|?rx~d4=663;k4&O&HLSDlk*(UlrmOd?To9< zI-u0whz-!U-!(O?6NU3hkC0E_)-yHbrlFrhQ#oON@9~4|rMZrj7G)>nB-zI}9MYr{us!)M3`I%t$2ERA5(Tb9pm#VD zWK7D=YbX4lTxtlz%BrND!fhRMHy1?_9CtE93**yWxsp57Pw5wMFxxiFlpXNChf48L z0z8F=ye|-8BAN2vI_bEgQusqE%u}&AGrlraj*Mgu4#LYa6n`m&7FfUiwNl*iQbONJ zc@q{J%!F*RTay`-fKO8a#GMx-v@Bi!*wPpw2b#xFDS5ROM1J5=a&77VWk(xoZkEYl zmR5cN-BMIm*)!3;L%1+8BV-B0nSK>A%zM*Ewx3o2u?Q9tSF`sQtPaBJ=V;i-7rKX+ zEfbK#!#oRcVPaZiG$l<5$IJ6CslTLaUHI$=|h7qy>kp&(T(58t^0@nB>PeqSy}(RQZT zLRNM>p_u8c>e7~JY8<8+yT4H9;u6?Oe!18^AYU11bj@S>Lzq1@W z4&0R=ih{t#oDObrakRhtsD)^7O&@5L#YC!U>as{xz4GeI>@NBAAlSZ}7V)#Qg8gT+ zu7VzH0SlTuzalzKF@SCRaFAKP_|2|s{fZK8>=jgbec=(4Z66o&WCuVS<&63;%8ItA zv{)wV(6c$WLX87>CCWa02Tf(h0qUVz?Lss3uEFG#yNx^;f?hczPy$xG=jG*{>F_)uyx`qZ zCC5>%;FBJ;zEJuLpZ*>yuFZ{I;5DVr?jrRFX-6+n<}8bfROw8lZ$^||%nqW+hT;+gBP44nf{YW{#{NSp z-jiW4`w+{Bc)H)Ym!O&mUBq=-|Z|^B>>O3(BP+(y~Xdx_b}{pF?QA za`{ocVcmCuDBKKtjZqy^D7cXPil}a*JZ?k-1xQd7BbL<`f%v;zY=M5v;=sZw28HpL zqW?AVA6rSLWjF&NH5}o+1ogI!mQS129L$K1AH`S+$fHixr_1o!aDZ}(hDXYAd-W)G z1ND89*$Th1!ZmDR4azAy5F1|V8rw>Cu^+w@-9?bfD0xSgw&kwrxi_EF(_H8dI=|RR z5kkj0_nRwT__t7vw4<~pZA244P))rzi>Ytx#@GBHkzR zV?^#prns*$o^qCM&*-8b>Uk_lR{4yRV+tXa#Rrz1P4WO z{a%lc$kVHrx6hid2ouR=<@wuS06rb`fyLa!{4rK)0`4Zs7mI)uK#6EjSN^_e z4-wkh6$=^1~7go20AfCkv|MRko8%?+X?MhbneM8Bsx3}_YGyS=E2)O z(uio>6_I}4ftUBBGrU`W|8dkBaxqFd&XG;eq#qh`o;D^J@}`PuUOzxyGf=7-HPj&G znX+qg&hFKbCjfq-j(WAFzphEr%8!beL*rgyXoAFXCpF*B#3xVw3C?X9_+ZmttjDFd z=D;rLNaAbJj2+xu7piQYkbsLPY;IIvbiz{DK6`j>2blkuu6k;QF@HdwI9q*tRCWep7~e{5Dj zAC0s`LvwXn+ZobyA+H-P^9NmmBikA`SB{=#=M(y9I&UNF+KOxY99*eIR^MxVSUlSj?wIdPqo%taJP_a}rSu)L4 zQ>0EoW+jP=h{%dXgzQ6Jvw?kDM4e`mq;0z4x@OO>1Ow=oJ5E@ln@_)+ru#c{2x(p4-kryO~u zPZ1>Ez!@K5V(f-Lz|+8OLih3N7Zdf=TXiaFy6LY|9;I~dc;Y+_4X1@)^1mAR`j8xr zQhx)i)s|jQloM>Yn8}kQ=CS$ zQ)Sk5Gp*CN`)}DS)@Ln}q{TyZt3=5(mvTH19<3XM|H0J)mJ>})#?50}nF|VZ?WF76 zm8&jUDb0>>TRH8)=ho3Libuy<0|3qmVH%?`A2cm<|G-wKyEnts<`F}j<`hmzZ${u@ z*(!JGUV?r41X5`jrW>&kqkkh|9Af~g#LpKs1L~Omdy%H)*_jvh@|pdo<9Aa>D)s`e z(|BhHw)fw@+}kc_z^oyE?*uKxrUo^};BReD@lF-6u$=_&I<4G`CnD(t)zfmPBa6DT zQ^W{d?EvHxb;YB%wQOrN01Otq!wrpOo4QzCYm1-{ydjIS`mkx)=RlD#3<~kx!kG9P zKp;cf$@ARlXe-Ao+jWLlA056{(dk?TY&Z@|MBkCeDyFHp^YZDrm|o_BnxEl)>&9O|ijX^btd)clJIn;(9{GYx7dj{VUL-mtiP2V}`=g*k z$U3mndnOf%6nkF<5iw_R=0&!4*r1G4ES!)@rGx@d?}gXiT@M&o3eE4+))?tDkl>#S ztoyq=pDVXx9hzcQ#`TT4?i1O~(wV|>(PGO!yzDh(p@bu8)F^$yt*-+u62oCdH zsovzChIPopvF4@7EJb>k6YlJ~$M+`!t(FcOys{+HYT7r<;!M`<*cCf>JLcee4pa2Zk)#6yVfVtP|(T&!(ZwvhIvoczJl0wbq#=MNzst?1aXC6Ve{G44WsSrGQ@8KJaCIvD~whs{ZhxdRI;+r*6YsCT0}gYc>leZWX8m}ZUwws;6;jx z#vMlE_$2$cXVQy8=W{a=bMi5350Fj-a=|;!Mygq6^ zBY?6ZQJ#g*4iOhbm-JVJjJ4|x^i(@XwB%FH)fsQbFEc8FKucPJ&<;}W`V0Syu8nZa{U{)0SBIJ`?HbDMrrgK3&bvPV~XdiQI*xx)= zLlt{*XJT=KWB}_pNu2;~;()K^2UPxM?F+y*mY4Mp#Qyh1AhUE|H!fx0+ljd3M#2b} zcH1*9o8yWFy>;EaSiWCL32R1Rh0Xfv;9~M=M)l>0*4t%M=QS6W$AHbuw!y;n7ofw-jgSP2+@Yc4(lQM4}Ix`&O?M?NBPpw`cbOfPyN zzA#FRUP#49D~dx=Sy#`6o&Ur6nrf&`?B73(ZPmThi=XW#vE*H)!(t_~jo{bx73b{u zXYD#?yl^we-ZvdB-fA(=q3?KWE?jx&@}_cSSA+#;CC|kW<1h3)|7_9I`3d`eKBf9@ zdcRowY~Dj$VT|Iq=l#s~79KN+)lCbn7kI*5zh+JW!KMO;c?ro!*jNSVc_G=(b)eDicA zm=+H}v>Gnwlpa@h{@Yse$LIM0_hjYCe#aMO4WA3%7lI_rOfF_KCF93K=Esp% z60oYVc`>m$^It{8M}Rgzs^M+gRP>yMBphL5pQ7o>cr9XuBSGnZ2J{=u{E7E*zIxN&iHy%aW(9IQC$+n5#vBLtYQ&_rpVR3ExpLg56IfGuV3N3{(Lk%2tt=lm$7HnZtnq`g7&)1%CRT2 zp3RMnlrfWv4(Iq5<+lM~^d2|E^ziB!EJ8$w25H?%nO@!Y3Ypp~`5@$g}i8qd(==uolptIRwS9+U4bAkJ^#VF_{b3Wl-tFd9P z;bIZMf}afnc8uFyb)TNAUa4m!r2wf3>kzR$}nT2cfzNr{gVe^0h{``55!M` zJfFnOH)R9m_10E0F&;59);G@nCdkuO)#hNr%P7JD2k3Emm)V6kG9WOu)r$q~z*7Y) z$5eKpD1%OXbTM|$0snip+#72P-!KVBKYoscKWyl9cH7e_8XklcUd^$Q^SfAbvkX47 zn4p1ePNI}3y6TDhoAMo1R#_Z(+tp-T8~p=MLZW#xjX_FBFm^mE=$h@22Sc?qk<5-kX*vjr7V;{F~T4@tnI#ub3j9{Ctu z4s*zgO|wTkE%pJ4tbytnw-~Y^i9H|(sD?dcrDr4r5nqI5)>Jyy!|AI;!85g(=NO7N z8$FaMRlzp+Ja}__Xw&~)jVan850HvBu$D_OQQ)H^o1dXUM1TGC@noS{h#d1X_s)1#m zKQZ4Oap0`KRoXt|X1GRQzE2;`8CT3 zYnA3~|AABZ8D95{K%=4vWOLgBW074b2w+C%!;KxPSMQS&)!D5Ep5)-Z2UEGF(f}Z1 zf_5c&PE}u zlod4hEz$Kw$M-Egg729f-8?vAD!#ylQbG9&vO_p^G@i_#5-RgFsmsnc8jphIQH0v4 zsl|`cHMs#(y&Riq5uz=i!Bukv$_0;G26m0IySI*k=TU+8M@^<0_&u36zpE{sg=?fF zch?(-NL5;!+=FknEsvQd??XJYdS_*ifU#w;LMQ8Fy}=C7BTGKj2K`i@;Uu!ilpaBj z=x!cQl8$zM<$3R{$Tk9+- zlfXENH0xUD1V3^LHqLYM06>_};e|?|2??(*a^FcO2}$IUsPKinpTP8ot`^;==Mxk% z@64V?7OGXI{JEQT?2L;_*yP+S`*fn2rt6{{jj5l6Mq(L$@>Fnp`bVl!Mxvdm#%|Zy za{s$O(8EpKmEv6s=HUytk+m|;nUYL97wHzk*wh{7*@2+BunhzybmKJ_6nZ)u0_}Jn zNnK20V?rA68}V?iQKv2<2%qXEr;wu{wG4DQD&I0>-?9ap$zQc^=+iIcc3WeWhS*&O zt;2CTCtTuCte&bLBAZep1-ui0hWa9gAGQ{U5U(*)fE(l%NZ3hZp$;zKq1d-MG2rWC zwQ?9yj*;VCO}p^tG5PbpQod!;rif-9F3?>z5S!Rac+T6ozV0HECwfz`ku}~P($6wG zY;I(s5ob;oRDAfS#mN}9l+&tA)+Sb@f16z%b8mFWdhrX_H}t!rB*$PKLIpd<&fd@& zdr?y4ny~ss?+~nXVpz)H(!)qxhR`?=>^0hV=+p;q22|bU9#U4CZV8MWj`l#Z@8GfD}+DLKJsJ zW&_=XH%XV7TuS{8bHIp9g_YnhnW?2moorOkmDBl-vL?X6^qV+ZAlMjtg9&%Q&15uy z^cJExS~=GvcHDDghD2Pa)-0HeDPb_Y+zvR&|3VKj{SWjIBMa02XHQ`wU}I$C_+R89 zc19M~|7Vwq2e?Yo)*{=YVCA!wfEO$QLC@Q+U~zD~00D`B5TvLV>IH!;LNPiZ<)UKc zGa0U6q!BJvIH^;si-p@MkLV1t{m7KecU2^2~ApOnEpg$6X7#UlPJ>9~U|!Xv)F zK|u*~2>YiTTpv+d1qOJ22)Hc&In*Ou^f97CU=ctjCnP4K01I>jAo$l;o!7xmGv>zc zjqWkDhVBc<7n1`-(gOc?NkaAb#eo#@^dJGRuYn-ozhwt`)Jlwu0n9MbU>t%v1`vz> z;$wpd=la$*e?5l30NMmH zKY)RQhfDy5n3MzpGBPqy1hlk}5WnU)#t7H)cz*{~K>abR${%$Ue^v8;+{Ci@9N%p@cl3jQ#^9d) z`|12J)lRy!)vM&^JIdn(l9`bE|Nj-ekO-)*Yku!C|zP=laV7KRO5MEzU-2Moz% z)A)xrFv3B)e@Zj@?IU%bK|=@##w1X$SD^!5fPsIeLvWbuxZY9u^^4!O{|wB}^S-PI z`X$cvhsfwCK><9ygBXg}^^DaM0|XeaRk(z|{i84msQw8QHqZgs#0&sXEM`5c8;=wN z79!B&r_)2g2^wgJvHp$*0w}^`41&MV3qTQELcR(9V1J8a8Q|O7U(5WiLbHwat@Y6Z zKn&vpk?Cd&|2rYt$rX3gS+1ymhwQ^A*3rI;ly1FaJgrxdCE{75gi_OmH}uUrSP_LvS-EX_^XtE^3#(l&2)K zNseuOZ|$~Jb52uSA*FzxsW_W?X>)9Zk+d->QV}vQr?A-O zSF7jMu1v-9EkxKYjCUkQ@$mP2JkRW$5JXb_Idx6Stpjx-dLlwaHlmBk?-F&X-*D4`=rQ@ON@#ZpF zI=G0`JM`#@zbdl---XkF&@QZ7~7oB$nP+osmqmwXz34Kcuk%7x=_2 zqaEJmEt)@FsxvWaJ#%Vn8|1lh&xGOk^s9STQU%}6Kh_OTOjxgGgg8ln943!yfl(y? zb2Wp(N^wX`NzRmGhFz$aE6 zsn+wXtFB>ql3UT5xMB?h)x`a0?CWdGO-+yH@&K$RxPh&0ZBnjKyFAJ!(M&w5I)*j& z*+1N#E4)pw&ay_p${ZvylJQ-hc>(U`!Oi`h0knm9f_}DF6LZ&ud1gU^pN@ChPBwFZ z_h&Y#Hv_b<;v?x-FPx9HwxVjC*=@$BY_5bAdEg(TgHsDMw}9n@6Fj~nNfo^d8YUq` zj=?nZI)C>Df)AX!5h!%^sFPoH13Y8RuYvjnVvpzxm>ODzx+M!s&Uk?{@4?6P)7@6I zzDz|O&OEc4)5cCwUK5ya@`s&jA*k4tg;}jO&jd8g=J(?W2?oi7UK37i+JdXo7=B-7 zNVx}APNi*$(ynq7;aY!#Ne7pwXzePS47Oo}jYsqKR;qt|RzpE4b9ot~kIs3L`v zK~r_*6`uLZ+G#XSDCLqckiK$DbO_Qn53;Me5?LgsDR{xR;AaBQq}&_!Gki||y)9}E zw6XX0WT;Kw2u`Mwc@E(XfIHuPhe+_>o5|~YK|OGjA++&-GKlM@?ku6eOEVeh5WP_PcHPMuoGm=rl| z*VHv-o>0@GQeR5b%oH@(Q3e0lTPT*2qhOUuZ{kRp4yX6nw*y;6fzuhuN*EDq#J&?^ zj#mUOGT@iCQ4~oMv`j|!lw^T(7qhbk+<}eBZYoxE)#XmwB7vHk*j1RtUQhq^_26hs zi$@51uE;S|VY|Ww&6(v=ZUt`K6|WGdm-KiLr}OD7MfG&w*cJCXc&L`}m;*}C>IXvi zu!>pnXqdk5$9|D%2MA@ZL(jQI1l{j{Z$&+Xqn5;+pIsGIi&a9L4zjb?boXN>tx$#l)+!I9ML38Z%(v7H;gex5vwNf%%&=- z=IhZtiqe>8*(0r)kgv$QIEcySsL!uQAjLw4fq;+7!NO3YPn$sT8O#JK>U=2zg2i|f zg{HD&oA1tmsU>2o1t0l1orw9?Z1YZn?>EoZVWTaD+4t17>TF`+N1ubK9kik%&aJG`c`lW$bQP#3%}fc;fPUbv?1xV4zJ;FX6|yb~_m*rxdOv{LQ4!Qm zp2Pvijk{$BJ5K6)#EXMCoxTW;*pel^*}AYMgrtwL2wF-?_p;mjAa&>I==ATV)h>X@ z`8@d`(ytg8`0F*FM#uQ!_aKD0SHMn0=&`6-Z@;j9x6f3NqROaw3H&CM6YSMWRx5-M zewYe%8qB-l?x|IsT%Wag)I3_XyR_pak3_N=taSAsC>(U5nG@Kb+$3070}D{-wzeJ zlGQaTqPG~9CMPA!L(^|L)lPsQH21N`scOR_-U#p6OJB@Il2HgZLgMfxN@CpSrmuJ^ zgu7I=a4d5dJ-AcbAQK`I;dCLmAXQ~~|7KLHlcU$x{!JXy<%+C(HZ?E+d2TvxPf+t) zE_H}OAo3v~_vA}Z#KTD=%*S`XjB}nkGyaD=nU||kO1OfOPa^3LF=kTRr)-2Z5ACP@ zjj2C)s~>7eY`{8x$SL;Yr34_x*W#91O}v zgTo?kJ1XTxIG}Nrv8!|NFna_V#Tx8B`FO16Pw-6-he(N}jzDaPl^8V6!Sf#d`vA61 z@H`-tyU1ut*AqNu4(t%YStdHLA-2Z*$po9@>;cmQc3Zz(4TqyIksgv;)#tH z)1fcT9Sxckfmm+D(U|hP*i^&09#25U&|WbQI*jElzDkii0W9g9fNTU^YymcwcqdO& zQPTCgdIotNtbXxwNTbd21V+2djjpVxeMlnU%azr!PiMW8D7&4J(_6*YFPh)^bLqCO zgQNa4)b=8YO7!nuB-$N%$Z%UBI@MO7lRKpytV>Ky_?oATdgrA-s^5Q_x*!P2gXCP_ zW9Cg16CCXc<}Sy{m@~bO7|_030+HAN(dAJU3b_o{qXg+tjgVHVf(?;A^UENW{z=x{ zdbS8mp%O7sS82))F)eom(MXP9o6)w98bduWc?pW|U}APv#t7pJw3+er!1G^@5oDuU zTd_9@{c+*gRL+1)px9z4*t$_OZR6>gb0sNKTo%i_E*1FfmxEMjdE%Il%U=ytQabDw zc`BNcaUZ%FtK8t$?f|DcQDtA{?#Y$MiBZ_q@$TSq?TVntLtFisQmklKWOG2{?m!NG6$44nb-}Ma4@`je_41& zKY^h}*3C5DxY+Of%uaRy9nDl%fD5%|2NmO@7KIRP`+@j6l(}nuprohc8;2wSZ*!q~ z!+JdT;Fi ze4O=5SPYbJw-_2t?A2(0cdkTpN5$>J@B<5io{TqZDum~^WkN@koj!nvgltMO^PB{| zM&1BZ_Mjh_>G3zir|?<&buj$M{>u$=w+~$ zBe^<>G{bK1Lo#2CYR5tH4!EygwDiTs3XzU$>cv*w%w<=J8kFYM~my5n4!7k zUYl063#(VP^zmVQ!aZWlJdVeHP}VB1IKyW=cF0|ez6|+=r(@OSF?fY%rb^oR60Ub> zr5eiBirpx0^cZ$wjtIVISRfaG^JC-IXOJQ8NLD-j`5eUe$ZvOPt{do?zq?v^rj9Un z;gO~Z@^-9nSPJzJPv-lB7K^Uka^~DHaa-l}C{P-fE`r z1*Nk_i&TU6>SsS`bdy~Bc{wpE0q;sUsZ3|DJ(qf(sSi4DPI=Q=EfwHNc|%f#$z17r zrg8Sr0U86uy6Kh4{4Vu)s)CbWca8|-@%7rI61m9rOEF(^MrH*UI##HmfEljEV~&0K z*0k00@o4Qn4^i)gb9=gg?yxTJ#QG;yd(ma5KR>3qf-rI0x<9R~!|h$i_NFJ8P_F4$ zS0yN~qe0#Pinp@Jw(xe4<1aA({wOn2C<;98lC@7m2gAAbDn?u!TFq5v%SEu*GC4ri z2h)6*!5l3W=SPea-JK{yo&DX$B{&IrKX$IyC)qi>Wa~GdNwS)D>-~D7GlnfINkeN% ztn1alAWDBYzoKVG1r9j`;R2a%=8-*P^D=5Iby_NSOeDQRyry}yYC99^=72W{xRNp0 zJjy&o%L{t67WNGm+gAJDH5B*NNO&bYZu$aPt*n4#i{z0x;{r0xa1F=axG#lu_|(Od z+poIs#0iFtl&~0vTnnm09Z#yqkm44O7AS|n*QJCMFHGTv74NI|&NVOeoK_q36mz;- zieiSGx0->Y^^3}SghX8W#+#Wlk}9g-OE{K5>BWpnQYV8|HZlC6Gd5cZFwE1DM&S4p z8s!IYHzQWW{96kC5Z%qt8lAC(cRWdSJ5P|nIj_w8nad*O$MYWF3? zVfQrZcQJY@$@)!h4jn2~<7LR7Mkp}LgLQ&IQnF;SMQcK`@A z_sOQ#2%zc5ubn<-yqiE!_EBkS%XH{peVmk2GBxF?B9Od-u9C7l&%;l2n^uZ8wEH9e zO#=EV64raK{*kk>K$@scE#Nr6JPIKn{^@4S)=SoYtQMmfCT%gpn3qC036E?wClO=I zjZylv5Y}YZ&ZmIS2zKwz8ljqSlfZ)#-8CpY);pGH)KcPH$11i<`FC zjIxBn2t}glo{`~iHfP9b%dA6b65?`vEkDyX5y`BZTs!i*y?gfl!lwpH=*4HhdFA)( zhe#)qxtV?`I9vNdFk6PXI{Qz2E=yV02F?bS?`T{JYf5}Hvaj7}G86U~Yj#JS&Z5c@ zK6&&TJ(mA}wk+iqc1fm6iKO>U+I%nFLc&4th1jtdBU$qRqdW7!AVk9B$p_er8JgNO z>)3;VK_(ZnLSNE5SS*5q4=Lz6u&~$CtIXO4W!!eVgN>0iE?n27?{(X~DyMbiw7gD? zDIA31GA@CpSC!Qbx#N{FrVBJZRSZgWXYAth;JVwyBIztNA4uXt^E<}2NG@Bs7d&hY zt3_B~QvX=!&Z}Uv=e`g!W{a#3h|x} z)FuH2w{tDGo|oEX5@VFuu=&g6s*L)T9c)PykE71$Zu-mFFx1I5wQra`emj5xYSW{3 zx@eNM=r*DsFI=)CmVgHwuPP+RB;xGXyRzO+58VlXs{$-wAE(@GDD@$Q`)-NaF0@Mr zHCq#Y8|K{M^Q$x8G3=pX)7%QpKw9e)E+rVN3@`rWdy$_@JnwlV)f$+ro&ztl#~fcN z*}vv8C&=P z=U}(ILBjTjY^^w&^c}ofRYWnP!?^k0-=;>sSu)CJK6ggA=DC?1&LyfmE=apRrpQ+9 zfVI!Qqm)PP-Q-5UTCtUy+A%wk;b$ta%&D`&YjbegmTA0;h8^gRzC41`57)GvTqV)L zwoDU;A&NEk&eRso?@W43%l@{vQ`)LMv4ALh%N)C$6TPMx*nXMfS*HWx;mWifb5eZH ztV$e?#}{!*a58EHgKCtnIHNX{zLI4-<7EElKbsbfKe6vG`Vc+a4nYTl(*z!|70FJZ z*@2_HWjB$09mJV0XTAcSmii(z5i`y)Y6iZKe;*!6s9N)vN2C|>wGQUc?BuZ>t{%ODFvoohz z^+D&H*jb=st*a&%hNd@`Q%Jp->k&_{v@Y!U9tAR$+ray7AEJZ=&^z2%;jA40^CL58 z?&Ceg7Z%)Z4P772i3y6`W8usaK*^MDglp|;;glWTQ&Y0i8hhHW#233Q%GG1erwhd# z%x|$k?LSO{j0nzv@ifRVD^`~yzBY&$y?$Qd&B6Eao2eORuNYpU0@?7Acb1lafWV2N zd1@Ro=sXr5wHj`j8%Bygm7s7k7ybrZJudaU9=W(}Nht8-<`o7uHTSp4J zO-IsPp&!#yO}Qr&?Ohfn!NcyU&^2D8o6|{gPn}XH4RUow+^#?WnTn3i@}`~{3;To{ zyY37xqVKaPIbij0PajlQ*&Csout8+p`)*5XQMhY;@-AZ@E8+C$9$d>V?itcNckJZQ zQs2&L5?;1=e5ncwQcW;35SUit?G8$1)?)GB7>3T(QA zJs6Q6GIoTuNslwfr4$lj8B-|AE^Th`Nr&&Ctc9Tsi?bvl1ibU2Lyh#uRQM zG9l=rb)+6nIbU3k68tc>Ep(u$T&LkeTN27?iU71AWYN1H)t2{q597e^5g-&MA?q|w z8F+aAF=O@KHl(M4ak{B$!en)<-*&UyN)swWce7HCm_P~CA!Gr9`2Q?zMj z=4nFA;C;R>uD*v*oqjZOl^ty5N%|G=Q_Ht5|05yoyEEB0dj$^pB;Ok|n#NKs)B~N- z56akTtsI|nj7EElW4X!eV3hYUO)bmdPK7%OV44fEb$uz5|J z*ND`8fF8uS(lZLDP~;#) zL{Lwopt?7#garW>71(c}2*BP(iF6kI4h`HsgNQ+c6qER)526YK6ggyUrGS%L7AFSi zx}6h9WF&wnsfkFb$$&ut1vd2^6+Su+w33e}!J&W;vMgAHpaIZEwYh>F)a>4W(Dt!L z91lMVkCKp(fb^XMr{X4P5YIpXRSrMGNuYDjJ(xc$fXo6D(eL?J1d`Vc40lFFO>=#5 z0r{`fK+qz(2dIx=9y!7p0E@UDA{rbB?wgh29&Q=*!JG*p0Ndg+VEA`HcHWIco&pST z70*s!X1MB4a2mkZ#3zcm}T~&FZ@5TJke@sPX zUQa-ukdZ+_LQDvVf{2I~FgY<1@bj-Jnh@oq4ffvOsYTJvB7Zv1?qTe?E|&KfD&Ve; z^F8pd+7&!Tofc^P4L?CUh_s*KJ8k^0ZP~xL}mizuk-L2O~5jMDYLS5B4Rh z!4{zb*+mWw;QUFRz}de{S{Cl+TH4X0lNA7?BaUSg9Yjo53haq@;eLL&B zqQJoZoiHmxN)n(11$qJZdvHb^-fgLa*h_E>UbQ6#4icbXg~bY7BB|ruR(!9 zbrC4uuIp3mxrq^UBQW$ogJ6-NpT1L7Ka?wFj3}RP0bHEX z&n!fw<|ra@3TYJJFSjG=~Shhpm54K;d@>4W^LxeVyE01DoPIK0G2k< ze@*p`oQpJjQMBWKzOURQFR3n$30Gz5jbex?A1^vJw0sB1k)h79bF98JC+az+ZLlKgDcYjAbBtJ$=X8@mr*BWgGfGY@I`lD8QP9+q`Ytwr$(C zZQHipecQHe+qP|M?o9HMS-e_QQh#bwNu?Hb&VN49I%GNPJ-UDcX2=FxB!lo>&;6>C ze>n$;p7P2YnM=(JLyuRR{Avl$*KMA08lzRG_b!|ErQVJ{Oj2hd9x4!y6@=#a;sgbr>FzALE5uPP2#P;5Q!p zbjj34r+kr6HJROzgch|AuuXPKB@~viM@j_KJwcBi;%^rU^8x~B?jUE6EYxz9(+2)p zfiY%Aom6ei^r6H<1PV<`0=k|;>2=c1eox<8)n=U)J}P-$xAiJ`H4VyQ-c%a^-E5cL zBHZDWn9!$F0}7*Zcyx1Q@D;W5UH)x$`MIYBu2W$+&JZR;jTB$AL=a5j+DYz-%ll0LCmCRay1xOTmB3ra?W@-7YB{f zq$AAp3dqR@1t?E|9M8h%g8M^Mn zn)I}R^gubJJc6#iwgz>v5b5nyCUJQ>|TV6-U!57Pd zc9@ddEh!v5!c=duFu+{G>(}%888{615Mk^lBuy`Vr>xtlk$wc42eWD9pBq*5xzUgO z+i%Kf1fpD3Aw4rUgyasOi?|^?w)Pow+t1QQ)iFt$Hx9v`pF@d8LTzwhpGja9GS$^# zN-gHC=8Q@Ndvzd`D)z5z&m4o=PXt#W8H9f}o-c7& zS=&^sc@=6BKI?ChaDEdXjerUXjU=X2D&-im&9;54cy-81uDz}v5`cF8uu>|u*LKT(_6@t)u_CrR zXt2Ox*B|%2jG0D8T0tT9bltTTpd?ce#`CXbhQGVwa-fZ1J%o(4NH`jf71-60gqY?G zY%G{;E~wwPy9tqYlF9?kYl?aM_>WC`5e1Je5M9X=Q{E=p(%Gi>;Y+uH z+@^N6?R)aMkT9h36_o;Nwa80W&&@2C<*cqW)I*n}{%t0#4+0Jw%T(9R%Oa|}2#|jH zGEcF*uDdKEzR3<7?@mTV;(?1eD&vM!RV^tc$AD*tdf$~!#Ed&p7836mE{522ehWZy zq09c(`n2^SiTPNaJSe@Umyw6)Y-8IykQ4!r8MmLBk#j?RqI`J`{co6 zJ|d5j7!!&T`}1U0pHi%qevyehX1)#72{(sj$EYEtb1SdLbl{>v>h{W>b0{<6N!>L- zj{{(JQ5^r?3>;>^!T^us!Eh1x?Qc4UZq7I<31S;FBnlk_3;!Y=bxOh_9VthF&{om#ff#4VS_o8!9UNd3A74Cbl)ngGwHUJ*0r zS4Y+4Npg%Vi4h3vnh4Te;G)W_I3`t_+}osiyg9|boZfZrd!cvIP-Z%Trq(0n{v{T+ zwD$3jSglSGU6UH{zlh&W++Q9Znbw;+uaEP?X(>{-J8w&t_H{GN_VnDRI1Fe=A2*TA z9M%YkQ|LYLTQx0^&;}AQn0x<>Sn&7sUE@R-|KgAv4=3;|rtO7Zxz$44MwFPEqBtIW zaY!||Of2GP4bUAs6dEc|-vaDR2D9Ry0C@_ym+NeXilgivRfICJ3+^LeWI4Sewxg*W zSt?K>RZZ@=V20y`$vGvNIoiTuC9vAyRtNMt+?BVbHfc?>k9ro{3GLb-^-@(f_Nas>*UAF+Z~|!0<{s{xHAwxLE%2E zRbrK})3?NlYFN3Rv+?L{r2^9iXC-&kv7==HZy|Ov6B4@Vy(DK>k#QMC*pR7OGB1H* z5>i*e+CPEr7PXIOOq!+`dEH@J*(qwbjg3cN=&GFXJAV!02>+ZpLhn)cY*H*;)w~z| z?33R-F3KFS*6ULxh5`7rikg0+Dg#>4IO-zmUWZc>hdY{~>zX(?5ddbFcr9*o50@0}Uur#QOa7`H`Ic(@zX3KvI&QHm)=g zS$vH&3uzMYc|RD{j8q%EWLo?XD+UfxS>obRTkn_G){C^{Se1ur(XE|E7o!W~HAcPV zId5LteoNmzhn=FHVuYNGB`1)?FYR|fUdq2;b^ENJt}3Ny+7AeVBKXl3Fjml!6a*EP!1 zA&pv0a+O)~9=iAXNs2E>s7J&>p+#azhYVxY=2@*3CjN1tc?YqzD6Nsig~qy&m`AcD zxn3_Voh(!-M6YG_6}fn|4&CcH6pNF>rfeL{dyd4Q8$5Yx9$H9RERIYbQ{kxM?aA?6 zq}%>+l{reHQIG#Q8?a|7r{>B}l}#sY_v(?oyZQl{r8$h^;S>N~lp13qsXJ%#%`yGD z2Ay@YbJ9-SPk*sh@eTh{#sld>YKd(GYxYp=m9E}9_v(sz-H2Yr0p}AC9!AutJEoX8 zeW^_B#RaBFZlZnlxf;)Tx|-Sd${Ka5ic7Fc+r;hFJwoQ0CVv>Wp?D-!PHX4d(#Ju{ zi6R_-H+MsT#wNbAMN=aP|Gb`jG)!v`tJ!&7Eh}yinGuRKZY(+t?U;f`3@h)$!QB}N zJuj;U4v^NjWWGb2S=)4}E&HX(_ivA2)UoveedPy1Oy)-Eh3qifxd_WFciS2!YB(qQ zD}jhoR9>?vO;|>?&mTWg|9F>t32zsQQUbXVq|yAC3`J)ZH!eT#Z|avvp?dh|d1~dV0=`)YkMTG{LtQdTI{}@}PE<*D zGd($EUZ7*U{E~fLq#d1)64&eYg5P+Q@sv#9kQwFIEiZzYM7t}{TI!|M%dB?S7Y1mR z0AAFP&LgSLquvhqlmwm-+qYPi6tEi)SQ|~te;w^%tHdEw#V4eZ4wB4gA^GRoQ|vSc zVfszuPcj6OU~z2_3~~~DDm>-p=O_ox+gBIA3b`dpg5X6Os%B)ByxlC%N?VNIhw!!W zZOzpO#TuPUZ7PvMVvI9l<1jP1Jrs0*!pv#rnt9jT^M&Q$xj!X!0^gCbZNN|GkCd5( z(p>-b0Qk)8G}rFI^Z((4|9t*S!WY|7TX*2k?N1)VV_0_xJkUnZ8C={P2) z0Fyh0{>A-D!v}#?(sz&%zylUPEzc?bc>H3*Y3)v34*7r0w4`Q}xxD5R=}sm?K?Iw; zX}W@h*ip3xw14ai3LivX_TXm%l3*;C&UxEJP2>f?k#t|zXfZWczF#^qsd56yyr#U} z;37Tm_eiC?8Ds6r%pD6KAs}XO=6HRjDN4A+6&WHKH-&Uqf@aThwmnGe4e9V#hM_yi zW^Z~aZ*?NlQhwNX^q(gufn+F3pxb4>S7gu4Hr(GN8xar(b=hn_08;{2-%P}Vo(z4l z#orJl6&qI8Dp*ZPl9wX7|2W0q497B|Is~31k7H|m+B6gjhnQc)Y^I=P@M)hsu04Ey zCbCWpcmanH>q#>7pH|3vTsY^imM{nGH@n%_PzUzD6$j#=%c>!KnpJiiTkWSe3m_#syt&vr@&W zOnfXJF&E5cD5JtsX6^1YzQHGv*-kdIx_yI+EqkhFNESvie4^l&>c+*p#|g@lqo4y>CZkjV7O_y5| zj-H6bZc$HQ=|ohUAUCmRU*Bag4OTDTfak}5FrwPQT6sM-ZNirITyGz0i3YowaqkTk!*}m$`s9yYCZbJ6^>HMXMR~ZkoE48sMuZGJYdYt+ zZzH*pLG&(oS+7bZqTJ@3t!q~hNV+CM${*d?9hsIV18G;V$18+JBGJcSBOJ?xym{^N zHn-_;-p)KdZbF_R242kuimn3#VNw_Tr-byJ9o>fKph~k>eM<)heQnj0*yq(Eq zRjC$Cuu-#^v;3$guYhg{r+n}cKg^q{k4~uBE6^7IbYW5@)<~NI>?m4J!bFf^!xyPJhHt(s2CVsDVgt<#Xi@g!cx1F z;~J5Ht~n)lJ(^ zGwreH1RH^8KwnW#b9~$6x=L-yMQa4p{g@c3i8rMPcN$l|_NdfAdJY>XXxyA8W*vPn zaqKSCYaqt)z(F;c)6%6o(d(GkbmfmKLg_AIlap3`wv0*4=221NyXKI%kLrHrfBO0>0`|L---7WAtv-aA;z3DHv^aE^(h ztE%xqjxMyx?M4mlZc9CMWri%|SJXDml$up?NHB^|7nWF7*ygEtHHwuLahswwX3x3m zhuU5)JB!Jt+g`IfV1nC*tiXXf`Zj(`=YC_CHWHSK>;G-q)y&9P_&!!8%;?H?c}taF z$$;bQqMk5oO01Gj3f-N&G@2=7XA4kvsdSUrYkyko-b~!^({#(o49?{5qcKi z1s5?VqN|^Clh|%d1+QR>eMog1bF-Do?C`#H^Px{IsYOY=MA1EOEmS&D|FyR^IaQRB zeP@ImhAKK}OWyK%C9(3qHV6D|`3+9hd|Q|!%cx@f)>Q~Pg z7)WdL%XV;MF+>OLkK;24<;2%07@^=hQ6s1)fI9~}J4hP{Aio;G+~`E?w;K?S9@?Is zj>F>w7Uzz91k z`fsxV^gus2Qqvd%+W7waaPxXjcaG`;z79WrC7*hbTI>PXKEFnQ1~sMNM^MT+1g!u+y(J%gy>eiQ?}68ffAIkUI;Zf6A@tL}@eWP_>L$MhvH@J* z`NE<5YVN=tfY(rd@o@n<7w|yJMc&~8mkJ)h1uPZ5gYcWnz6Ihnmwv-Q@SDxQ1mdqW z{osi^U+_T2YaYP`8ZLkSl|28q^z$2!52EpJhbHwezBrBQuTU3!PD&$@ar~?b`mf@vVD&((gZs0olLxOS^|o-%i}U31oBM zwT-`cPHn!>IXwO=sj4hZDquSy=} zG^0AO|KhT5)0#&j(WTKk+UBnNoY&hcys(={^02Y|Ck3^=bB{J!H~uPfJ)%czkoSnv z%fgYq@6?|D`<;|X6T+{)2WDr)G3j_3q~+{8zcZ3-HP+FrG=YO{Z3FeUNW*IGQ_(Op z!GHMK+KbgoPdB)x{i9^!1|*^9J{$(dl$sSYL7v|q3x>2yhq1F;vqII;vqUIig_Io;HiOz zJGg(rM_8*!Ly9U_hJ<45<1)Z)FZCcy9u3qM@Fab^sc}f_u+9YfJf^-2b_7#5O$uSA zoXgsCx2U5Wh)883D2jShsZ8>?ebyBMV_hZ;>*kZ1fO=RON@{xF(i=tNdY|pel7?^S z)Q=r@W(e9PPol@YCYPDxnhC$uMid9anRQPSWlwJs+v<@)u)8uLhy&@2`3D{Lh^o$K znM36xrh%W7$(@aJQJoycw}r?RLw7WR#12b{a-~_^*5;BYRj7m6FyAocCAm$fgG*b~ z#BxGsEO&eDCz0M!pgUIZGl8=JclHzI*MbakA~8g;xRn_^5degeh5;R8KS$;)heutf zC~>?9{|bx}gz{=uFxdbuT$MO8Pgz1I1YJ{~epin$o`$cUwX&(XU*}u;T zInrNK;H32`3jSgSyEbqo#wJ(PR>NYnkt`5UoUH!*T3_p^Z3gy`*p&OHlQv+Gj;#2(V|J6uLx}1;hnFrC*+Zi| zdlm~+h-D5m1nqiBdFu<(U6nWSgY+2^#w4;Plf=b%r59hItA=xVqyo0Z>c};a#6d3d zjc9(2I|Y3??IQv4tW`fgoJG4XQd79Z7FesTIGnPw)+hnLYr&tZPxSi81)w-J*ct~jO z5Na)k_RgylX@O3H+;L3Ll49^{u;cTRQ%NyBieaEl58SDp>7lG6D3d!pG%1|ispNJ3 zMp9BWP6OLa+>cELsl$s7EFS$!k?YVX66b}tRYcLE%eFGc-p<)kZ|W#K3syaiYi2i+ zW?aF2+Ma(_4wlG=yReO&C;^ijpLiVpQduDbq)y%*kH1Q@7?>ka5=;9b?83Wc>qQh# zu!+m;gYOm+Xdxa1iDE6;Cqo|T&Iao~g@K65HK-9t3O|A|S!eZa7z%=I7^@X~wdfT= zQgU4T?vk-8ktg-5k-U#|O(Hum`^IFLD3*>c@hRVbU^c5YS8zEAa5P?fRRk?Oc;gLD zpjNbd?ynQXXp5d6w&$g;`g2Y!8DO3iB&vl;~TzjUUoI)X*1<85)ioy6(Jwh5KHgm#tetwuChO#bQUcGHe1}x&DFn;-qD!qm0qk}3!bu~*$iQKG?q)(WE zjMwA0t6U5%*oorf_N_cbirm0(dJXFLFc1bsgoqdd-m^IbgS?uRs}IYI9vLn1*+d*~ z|E>Y;vItW`IyAlwlZ-jb1T<_AZ~Q4>f#jRbUWf3pPr^g00e|_FB&hR`m-+?K8hgav zni5_IJj}WGOr)avYheYCD~Xe$FT6v&rN-_<@dM>2L^Nw(50>ORkKB1YG8Y;r%(9e2 z^#~I&NM>o023%kSB-{c^bGlLDC2EvkxT5ai!fRvS*)Q%Oigr|*4sKqy{cb3(xOk<% zIprDRh$-vI!LBKZQ2U}^YJh&);Q*(ysAtr^Wj7v-jOH{Qi6etb^Ll>)oKPtR*&67@ zYo*l0i8QK}wClM}3W~-H=3Lk`UF}?xiVp&n8RyM6mgR;H%_mqRw(G$gop|=1nMEC) zQ?mK~jxmy(#Om9nAWUP=gG|~5zzuuNZ(Rr;DURhS%2f6N&u)smV;(JWCB9=%eW0RC zKkJ)URXH{g6PB7H4Fz%i&0Az#e6VJfJx;ivJSmKGHWit)qD|o`dBdRZb9d88;)2e9H6Jbo*wY=}*_K2CulmK{A`p&xMOH85+a!&gzr3bM-LApX zx-6TMrOi+f_Z^k~Y5WaAg`Zg(gMZEhmgtkrs*9bVV|BV#fWuS8hdP^E|0;4mXPd5G ze|g!E_5uD{`3F%;8FjF}lHr)lJIICnAmY&jNz1HS>9!BVe3&3!N+Sy}4r8MEhyd@f zRljiq_!)2154Kqgr z7vu-Cr?FrC?HaQFdmhu;eiEM6xm+%|6DS zBku?Esa(Gu$=^9Vkh1|&M#w&^W?u?5v6Wk5+)yuuM8D_jD1N$O2eVm9mF=%+b`%pj zyxkxOqTJOd=5qxRQgYGShs8;R&^2H#QWrWMZ^yb-aN|a=lwgaP80soi`Pi!M3GYvb zU!HX=%9DQ1jpeNmq&ra3m6mWx74Tw-6Ldm<3m?;|bpf(i)lNh^rMK}`4w1+C4YF(x}Y;F6?05lJI zJUOWXQ^U+Nk5DsyI03)?6yh=R>GqLd2J9J8c8WF#0-Kk8wz6{z!X{xeB7OuK7h@K) zFnu0J2H*Jj7c;YhPx#f)HMbxm$bd6)h_zp{yuq|S)|=iq-BFA!bdPS2luZ-u$0GPK zImO>~gH5M{v0h6AZroP1@#>R0Wxspr&XOdXK8p&9XY$_qw&?IgWj6pLs>bgp5auwf zeDSEX$-_-ypEdc0-px;L~@1n34j9A=O#;zQFD3uok zneoprIMnecSv;xDby=X$9jL@zJIoc-Yr4|dL6DNhz#`n+T{1le20Tr-D)2CLc=Hgr zB+4E}K~A_Bitia3GPR19*DFNA>9KKVx1wS-%`n4TVVk>~k0`>7JbbpCyYw{7dRgR^ zIpnbrZ@_66+}#|F^eD^z3a7a`1?njO9#I+i#qUFZnw2x_bK`_u$6wYbE#oUo7>8r&>@HXIH1fN8gi8lgx*oY990-WisA!=q}w#1pc| z%yn5X80FWjnq(|4rm%{=pNvYf4?q$4>T;?sebOtH1+-a}r@U2s6xbMos^Bfiw|OEx4__-w?+b%* zTEKaNh#_4n+f5$30vS7keVHhNDGlRBv(2B~|G66Nfi5#l3O)gepdaIZb88kAA$&O7 zgfkf^#~fmgwP0}e$x>e1@f2JVkoBQtdG?tO;+35{&r`79xr47XWjZ!?q$2w@&70BX zLJ=YUxV2Q5z2^}}nui{x9N<6C(eTGOu+LI|mizu8_Zj#?J?u82V3LYy_HGHR_Dy~d zX+4*44v05EwBbNE2uQXQ!bD*E$Qg2v zB4qIPpN$Y%H=1`AuHgIbTqj~1!)}IbEUIMu)t`DCmod1oxM}3}kx*Lr7lDcMY^2gT z#WpMJmN3q_1zzp>Ei(#DoH)W<2^B1cNAU(t7emOIVlD*$xv6T^L>LTc545^1u$?#E3 z*Ad=1_$_;P27Wr_05IK7wcIhHytIaEnredC&h&T{HgDFR?(RmnaxN1Qw8}k6r(!9K z87&UhS-+YW;a7v|4CVs^h0Zc~Qt#A~MC3`5PHbbyk9UqZ^%&}I@4k;+E;uF_nWvfY z`)!Yd>JI&Jf**1rzbAFKM{j03v(t&93i8X{JiX}jwPa}kcd4eAPF-dw`$Zr4_o|&* z`961<%@PBlNV$w_E_411A@#c0`CzQ=62#@O0j`o;^BS`)DGKxm1nC9CK?>?eLvj;S3b-H8B)adV6O1vg1r@r=90-v=R+}9|xZMYED^qy0-)w z6*aqr>O>cBN9K^?SgCW#_UCXO^vl(IH_*vfB5#LYWkJ*rZWb9I?A9hETE8mobaM1^ zBVFc4Mrrsk1gZ zN|D)5E@4qxTZ`#))^BUd$q3^5(z*^}QQayoE8Us_ZndMHB1KK(v%9}rMv#r3EV09i z-U(g6;4)53PQ4G4uQWGXntQygl#j=i;d9}1LlXDYBCdz!*Eo3E-==Wfvh{n$h>E^r z`CQrV+*Sp7N(V3rmbj7}hV)H7rxF{ep9ILJXbl%C$DX@Bg>Qvp#t7gR)QP&J$yG0a%E|_Q(t_EQMX-$jOS(0dg41UR`tg01fjG2f>LfBwL;X z8xK2ug(iRU+Ryr=u{kEnlHgyr6$1?_7>P=1bEeTY4-`6&53pb)qUV-Ps6JS2pmcy# z+k)9F9`rgu!IF0BUX>F3t`$j?Y9d|ER6+WP&EVbI1Ehn<>5=j?J5>3TbD37M|QPOZ~jfT%gf3g_>k9&tF=cIW??SRguiQ6PdZHDOs^Dmp9nWwAgI% z2uF|8N6)f8OWKIkykC1pno%sFcYV&1KcVLCf#4LYlUgIC%~5g)p2q&f4Y8VO-$t-0 z<$XL(D4XFCMwwGNVVy{ZoxZ~BWH%3UB@Mi4yXwxX2N15Xia=egGe?r*P;HzuSvK92 z0H5iG!hA#JftBlvQt$)YoG;N6^4Y)TAlBJ;f;Hjp${G%cAZ{|DHnVHLD$4mjXYFGm zThNxr%VS`iY)7T|5Rv0_Y=z?WJHm|4cIO#sxLDvX3Ex+6-9%6QA?VKH4wc4vl)*7w zSzTTIgVbFXIn&%2tK}S`+b-L5Gl7nljQ#n;xk8_ngjO5?NmpI~AmJ|J6M` z$)IU*jI>Dgr(E~83P?ydou9hd}uW{|V>Y%t- zx7g1;S>mPCn+vg#nQENLHR-z4n+%c2UMIrWzEcs9H7i*l3w zRg(MfTZ;XFFyBl*cYW_c$VxVf5$~ME*QLfUrB73O_WwF48}imQAVD5Pw@c1=J&_f} z@|&4gi9L0}KP4#UfTK^*%G4zMa#!V1-;5DU5~LA4^RngDCpyU>emG^F%_v7BJfG8f zGuHG4zivu29`$|T?cztW&HY0eHLBxRUyv|_5C{IaNWOcC@bE{C3Nw&+$R$c zZMupUEWz--5yC}@Ep(Xj3_&pA4PxmX3cmm`C%}%4i=h_3+p6X7*VXhjJnGO@uP1)2 zt^OS0n0MKBT)fM^?K;pyKulMC-45}X2koI~MYUox9?3gnJK4nwb(*QiJ*= zEFui123SRl=xr-?1Yph)Tr9udUDrh%jk!olc{)cc%bi~+2*>Zo4&Y65oqzVz7E^$i z{6yGdbscz#kb@8}J;9Rd57)@@OcE+VYH&!>!y7VmO+bAg*HPCZeY|v%@@B(R$``V| z=|>>=u=_JMI8^;!?!6^Ecm+@TVr|;MuT+9M6iJf%p?0k3RVRk3Ut*3d6#3(NY3uMHBdDXFbZk}*fC$)1u*JhykM z@pLVLpbRxgp>|L|bys35LfOAlWA}BHHH)4N;?Ue!ai|6U*-kl&TlLF#G2|gCq;&6cjh#4u1`I6Y(={$xKC7iV&xnuK?dySl`$JKi z3LzdMIJuupTqIf3FJ1HxKQaG1qQhOvdke#qs8ARU-`wc-xN#Y-V02DleT4JgN1YWV zl^bKH&;m4VeWu$ZjwU$EVEFtQN#nv0DODPF?F*I7)%|pB?3-1@HW?D6jZkvVLx$i> zxUmGV6wRurukG{^+*UCPo%JM9x*c0Eq@Q=>BIzOzD)z{-LB@;4jmK!HJ#;R>_Ca`3 z?rL<9g6VGH*p6=mji-Y0TNU>yEjM-7t%|j^y0vdY2o8Is8;;EyWY%?KZCu}$3?&2e(F{`mt zz_e#xc7&H78AbX`!e;D*Y;}NmrQ{}}GqtinlWNmI#$v!0#-#|Z8KTW_l!|dei@`+2*nlJEH^afgirQ<+w$u!)&!FB-f$^9a=TCsOg zW?ESTLksHN7wlIx#(9qtp%;D(7>AxVF)5ofcCtme}+nsvjiT#;ui{8a{Om zpdg8nX8)>#oMy62y3k2AqV_Ntp_C_LXh)>`? z>pwJg>_4e;dnD8{)bjxlM_y(#Tejs>+?t6@bs?#f zZSX<Q49sM=lyG`axs&zd~q6oavB!kDtS^n zH+Fg4b$G~fvn|Fz+Rhu)TQ*^(&dXtm*wD--8#rr6FG2{$$o3a?C#8LPFOfpoB3-{QS2judXv^Ry&XuiLii&sqnE2Wz2?@B4ytUO!UCwAUKd;SMS? zwwyt?w|N>vT8x^fMuwimdTN&F+qx7g%k{bZ>gpA6Dsci0(Mtyu6;ZEF4cP^)xZNmC zuXfw~-f~vuQ;+TG8oU7CsbX)kI-w|6!;s5JY+QN}`I0#;wNCkvbhE~ucRFpyQ9H=H z%Ih-m7+OgDKRTE!%_wurgV)#b?ZrbP7ESI3mVuw>8#2qrHO@<+I8vVF!49J=TWN=z%dd-0N2(ESr_ zo{zdV5gBt*pmathE2kw3+Zvd1wu*kRbNf7QQ=ypLVfTyTgp#VN8%K#rk);oyIX;(Z z=oskisDcq?tFQRvDj_pkFwvgjYhT|WhVNtHQvb^EVMU!6F)&=dmhGLkXh*`U{t-v! z2N*qh+bMTRNEXEAV=j#A7uu^HSJQ~q7@RpW2@g4i?DP^0q~}Fzx7{AbLUG5)>r-+O zBC;cmLH_b9lY;1jpxAKuVJ@jyNxbiR2NZ1GR685S6X9r|GP9VCY?XrssW%mw@b(QV zan@?aY!Hn!GIoWwkEsCzM6)Tg0FNO(x*{ay?P9gn6ml^AU2YK)MvNYHFkBN@>KDnkWn}Z^8>>ofzJ$ z8yVzuoKZc`mrU61WRj%z7Z^%bcsw{p?+mZ-kpo!k?6|oa_M2E+6@2&Hx>B!{HrdDa zq$N-(!TED`Ez2h+>y$uKauy+Be0*_vy=z}q_2O#6qLSmA0uB$-ZfNK6*{cRtpzZGp zeph{e@aRQ*DE?CWaIm?RnFH&spX7-lR9fbcXZ(a+@j13?_aAaa#%URI(~*Q8{%zec zeyPI6jcNc;T*Dsep`VjjCBJaEDW!(!=cAd(Gh_d28T0w2v-~Q44j9u4$oplOsvQn2T18|_ z1Mt-*h?f*bG+A&e;WCCd=If#}DoSqbN~gjUherRz)YIGdKlam>o<^l(s~n6eJxhOi z`HF2qtX^HeK69T4RLdsWb=^Pz!nusb0FdV*7e9c12X?0S7>JmPphy%>`!Ivcq#;+O z7VJ-XB>bJXF{Ux~u*y}H{B_3#KoT}ojTcGu~w z?fqr97_ArZ-*~*sMc8d|GFT-;I{nIEoDU9|nP6zEkK&mNb+^;B{5!$?Y`AInss2E( z10bh_&2^k;Vu%GAeL%!m$-Hk&IC5IcV14%`DmAiJoCKl0xHi_v=K5@Wn)O|A$GP_> zM@;*&lN!-ZO59-v9~6Lzl8}7w86#3Z(s%i};8k!CYb%`KMq_$Ir4qGs$Sc=fI~l`? zOh6$kKBJwp{nBStiU^ET2Usj6I|b)%>j2G+E8V*BRk|75B6FXMzj?e_m{YAUdwG?7 zGtcnk#CRqj+zc;hI11A`ok!60PhS~RVHB6XZ5P(2t!T?--{ccOCeIqw(4g!(15;;*}21 z546k4P)2VH5W;@a^cYW4Z>$L5)&;P?WZ+*EZ09fzd#jzVLF%&BJ4u}0Q< zJ|)g0w4MwgOvUsreOos5xze>OL)3+Ym?S9; zDM2cA$GcD@_oUs!%m!-%NV>ys5xfGb349(D0i^%*9-yk1Mjt<^d=%>rI=QJ9Wr(0} zLUG)?aThTXORrG%(z)0>teJdjSjlW!18F!>XRC}$_L49;2{g&@b!LLL`KYGtog5b^ zGd>b$-!mYr(NEReL#PY3ws&vRdMrA?b-he*_Zd%Xfe{kO$igS413o|R%agcW zjT~hC!)5vtpKRW_Zw&`!cZF{u$^M^|HE4@}ZU|qBqZ{Ym+sRJ!f25bzO_nBc&=y#& zei530UR8-}WYv1;5~gA7eCLItO6RNN@lTL}(t??QGlh{}*B&#x!G|Fu1d@T1^=GTu zbrVKlUwpV;N4|dNnEhO{i#qeKknN4C8b>G5Jp(+J)VcB2)M0P|@tEBJqs|9&38;I65=*lE!nCo|S$;d~q6+%h`1<`NhoC3Os)wPwqZ8(jj`7me1 zXX~ta1bPEwEmHsu>yx=ntmM%6$=B1jq3|X13yw{6mr^RDUeeIw{DFhMuSRI(#7?pU z`C?f8O|8-8>Z?!Sq6ifV5Ydrm#K`lbi&P}kRJLxQl_P?E(_pOOMY0BN<4>+zQ{!wYm0FhQX+HvpZ&^PVki&|l|~aY`)&rq96!{ONHg zbi)D&4?OI`OI)ZK4(PL>bwk^NxF2I=a$O1o%si##Yoz5V$4e?Bah$+78$G=En5->0 zJT7CYlFJhgNM4v|vK|ZMdKx#^*ASUU#W-ijSlj<$>>idyQKD>tZriqP+qQAGZQHhO z+qP}nwr$nzRK#mk<2CaKG9z-WIc6#i(j>;^;mwX75!dTl74$FHf`)_S*1JNNJIMdnHAOzY)EsW&$kdKl+UX`FzBv#!-_$I2(MRahJVq*28B}NvQn*@yau}KHT4tG@@<#hMioh} zk~_c{2af@CrL2!MgN+6A1qaU9XuLA&3WKev;zuu4Z9$12{0j15%OcZ_VT|xXi@oC4 zh7a2WZmk|8r6!!af$P<~V`1WRGS5w6^Y@=J`gsAj!Q@tJECVurR`a+tYghd_y1h%M$dv^lwH-tb( z9BWh~G`ZGaI0dWqR@E`)v(VYUg0A}kfY#Jr`b%R5=xRwnrZ(mW)Ie-@$!%ggoE)o3 zsOiOc-ZOh$?VJT$C!{0~)~D_?@GPR#>V{&F9!6}@C4Q#+0FB?Qn z_kML}&{+`^b|M8mIu!T}$A?}MH3W>B z-ZFzuuVqDK!xf!7=-JV`v+n_2MiWT0N<%+ur#DU1b;Uav98IQI%4)EtMJ1L0<7u~R z*SHB*@wZbUs_tlj;=)AV6MSE+(%@x4&i&KwqPWOTuwLH40l=xyYjqv=SY+4Ngqqs& zeIK89<)+>~gEmYO{)wQq&z3FH3jEg@$8)THHr0&E_*!Nj{VGdIAAU?tq?CLB1oUK` zAEw0NetF}3@_7GJJgmu^UK^Kc^|fR!ipDY+Cd4{2lBw+#%wv1~AegBdbP9SSJsuI2 zfx*+ZFhCgLb-yQ;Vzcu#( zV*_{mo)z3W$z!ca0VOBV+(ib|{k-Dh+gR$(R7oiTTv|U-Kb!b8879P>VRkCK68B|g zP02gPjkGogQ=8qr+F1B6r^A_%HkpA-&a|bWQuDpO!f8P5@~Ak*l3L}sh>n!FTI){! zK#x4?D)M{?zJFGW%v#Q1AYMX5yUpS|VTXZ_0<9s0u|+`$c3r&ST6TsHkMidQDa5_a z1X=);013?C1EQk~bvht5@Q?MU?;Lp)e856?Sn?o4;jV6FI*QPk(>Mo2Gqgn930V>! z;@XBz`_yyX`u*o@eJKzhs*gYe+%i;bOI23!g{W#^U|1m?KMi*}UtVkB@)u_wJ4Wv2B4L$4~Me}Uumq(Hnng~7(Y-Cv#=;I}IPba-GIzSP4 zNb7SR-VGWjzR(td%VV~BwY)qiFM2mlSkTuzoI>}bZ|#CeYw7^}MJ)$umF*GrOU+$~ z7vuM*no7pMhyA!G;8}rtmjr`*K0|n@LPMRj7iSZ2QYEjlUP}S4C}1u6NG`$06Cmgb zvyHekLs$VOM4_?nsAJLW?Jhib&I|4BODxzM;ar4u8Mu7rXPh6hOsHV&%~Z-4+= z6mo*ER;xPmkm~_8ejS%<7TJUNVa13f4B0D;$o!+U9Prvv-;IiRu56Wmb%_l=8#H0Z zRp{-wy3CA>p!1GNVwTB2lvB+AWpmQnM}1H-35-^!lxXJZ%PAW`Fw-VO#95@4_Q{2_ESxHBvDc^9bb3mVLnB(4r%VcE2AR*|U!6lr4 zU^BYMH$9JyH7A=Hi$Wu}NY*y%7HrH~)U*!mpyxUb?zqyR5aK?4+%w=mHkiTRLwIpH zn_sj?RukSgZ@JV(gYjuTAF<^i*Pk5rNP4V&F$`HguG1JVU1c#^h&>hukz+2G()v;9 z5qrS9ViYZ6s#)w!udZ6O6L1hd0l}{Gcs1AOrM8>-QeBH$su6>V&dg0C6JJ&AwL1s! z*(a`iJ^4JXC;`oVhl=khFE8<^`Ol{-mu59Ofs0Tha$Jus$!%Y)o%dn}6^iMATY-h9Cx!SM%EerL{2^@{_)}&N0{k3$PGLyM4DmV#F6I@CdK!F=Kdk$zE!< z3z&v9`CE&+4rXt`@ui+|wN#i!dqs@*O^_!g$CX&?^Ec{=yGCMfj-{g5w- z7y}|q)c@@%NsAdgfW}V#*cRwbsp;5;63RKdd_wu|q<_qxqxGGxxYc|@K4UctO&H)z zAu=kjYM#(Vjo%w#I(gYY0605^5{gOFzwtCBP^D1}>Ho4WMh_+WY7(MtY||K5rIQ%i~~bv9{0 zYfM}s@VTO_=~d;AVkGS%x2aW~znL2?0nSwmDXRQ|hk|)BNCCOJc0WnEg@dmYxX8j| zJ6mn1p=z6$)vFeaS_39B)-B7{hWK1<)S4==6=nJbnbXyYCFx2j*o|$z)?p$2p~YqQ zO{>|Czgr<*oJ^}dS11;x*&8q7|BPUpo9fg)cn4|`=an*3R%sZB3SpwE+Cy|_z^1Yn zB2}5eqWC%XY1I57A9sGtzL3rH1!U)Z3Fn-82Il!H;S`a$HU}PCsoyhW*Nn}1hx<)0 z7S~k=OWKRG_8PFwtt!Z;uj~`+woxCCKP9Vx1YnRW?tpLDC!85`#*~r&>EA;88*uue z4ke@o%wX9mAK>*5FRO*n!48-qQe^F5Tu3EpDR^|LT#W5g!H>TNX)UITdk7{vX;52I zFp(?Fg^Lqn4rT7-pONRbVmunV2ixH_L?QxB0!mHu+Dlcazw%J*H=okPeEsD!H2kRS zMtENSd<$?pPi}zz^1fU+gQ}~oMwK~)i#0LAM3RKs_>k)Yd(3Ky=`|)HWFJVXm>)Nk zv}KF$U10E}5TNrtL0fpDR!Dn4-E({r>H ze{h`bVnU0hFN)#FEMw)#dAi`K zs|_I>R%avr$&j7qoDIZf;JnEO(Up8mbk6g}D)-M_sGh!gPu6W?6VQKd2A$i+a2`j8 z{S_(Yq)QIWKQ^B%P@5SfBZvpH>HZA3;`>Fob=YA1@&t`0anDSMF289p1IoBjE3P*&nll(TX{|DCd5 zP=`K|z7i;B2(<}PAVV$VI$w;F_;t{n1t5l?#e`q@M z!CN++j)1wS%f-?wpu59Oz?W^tzQLyeK)~b zG?yAZ>zv#S#pRc&e4G z5bf04np~2QozL*Yw1;)nnZ|9OweF)ycSJy2N_e^LNcv{c9+j8>3_hf&v>MYZ@ zCO8}n>-7o!;$?8FD^XQa>6UdAP*W`=4nCOzKV@oxZ%H9(jM%kh-v{~=gOezcG1$@c zXZZCJgTzkhB-ZpxkuIqY3eP;KTk4EMj3!Q6ZbWyt>vA7K&;jJIAA#7~G1hvij-_kX z3sqz^ioI@yss;ZC?B^ss!xU7Ddyp zyYmRp+}g%HTMtaiU4lxKWBMalJNG4Oqj+CJtW%D>bq6-I`i~s7e??<8KXl??$T?&-@FHAKHD%$Us$7 zOS9;TkSH%uEQK#5y`1|uM@@0KO-;Frg=uU3IaM+hT8yVpe7Q1?-6*rzii+e;c`D3E zKSuMpui|N4n5;$E5@Yz3KiKl%X*J=%=h4o1#r+-tp?C1W=<9z}r3YohzS2vy2J2oO z>v=c>rMjls)wCtS*En!IH8)^AYv#kTw7^;c{;&Bja75!&Wwu2yAY_Vx7&(D6TAb0J z&d#zQN7_+JuJq6aLUFnB;nBS?VWH$m*S$jhs%Pe(9mTdHzO9rw7&|?(nPItLZzXly z;<6pF(=P*>(-o zqcG;4Mu;u31i=wu4R^R8(O+I}R{HG|b;) zlK??+bYK$w52$ca25SVxGRG*J)#(mzsFwDyUg2swqe}SR44=Kk)$GHh<9%GRb>r#5 zqeugm62H2V=CejumSO&IE@ zHt#A}?=my~V_LbJkl@Wb)zVe>!$(s8ZuzR|JNdfG&e$T|EUJb0(z`2=M=kN;FgJQ$ zejm-|n_krg8;5#CW0d(%)!Xd>%vH3}R2$q6g!p4TWO^<6c-5MY7`eu2&xSM-u$XRdx`T{vcS8%uAh zdBN^Ot?vLvpFs`&-xWz(mE~%*jZF_9W_D~M>XL;ClJ4YSGp({EdXT7TRZ#ZjaZnSR zkHPhf`!&#>mCdprXV+!-0>sZ7mV8`kbuj(! zS=%|bgSH3s{MEKeGw1R$jRL;RI4D+LO7gSlI+M~uTr;?ar5rXR6C_* zlvAd0n(LPHd+?iMeGXAkBIVO0Ck>h)1IQzfX1@7SgrD@NYIFt8rlBybV|I@BDPzMo zoFWe8*W$KDRyI%pN_0)MF&*c^S8FDpvy$;N?diybi&f@UP2)FC<;?>Iv)1!M z#;-mSa85Y$3gBS3@`$|{O-0Vq0;LTgCU1tO@lZZq;D<~JzE#~mDSRuW@53Q~3H`H3a=eOM>GNMdE zP;Hmt#n$z?5`@!ez1Rj;S(41aqW+}a>T_ox_s}E$UJZ>hDTT?$y8?r=qT`w2)D7Po zRyK@ofYfzfnxLt7Gqg8EW<6)uW%!m=oiJ54;zl`$Z{j1&a217}BFnG4ERz_<0m>at6DK{3v5KT?+GlEdQSbwu`+%w6oZebhVeC0Gil6_?st)A*m0=NA?^A(m5x6_21k;;w0 zW}TtrsAOPag)Xs>_H?Jh@x*{2q}?2@W(Fp^*!frm<*7%ug)T;keSP_vEO56WWQgo> z`IdUTr9l+?=-Da4@n59rg1H^1 z^W2!Nfj8+mF$hamO3GX4g5O|pAsu~O`$Jd9!fWzzZxi>UEQ42OQ+as(&* z7+^Bl=|jop2d!`bAo?g9e9Q?6S!DAj6ZoG;UC*k2_D}Ct`(+xo3XIhEVx)GI#5Q+S z`T6p)`vB$Plx%XYW}bQLA16O$csikORa=KQ6}wn_y>}2ysf4zK`5Vx<+XX^))l1if z4Vf0VdHHkhPbVNL!?MlG2e9>@Nvh5ZbNI$$U^?2|UWMILpVty6)Py3LNb0b2P1i)N z6^rZhQBk*hA_BSyN3}Xy_-fNCW*y?W)Q9~M+RXMdU~$YeWaw8;5v1ePrpZXQym3hP z6GN>fuwogu62xfv2MdLTC$g7y9{6)F_RH_7*bSf~Z>z@%UYg*xO7?gRTZhl1fl2zz z`jwC(`qw&fiCPGgK&(cG1_#0kZmh%-$-}z?m`W|~c~C!3MDB<$&zDX@o$3T-SfH~2 z))65Rq0Kj#R^jVYlo@>DFrUlaqe?T`N;UTy`y=l7-nWE6Y(LH@f|0oVtg%PSij;NE z6(W|l9YN-5hdA$9V$p%7bIhGMV`qV4eFkWD1$j4sB7GupTog&cO}o9kc9%@ta$};6 z@z{)iEyPz8(J5-&hQaW%A`n4{X)5o`I;o;Nn$AnDY3xy_e~|9zVdp?{#E;oNmjgum zdZ@=6bvtuN@Py%R1PW%g)RNQ5NNz4@v>+THm?P_e<|u(m`PJwM)t*mssXO({%&B+Z z4^H2-b+Q~D)#;=BH?A2gftO#7tnJ!9Schf1tp}nAaTw2^Tm%^;l&zIV!!V%g^zIir zrTIKXeJ=E%QsT&E*Vi*k+}O7jO*ZP@S0<~R5B58T5^o&!lP^Y1s^3j zKrXjHN&gZn)J%+sA2j0+t(q&OIb(pi!iv`+FZD^rGk{R{uh{Md&Aj3Hgi%@)bZ#oG z8n{aH`m93E-6iXAbz-G2skFX2PQ9`isU1e*=0hB(1!w|G!z3CxEy!Q~hS2wOmeo{F z$PBDG{5nUS&VbJkLJC(M`}1MkHqEBl(b`ZZWZ{!sQA$I4n-YR14HDMcoDCu7#|{#K zld_f2Eu7655)s(5Ht+Hzks#T$ml2RIo$LA0p3kR&_7Zu}1HDR-bL1z55QQd&AD^a8 zBYo$okI|^q2uy+-mT^iYC_jT6%Q|OVp!gR-g?5#v8;BETM*`=?Km>N8;br1Y>Oh2%uL+Y-C+M2*p< zNctH7C%tq#M*b^m;|$TIc#FsUb6OG&r=|Aw`VIOU zrI4+tfsdkg_c0xrh{f*Ki!2B*`?;X4u)OVU0Bhj7x<-j-l^|JU5<5VI6qh21Xw(wC zOK_{=U~~Tp&VE=pjzy;Z_lpqqPYC2_PhvrkQr*@PzYiclR~N#>Z!_I4vs0;MdGBc( zM-rC0B9++Tb6>(k7P$uir|e7|_%_g1oxB#`WQN$J5q&OKOGvP|2C?K-*`sDI>=ZmR z9KWC;hUW}Z&ZuM8h%z+CTypx2ts%^Wha^fYvv%}c&R}P{*kT*cKTf(=JZRrSI!2NJ zYeqImzk-;aIon(Gzt8)>vpT%t&84#_{W8Foi!~C8!MYC}?(o`V9Fqp#09_}Gt0|t( zr*aK<_2k)5h@}H3YFiOqEnsgGtnxMNe+oiQZkFUiL-R>B=E7S}oETI|b;VOojrh6= z`%02Z6SlE)n_Ayg{>Hty1h%K(MKs=^RQFFjx`SBwUfvk{sWx)jNA+JU5E#zHWk9C_ zg>B$^x8TLmxiYlVr6TeLUye8hf6l?w^RAJnlE!V%YzZnTx@lB~feo4LnlUB<+-tla zKG&Ee>^Jzu<;u(A1SHL>#OORDrdttfm21=Rd8HS(FANNPha9V+L3(_j4$- z5?zLAYPk7wFDZ_I?uNt=Zg=_Mn*k{P>WjOx67V-h!hc56nW(79cgu=6XAL}xFJ-h)W`HjI!Uu?!V&x;yO&h%bG_ZM) z@x0`qVKgyQf@}!^VqVl?qgb?<_&6GTz|rJ*yO?QqD*N#?m9oWTy%SSFe1=AglFxos zQrSQcbQ>OvM2ODj*VgS4F1;7=<_VR3IzHn)+%U9ds=dXp?ma(W7 z$7swXoV{?%%;=xl2}*m=!4W!&S*eB6#&8I~pxoj=k_QfmP#2(bbwK2PeLCTg zu%q|X%Vi!_Tv3veq+GOID zK$VfiTfp>gzDoaN5jr0dnn2*~ufu!L|4_Ai)-uhb2Od%(mj6TVhhp@`T<=_bNhEh;VplHZ9{hYV}M&3ol6( zEv1b>>!SRav_f~-KYArrD<2o)+De~^s%3_nz6H0Y7~fHsUQOw_O3VwYgCJ+=J9|C$ zO%SBQ6teBc98&wVEm#mc`tJ#1KOShSY;C0b}Nc|-~10g{4WQj@;+goAn69H1L03ykpqG2W~tx@G;a~Deb9H< zGw%k9s3r_&4ngksfNMnG-QM=b6T)c1_epCF8z^`Mt8unMI8sRs_=X0_KlG-pjNqw$ z8>W~_chmqE zK(f07=nL7jMpQUDkqil}jddzTU?~h()bKZlGGZ`qR{TRAc3CoG2fw-@1}D@buDu*% zc@KTcNO_2W{)v(+nq#5WG9%t8Murw7H6W(3qK^vM#h2lEyl}`qnKw%*) zr91ih%8M6n2(1Aw6@V{C^fpnd$){S!WpN9xmlcJohspC7TuFrWmVTI;(X{)-^MKpU z^z*x^oYB=OAidpyARr=~q{kFl@pA z_Zztk&J@+Fgt04kDjy~AW5Yx$-O0FZao7OJf0=?NJ*VAN9z}cx#ue-AH;~cd-GYUr zwN9^%QJesA(wYqLZu?@F59r}UYz(VXkb|!n_>D+1~(^GE5#uAX0Me1k20+j4+h}~1;{?NRjv~3mk zLKxWVuf6LuxML-V@R-5I!l)8QCn2G$YJFBF!mEx($U&tgoO6^ZOnS z+_;$ZWB7_OR%jATtdbxs$|f^9XXclA^bfe`Lw#H0u7v?|3ZKNVjiil&JI2j5~yhWwF6@ z#q0oQgs!q@Y(WnX=r-iP0BK<-5?I&e%OlqcwdCvO&PP9!r>K7iTUjso6fBsm4O6&wt;FXOnrE?Lc>%;Dv6f# zhdw|e$Dr*Z3o?Cozwi&^3ny;0p8#Go@8P@Djoz#aEfj;zSY>k6fmOaL-|{uMjf94YKn!33wzGk_=Hi$kXSo zKALpj&(z&N@#0ed$x3rfIdU$quTGERWlMvVRJ@$)s8O5R+J{vHVzo!tXgSEIsb8#x z$!jJ)>0TlGR5-UNPnGgMz0UbHvcFw@`3Y;8in*%;gr1_xmJp2Vq2FYZ0n>^9$nROB zD=)<6mfy5XqtcUE#VG)t8Qgk(x4gITudxK>v=O@K)G0sLob)GA=A%qpq0?SNRE#}PjRf$APK`S;(y6n~tS@b+7+rftAE+T=tTP%6ufS+a_E3C<_sCQxDG7K-a(qjG9ti&@yk_+!UxqP zXAcxc$w3V9$EnSun%p7}n6Ja2xh5cnU#XhZM;_&%w}!gHJ}Fm71zkW7+5=ZsU3hfXudTsQ z%Lx$+09b6B_!*`@Z zBZ%)Ck&jqc5_#Fi(w711H4uBG;NhcxzkTCVGx?hM(tBr- zN`J*>gvP+o1o|##pl0h;CtW{0j5U^p%f3oOy83a&9QpR-!95(AvK1+?vC34!NecJjRqXFPRiPbGj!g66>tEG@;U_qo3jkJYk=T z8?ypt-(m}HsO0Dl(a`nb{(Hvd?JzX`w%Jt_aU$zy)leNk!G|zi$Rr#Qt!+S(VuHtl>ctENpFI*c zmYd$%Y+x8n3~n+&7J%$}PCPIZCnyK+*J@~+l}Q@5nR1=D>H9GRuMpm!Ey1#V6D0JI zwggh*+zO9PhGweu84MiQbSkme($|A47Gx?@}d z$B-|lkdXFe|qBrz#W2+2le&lWl9wSKSV23@?)2$<}B>pO|$OSk3=&9 zwgJCr-FixF9M!0kd%9ei&g_$=gGd1xVY)$QN+V;wG?2Fo$iCXZws%a8qV}{EZ05o{ z4E|&VB-=kap(~1!Q3J8Gdz&XjP_(^A_amSrlDCFmUwkqhx3wt| z8a*p4EOUzbw`}+ZCQt&t?55Jx2dmfeZ}F&Nia}_K1eYvq3~k$eVd2L}feOtPmf{F* zHa&{`Qx;_GYRChjknAL#7Uick%I>kA8GtCkvJG&%%6^lCEa8{O#QZ>cu|W4mU&)0F zd;|z76OQK)wgODW{R;V}9NTH8k;j&A)K3L>O>5y`4csJG4MIXfx;MZD>iz5l&1Jmw z_H%0Ir~o@4&BMp!v2!u;wn@HkFT%ktuN%gfotmaB&t1i*Fgt29crh_1*BDJ*F6#I3 zNBHb-Xd!t61@^62@-b7{P2{`!5rq_ohprj%Vb92b$i=FMS;&4nmtWO7>Ma4Oi87gW z;!3rA4F>uTMlGAX{X6q$Na523Hr-eBh*iM#fDxXP zw0QX5W$JX9`$poX;ht0Hpa;kf9IdzH^~wDDb!^+pO+=tOpKF&5^?B5M2A7OyJgCZ@ zRKY1{r4I1hNHqUd!II4{h07T1mwS%q=mB&CN3J(a!9De(JPV^Z8gx-%1R zy|bl(mEs65UagRuJ&n~twLMgpNg26#Z?Eyzj!TN;eD{<*bCegL!^Hnon@m-Ee%%+fXuLH%q!sHLGzL(H|=& zBV6BnpqoSaGLb&Y_uhS-WsSqoves^9v8fs^uEnm@;^>YUXn29Sbs^L+2X-_w*?vF*zBOjE5u=!3|UrsAP<_`Z;7O5F7R4aeV3GywH*8-g$upa zT=&;@lqkjzwq!9BRe5o|F}A`KYy@bxV!01_ZwCg|vbW*ivr=@c9W18^O(`Zo*VurM zZnd(y$F#m+F1Ex}ab8L?nP2(I<#VdROTZ$DN>tg#J*E%U@kx#F`vDW8HrU32wAYMd z9kzqqpc`YJM2o}gQzU|BD56`w;WzOkmvc#n0GQ|sI9yEK=d`BMOtcw1Y&?xiBqh$n zSb`0}JUPJfba1~UYOG%+Yi5MSN7S*C%GzbDXaQjWkvVXt&})p>48IE8FP31I*lAG} zX2+|N!1g!Ewl|}dEEn){nyUMEQ!@njw`=<+M$dX2dO0(0L3=cDg5mR25ecU8C4`1q z5NI9pu{h23*!u#N!p%l|sfOaz>)3>^P^`hV4C zMg|Th_Ww_{Ii?j<1?3WrO-3TjIY?NF*%@MSXJ@b2|>D&#&fAs|k&H?$)f&tB<}qn(LFNQ9m741QJa2{{9Jw6yU%27wj7d zfIuN09Rj7_#Dvveq+Q^bc+7+~u%8#lKp}#k=J)^s0f`3Sr$8S9$A21bUVR0?yh||O58LY7rXTjbj1$0D_t$Ue=RdbOp7@#GEWUut0KkS6RyPM8W|4}61D zXu!44!k>5N9i|}mHWxuP#LqgOXq!8O7X%Oi3@E@bFc2V|8{p9Bs{f5UAfOKWn4a{{ zjxkX}Vh$A?Q!MLpdZ(RlB@!iNuus<%Z&YyZ!BmkkCAe!afqa86WZ!(S#fj+Ky zaLtfj(?i?({abDahY%qb1N_(?JG6foKEyXZu49LFPC{E4xFgA*l3;w-(;g;jK`X=k z9Vs#b8leBaeLOuK0;eR<;2?mX?=f2-A)g;CqkkMalrVt~fP`xR05=BC+gm~;1n|4D zU(~NyApqY8HwoCja=Y*$0N;pT0tG^UnRD1Xp!cMk1VDe8V^}=q@*M(G&?&Bv>3pZ~ zILziB5u-6!4tH>Lg$)w z=f*Z+ao$az1V%+uNC!dOHGkqB`55>wFwqxa$JMso@ofaoEjKU;ZdaGT1ea^9Zan|9 zM8pq1!o>W69!2nVc~D&z@bMf_JkuCIK_jG`p8yH_89zZ{f8GeeM#rw8FG(6d5wRUL z-@1-%G3_JomvD^np=+7gXD4AEN;&i}oCEpEpZMpS%{_htK_eAA1kQ)^w*l2-+BD*; zID0}JP2N`ydL#O6E&~XN^%FR_Z!yX*xJc{JQEi_@3HT$|4?84NLm2e<&;dNS@opDY zJUS}=ARydF68YlT6MVRJ!5Tio2HG7rg}-&s96kc(>>C~T;tD*_Usu^*%XDk5y>SwP z-%i`O>>vE^)km;^H$iKAo|%4Y++fb-;1<25u|L^Gs#AmFbTz!l>YDEv5^UyIV zeH6#IJnnWw8lpia1mI6Rd9u1iBu#|$gt@&7e)5k0>gB+1d%T2)nWqs}L zhJ`6D1n8FXb=x_td)~N=gMwoheyz{sZvoB3s*;QDMSa{$y+s&0z6}i@T9lNG3<5rF zw=i2*kyNdktL0}7Rn%C*^!hEdX9<29X3r4NAhpla$#!|{6lSxQ7KrU7`G!{95_^}Z z`Etz|pmKH9PNd43Q1L>N*t;kNUyw9 z9>)1ic&O867Qfz64}TqDn9fQo5W4xG+@gV98q0}dG!}efQymsBb)qTTOT0(TrQM>( zd%S_Q>8gEpjxa+!f3nKDhx%R8O%TOc;PFsW&o@?ddHYFGRfm5S4|@g<>&a*7o$y;+ z0{!H|1)|g1{NvS?-E|TlH3nPhg+SL2cm&zm#|N~Bw`n#j_Xrs*nsORq;xHqhijnfe2lL3LlH~klTfArBe|33SkvqcyWZ4&B4$>|!Sd6*OkB4(h z%dM3*IS4&bo8o561h$=%8Wj^v6{RuX5-*@*ypJV8d0iRGh!@mRo*r{lEH@F~caZ!z zAJEJhcX?CW>ckf}$uxam0f+OhY>lLZ&uF zk1u@@cDrc|++OJ`noOb`23-=NO%HNU<~?yuw8fQrkSs%<-?V+;Cbz z-JBxCjL*dbR|_|e2GRuK)F_v@mici}iTn*?18HBPC!scK4T2G$X2tx&_Nptpt0UUW;!d!L}Bx3A-AE4($|?4Ot2#yLb|DrS$gUVzFYK2xG4 z-qcOuLt2)<{5eO3SL-w$>Em#g<#na{;zU*B7a?{8;*wCL@JNJNXQGBvYfl~~H#sIU z7o28cxZZYoc`JA+bWv>2RPY{<($wH?i#2}>v86O11qi(|EB{5MP_RuYkF&k<@DDq> zY_GnJ5dP_#nYjPE77ijF(w>vIf@w~ar1)?DsBGhwZU~2hYpM>pVsosLle}qenwnE; zsx&!ToMH4oL2cDL*%P&^gRPM@55gs-7R}B)3R};(b$eQIk*nfPMozKVyOK0T+=i=c zhXHq5;!>q1gnL}72#({Xh5J#OjPOfBp>n6@Z~U+&Fm&?g)-7Q|DQ7#yPOuC6H zvULYOJOfpgg@$jhsbOIZpYKzkKX&M3YDAxaAMJ9y{a$Zs`jyTra7p;&O<>DVJ4wBy zxb~?N&H_`RVt?s2mvqR+P*75EqRiikOtqOz^c|nA$osozDUUb62@m?islYZJZnMO^ zngbdsS}R{F@6y!QZ~30tTKR~#XodP=fotsTa6z*Ie@a7golSf4{prUSFL-3s)t-+(}j{rfxYW-oMrEjlNbGnnGrQ)`X zLg>dNLZzJyf>EV}o)>HN>K3>=E zLlTErE963X^pwJPnpJSt1<&Z}gl7LxR8+DP+zo42arIyfpG=y}Y(%3f-=&#Ru_;)jhBN*uKL5e?+9ga~QYs3V(P{5*R&iFEu9Gk$xz zKzzwQXQ{aadKOywi+A|U#7_0?c2nY2K*Sd><4fIhAlx1AUBWg>9I}k^aUrEi!UmW3 zMq=DxLwMl8^ELBB#t6GLLj|Y**u8kd=jY}#99!i856ZIFpTN*AzPS?Jo5WFX|6G>1 zBrWuN1x$0^c_85W7(a9 z6D9uZ--mNnp#49mTmtiSn%}Vfb2wlG;_A^aSs`ZRvn}B>+CV#Ich4#){=;kS==##g z*CfB%-u#-6QN{O^s*v1h3VZi#ZVPFbS6DfU3`#&QH&R)Ra|Qr|uu6IJk11n6Dg5CQ z?E}I~oYEco+(3gi-(%15NCg@gCRC=Vm%_~SWTFU`-uu$DhN3o?Xm@WaL69(Z68JIb zJ>fJADRpml%VfZ1`<^(9oq>@Z#?uL8b+x>N8!r10JX)i}1xptwt{hGd{Xf#6HO{QJ zsV<8X@ufkq1O`fgA z90XQ%7}mq!O)Iq!5c1K`*IG8%9M<)-G0AsyeuQXHQi_>Pk^4a-`z$OxSnPXYjHT19 znqU8Dfi4GCPdqpdD*E7fNuK>JA6_Qhj5i(dxxo?WitH!XfnZ4@>a znfiD-4iRzhgiidrrBShLIW<`^E{Aot%`z(?d&WKvvRwMtL&=BNGK+SQRJKv%M^~S) z0Y+(w()43DrI*X;EArM(%aymRgWq#BmX>tdXqS$4@8f@n&$xpAl(dO&gXT2l`+?Ez zsUpA0DHDh9l+Y3?Cbl;Xj*uNK=+)vzf-M%B(suDb{h{Q|W74J7+wE9I!yi~@E!71r zgbBOww?EYeB0TRymfK95hF*6e+QV}R!|x(1D&Dx!GHl>)&BImV{GgXZSScR3hB$_O zIoBtXvT|BZK5UP$Val>PO_J{2U~wR_Un<92h);$XB_IyI{Nj+$=-gUWvV*MGdb zfN_=^c7y+bPaf*vg5R!{g1{L$f8Zvj+N7lrd^3{FSMNeY)FnhxJaSaFGA44|C)KOo4S|!oR$Oo=u46So2Gdi^oQc29F$_vH2y?6N{AW zq@O<&;+D$6^XXEq4@)rDA$WRiYzJ~utE1)B9ox8Wh40C4GI7;3$j$J0a)WRoi!(3v zw@vqr-cRD#%`4f)66z^HDiA9W`Q46cPuj~xM=+)}B{dV%wbXp<#I@5)px}bCpN~XA z)WlHUfl3EIv8L!ZW6hD%A*XN&%q%zv>r2nZI%)V-tgRWlHB}SsDz#bU>K3(^w-H5c z&b_^}cj1I+uhHaslbIUYY)iewbX|Y;3>{Y$6*%IDxpj7k;_P zU46a_h6S4 zHng#KlMo$(9i;+Aj!n`#X7^+(I!H*2y}PTuvS@a<=$DGeb`gv6h@BVy<{ZTcXY&rh z_7=BTtj~kbieqDkQLtUz3$-J<?7J3K-t##Z9A;_?L7^0{dZxu7xR#T(UzzTs z=3DUDRSCZ22FvjXa)C3)>|c?b1-&aH_Npl5v)6rZenjDSXDUZUh}bsv5x5gzbe?>E z$#meVcAg*Df@wK5{^@oE2)c2RFXGv(JG7D7doS3=z1?`CeVds_NH9OH1UkkmQ4VKW z=v)04EDCN#7@Z(d&`}EaCp&a*O5}$kPHJpc$W!h4Ib^-EiaXbl?=aphE z40L97{|SLDOhYXVVnnN$qJ@V_52J`Z)-~o~nPk>@&-jZdR9t}6$7Ac>c+yxqcNDsH zS$MR;FiY|=7RS%9N}#I*)Kijp$WXqiF8l@@p}zZ}fV|I_e4gAvu1*cZB6FlFsMaJD z-LxeMqdPMBRYo3d{Lp)LVg-eQ>K+N!;po(@-u}yHKhgxVyc^-QrcLzW!Yn+=!h{Yy z2+1MAbg^+Q|Jb(jrF1Dit%I$XxqVePOvtVIO_AUpQ+G)=FwY0)h5exejE2*%xJsGL zFH_D}iospFn&JL9(&ItfA9lB4VOQ{v1og$m8os+6O=v~NjD5JwTOSuZGCMEYiF=L- z)w`jB#>@TdSWUPEkx!=bu@jXa-O8SKto56aR_DSpAyY6GW-4;-iEOhWS*r$-dcI|K zUI6X#fj2XLXWOl`tFWMWkIORvymA-R2F}LWpNl+nBtwnXRw5Kvg;<2&f=U|u!ou#V zkDl~5CzLUBdmBmw#OK%TJ@aAr6P3XZY9Y&%sY4rnm9tW9-%U+tFra=cN&&?XIV851 zxT#*kmAgyEO=obbnY%h&ifwdsx=GPbL=Q{&7W5IVKeq>de)93cF688Kesea`ouCx+ zb`2XEUMnRd2W6&Fm6~B=j=lPu+<%z=yc8}o3;MyO>phofg;mv^7IQ3)oy6d0*<+!{ z%d%^1e!$-pqR^cv_Ce!$oQf|lqZi>=yWibwta0hPRbvXR5}hw+d-a7mJK+PgEsrXf zg)m+}m#TI1E0Zz&JZG(y8d)skMFHQJbUI|^P-I^f3ZtBTH_o{pKiwWC$#H7}qW@ku z8Z_xur*<=w=5=q~LvCjxie2l8ayC#5SoC32?$XcdSa2gUh7D1ErgMufD8W)l}H&oIGYqckg&> z)>i&DW!)`0mdhyNi;JJJ*du&G%QtOa8&bj$=*g%a%zge-$oC^;p-GC}x9%k0A%rrP znch{6Xy=kufnZju!qB(6@pRkxBDQ)ol3`)z3Dy;{%WI~Z?IsPeLjTX^5;eD20&KsK zhn9LW%T4uRsc~S6U|LJ1b<4%jbL~1rWS(+ELChU)lQhq#3AJehrqQ6M$LgMt2f^X_ zG^(>FcDk4{IbQmzbsc_j&4EvCeG+D-**SG6Kvd+ zBw?1??>=nV6O8OqoyI1ijZ}sO^gBpu+vfIdi<^@{ibf%W(J5gA%(GTv2!n4b2%rE5 zRt1DYpzmX?%g@X(8fikAJH=iG3pyh_tVI zQ}gz`@6Ph03;m5l{En!h1J*>J(o5sGA-ZbIvDL0!x+ywv^d zxSueRwvqw5VxoKQm}I*4gK1%=zaDl%eX#|=xx zwk6@LH_!DfZT!)Qj8&vk4mHnK7(BK%XSxN$A9PgBz$C)*){JV?2-@@>pHOXw#_^=FquF z`F?Y2XG75eo|Jn}g%PC?2F#|RZ;l$>IUWI9gj(*8`v3*kJYe<8)~tN+{2EOvUGz_1 zF-7HYdgxGh!k>7-DYHeBCt$hiHAz7c9z-lSWtQmvrGD{obVb3FykoaVza?nKNbp`q zkB{=>1Y7AZlE9!CW=x&W0@abu&=D2YrL9&<;O^X8R&Ottb#{f83~nskug&Nh@!!*c zg=Qb@*gfqpinNW=H{v+TPtEQOqhUfAeR6xQN?YBc<|E@cw$a1DQ%TOcMoL%o(c+ zyiF=Jr(O-4Z)iz@>lpd(q448O2FKaJ!JtW$PS0otp|Ox1oKf-GKdy>=;25;jLaJ%VFU?^lu)hd&4FcScqn;myjbppd94%n{=O`W zxd0w);tjxdY@f?$ww7W}9@Wn0!_|m5>B= zMB6nGGBloJBx0UvzWlI0_4 zIK4)>R8=X>Rk@wwyyA@wP%ul@`Jo(4I4#}sx)CbGD$itod+W!2Ne0dX=vh}A_7-l?=`ZBv>Zjb;Gb>(B6`e_zuk1=+ znMVY|>{9!nBPOlOukjviZJP$K&yYFX-K9JX2?c+RuEgSHGY(uvs_Bp)dIf;(qAqY= zT!51?u5{8|RNMSbMQ{gm#%~VJq~{;q*f_>C&Bf_pj{-JUX(il$R;ei;g&|n7bo(}- zQeq`$VUM*K{dbw$IHtXSaB}ErwckS%P`E+ej(v=b>!QBjF58Bq;mSN;8+lk9r%Uz$ zr!;501SoY@tQKhJSRLPdT#Gd#oNHK*b;CL$dg02TY2Sg9h{XbR#;?D(#%dY7=&QBY zCJ{tZ3^h~KoS10)s`H8dX3>mf;EUn-b;|PF;$D>bWbW*CbLAt|a{Ns#_kHT4GAP%2 zT`;W8hM5zMbntXv5WUJ<6({WEyhJRNwZ!1uG^|Z)5AYOb&cB zoNiWvtJ|^TX*2$1NN{#4p%x6K`qT9Ad%i6VL0?^gxy28>0K0qGB&uThXGolrY(7Od z7Oai+lSp$^Xa41N-fYqw_{Cx^v)wkkP&a%~_+x|&zq!!iH4kn!(7hdfOIKfRj#6K1 z)w0_ja{Ly|5cYv1?gdf4CZ>3yo)NaCB5lNW&AZ+ECS9WqsYgjcq@y3-)2Lv_s3>R7 zhfz~K+jHs@-|ja%tctBGurGLAe||)l^cWI9;`exvisXuR*sZtwFGKU0nf^4tcrIa8 zO|RXy*Q6lOj!n=}I2a2~KSqd}-o??pE~WN_&W+h8GwH<1s=^dGE90)xtao*i@XA{+ ziizNXg$m#d@G3(g2#rvQ8=JH?c#42xaqVz%<2&}z=4~*UX+{M@v0BO{gBhgC=olrI z^1|VqCW<}DdwCY7yK87>^L)9acJGD4Sfh*OVHGELyN)%)3jT2AJ?GOX&k6CSdC1Wx z<$;QNVjj4|Y>|2{TSPG(?=V1VyqYIJt z#5Q7KUP#9eCnUzHsTtMU-@)&EW`8ekNsF3XB*R=s;T~?d8-*{}ek-z0CGhq-jmz|U z?$Xkp_tB(*m3|f-9&E-Et>B(q3oPVju7Q?pT5jPb;x<*$mS~4$E3fiOCtJ@zSsSOq zcf~t6SnT@*ZRg9oeSWYcw}q2O&6wFm?K*_23!CEUFkh0A_t+aFKm2sI`*@Z{hZHeT zh7aCE012Ws5V>?wQg(*$yUYqj+gJzRM+?H(NPHq)pK;j$&qG8e33#B1?SACL7%>rk z$rlEO{ykRiO%eTxi#_)CQNG_-c5WS`l%m&=%9t>nd@8f^``nR40jU zMjWQK91#irkOtRF;*pqtw27VQSL`H;?+ltxHkWSh)Pg`a#y`hTryxDWQClQx2H$|j zkbx6Af_mLVxJuT~`wJQm3@(=;6!xw2x@h?vn7yoQrfLy!RMUE?-&h)$I*jlRuMfw= zv4(Fx)ttNx)NO#$82rgs&tS$2Rxht*j5Am^=%*t6$sA5?-7ci==0XPvVhlkB6?{;v zT{mk_%z6X|NEa`V-U-uBFw@U5ViNqv5XnHzWhn3L24L?!K%0xUV=z>_9M-eBUwt_R zQX+@&KGrRi1rGfGC{ja=1u=%7<=+AK_KBS1!vr52&EEj5!!zfp-8W=pog^~?o!jAy zrZ>t9a)Qy=O1$-MoeQ_sQkM>9`ynnNT^xmuscc>1k6MOyjOn5PIo5&=u0cLN{}6P9 z{{BIOjD%o+G1|D9)RlT#B`~P7pANN|9~PUQU?C@5yo+z?<8$wSS@MTHpI~GulX~4} zlrP0?OKvsmy3s9DiUzu(#o(OO2b>psx~|xNaJ2JC)}WgO2cAUb3)XyAv$V+4l{tW$ zqc|m9>)$3f=~@6(zi)(T^c z;cBGw;Q%vdagm`Ho+i0rJr->qA=&m{k` zR>|uDL>y*~^Q2^3mSdhed*7^FwpUI}mF-OXM#YvWqurG!OCgzemd<)^4Wljh1K+_~ z++Uyk{SXl>A3N60J*@Vu-H9r-vCV2rg>Jas>*}G8^!$=(t9%VGh(`uLFQAYgM$@(F zO%+KDzZk7G?lfqO#Q2p&eQ|Q2UWzvz=+UF1voJ2zM~(2Ba8V;a=x5_5Z79j4mI5 zj^fAX&_rDt9nP6{MbTrwKEqOFF*QYtqWZ6d*p2|`xlt@jr+3_>cmk=^J4~@2Q5tJj zZC5IGq30#7t|g4TatSXa^j4EL4R;?S$HRUv6^KZh(xr6vt8hj}(%~F8Ni}Nw%RaBM zPcS*4oLuFIm@mFty9c9?F!L>8=KX9)OX9L$p89Pz)rpO8F$_al+4{xjDP>q_8!qU$ zvJu&ttV+x>3Vm4GnURaqilL_!nKQTPDwg5xo7frgSb8iHd)9)yGFEJ0dkX6dD)4s8 z25dS0Q4W{`8{XFx3pFgDy#>xLfx48YKmnX3nkw-8hG7|dX7+IUb9+~_;0rNjyn^$^ zSM`@aUil1JMYF&n%R{?s3M4k)dDzlT-sHrh0~`uD7p7hdd+Wg(M4r^L%V~9&#BMZX zz{|7__C#HGOu+L|+Q7^jjx8jj5LUx?t0~JDh8-z_W!=DFwYgEqTpDqU<4qbLb7D6H zu4HRx=NcoN%+Ny(muje#^!)2b(S)O@DtSk}ABO%HlpvvNU_!b>6#kqYbO1N4Y12#K zuB*Ak0kC8ncm}8Mf+U6!ugc^qFVua()yR6=(fP8U`w_Y3saxH_CaWLz`^so{14UWv z)c&I`-ycyaF9n)EATEn+Va(B?fC2GQvKMeHXjYl5&|KExo%xD6^~`#li{Ox)!ILVs znzzt7_QP2?l=xjD&kxez>p7MnCC}LYi7Gv5oKZK|wNAQWh(X5_`RW$G$56pW1SB;x z-RmTS80*7$nKg|DPRDU>P%M|W02+BSbI(<57Dp|<)!Qllb~UyT2&`r z09Y01B^OEM5yua8O`U8I)|Rgy8{0GxP*tsV^{;;~yhTv8cQ}zStbyBTHCLTK1OMDi zuk1UCkW<+WS!jHVZdCF}1+Ka75?aH_H1Ul1VSq?$&rfdtss|J<#{fl=bc&EMsphvp zF`ubDcU=e3-Vd<(Gwc&q&L$*uBFKJ_ty+^^D64alu_2RUdSs+=obD~x-+g4DsR@#& ztx~Aw*;XionGnsjVU__{Q-NC@Z9qsF)}Vk6`|0N}m(7=^Y!;Mg(~eV1ZE0tW{IBHZ zVS-pd198$LsWbWs>KFw{cDH%029;l@t0Z?3CVt5bV}rz+F4)u@1)Lc}CKh`J2g~Q< zuF!Ex*xPaqwVaXTdnn;~;{msRU8WVUoqYFamqy2(ub8wixUM_{-zC+LDrmHWv6;Em z(bCi2R0^XW@8{rSj}2^v^Zi#ovl^vBGrLwp^W8T+?3Joq1gV*b_kdeT&M?c#rDE=ZBs5qTIYg4L8m%omhdaVbMGCj!C}-Kv=s+muuE^;$p3C(rK6$w zHw!B+GXw2EoT<3D5Y&Pu=Jp1*xYUB?I`#(q26|Td1`wQ_5O(&q209iH&Z`y*5>~r3 zi0wy8&aerqzyprLKuEl`cX&3-77C>cns~Gu!qM4q`4?SUQChx97O>ejch4?^*y9^b zNfZ1E7n9<>LX(;kEgkj+DfwGXMcbEDW@BSB3?Hz1WO5O8Mz($X(YthtVfN}%nGa6I(wMnV5H`3p=(PL3(&61FsDFxDhf#-G`3p@Gnrt>R-BU^ ztd(gV7!*n7ucjnZ9#1FSjSe_MxH0LJAM_O2uM_U;BlAcKu5E2sZfs%B9pP8OGC*ahN1bS)lmKq)hmj~*8i zIJmN4Ha^+`8e>5h8`0AiarnKBrvuGSkPa+V4-6Iw-p1KCX-R1>il-q(_iN}`JN@+r zXm319F8t2xnV_h-!iOXs?AI>E5BB9aac?4<94Qu9^(pTRuoSk)J6w8WBztIxAPV?W z@@>`%2h5`gqewod=)N%PP5vKWzV1koQkk+tBOwGnBSC9Q5Um!+NwJl*z>+`voP;@$ zSnwX8dcaS@8SeS{epnO#vcb2`O@HU*laHpp4`;o}!5b-t@#KAn@_?T`xs@Bax!-Q+F=2EmLt+?wPWeckvKe zMe?ug1$Ac`En1y#1$!2~Vy+&Kqhd=dC^H;QmpOe1O3a6^_*tt6iY-xYqeEXX;hACE z)UKALx_?Iv?cY$tNdG^k#^E2-z;*nS8XQ6&2)vb-Xclv|1k(m3VR2@7LK*nKsF{N5 zDcHac&K$BDv^p=^IUKq(%pEf{B00^hLs?@ks;rLNKWc=ww45zu!Ls0BS5;fJUBAFt zwp&8P_*1>6pHX>c+HAM)pn)k_fTi+9g7&7kkO_2Dy{7c$CX4Ch6(o8+aw(4$Sej{_ z-KE&qGt;RJ#EN*#fEWkisBjTnJle#AD7Tx~}A~lA%z8Wenx7 zZ=e8mgyb2P*M=m#+5Y6mQF-Qp@E13`Om8`ELK#jy_a)>ON()x!k990fEvyX}J-%>* zehZ-qj#s9wZqDo2lxd!GCjk$yo2tY%l)`E=@r_nZ0T+!AY0!!3Rm*f!N?mi6!Nc`e9<2V z$&eK;7m@gGY-ndZKL9P3XpvM;a$XBbdSU@*c4Pbdw!uvUCy{|tgh0Ojc{OWsAlPPdKm>$pK=y$40R$FFAo#rx&^IvOLzK{PxPOaLf{<6Mb& zznZNj>NW0EK>Hz|yGxGAOT!gZ4J6bI9hWxxNCuXjaK(UHaA zBJ79laf$U(@{}@J=}~xRmWAedK~K>6IzC**Nk95FKHfehe(Yh0e?^brmX*@-l%nFC zg?8a#82*$Z@_p*m!;tQZKF%$RX#Lw@v1t@?&nae@#M~F+jR0VXEWR?vO(CgqWJnaL zi|I}CTDq!P=&)8F9l(Wox7P((o_V{CLnTrt_;&=+{X0QStW5u7f-L0!iaY-+LEtg} zLXe2`&!Eq@2MmOAMGiQL%&qbE8%VpSGlxGj3QbK255p^I98O7{^>ODBrv>Js!21WhFcBr~lF3P*hpc4!2|I5Y?HJ`PPGF zpQS+;IaKhm3J$F5TgNI)mz>eHCs#GiY6Ihu+pK%PF{Bud2wH_Ou)o)d^ z>obqGY*>`~G}sl2<+Wf+GZSC!W)x|kEPD}mpDwlHvu)R{qm{KsbCvZUl^@kJv3E#n z$OB}x77rKVwpq1Y;p_7@FPN|edkMf$yNrN>EZo_t!5q;5+@aT`#9-9`+2F%00P}Dy zV2Iwm1G3Jz4&MB&>TsrX0K!Ru0Hrz8iatFpKs0==;u_CTy}H$U_Aw$rSm3B7MughN z_A_BL>4WNm{yX77G_ZYW`4|X+cBh!pK|4^U`s`3iA^@jM>1_`;u10q^tz%wopQlht zkoh*l(Xn~tN!VEgj|s@2H{(NQ{hL70&{QG1rFI!nwkOB|`Kcf~Acx5PM}fLm7Lm%Y zo&I85%tH;rTg*aOCvV99Z=II9C`FvaFs7r+(GB|!={{BMXz;HIH?-@DDNhNz>w6%( z4MQPP4ulI?HeO1{7;@A^v~KFpsLs5Yc-}?xV!w@rOOoWL6Av@}+C)oWrhKNV9D$bB z&w4n_e!b7qJr^iki%t1zvB7wr5Iz^oZ6_`EP-lR3g+qC!2wamRFn1k~GO^wMWVtSF_CMYdc9y z?q=#UeLa>x+Fm$HU^=$rE?QP#CI0MecQW^F+IFTHHV#^vj#IuEyvofrNQtxCPFmy` zpOydV?judsw68uXHeafD@cPx)7>)dMs&|s=Jk#txvzA?fhV>0BrSs- z`lCqh8BSd5+xaN*)^lkUzmoFs;=GPgtj(@WrDEaOnDg-2&Wg(PF}a;^`{}|U-dcmJ zprDjCXCs&5Ip>*-d*D7YTdNLQSn-u6W<21?EQ=s^VCNn{rV3gDMVUwov97DutzTS1 z(-cr$&ywpqz=IAMhH&sL!%Tcf2>A%8rulM$RMnkoS_MHy+~wszR_;+Q*wwYsg!D*V zpeRaUK-dY(eT@7ew}TzGyJ>Znj%w@FAiu-NQk?j=ksj5^v%Q<;X%A{|7w2=B)XHod z{D{Y&1riYUkq|*!MEw|H5_pMT`-sOda*ur&Qx!zRuNN=O{}jSm(hO_JWv^v#8Vm6n zKb`qkPZzFM!CqX^xDCXvu=efFV}h?G#%n>o*|-lv?=_aJ6o{lS2cI=B8v!)O3=t4Z zTrkTo78xGT96YiS6xYG@=jrHuc`z#M<(63 z`~uG^5`Sek_!s&0$JOJZ)NX4HY=)!pFZs1pfB1@x(?;-@{2CtWf)3Xl-KM^|E0+E{ zW*Gi|!A$i3VdnoZ^Is{;|1M_8ql;Dljv2Z zI^WTUi&L%nXGRbC!AD)4_-hTff`Qsf&RRm_79$RmFpE}Q{_jN}i_)({#L8tH;!<1r zPX8)tp)=f=oIbihAg5&{Z8`>J{qYK&+BbTS-`|$5RYEBZ9~9ix9E~1XIMJm z3ZN>UK0L?(2WaOIGmas*y_BPEto>~PXDfW5FCY)3%&1J+sgW20z&zNlDn#eC?|h|w z5UQFNs*~@$#0Nol3XTOYGY0UlQ1hMX+v0EYv%Z2r%lLF=8;!@nqRR(5Y?JRBO$+5n zdL^aiNwP>${$^H{wX_#*nSmUQi2TjWR=92~&@z*1n@3BIet#V1G&S{GimhuRj>_^* z|Jfol{5h3`pzCnNKc1v6n>{7?_S+#9C-c3N&z^#GFk(C!M;lKPCz~lQ=t9SOFoGCg z6fc)4%i|(SbU>=1el`Ae`$3JhN$f0CRi){2>Gp*CI<&}Tf8D;+YchE5X4K)8-#MxS@d|3ytZfguK4<6?0)zY4j_5V$rsr;8XTljZz z_D)yG^uXfgs!Q#mR58(gS*Orxnzqvut7Or@wPdL8AvJs0VZs6OR{h3&XC#aD`l9b< z`N%gX$x>`JQQZ`8Wtc))GevZW{Lb`Iahea&yZA5Wu{AK(22D8=HuHQOp z#}NMd4sy^*F>jSHkA&6EA2#UYIt!PrxDqS(3zC_EpM588KMSWkEM5yejGnRrRzm(E z&Vv6b&axgp%cB2CAALPko~8&zxD*xIRWqosReW8=fUqqA<6 zhuX#Y@Z|*4R;!`=^1j#+8~}{~Ph~@X z?*aU)tgZ`5x%~}O&~39-52Iqq{PR!?M(-BHGvQ*-M{noRq#8o;>lHwy>28Rk@SQ}?Of zYkOwq3FT1Xc1tT(Q$Ylikt9+T zPKYujFES)g6HX}OdxICs_$Aaq80jj2*?=FZiXdbWoHrSgM-`k$6_U5pKl5|fn9zoS z)`m6=gZjPqI?R_E({R?pT|TMvDGme-B@z-9F`}+%Ww8OPycY*%zy!H(P7Lh>pif<^>t-r|2}xI z0be8R%TrXHS(-%*U-*J`%b!1(flgVToRjWdUXk>70-;(r&0JC%>5N;ALA#fg8%{?W zm{MDM=Xzcm9G*yAMIKM6QoEfl#yEpJ_pC%*;Tvl$QA_1falGzWgXzAUyPu?=f6g3- zoQRC3j&?2>!s$cyi+1i4V>#7Pz!ANCeVINGiuS5;x*#3|L{q#UmYLtM23e5rAdzUh zV%^XNJzKATO=&CrrhLu$qtSWXv_`~H{c`i+9NHFiW1vNmr29!ui7hqv$L3vq+)Lae zJ@7_h{Y!+Bk|Pr~sOcv>8Ei z+6^PJ7^#wO%Sw4D%cEC?$&gy6R%MIBXoAG+jH^sVXu&M0Vxadb&~HHbAItFGAlm@W zU4YwwKQ?m#volRy!oaFQtMR%4r{@6@8dOmM=DX>(0i-$b3sv<|=V{dY{Dewm)pMHS zNvRT%=*T9gi*@k)71}}iDFW-}*MzUDv$e@VO#!C+r~sN70l*|DYq*bhtB7!K1x~BxM9LE?15lTcT9O_TC?O{KTm3$av+9;r-GoFV6w9?{h>cWjq}1J2ub?el zuNK`0Ka|E7x`|Pqmnru(fum*tYc(6u^zK!)31?Ha3-3r}jn~gkt>Aruj3nK%6Mvz+ z0%7hrboNP}67@;T>73QB(#ag@AT=x)Aq;GY1*A=BhR~l0)hp#wMdN#6&-=mLM=v^R zbmuy5`TQx2I$V8$v@@>a<290sx-Xwzr4tZ(RW(R~q}7I=PMMViEOo{L$~@+Qb-E^o zHDAA;ByQefC3cf~OFn8;Av(TG)d>V9r0N2E*BYjH9i>;nz|tIh^`m`%S;_uIV}ecE zQ#)!W9P_3o(N;CmhRSk+R$~ZyXj|^<5RJM7W=lHZy7P=Yxw*tyY|a(@)MT$jVpfT3s3(mi1@9G)+Mp9eL}b?-IDRMWskJA17pJ_bvBOo z3aJCpnIMpMGqb+$(_7gd%Y}jDlZpx&!^5emkqde-ibOQBgyRPBx$m^1H&Wj+?ZlFv9 zW^-{`=MKU=+ce;T({Q$;ikjU}wH^m-Ne{bbcW)Rv5c;MD-iPLgOi;B3K5)ioNB3T= zox`;8^lMZ@I6L!wO&+h-G!XBiK-&x0J8*KR?<#~S-MErn-glN>w1_j_a2l7?}w%*>yg!HF^34^g{n5P58_I^^~jT+&&%T@tG z|B=QBU5s*CZo(b0{f#qYFWWcSRWT}apmvO}g(ht0XC;fLuA>`XYbUxtE09iw1$;rS zV^en%bvbPXY($;|bva)^DGClcqIqfwA%qA{iVfEAi6`o=P?c~m!<@J9n*aCZmY7XVXwj>nJCHyO+ zqvyM>z3?$+_oVln@K^!`%qeUR^^%Md8Ly zNXMf^O9Mfh6!S`9$q*RWCd0vv`U?01$MAAg>kS-B&_>tVR^2VbBJbos>mbkHbr6{Q z3r;%`$25KA`>{v94W@t8+cog5O6xg%!ts)|imPAvFQu6kymtB?yQ$#*YCLHwK2u6v za%MZfv4%g^o4{B^6Mkz`rl6WH{z%xK_kt>OE#_}+nwYw_n%Q<~>;t;j<2pg+W{s8c zd9*Q7`_hD7mHZEBYR1z(+E~liL~hY##%UgH-DxNyQx3c)|8+4SU&AFnz0C6h&nX)8 zR~aPC;0HJrEnwQHwC)LfskMzTI0~2up*p6Scr%Nz?sB-CaesgA{@B}YZ}L#+e#lU}U_LeH zr%U9@MxGhHY+qnpI-zTr+doXCY$3WIr|>3}Q$vu%5e zzDu-#587}8MSBm*YW%BM!arLpx)SInpHx6x5*3{Gq}#3_dkOSYc9z}{4#5bVQSqA@6s~)+H-t`E8t{V`}`nm1}%OulKJl+mzlnii1s~ zl50+}f+v;M=1w%pAL#%zb=tj}Lpt9T5oS^ws67Dx871aNwvx6+KeG3o;H*zM?FZQ( zUVwxaRWN^t&)*myn3(DQ=R#M$$MQS5>$pSOJ0ntRyDS!S72mS zR9J9iz(}1pSe0d9t#qs$sx5`XX0|&loXD<)JaCAtMGgJ}d98kPdUI3E7?K%Y*_llr z9GqZ@-b%aK7um|B$I;jFgNpN9@ibO_>hC8wS_gwrN_<gPM6K3z#6W85&&pP=Nj8 z-6|}s$h_eE(d*>!_yjo8HlA#&zP*>eCwaO=|ubZsNOH^%(J{{iId!lDAQC zCW-vrwY6X>Zxqs+nfhwyQe*mZ9K<;KtBAPU^s}|Jv5~K-%Dl@7he_40Q|GL9oquuR zD}xAc&pN$0rRbKZT{pozw*e7*3=wEu568PNAIhfyHotoq0xG{-x?8$T4SLI=H0!2B z>%nMCE8)f&>5bv9ybf>-F6fmVMj2T1hdO{HFBl99H0qi9AH7Q7R9;<)2xk!Z7I4%T zmt-Pf%t7g4@c&5b*aV<#0NGl?eW>+OF&Khbu2H;YiWr7@GU&iK@YVBsQ&ob=yT_u# zFz2W2_J8+-9HXE!5@xHQh9?cfzCzoEcE@=8d7hK=bg2suCrm5ZZVkE(0fR~+KRa8E zc|C^)z4@Re;C96mcFBu-mVmqzz~71$>?e&5Qx|-1eu0?Q$+XvYw=zikd1NC(6C*;D z!$OoJLRfK8t~{Kl{_gXnFU9n?GSnkN9r!p6-y1Pb10IfpBuBx>v}Z zGXI&s`@{ch1yI`7O5Z{6+Ya~JTHjFKK$()3ih+iTj*b+9n(x~WXK(d==KJtJcTKD; z`M+IrxWDjNWQO=u+snUE|Is@v9>nQ|F+=inA;gZ zP)o}TC{qbKn49bBSX%xy7I|Y6JKXQ>KgL9Xt7u?r_x;hh^i;GAw2Z8DOpNrDbW}|L z9PB@ah%0X5Xn^}YDmrR?14n9p16>mxOHy1jYaM%II|ChD$N#(s6BRub!(W$>{^j~J zu+;ynQ%%eKuR&Da#o7RuT1wYc&fWroS`wFm?yr2;*3KT6o{{#yTH4=V|36KI|9v)m zp{gQ>Ap-BYR8>5*v0Q!HiBC-2tBoodH8jk$yaR@YKCLRO43F*R^Tts_+qexN>ejVCA%oAixf*jAXAPbx5KPKSBi~<}u~WV)w}=0zo`}A}YbQ zx&-$R&gpd2DV<1VLv4f_UqCh80!4mNT_p4UaOJ!oJr|4g_#ic!WV<1Hi*Q2eaJtbL0SG_WNB383-}~@w7+nw5gj3HGeY0VMhMS z4%kL#F=l$!M+M%z25_GY#;fNR3AFVAWm$8^RN@mty!1gr7Bpwo+lPbsSrOm@7)5bp zow3LlQ_PY#n^#Aha^ZSPPzUyM98S{Q3n#vdDH@;$Pl}Mw;M2GYWuZz!*_bMX#c~N% z54mRpZOfRD$TUt6=+mZLb@!6_8$|7l7^9XTqL{3skVx%m%z+H=OmytPja5}TVbKpv zNMg6NymI<5yk^Qz3O(FpnCpVe!nlV3>W3d;^1}7aEX2haW|eGCRV@-)P^hbNZ&7`S9G>$ai5$ZXOB5VyIY8Xubh$QJa${+}CjN@z{ zMQR@#EfffD`5hqNZ_SwS3L7)E9adKK7D3ej|M*yUXFmKm#hISVDmcvSR#kBh?6otU{2_)nbY7yOaHDG~rwdH-ie-+%+@`LT*ZO&5{*v z!^Ek25(kF;GYuw;%($43zh0~JRFRHE39_4GG^Ye?2(Z@1e|6XR^J>QGa75M5UZSiG zlUPwc?d&J#S4R*SQwf)##{sOX_?e;mYqBsvBEW4pb48mJAT|r~(D1`RSY=p$=$UZZ zK(`@4J8J{bL!EpZ2UkhXi!8pEPLh_rRV`8oSeN1d`W2czkxON>*qS^8o-vETmRC1Yo^8h#^q9CAi85fWs$PlT@YB3=21H}l;0DW)l7#Cxw6YRK zfJez|>ICaU(rY^F-fKYj^D21i@<;vdwqAHU*8AsM-1X+#mK%pF(lz#log>$7FTt+$ z8X_|Fn5kU9mUYA6I%nt$j(rq1;rQ2&`7em#t|BAG`Kcpj=DpyLbKUm#%I+$jPpwzR z9!@XR3wXW$r>kIVEX^gb&5ewWiaI{;C+DuQj2%B}-?44ICmp9msUJ(nP--J~<^-w( zrs2O4!$#hL*v}vW6Xp6ZD+|l?JnTQ3pKd13=%*hnl5l#K%w-c^v}77W)%X4&4OIKG zl$;T~9^1LMGP-&p-k;EAgr*Gtd161cU#kcY$%qbl#6`iH?xSE@L_6Vga{s|BB_r$J zh9MupcK0yevo^d71 z-e{L8=KN=XOO92A z>)u5%c%&7k@H3r%Nc$xGj4u1C%kDm0N4}LH@lPLPVzle@<^)yfz^!}dZ|qR5t+MI# z&JXLxD5w6ox*$u6P46e$cc6M#3w`T@iO^Er3`NjmlC7aRT}RF7v=OLj-eDjGr8=b# zG%TyGRW=w?$ARC@wM)-uK6B6H_jblBSOrhVbL1DDoGY9wtbiVZ8%}s~nin^APcE0d zw<@Nmen=0*iu3QUFXKdvutMSex-F^Rgs|~ToIJ16`j4e#S1?boFi$i6c~l5{7ARhx zaL;c+n+z|U7o+hQPt`h$PYRnB?%f9_^>+xlSq-=CD|9@dUXbsRyI`YR9x4G;Fusdw zdj>xA7qG{v6}?&u(L@^=6p7^hp}5mRd$k1FPdc_Wt(t*dvjGixFKuYgGXgw6^gf78 zy>yujZli_?m*~*e5<{MnN?tS;qla%2`I(Y_x+U<}QpOxp%j}(%kVczBoLs*UWg(yc(7FE*Dj5}CR%j)9E&onzyy;<7ST%1hYX z^X=;68H%f+Z)v3-1{9M{>XoP(z`rE&yZgy+ly6Cdyc1x@qrh$^9%u4@7`iJ^bnQ@Or_Bz#l$e&AlVa=VVCS?-(uJ; z$#7x?Fx`|Z(+l4d=rSE)i!p?By~HXDi@c|V^kZ=4QVF+5Y-@SU{Ymk& zfZ>Re4E5rby!weCHFz-6s-DbgG`xz2pWAD&n*cQt_kV5aPOZJ(_GNAFY@AhdN**mO zty*4qSpKnYF>6R{S9sR>k70Sqhv_F5%(6Y^a#KrbatLR~v1M-! z?_7Nz-n?VCmB6(>^4#GU4r6QOBG1`D8gvmsscBsLo_Q&$6$;S?3Py$&v0VDWnN`46 zqapA_^KgY|Lj_|6L!e+#etrp15VXD|2&5w|AIJkWn?XDmg=iZK3sW;g3qw;ka|0(! zHvC)Ajf`~?Hb951RjEznTV<4`-c1-m!t)06i__@q1{qJY$pP&1)dB)7Y#b;~;nj<&;3^nOanG~whyGF56g}q|(ihVWn z#1!~+el#E7Bk|~vk)7tTi4hSGHzYB2wuj!m(oo{o&yrZ`e`S)3P~uwutD9z}Ox_yf zp{5w|eaEa|Pp+I7(-u8jdeATQ)twl(yo0A3VtH1#R)02148ONZwb`&e<*-_I>6(U# z*S{7eZjXO$l6U`f^Vw~hui`!>a_<-6_GHX4jNLdxPdSeotZb^?y=uqIoTk5TuK!Rvrn^MyTe*ks)+KBm zZv*aly=u51d9-R1?}@9z-=#d61O7yBoN2=N@VwOey;~=VEvWR<%}uesaBt0y<}J(< z9gp&D5`OV>(T>?~nAoO^exJZ$@GI!Cx5W2Jeh=QrZY`Ot_ux|bQMpb0JN9haHUA6K zW5J_#MGQL(7gu!1u%xZa+!J?{;U)XCFu8MyMjSe6nsSQH2NsJQO@5;BVUgoU;GP=w z6C9GN>X!E&@Y!5w{o(E^d+5LNt4R_GE4@~4ENHsC>d}KQ^TI{>)7Q9$GR=AE7wTvc zKY!&X0S?g>UM>rq;$uxX8#flOa^mpW8$Pw?z+H(@&4RnnR|P$6Q~wjay6#2r@8iqw z)y&GOiSqxw<^1oGnlGJuKEIsx_2u3Nm&^T`y*T&Z$dJ@*zrFv#`o5U$|JQfbe#Fd0 b#U+VFB^5=$Y-40$ZfIc1rK;-c@5TiHWExC~=2>x%MA{L7VdaxrUJs1)6Xw+Wx z4G-s_fbH6vQlTKA(HLYYZ%R3>%G6}0>ahuZ_dQP9<7Xt@kDrL!s_L3NQr<8#(&?np zXf!iE{6G9Z{`yzH`1sX#fB)$h_rDDEAO6Gg{iiQ~{OWI@4o)=pT7M5)1`#J z{pQEtfBH{%|J&=|{vW^j?%RL-^zDz|eE01yFU4Kb{`$+`e){1*-90=q{Qv#_t527A zef;4!pT7O-LVNk{kKcav&9}dO`R&w_h#a{lk{@1O12Bzx&I-{?&$l`R!N#pKreX#O=8I#qaLE_~Q@X zfAP(CUw{40Z$AAW57QTa|IN2weD%ja{QlEF|MGXAzP!}q?ib(PO}~u4{KKW3|5sCg zF8Qy&{Qk>-`1Iqa?=LU^>JQ)izn?<*U+;dyZ@>Hg(?5LokDtEcyubSHkKh01Qh~d_ z{q*a*@BaAXAO5&;@A>(cA^viD{7+Mi8^2#Z{q~!0|N8sye)FqOKmModb)+9YegBW2 zzP~X3$GgAZ9=-7B(?5UnVuPrthRt3SQmeet({^V5&tfBEeXzyI>bPg|C6e)#~Uu{p*-~P>i{f94q{QjGNzWdD|zyJOsg%$hh(%*O6f8TEd z>h6a>{{7vTfB(Z>_+_~Krn}$&_)|%^yzo!-^>^QY`TO5<&aJ`#{ueU<-fBEvq|2mZ}`bGKk-!3oBFE9T!aJOC1Z#Rj*{Qb+d{{795-&f$PuYd95 zm*4#UC)%>iRc8O{?gF#0=I_4!@o)d-51;PhmV3kgOJ&^G-~RahKMU>UqIvEvue|%K z%UJsUqKWRl`1>y}o%8FvFTel(%YV5D`Zt&7E(+z>7x0hY{Qj%ucVB-0)AmonK7GIO zY5n#$-)&m#?o!FWqEFkn_~Rdb`r_;FzT1Z4*O%A)>7L8GH2?C?KW^{;{dZq}b!p0_ z#+#;)hW*!#U0;2DY1R*4U;6!0daduqs|&`hu9<%LX%mhun?vb8QyqE z?xFRkKYaIx?VUwnK7D;rnOps|o|i1Y`Q0!6>6@>9{M}C%;Hxix{Bp}mzyI{@Z-4yV z-D450-+q#ezx%XJ5}5Ojb3QaV|M{Ziexm7qz8hqYl6m{f5B$saz##AamWQ^y(|v;c zU;Xna_taG5n}4{>LXwfTj6ZHN@gU8{Pe?@|F}u!bQyb>vG}XP<0^UEd++Zq%JtJf z7x-5ugB7h6M!p~ zhyT@Gox5n`-DW%d_}$$Xe|I<1=%t!dson-{Ge^cCa8dQ&{|JpP zc*#ojNh1_iZnBBf+C@(}Y`RQOh4)`vOsI=|-2Iw1O@IIR#QO5DE)AY8_~TC(@~^fY ztXg_wK)4K`+KpS6{pKHwBK^~4dZVYuG*7=wPqpYem;U#bRT_-Sk82K|kn zk*G{TFR$NShU8O_cW)#Aj~DBTmB0o3mkR#lPmKLqa{T)4?>0&KJNjw+^EZEY*BWe> zmw*4=SAV>q|Ma_0-+wCFSty{!(xumb3U_JKd*6I}u}A;$VuxSS|Ksa7Km6hMU;fMg z^6CE)@p!m{Jpb{pKK=S{x5V$h|5GWuno(>V{fVtAmYB=L_0vzkErR;De^bWw<^LCG z+StBXLwB`NwT{d5t| z3xkTSdFlFpprKuJ_l;(daVK5){qHYx-~|s{rOUiYLaTY|Oe(amzqu&v-+~JJ@~2v2 znUg>L{-TtNFl;j}6%EAHUvGt2*8g&`l(#8(OWKC9%+P;d=z{&`r(ZnW{g*${R`yT7 zh!^kzefsGa_Z9kjGj4)BpJsKlk5%qK69o(@(B0zyIkEcbE74 z(+2;;2FD72yTSJr{%(V(3jbd}(WaaV{QXaxIxXN27m%O)@d9#z|9pAZukdaEa;eXK zo%sGhUsg?eF%5t9D^`-Fi5JR+U~I1bB6u?1H;K>F<9t6~9LUEwhv$nISyamt|Mwqz z@AQA6s?9ET;FlNsahfhR})~#~=qrseh>_A!RvDq2o-( zt9?r5f~IT@QHC~6OS$2QjpfK(qoiYt6O(n*l>IMYH_D)dK`SI3*{?#;l7djAq$Ow# zA{T_cEbq0}P1aA_?_qkF()_Y%tmi~UV`(-}NK*XLQQQDw5!^beB^^=vf?{9L)MW}c z3}QNy>EByH6HC|%gsu01l0b=&<+(g)nOH;@qlh>v1w>s6ijD;30Hq2U2!xCpc73y5CrJZjkFFv+lw(fO>$7@g00jM4e5#~7W@dW_NMDMlyS_ZGvh-usLB z{M`@ch#1YH5F5qlm(YGQI&Mr?AJcz30s#kqKkMPvl4 z>@L0;dTcF5XyI>(=C@tL7Kf1qn`Ur8Du!W_#5WYtk&6@oq1YUZzl-rD>EQCSEXSOn=jalVq3w_mN9JQlBR{iN$umbVq7dO z$oly7Fr}vwqMN2iiP1H&^2zxwtr;}J=~{-fINkcd|9~2yv32SA#+F%I5BV%k`x)97 z_!|6MyHG$wv^Y0wEM zLxh>4C43r3@2 zVt826Mxmb6Q~H8>3h8Tc4dV?4DZTBnp9}iAp#6e+fjmKBeSrl<$fU7~hJwNr(%b=~_fRljX>;x61t-rBqA2QQ^t-dbLo@0T=xe3|lx#~GG~ zx%)z5yfUh!AYo&y#Q4=M-NHlx7+ynXQfz%>v%yh5E_t6VNTqeC1!WNLk_b`$SV0rc zergMW3(C-sW#)9N1(9XZCa8(+K%`LiMHfTgl{}(cn%}HUT)d!`6tlF(Q`Tb&<;IKK zdgLUb*v3t9}Otfb#~|kV(KYtD_IkDq`0;%iA!7U zv^27Q5y+)oK9hYYTJ|I(RLhRJ{tNE?s+v_DclEPMehSnI2ZmNL&sp_MR-nxZ zR(&S;a13W5PjlN|pn(q$4^Qdi^6)r)cuIG#f05x=b5JwWSP}MH2Q&+B@z3oB!(dpu zKrIWzp3NpY1iaU$fu@}m`}aAmsk;iXGe$rPD>TA- zyNpR~j8imv>%(h9qY?2fGGVoU2S(Vj*J89P@gAJFJ*#%nZs&%;Qgvl*|^ zy1hYOGV6#*W7>)|8WHzstsG)6Q$x&c#_1%DoWx3w5e?-5Xi`OT3cSsTuT~h|0t1(D z;j0_r+Ey@ot^KZFM2$Gz((9l`ZI(3EOC_DxE`l||Y~bjc#f5n;^V}JB1oB~FTX_DE zCbfm#^SoMZE3jn5Wsq@1TOCart~E-wQav}cSFA12R-|qFLYJ1f(Tw{>gXp)?#QG*A zLYT{FP@<#Fo<(-LjO8EqQS0??c*(q)V%X_Iby17a9tG(Qvr9gap? zWS4P73nN7+rLuOm}qH<^6u~TJyGU;Ax{fPT$6FD7U zTdFNX681m)z#J=x^-#m+8K zTnP70lK@1Uw)uvKphKjw=QB*VHqbQdY^H=_YWmI2Ljx=fu(`)*lo((;%m4#NOT3$E z`mG>hv0^BKGcQYGlRr!yGRo z-mi6SuNqV6Os0;qYR|^%46kx*iUT1ss@TN6C=tRu%ONOOpD7joutgZpdi!V zt(4pqG@Gq0meN@YUCR_2ivAi3oyqibK|dGtXDTRx8WGalC`~~Rj}P%;$lO1yAO2Do zQD&~d2JX@5Y{r(>1ezjs>l>jgi``74_dJX=`W&?hflu;$j^@$ls7?5p*{5j#_ha^i zv^1ib*IfxTA_g<|9Bo#bXf#<=F6zjLQY$l*GxBStMuoNp|3lyA;q~EB7gl_^s%cx|TWH{kv`y5tuiVlG+YAEX zjsfjVPtz(a2Q<5d1 z&}7W2s6xGM`gAgH%?D#9?0@oID(QxGW8-VC(cafOV!gf(Z~{_-(M(NDzSM z5gdnb*>fi|e^%xdMWXS()!wiTH+eeM7;5GGTzf4CQ8-1Q`3wn9^%mE4fkg=Hm)!Tu zeWvvK@ioQ|Qguu^Hm-oHFJC;68U zdC$N9ItAr}-$k&e`Qhneh&|qi4{|R+s@=tdb>qLRgj7tlL9TN_<|UDWrp)Xsn2qkZ zMf-z@=(bqABKERWOt#O@K_nO!%}3a}gHfNjya}VXKGD#{oK$U*rp=~ScR2b~lrS6b5=wrbR~t|EyWhP?cV|1;g^GVRU9eyYduO)u7~T;}SkRefmxY)92M z=s8ck{_(O-bq2m@LxH#)g?PV^5-UThs!LVdu66#Ov^hh_Zc3Tn2hry#W-(sBw?IOo zb?jN_ep5iNAD8=;E*j{eXrNMD)EO|%Dl)T%9Oi1I#5!MAS>2$yMxbV1ncDoRvpW|;C+jSndF+d-bZs~5e$wV~K^D%gjIA4Z^a3cW$r{We z2Z*(7<6d=2U@H7V`l_NDLwO1)hZHJr5&9pQjkP zL{Ny+`eul(WQ?pdf^JAdDDhysFuH6On`(eWd#FH3$C6mQK}?YQ_;8VtmvxiMa*Fde zcS$b^4=nkMWM`u|P4&%NLK$2&8Z>Q`xDC}you_R-PAU&pxZ2;FZcDG#W1HjUvDaDQ zTD1$DH1cpko&uofiPK&)291PrLaMGJ!j(Eg#ybLynaO^7yHXJ}Pr z(}s~XT$h$+z@o)wg0dPVL@=^DK9HGFO^H1$Hkig8k$ow{>-ojf*LzrP+G_twrGyI( zn;L93j6AB%UmR-|+Q>o*FVibm!p=EOv&E~;-yT6aGV=&3{}ObLgO07I z{%v&j#n;|u2SJZscj9dtj6Hbqsm8$y^R>nuzm0C@DnR%)Tdg{J-HEr^*y>(cygvOLNtz=4zV{Gvl0Xj)s)MIFp4YA1jBLf=5@0U5>V`wu>9wXPL z*kCg}pBgLxrmkjId#1b!OkKugW}NcNY_`IbpJ$uoFPWaz_%GGSW>41ZLT1CjT0{d> z3~K%Pl}(T3Bg~&l#L3)Qr#&{em=zRVqQ??qW_=AKStAi&!lEnPc5og#c#!%_b#bLA9LDfrxB%Q?U~39B|{>%kg!GD|h45*2-k z^M~y+sn@$LD)H6!5z!KqfK9hq0^fKFpk0Lw?E{xuJf!*gyB2pPmV&7top?PGIS^uK=7elF;LSV4XFw<1F4q~_uN;r?PkPZt%oeS5Lb zKn>+KX@uX-)|T6U|85HUnc08Ff<7CwAMWp$^zw1NjF>4=wJEj_lRw_qg%0(#cpY=) zo5>emewnNF%Kw!yeeruqfRafvU;frrnf2;lURhxq?*hE{@6846)vWrimw(!($=}M4 zS$U#7l05tpQl^TNy;S?|K7M$ZZ5LRO#ig>|%m(~sl9t$FF&5e2Ef(bEZ?R4jFa0KF z=EZE4#f~hsX#NskIxiyD{B_~enFC^ ziO)5$q^jf!Q#ncl5?|L?zb@YzzbKxu8uRe|#}L!U=P9PmUn!dFx0~LY?jNe|lArAK z_MdaBGQxi?BpFz_E&*1DEx&AR0COhRjX@)O1eTqsuq&1`pgu zrya4lwfFTL&glpCq=I&P|3;~x&(!-`(0hAduFPpW;h^`!m`+!FdG_={A8cm#wkr#d zq11f57Pvneq^)4FKuKY>AGgC%w%%>sKaK8C8nb3;NSBGCnJr_PRs4$ZHUW7G z!QH8n>BUp>Mp+w5Y$=qZuVdb>0&Aw}#jmAr6I0XFn5k*2cT}|ko)DUshwv~HPdyPf zmIZwq+#q1Hf~i`0VX~vM`AWX4+3J|J;iLI%5?VTF^Z*T?HBGT}5RO_?ad0MlP)c83 z?#r8;riG&Wk_Ae`cQ8cmB^@-iFW!ORQ+T5)oUBZco8ZD8SIhUzKz(CDNIo=DM=6K$;7YapItCBh9JS`I9%iNSWCo*q(IKHldK4?(R@ zyE`9@Ml0Ghgiie$U&QKQ0|!m*8c43KXL&n-$Z>gnJ48UtZ-%0M3GsNS#euGB^o=vK zuR;fdj>4OiMqg<PQu_IU_9loRrHuM(FfYM%K)sZWp4gi# z392iVj4hL46Jxo-s7lIgXYYD!D^%Q`ylgcF2WMu?oHb#c zUQsh!qcG@%qdw^SWP0?LzUz(&IwN6U)SlIuKVZ+)L>G8+Z})z_^D3zIOU5JFQSA>ppWN zLApY{lH58fxzQC}|2AES@%Z@g@IdKlnc~ZbM~=yjbfT)b9R*naCHBr(?1ZIDK>d8? zcguzl)`>d2xbsqt)MZ($?El3k%{C-?C~yBh-7o2(tjyBm?BbQ91~vZekV3g7wbAPl zWe~>e3u>1ZLde%C)7ZiGl-NL_*9uq8DikknJwmXep$D3Pfw1bfQd zP-6F$%8k+#l-4#=?kVm|C-q(R0^o+^C^5Iu^wd+)I3(S3>;_*&2cZFQn8!`CROr$`jSNEW4I zXc|!FhN31c0m*8=x|bORgT1SJdogNO_j=8Q78dKlER#~5B9vKEv#`Dx?*oX5Bf*Y;D<=oPWtq#ORh;u_jeVBm;b(9N1 z9C1EDn_3!;hyiW1W10?L8WEv+dJD%DTEaaH=FA!mhWz)wg|@T>~1adO%)xw2ZV?krbL8)0zhv zTiO9c9We|t8k1&)8X7gkhFwFNjaw5cm}iZck=E*3&qCt@iI(ko7#*+0;Y6j0Qw6!Q zp*$2#T_%E`jsyPLmHk77=4c^w1u`i-e58kmi_VRY6Ru-OM@6&l)M_$GpG5NEvSoSm z5-{EW2N;0d+p*Ahay1hys*p2YPV(wJajV{bSnnGDyKmOV{^5qNt_kjgA? zpvlgIE~9VBb(-_}(jcR6@ji{tX7sG-X>>N@m}bHNP19(Qv1=OAdP(9#nx3c^6**ey zXvxz!w660R#D4k#WjCW3O{`q^_37Bn8)jC!`M=e+!Oo0hy)pLNn(5B=foj>8+N|+e z0(9By$F&JbsmTC5GCdqZm$UbiJE4fpSbKPWm{V9DAMaU#gPM2onwHocWbHG;>}bE% zU@jG>*%RHMKD73tn<*E&mlB8ywN?p)C>(8RZ4n8v?bD38&743$jDw6jG>n^7fCj{L zo@POB8WC-hJWJCV+Y^D??XZik z@X)vTfQ7R5*qaC$Xwc%hp~)j?99#NzG3;nG8kkxyqgfq{T^8|8c2jD^u{s*HBj*I) zMxL1Qb=qMqKawgXQ+#hQj*(m z^J(Mw&|sD^yS);LJTjCvouH9>c=o&&Hf28>3~8-#tkN=~HB*VB(TEtX%=tb5#Ky>o zIEobcpioD%Z;y`@v=4$TO)(&8Hb27p4YJcIFMif3xe@_b>@&vJnV=0tv2WJ3kW7=p zR*X_)|3#bV$C?_<&mYp%1eq6rr2F~Z2Hi?`1uZDn+PW4R({-<4udk2j89RH0g8Br- zx`}*DH+#Jkm@D)qhDdt}<>#Y?hjdajG|&-!&$82(t{j}^wf&$V zpD^z{(3mp~o%Rxh-a>nHIVEy(ASl-L0qrF)!%p`>qx`Q8gz_8q0qrG(eR&_!-7Io_ zASl*CT2Pk|ZtMfzc=S$X6l6c;dzDd8PxQPp5_P`IE1s7CWFKCHN4$=l0811zysE3>RWgXZ#Y0F zSz!1M2W+Cf#*!->ppz^Zl&~CJ_X1LdQoDMQL!yPIy!tJ(dAt0WwbXk2k3|eNm?Wx?7V|(#NYep za-S=$S3-WCn&^6IegkCM+p=0?Uz|CpG~8(#ZF(b;Om1+E8k!PM!B?ST{+@J ztK5_){~+9O6xYdfe@SMIG>H)(OiR3)1_$yikuSHf2MwR(E1bd%NvosZn;i)rjBA@_ z1(mBjnwal6ajP7pW!x|C)22*~O;clCJo?xs7^KD3r%l`3FNtXhO-CK2Et*#C5n4xy zQ^w_-P<0ffwVhZEcCKlYI7{6*Z5gGNYnxdgNoQzcV+mbL3!FyvF2F9DUlQeHj_W9< z$mgykjWDr<>f27_v_fn<5z;8M9aR%cMY~R{X=>?WKT(tDJF&7-zDeA4R7oopjn2>- zB>IlpNUyT+v9xJ4Oic2i&P=;2Fc=rdNmvrwB^eB|H=S77C8pmnu}_mJQ0GcbB1>Gl z#7Q4mT;W_ZPOF9ic-cv(9g(bdiqf>)ro#F~c#^7mUe94bryyT~EeS5- zXNIlkb*0a~52z%>J|T#EzO;R?s+vGL$vL+Aoa>xm?!D`SAWgtNpqh}ngsAkn?Smjq zKsu=sEm6M!511Zo`XERXoO5IwK-Z>og3;HGUYdY(*fW%;wsW|u&ut$l9r{2bV_RKc zJ6_8O@9CkX>-IbtiuHj-n&RkjkFEv9_JJAb;VM0;Q9CsWyL5qac6~rQ31GvYqnp9z zW?JakIOKJwO|3?dVDnO~#H`Rh-A!xb2k;zz1_s#q#x3$qP3Su`le<2Tp$hN7mDb>F zBoD1F1i{lZlpfVNmd7#&4=0?C8tXtJXjW2}G+C@h3b&WU#`v*!nAidZmqhIkP{3-l zCN3k2TU&)OQYq63iUl-U5>VOz<%p8eS&u@}-rJ!}yA+6KySPd+23wpVD=+qM2V#fA+ zc;pQ$kNpmB&6IOblnS{@G81(Fj9j)m(-*@6b zOg(Xz7WQd`ecxqN8fXs{Nl&AWm**Z=dqtHd-Qar=fXDO`4gvy?H7n;-Y}ggRp(w>{F~7#vZen zr;Dd-O{b1D1I3%C6?RcJqmU!H2<8qYjVO!TneR={1Ip|ul}tSeD83y<0OAgXjx)J- zq}I+v3SG@)a+%rm98zd6(^(ctzq%7gm??@edO%5HjLv9`ECZQoDZx>kLrDCkdl-Zq zFB0=qL@vaL>_Lpg0>`e^HRDd)+; zVO;j}>gJ6L__}tM)=GX5rcMe`J3Uz@rew_|3R1t|yge9e&@I?6R=zn(ZbE*nB3GYj zNa-Da2>A(V?lSe1&{8li){~eoOGlYI3gF{uEjCXDi`_gtK0Z66z9d!)G8w)Vsa)pv z;9W0Uz3@cr_JB^0iKWzHa261)`FVXnO;rA~FkTCS$wtgf`8y*%87|XJ;OD9-8a<)I zli~bl9pSW_L0hU#tErA=i&I0$`}i`%g%Qjn zjX1OXs>T}N0dXBKq=XgdDxQFg(dEX8xyo>w2Ffe=VhxZGILsT;$K24B+)A5V9;QOV zlbUj34c>%{%fp<~^7t^(hdJ_B^V}~qL4z@6GK6{(P-dP|t58G?w?w7{rVw|8Ey_rr zB&38&K?PSSK`6u|&JLN>$n7a2QFzA!RSw%FWg=$Hb~8trl@gRfRT%v_#1Un-&ld3G zvwj&}Oh~5?g8L0A>xkmqzT9ORQkD_r*fkzg<`Ly+bLJi;j3`*pdsuY1IrA14VL|U= zQM)4-fGvFZU1m1O4S+;8ZNOBNkE9mqVIu zK5#@dMzM9Us06cb{Yx&R+BHC1I@&a%5ir24h|tjz@>jqWXzKPYT}J%)gwf`X7EKvjTDe3a zt$oJU-dvFTOk05oc-+w6DO+j&8!1&D+%I;nq2JaZsQXi*p<*h_qpPoK0@$u=yjGvgFx$0Z3GOqQr6{#xs8Vw8NO#>Y8 zK{KiRjgB@e&0?`ZY5HLja1qR#fQ2S?5ay1S%os7Ei>78@9?+UI@-qiCTA~~DN`u~y z*)$seNQonjhDu~oX7)6{av++j^`JweFe}}`YRw+eXeZ;Kz=!4G;pwHS z+-Kse=-p>Moug3aJ^7ANoh>b_m_;QKM@{tEXdZr$fQZyhf*Rrhvw>Y6BQ>6bQElr1 zV52{i{oze3OZO*^Fbn%L#&w!mAsX6XN$Wdf4Snq^Mh`Bs^ms7{UOqf0{)U9}LM)`W539-OG;7?4=MgMHHbMk?wS)663;|Ubiqk&4 zj9{%a4O$>visdtl8eUY^pwCWUst%q~$2T97QpSYS6* z-;T2`5MW+4mON>-Sk>aufSkhicj zTy_WrNH_tn9o!E7>+lo6s*6hiV^$;r%mZlyC!TwQ@GE5C&c@$K_J~B#5$rR~!VPqN zn0o)r`q>sQnb+Q@@LtCeX3v$23BgdAj11+cK`vXOOuQ5H z8Ge+EPfpe6BIxyr=fhd(Dwwl}xi#hFlnkYmZCLS)RN1>KoVk3Rm&8({91K|i4hBop zaoz|md}R(fKPBwXAx}0x;9Sv86=DR?V6LcLp_YlCE4ry7Fpmavg=*3JkKJ4$@6%s$ z3t&F_+0G+NL_W{v7i_hMi9ePPlrL6Fbd@`YZ4S)}(?i98@vyy#$Gld?>2&AfA5B}$ zB-UYmZEtrH1574|T9Sqb$fE;HA8bRU(q0#H^5KfNVC<@&oZk;Hx?#3I#Qe=d36lL<)pS)+QE*p} zOs=1`y)w5zdlh3rBoxyO0RZpzTNpt&v)E&WF?LQ#I0JDIXb~m_;Pv+0{lk<#5|}I) z7x&(3zJkFH0T+fGP3;crbuDb?f77I-jVF&Xwmb3cJK%Jy@zS)(%&Z!_grLK-u`<?Ud&1Ah)9M7`Ep3^TR54qd z)@_>rVCqkA*koAT>NQK#By`_Y<*#w6K@xkwJ5Rn?~H$= zIL3Ah28`94euGUECw}Vmsa{=yMVm8Necd(GakI$ZsSc~z{f#-u-x{M&Yc0BdzgYeT z{(5n!(GgMNEB>_!gFPBx?TmkcrTqMUd6ZZczvK2#BG<^a0gO|bz06%cv%Xb~ z2EYH21;NA=p^-3qPhjFijsesKx)mZI=1=$th)GjIc(r(6>2VdY1;lI~!V!W^e4pMj zbg|+-JdfZi{)C9aVFc%0c&l9Rs-KalK?CT<7VI{F5INC;-3HL97NkWhJ>?a(jl#sf zm8u<7JrMUoc(U2XqD53=twAIMM?6BZ4r)+TdtaGs# z%Zjbh%FWelU98t^g>I_MP;HIs*D$_~lrfHE7mSMK$h}q*TP!J-JhqBl7H7?)we%qd z>c1@01;A3aMI#yDYE`qvk5)~TmR1vJVte#aOb9jhw+z$hucWs&lf8xPbznqYGtSgV zv%hex_=M5sj5q&&dSeB;_tP6I(7m7DSONF`O%%2{WUj07vSv3hGjpv#H9S`RTSFT4 zFQs~HJjmZ-h6Y$8$xRjH0@IHdA)1$msor5>HxhC)EF!d8VK#Q_RKyciv>{$@CgGDV ze5R0{^tIHi1Z^jrG&kkPBaTOGbwz0Ak-_Pc_{THrjHu!b^x6j2%~i*d7>UvJW5GmW zN9Wh3J^D(GSFQ<-M{acQjPqkJ#W4o|73n5_EuG9y#}}dhj_+xs6Go@K1Rub# zIk={IXqoK; zF2Liq#QXeld92FD>6*~!m;#Y0mmnkE3<=aF<$2H{1vk|}wI@Jlghp2sDyWFLQ4Mod z3FHVX$nD&#fOhC-z-u789qQf4Ye!r=(saexn?Z&*Qqv7269#ew9Z^2$IK+XhP2&+0 zK=ApzS33}(MSxEw1{Jf9H}ck%7S|T>Ou+ce5t-Vwi9g<3+n2fg~4W!1~zFI z5LvNW;Zj>35Uj7+qJFJn|1jyc7}#?*sXVg@osDcSUR=LM;lM1X+_@VjwO2#c5HR?4 z{MOiBkThbd+;{uvWJ5jklC26aPsi1)`f?IMl-x0 zt}UB9454BR1eMNCTu?GZdctgpNpuMdj~mRwk76W|>tyVF77Eh`_gS~N7vYgftHvOt zLT0BfV*gi78pZUATkxSQR~KnnShRaPQfX z8tKWL#R-yq2T|R6KSoff&}y*xl{rX?=sX&GII<^)uu6M{36c6IdzjfDMsPNm1ZivC zwO|Y{!Yx81?!(IJ)Umzp{v_=uQ@6Q(T0t73vJwH_3R%H6Q}>}2;09hsFiNr8wjN<1 zBAoW&pfG;z$z(r)MnE|4z#Q+pyyAYnj0@Yy8I^C^dt&F$%)VF$NO}BX)R*sA6`!Fp zc&U~%f5#%02uC&4Q#suqrP!FYUpF`ve`b4;%rFIJgQplE9V?k9l;ZR?CE9+^e_!u(7;*=|8WV+a*-a|Y~^@wO4VWwYQbfj z@k9%)L2h~qMP35Qhd|h9V!b-6+U%jEMjJvjvufvm?70B87dWamPb8?IUba?iy(*No zO`twzCC=JrH54~7k_}Z$5Dm0U3aKR`J1R!dr6m>>F^#PD?OGcg@Mfu*xvaAgt4f>I zY<30DhW+YrKVPib`O>}O3PoiehqBBfSCf-_AzalgjN4<)} zs*bn%+u~QJtQNJ71D1{ry$=r)rSy{TpQ>;;vk>%+%|6U2wRc%aSKin^o6Q|q#uva> zG^v6ewUM#8SuErT6K0TYF6Jm0zSQ!g;S_WmjL%>Ji7eqHaUd41swvT27YvbLR*Srq$%DDscOERjw1nVx&k>svH%KD~i2DrYcKj z5%+R(USyrk^O>E8RYbhPtHCVVG1t|hB9DxSiNA{YP5_5$7u9M+foBKTZ z)Qm10NPd)o8)Bk^Fdo*$h}{lMwIgMgg8_grssaa_P!USl1j2)Y6ju;36R4o-I%qtc zEfA53S!ch}M>c@a-FD9Ciu)HzktICzY+H~~XX2_s;Ls|T!$;p{Q=KeGv+B-Coj%c? zKv`KonrA2~RNO#2qllErQ>w7h!aWw1wAk2OR0>C{=^3WZdtov+%8@O6qFSr7rYOrFvTdrP7H zObNzC?ng&6Mu@F}2}Vjp3LBgKsATow2Ur&&3`2~%#N1%(5KD`V+R_rT5xa?y zS!O9i!OTC|ZQovzUkSb%eTyLq2*?*C2~*EL1zbcwH~$>4oKF z5=W^@PZog`EhutdrGZV6O*B7V)crg#k-dTJ%wsT)>!xKijlE&)&0=R1+lljBd5s3K zH;273>`h^B2d7o`h>zBhcNxe^xEghyjL7SHqEut+pVNdB$ z@s@7I$fL2v*i&1tVEQHXgl!sPSuwF{x!&xAJBzlCqX{%vkVu3xM|Aja*&_{rWH07`-zz`eV<@5J@AJtq@JsA#7Bff)#j5bCT2yD=FU@=kC(bw47VOM z5hevFF>DthOq*3kGESwPgVs%@T)Ro&8Q}8T zl@~DzvIr^EuRr5Wi1>|K&@k>39e5KXo*sZar;3EzWBrMTt3`4RQsog}cd$S_V9_Eu z>SvFsK+RI5ju*XwUie!j$rzjsp7cij8W&6Y=p5>*+8T6n*Y|SBkS8N27~8+ar2oh< zGOLhI68by9E*kbBJfx@RvN}cU-52?Hcj@77x{r5Xef<~wRr=cYiI?=sm!eC3cmEP! zsj)~TwEsKX->iRw{pHRQK%%u-{;s3H3x{&y?^6E~w;YVGwyhMnwuq{rsLZCd zgDYm{My!XtF9Sq4AVuOhgNFJ6Lq3oROQuB>c$6r?~jx>WA*^U|7n9 zaXeNlhh-8XeVKczDt6`TIT>Hk+shOKc)#(koHQ$}N~#RJY`0=1u}P|?nS%%u!mn(w zR7=j_P~2!PRF0z!1#2Ko!3h(!lC^d)2-aFRD_6LhzdZfa8}VwB>d0Yu5iMLNPW|#u zsL~a9eBS4W*Rsp~fj&gO<8gavm5y-zBywfJZg_=sn_jJ_*+oR8C#C1K)NwNHfhNtCzG36T(q8gey<$+s|UB57u@_;3R;@KZ!V~-s{h+IQJOyKUuuiqEkT~>GIFm) zWpmyus{SqCySWb!uTLL00sQa~>ckOsTqhm*rW{OVm;x{rF@fedm2)1t*PMXl%`~dc?NG86<=blmQPb6qGxn_?aVY}SREu0g_g=^4nGeyIM^;Mi zOBp>ocwoxTY`U{4y}=w*3w6P|I(gfz#bU;+a>Kwjn@(k4TVQ+s1p;{D)|q=%ZFW7*=={J0 zUSP~PQk8d6Tr(zr9$^F_pd3vh$120+`^+Fi1Vf6 z0RLU(-vs}%U+q5KC#Pa&Wf4d0>T9YQJ&W;}*S<>tSX>*k#(gZ?$VJ}q2FW4iburRxJ~ z_O}0BPq!GlJw+H1MVa-2*I;d#w#3=JqAhdx4U?Gp*Q(p=bRaPMR$2B~XBd|U z9TuH(xJ>A;=>-kGqFSXw-V_4cHS7D$;&H+Ca*dO)sD@K5K!Oon4Ir2nmw#tS1%1td z^F@n5m`xdmtwrjh(_$PcQ_UFG8EzM*)$FEx241uhlY?n5VGP%$NrSV(gowfQy!kO~ zksS8%6GT8#&&TNCS*0^MxE}eMI6r%ehGV4&p14iZ+Lv)cYkUBwiCT#-W~-E0>cZ7qM44 zs?DO`Td?R!|Cf-k7irr2ixDV7uI&9i6OkeyoN2*s1286XrUm+!O)=f>t1B<va|%TfiCOi1)EF+a1EY~nMFC8E3D8l~ZXSpR z>Q_Y1_y+*{NdNqf=)RRx8KM!#+45(|zc%?L%WPRJv<5*DS7d}DeU_!qLMQT^%0h!Y z+49H@Aoa0Xv*2a!bvjVN_O8fv82p=c3(nI;{y#SIZy_*KsIoFDWDUr(_*3ZQU+O%` z4vblQk2i#x`0G0FLfWl7|KMCdp7(V}rjy8VNMz3!KafsX|H4j1L&+d{HYtl(T?CC8 zTSB+b4Mp2k4&ycMW3)Ip_ksZXQ3l#eKv*L`3BK)`vavE}J@ygmTChyc2rt3-w7`x( zGhDX3q>3;eD|rza%_3K(jzI#+Mdv~W_PWi4w#8+=m;-JoH12G2W8L6v@B}a# z9Vw?3#KloMW0o8Rhak5m^#>Cl4m&~BfKgsTRzN+l&!IWVO_M;y>zPd z40c;C<}pn6O<7coHXbK8Huqz!gvd2`cQme05J-_#ZnD8}4JQBa{p)2acsBWWTaSb> z`ml_{T#c*+5-?L37z(>E9cS88-91dlnfAh{J#7B9ni}=YnCK{F{A3MQcs7{W@2Ron zD#OgQpwmB(npGPcWs;qBlB+v-f;X<;0w0FMMdjB;8H4RR$p%yPJxzgRP#_r;h%;+@ zcQX;#!PyK1VNrDxm32wXC*!qc_|0ezzYo*C+dDG5Tn7!P+ zFgeMfEVFa8U1UM_d)8u97*Hf)#t8aS?`AQ783QW3acXS1A8f@@`k7%-9#3q(dDh=QM^ z;7jc|M;e?RoFR|F3xp0f<)PDr#DjbG>c0vusWW5|SJ^L7;RQy_qMV;vnD!DN0yl!I zKLD_VpkNxq%dKPwp-yapPc5Lm#L+c~F4UB{hmBj#vrAb{s+3#VlxtDcqgExPhFut= z#QkbfcP{1Cgsp*9b`D|c6J!^33r7cIIKP}@7bZZc(56D(|IN~}h@F9_U6{YNdm*R@ zVk8H64S@-8WYt`{*nQ|OxcC=NE7on!3`_EAlJ`k;C$APJ%F>10Qd#z&gPO(aLDFu3 z=tOW8VcH9p-U^eAuQgnQBftr2RF7%OShFU?Y)z9xY>W69m37QiaM@HHy~uO9-q$P9 zKD)3wo57OjPUl)Q}VdH-z=zV6SZ1cc%=Xd>%3QB5ftG~gy>y*;`@%h4}r%6vBz(SH7&d^ni#pr&%s<0FsILaH96CCzZIdgQVm|zf@PabwUlL?*Bi&!&|Q4&d@BUQHWtni!?h0hB00peEM7tu z;oNiegOROr!7X9lzv^`L^q|9`nc>k@M%Xq%;kn_s11E%651R$CfDdKx$@B*=Za9L@4%xt zHo~^l!vZtf$aVCu=s!Xv*!(WTI_PAmFnbkjo63|+b-89`Qw?Fuq<xhfA#Y)h+dv z9m11X-R!;zzhuD7**m3tEmJ*ez&+kr;E>AKu`X zi=xP>muU4VxDqJ)f{}0p(xjNcSd6b})s;qY z!r&cb)0K_5BMf~x=x-h&lk4y?CPt{ow1+9|b$@i&v?3O;`|8gh4GY&X-NIgt8MWC+ zoTZAcr!nlCz|7VO5fi;CJjRv)=cjPL8UujMz#oocA{O9SD01FiWqrx z1v#42uh3ZVUX5~n_c_HOEM{L|oc zErS6tZW??I{z3?66pe7xY&}k(?)=2qPKF0~W%jre!mWKkr#tCNAJA1EpX>wL>4aPQ zfSKJ8yjp7&+rFF=NB=Ykm~2*KYdFgyVIJX{LDn@QGci__$Go6W zF_h01CER3%3;whuHy(4MUY(_bM8; zfWN8>P?a^6-@qvAHFvTh$e9O^EhCH-yC_KKZ^*i_$MQZVrxtXoMp5*bey_H=ud3!3 z5$wQAv@&^k%xCq*6nl~sT5A-l0 zavC1!VM63AJkrDNr1DvKG*uBIdvLwllEq*(7G=;(gvbfFJF_6Vw*)iHy~mJTz6qvs z`-@IL_qp|=*jO}6tiDqBg7IIcz}-oPsQBIbI}YJ;y9_$ha< z8on}43lh?j5AodM&#K#KBTPV87@rm2=9kIgvBkHmF_BQO!Ib{H;EUmNte>q5h|h-8 z9P792AVZJC^m|E)h+qD{eJ)iNjE$i9w6XOVx8 zCi*Dd3^<0QhB#fzAgqx-XmGV-s3qKlLmyB>u$B&RS#C?4tF@%W9W+gpuM*Ra`P^sGD$pmEC|)wgFbLiELWiddP3UeL2iHmS{wk&1h= zVYaqdv>~^!Ew06gz1GAh#j^`Rp3E{*U7!N6S`J>L0C5ay2fLZj;>lq_E;b++_`>}G zU}Pf}b}-saJh3o~7`2J*Vt4f_M10Z5h|V*momJaf3SxL!R0w(!K$yq|q~|d96rQ*> zEClEpiyLYm0;6ulx(nRG8!@*B)^E2(3;Q}oJBjU0Y|9=daX~y54mCTAX?T&X^{B&b z4j3ck>=C07R_-B849+8MEQ z1E$Mim?&FLZo#Nel<5Q5*6+rtj*ws+rc-o}|5dKNcz z;(buK+mWwunHAO@Tjpf*CldhXA%c5e+r6;U0Jd)&Y!wSu5OgbczrYwu!Qg`>R9k`z z!o!IH#|AEsPJ0Qw?z69BK!U0bdwXuLJQO`^ zaM`^!+OAr+O^QV@iv?GNCP{c+y_+eM1PFzZxC={edXqsUT&$lRW7}db4CU!)i(9<# z&forwOf3v|dLV48Sc=pKG(0|?fa4bSDONTktLo1gVU4nHVFXnj&IV)V?wseoJ(fs_ zgg3)E?-(OD!jnH?ep~#oi6xF#`rZ(39k2AgA>4ZHs5;EARZ!O*nEC{KxP%fP+2Mic zfXy+yfrAf-1wOn!z%;8bru9Y0qa#dR3XgF^*g?8=Q*OIc*gO;4FK}q*jWD@s#kn-z zK)W5&a};RVUGeXBFuIYayV&8T+!NlOVl+rRGKjOsj7Y}!BfD4(LkawOYqOf7YT&M5 z7l*7Lz^oWf!g$`#hDdQ@^Vq|Au|BS! zZ=SW`12{PtdIsT@g6}fia@fPs!H!}Zs<6)x&IqqVST-=A&CK&6!t)SDITo`Y&Bmjn z{}|{-Shjc#Y#)PF!RISiMMyy=da&AzFvmL?jt&k|Q2k(+Bsh2-!!|Wa6q*oWcpkwK zh1r^5)NLE>;h{TjxBV=xF93j<6R z#77t%CCXm<>lm*Z)epA?y-AU+YQgH|4GUtzy`Vc8$fJtsT7Tg=U~7-nrj;|8x@1);{#(Xq-m`MDO#}wv8>pNSZc6sw+Q!FSy0Tcid}MP!1{vC8^)|%dW;l( z-FW=6p21ecs?h7oTJ?69UCi%X9AFmC&@o@QsTX#|2?k%_5y$(Ds_n)uf8hCM<{e56 zW?5xSS}{0}V1=-@swViryHpR|bmt?IprUrBjJt#Cy|Mu@n|?V66S$ylSC2UYa@UDe z$8VqTzA{ef7n9bZ=#w{LmktZ?%T2r&t%f#Giw{PJ$%=@f@J45tOdWm!9XxV^(GWsA-F*})r$>h*bKbl_|$4;;&e_%)TkWmP&*7P zp$4((s--9QzOdaVsL$$)i4U-{B(yC@)@)R_d~iW)3v*cuKCGrQj4L$P`mZbaFD)$k znA{v`+{p5k*wvV*Aqedwgy*g&OCTxh;2$3|j&RI4@~5!5YZaCliC7RSV+i}p0ya95uPM6#p1~mQv!;2aRqhMQr2umlAA*+*ZouA`p?a$grnJ%WW1&z?U;dPTq8- z$j>_~lp>37DBme(ym-$i<}RzrK3-!K>SF2(nMG&YT_ce?1qacHh-`Pzx)0&er)dbCY%%RL`Q{drf_>|5YH_Rj2hh0|(`}u+57BL%yx(Fv zC(H?E26^KC>EZQh;qRJJc*DqmVzu$@#Q&C_7hb=m=liau^F4no;dI>d=UePn=dZVJ zzs1dA)sxVm#q=hHHfSB)*2()&Klsn7F`Zdta^PKOHfx+48R@l*jOR$NH+<@}hb_Nu zq^b{>Pn|AlTULF3l{SIqQ}46Zk+Adq|7G((7`$$6}1+ z;Nk39y9d)=f`tU4RTZ|A;g%Sg@X-v`ep#aSmNX$-6lIbf`^CYmDnC}&kYjTe4#MyX z23K^k*~8=#atAvu*JoELc~VO~1i5e#uPzzfA6fm^$}Ha2EON(1$+UR?dcUUUkDFcd zyj?BvQ7@ zNYc^1iio6aVG-RE8`R^i-A*E#MQCQJm^ncRFUppwfhSvWKDcK53e5YT7JYLx_6*_V z;5HT^8;w^T8k2(;7(NCk7*-tnfqirtIG8z^$Lh)qSM3-KX2nC=XZ4%!G8$(WlhYaK z{v#|Da>K;6(#UJjGEMDd(VMoFFLD5jD_>816c$Z&pmGxTjqj!rP>>{O4bWUnmCLGMpoMBxR77ruI7fjM#6nGzn$3{c}>FaS4{ z9>aVY1O^K*RB#1`^=bZcj4I_--^OYzXp22O1B;3lnT%h`Q=2$K-H|5nqEd|(V(BMQ zYjp&qY(Z5yP4D>rU_1ob8tB#Xd^L`k4Mh+qhROr7FJ^&Ag~ao@U+4@M1$-J*8a<&2+Y=3SKs$?FtV1oc^4PmIXAkq>Q&j%Z^8tT>i(d( zWwEijPO^Mqvtww+-8QC75XzR)u%ko0tA7r5`c&o({lx16!>}2WLwE(>jMds3HcSGn zfu0ldOaqUn8TvddI`Gx@?fbELPoIfJ;;Yr_o1()0I0Ne!ev7fUZK8Q;t}6zfBsNIR z3w_S&NWd+5=;Q|-l38crYMS!2u-4JdwV8R^5MOWf?$=1!9UzZRpTM2D@2 z;k6eQgz)kxT?5kT^@YuiMTsqARj}XP&71xQc9U09#a1J(q8Yt zuOL5}H(O1es%`gogn&PW9>YTFgi*^ySm-+ZMLl` z$nzV((c1ArAzOC1uC$L;I0@P_O7<6LIba|*jUGj)UCoCc7KbiPR$olKkFP1aZ@b#J zl`xXA4j|C<2l*xl_qow?^N+4)QN3aouAMVeh=3SSO$|aZV~C%y1L5I3G+K&+Zw|j} zVa5(By>ns4R)S9qe;Ka1EXUaRASYY( z4&$ktF*`CnRSQ7;kVy@K4T~8IE#Gcu}Hyh|<|-)x!O z&6eAd1Z$<+pB`i_md=m|{V#9b)58rR1J6MFEv6CCL+4sdJ5Ang zvBgBP`%&Kg*(3M!MecC*gJU8bU)dZ^zPNAux7=Bq+=^qys4fThV8cb#TGbwVzLr@b?nC>Ox*lUo+1O!_Z393^Y7W;@rF;bGKwI}9K7ysZSab<;F+oTz zl(Km#43S;H6BSBX0a4MsJ!=1-K+1HY~eJwo8PO?kTp4P&) zciE0NT**XT!(l@aJL#ONP1OUNQNA-&5BI-qL^6L*s~MGzBMkgU4V13v@xSu}ebgl4 z={>29U(Xt2Y^~w8xE5>gf!v89SGBf4^KrYJVygLhkTQp}aeY&3b%=@{&bo0^iu-}0L=ahu~dvH0(iZHmZ*T3)F!lKVJ%=@Lq5`$PZp}QzA!Z6eA z?yHPjUiIfsA=NvwRp;r1NZgV{zhtL)`87tn2ly9SR?wbv?1Q zLp2FWvoTh;b};6Vhwqn%`}EA?#gvG}1US?nMe(>{Y?i@SF)6?9@OcrmJd#Roo#l!M z8)hm^WJ5taCtxHFVLu^MQ_wslP2CZ)c z;trGXCQq3W(_^%s=!+>P^a9$tI`@G=%5sfKvN9$p4;RUsO-W`wWnM50X6C4quW%T` z3E?%^wk*+Ut343Wy5R=Q0Q>oiz*T6r_05p5U;=i9mQ2xao51%MoyqSMmz`@JJD3Dn zAfbmd!i!o&9A3x)4)t(X%rdBTCRLq;N?hVk<)O7MK2!s4nUpPY4%*(bdGZ+ybY&2J z?r5>ClT=H7Z*d5dPe|b8F>Ha9LzsL5tFK5Z zeaselWwo&p34YJJ!n|rJe3zV}IcvO!<)GW|&+x=UQuWa+OO>j9@0ZIg^-Lc>>RpFb znX&m(i{Nl#>|o4O&V|aut{ddbcZ8o&YWm88 z)WEWu>S5ACtJh#K-d=~W-ZAH4a^IFi7?c~r`l+M_Cj1WG3rrAhiP5Ya58GUtZN&&w z-u-(2l9HE;$l}C0MCFVDG6XZ>!$am4c}AG}1Xz06h5eUof)#`bvO>XMcM0{QE{lUj zw6em4$Wj+)XMtCNV@?j{?>H{oW$p~qk_@RZ8v+%pa7K8p-_#1*J$18O1DWV``%YDs zgzcMMa|hQr9>Uzoi@}6Qw*8&%2!{tsv9O!3XV5p7%{QO>d7$hGb|3vrNYLE3Gb^?uUrbsXov>f#bpsq!Mx|>Tpn_0 za#=m*JU$%W+Z!To5ysm_L(nT8xlORCBSJ)^GBEi?Uf}6sDlzzd;Spx>X1=Z?T$Tlr z%@Z?gg&@7}F`tusF&8)LCA~U%TQsrGfBYm{MTs$TsL$FIw*?r*R%SE(4`)a56v0S2Q+Fw!O`bkgSRu^F&u z91D7FCwa7DFAh71jKCP8&=0T^0f!zVY;WkI2>Wr8X6*YgP27JWyf3!84R>W_z3P=Q z&~$mlGB-5H_PSiPr3u>g|| zu>j0a)-IQwOS0UtLQXZ#!FIHO-3%FnEQIgKn7triRfZYqAm_mg`i!)b^L$mZdHj}B zZ)0QT;B#V@b1;Z?o|*|b)RF~8f=&-&Tw{m?CJDvJqj@Lm8d@>INrKk9FY-5c>FI8| zf4uwZYumUjlHz$$)vJp$&09=1QTc#}aB!QaHu?kuX`5EI?}qB#y(O4CmHUxx_?105 zu^1Z==PL8n}E1orwm`Tu5(5)1$2Kp56Q{PMU96Ih(01~H$~eo zb~I9~$+$~PBU&TX$_=BfXKDSygL4w)mpfX{UfyU#+@ncJ4O^=&8WBhQzt$%-?38mI z<@$t1#FnDpRsKjnks}^=l=t)-8V!ikkand{=%7<>8a!ZQ9d^o9T6;KP zuTy$j!+#nP0XaQ*=wzo{7ut6;vt7<6uVJ5G=_)@T)~W`t(1Q4qUnk3%7D zLJZBMi8zFVD_#Y%T}_BKA!2Ke@NUTfYkUjdoWn{@2Jc!N;WW}Pln?x(n^tci;r+Nv|1Ka_S-fl~H_oc+KnI{g@DJ6i~=Kgd~K4ErTin?dwB z)X4jDA^7)Q+9NtV;ZfCQt4O_-aE?wROEkKqmpCuP0ZXh8ltU;J$|^16b67f5>hnF0 zVXg)*6Rpn>UoA@JZGa=3fZ(H*(A%wTHNu^p)!I`Xz7FxML9y0;YH@Ghn!-jL?p?AB z7#?&i0hb*)5<);t>%+S}4oApl*Fw>>p3vjD#gzky(=Lmq4xe{$X7<#uBzAZ>gIg>s zwJce0!polZi1L{EDjr}-3fFOUvP(Q4csS`4VJvCz3GvORVGS%i7hJ?}<{3c4Vael5 zVjhXYmTuM+ff-q%za5-*Gu&iqO$&L2oo-jyTyfjl>nb!*;NJ4;6gYVeeocWt1XkeB zjRGeYPH*Y)KB?xi8ucIQVJL`c6_qVPeyCYpYGCGB6J*VYzd`2WTjg__7+039|Y*HAYxmTHO%D%o0WHT4sM7hsJcCu*V6z^l7i4R-i=*4Cvv2 zJ}d~!8_+#jH!t-G>&^p1x=cKFcyjd#SLuz#Or7ReH8bpGH7l**QcxQ|@V(rc*8L2j z>ZqL>d@m2^O-#o&WcyjP>E*PZq*FWnHofduhXPI=LSLnzal$cu>I<^D$0xMYvM$Mt zJFOnWbYkp!AEeLqeov>qbFlZ5`a!<;H+RR=$r9WwHto`9PnRK1ss@uJ;Bz|9U!%*0 z0iE~T-e1Jw^1YvozDhxyARd_S{kTL=m*iD~Cqmb~Il^t%X)w1mbp1)+xZois+@^nn zUsnxn=O8`w^r@Vuj%6x!v6@s?y>Xuy*D{ySUan0b5$LJ8e?xuQk2R0T(^&!iml`m+LLk?{Zw2>43J-$@agu+M6&h^nL{gB=V?0ymb0kiLaBF$nvu7$BTJo?qK&R9dqQgnBwb$y3#935uSSV%u`c6IMf}tX9sYpwhB6*l?Vrt=i*C ziH@>W+I+uK&{k^;fr{Ljxv|y?nJjIT$t#tuX+=DZ#1aAnH#a4hS2s$Fs&+?BTvdQm z7zHWipgt-k%MGg#^vUQBrCM>g7;UTJrN9W3x%@Y&*B>Tkz@)Xg*y8p0Z-6OLKuDd$S^Ds$S3vl#&eNLj4t`5LXYzg-9|`CZq89t}nKcqSmpE z3+Z^Em@W~|n?=e-%BB0+%GQ6>?zWc6NL7kXuUcubAqpuv$1rP!M5MA-CT;jxTd)l||C3UrvJa1sQ;N&S$M7L0vgIyw9nIY}(VP#m zZOoAM88c=L(&F3`I}cE+7JTh-)DB(-kJ9n8HlhwoG2;Q89Ph24F4-^i_?PUzWAHW_ zzitGL8fOrltjuAxYH-EZ@Cs-4A!^z*lBZJ;z2)IFfM`Sctp^7azmN^i*DgiwnUi$%b>x=_aVq;cb+qTj!C2906YFgBkq}823Ip`NFI3qn5Dq z!anZQypK-4@ap?)ufXvOufC5?zHkrM_t~M2;}>3iU%LtFS_60AM<-u+^?lS5c3xP^ zKWK1cuI;|?>ie)pM=!kkK05ir13aorEHAh=Mh@RcX9wlA_rV=5?`t64Dnff>ve|FC z`@*a5Yop1`+T(YBl(Q;UeNc^tE0jpR`WHWFXI`}~HaYRsnMVI@J4-e^anP4E9$vSjz z)n=oS<=t+*E%mVRmo-rQuslEk5tw{Y(w*v~a=wDp)bPy}6dvuc@O%X+9I|k)S^}9YkO+hA+vuQ6;rI6s_vy(=Dx7|w;3B9^ zE>$$MsrkuNC4Z!DmMk&^i^i8&_yogPcGl^$GSq{?YjMU3JgqfyRL`$3#akNpo>OQ2ilZB&e5rm_{R)p9IWyxcqQmDkkLY}Y2_`eK`)z2GY7{09?I+RYDpl{2VK++2nB zFZrcZgPoO(CN(~QOT4ZY=iqq+Q$Nfnw6)lf*@5}AbuFlhFzbIa!r0a_1k_xzP*Vec4_khh@_K)Ua3P>|n6_tSD5Y zae?J$fV15zT)b233o89m(E%I)ZoS8I@Zi&Zw3kq; zU)zVU{ojOsQpCzqyv62Ou=DT;^Br@HHM^ww(|#0=A{IMY5aGUEZ;3Rcv2lP1EU7 zN93Y#wK!ozxZW{p>`KV|KpBI_8a(kcA;e34Jt?K$(;ls1^3FLyie#9sEy@k}G%52$<$HKdruaLzSuv2aIUUQ#{V66fH3 zouO)Xnw1;b?3L{*NwueHncT#s@tnm3Dof$UADzCW;(^QHSqi=JXv%<(se@f5zs9zc zADW1*J!+ed-eZr>7|pnD!R#3=b}ZPxY=2tYvz+Dtns(>KBL~;R?(^<@AA>#fDcpn; z_L)!K=3|)pTBA3u4&QniNB0MT6T&G?-?JUqfjKb1+~_y4^Sm>gr#`{I99QY!w2Wcy zG(a32!Savm7zUL^5(3dc6}Nq2*t8fQ_z=6+thPE~n=PyN$81;q*_3UZo7^cr2#RbE z*idUe2q1RkYEP$GGZn{-;d0?&*278`c;P^t+Uhdbv1*@{)d}Jhk=$C`qS}F1vH%v7 zWU&F8XHyP@2*4;l6KYF<;qqMGoWY3g6ZbH*28An{wIh2fCJ-6ezFvl!4q@y+j3vrE zyMe{BOM9XG00rRmmhC?0GKaw1RS&m4pu#=}oo8}TnC(8a;4+H&Vl`uU)|{z<^%fTw z%^I5)D(3u6DF*3j`~K7Q1_hmI!=OX$r!d1K4oUXb}zJ3R6K%9s9Lt17HYu>KqD-)b1w z@KkMZ<>3Mk(LBt_Yom||1%xTZHGPC$fbXlQc8_`@TD^2L(Mw7VE9i@=8X;HMiTS{R=q15dk5319 zjjK0`JT-Wu=xq#(*&zyh(j8B1dlql@3Vxa6FTT{v^7Jr&czR@Yq~|!$d$z4q4MnAK z((eMo`GD-0W5S4E)D_>>3|4Wrb$^dt-PO$vS|=Ur$tc>J)RT58X+W9Vf|QiIOnCr! zeqhs#v^ErN&}5XQ$(&TMdxiiXgD(+hHSwb0+jGqzbbg||c5$zg)@Kg)Fv*W~qhQ!; zwx{v6uq9J#f{_|2Ff=SmZR>+gB962}iAEVE(QfXRk_MEyEhrrcZ!h4BZ#~S|j^ru! zmFu2@0!9s~1$xdv6ZNIDZaP_-w@Pj_q!wR}YDwAPW0f7<)HSyh@DiM$T?%@bF5d3; zk;+HhZ^Mzf0M7-~>9tH9p_^Y6?IvuA-%U_3c(57;U!m<6JA6IFsU@Ti?=9*PPJ0=e7F@;YC_}k^`WtZ4zu|g~oY?lUTgE z3*#f(7OeV2cFvD}zDt_sgz+If%;|o4e0Y+$Q)_R5k~50lZs_ab1zWerADCd&&9Go# zT&LI-#$>ZbJ2mu_&`q97Nz{|T_G~F>mjWM*RFZis1Ij$6On5**Nb4;W`Eqimp0pfM z!kCiaR=Qm?`wS)H%_4553Z?>2P{5{@>NlpUtUyn+L9njz>?tn$v*fw1PI0d|w%EWF?fsahg#li%_!P4sJBg32=uQKJ%UqW2 zMl7+)@}4HfylPudP~4u($|Ah6p)w?Pl+;n&caeBOaz=_5t9^?5 zgefmn2Bd@#rj8PN3IXXoEShM{BE3m<*{8S=z*Ka9>xKbEHF{6+!HshC1--ZWmNsx< zCzP64pHkFbbEZwF)26kh#D)UaUB6xDJGb!z>KOFb$V2NdOSOQD`*2M?6_ZLnxSp@U3mK#?|f6jGf5 zd%BsD2b3_RBrHfr^`w$=%Md0{dYhJmYxhQq4|f7V>v!}X^(6VR2nD1BunCmnFY@SA z;f-3(n^|XAv_m0BiXDq$%Odv!3hicUU1rp214r?q(NK86;LEYR0* zh_6j}k|s#&&*CKghZ(#O#13BuIPGU(4cU0FDArZ%l}G)Oq8}F$An?wCj-x06-`18g z*IGs-2MBcQY8_y!;9~dB6TZqmhG6kDbMc#x%(zc{$IS7x7iWXb+OOwmgo+9q1kCYK zbM-sNMZ%Lo0Q(JAs=4Y8k||*rtXOz-Y|j4G0;uh z9u8Mva&6x`_UMm5BCH)tP8^Pp;2y$sV+(d0KsUBv7p5Cqu-gE-u?4#gpc`AT+W@+; z1-lKP>n(^Ux&U(a_Bo+NFIfH-`K*vUc0%4-P-e8TiEs58(LWY z5=`FECc)&9E7+tnKi`100_G=GTPbt?lkJVw%@7czp3zaN-ZsYA(ZrbDWjC=6r2xks z9-h+k5@T%K6du{L&}c+#Y2dGePJn>-Vs@VqB_Ggw;|3y=7KNAwDl($J7ApvX4B8J$ zvUwH{J*x-l!hK8gtajek1Jn}ZK;#CaXKF_cwa-dSdG2ZMSw8co%>V`IudL-rK*l4FW6+t|ye=#+uyIwiB=|EiTO@vhi@+RNSy;yTGue#ib3` zQV>d;cg1R50ngd$D4HjvHMA5^jyi{*=<|S{{C%RU0JT6$zuL6P#pI0cY?I0E zyjtA_ewF}P9JbwrgneLi7pIH5F&J+lKSN)A$R|lcbLudOFH-1BPv6%1_7Z#_Fg%kyFFS@oq z7}G_-p9X)vZa&=%ZQAl`UreI&uBJB(d&Tts!9M+kfYg7pRT4i0JyS zDW;C5D;fUFh>n{ zm>V9}_O+kNY-l|G7P%D!iEH1fmvwHVt~4Ixr+ z>{m}C)1?8#@ET1@w#7z*qnyoVC3oJcR=c~bP^(QrE~}d{J%qM38g_F&k3C2NSi&-V zPz#o6OnF}Lh=+~rUxP4g5`v%YREyN6@hzzk&6(ZAU+p?@21HmOa zAV#-xd&Y8u6tvHR#o=~F5aV1Wn&l<5jr7X`l!?I%#h4hV;pc^-4OTYs$q^1J+p&Ou z*Ei>yEz%OWMc44T#eI$THs=~1_VB?f_K*dKxE-Pdg2$Ib99!IO@#G9`ix!9ZU}eRr zB^dq?c8(U!?tw}mN05e1vL<4WxR!~FWQsTY3r5<5v^HQ`Tn3Edw3l#*x7UWK7{#V0 zdw6b!ET{@<2@W@hy}?zNXp7xmn0*$^j-I>@v%B}A1j=VuF}M$84mf!x?$e-nUC096 zfil1~kF9R83>l10TUfOxbVK)Y1=m@5g)Q(+lMirE{svvL^-@ONl2^a)0M{s3 zjg$&LJQd}WKFs%bul7xi?OB+`EH;pe%!%7c)sB)+kQVz9fJSMk_&pk%&n(8pblE*|moe>#n=@au9n+KQ$;+h5e0X91fXA^gJW^=L- z$8RWA3)VRc;48xtdMU`+t{nIDdS!8m{w?TarFJ6}Hq9R$nhuyi4NuZlEDbUCvB!z* zWec|U*lHL7TAFC#i>1e0$iQQyvgg#T$mI;LSB1$l!(yduwkS$+TCzRIr|IkiUR$wx zr$z03?!ryyz;=idmo$g&>Eh-*FORVVy{C*n*&Gfy^%jh9ok9ak}_{CEJ=pD>0?^l*OMwvEy3g5 zMJH@zI>>O7jZA|Kx7x_fHrP{c!afeVUw;Q~!U&&a>Nmo%n{dp5ZX4`vZbQXqU!!X* zIpzSJWC7yC7GucoSP0s4n@+bUf1gX8j7GTXme+@er+l$Awq5YD|9yLD`Ro@GMmWqR zlNdZsI|sOesa+Iwqj-b|kCPkL!G9F(1THt2sn^$DfO`!WJ*7B#J6IP)_VDBs8?=yG z!VZ3w1MR-EJr=Zh*J5f3J1p?7!+-G#QAUp9q z4!T?~+vcB&5##$KiQhNx;>~)fw5Y+oZ>>fO;%L;2F4OjS@c0TiwFG$K9uBLUCRP}z z(=}qoRjlF=w?%Ir{BiLaY!SO#+aKb!e2e*jA6{|vrdBL}h`Z0L7X8-ZtH-An7hhcn zc-_hf$FDMMcy4gK?`jiey!EBSFM{Jgfv_DA2vDO?zcPs#TDzmgmlnSkf)2+xI<(ub zIf&65^cdiCgFC}-i?KW{sC-uOxr5nBg_=4l^<)j@Ocou=U?ADD%K{Ui>PLS~uu;Q^ zy7p9J3GOn-F?AkM>(+yw93ZnQ$@L1A4`kJ3XD6gXs<$Ds1jRjBw&cneU+Qa@L+a{I z;%}R%l$kx%MHZ8|$4sS$?85P~{45`{6l&_J+tqTFygG76&Gjuc_`D!B&8)H;GAm*o zkO^)xkm-Xraw4o>LaItsi>T7M?p^|3llSWk<3G?n9_KnsJWX{Ou$Wm-JP6q;kS|$99u~=x%cr@PoorOou(&q=c3b{5+jHA( zUoqI<`npR&Tio$!`|49M$=4hH&LOcGU%%DJrqPtA*QdwjA>My@mL&-M>@0*KrfrDG zKD{NxXIVAehW@KeiW(x23%IP0C3LtbEZ;8-wqZnKvHFyRkyraSw%B3ogc~ti}Tdfnq!NK~;*F=D1hGW%(Ni^lEY%ljd zDZFskEE~)RjW2UxRcd+riddSfR@}a{kQO!tvVCl+zGBOA|9tjMj02g#d?C!(N0O9xE=%{x#zZ(@RO3nML*bREOi4-g`i>5S5C zBeH=?ZwAhaAt&;21%GU*Jt~p3pA@HP9hvxQr{sOW5Q$5gPnq{dr?+X+tuMUdq&IoBY z0!oJo6m}pV&p^pPhPuHp9w9w6&AM*9=z8)D5XpCsvGaIwco8TogDW}*j!@b2n0DcL2us%zA{7opc=u!Q zTsgY21*J9uUzD5Px=(pK&d7aOje+{yp)D}$>(v&p^?5Z08u+9wK!`{SuGs>rcVPY} zMC9M(V5ruCEkJ_w^Z%|bu#Rj2>WzW7*#guX18=hh%6_3@3@lY`+LyzJT2nUrFCkd} z*5gkxOIe`oOTP?MeElYT;3rrPHeEoX7A(9i{jD9j3RHD3s$Kc=ABTJ@DB&G$&y8A> z8sZsUX%J{lTJ$UNGeWiKEGp4!&AJOh|0xMnSycz5sZG?Al~l;Q+7;_g;thTTkPJy-wbTt)Mg0{C9<&cET{U$ssM5v_F(XM-mPS61UU4c+$)p+C?Bbf)1pA=v%#7X=C5 zS$eZUYbOo)cg>(Rc(QO1n`^o;)+&+J4`*1`D^@GBt|$d%gia!9vEe*{f=6k;|M(En z3(U2ai*S4A9_$p%j;^JfoQ)nr=h4*Q9TTwRoEZj1g_JrD`FK5 zot~;8%Bh~ms_uscl{{>^DT9VuK`X5?VaIG&$P_(quv(|F_(lqAY;-E?;`-574EL1i zifWw80DCQ)?9D^Lo#v>qCqaw{erf#7kFIAa1y@cMMY9xFrXDybRW6_J*uJ&HToE4B zf2-+VIg!{gtZv{`oIXEaz{1L|?~D_YcZFd#nd)6WwLbc9UOB20XZv^aV9TOik%CZN z(duiax!k>&mU8Tlq}D-65MU?PMFtKadD!oV82(On+V$W0g3PK?r)SbL zBjvgyPF254HJItLa)n0KY8DTZFmG9(VjDZDnpqKIbz_kF*`-XVqBiX~qJAcFScjT< zLNM zTXHq?UK(`&`1o`gk2dJ`tqmF_oH0Gxi5nfGn@o?^@*qSEKI5i)%YzU((Sin+nYQfB z$Tl6PnLo`O%BB#3FYM+}Q((Pa;U_k$%ZlM@$Rrm*)F>gxuU7l!WL{v8#xjA}MrHN| zfe?*tLzwcjHZA6t*gm4gJVYM62s^MW{woI5+GnY18Z2Mw* zh;1KBd@?uxHkKP5Pu61#Z8Uw;*h}mk*=yy~&I@iXpLSkwYx#9Q6t0%HqQe{|PI!#@Y&9ExtvnEyArGaIf)(t{cI_2JyY@SM*iCY9NOH$+3 zI%kh&?f~(EjdqSD$t@n}+K8#|ulr#`v?qF>$2&B-ZVS%LB-3n51aX5Re8-`d2@>vW ziP+ER6=}G9KqG9I#&nUk01cfg-~MvXVQSh{9jhZ6I$d0O3g~@#cJ0t6qk&UqV-~$L z&o0)REiE~o@S(ok@{pCN7Y$h|EC_mY;o+I?U(@{YuDPmrT&Y1V{!s zYaRL)PTdNXly!xqn{RA1Wk)=5J;lJE4-F9qw=<8QbQxIQ6p`-FP5JOF;sb} z$7+=kyeHEJCfrBHqByIc&PAy;5F}g^Z1alo4ZGBqvqjng(y;0*TRh5Q*Bl%abup|! zVGr8&offD#>atWjfU-g6Gze5e2@H{e(jO5Z?U#(=W=(~TcOOz7TiKe_t6hJl}Y?8tIex{|UpZb;F%?=m$IbPT=4)iw0Wd@dE-NTKo zE;DQkGfVtP@GuV9U2Ckj%vx9m?rOo_X?I*N!KMs^(bAgDzeZ(r<620qBEgfXa&8=XOzkTyz`YUI~>=; zQl1=J8jSY;2!S9kSLD)(y!)HVv6Pm>`4E zuL5DyjzP8tX0v1qaj0Kp)n>R~mV|MiY}fQso>wal#ZJ`k&~DNxG`;8Zp!>rP6u`&6gebqR2@Pj1d0^Fym` zgBZM%V}1jr^6_wrV?$ZU&8J%HBw-x0(mOp*+ZrY5q8YT$;r@b?h_Z)!BHtsO|aX z^*(%{O&DO`j^5I?wR1`8874V#TB;b;z+2^r1C;}TQ6@_g2?_VM`8`)V;o27x9&GMVffu?`9QVdN+wS%CPZQZ~m{Z z*ks`Tm7VH`4<;{h4jLdK7M6OVy{h$<1!aR)fkZ|MRC|9_nOL_q40gIjIN~azF|8sO z8^k`&szH}!Rfy#=zzQcE8=2K`oTY~hDqCh(83B2zo|?**L0QFWRY6i{GH6kVToqc# z!g~f)8Ff&-5u&bxSlB|vl?U&llXKQnk>*XXc;6=ut?_yVXo321mYRR%qFEc%$j7Q0 zp(rE186mEiVU;sa6_=i`j)u9z^3Ns()Dk#v;Psg#mBHO1h%ee{lNyL3DWX2KW7Y;| z&N!TuU$(wyUB~N4fUg6bBeY3Wmv}4EX(D>NI6Ejo)mkN(3R6Km!eIkk_0uP9k!!Nz zaX!XbVOXSJ2*NVcel1=P!m-W14Y^|K$pz4tKWowbY@ltzQcR{QottyJ9m2zJ9Dr#W z%3=5d zQ;lNA3iyLE@`cy3OX8&nP4(2$GfT`CSrp5tgpqkZoa`FK0Gg#v zcFU#~snv^bw2p|o-Sp!v*3oL#5eaf=9WB?bBXXwFtQ~dEI;x#kv;`ruXC3WWM9w;r z3pbmZ?!i;v0EWpBjjcB@kFeTZGTPLdsn(Hx!hExiWVCJ8k&e7=Zap;$4^Ib7t_x5z zHE_bqsU+NygE)WX{W5B8R$`SRp2=LZSJQY7x-D&5LCov9 ztZ!x27_xi~X)69(wFz_G*RoZPO_Op)?3KjQj57RLf!I{4=2_4tO4a#Sz2a~x-_>k% zGT+r~BzV+GIf=E%r{h*BZdxT<`Qp_}DRkj!it0?cD1#YpfoQnO9gEIAyj+m*Pz2x~QW_)q#2CvyerYrXCIo zXM+h5{&Ly?u5(e$@>$IKVuVSqC{8(Oc)wOYC@d$0^U}bTQAM-J$0o(a)Zqv^*28AS z;C+SoX-9Wk+0KdZX$Ow5vGm-F9gOcKM6SMg+HEk*Jl%`sWN9|5hA%doJGk77mwC;m zbd|w*Wusv0o+Syq?LicgoA^j)xzP#{fqAj1hgXHC7A8fiiTPWbG_=!{-hQv8{UnF4pKr-77d_KP%%W3I z?z3G>Yz`mX-qZ4zu4Op;mj=V_;P1}pn5?&thhzJmZd5N`YxK%T-sg&MlB-yd$w~I< zT78kTP5!o>vKRUL%fHZDGuNcX;5s$|<<~y4LH8Fo_hNren@m4mWLhq8sNe5fyZ_23 zQ8T+PFX82BnLn2LgkY7Gg&&#oM&~=?0P2nwp-KhFreB1PN~W&pH2oF18`v={nfgkHVaeu!k4q*4_bPp^WV&AIb0yRDN}ns43={}f#LZWKx{TD1JW^d5 zd}6i=w(5zjUkBpF)?k}GwmW2aTZ`CIe4SN^*>3ox)>JT4zFLR1EPym}GaDqQsU|%o z?jTiz>lH|Bg|w>bI=rFo7^3$)j1bwxe7Y&p_!SvK1~TiG5j#h0cil6SdXG!@kAC_T zx*o*AL}V&jU#1Lcqr1O-`}h9Rz3HX6yG0fua-xFWeXiF-Lv;octqRiyJ!@*+;> zvP7$vlzF4>EAvL0V$jU9T1l@5HtUWUnHlYzo!lZ1Gw%{~vX*<-L3)SKNiBKNYtt-5 zUCsdr_ch8(05S(uZN5cS>mpYu=||jUVHKL3DjVbjE!i|=Qptpi%}grcD~3@mg7sn% zESn5QiKFSo2)22#(Yji!y{a0cd10F-sZEo`oZ(wHir2U*v@pawGqYYQlf55u)~}g6 z$oU~*`|@yL^Ow09c;9Kx6EkhI^=g~Mk(M_i+yfaAaX*{ ziA)+oX#mlPXkm-kL0Qj_+qI$_M5+iy4K6T&O0;$k_NBwR;?IM2=G_}QWpx~X;J3PY(eo@TY?Q!*0?k8I|#h6#@bq@=b+ZaqJym##B&C$ z%~e&ezWSp5TBmb_ML=lL5qulZ;my~7kr z`}c>b@zj89YTg+uK|qR903NgnD`sr zAV(3T5Csnf5(U)+nEZgbU*c<8i{tf|e_@8!3bvBp_$2_zbRr4xeRA`O(sh{5WZH%0 z^B6%a$D@eODR|0I&~Tva!OD4oX)oal+`WLV*J8?6>-}baCLYyE$c!*Q9oz)d*#|(r z7LdCI1#-zdnc(61emxX<+L!s;M4rO!;yi@gfoz)*2?c3S*yx=6Ux(?O{9lLZ zocw<#9{rgLs*!b#8D%_`tPwPXyU03QbVbX*63cwesUK!6@h7=+r3@SkkgRgmkiA7d zJFUrh4F^&qnw1lwSlw2Dci$slL?WAN8ZTkU}N^y2^4(qV6XMPDe<}% zj`HQ7)r#xnC|HzJZJ16VC*Eje5RBM{7YA_|az>o)s3or^g5@ zEdD=LL}yQZHk1pc42InS6U?>>Zw19{Z?HAFjd@m!yxR2<_AQKKx6TjLnxBkVtKefscPA{o+l)@_%yhtRcuDz`lu9sewiVz|LxPH>` z3cUSGi15FgnUeQvfvoe*Bea74J6g$p6l!NrBU3eQ#$AAcOc%NR}wuTW4$ zDjX5cV|a4#Yz65>nGGBqyo}*;<0T2QLIS`M;XHCV@qS<82=SGm4aCBm)WPfx9Z*6=H=G%TyP|1jrG`0CRv)=v>gvH*ja1d z5<+#k{&IeAY+}nSl634cxlO? z9WmU_XUMA1xxFKhRoqt>nqL9l5_GSs*f*MMlJ^C(vZ__F)cPA-4|J=6xA@SBn3KCO zSkoz6K`sJ$E1a>Qoi}6cu+|CSvY&U*e&Gx9%UEwa`4nD5jXW!z;aaO3!*uP})!q#M zc?*JFcmcQP8c$^N!`x@Whus=df+s+*6 z(bUe#i)fm>fVgL|Me5$<@|0|{6CN&g zxcnl9>^Ok&WHP;=l8MIxY~wCy+|&?F&KG^Cs_Y8eTW5WjRBBYcZp*dZo5N3Al=wFL zguH=*TWMzxO@1FP2UTWn<92MSwmiXtXjyyOua@}e_KfoU&c1SdGl z*1|MOKq~`bfA+1Ge-Zy;fY4@vJk?-#ok=o_b6RK8$ zS5T)0S4}CftXlFmF+I4F^aE1mbqK3Qoa`!v0)ugknr}5Ytxl<7G+}OGzEA{}*PbFf zhUrYEfAa-RYSOTCGY4?i=O&?n^C2RDv{=qw-45GGh#Z%*SGQmV z_2o=9hxehM!EuagXQ`_`tN6xj$EcG2ub=3j7u$Be9IxS(qQ zZh2;V4dTC58>MVAIsqIcUc6c_@=R6_s3x25`mQ^t8EBYbjhL)KaE_YWwi=lWszDZ+ zgVYpfZgE4hzAV{fCf!yG!YSTO{Wri+5T0N|TS4{#)|@=~O2}w{j0NGeYsVuFbV3QY#fwn_LS6pjB0gwVGFeFo>SjO;TJUZ`h?!xB_iVkjTm*KJ$ z*EHAuSvyIVpbVA`(S%IR`o+qeT_;#i)m~fL-b^dmVoY2@c7|9X7ymHB?r&9%2^9&+ zf`q(B6^-4BQ3j74x(`)_E(~!CNH4`mYN$p#JczTRK^S0``om`$Ay~WK4h^99`L!yAdpw@}yg{BeO> z4>6RGP2o*Da2fkUyS|y?PQd3riFrdoQ?`wnDeDDig(bR3eEB_#r>2~g>5<2sok`|v zL73gMO_EvIDsqd<&o#bg`$0es78T&Bb>>En9lc+n-CEF2jl{kuXu{xZ@uOQ@M=2BKL2d|q|y{m7HxQb$#^{*@KYEJuU1S5(o3x?(E3g|=+>sakNL z-gIftl$e{!L=|VbY0>U_D5WVMNgy{1W^VyhP7sHwF!b4kiKZfjy?RF~Sn=<&iz!j8V5hw{BvSVWY>taUs zHX;a+ZJ@+lZ$IKIF6JOk^!CqFaKXf^A}Su%1lHx`lF!4mmtg;1S$5(Xfgq5e&8ch= z$1~uDm1M&|;I<*Y*{*hA83;TS)y67t(4T`BfZ_WK?#n!;{(LQP%Z=H*!S32K^DNf| z5F460C1eb%Z=GD=1NEoo z-w3rgUA!5h&r^)wlzSuOHVRe)a)o)QhgNvx{$cxet{E>|c}<1nub9>C-!`9p89oS$ zKEtn?#^V32tP|m{&@s?I{!4oaxBmUPAo%Mp`Ko`XndL}63RypOwk^PiYRzyAxSB=w z^>#7%B*aw%kabE^Go)IYkB_>YaVMZ14`G zR1FpR*hjtJ*^YZuM|v7!bzF8B9VLd^8a@bnjM4e5#~6K{VphHB;+@4*X1}YL#O5c} zXvfyKa!JE-1180k@XDok@9&p6aB!fAG<(yLGgr-Q-t;D6;!IcR=BVnWA?wx76#hbpr6ksv}mPeLoUu5yn2|2J9s*`rk8pq%%?j!|DeuQm`y6)X_qs zkh882*O%}h&!`(03BAi!3GL-c{q$PGs)IBE%-`f#0G?}AR`ok4^wcJRLsc^Xug5T* z$#lLTcysDu*MzodcDIBf$d}aeF3A29Wp>g|*tSZ@;V|ndR;$pPvnt7~OF%_nR0*uk z4SAn{|7#VrKT;?UhK+~o!}s|zUKX@5Svu>cDw@pH2z5gN7$R(f)Jdhl9EssmCcH6X z`bV(VlmP84rqiDHLHyWq%lz^%#L#sft zpIRo`N*5W4I#O!A+OxA!>W?;>?90-oDqrOV2|uW|%0#2nOo{i3MS)Q6A-Vb4E>tYR z;>w6M=%!b8)1pjBllYODhbl=g)^E5K{k>@S%wj<#{V?PGgWoO5*C6x0-uJ7vcb+j) zXGm_j(pH^-(ewRmHTs{tAif!#=hS%zro9A%^~#N)ra*FjK+yuSKYW##L%@9zaK8oj z99LDhuLYSXNK?wA_s?-o>&p}DotUK0*ZPtWObQGpRzjL7>Yi@n*BMWqK2O}>-$g-r z3{QlJZ8K!oSez?{S6v*BFoq+~McH;3*{Rp!9f6lWq>GRZmr$- zcmS7U3A4h4h{3zJ3m}YVX=l26qrm=U0dAqvwsr>&CX=TCS0|ZHA zf&2OJT4|C@_^{_K^aXTM3)5bL!O-(ViKzdOOUR4DQyEtcJQa~vSwxr+**L85GJpfa zi*%rL;&!$i5C_j`0M}x8D73#yfa4$oLtl~Uwamu zAGWc;2ND=F;lnHpW)!N)-hJK=_afpRMccdR_B6s~8oVk&jtA7IVKCq>L4T=(z!Yr6 zhhQV#WfHf&D|EV4LJPvLHAcL!VGRgt(=uLy(R~7O=5maA7-(dJM4si5khCk-z@IbV zL!tRPj1OO2pqGmaRON#2*q)l+c2CbOEP3!*-HgqF!&z%Da7&R@cQnX?%|&pdxWYN8 zI*GgR%q2_jt952@-3T!cU_W`(_Yxw3Cl8ec2k*=gvt7P0iGC@sn|CIR1?B~WUn|V| zaXd-vO&BAXgai89^~L8Qtb$#U010E*ZPgx__M}%a2iEd3ua{fR6^;NW>zQZUyMk7? z`OIuxX;Yj+K_#Lf-M>EE%YeCC@4m?Y{qFJpZo0p}`|9hz&}^z7UCbw_{TJjG2V`pf z$-RN7hWKYe2zv>8#V`7_6{aR;2`{cqHRG2uP zH}AZ^xSt2@^^||zL2ElD4WxB9>9!Pl%R`v*^!W6cUe^bDTt7?`f6u-C)a-A071l&4Q-cfl96BoQyRLh^_h*MgSIl$`-7^inRUs z-cCHGdh<|uWMt$~?lWW6S}Kjnhg9@ZluFKFanG#-YI1fBT0aB(2!1zE>HS4iYJbk? zjRn(y3@4Lw``e|`Et!v~)RP0s&@`kDz;-%%Gf3|*!lvxcQK6yoe-lWbsY2_qHF5*E zXCK`R_7>Y4LHbM;I}2z8^E83$aN5g$jIV257HZ-Yrw~es&(`D(ROYHh?3eg zZ!?=F+7g|PkO=waQ9jAGYBrllHNXl_Yy zF0KG%lZ%I_ExLmeLURjcfyhAjVqNyQ0}(u670OCp(!t5P~W!WS~H;s0{_tD$Alv0j1atq^@Fa=(Piha7?+4cQ(Mg zw>U+hQ-1<&(`$kWZG3}Fd+Cgg-@EB`A3S~K6FH`jKsTDc#E ztCwNAL5N2KDVfe5jh}8LwNS4B=;pz`15rQN-vljY9lulE{@YakZ{^K|5Z&HC<5_@u z;;|XAXHZaE>(E1cmQ29GN3IIpm>OG4qj4m`8mHr}ywE}k#4LlKn<(_Q$7I(O z6qGyC+bBO5^simepH+?q1)Uub{V;N4ZF$oWozIQ_0cEIakiwqYvC;}7RoAFPZ>wnm zy<*iQ%`7G2fsDQ)eH;6Ja8ri4#Tn%(3?`q!1O+X=6!h@$6w=4n*T?C@!widl-NK$< z|4SaHsd-2A_|(*)Sx`(`M>uE(2HP8OZ_n22Y|sL1mo#^Ax|ZQAPVX(?=NkQ7qn~T^ zf4>@~j6e5Xtg80jhLaKE^W&Nx$|PV{Ct-E!EMgOlJ{dG5T7NzrC(Zf;?*wKh4rN7o zVzP;}q(pm;C;!>Xoy|L~=Z*N=C{O}=v|9--fKVp~3o`Iz!A+}mvfi*GO*X~X`$rad zqcyeS)DjvzAK(dbY6*gaWjDmZw=6h8h`cfd!vh;W;U*AgQ^315x?+SLpS#Mxhg#*3VIpG9n^nc1^RGpOU7o{h{NLP6kOS z+s`vwe?7&CMAMH_P^*-E&@X6(d^wF~mf7e^IjuhLXE17oy^4RUwXY44&a+z*KBCw+-;|_ogi@LieE`=&DCV4KLAV=9EVIy1Oid|8f33Ox=HEdH#c|yTsHS( zhNZ#h1`i&u9uLUl(qxz$Ja~NZxNPxK0_+#F;tV%EiMdOS7B@jpO~;?c>DCAKaC&b6 z!y1|4r}LXdQDrO4DtJULZZ-1V$e(1vZk7p2#DM)u5fj$^Nd?Qa`V5k*ICSR%P2x8J zkGKvlc8S~oSs3BeSwrpEAK`;oBwfqk@HV&!IBCq*}A;40~p#}?#s*cdKP**~?tK~l1AUz=Sylk!^20h;FK;f%bP1R_bMv=;5Tygrq}(bm z8!YaqVfNc%%{I5iIQa=dr$79%cAH0v>Y~#^fH>avOiB)$xk zNdm_}sq4a6@)GnSM=RQAk}1XfI~f}gZCPvgTi1)-XcYX=r7IZ^4wsVA&WfEJj-twQ zLxF8IO{J02m)ugs27`iH@rKYoysM{#rl8nWC=DoWg;e=cPlA^%*C^(nQ%`~l&DSVd zDb$m8DWWll6xg6KSn|6@Nqs@-DAO(_cLp1h&Qqqopnfs+R?2Qc=&HkV4-!Ah5s%cZ`wL<;;UIiF!fZS}U(9vr{ zgj9KGNvbMz7{fw>8Y~dVrZHC8=bmHWNyA{~QEIrX^*x z*`Q~DVv;IlCvsu<)p}TR0c8?O+cEWvM6ja9OZXUz%d?237QWzwtG=E8FoDm+J_UBx zB$Uuna06qPDWvlhEGTxFhLm|ofu>cGGdLovN~d81a*ZhwKR+Mv>v_^xA9rEZh*rsD zqaSq5975AcoRG2P#QbZFYbs{WYG~A#u3KM8Bj6##gNZrmKfgkRv+sV;cQ>iU9 z`pN}%Xk9L?-h#!Wp$JZDTJi#Gv!+-bd;$NbL~^;ZI2xmbS|g{lvpXqBS*TG zg1wfIgp_GW$&M00jxv%QMFnyaiVCe#iiq2`Q)l<2SD~mtD@CREStw;Kqlo4tq*|wm zKkBQ4WP>F-XA2xgDYH?1Qb`yVv=!quIEw53`f~!=W^$A(@pkL}({6`TL(HPZ4_Z(| zvfSX*$POiIuW6z+I*%x8N6`+Nb>HFQGO6}cHmD|7ibXP2vXQj7m+35vt1nGxG{|0Y z@VO?lpPU!r`bX_KXBW1cg*gv&7|$yKe$6w5L`d1)t~iKq>aRv?=qPdgi=C$#0y?yV z#OrA`Ar5NfY@g=Wi96P4Kk+cF8H{w4=zg*9DNLi?#N;I9Y8o9S>aP{8LE5nUr(OiF zrqNO2_A@W=IIXKus^1nFw(%&95-9%mr(c#DrL8^Wqf2`=cdx%!xV~^=YY^*<%Kz ze3Vw}lt=o=T${nV(Yl+)F?bp>KmzpRG@Fe&eS}k0ANrej)8I%f4T9meeAuT+tp@tY zEnv36;&nI8pTEt%``bjH=3)pg4KDKQ830#Mb!m9RpR|zzX1CwsAKXt9qQQ9K`$@*7VH5)u&cCQWF{RiICPxPHxi(7~q)UnKN~181~s-Ku3wL zjZpC}Z3eRj1;tPEsZWELfIxJVSnGyqxlgk-v1Skp7PT}dRb%o_~GgC?iDVTIf z>?R3nep8A>vqulzltOQLXp};4d1#bEgQOcdR2fdYNjLT%otG5vK(}(|LTrPFPGo-jGAR zD)ap(yH5t4K2|9ZXc-RLp?EyKWl&tt(*_zK1WAD4?(Xgu+;wqxcUW8!+}+)m#ogWA z-9m78*URt!-h02C>Y6!qW@~4ry8G#;yU)g$Z*4Kl>fS8~rH80l3!C4V1V%t~>dKxG zXZ@dEww8ocF`+!@|2*`CknI+iC!< z`B+-lZ2xclDGWk3p^LNtHMt2{1NQJkq~CCwi^>MHuuXJVPHFloNt|}yapdQ;{`LxS zU@FLZG)$Kn1L%Ld+;KKl;C^EuDgYAATX0p|gdeJjx`@Qdqc=}YB_n0tkZj665a?o+ z{PE05qyvY|F3aa~oFi*S>}HYK^Jv@+Qag!sGYxu^58OeCvaOP(#3KO~4a@Ct>hC7jDy>E#qe*r~nOuUioKsoG2u=qWS_~34eX_LYdh2+&j#-(@_1GOZ%_niaW zm*d}Vn|DdeS4|v{-_{qm%vL)jU*#i9(d^S+OR)&PIglD0$soXFm|kxs z9BrGES~LEqCP=95x)1trnqt*{MIqn~=4}esW~Y#%!ZA7fH;C@oSt0@x8mWIf5=8!W&BP$P^uN}3{p^VAq42X~{>E0r9F4Died0H95 z{BoF3WPJR-k`Mw8lu2e>&#g^kUGt#3@7*vm-KAOc7tXkVn&XHi7O(F;)%6TjjSU7q zIlbp<08D@KCIUVkT)jV0W_?!y~Jz~H#&($diIL6?eZkGvb`6R7{s*gnS4{)yOoGHFU zxUPSVZ+A^dz8NnbRnFN}X1R*#lISil;j>w*&`ur|jbbmg#_i@j zB9*Uh^fM_p@mf~&Y|=hDTS5p@E|a?gkP90CSimsdQ+vxEhKh5eH5r?-083lo`o9JN zNC@}7--9S4=i_NIHeXqI=fgr$1=wgPlANh0LQ{*-+Dlg>*^&!ZJTh%I>G6cqbrKvi z|7gNG^>?)}bF8tS?vhENT4Ookq<0#31X4IZ@psr+uhb7FW_-zs8O`dmNTkZFe(#L_ zUu^92x?osO&9!5nG!k^87fb zLCxNhHT-Yx8@ArrHa^8JsE0v+ZMU<7Ek}hn zkI>f`ERyFqD}^HFCs!+8(LsVuIGT>G&}Y&+%=Gal#XaHZXg(y?m=ZX*#mM2!7#Q) zqOS~#js&!a!0y8OM}Nd?4WY5$5O>zsH)(vF7>PU-a!CUk)4mjAsJs*d4YU@5343Wg z6h%eyjaMKRX z9W9g+bJLin0%U+#q%`UQ|F4w7J_sGf(%@nBk|e-jk$7YaZ7b`kX3E205es4*W{_B6 z{7)!xw*Ll=aD5BiV(tcyv+H3jbS#FCK5u?})6ZeROH(p6e~{L$ZUrs3;t?)K0a1%i z>t~zA@X#rlawMCrB;;iGSdz;yRYHVaQ<YmeWcQ8qp#A4zD~r!XGJg){RB{CBx3dweh1d5iC(!r z23{DJEEBQM8nUR^#KNPWq@B@lcw$Yl0>?3uCwP$*Lck41^{3?41NQ)DELL0^4sUlJ zX>#M0OT&s~A*)TpMpc{FOE%dcNuP@4M?>(kuT`3bTYQTVFE~M=h$4-)@rFs=CggLW zju7kP8~NjYu!xz=fSwxx#);9m&AI}+=OdS@fP7N|UNqn;&}_)Y<8HIh3v3{KF4!(e zsX1_;x^W)bmhM^qbwvO#T6yGbpO6m^2>>EBGrp!`$7AO0?tK5(H>2(QmI1eqIRlFI zj>Yk|YMS817r=7^Y<%VR)$m?oy9(O?IvA-VG;;LF=>`N=5_$N1{L*tfG&3^ESa5U% zTCql*3rjt+O~aY2kbkWJh}b6|)aPzkL!|GuTDzp?o~pk>+ld}*9O z{C>zY54R?_y13wbZ)#`c zp}Tv9hXKJ{(vdGpd*|CO+1r7?FT>%1|EBu10~XTZCk4Wu;m~m}iuojj`<9UI>YKJb z^pg*)ZK*;fS<_ULbKUC&Q&M_L0D=73oSqlw0V!AtabwO4Ae+Rk?}UGnA~4V=HNQE?KmFmtooLRWX%WT)7QPs)yJmes!1Q;HDnEbLO2Uue|BP>MBIF8MT8t zGZ3Qz|19J)_ypcO$@E>6UtK4Mw2t#k^u4eEyzEY6eR+nijG}LS>HxP?AFmmgxxO}O z3fG6kzqd@q(P?kN`NRx*DBsSji{zmwaWLwlu zrBar1Hu^p!nnrR??Qz`!8`~tmuT^>?RtYsT4JZ=YC}0 zEeF-jkb(jkOiQWDm~i7EDWa5P6uM}`ckB^M)#$rvQd4Xs3%^J;6;1zVdXxIx7H@yX4=o&= z-YN!W{y%G6dg!zXW9jed_29O6z`fl6O+XQJRQz91@WTsfb+H{A^uJHIyyt;5>Fo43s$d`#-buF?N+VL-Ic- zvEV;*6and`t7tl>?c?D5qvZu@&iDNN#PUPFjW8JxCT!9}K-yY%5WlKzu?KH!gAIp* z;W0NNsthc}AROAgO0KB`Z@y~{8wn%K{vH7>=Dot}-ZY1-3+-{aQ!e@-vHS;yV zzZ*=KiA}N8cluFRvl{U2+0vg!8gP^x0)3S!y=u9$(6JA-rCSlfaXu2YF;_p=G@cvcOElT@;!@5{l)+O%8T?Bz3hwJ$42Gd(fU2~~y1^1q4e zm_2TBn;u1{Jzk>9O>%fRti~7#>cI*{J0>3)h+48kBZk&Dz7fR0X4gHR9A~haB1l-- zJN)R*Xv(jB`6jcdj9_iDQW(YWyqumZFm?r;yP%4~cX;y4nglc`><#d>nj(Ibcz@HhmIvQW z_3cNb=Ce;0)CYQ|VV641noJ{3m|7s`i5hmUWn#bEm$4}3HGh)inS4g^F$UEKvzgH4 zdx9U#MMc5Iu4-!&q+0i9E^Is*jv>WFQK8})c(#@nYSzoqmbQ&q@q-z+=?f_1wO!|$ z7K5j$Z)HuOCaCMQoNZl)3^Nr+IM^=x^u*vGuDI$H+F~>&w@lbG5Qa zCm{zaf(}d25$(TmZC|->IHxTbjZtxS{#wYYBX}D5@<|kYrqg{ceuYzEe6RhzunxOw z<{f0fv2UK9@Y&>rwy$1x;J8VN99#7-RE0sZV?jrCMiDP{ep2C_pbf+V#@72*WVqOV z!;ir~D(cdr1(MSJ%N9W+sHXn*mM5y|_j^!lKgmBi z_^X|ZRc%cl9s?k#%p<3;JZ+p0h0*%o60zn93j+7o?U$1dHA7jg<*(X2on`*TFjh)Erwt?wcj(b`^ng?EDdSe)t(ckHRD2(mcan`Uh z^jPylO&pnQW?7*6GeZa9_o==a1`bWt8_fQ~0zhGul+;4Am7ghBvE}|@u>7n-nktTD z>G-Pv*-_MaZt%CbIV(K<#q^=gpf%fPww2`PsX0yc;n*=Zzt>}&ORE$l)pwCI!l#Fz zhnk3bH!C?z^>;Ko4^)*Hw73%;Zx2MayMko3XYZc=`q1Qs8mVKO2QFuXpD>ELJxGL~-Aeq1-EFS-%+G&hZ@T#Ok#(Yw;wCS+fF<*_Z_F49C=(>FwGFCUx|8T?+Hx ziA$%An9K4ky(V;EV8}Cb_ts@M z*ho5|Kl!~sv{pJ}rr4%t`jPON9Yw8q1}x@(`f$Sd+=crw>s}iR*;S^?{)7_v5imHj z{+gP+sW=*eh+f#TnuNpdYF{yqYvGdW$S*ba_H7U@!>ZnOWc-KZ9U*@WTpDM&7)~W4 zOI;nVkL~24g|l{+4b2+z(gupppX0A53a9TBIRF`DXf>HpO5Yom_Xa&weRNMoDOiZd zs!FQqa3g@iDp1MwjV(VYSe$fQ+hTsHxa_`k=BKsZ(6>?x7t*00HV;P8a`cLN`=_q$ z-;3E*SR>F_=6!i6D{{qg(wu8el57LNhi)Px*sZ)?dp$m)akw4F9URRgi6swxI`K3c z@hwH9N#AkkaiW$`?@L@{_DQXpm|rTO^;ddav!(w72M`;3J1VvSYe!RxN6DYaR91z9 z_#Mr!65o0;0urDLgFqb;FRW@x5gu^uLf#h7A$zskIBzu@wU9zdqLsNTc0xKQ?l@o# z-Mf)G+*&Dkcx*NPd&5nsXxVlHHLLrW2AQ>g5VQvfS*117u!!PAae5`cPv;kgL@2VJpSqT|9&h=JYmy| zmf;5;jTC}?fUtjssY9&Oy&|V&b7XN za=!TBoA))lKQL*GJZK285WSuDDZlR)gO)XPg*hEha-N~XJ%%n3)ev8FakwDqn!qm< zQDD>trxQd`X#(T&tEsLQb8w<)o_c1Td~@j#MCibtfq2DD7l8r$kO$SCF$fnuGP2gj z4+}r){K2u5Dbbb26I#L{``rGhPgdeT3p29rt-j;LpFcdg&=SnF+9+=#&K(hnR738U z9nG6}FWX?DG0B4-Fo=#s1m`gw7_v6Cf8P*p0=j=oTv%^U@ja(W2;Lm}@x?e1@Vv1g z=vk3Ob0yEvMnOv3UHA70VnB{I2YxodHWBq7UJRS#C# zW&q}HQ@2@_Y1tu^6c)&2hJgI=L>ue-w=2#iR&|PB;qe=FUd1v>rX;ApNODfzN=8HT z45Bo!Y(snxi*TN$pfO7kf(!^aOoU3#V1vSf4<&;D{W(Xy_!#iKEr<)cU$%p zz$}5={`9V`;A~C_{sh}g!WV^1<+-}o4Su5$)Xww8Lkv@>!Iiu${H3pd0yh&cGN@2> z=Y(ygaHk_Kk$or93VHwrp!&tGzA*_0GDog3{rI8XFhj9Uq_7%9m&ug}+D=K!;c2cV zpo!uyT-u-PNQ>V0S5l<~zj+Pc5|UXER0eZY4nqNPicvmabEbhbp9Re>p!3>;@M__G zfXO_w!31Tf#1R$FK1n2u;akh3K!(!Hdp>t*-4PX6Eozv3YT&uGUL){a`}P~Z44HIAwPXp;YI&ye3UqZaA$tE11g#8&a-aqxRUL}h4*9Au2r8J+_fld9sVXh5XmlOmN~>MYBe=#zHuPeE%HQ{%1- zQtX@NhcM1?#^G73sN+cO+T1yMFwY%lC7%G3qjk`^Z>qh7 zfOs|jZKLTBLEcq=ua2_KUJ7ZUXiexo{-mrP=;};kbh=VaoS9%-asYcvKu8OolkR&w zT&aUmKSATe$>&l3=HI+NjTnzTrzqi^fl6=RUo@{)aqWx#N0weUy~qhy_CWnxq>F`L_K7810Vt)kWVe)*Ge?vtLYQRML`%wMbbeRd&?OLPqB9kMbrp_I(_&gidLGJcJ9-*3gdR zs=y;w+0u-6Tx%;yBu>-DgL@ntw>vUnaWBIAXu(jHb8L5fm@(p->dGq`nA3=AcS0vd zIlg2KiY+q!@B zASHh|e`bO*S(XUV>qY~7NDmGv>A+4Qx$^QY>8C+P6dse!PS6l&Ax*Y z*Yh%t9tde$<_P)qoThx3>01suQ+^pAi6RpwOte`&HS$8MuZn90Gp5mlQ*H6YGyT z+nZ;FG%zqmKqp%bS5%;GnivX>x52~Smbs2X)>Cux*{Xs}$+WmorNq|NX3_I!H9Vv& zd1R4VShA(cNuc^jzSTd99j&-KvIe1SVt#_XzCXuKy&(-jNB?--jss#AxGDDNF_~o= zg#%A#TSXK3BWNt85rh?my6XWcCveoZMTR`U1o3pWPPPO(+_qCN9qAxrT~_ol9dr41 zWav_YEgCht=x3Ap;?<+u&l{+7Xynlz-Fj$KCD(`bY$5&7_JMV~T3*&V{W}S1!{=f@&F(9t13X0~%x2`GBbH+?aY=spxK97@}X+2jf z2Ex59YGc;(J664%z7jGevnNZ@>um zgawwC_%M-e%^{ffq?FZERA}Ifa);lzfyGci{p^0|G1-aNYg+WCmYc^d`bv`fbCJ}_ zpLnNaVXdc39=FpTC5~Gc2>&7P*sKnI zVO3ul_}Yt}P(I$UQkO;+tyN0pq{^7`$dnhh(%HMXBxGo6;1JcYPH)lLs2+T6m3&(& zE<6hbZ32UUuJPC=enMy_e*-?UsD|HRc=2VA(F9#Rv(P($Su~s7aLmfCyr>sj$|kst z<}S|#1Q^2C14DMnuGQKHw+<3*capmELg0+mJS9`UAQj9hwH5x=ZozysQw;URIhUhw z9_6GReffQMLyPAlmzA+>)WxO;#~0V$2%sCX7A!NT69?~$T!-V{BOT$7DU7iXVtTo= z!fEOSMx)B1w7=ZYa}2{Kk9X#0pk`SGR;9EXyxnhFDJC|FBW}ViL?nc45iOc4ohgJdlG-fJ!>5dpnK$~ zr(r?EP`vVNGy^tC-a!5M`vlYinu*7ZGX z9exar8jz$`=y=!+m8?c7rR2m@5SoQ(-=~{YEmj=Hkmoia`6V^9-)U#`UWe=X`ui;f ze^uX0#Jh9-1}T-`tHpf$eShTfipY1Z!uS9ryv-~_iEtXeZ>}_nS&Gmi^HkT$kx;Nj zs}-1`CrnIf`SIY3)q072@1bKWz`=EBx(Q-P7oHNljZeBV>i=D=B^fk+(0}UCAZFX+ zP;tix&C@=sn4kt>pC72H|EaWf^!J>PTNB>?r=lnRjQHvWd|EI2OC!FkCbX=79{PiX zo}eSW08`QPEzM4k!CzRvJ^sSdlnr|~2)NC(opPSP9J%=MZxxNpCR01(6$&o(KeynNXx#q$U4D;EzL1!`x>wsLSNIjk zh&=OBgTrCI1#}tHt&g<#n0Rz7)#P`n0F=vPGrA0JeJG@GAAx{~E z5n&Fl%}}rb$WDEgs5gTP^-`A<7ijHLhiU!VGpzqc)&(rz=|M+bWCrnWWtWvtP1PT( zi}8M07_B)Tvt8$4Mm)=swUqj`>ni+od?!8DQ+QURgNTna-=W5o8*KAs)79Jt+VOB! zIOAkK&rmQya{OFn9)7=D@sm3r-O$_;RBms{`Bq=K#f9&FTxa4wF;4y*vw&YZd_KzD z3etyK$H%?*fwp}CWqa)X!0kMF7Lo|B%OGN=_R<43t5PHMiycfV*NrUZmWD+lB(=W* zwp}CV+>#v=2b?BI^P1A-set`FG34`@b=irS;(rYEcW59{_^D1=`+11GG~{&!28W6k zj2Q}yk8;|V{<|GAbiBuVp1MQ+J(c&N!+!NOcR4J=<9Vm+>BL57F~Nzbu#HrNf%nU^ z{;{L9o_W<+xFDQQ^?f9pX?8v44qP|dA4OV8Eaf9X_k!ACs@(HpW#?H^pUDGf)uYNb)X|(&a@i4k9s8(ZO19rbUo@kE|qW66&=t zoN$X)IjN)ro3T)W79HQeCH}i0ecWd?1Zc(_%UtZ)zd{^%?vpr;pYQqa$)+JQWDrTA zc;9)(^Hdh9&-yyYTavsaNQ!uH-oio$_z{elMDy{Jbj2FtprpK~dOA*5r zlx^403Urb8h+W_R++NjF%_^(xD4(HdGx%;CwfeF;Hmk;RQ;M_lSxfmVFSRY#vJtd3 zEhLP~Bdw1J1EeowBA6;+sQ4L+$@+C&g?mH#5u-VM1f8|pgM1YFqfbU_%Xqnf1+b>$ zTtrL*saZ;P@@HH*fcg2e^e2L~5{x?JNdA@g`JQCCii*=Z6Krl7*t{O z9$4KE-oR?iK9YZ|KOgFi%;ha@KhSxi-3;UYMcq?ZTM5#inCH^1%H-jM(3!TwCR?75F)h@6Y|0&g04mP{2VR;IgSl3HEnblc(TUw;33k?trvOe)q ze$YSIZ*@1up^+6SY9Yw5Hb%}|I~lz#sn%Z9iJ7=#(>b_rWf7qrpO$v!=UAF^$oiZe zS2El$VOuhG=KsD#Mp%W8XeAvs4-BnZlm4|V0Pt{PQfM1~|aiS@5si)>CNzaa^p>2_# zJk#iv9hPWSKYu)%t6HDaGlb=yB-K0ts!|6hHcG!9}pj?$Xl<`nIM*SDv44jIN70R!wN62NdJ#n!*L!@!c<=F*Bk;> z?}S4Mj)hb+qhc=8+b`8V|HM3Oa>r63mhlq%}$lHQH)RA zlYY!IsMoaHKsVT+Vz;H_q(<-ROYPQIH1EJ$X;-Q9TNZ#7Q44&8!|5pW{_!ns9xS}` z9#2voUx%>1*hcaZ`s&O0V~Hw>bct!iFYZu8e=Z37&`QxoL6Ki$k28a*ze_`%kRyms zBRUob_Csa;!w_|8p%m3&D!HA)zBm@e6br@1(UKLsKr;R7)Jd5pr5+zgKo%R7EY8Dk zNp7r@MLQ_Y>eS>CofI+FsJY15K?$g`w)zl*q&G+=#10l&3!GgkDBs(A5Ba=xnb2^Cu=Hr-!sYs|f zGwxLrN#R@Su){<;IoIfE&&WN&uoy}> z{@)0g5hy`S2}!b{KVgNlu*Zgeo-5?9|KuoUn8@s>pqlALfynQNm@tPthXoZllcR>6 z5-8@HQqXHvutgAx6GyNm)EQPDv$)7Fr)=$GGTh8neN{e^fEQ%zNOabgL~?8%IsTU~ zYpUd*C-o;cCr(VG`BwH8-CWZc1 z^JCH_USsb17Xl|4nZ;of9+;GU)DwoAkhQG%??$@tUwJDQwE$qLY}MIvBJM?1UT~{X zFRE4I!DPTOPw8=0oP3Dg5PM-;xo}c7zf{yh40BxBjiN&;cG?5chCq=S7gqA#+L1fg zlOvxKaco=>YhU`RlROD5N-c+p=1Sa1rg2#F zBTxzqZIqptk}Qb z3p!2td3Go9Oh*dMN>Qe4BEZSZ$+|j9^Ohd2Mk}AXQp6Nf-K8xvd51Wdn#ykmMeC83HoJVq!U@>twoz8_i8e*4|TU7jF5zBKv-<85s-4yCj9{*>Mv6g~s{< zfoyDv9Fm%7ylrMiA6Sy_UT-56R|5%~qL!&9A4YapCHUG6HvKf5-);D%_o+_CdyJiv z3oNqI| zryFxA7-#h&Sw%C5eBe3u{dtYLZP^fEc}Z6Ma&;deQkX=A5?{;?HT#Y%xHuepqE=Lt z_i7Q_#~=gg&LWu4E5D>4a7>I>u@I|L4XcRVEF60}mIxH#W< z+E*`Bp_X(N;|BCHRpy`&e&vc-i!{VqiWV7l7Wx#a7RgcW6kRQ3()78p$^dY-7K3ff zwC5C`gEV)<7()e@RyIC^am@AZ9}?=NVmwFZ;!{NaSdQt3KYV91H4HgyyYJlT0u`C+ z#S)q&?shAVyO2<^vrewrX_;(jA{o8#I4FRb%FfG5Nb zM}V@b!h~e&rewF3Ga^35#Z}R0E^Z|knjZi3%LQtSoVwFo+(vGkBwZ(*$oRr5S=^S3 zDHoMq!~oX}vbvHh))bu`<$UU8ztI$roYg5wp`wOWAj)1ISq>;Y@!+D2`j~2}RHcZ$ zrk;2OTWd$&gX&#|oK)pgIP7u~Y5B!3lO?E?0DWsl(e~M`Y4pSeA8BD1m}s7)8?Z@43$jW!Ojd7>BIcFsXc*;zh_N*QwgW>T80R?8vgTX z^}S5So|m*;H(=i34-Ts;Bs`n(v(Edr^1ArOQ_(V*Yk0h4Dy|8&%y}sKYaM&ayyzPy>8QL9RPU< z#?%7NkE(9g2es>5I8K*xTYvELSk}U2OT1mX@wx7*sNz40Vafr`_U}5A~AW^1! z@Yln?GH4cxWS{q>(JRi^i@@wToLZLWV`<<-P@n(&GxB)pMp*k={n)5QxSuKu^eJTp zdcv~fulPSIr3LrX)AF&&SkI%dm?14{1s9;*ksBg39**2?3pT= zJ3oU4Cz7n?g4koerS*f)B}xY%nk?{kBwr}XoJO4l}R!2FROlC8DKM(Je>i%y|=1z@LkIGz4|Jr1Y9b4~=ef9@Y~ zPbZ)P-11pS5I=AiK%j*WIWEf@hjP2{<9qRav5>+Lp_?N!!$5qed&CNafrvh(+mE6T zVU`(CGEem06g{y1c#8hN9VkZ*v&+__sk3-8hI$00L^gOt5+Wt0Mjs|m)6=zT5>PKR zqTvR#4qvlit_sOpow_)c?;Z^2Sn$mT!G@4{s{n-6owZIndBUmVDUr6r2(=;p;aa^p zWXaGVASN;KeIh&6e3ez#zp`p*GF(4>{UR{yx&OipPgLqpNQkI9z=>>ZPVb#6LD`(0 z!VXW0nw4k5u_Z7XJ#gLn=Cdr=HW&9fII_~6ZcgG@=OWC9bFu!m>qnw7s68 zV%jM?vF^P#6PGBu06v(K!o{#>R^iS659?!yX^L8E?T0&AWJr)P#H8$=0q9*iH}(5yR;bnM0xo?3vlChB$gFl*{-J@ z!MkrA5%Y=i@P+i z2pP=-FE0}6`jCF&s@$aCw4aE8U5Yq0QnxUQHLM)%8g5zNlI z_H!m5@O@!RjbP=Aft;!Jmo{`&i~*X<3++3$J12}KMfKnThsfCK=NcSES?c5Tz`sJ? zi~ER#v|E0pW% z(E456Ka&Z|#F!Mz%A;4ycvjOB6|{+fWV%b)dHFF|Bs4&ZTHxaa*Sf?Q$usJ{G;Ovv z##Mq!WR7++0rZR^##@$-YYe}G)ryZTRM-7smtl+%E2AhJtJhqwd-8NKx-x@RZ5BI( zkOl^_GmH1xHvE{z0mk-dL8`&c$b*m+4UfAgNBR$Cv z?)*;Yl!~gIs?-J+Qn(_=Sgpr}T(^4GQ^*{#1AyfevUd)L#Q$W8^Gw1Jr51jc(me9k z5M()+uUnGhRHGI4N~;mOlI3R>6?l;vYpj8D#0GjH3p8eNA#z3T#10L}}xkU5ni5WNMa(FY2yeeQm_Ar2MnQ7VIrQFsussMWLjrdV0hJjrxnXQ3p?I zG=S)--*e$go;UW^@-EsuCVxqE8YAaB1R-eX1iwo`bIa@=nx$@-N3pqU^XkQg6moW# zfg%8osIqGl-xngnQ)e({zKxXe(TI!TQ(tY&U>1Ao^)Mw*9 zDdEbXbwlYYbf7ovv(>eyYx7{%^>h`&+X1J_BK0d9$=9WdVHR9_-vu6O=hQ znFMh(Z0FhLSsJ`f(JPOvU`W+;rrMS zZ*+7xQrQ|HS!EE^(kVYJ*uSq~SDzHV5q_zibJp7J@Y@~JVx)YLn$r*o%~EhE82(2D z!c`dy)!fPSoWFH--YnW^>v0PKE5G;aYM3YE`e9q8yAZkVOk`CYpEv@oqb4`{@Yksu zEIbyLoSJtYOclFDXZY~3p7PAj?B}x`3i8vO%_n~EjRn$>IzJmzh35CFx!Bmq)O-yt zaX$@^r*AQi8cEgP$$7zU%%8i{ojWh87>kxTL=-G6uX#>po+^v`!7agc%YOW4dc5zM zqd6xlH^)9AxrXRBn+@E7)b|=S+4b%B57!8}BO#4UsK712F z-9f@;h6I4+MJ@tXXnz)Gb{_r73aZlIxIGEV?jf=HEn5d0sT#vFr*z~Zy$>9e{5k=? zicF(-$zFKo;1POy zOZAHh+t>ol^G#e*g&f4%2jBFP_aj|~e=5btJ>d`n`7~JHoCt>t$*xtW8n}sf`Aw_( z9F8TbHmH$&OhrsxHvXPiSh$Xd?28=wf`8Uq?265=v*oUYdtto8aY7bRUQqK>RbDdys-@Dr_oTmA z&^r_8E@aVSs$c8742XCk=AQAK^C1dmeVJ~@{5`uM^cd!>aEYoe2$5U&$_7KwSdRG!FzfR=yAJ49cMy_f-IqGLggs!y3z@nYD%{U6 zsu1wsp);cp;KVC6uTkD6xApVY;Fd4}y&t)1U%Iyvg;0$PM@q`YeR-#{1BZ+Y#v^%O zQ$o!?oNb5yWA7mMHJ0WkvFS zu+;4MO#~zExRAA9gIT{*WN?nTkBGiwuQwDQdT5zS*Nf=(ytK}{XOfqq8miO4{ss=o ziGA~}_4qqGNR=21lm9`l2^lV@_k9X4g)+BuB4YcW^ZF|8vla z_eb83H(|Fw+rPc4r`l}TdXton3lFD@($#J9EU9yQv!*87>vv%z>=NzMxB)n@EBE^I zQpKNkEVp6 zML`6`!w)@3xu`LvE*v~5*;Qo^E|FY(g_KA9e`tEg@W`6)YdE%TTNB%!*tRpVZQHh; zOzd=Q+nU%m-rT?c`+TYHtIpMZ&gnk2t7`4F)|Mm_E3!H*Ok{Qco^5UfH$^d2kneKU zzrX7v0N*l6!N+OHNz*y98b+C8Oh~<8WB#ou8;UarcmB}nZdm|d6$$e$?8)OU+fhZ- ziuP(eMXhh2ac%$yw(3%-xM_@ zxa|#h^@+EvAc)VURAG53;M_UZ%X4ZM* zUDGDEuy9NLGRwFf!w8*=1jKQ*1BsYsw5wK+%B>*}Z`7sOKk4S|%~?a(WD%;L1zu#F zaC*^bs)%4r#!nPb=A^YWzyMd59c_iXal0Yhh z_wRQdJ~FyDM!c!Sn)cMVo^xkKq$RM5q^j^cxrv!@rAy!{Gh`f87^WRfrF+mZs2^2hmATOPom@^aJm-u4a8<_x|EC`q=1s zBw9KFU4lt5B*rD}`gf%e-{Q z`ilCcaCgqlVK(A-4+@ZT+93yw5mS+-bggQg)-eUXuJ3#wgqFsAcPJBzCVGyq7Qg_2 z<_)*-*MUu(38ZsvGSq`K%!aD7yZ87EG+5UT!iD-{^t6XIQ^n z*i;03w_G+1Nzyd=*$wWuriFY;~7Ev|H*77V{|5@eGq}7*^D)Xc`%hhJwjdiR@f|+8^q{Cvf?V|2K@lGsmF?ZCm zxO5~J4X;eQzo`#rEa{DKr#j5AWgf_i^mVjq`MvKR{nQ0#B0<8`X{uQ$aXI91#uu}W zYf!KdEFxXu-R+w!dRjYQk(B9EvnK-)-JwXm<5d&cpg={F=1D=nPkD8#Tc z?E1$H3nDp|S0F3}=V(DK`SX#Ga^=Q#XoK2gpG#|^>$8ZDg;E$gSfcx8D&@<{PO<*r znudYL)SI)HLd8qBI6K{V_y|i9SSG{Z{V@tlKN|(wzuP880SAmBH4ZtIcp;4;4)oIr z!sA|lN)1uD<%*x#FSsyPovi{d!uo=^p!<29*C4c&!P=}=;4Vjb<9I#sWq~0xs1DKM*N}0s~eJ2iDl_5{?*7LE7L(W8d#OCPza0R z3z4CE4G0(P6CE(E-1e)iD>PNxfmV-G?p+qb=+nwd6F=q_WqUe?2C~A8)MguEJVnP4 z%C;p0)4sAQcj?FsgpU(k=s>g5lC?KkNHcOPQU#xr;C4}4fl3mJw8ZuuEhN|w(R1Qz z1sYtsq>M3e1zTIr9{Ae8@bT0ku2Ddx)^Zs{Mi9=$StAl_yxc{r3kdyfEF;nAubal` z-;LEDfpW#c@q`Jy)Cq&)*TfU(*OPy*v)P4$?*h%zOf$gX4-MpU;^Viczbu;7fH_<& zCYuSYcVl8xi&~aj6%}$0(p02XBqq*?(2A$+|6ozYlfi}wvkBstcN`-fQ=R`k{dS9f zzB$B4<^aB>{j;Hd=e?ZvlKU$GC;0u zKokn}BvlO}iiq~%t`eiZaxh0qWTW72@bvU(<&0a2`0Z}K zx{Kq?bv%_sY)bxS=##u&=Psad!nhvXIGs^SDsUtX+)g$++J;{b;(jq6GNGhDjXpR^ z#0}Z%M9%-L$?AmRj8@9$!J$z2RypCd&3O6>5}TCOz75A}MV%2k<INC;r`sb;##o0UIbXtFiCk zhfs#+ADNk9YeJJypZrbx3Lpyt0rXH&Tz*`aKbvBF)p_8t;Lvlw+;0a-Sk>k#vZ(YW?M)>Q50N|$xr|23_a;$jXzmd${UF@y-6KSh>(_kJgl{=?P^29O_$kPE7}szkzH!0yG?$pDl`m+@#FCAc)e#xF63pzT$XdvZ2zuIf zNXVmv0M}0r<^=_3T`C&3Xz84a0I$iUjdQ`Mv^kgjiPQJ}S6u%1ELXY(V_ncfs=bFc z>0Ihd3`Qu1xua0m1MS-Tkn~5G?kIxDZY(oHi(B_XhjpRVh>W3?KE?iN6Q7%PL8}}j zKi%bOjLE>6JEvOnX0!QTRF(y+98fqTj8KYAbG@jilsa=W!oxjKDdF8_vz`nX%|s{X zD$8`Vk%ZdPO`Q{a-LS`B!Y-VXV(jA+9PYCc?Os*28xCO8ACeW}QYHOt%8Tqa)IW(J z5S=2mi>Wr14JH04sRad)S{EFf9U2-jpGF~S#-{}8_6m9ohl++0W*mF`6f;HMs9rk9 ztQM6ulmx%hsmb7#jI}Hi^g06~2C?}~B`G;vcna$tJL%l*J&X$L|*MEsoQJ>^rJS_w8s`n>FBhpt7 z5^-spzvI>q?A4K?9G^AK}X|JpwPk?Z54F}h%!ud z(W_FTH2_HUz9pk@JE<|6GySu-+|hKq?q0gfz|twbgtCZwusT_p7So?AV$OWhbz23v zaXKrf&F7nf)gI$6g4H3qJ}eS01iBMoD=-%K9L?Na=+3%a^x@UfjO2PVFk9eUx(jb| zx0zT12)r{HurNzw@~T9_Gg{2&RHsMQ1;^Z~>0|Nlr_E3`2J)f^tkN6>XLTAr$Z zx_@zuE+@p4n|V$hEi^Zg8sROwMErrOdU&O4WaahKOPeNpgDWOeSQAeu9_>@WMsivJ zn!qB#@*G;DX@)uPD!)y95C&w@rr--|RQL}FrWOzX#W36|{Vk%&*ubqbU{l8CAkrZ^ zVpYn=mJ^y7)}U-fU{&5)TDQW0Z*w;rfPFe$Bz+{1v)zr9!+{)D4jQ74sfcLO?CvRNND7q4LT`4zzl@469RMjEfyCkTnNDK{D_b4`7 zuQ_F3yr0dhD>$1gK=(m3J>@NBb8%}pcW&r3uknf=ztXH)IQ#`7f<;jAVAIn}iUO=A zj5w#Cx-7GR6S_8VYoXBNn-v#HIPy3K_xO8JllAzY1VQHf0WNgxO6^C0L7t*I6GXYxUgxp4{J+M9;=LXI z%Y)*5vE;hS=@ZeF!)Paz?{6cIMEBxc#Y1*jVDk^htkP+~C9@v{si(^H&!t`PI9Z%6XwL)S_nAswiP|_woQw+EnR9dE_4Qz;vE!CdXXSMBlY;_FLZ(|Kw3H+m98zKX z^a2)bYC|D>caQFbx6+Ef*tJoKGDCeZ2egw?opPq$*(BO;#xBKNQjOa9m5o6Y?!X{@ z1T^=#kVy!MckN}<=zJ-^>%7jX-0jPMgfjA#6KF!ieh$7H3+Mj@%k#|&M1Hm79x=N@ z0+)MRanvl>RZf*-IS)oXPo-w^jnG1G$`sSBE$5N5!W`Y@EmHe7hFr*_1}~Z<8}GJh z=Q;m`lLZ*@w==Vx4w7+YLNJ|S-QBXKEN3F$T7a796%kg|ZOYU?u zXznU=VCaKjDg}Oy1yxuKZhRHc0T~6Q4ww7&;9P>`qI&bz$hn6RX3CFJ+`HuS{D3A= z!50&cGZ#SW(@N+M{c>h&mQ%+nza~TJ3%d-pD$hwa5lXupK3Lg8eGVFvle_?8ucjZ$ zV0BK6h*B5w91D4BlCGP7YCQaCh?Y_dE(fv)Z?TA2p)({5LIk--R1|SZvYW@bsC65p zP$!p;6c?U2gaxb$ncth#`*)eI*_rN4jOAn(I^|Ml4=yrFW{%qRo2@7T=fcRjlZdE^ z82XO8ln=&i0-+|xj=Pdi#RrZ-!OL*cS(Jb5ou%W?*FsTM6>sBr%xM!2tC{`&?^FQ2 z5q9bGpCI)47*?nI4UerY`S799eGVwNA?0b_#&7F0&dSl!M=BUfJ$M>MJ=a8tc*4Go zpGiV=Cl3N~I}nEioy&kU{ii80umPlgLTv2ZoHDdF)u9RBtc zyZdj2B5W#v4@Nv*?ZS=&Wvg9xa0HX7Tf(kH`Dkg$Uw7Jzbc2Xdm&5;5R`)!%*K8Kb z4vt-IXKHu&^JyXOiO%7BZ>$Lsr1O!`-fKj?MgEZ;qSi2}z*$?`MQe1%Y77wHj?89@ z{Edz3BNa{xe1%3t-R7qh(MLxxyQ?Nd&R~KzOr4xfZxjyD^n(tn$D_`G6VP;SEK?z* zO2LS-XJU|*r6wWb#&fZOt|1}$8yQ}$wEtTgc}H7~SZ1?SRK&3~?Zmqj=CvWWi!U8+LJ56v_|niTOm5)KC}u1@A4`PTd7WkTwFDC0GA% ziIGeO@qqB$rNz>pE$W*?O|w>z1lWf4zzVE!#NVDwlYY*sdjXjjtv4J~xoXGUDU-#} z54n}RON{il2cxQz1?HliT{)X#M{JDu=k9+CF{C_uG#D8g6ss zgfB4?22aH=Gsv@_L2R~W&pkeIx?#-I{q%rW36CcH0=w^m zD?S#F>QM6+vW0z>W)&WL*u1nqdd95yVwt+MV7-04fA zRy}6yF!Jh!9R1z91ASZ_oXAi1J5@%E;-B=MVVPPNxLMBJ?MtEqUbR?lT@rZ=FGZ!T z%~H+KZ3_N9XzkUqTqAY%xphKr;hK|3}(0 zQMpKSgTMmU8H+ln3NAF7j^86`vzQxJEr5HXq00_0%!+53^2U6GiAH?D@rj*KRX`2J zj<{cD5*g>yDIGFd1squkqu4QGf|$v;gq3XUMJ3X58||H6JJGc;@f=X0Idh3I~`5STKlw*9$NF#cJ!!C3vi;$A6D+Y6;)Mq^hZ0RtI6jSX zAT1K|(4aD&-3~%17f+)0cqNP=r}l{+=9~-C_t9K4je7N+ zEe3ziOn=$^UnMjr8MB^!wVu-x;|Ivg5O}|r~!fU{bC4poo+M!TOB)TR7Y0_K` z6t?@#=vYr;fN05MQqW=oT#z``6uFM@Wkx{k&{TsLR`n6eI2q3IQpP~z(u@Z!V{ugf zPfU~K^}xw>KZ=K+xu_6}*~ix+W#%=;p6}Q|pS{NMIcT!#{!oxH)YRmlDJRd9Fb}ED zEPF`D$X?BYFxs(vlVTWmhDZFAH@v#jYp~VgfQB8je83qkLlm=E_4~UySydhl+d3d= z(2PAAu438T(}tt~>rXktMED8=8bVk?MaK%bOhq>W%xz^YqHbQ`c-s!-8qJWyNe}O! zR)TlHrNK8)jCn?4)d_PTWelQ#PBKf3AmX0?$%08*094_=5UP;2>Asq}qMRCZ{`EJ0 zs1Wqt?|~Rm&D(bMeX2l&Eps`V|1^i*J-8`f937}i}^@#u`g;5J)4Wx$co)@-i+SYq&L zg&{V?q`fXKhH5jIp5JY0nvp;+vt+T8F$7?%+6VL(f@8|laVbal7y_`AxRmUy0pzdX zGYMNvx3>Et)M5Oo%ueJV9g8^f0_YK+fzY~4IMYZyU9U_tFwd;7`+JUcvA=jx(&8e9 zH;oaqwOr+W_UalWBE08l5XZHc#Vj)p)&i1ydpA58r5eCFh_5O&qCsDI8U-&TQxTOh zpj%W1tHrU6b8{UPXxo|!0Us}QKPo5y`vc~-0>LOHNhy^4<9CM$N59>zrwUeRA)1mr z70gkXSeNgaT{s)%aamz#&)kkWt0*dUdheLn8!n@e&-{G5UE|aD6mDdTjmr&3UNkQ% zUkzjvtrf{HXZcF|mfDPj-Tlc$*fpH!hWbFdkpP<<5Stp&oN*~G_E2xw!&}LwxvO}2 zt6l@<0BF!9IzplF*uy}?Q4`PV?GW7kEHX&_n<0%(j1>t9c1aN_@aYaG+reg$ZeFR#MTYEZkHYS=n(Mp`1W|j0*cQ~L_#3TbIZl~ zM43wIbX8`^sZt@)#w)QoqAB&3CKG%{xQdO-Fw@AL;)LNcceKcMH6by*6(d{Bjg&g| z>e;Aw{f$G)4Amn#@g`9b4SE+U$iJOPF9oX_^e@ygNV(O`LJ;jBp7)_51oB!qeOtpp z1u>jZxJ|V#yhtX6dTkx^)?k^=|G0vOJ+drE!QM7O3^iy!OubkOJ3N46&^+jPt6gN@ z#RxB9q76$eg3TWQIj1k=U~tyF1nge*$plwSw6$+JrZV#03JwlRps3uT)M(OG3QJ16 zC&Y-lOIzj3_!a6=bosjqcAE&POD*~Z6sJMj zB=f#15Kc~2whZZeEuKCMc2)+hA)+co-KnnvzV2<|EEZmWKiYsmCAIHBQG$PPB}A70`qp0e@_v%nuCW0v6iAa z5-}Siu*H*^f;+c;d|vku+W1DJw`1!2H}yt`jdkUv=+t5WZ{#(js|m%(rXf>CXe)f{ zVtP=3R2!+VAMp;IyS)_fKY;RkthWQsR>i;~QfG;Gc#OG<=tV)!^d@skJ*p`e3lz`K zBzF?)?%?P|`Dcna0;4>^J)kU=PQBz_JnO9_ViAq?TxuhXR`aIj(V>o3QKV(;B>Rz5d`dJu{h5%2 zE7!O&ldNTN2B|$Oo-O0~_A;+jE_wFihHgvBg0!WXW`eF+RJjZglBt1hxB1zrDdu%igd@NA)fU>AcpLm&X%@VmNDGv@Rb#ubV#*m2 zH>{BH!}yEg^jN1!-b^~;j4W0nc0K>1)s#rviEq@Nz*V`aD~q_ADZg068G6nT3?lq8 z3Q&SPA?P$WUU=NBM%<>ofQE6LlXL5!+I{KPjDK@RT)WtfD%p~x=7#aP>~X))%hTvG zkLJ^uu9Vt86mU$sayhKaVS4)e<5ZCAwN&6m%%e6lC2=Kt8N%bb@O3zHJ%vE;C&)>; z?L^t5LD)fAHP`xKSdnW#!}l!Ku72A+eYcS%-+r9?ZshsXHT01Cj?q0KXPD-5!tKo*64P0?2A5ZhwTXKxo59FYH?18v&gl&>fH`r;0w_pYm&nk3;!#w-LS3T!oqP z?_OB>a`Ww%|6aa_;n5*2GLz25$hV#TS@@(rQ7zZxJ@n_b7wjvDSo$B0oFlTT5V-Ft z`1s@(Xvh+rMS(s(285zuY2*2KEBe%#T;Q@gmrEEzg@t}Jag8G^rOE3yQ~vr^exB3c zMt(JecOk2W-t-Xu{Gmgw&-$r}n8sa&pL<2KV(q$YLkZfuHf{eXW+ZTNn2`Bued$9CTdCO5>CvI~B@_K5rWGRFbBusI?E2s-#YETg##|NiXL{c5<;~TX^5+ z%c|C?rc5e!1Xb%UF-=s4W1Bo?YNjTC;d&rjrjxCrS0mcJ+1cc_*}=u=+!;Zz&O- zB(&3@4=+%u73|>KRXb?-9^)eG^9A#;sOS!|z@gq;Sr9V^;`TW<;Aa#u-;7)>tqE=J zu*6DSwOf9u%tY}y2XwD6Jv1jv*8OgCF8j_szpO%6E8eQVi7>@Sv7KAvb2lQ+4g~h4uGw`}{8N zU#m6u)6+=aSJ-JHa6D(|RZZE}PcVMmL^hVXjTr6#u`6abZmX)MqU&L8@OVXapvHZ>vzHS?6?Qn}8(R8NFOHn1% z`KK-4`H#SZ`W%qta`kg8*OgNBnZJ{3#m~z1q~vZE|9>+@+lc-+!dAWCXDwG;_oCZz ze*K7Y7LVAV>wFG_;``)L$H5#+);Zq#F}s`O#`nR|JeQ}d(%zUweBGm#3N&fM9*(`oq(L5O9B9;Im!8SG>pcVEv3v6vtyHChF+2j{n1qdb?_K zP>mhuc>GJ}TOC?>O06Wvbj=zXBTB#~D0e)kESZ7z#`ABNaJjtpU%~et`E~x2`e;^e z<6I&K=`CFD%W5hAi6M}POkwE$(xymw5^dZw=hmVccpRZFK$IP;C*|q z)S+E}nd?CVPO?4}RQI(j<@c`sKgQs6brYv`1J5T*FVRxojv-(1gdFR4wtxC~@rP## z#)wIa?sJ*tLZoShx!4&6v?zw^heDgC6M~s&ly&vj@VyLNoA6NK`6B-vr)(+fzpy-& z%XlpcM3klKx~VEX z)aN7sr5_A&`;jJ>)ow&fOwoLNcS`G@=M5K25@ul)bw_eFuy);q+03GB7Ew zxiNOTF8HQY6qM5G+5B->6>bSFOTqeVPac^H+Wpgfee?W06|IAKtF0zwOiP7qHMP^A z*9e#3C!I3pbC18u>i@}H4huTg61W>1&vPN=>zwrO8)}?(3OQ0MU(fd1M<|%QOn{50 z@%?TVFBL>#PAF{$+ZZk`Bw__;6eD7jrD$Lr$EMaF{2I4`h!>}0X?7w)MYCAIq6!OUl+%kgXPOQ- z912}k3JxnbXQ=#Z!p@qLO4_%`ibJzTC~7;nJl{ zMH6YEi;821k%pSO7?I#ax6P!+Mh=`H8cP_*O}^yc661 z{NW(+eA`=Z$5;1jMA~or&`)mczk~C26W{LC^0u9*F*rX)p+j9+=N|UOf%xeKpsb-jqg(qY0U5M z*#$qC(2JGG52~MJ^9`0EeJ6Ir| z8>xPjhV~g!m%x&pLYlcEXfkN{ni-2iyL^|CoSOQg)L-!`D*|w)ga@0fP&s8u!@!{o z|DxX7mHu_(!0F`k;3{Sv3~qJ@m4Nki-mKNNzLbbOTh=wtoh1s{RZ(uU5rg?q*;2k| z@*x}@z^M<>!UTuvjL#{0RznF|GqSvB06by$HG!AlHEzaX^>NP+V5)rG8H9+Hvi=bK zzQyHf1nRiM_tabQy!BIw&9zN`qqok<`3=%Hr)cE*n_D7o=i81$qCI5I2WoItADh@+ zZ^Us}3Mv_AKnhJ5@sZ#D%;!gTkuUq2wTG+TliO3P(Iy`(gFGdj-wUT&qLPN9G`MHn zVbOfUv6zyHf*UDifk#njHdcr|{u^KuUNXv7aK)fYQHto=eV^OW-&EwHMtauPP~Cxu z(;@!rOrcB!TGU3%LwJkzKleTPa#+u%|HYyImsZs)PWrs~pUoWm!?}ro*7J@}1AJ8$ zq(^-q9PKelcvV`#^rI?&x9!U^{0`6qM=BjYy`MVUX1zJZHdW%O6M@}5npJ>^av_wu zG1|8~IRnz+`#+W`F_?OY5RV`>}QkJrmJm=?4mH97CAnt?nsLUiw8$~ z_f&(06(jfQ;(@F1E&Gz_d%Yw z3DX7=AGWaNIhU%~m4mVA%C_OPN@aGF>uGjZI~7xOzXxH0C5`@HuzAqVsRW}}1#GS% zj}L+|4!tm`v-phjJ^CAsQ)?j3w?kVu)$B zh0P1~5%MP2>NYU-Ss@y9mNR11kHmxOt_f128j(yX(6+w60#t9$=^4H>~$rd za+Rl)&Au$)O|%>({4@g@xyX94GW0EV==E4q^Kc)@RdEd9MOOin`T+NWrjD-gpQ;v% zEu!hl_?;PI9JH7zH5wC7KO7ac-XX*d3+ix${+oRz>=c-yAe6lrOXe0cv6pQi#>S5*UsVyOc}pLD@AR3)P|2&Mf__9yv%= zt0qiX0BZBb-?!{0u+DUMUqx;7p>fH#<;>FfuJJ zfg+a@qOd9*J>%)A2es3&Dncmy$1;a>*uSx)Q9xRh%fkFH)Ss4A zZzY*EF;1Hj&lgO{=IwYAMO@rpCdw1V6f||z-OkE4T7qip5=Ro(D~-1v1apSsX0O{; z!p_VzGZqWngf^{k^jmA?BzbTmUDsh)k`e>~lnV=Io_G8Y(o`; zrQb{0v4}*C6pvqeowti`YUc+Xj~Aeu57%sms360_rbT_E=W%&KqK#0YiCpn4`97%<-T)Ro$ko(KFT{g#k|Pc@yT9oC;R)~ z(8Q+Szpk~*Pi;)G!@7I*p&@rhXH|A?FV1hj)}JeW{22B-ZtXN{-s00aD4IK|#yg%J zeDt`QNE*ZDhnLlR`_;YH^ZebjHC4w@yY>2J#E1nqfEyGgw8v9mH))4~S#>YaveE-N z^*S?4e%$Ho=B9Y?+Ur|^7d=B!M2yC9d_8^dhKUI-Gqf{_nRV31!TiW#@8jEQ#qf(i z`+MNK%RcIUhr_Msm)|qP#EAOT^Vi+hYcF6W``hsCy;VTS&hNPs@VO<%kEd$xTT;~s zdHkM?c)g5CIO8X<^Gf`JTs?C?X5yHTbNCN${oXUPRI# z4aIL?z1H`r!Uy?_>kp)39<0N{TYT*87m1Dg_UpmzXIqyZ-8Vl2?7B~8y#eM9D?N$t zRUW69Rn z=kmRQ`#VCQGh3g{cYh}b@HHTR&7k*sie&;O+CIo_WGWB`uKV> z^6Lq?-24e#6)uJA5K=Bb1}>+RO!bJ-0eGd>LMQ-TCQNZ@t=GZ;0K4xJC&ut8y|8^y|Ot z>AhiE_1ycq2|SOOa6WXdUT5Bx*!jguzdIgW&umOd189Ok*ymqAK1sS-3Cl zomww>hjt91pw9%Z`H(=XcV(^!x@~>%Pxjs2VUK#-Tf&F|Mc5Y*I7u^f+iT0J3A3D?otoCpPg=jHVOvHTmQAW2 zrVoY9FR7NNDUfF>kk1lu_cJ%)*PD>p&ceQdzi>3fL2^;!B?r-CHQYuh?(ow!~N zYP_#Kd^>_g6V1E;+pD9te+*KbEOi{d`kQsjtVwQGn|fb)lcasf-hPO^h3$WU-RFYQ zjK8-xTdQZKju2z%;HT{}OVIwg+g7vqGt^2`cQ7wJ~vU?av)EhVLTJl zKzEV=ozD`JdF+T9_SvoeuuuN5AHv1}a2e@lJ7|Tc(u`{)2D*=?KDs6Z$_C(Y-akD} zKp4w`4C_Gq%iiuL(;v0B{$Ov83#|6+NMQrF8tK@jYhqB*kZr*X(y%>S&zL|N{H*G} z(ng@_8}C#79741JvjgX2i8VT?8P?bY$ON1(5ynmW3&-?4L%kP%;=@^*6Fr(z1(Rd2 zL|^|o`uRDw#UP$<#@fOiLqc>*dL)OAd+>|k{yc_t-xh(qjVW=Q0NWUtlhhIApAkFY zbR~a>!rt@!^V5RS`4m||oBLeHRSG=%`y2WZb#c=U7xgoB;lLS@#ofTPU=l378U>tl z?995p^p^F>0lQ|Ow(u47f4YVlMCQj++I&sxcYVFL`UUej`Z}DSAp+I#Wdd0%B_AR| zEQ9v=IeC45aS$-Qhyi8SYX}>5biX>^!g{~fK8k&FReyG_eCfYH1bPeZfv&>)F4;#j zqiwb`)?a-Po>=wl5F8lFS3*Awia;V(9kZ5$uy5lP?0kJ;jI@$!ff^R(eZJ!Y^25bj zoIz(|a>_gLiByB@eJMdr()T%bmve#qA&2cnL-59V_p(LwZ8>roZfAh3Kce+@m{|1| zJU8f_Lz*VA2?_uL`pq)^d(@h_OP3x_KApMOz}Q5(<`7^L0}Q-tUqYgyOEt;f=zj~+!9VfXS90XydDYDo;- zJj_xxj^ajDydf8Sb)YzoS+q>u7s4M)i*Gkphz24rYXJFnWIk!6q1TNyIf^(xm5b=y z@bvgNV0m>0VZL230bA+ax6a+w_xN|}^Rrvq(hU}neFun?A28}}t4z?~xa|7`AAus? z6}k;j+?W8C*{I(FaCim)k}-}4+#%#zg)LAHnSW3>-2SxQlHGOTxDq^?C;gs61Dt?C?TCe8jtb2|*&FMBNMURa^t^?}wy0 zPwvB}mKUsFp1wPEMclsIY}sKs87J<)-4M|IR&8r}joO2^-p!tYUhG-%i1Xb3^%3GV z@91~;v~hj20xSvqEkCkeuarT59`BDNLI?vjYi|$!9BH=?NQl`Wfo--tgP)oE9*so2qdV zRki0IxVMX#X!AoupUq_f`Szk%$=3G1;&V0-ktxLY;u0XI=gUh!{pZVE5)nxc=u-<}bE!=H<}8Pe+S7>?7n>MrWz z;1&x~;h29%r3Z$NN0N#%7;ErQ$ahb82ipJ&;sn~yTYi)CR1E6-jv|?fc$`%eHl`C0 zN5`R!==eguo&j5~Z+#7{(<~g#D#_{TeqDcZiSN(6N;3Lr6aR-RjKv3$5|Z397yYK1 zPA1`!e1UCn0w5d!8B+(+a6It(4%rKOGc}F1H%t9*gUH+Ec;Y>aws0=NNWb)uZGB}YCP}qV(*oXuWV-+?4N2adAi&D7i-*mz|B8yP zdxN1pT7B9r>P~k_XRu@s!kmtx1a99hwKzr723tUdhJA=T48bus`->Zi*6w;59oVkf|W#P(Zg_Ig&gmMjo`2(3ITUi^rO)4Pb;pLL1{7MKqMJ&{^4H`!*B9@{^|}O z=xyzJETRlq$yv~$r771HQGE1()1MOeRwvl?f2~N>9`UivsVsop%Yb~Q8aiDmH_Mc- zP`%m*gLbuT6Y^C&a`K3hG37g*PrXG{LY|OW;i#tdCL~j7s6_s(xYY*Zi_b$G-6H+2 z$3mXl#_}K8SqG;z&=pgvO$K3$(Mj7mT$%b9CIlg}#T*RBMwsA$A-a}r+$4$PEDxQ9 zH>!6VmD$qoLw1yeCYq`o=qTQ5f{dWdJ8VKtpxQ?eY9%BoL?Sj?$4%^+QxkHBp!dIV zhr0v+14Wq}Oy^cEEVcU%iM0;WSIEc@cH-^^tPqxv*yRhVYnKKErA2BJ6AR>fkoN8> zp<(6w^G8d@QwQj@<1nd9%1Ao*BYWs4>ezY2p1k=wt>6&(oOV{)V3t7f5`?qrQBTQu zPl2Q|AsNoY!L&hQ-#z5sgITKGi$O!Kjc0iopxD#C=XAm4**G-#_ZYNu+IRwNc=bm~ zFjCX5_{i>5F3I z4>HYMSK{!%_IE`@e%&xzpdgkCTtMW8E*Mlc1j2O(zcWcu*lRQi@|nX`SyM!jQ76Mo ztfl4uMSdW`7v1oZ;%V$&QSIA0rLhNFd) zsNTpX?z75}G&2OAT0q@MTtW0!IOI#D-qQWU%GnC8-0fGrZ{)1aH}7x{BL01z3LWe5 zeDyi(l>h5_>e73J`=4wlV8g2eig*=`Oho;e87mtG46GB}>h1$$m4)KX-_zcm&@-BO zZI&@SsS&0C!WrG8;u0Pv#^7F`M-s{Yn`0~$l55Nb|4AjTni z9M7N_YbQMkh2(TCR2O6;^_I(1EAS2pE*!X(@^H8G&u7;y9=D^pb3;6A(VpWPzrnDoDb+kknZ4MgFrBTQBYs(lW@ z={l&3#Ow!3D~^Jgr(>oyIGmXNRTHMbGcS2x7+GmF6nLc81a=7I86?AT<*{S{sYtGn z#zR&_=r4w{Bpv5`xu52w(EgJ547lZaOmkA4kZ;Ym55NwV5{v_6|I<<#$X*6v-$&X- zhSu~7_Vs5|tcm@wU_^EI@N((4(6?dd$WpJmr2LDd;jVXpIfLuxetp_az=&ZfGYx`wLsF+M?DhQnVir+69EBn3`R@MS(7{3Bp8SX z7;^7YJ5==t{7Kz$uiJ&C0wQ(pByoqehgJ;t7F-oY!Y2D$&}?#u!j%b8>}Oxi!hr+g zCI%+$MEjC+fDUcdA8MUj+a+>y!9gYH&J; z_zXXIe&RCV5+gDu=~dX-giBLgj)oX*f)Rm_&ufKf4{@W(dm)kmx@j2oYLiZXhGkDQ>@|`B0yLjdF+_Vbbohy z59jl`U9E9JsbQj|>~?5rW)RmdomhKpnTlyxVbk8+F)g(ssDM^Chz>@m-`N-=Avnk+ z9q5;`&@3D$ZxE&Xt&17%j>T`6NxgoFJ8ZEM3B-gMgdp96Jx| z2*VmFJHpNt(U#fSWWXX=d9R5v^MWA&A8()tw~PfJ$RcJ<#w5aj$NcF2dpaOs-0`Vh z-1~1{v4Z_W`Y5p)yw9_V-Yj4ig0_ED92NtS^1L&Fr zZxDRMRT2ofWC)3n1++8|pq9KCPTYHI2iXNQd+HQ$+3g)|sSiSvEdU1jWnbK1g&QfM z^Xg)fG56!AVNiJ*&a8-NNjmLE029rp)`;_9lQ|F_A*eu70H>vG7CMTc;|93#r$&T+ z%*23vopCO12ho-dvz!hU629-=G4}QVw%NTPYXbb9y@MK#7xc#am`p)s3rdKnS9hP$ z{*Lt}ECN82hrLihOhbU+o8A=d3SeIRhENpPwyD_OEbsRSnku%qaZJU5fNhv{rtm}5d#fdBQJLf$^&Dh-?zt4yq^-|i?C;4&9lAfRJyRd$%r<>JH>drI1s^E%a%4P) z6dnOdP*`LI9TU-?_QOhl#GJtrEy&JSB-OWIuvO<`oI zxX(zS4?2Mfn+C8Cfl6F~n4_mYBqaZ;6`q8JwHp$cFZoc4^boAp(CCRcP9#6*Q8X!8 zv@3snFZDGPfS?z`Bp>^!+xt03c_MUc3Q|W@C$^%pPDIMc`YE-6J-4q3uGnt ziS1@&g$F^g)23fsf%!F<`EBp_ND&1M4{8J3B`OBW#F*}qbf4>By9aB!XY=AFhz>|d z>Y)l+xL&bT7Uy!`Y?d#XXHW*VPrMDF*ssqTb-9N5C%JKfTvMM#*9R~HZ4tk1Z&;;A zctEt#!K;GV0(D(WVH>y;6=lI9&az|n;vb(_Pza(~P`NuRnc-_R_C=OFa>HSL8IZ4jvWcqQ-P3dpb98|r-+|8j!0f$aHdWU7^1`(SGmKGx zEk0Pm;YS&gZlmr!Paxqa;m%qzzZctg(tUs_F#5;DJI%4>`~t8{3qt3{xV0&QAyYCQ z{yJV?)#=@Cl=SN2R+J3~>uCfG7qL;IXN!mEA`^>I&=>A(SdQ81pyJIt-GbwwAMOnZ zzDA&@=>5;8|6rWbm)zNP;Ef@A*7z;&cE7{4K9zZN#1-RA`MbyqMffkyh_J`U-7&MLU(KC z!Cb^>fP_fPhL4f?Rzut^*{dkILjnj<7VyDahlA(H{Nf$@^)yNU+H^3qHP9oBps|b8 z9iw8%#sF*z0hB3y`=$Zj9nSX$()t>_haO2tg3t*j-8Y8 zuCOHN_d*_Agz;40&8*;!HDo<65()6|Z0Lq(MJfTF)8h2*^>t@20eBMgwpLj?@;F4^ z?|SPee|zImDX-hi(_2=lZDH-wyMPVEKe$`D@P<(E6ZhxiLqZ zgAc+6Hy){a1*)zSXzzDDdzWv-wvdU8)I=i-Vy-r}Bo<4wt|DtmGN)i~a*5WaAgCqr z(|Y*3==tfQR_iMOhM$9ZXqSc`Ih~$V$9#W~s5V-*nwToTcs!BI%zX*#%4`cpL_~U`G{$j1>YJ zhmZ!Q6ei?t_&Q3%w`!0zCo+8-{A|#0ffu=P#L+H|$TUX^i%vC)X+%fx>Y~bN`Tp2^ zYDAN^p@-B>Ol_0>mxp63pF_bJx!WTH$@)?_C@lU*kN!J$OPK;g%ew--Hiil$KQ;7m zG-?I$COIX`(l_8bd9e}|TWfnoGQE+r{j$ykHz=Qn^GREi>_Y&&`Tb^SLwEj3)|3DR z7!Q%r4KKVN>9f2TkyH1_h;-S)Bzp2W@?)T}`YC{^`_sAw8V(hh>qhq_ijVzfMB9=1 z;*%fceAS4PObc3&_Vyl(HL*YY`zs>%GM5k0X<7S5A~*D7Z#=d40Fr{wWQ(L9L(#;R zbewuhY18YsX+gnGNuPl!Bum3$B0$)Vb>jzc_dxCV*Tf|dWT))RObQXlDy>^`JEeWQ z3}UTSKcpa+0`PVlmn8i82Q>LW-sq~?-7%_bcYf0Wk#GePj1Nfvs*#YNure3VYSQN- zO}u1FdA3CAm^A~yHf*gwWywkY(2=Sw&j%IJBj7bp zx?E5I1r{n8)FG~rL1~!c4}kWm9GBF2L6XV(LT(IY5X{tp896#c!!Ghn3t%H~B?oEg z&*JiZiR|0D@hisCKD3x#G`Dwau!$1i^M4610HXx(=I$P(xn}*GyP!cg<|i z5>K}?J`tDVuY+{#wRl?G>{-;k_Ps=+xZPh^D}bFx8EzG0AqkoT_Vnauz2kE=e-QYx zAuIJ;*jacokhAwk4Eaj4KyBAv9DVN-hyG2oWV=Tv?4)g++7@|vbU<)WAn_I1r_dqo zhGsYJ591}8OcrsZ4kp&v45Z;?%ZWS|fE`N??ggBQ6G1Dag|p6B_(-dc3~OdLm**eb zn*|@tHPL&3YlwvZgy&e5l7I|a0b*f=QAr8*>L@V>ttsF+wMi0`GYsa)))>lTL%5;a zNeBQz7H5boh7>kO4b@Bwgv<$;ug^b&Pbd+_R6RsscW9lL-_RmL1E2SS^*t*C1h&E69D#c>%=I}fUC&-*Vr3^Y~2IAt3Xx} z|4Ym{nmc?_Il_we5gR$G`7nyM??dzna^q#9dqv(GQnyCo*wM~+3}p9?Wq(6R}|khr~dRW2cN1L_BE*W1m)aA$O~#Y-@Ce-wv^_h`2F z8Rrwk>Pw+Wol>NP#Svkn3hT2~Dw5ur z1GGT=`CtC|pUA}L9XS{P9%~6x0v_CYCL4SNi;B#i5WZ@hGJ)F&%9o&{l8`DbyZJ5~ zcAY>IaD*qF-V^aKzo8z@zqI?r00(?Y7^8?40ccLpdfY#RK?l}?UaTcbM_ppz!Pv0K zpfTKE1T(~y`r?V%DyyW`ux3PV1rpnD4wG~?%+Z8gdi_BxxRGQ)*VMIjO{Fsqk%j&P zCJaJAad05@6^evMDA6(DpW1%#2WSEXoC+E6sq^WDsJ5HAxXxL^;2rV4 zuqhB&`&~iOS+-e0M;`7>XNz5ENEO6?i}ocN459@Te}Xlt?JG{k#B2Htt~xqwifG(rJ*1%FVu&F?N}Pdn$5ZT_js;56+^GY-hEi*WG3^HO-?rA zO1mUl1|U;|ku}*eAYlQGJ@@rx~TXs zOc=wJA%5M8*mTy~D2IB*VId`SiEcrTq=eE0fHgzQB6cN?1Ke1!6!?YaoI`uT;QZQ_ zXz>*00VzY{m3bAZyw_nZ;FLhTIOfTh-OVH_fwjVj#Jq%~d@US7Aa;EmR$%aS4vxuy z&5Ec)shI@|4Y?*($ph(})#a}FL_Ur=T7&ZtHQ6`UUQrzptP zIkx9qQ4pmSYv90w!7)(3+HN-UTQqfEsq>GBqu#xvAa4+m7S&x>%-jh$b-@J_c8DiF zHfT{a%FumR>_@~y5sW>^1Htf0FQKVm6D+%h7iCRfT1Q6z7vwFeCdNu;g*RgYiRj?q zf&ifV7X6^x8}hptViXmSo)~(I{oWB!s{=-`uVBiBAoZths~ur=2>{f@IIs|_LYHb2 z%3h+H>S?LwE@hBV54uePd!(N>A{1S;+}4;Tz& zNFs+7X(5m)VCZzi^ywHmqZiV}M-mu( zKKJ*#=7)Sl)np)lwY~Xw7|Rd&_t-BoOIMosJ9yNMFa!_z4WX^bm(L!Jz0;R3xp9c1 zdb#oc&hIY$zZ+1x`QPs^KA!*PHy3V1Ubhsa>}S8z7!i5o19BCz567mzgCp_73!t%H zU~k!6o4SBJ(mQIN#)3A~;@AYU@eco?yN>Oc<+jYkyKu$Lu#qu8=yZHxr0^t&q{$MY zEv_ut-`rEoJH+wVh=ucU zHVS^IcxZxM1?kUXb?nGRr1H8~M<4pdlO1rz`n*DoZ_fC6#P~Tfm?Ctj!C1hab}(XF zIuzZNL8N;3^Nu{MR3x|tvz?| z*|VNeu9*fBmgCOnKp(L&&>*YE9b+I7tTY@RcFjUza(#`R(^)~YOmxQ4im9J_VsA;I zBJ^k~yF@;1gF#km_zwC^5KVAYV&;51O3X7h`4Q7I8^%v(>JI6iKoL3+9nTXk1I3TX z1z|vJ=#eaihrAd=PCT-SHhLK`FvpM7D->6WG>;tpjmnC-DVCw$t;Gq6bJg z_5H;N=E;*ji)bq>F#NYgBC?^e8{;upte9k=A2O{oC8mVf*)T&4DU?`0G3DEG!+bb0)A1c3secDcmLFdD zI&gR2gO~r#Gh09+!smyB@z{{6-(mF1A6|U?!~BI(r(+H@L(?)GVZaxi^HKIH`yOKI z`Y~Uf|ADJzy#>mSGJoY;_5c3h{~u*gzJ5R7AuCRnNvy-**l!d^lN_@tG~Zfc_*Ih$ zz2tE*LE>4>kQ1P3g=hBBs53xnoRt81kpy3FKa>7b&NcA9 z4$pORrkJTTd`)z$s4JKLi2DtCxuNhAyt_Tih`amHNx?@(5U|Nb&zM7iXZ0F_hwpFw zzaUTj<=jvANBaUb|Mc{jVc)vK;e1TkWH!>wkn91kQ0 zN)g$)FGH*XcBLbuy(^;^YHsZ2e}%5MlM$!M$qpO4?;rMlOCaBZh6!vmQ3#oJMC2D- zC*8`(e_ygB*DCN|0^R`Y9)x@@Vrcc47>Ol> z7gwJeaXe?Etvv=j{X!;aTXqd33SrqNZSmb(?b-!eagk^jhzc?-mT0~Zwh@gBW}XvV z|DO_lk9MwDx->AnxmpkqB@fdib0YF@La9k;0B{}O0FOzrNVe+ivvkIgmRYcI37vV3 zJ~J`HNXrY!LC~dy*>dUjO8R^eWg-#B`#VtSH<)$-qD5E7?_iD*p%ua$SQ~_NUx85Z>znc&t?~Is7Hm+wv+#(gzwANcjW?;$v zm*tb7YH;pDxP~xf-ygCeK(D^}`@IJGBE4!S&f}%}Sga&kJx3m(;dkw)EYnT=&AapL z)#W4u-fB$-Cjxz=jkDmfK%%MOp9e!iYm#CDbw437MB)Jkc$fU;Mo3yo=i^k)LDSH8 zRu)?o8!P#N8`MR|W{EqE+*A=)g#<;oOG8l2&+Ka=z9=m!B5;0aTiWoR~fEst~@JA5L?6$YNmf)AQ=_38HlGR{Sa@d-LctV~$ z)&agW_>!lQ2x>|ZCzFZl<)6@mj1=#_H#b);23^5GcAx@T9TYDV+E#c7=*+^JJhK_0 zpGV?pgK`Eb;f*{9Uit8%AI9~*-JpFin!O0E3<}W%ILIx3D8W3)&+h~!-(b0k#0ZLw zueX8Ro(N}$jq?|b#0_M3U)nvZR@bD0ZV0E6m_jr-ZIG9cXj!69Y(KSReOHOxmN=7^ z#6}orkTm$nY^Re7xZYaXAS#YcwLrrd_INd!C5myG2XQ)1cgSDS$=iTCn`uSSdrO`tKtBBHtHoRc!ha0t z`-%cVVm7wzW+VP)oFM>3d+H4WNsA^D+YNz+xMYVwoGy^4oUn`XrU(j%!aZ>c6~yEk z^e8y9K;WD6+s*wBh0-AGB&7j*_G^Nu30i!M=QtlZ#3CB79+Z1Gt#xH5U0WQO5!JPv zN)c7uU=TxbHOPQ@D3N^@*6$em12!&2P|u><2u!SGfm|nHv8DmFM_mAL;u4n|JOi{! znt?$BFtL5pD(E&@0;{HKVC#Yg-l_c@v9hSKMGLIY%+X6t%vqio0{5SR=n>asvbqJ&fPHhP*>;d0 z036m6oq!`n%r_y@FD8Zry$^T{Q=AZR7QQ2>;YhXbxp)xFab$Fc1tg>9C@SFWjXc`? zVlW|P{59tS0}lFw`b#i>CXjvp@-t;PLH89nb01e8oWs8MKVxV%M+ZUJ1X%_UxPc5H zWi`exQtq3U-Hbf}qS9vFy(`a=qr4DWMxU08L7A7Xzh%rwNn@eJ5$uW$h*(XZ=Qk3H zz^V{p1c_33#^t=WY_%Z*56bEdNE;GC z%VVW78&-%H$2AZ$Beeupm6d{R&I$M+b}k(GA53;8Ak^jir2j3?-Um~6k=vL+*&4fM z%{!k2w@xiNw%53Y*r|ykTpD8_3f#8wTj55%7|>%pbE7;TbKo~1d>5F)kb^Ut_&Nne z=@H4AAHX;ftCeS^E6>JhN}F7A+WcV+POGHKM5+|km>X1MZh}1`;crqKjHV%C)C19vz*6#F&~(47DwjKuChz!8frT*s}P6NTj^ zLqW<%@qxiq!l5rDqYMJIJBh8>vQ?jQcB-lso7_O@Xe`#{G zXN`jf z5kD>{5(Hw-2BwNE7A2zTE$L>GFL{o*QoVC<+bkn31Lime_yA;Kc2-R)t4<&ihB@_U zsEv>hC}KqC&MOpA0o=T8c2^+vMKB43InZNQvG?d$*764J#+v}7*)vwmTGyESmbH7$ z9R*Yb9HG{Wy*&dF-!gT;1(g4>i4a}Ow)hzSLtJREjspC#?I+#eZ13TuE`!ep-!v-K z$SDy#nG4FhwjeBnD(LOa_acc{1<%q?O&ZQ|9bBi@MpZV*tDG7pU6~8IGBkgb?%{;9 zW8+*s#eyDA&=@S?YYy74wow$aku27T$Pr_y-;iR9NgPZDGAE&}rJfJr-9lja9T6T# ziQtO3WJyM6#v?ji61q*m%Ga++eEdLR{*u&;pahP z{*k16V1nkOfh_Nfk%xhJ*9u}OoFWTi8;%pLp}_^jFQYKDNl1JOrW5ul%eFrfG35CK1)F}9(e-Ig9tRQVs{9hf(i``kr!|_jCB_aN(>&3CAvNU&YcEy zn5M6%Lf~@^(r~!&5egFK(DzE z%!dkc77&>B`Yi-=jvD{YEK5lUKx!HI8)^ zIl6O%Oe8koOqjIdEani=B!)x*84cx6JcYS3SyJGG;9ah zwB7HnB)Iefg08eE3pU6z%E3umCm0LRxAPtQ-GrEetc(O(FElm5Bq*3dEu|}vpK=O; zqk!qbU2}-W^eWg!TtTVa)+XJrljI*r`q^AC%cTF3^bIMrckhzD>Nps>#I91`mE^fK z$4m#wdj)~xZ4Mk%I79<#HL|ZNwSZmWE;4_qNL9Rb+VrcE85>Bg*JSI6{LZFX%_WHI z+LFi=ShPgDdWie2c8$EVQN_t3JC#a=h?B&k1fkeEfR>eu14nbR4se_p5@C&mzNo%{ z3BJ`FtmGz-_FZV*KtQ0qV$BS)Z2dv%IYRZ{-N>UR+z+1#3Os0qXG@GJ9~i;44c!$f zCX-b1g9#c?kmv~A{4P5AI*8}K(Uk8fVtCJp|EC)hpoqxOka&+{nO>#20>=Khz?WT= zD*z~n6aXil*x)ogM&J47j*QGW^7sV>w9*m@~)p-{5-;}*N%IcOpYUn7;QLAq9Wm4 z&pE^m9$~Wn7hnWVy=V?geug9xTJ?y`m#nm9*Et;MHBM!4O#yhSi)T^u5HMX*doxB->b>n>>{2V;8~| zyZG=H!%2RIU5XAb=s~KIGXMpk38L3xl1V;_Lacc4v@Fs{nd#0XAw(eFh7t~*lD}Z8 zy(6(y>sdAOzr|T>2hl?_#VM-2V96X%D0y;*p==?~r zn_-twg2zM91{D9I&|H$9!I4^|z+}R(-6ZH~(;=0Rrdbe8gRohRj1tmDXYYzE(bR_H zQl#w4wrt-^lDRJ7=esq^nE^wRH3*!{G6ENd&)LaLGg*u?jkdxLo;J_{#6iSR4(i+? z0)r?}*(Ixo)}f0l7aPo?wgT-$w;YlZS0nJ}5UXP-qJHXygi-^DOEK$sr>YbppNxmt zu7%`Ay zqncBy-C(Wh$V~Ol+I`ROF45xqT+LGl7%~+fj4ZFglH(36S;3S8(;#%LP0vQfc%(1O zKz?{Jzq{VOF{C&Z5koPP`0_*f+eS;5cj@cJbjpseFSfZjxLj4vU`L$N3h*$kBLlhe)%BD zP2fv7(`3*LZmMWTM@B=pqZ94|wB=muh=Rtby)O3uO6NOMM zJ=2_D-|s2O#HWx$ggXyM;u0hJ0#X5(hC+=`oYA$3r-)4^ObaAfW-+55w0vV5?QdQ) zO&Xy43}%#!@nLh>T`Vb`4xc!bFpJX!rOg4c2xOs4%T)ZeLrEGxtnPi9fi%t-)NU|-343ARq4ymp zd0t4wB1}w(jTGdy0!MhOxdYB2sWy$ot$3#C0FoX93I~)2Yn>tru}m*ewgPeuxD#3J z{mqpa^N~?B7=k@$hVCTnkrAwuKFF$q8ijTa(nlG0>dg(v9&h!Gk2dgz;&1HuO&_K?k_=4%FsC=|NJw zGy|yLzUYm@8W{@PFNS3HjKhsLsv1unKpb&=em&XtZy`?n>qr8K1h>SA7Ma=FQF{6P zojUMpsGb%L0y!HLOmt&t`k|+<@K#5?Lf?^4m5>6B;^DxeMk(%`<$Izutx;%r8gXNn z@t@=o&4PAKo zp%+FasSZ7mn;5U=n3=m+Tcbh2w1pBEL{kokr2yx665@hA@EqwvB>}d(J29nagJd`) zR@^O1SX7Bn8^Qx3SxVw7k|2B_lS6;N1^V*@A5D`6ZV1SE%bNK?VZ8+A{CDJWptphPh*2V46f83HBhJ$FqL1a5cjswJt{ z6$?_pk;3(0W%I0I#JPhkhmB5A)^*ijcPMhl0tJVpaF_EJ%$hm@ZKzm7w7lXRZ?+AS;y#U zznG$@Y8p`F;H>wgN8;=o!nlys{t~mBRi)UWz@(*osM>!iTx(Z162z^!#;jKZk|C_l z|BeN^H2;=@3lQx)g7mzyu&YH3d=VMbg>1bMim@k}snB~?0fXB4B`QYH*(r`1A;cxX zmPKd*#fKr$A_@y)ymHN(6{XGSfo9$s;5f^rgzkglO}yA{HCnTm2I(XDTXWZ{p^4n}TQb z%@6+H7t;+2to#E10#aZSTm8LXp84)-O+)zyBTmwsVe9q83^oyoojU2z=UOuF7 ztRXxSNVj;aASZ!Jn_`@W^apclFU5akjb~q6A-~Ij1fB*V#!38?= zwbxXB&e_0>6eo4v3MF5 zf#8QoIT{ekVkmux*QM#NtnMaI1e}9MRCTd;5geS3vHPgR_+EZFug<%i6r?Z;p|Ma} z@HcaD2jvKn(?-*|u%JwRtC|hH6Qo{vA7l%@cd~1)SPqdU9YHz0#(f79hMYOoySb;k zT9LOV`jg79x7D9a&L(bhR;BtA50?BX0Kqpa`C{UlOy60_c8T#(QAyBP+iZznF5nuX zd*Qj~*yse^emng=+U>P-Pq6I)u{Xd*f@fPsQ7nUCTf)^VhtXi-TcmkJd|f_gev5o|UY znwygnJL$Vkl<~};Wf`L)gq2QG*fJ(OiSnwT=;BOie6Es3~#B6#F9(Ew%Y_W7n2sMyLi|Mfeb&Mdy*UtWo6dM0`yNm&3>C?=!U2?=*>;@D%$DP(ulqi;+cS) z%D0UT|L{US_(HAI+4z97Ax3d|Y)k-%{m}AWz0z!>WYfS&ygskDlJX2B5(?`x3HFGN z8jKjkM8x>a+pwPn`uBVgKUS3CY()sNCIc61T}AB&(|TvZ7lAw(eZ9lcP)6k4I}PSx z_dnQFvp#~%SIpvENwgu^lM;hyBM zIf*C9R9^H?J0&j#|5GQHWIsKY4fGJz9YtQfX_m@)Bu{0Ny61J z(F`*u`MCfr?9dPi5Vqt~C04{5dI_F-yfbp%nICQAk?O_+*YS1-+tNX?bJqdL`QTbRkv;pF{Mmkf##|i&1NwbjjXpSvReyTXyuS7%LU3U9QgyGF1>1=pcDfByyaUhCoT;NMc zyd_K?*^q#7G9W1QvO%8*A)P4W7T6SAvxSWlBe4)^G1=r~Hj-{~iUO z4Ukx9V*ZzVif;B0#Rx<1_KWv)E!p*GX*8SDjfsI&3I#O`-B81i^A4F~|)z4}0YE{6QHaHpJr|fwKa* zDWInTMG?CXhQrh?jjV%nQOU8R?b$!o8joF9Ok6-ll@R!qdw(zl7-1;HKCAQWaxl8| zK>gAK^h*!SLqe~FwL>k>syj3do8#(*Mh0nVDg9McSByG!gCRT@zdIwp z25#Lvp!tAA#Y@ayKyfh|xY(8Bd$sx7u~#jG??cDY^9)+IGTs(A4&O<$_=k{m9 z<1TRK*6bh^zmyLq@mqi>C2JqDG0{A%JJUcY&q=7qC_#jcKZTAM5Y`%5QE}cuf(bNC zH_b}p(5!7B7xCcZNCgc$6T4_R2e~IPvFXm-ve2ADiiM+&0GwlEszN8qWkQhANcRJi z@`DFXVmUyeOG3&z=9t(c2?d`??Ay$Z2FPRDL}V!NB)L5R@&&*@+F>N~1j#DWv`d_X zAYMLy=^+~v(m@e38fJhijdb0mW~iS)U1khtZV@+n(Df2S!C+)C$>)e1kYToey5Yb< ztZ#n{5-^T8i3*HNdjeCnG6vq(+-&Dkxx; zpI8uSyg-nlU1<7u;quN7P5;G`Xi^twQi6FQd$L+pi%fzC)|`uD1m|9(7`XWK8V%aW z{Ul{>bEW&8q5F}I>*#EO1jq|;Z)#JHP-HJx9Eb&&BZO7RaDecxH7TA=Ah+PQsmB-Y zMRxC4!!b}u-vdY-kGXW5NM@*!&40&-=@l!{HX$GO9%2jLpieRC7BN`FJxCqMY}t;h zD+mM28#HND7fi55zni|TLH@^f7^3b;?T42^5_Yi+)*BF8H-Kg!g3u)O1H)pJ8Bl7; zlV#yL2m~Ay$$oYITbyV#I$`cBV$S>`QnnRj1rk#woA3=}d_;q9tcA`GFQhrJtr1_g zoDxmIlL5QE%9ebWqSnIQz60mm;pYYz9%d5^ZWufNJEG!W&R?)O{m>jtP6bBB)RNrE z*qq(5HBGx0G3GslF!+EzTsk!`G!^}VPdIz-n%LmMfzqZV$VZr|CG{J_>BM+1YS&;m zXpDIgTc!2p(lIfSKvO8uMr{I36MrPcfcJ8;$V)~j0x=E8BDlu>y zOeXSR0+wLZkW>cYQjr6_;}k9xXDM8YK!`+~5V*7=xYQC{;$8%>bj7Uh4OxUKW^F^8 zOAD5Yf$9+cLcx;95CaBFV+KqA$Vz$4!Z~O6MNXFkF+Ih-VBz=Xn%uQ4>?*1OOIrQ} zfML!_HrRF81oByLckO?$`3l;6$BE$3vHE_0k5=l-PM($4qzna_8w9)pcMJ?q%srW) z8xv9Xq{lM`QHx8K7yDHW=2;C}9gc)>N<$f_<&D@3-cza(G*S`uCh@y8eM=SMi=XU{ z(>5;zRy2%iK9kaq>wYC%|7>O$} zCad7Vv4|E;lyVfbc^E~bD%fUyNksRD^IH-tG>KW=B}$_1D1)^+kmD#^C$e%x5DrR4 z7`H)zjsOetUord6X6K^K4%VAU;sK$UaXRf`$XXb-%fi{1Um_JDm$}GibimwDY)OAX zNt)Njs#APkf>mdaKRJ8Pkj&W6CZ*b2wENr5JjxS*$Vy@Ga3p!r{AryyKCI~BLJjsL zCCn5$V|uxA$eF^D41i+58zoW3H^_K4&6iAes{QHH{pB12)!0Dk;7%||4nq1q8Y z0w8FB|K=G4bMQMXFy48VSR>{T7}UYP8M59W)>7shEj5Ob=wp?5GSmGY(ZN1A-7kf5 z3rK1i5-XDki2|&VwM885p3-(T8|6z`vH-$zqiChb0Aj&t;s%_lH$9;_HlP=mYE*m} z%q&(f)D@eG;<}g+_~@H<&$y7?!P|3+%uDGfC~oCX954Y4Pcg84Bt4M`4B(S+ z&vCIbPU^<_5D8j!l5(?8{S#Ye~)p-(rfgQL>7295Km#r#>Z=!J`C)ST=t)svK9s2VtZ@s86--v zF}rYNc0sppty=?F{0?N|$ijpT6_N6ws0KJuR&KZVK#HC2>UOukgFG{AH^A~wmk86a zzZw>sBIP2l-+Lk`@h@$|APd@)mxPI%FU08oyAA*@=v!YA5KbHr6oH9T^E(nR zPUh?_zWoe`hOIL&N=zrL4&Ia)JQ`kv04Ll1sEv%x3%-D}@__@4k=5W@J&ZCANWtp) z<^7IL-q+3MRxLSN-LA5nDBaT;w7$`8LJqe}5i9u9>gi$TB(JwYOA>`w;D|T#kRpfU zQW#iE;UECmMnQmbJvy)hG49vD<6IuccvU1p_aVZB5mkpxXIz%S&trFgMTvW0I$x1Puv~e+`yc%dPN*g9*=>ijX4J?645GZ&OXqN8o5G4`~z@6E#uhZnrBJ50WTi1+z=(k>K;wehEdpqAr#2A2O(%2%A+nK zKhQY;@qkpweu6T#2-ENrXe)~jXtoi|e&kR81;-ttY)B~N2^Lr4T)-~Jlr1Roou5;% zYDI>ga~{PB1g<%Gz@VytKL9jY@vbjAP3Q|yXA|uE$wBw;oEYY64 zMlr1eyI~-8ZKE-xrq}oyaSon8k>4R(3GNuNNmgJ=i;XVF?-E~({R=SOaPI(6K(D`h z{b8(M95Xe^+bsKCxe$g|?LZ}4Sw{tyNf(o3%&yzL&OAeZgPv!=%{#LWnZRQcQk=~) z)6T1TCxtVIttwa?{3bEJxkx78y<0c)#fyt&`oUm6Ob0ssN)!^sg=Clp@}+|YxjZtP z9a~4VC&GQD^R4lPn)COY`OU|d`vy`pU^;6F4bJ)nG~j@b!I$nR@Bl08uA&Sf#^h({ z&$q{6FnD57J6`l7S&hd6(nHxf+9R%~tiXJuuOkw|8g| zi%cb0p&+4SJVj!wW^@xF10D!4BZrfvXzc!KW!@aY@WJZ2#1w~vSJ`l}lQTxM^CL2g z(nU7nBm)5{{g~V;Ay8A2tw>~IA?hGYulB+g#@@ahGh>@`1iV{`PVGnASrp?SMFR${#IG3IQl4_*KLQ?TsWc2uIKjOXM*-;HyxVgJlwf@qRCmwP6>u;LK;8UYEhofP@dsXTRP{ zT>AU@M&}SEl(E4LhK#R}J?yu=LD67(u8r#(h~&dj=|M2)kpM}|(JSSqk#zSOe9*el z4cUzjO?Ag6MN1px76w!s-@35UEdKV~z%@(c;n?70bvoCk8b*`jm~EjrmYX9yjDpP! zIX{i*hPwX%mD8B!Y=hluEUKW!8Vm?oedO7R!pn(GsKljOxi~NtL<7MbF`U_e(Wkz!aK8HXhg&~!6Nn~KA#K=rhGI~ShuZ1Q`TvcB zr$&kbqwcF|lYx~(b>Cp=!)S=+x;FGMT2}vd`ujcS5_;`=hq$BN%{DDO{U6nSBTfsbl-4S+ zie&DaS(#3(Lwy%}_z1S49IBq2GMc|UwYeSm{~Y|Ar$)C(?&N0QVV^&d54lp1b0ok7 z5sKHO(iaNsi->|;LCkSUQM3+GER5M_EW`6`=`Z&U3e^@O-Gwtk1dM3gvLO4LUwsrc zV0**Kde07S9!Hx;_)pl>e_|Z}xhri9!YPvk>7dtb>U;*yN^Flb$~)#r&E5PR{&NAz z9)I&2@pcESVG(t(`y6@#A8?a@={tq6sJ=!eG9*WBS0038^Aw_Oh~6tN@IybygTJ}W z{%`-}#V`JY-xF}V&3^E#FTMHk4|7Z!039D=Py*`p%Ba`5-=NAa(HBD=!5--S%uO;j zqUOAb3cve=PeZc`cSc6&TON+j-W&QeCTS&3Jn1Rhn~UHGdyzP;X_7~<)@VXdVf2qX z%d2utVGr;j(=&m>(D$hv9{YRQQ%R~efO5xq`jgx4y02wM@| zBZp*eb-_-Ti7b~QvO_d&OfAE+tJyH&1c;8KkC;XAB*Z+0UNG7#dT`efpFXDj=P~I| zvMSkb^5$%B|y(dQKBK&&teUa zFh|Xomck6?J^e+LN$F~eET1@%;v@X(-^YPgj*;Fi$FoMoB!u#6$VeAMQ*!jo`Hp#Y z>il+tk1Rx4EH8`4gVW^bpe8^8wGxjp`=2h&&_q;3uppVd;Pb3pzuzm7BVq1H?!Jn= zu{h1pfE-zPcyT>*hg%&QNS@fPxD$Bo6fQKA5Q%1(~Tz zpm;CT)SyO2Ow&L9{$iEbA$j@{daW@mkoQN) z90{ZgO!#g1Jv4nnX`N4 zO2fMWns|eu9IY4%G$del4^4jroyQ@WvZ34{HeSc~5ay}EC_c2~d&e+wVn0nbY9y40 zlS@stIt`HwJHS+mU{z2ea@1qt@IO%Ct5Y~T@z)T@jABJtc(C^7Gl2=nN&hzNYK@4B z=5z2NEAE>O zO#Ye_HvObQWPsnH<0eeaP{L-(OM$7khm@tq;jG!u{-f@Wa|`L_Q?Jz$77VG7qz2I-KmAyIQ!0fw?DJzr5h2h640}55sS;gWsIY%@(ju6xE)Bmihs}{2<$jQ%JeNIZgR|NA9`!{pD`U zCKs6_{}plRIg4C+$WR7-cPNo;d;WWnKapUf2k|1T3AoZGog0WQIGCq;boNrAC1~ds z=bXU1HRNG~=oM`o*p&w0(&$V>NU4((*dY10HxjUbQyu=@4coLZwR^$Vd}#liNVrBk zVA`9pIT2xB&PdG0@(?XRB}(J7C@_idFM0YP>j1TB^3CDd;TTy%(IyB(A=F^2f$k&i z3p#g#_Qk(pV=sT#@e3NH+JVCV=iuJ~`2}U85UV`((n4ALW`6g>BqJ#?r#%JH?+^Ke zT(&5RL+r7|5fe;`-7lX#qurI1DoS z_8c7&oA5>4L}n;xO6*625=)9(zm}cmaI0}1HQz!rMq0GXglC&C9< zm-iFLCp8Wm0&Dk}=W0Mnp*{>_{?+IJLo|N0q=!&a;`qp@V-8bKDO|}@ z`4G2y;L9M5os<@s(y}w2l+B@lOxz1>*cX^&$oR$@(>1kVqEV7(?x4(AQJKJeH$c~V z$bIw!_?aSsi#p33EN}OOR6kIRcc4XZ0clunsUOma!VhGRr z;^H7|KuJkephF*cVxnXfU2vTH&&*Uy#;l8 zK|PAGl-WLlsAQH?M)~*1F4b|~?coFyS;TTcl)UDXS7=1Vp;Sl$4O;~!7*6a;BC|Ja zmml)do{4qIP$Z#bS3~OaD88IJ9L+^)kEFmqh2}9|)jw1D&>FBS60)pi&cr^(mm=OS z4FH!)K9NH~cDz#SxFI`bcbt$(Ky^UI%kt&sQ$rGF%(A=|UXopONbo$$Cpl4l!13}V z3Ew+VS&cn~7{^$F#ihiiTpB~Kfn=-hv#%gm<{%ndQ2=8CgB&ZvjS4h5VPP{;LJy*- zum)s;IN{FwhYQKsa+}7(Mngccd?0iW`-PYCBQ-$o=|VJo;Y3KDFD?cl7~`2l)`)go z*n{&sUTU2U-dbIBk4M_*!h+C7t%qRy0Gbi+oR6$bUKlc<5&hI?G(u@9y`})@rRM2~ zqHz;=*sqhwo+Qd!cj@FkLk&=#GC5762^}XUv!4fN=-`%&QK?aK(kLw}@gAHQ=#^wa zx@fH)gL1c$yo2q&l8j2~Xz^LXW2<)poS@>Y&-0%mMFoeb;6T0V(+@?VBASksU6Jz9 zz;QYFw@9t*(Bn_G4=hnjR(8DL3*MZq^-sIH-$`s(LvjXNI||@J@?vNSY-^T1Q4p{V z6rivlY2T(~BRYv~VNsn!M=j%$USS*rb+#M!rLr1PV7_3BiBEM>C{XRzJcX|r2csTx z|B`Zy;Lxwwu$-k3FwMm(js}^+M^7#NI*f`;SIExc) zW~T4QJvM3IaTcGJG|(xi7}%L-QLl?8jLszaHj;KoebjB@20Y|C?Lauda^LUGckpN? zC>!y7EQ&is5FzLFu)k}r0c5~Bnk_OoGya$zshdU)!g(d(@m|e3&@CB->0^|pi2?G* z^Y#5Ti94t*hSw$A35jNt|cmu8*X&Mbl(H2mW3>qj% zHbD*Y+aScNdiiaT+f~GJA-~OXI+kDlFc87(csS{cmz9^}NteaWt4WYoqWD!#Pl%TV zLfA_XIBp@nm4-^pFutWcZrNwg)PFqx!ff3jFEYae9}dSg7>qc{^2yP-1_jX8Cq&~K z3^80|8jb5CM&m+RMr_q=gWfVnoP+H7QETi$iR9tqY%xadKNKC;4JM`BJ<2l|8Tw~X zx0{Z1yBJ_nu1Ce&MNtfou*2>bm2Vfrk(C^LF}cfCx;=Qy)C8+HKSr|cWi(-s9vvF| zVmtT`{}YXQ7=$etggFep-`=eD(wFhX_T?e!3uQUa$9_jWByE4_woM@%xNUF=%{d_$ zxh0{|vJjMPq;Ris-%)xi=ip!euYdd3Jo~qQ9sL_BmsmJF&>%P|lEbzYl5k`JN~MjK z6akl+bEq1m`Q0$bNb$PB5vpF#(Ur_D5U*QWnK{^jjO9BwfJu{pPNX zRE61Fyls72d2%#>-#wyLyCSiM6*DJA8$9Z-u3*%YSkEG4YQV%DEEVmd3aQ1A}qO$ z#&WAh_@(A<*Cu2Er%zf*c3H@u1R^}ieoYLBv$5uVzyE(oEc1p+%-5hQ&mj5~G%D}{ z62GJPvAyHWX(1_DYKE7cvGJA=VEYG&-hZ(^=d{(SI%IKBVUfF-aUdawc*l}i{u!xE z$n~}*tBV1^QL8nPP@r)trq5|+2;`-1w(;ic>zI|B?j;cMN=}PJ@MeF9+YL1a;v_Pa zr4NGyYgn?U%Y(@^N->cVE&E~h3a8;RB=Dq5@`{uOu^nttV+JdEuvD-JV=5#a+^U;f z4N|v^q30A?NkS$cfNIfX5&O;*wd)Fv$c&LK;v{$2_7hQ?7^bv8?kXEqd=ZVl#3z0D zVL1A}BlH-iLJv=HdRle_(vObuKth9}h{1TlZFk$%yYq+3FGzhkBEJw4=%~Qv5#249 zHcZn;0u`tVN5&soO7%LeRqEv7Db(q_e91~3vP!@^hIn8FHqQ7c$cQ+G3rsSGcmXig z%wM8{Rl>odfdaOd=!g>g1Ry;7hHfaR)~@bW(ow8P%&`u0z_tyKlA@Er;XDlz8}Ly| zl+6_)I}09%F->@s5@p+jM<`LxOmSEu*36Vj=cE!?kjXkDR6|W48`nZZ(tJv z&V=0JEU&Ahs$q9TwUkrvqR$)~UZ+oDcewW!rD6|a8NG79*B7bWm#DeVTZ_y%3&Brp z+tmsgM)E7P-@=rU_(jR#N%63iw7^(j=*x*_AQIG~;qaA943sV24D5XI{z2WpRwm3- zKAvR362zJCS1jn4jmKVM61T~7!7K|b-g1x|gzV!8vP*cDo&W;-$XsI2C^HJwk7|#A zdC$a;(Y@Ee{7a%1#9r+}90@e@X%O_$ilD$8f&+1iSXD2OX-6ak{0L(^oQ4zWBa#GO zvT@!j5fHOK!c9PpC{LnLzC9v3Y;!u_`RQp<)=7dnps#P(G!hbVkPQ1_ z_Kg(Tp@yO!9HtGTXh&zp+Az=rE}E7@M32N;=ZIoP`@yyyh8EaLgwCZIE`j=2q?i85 zq|#lR0iAVB3IqL?E&-|QKxU&h9|kS^g}ZHb^E(vY{Kq8&_m5-)=3gxiVd&7~4~m@5 zz<31Qr$VC^QG=Xk4mhg4hu|N6P*#EIRWZzuE4~0eS`LV>&TwGk$mMn7E6i2c-zea-*Vt@kiKI; z&4P>;@Bo^%#^liwXMyAfA49A{a)gCFPA0Gr8=BHhFRW97-$j5t1k4(Dv{k?H_@116 zfG;p6(?g1HEgu+aH18n?rdg;?h2vzk;Cy$px?7_jV|Nh;p+gJG8eVC6Z*u17DcN)+d7jzXOl6&$OMvC`5cYkZ*$zbk4 zu+C{kiqTc6QFMP$0+H$3G76mD++01d;{n0+brWmeIYin3Jz2YIPBwtX9rlV6M+~b$ zzkq#+*=MlB|9-W-2Q!Et{^4(b_tWox`g!JI(OXVDLBXn$#1q5Hn0g_k+JAl*HD&CW z$uLuh}}>^aEw zYAB{Fa$do}AZ}EkQ_2fZ9_{WvgJS*zLRCRy%gy~4fEh4ncMtNp;c!HsV|x8h;KloL zz-+;2G!8OaG#&Au?aqtCeu*`)OwXF9#}iE$%d@>f|9BkzxB?dyYOga%ezA(|8qmiL zDGQ9PJ9t&(qm*48`~MvLJ3@99KB4TYe8f&OGObQ-#Vkz&iU&3zTsNroz^Pj%Q%5{1 zH=(ol(aaw4r!;vU%JBwBr!VM3h->A?Y_H&0f(FG;DowR96eQMEUXnZ|J=`g(hv>P? zuSL`iK-eO7C%!X07pZl}qyIR+}5>&AhYlW3?R^Ah^>xF{lqXxY)>1f4V2ab!o|yHN#kWJnmi zLe+7@t~`qt^&0m;k_WFLaUnJ-w)ZBZXoyAvc>QkMO%vIP`EaQ`lqQY>Mij+5MrVj4ZRe=kh9z#yhFm1xo|!nC^QMtmT197)&O{vWMaX3roI{U#Gkc~l!@xVi zl`>>5;lU)CQ$k1;Rt0FvU~<2FkD-gHJ2t!aiY&}KYtlwBBq7scc1AHI_msA9!YQ+U z2K#u=tu@A-4rX!yyzt-5MYQoiF;S+?O$g18HNTzyK9Y$H(om{*{NSB_hy>{hU}$+( zA;TPJA8`;}JI4U1caqs}(SB+d_XrV<7IM#Gnx82a|HG~4J=-7%5`97Fz$!`+0>9Uk z6+*x#6O2{zk>?%VKR)j4koH!bpp?nDqT|FB|5px27sWO zx#oiq*GQD4M9v^0^y07*A;9HIA{YA|LYB`6LqA#(uS8tj49h0lpa*9gFv$T1BmN?|3{k6^j zac|vlQk77jU!UmTkU1!lbqKiB66JT(`)PI1C2+cH<;fjN^u6Z@pU8;>h@RHLi0fEZotXO0k zEln}<>>qYCO;Q2_@C=VrQ>2eUS%aRu%8eY*y$=s!4F=Y#Q(Ce5Xibtb^xFpuvsWF@ z`*A9bDE)s9{vF321AW05hQqUvUsbH6zeCL<8{d$b4LM~seQHCcWh9CO(1A%VLL_Vw zJ2fH1&~>D_Qxei5_V4LUN`i-q7&Z7N+wfk5h5I-)2So1OdSajlF2Md(WnETF|ihQDAt$f!4zq zq(tdq@}V6Qm|8?|q>b_88D}d~3*VfLR2N!DVs(tU4LvBwU5VuxrRNe1*<<|8VUnI6OvP= zr^txHv~U=;-kDj=ht)7ecYTed2rp`5Y#eLXy~fTU#i~Y}vi^@C;MDl>Lni;m#NCh& zA2^C5Q?VnfyWsJH9OI}+GytkC^(RtT_}}p{ri@m866S0}plK| zOFRWAcaOb%3WAQw=^`o=ov;16fkgu{Q~yZrAbhSY1&qD&FdH71a7Yo$*8t}_59u}o zGy(wdCDwftmX7((5J1`wV8_!1!juVwcC@bXgWgGbuucK`NV-TYZ3%rJD4K{fXJYok zV0W+tL$|BwEt350+EsfaMskw4DE2>6!$Q#1=aegT9b$P&_6;H@*w(Jo_Xms7?mC0D z`5Ezx?tRMJvh7Ub=pf0Bw?@;$;hb1wk2XiCb?(zqa>u9@V-28Fes4g7F=}0I=ew(B zlkh7X&_u+Q62G7&a?uUQ+di-?s79ZlD&~6r<=;2W&A(H-L(ru%|8cl2X$3K zl$t;Z6yK6~o@@Zl|7o}c^t4xq9_j1p+n2t7Kp6Yer$J+i0cvOtE8_jD{$KyL|JQdO zVzYp)#|9M}0n(G2JZ^_<>VDQ}2w8COM8NW(&r!?-e?ev+NSFcJkc0_+&1l7-QtYjBFBHc~}_}1$>A5P9A_j3pqxcOo;3rca&EbRbl)#S&~Ot zlr;xK>nZtlbKNLvp*;`{ojr6h_@8A)hfzK-qzWQy1?CAj<~S`cm85 zX|SbL>{=glp_B?5<1)muwvdjc{SUIlNYG%`IZBL^G|e)0G9M8hV5=YrDzLwZ>@c^O z-7*lVjoOCYkxzu-*8~1M6a0xxJozuE2Iin&?jUmBzC;7`5>x$XSO7D&7~h(Jq{i@g z=^|YUOeq+v217W^OwK!fkGaXka?WOB7{=KcCTGV}ICBnF2y2MWVEZN_9`9tX%#1_{ z{NCA2s48sfk6`o&*2rD9??j7aqA+{Nag=?@NXH5Mn0YoGO-BQ-b0k^ zhXctp9HxOldp=5tNj9$Q*f1>U-p@pTN%4Q$)f$!J&H0X8nOYM1M``Cknngqy>!z)) zvT;$n9`6sy?cqW)KVaeq(I~EHjs}iBfo`@pFXMYlAcQuNs7+Fx(Nw6)PlE&F3m40f zPhE1&aUAcE3<3kdZZ_y$h!($lj;;_5`GE0(j4`D9;507A$^Zu<>_Q_JkJJm~aX@Hg zl=Lp7GKowoF-e>6;JbHW;140AXRro%!CRL=7#bGyqP|=cM`eN~fbYONP2@9QAYK4~ z5UHWyoM>EwyB+D#DW@jnF9WXsY&>G?6pK-GD4cMByp>#EoMd92ObBR3v9===c$?Mt z?mR;Xmn7pQedcnv1(|RLkH7LLenqTTS%Ma1NOfg9f0b^q^AO#Yls4khK><`%L2OiDpJl8L5=DSsTIxorB;xpz#X#6JNZoGhW z9BAbE?j63jgABKnKX@Ia%6Np)-sAonET_q2CqvE|%;g}14*&)<047Z%zJj^nt-&f_ z!5sxu=4_3B#JNM<0=>Pbz`Hv(J_W}M3A4OI`U9+h&8ZbDi47X-P+kW=GJFZlXM$Zr zO9xUS)j(#R2MJfa7D$BWz6oG##gwze0FNNc_v{!Up)~M~1;>MiADa85_3Co_T&DuiA^q6f6s;^<-MpbTF2n|6Uoc04VGU$IGcuG;jb}U6UZF~6Y$s^2xYs%NyAu0dyuQ%Z2J~XhxHQ9DlfgCh5=h;ti!^u z>>@wJ`vD*JM!xDF&fi1JdtkRVo!mc*M=ac45rgq3 z8VLKGXoPWWORYRO2stA_;C3Aa@y;m^1jQKVsA|?8FHK5g;V*(LCmyx8J49n&Xytfd zehWkPaQq4_9i=C!M3;s9Pj`yV#s%v*3J$#H3})_qHzq!ywk9O3inFeQPQ67!h3>8- z%XwtG!M<;ridtliy>K0)Euo$v>MeOckpN;Pv{@@$FI$Jems1ysD9~KLQxxlo_)8=O z8x*5a$9E+Sj4>FN6918w<%bva`|rMr>}j#B(Ik>sg=Z+%#DL^01GIPr`M^(*@^(r< zm0iz9Bd0MKvVXf+z+>RqvzK5%hgh^xCP5(RG1yAGy>$|AZ|*K}e9*pZFIRxd1-ldl zE5^@5571$;rEG2D`uQ375R7{z><4Z=GQ78~LxAfEz4wn%!}-4foTi2w1Tdnd8ydef zZuiwi$!_NO(GKN=bZnTy4}9`8`6uQpVaHTv^hkt7InAt+fTx~UA3`t^-o1)P1#xt+ z#W1RO=K~5qE`gjsa_9^?1M-zk^Jy+dYQAud>D1SWq)2;_OQKP<{K>}|ZB^|^;kg^E zw*AO2&1)ayj_01qffrod@jN6n1q-7hr{nk9Lx#P>ezB;T#1!&r)cUgWHfc*90qE!9!HSd(5F%=rK^DnQAboT( zRq?-{+h(zS$GyIk<{-9h#~=n}7>p()YXZJb_CgrDg=@Nw@?{xr+Mx5`6pu1`5(d!| zA}=Qo#|T10H>JC3l{L;Qo%s%^26?%^<3F-6;q8ZIVcs!4!VP2Wf}~rG$STi!*SfO#LHT#dkSS;pnL=R z7!s%d?kUk9@@Aq~bcEid5&=xHXgVvfGPFNoGI1_d8stbCRHcS8o{oZ46Cf7qrl^?T z&73xhgG*18z)$vyzd5?h;n6&sRGc~k$~oPJ zp85Kuh!X*;1F*S1e`8h+5k<7_Vg%!2 zy)xlyRu=p_jdn1_nuX*$#>@tg(G=yr3UfVv+srYP z{CZ9<8O*`4*p!|(RL%FS47ho~aJ<0Et!=m!GRyZ<2g8_$P+}Bx=x!@2lO=taG)WV% zLqLb78;l=C65=`Nj8GjF6bUjbngNiKDkEng`E3@^QkR2E80C<8QY3#%7o>v7T;?bi z0S~4^f}pH91WphLF4AUOC+XUD@7uebpb_6gk`KZuilDjyXQ=>1;YIz-@t`-wFVtxR=`C5CAm~g zbF-cxgng!<$Dij96q8-lJPV24j2Z>-s;jaP{HckeMt7spVYM7#Ku3Ft?GB*=RIhSD zfzz<52NnTh0umdR+8Iscdl=LE%|rYFNuhDE>Gb=Fx?BG2$xm>Bs21;n|Z_FN%5F}sd6${hj9T)QA-jPkwtp|)cDPjdfhW5|_ajU(4nX9XDIZ#WID;OG%Pwig@Xw%?p z!n>OPB$r!eT-3*p8Y!#O&5y3XlcPdfOUQj+rIHy(%$m4HpNB_iRZuX4OkG(H*{SL~ z%YcO!kZMAb>0_Xm7ECV*k>-DicyfVyLh^{Bg<=p^Kd=8^ek7z!#%G2*=5S=08;0iR=={eqX|w67RpI zC3|{lC)<~`c4FmZNhs!ROn)qxI^s+W98?4iL^0*Jq$~v=;1FYGAChOcV6Aa~-H083 z6!x=2jb&~)M&ahB1#3PGj3?R-Byx6*W(E**RMs+&U;%ZF42o_N}Tsp=Lo6yG#glHo^}U=?Q>k|0^6U(IBLdL|Ua&Byh|n8+x&)w)rRTy{89 zg~qMt6Q!zdR2bvhNK&HEQ@1uT8G@%ctCSm7<&q-T1mqzb1IQFJ97l?V2hs>d93)ZC z>xcIgS5_?G#^mTS;@Dq3>W1&ogFmE$Ldbc&m0(*T#?I#cnkf;ApNd{!=QiyH(l+2q z9b)B>`80Ea2KqNqwxALj);0GL(Xe9yGPEXxZ*C z=b#OmBr|;>ZU&gw=ziZy(rIAU(?u1Vr}FHejkxAxZNamC)pZ~ror&Ech{c7QZpi8* ziPkto(L>jcX_82;2H<7j$zKp*ot9j+xtsq`d`~F?COvsnO{|!A@UV(dNm%4-l8E}` zDD=zuOLT9A+>=Or(iEzMZ$h8`PGbg{cR9Z)Q*Ib4hCY-~ZoVprXDK=iJk?1(3iKIi zq>NxVw1LAS(yp6o)fs0SGkOk7+8_Ae1SWk-)&Yo7Y7-U2qeb)X9a`DLHO(;(j?=0p zlH{o4Fg}%>LY7ox-p$!$?eHl5=_46mJ#SWp$CAa-)z?t21M=`nwVvVYfxI5%gNVHL z%oaLy!Dh`^l9|04f@BjY=Zs~n1M=4jOA8MZRSiUVaS25v!htj|INsl|8d~er&*%hmUws&O73!YdMNfATjQ66NPkq<2VZO&DoJD{MdcA~?K;8%$=DdnTS_&vsG zz+4|zdq{OQIH?TfUvpIdj+ifLzFb)x<00K1-5+E1VmaZj{3F@n+s5*L(-7|(ywX_q z4>?5PgJ?quGav{S;evJfL<7++szs68K9u5xW0kHWrT1iz(|fE2g?N#MH}ULimhdKM zE}hb?X!OH$>sskn&%%+xlSZ?M=u+pm5|B!Xx6)NYXnyEMPz2r?pieMDLTWm%Fy1c= zbT%hxDAJ*pVX-m4*&L(Kya|+$?GEECq2n}I>dq6R6-orl28ja^QC9kO*E#?KPonx*Y%5_Y*<48E;0KL# zOmKNiV9qHGD;?ZO_ItP`RIOte)1#IxX++}-lb3gSFqeW&@x}VxY>*hXxc6zCZO*p( zm|=^Ha;>f+yB6Qr``00)=U|%>iCYu}z1h>Q6`RdFjV^6O zm1KIdq6G)5>Aj|S1YcG-4PTT~dBAW0nuG%)904+ZA|VC+z!#e{_H!Id+F3gqv_!caUygz`$_J_+r1eY3 z&ky4Il{_!?7&Tk40|I&SK1GKSWa;v8K_UMJIUD%@eFJF7!{g04i^WFFY<}>p&3lY7 zjNov>Y(3j|jtYwWxxw_E!ogjv`y<>ES98C=_q^u=3aj|5C6(a{irYc|%Vz_Oun zS&B4d7z(aNqYTq_qHDrHoF-b18_2=XcFlalA3+GVqX4TLOlSmhBJ8ptFpwiS9T_>L z6m~1LY0p_+ARK$spuNm|@nO!wA#?>n&;CR;wr9MBcI=3=!qgfgQRJy0f^RFx74Xmv ziYm4@iH&%O-?pYmz2|GTWWL03r2E_Ut_k{6*#Ra+1+bq9|CRgy9Q>Qbek2>a3)qS~ zqjOdf<%0fY*EnA=$G7)8GP|abSlEe%qZ5nNJU~a&)u~wl7ZFPwEH_HG2d_I zHy>Y$Lj$PqQ69Kk?1K#e$A?4pB)sg8I4s-?l0cy?`YjkWORU7r`E_%*#t@x!;{~|d zlx#PF2oQJqfHR`rnLWPj?5;s>c9#I&w&UVr3GvTp9{s@|XXxAu^1<_GN|}Z=2?0s& zVUK*X5VvT6B$~zL76VO+=pTrpcRIuuCMX_xWy>9~9a7(*HS#bH(Sr)#Z7f`wcZ~nH z=Hy8vdZFhL_fhz9b2P|n;JUqoz9-Q=SzYT1zfHrr7yK-|CQk#u3_!-c5e+o=(qrKb zdb+~GLBw?NVE71Ye(k6J`u^H@HK4=?pKWjKEeiki0kzXRv;n$xcw+^%YA!-Tf#wr0 zcJeY(+dvlBcdxgLMjl*JFBfg-Cq9Gz74{{JU`3CxPu_ICN3(sBMgcUZ2pXW%=JaQ* zIdW{sZ{yyxgPoD}>atV&K-T`Q{m4Au>)T-Y>Zcu*Rc~RXJ2ya&1E)}pF4%97d3CU= zLFpKza6y{YgcC%MJtL+@&vl~M-vb)=kargw!>nW_2o^!#$8?1MxV;vSdKnl76OfQ! zNcCSsqL)=8O_I5)TT;jt&Ktyvy*>zj$$=02+K0nLFCrMgt*hSU?$LUs>(KJOK5 zgqBRQSTt?yEhyeR0FhICv*j{TF!MCOL zZRtz_oltflMk=fM;?c?mGvBDLFhhYOdc0Ylx3F4V^0mWcdh9E?RS_EkSIo!|!8yv= z5EFqxTHc{FKSYG)vO{5#2X7IG_LVx(3KUBXlYRo zYDsxKKuEn>y&?+IqcPCtqqR22BwpvTUTTaAVoRb63T#AN7 zhswGpJz8jG7=@q3K-WYo{lrFTnAUQcm~9E{xNfe&7{ZDIDNDTSq%)B-(^#0$K~RH) z7I`<5vYqv37>|HacQ>=wD+(q7ntW{J;;_BN*s=ETLPPbg`m%y&BKj5qgd+h2+IQ|r zh}f81oY6=)eVHIUbW+jj-lHcLvwwYAlP<|Wn8Lbis6my!;BJ;!wkw%XlEBLsDLn@V zZ=pO`8Wz}(aK6#hRFMwSIPWegBFh;`8pI*3J(u9(yqtlBVOI=sD5HK*$|e!@LktrR zFwW`THYS42`#tKx*JK9u|8MQznjAT@YeD!ueg&*vCUGn?CvLcrM*ECSN@|{ST9Szr zdo)T}8wv;jk(CK#1IQ}oY0ZBh-)rsdzCC~}9*upnR30SO0$~3lTxpPXC zirx2YGqb}hj*@Ddn@>GnH6jO}m)qAcjHeS#-aeej%g={?O?Ro7Dpp466*m5^j;>MK zbVtqu3~awPU7dFp+BRyDnQie7B_!NazU@K(qbU&IzxUK|QaCcyCNl1<4Q2jCZ-5jG z^4p28=J2Yh5rLSFw?5cu4finDX2{l_#PhH8&F3P5a% zrbf~`wM12|vGysl(2Kr^-QY<*qyO7?kG!l zsx8%b{DWxf;qf*+)ICf9YQ!9Nl=zx}M|-JDimX{r{n~SlSlLG~$fbfeM?SC8T8y&> zozNInug;G7yGg1B=k^*6vU{RgC`t?qK&8caf?LX zxHu7xTWGt~6wx+aGFEBd!9rKzXt~|Scoy|txRX2_dfoA}F1CeV0EomN&)^Tkj_4WZ zQ!_$u7GO0;Q%H0f#ldKYSZqhDMJ}|DsZ_Z~pq&A*-IYFj%%#2;txr$nL*4W8mR_w} z;jjCr?sOW*f{eBV!utyjY>QgLbadI0XeR>LihQ&fAAk4TN=nqQCqSpX^VWfR@X1B< z;be83T3TuzwJC;B7i0CW`_)q?eJ)@zjIXj;pd&S)=CPuszx0%#j2HP0PJ+3gYH}$V z6*v=Og+R}zZavlbmHZ@Z9!DPJD=flp*>CrWeN}Z4dYqRnGv2B};Wlhtu3?z|f8%Mz zgzitOSYb6myJ1aGFL>T~0Q5s*n!>wUaIuE9P20S*shu=Xr38bzyyX}>+9t4av?Hy| z=wjSAaoiJ#AQ3CM^z!DSM5)Uh;sPGA@E%U0M#MF0Cjk&pK*?$(>QXQSrI7e5^}^4u zTKJ9K5wT&sg0MtNKpV>gaL}bY*rCZ4`NiD5 zWUJZPw@JSpI%1=b7u>x|TCg_#j&#=uI#pU`#zmb@z2`BH?V}k66W7ZcMvt^%#GH(S z6JV(%*uS({j_lnn=kUAsDA!rqtW0NKO6cP(&HPY%K7jhLPwdA1VVwByQs?9mU!%*L zo84~?7l6Bs#w4*%TWFQUz0U{oFDKmQJMRa976#P-ML@d0^lS<1)cGU{M47gCfvw<+ z?ecXiDV8juZRy!kv=~qmVkWP3KIXIAIkH=#p18|N=fZ!dazfre-y=pKfUyNjM!$bK z_D}ftVtyOPU(-_>t~|c3B6XgNeJ%+lj!c+ZX=G{-SD4-@{zkw3EDPp4VHcK;f?aq5 z&)4RZ#8mK|BY_8g<}BumEbR@n>`|x*goJ9uYWG*ZIMAJgweYn<$Z^oZ^s)xrF5$^g9uoR$OM**-Bqw_b} z7*bC;>Qc^6PG&>!FbkuX4?0;W=GDG+oy4y5SIO_lEnkmL4MwvL_M_O%VGFbnRkwy5 zA8S9>ZvGGB?{8gczr50dm1B7)&)Ivl5V(|h4ZEkx2LJylCDSOe+-=b3v%OPV1s^h$ z?6AhoL1I-6X5|<1L;8ozFYq|q1kwoegGAX1zhPvMF`=6{31zjAM~Dc7&XlCwQz&HC-M~-C zxODV&FmzV<%l|9AV<;>3&TzVUCw)U@U?ZAZVDlPcN?>J;-S}O(yqA-%s_zAUT(|df zFf`NqTO2a?T6r2Z*Fq3xR@Z^y4c+Zn(oh$c>_+DRkl76d$Evd5tsM5IQ(9 zyvGL*tt-(IpGomM_43pp73c%rS9*up{_7}9B8y{)>@;(4wQT06OoZeDo3!N}Kfll+M3%x|w5&CJ-U zF_#YyHfm>f)sfk{@5JRN^hvgb2WoR`T%*R zWq=TQ2p(^-)k@x5m!MIaL>S<33~Vi$u{XAmGW!#(`C^*|)9w))fWP?DS&Py}+Epfs z!YZ(r#^jStWH(_t4WV|wx%aLn@XSnLv*kyOm~cnhlQaMMnWy#ct}^f^(n8E@UC&_c z6~wZ0ku;gJU^Zv*WdjK~?quf6vf4D;$=X4rDZ{Y>Y2tyWOVqBZMyJq8$W;dGActBz zY1NPZB{c98@*b-&0k@r&3Zg#Y4627GM$FQB1>G7$p%3G46g#Xcukz4`hqK-E!+2v_ zTtI_dkEhlxhiixDPcEJ>)xFSs+96b|(rd#dMRfm{!!?gxGu^&W@Wk6V*u1aFD?sU> z`)#k0%_y(hlic|uwjCe&?SL{0?VZ8+;@kg}*!39V$D2v%Ld%*5Z(6oNEkp$#WIIe! zYRJuM4K7Lpb9u{q609xe3bh{;*-a{^?=;cG7+;(d&JwwZU<%$o9mI)9OkYSH#vf$H z(oQUw)hD^Wld`c>qWH>xlq}gJFQlojw!5=twdgec(rnsVvinhDUh6ZXpK?vIO53mg z<^N9ETaoJChF)bo^eUV1ze4`3;Z@3`qF4D}D9a-ncIc1oqCYltK8wrpJN^x~6~oMG zASmC~-~2Jo^V=vm7%QbFvnj6hoSU&&ibWEUd$OJlNj2v)K4(~E#kzz>;S?@c=h&yZ zgn-#)#G6Rf(XwGJL#c{$S5ETDC}XoaG1cSBZ{s|YXrtyaQGyk}kjDNu`JZi?5<8!^ zin#wt^rj)XzTqCbQ=qqjF*}D;Ar`h^F2F!4hS6D*3JP^)8f8tufXQafbMl35IM<5R75nhphUlP-UA>rfjpu+v}e@ zC$*_4kkVBC%uEP-G7vLabFE`KorH)70E#x|*9yf7^)TXmbJsyzBG&vdpki|Fx^ z|3xC;!r?<)C86BvX$kT#stkV2cm9}xXdAA#xOa;5?N}t6VzgsCF)LlZAZ}{8{}3$~ zz<7iwf5lX10){+Ui?)#GX(L`p@o-K>PAm3FR|Z%AM98=%W8<6Ch;Y(knfmv9Qbt?| zBYd_^3FD`n<)?_LSvoT^-1}S78)yQ}J(uV|Mi4maIO}b6cFE?v%O0uc{6O69X?f*U z&@iit{8iYHBl0A68QDPjf~KhnACf7iLXF~`Br$Ps{wX9Umr^OBV#~nFjnISCNVQ|P zOg!IM{uJ2{q<+IPVS&RE)>-stB~G8(6fQA8h$M1{hAg*!TVOvjiJ^fmP0A3jLUFN%`7gL$}AnD~<|RVTwUbfUjLa zu;NCNmsY>E(+~=ZbWb|?Fy%*e31z(m(p)Ko{k}Nih zl=x*CNo~$x=7^<=-`vRpa3=w{M^Ps_0qPZtT{Au=3?4CeOy$uRf&OEMP`_knC(gUL z%B`L`g&sQmxd~%&Mdy3_@@)GFc9AmJmXMQAWdIZ3>JJ9HR;aLLeqCfuU9m)pzq@vt zAW~rQ~uJAnS9d0;>*;{3XwPmL$b`WzKO3|@G;<2=#nb~! z{26%3AwqE0r2=bif%TjWfKNBSXceEu8)#7ioOsl2Myus^sao!O^_TXc35*(Kb2N`{=%YgzZgj=C$DAMM_ONOUoA6OLc*IOdQ-DdG;O`SBZ5w2 zd$U;_kLl2Cu(P>{H@ZZV>jO?KZvNnwI1CzrS_=DwNP>l(+Md0B7=OqpDgca$5NA*t zyj!ri77Twb+ymCy9cji+R(DyXS-i~YvP>H#7G3TxPIkJ0`N6 zomHxaZ6NcFq$y5f%NI6XEf-T=>>Y z2ce`;i)l{hZn-8#3K^F9$U?WVs5)*Haa~*aGAfG-+d{@#%1Kzy5o_;t6*DpP0-713 znJP`FLBljIz9MO*CEb>uL9=(UbWS^NRZx&o2U7-m9&A)x>Z2hXXjrHxXN42YD7FPl1-M}(5Y0i4p(esPGiyak zkYS1UuqI+Sz*>%dCFvn%!km(HsCCCP+1uT9{EYtaJdPFIyFogw8K%<|W&Tm-%(nj= z4YaY7B}I{J;p8bh%M@ie4kmtIELl{nnf z&;GfKmrv3t4AP5oBdKWyQwW3vK1mRgOcFmFKq7IJ66-ZtB%ISAq!{9+$>#7qV{XdX zpg-7I#Tkr6?m6LDQu@~`$M#*rloC<8NKH@E|o08v17 z;@!Dj<572<11{Y<>cjz^qwH1inBcLPc;ru-kTvd`=jYHX6mPAT++{8*pbPeXb$B8? zuBEcl%L!FuQV|}ec4sPkJ10{NJ=P-I@*Yf$5DScyM6ujcsaQkvZWU3kaL4TpB-^$s z|7S&rTC+2$;Lr^OIwi%|$>L*rQ0=E9yZxl?-(c{wG%_ZEecu%~xelb25gM5)NET!= z?fMtQUlDDmhWswF4P0u|5z22<=~1&M^5VFU*ai(p6Xc>&@3u!~biBQKH%zEqusvWQ zJ3^wI<5uV@;lD*}|1}*^n;?)Aw7Nyz&91JIjd6Zjq&9{~1XVh|(3Es6K|(qJ4~{dH zYz^BeVd~5Eh0;8Ke9-GC(j!r&U?XFvGAUtl!GLzR^)$A!xKZs zpcP8KawEy{u&{0{x-rY~b}(wa?JU`-J?0F^%J6kbBI1)SH+3`2xCr#8kUyIt_ep3? z)IA&uC~j^LT9@HCy#(u%ZotRHLSxa7lNOr9Ei z)5>0MF<2AgId;!O&b1*ES*du)ycDrWz(-aVby`o~fr`7ApJu@bkubCZl#vcF`jTQaslSHDO zAr(y`1Ki!z736ola`QV}Rq_wD5H5ETLSEpD5|+vd*U@QIZfX||iL^aD8!st))Kzvf z?JMmWZppjC-cp(~M7^42)G0cR_IL-!DOPo7h2}Heb+OX+-4`+>xOn(e7Jxxo538>kX4hiY&=b)@tD;8P!dpXMbq#)7!xZD{x?d-EyZ+Qb7gr|pnaWw;B~#&8e~YHgyX4*MkOl%is{ zLVZDqv}VO_nbFaj4FRim=hgjqX2wXXAQ$YFhc6jPlusNzA_M%8CDA^FtFr16W+8im z$9NkhlG+TZWxyHr2XOWfgiU8}?e%?}ET~fP3C)aF4VoHFEGXPZUyo$5e2*ugeTSCl z_b2*tTYP`wg|>G@?qSgE+5KvLZ>g% zejt<=%L6r;m$C}6|5@29?;ncO1_95W3WH>PnHBzNb}ZSU^gf=^F^`g zRTA~ANPc-(ssJ8oBf+sCL9309lP|oEHD8L%N!m_6&m)1Ef^bn4qtA+%jR%5@2awV% zO)Q!83@_RGGUTx*an`UG8AFPknQB12iYvfJt%J#i^e{YTHBsQ;w8K>EeHXe;9f011b@WhqSOffbCK~xvl ztmcr23irGo*wlz`Rz^%i=7RlX?BM3v>?2g+Gsz!Iqa_DrSOr`4bD9E(rAYs`KB97M zS}!fW2e)!(rkNc#b*X;A&dT;=L)sKWLIl+-s?4}cgebbVg;#wK8gUn`_oN{!Ld!Req)r3|-T%uVS>%VeKkQWKOzX&fssGgR&Ucrw@1I-*tzzt@wy zR*P4xUi%sWWTQlleQJR%XJC?@1_okzje!V@e9on-%g=0l zL$A?7H#w6DFCY<_+x071I$Mh%q;<56Jn`dEOzvB$dUBRBU(Nz6O^E$zu z{j?6Rz43>m@RZ{{8}UGoeIzocNVpe^`$4Tk%5&y34Hr1WS2awFLE=(swWkag^e0y` z@Fb*u8O@$-;5Y=ZO~?oirB321*F#i!QxtL08cogkN=1iqOaF0+JqFg$Y1t!&r#2)v zY}t7QgcfNQI>{q#O#Xm|Es^$MOQbUViRmKIKV7lHClaGz3$?R|AagGgOZQHZYK*mC z7-0X21sYe3NoAUuFe$;`d413QPdSYVvgjBIP=i?;b4GMO(V@T~nPlgB#LN!1`b}2I zB;j;VbQq%ug%w_{?9;<3$+LMY@?GmOA`dm6OhS~O3SvAJw}@oW$JNl9Ohw>*)*A-VFysolY<_?(Rnkf$xE$c z?p)Ktwh~>A8(OQ^2nj}6d?lV4tYpErvX;!@$PU8Da)LFdF}QJL-82=u$%xYL7oK1~ zIVXjpq4Fao9%{~F>r?Em$YOV&_J=Q#rLVuq*sLNDP>5Cvvn4R`RRrq;wkS`q!d$*1E5?i+7^qDe&kGgp2uePg zNfh;C&xqrZeo+DsA766|m(aOq0S%oG$2TQc;i`GOXhUA3^FF3E@pEPyY>{DykXIoP zEKvxOLPdgH|CD%p1{scvG@PZxd+y?cfR@G6HK(LAVZ4^GnwSbN!D_4Z$sTeNv*j#E z^I!u{b0sdf) zRcNVGV?Ib&(EyXXtIFYC7+}f}ILD}x%@p_kbuc`AF#tUca6e9&$EmrL#iWEHE%2Jy zEyxwD(2Xm(EeRquMI`xuJtB>&aDBgNMAD`R_f0I{N*7$%o60AJq@kOJ?A6-u$%3Cr zw~H_FZLYEzafR0#&}Rpz>T`zL83#A?Sy~<$wS7$5Djm-;=On{XWb{Vb7>qF*^l z)2&LHj+GXDlOU4XwI=MBxWUCMmDUWHmSLhBhi3tmmq)8)aara~6wKI4Ex5#852$Hs zW^M5n8N6lL8aub#c16g2@5+5-l6F@lT9MC0TqYO_Zm zB?-eE*hVFjzR)nxQEN7%ChFBu&l z0#}mLt|Z9d*wBE;62>dg8b7iY!I=9ml}W$VkeNLFbcisKC`~ooRr;%$Shw+J;vMZU zfjtTTE9B31n84N@LV5MMNr-ci0I7!ObsG!bXy`n|MKuj>#RW0l?3@2MSTDIw1#%0L zHc=_k@53(zs!QnUlGQj%gqWc1_73vh)!FgrJ?uL}U$CAJ2R1a&Plu;h*=W82hmzG6 zOZrv{!3kGWHc%5=9ho==Oo=n0&cc5PF# z6!E#LytBA_dQj3G8c~-B<+vi|d|kkOf|y8nHzHm2T{Zu9JW>5W$z$)YC6nZ5(|P>a zVgFq-Wr66tBydS?lob8q*y)0DNzCk~qXi%;yrFU+Rb_jP1*w{*ircWLPUi>f!mYL` zGUe1jzK2VcDJP1_S8ZwEc-hBJfOd2hWXl#5rPM6G9wQgf1fGlhF4iyT!-e4pfaR49 z%{OD#$|rUn`>j^juDk6IR$e=sCcdMMbtv{sdh`gctHQsT$5czuDMyHOdk(y`wQC?` zHvTJCWAWFU@U2k27{5$e`CoB8qaMUAs2XuE?1FGq{l(6u@2dN7r*iWx%cxC2(!Km+ zS})sz+hJ@vZeC9+_jLDC1rM zQ4YVz^h4`u<@H8!slJSGldsS3t8b>BH6lp*JMwa)dOmRqqM5v$%yJY~7tjw~%nTuF zlsq^wZAstFCz4LOn?W!0ZYn~eH~06Eh^KNgqVYG2^8WTNv}IlzvCf_R^=?*zS_l8V zn}VgMkednBk6aLbXN)~ft6Ra{d~pxW&b#^io{gB#g!!N~PkpG0= z$}{9Y;n(_GxmWmZQGQFzdaJ}fJ{O`c=7ze513KSyl61FGfqA@pFIfm{Hc@9AVQo$Z zg6An2$Yr3(-6&Rw35>d#&5Ze8>iV((D8W9CU=lI8pMU;E`1pRJ$G0Q{P2ob*^FhV# z3VJQ|q-;lcn&w}Ek?_UWRPnN=4CV|jQ#DZRWvPTknmk`jd`(d-#gNbm1?RRme^PFb zr6!S`PdR3-b+Jh&_cy=~C!zp){IMLgyY%^GXizbA=e<{X;_mgiU!(fliw6YfutMqU z@??m?$PVkSYYVefC zgi?Ig$ zL>jTOX&5cQBp07yX2bIB#4KXi7$&R}?{>wEcgqfB~ED1N;2ho>LI(9-8b zf`U2O%-xZ0hHCuF*i&lSd%LjQ2o4v^4TI>F+|8|@O0R|g_$Po(K7MF))CrbJduR_+2y${CNFq-Iy((_;_(Ky&2T&h zmYC6K4wT&aA2areZhKdxy+|tUjz#lp((iaM@nVDK`sXJ#cg1Q&Ys?ri)3&x)Ov^=C zMcr~@iw1LumxUIDbNF7`T^hgPVlgWl_wy>>cjxwb-O90%5AAm)xfsh#Lf*8}7Yhoi zFQ)q6S|#V_l%R*j2Gw+$({pFB#e2*}a9#WW%Fr*`TYNY!}(?yUZ}S5O3!5rzWsM}K2yYpk!(G_{?rP4eFvq}ROYE@Oe& zSaw?(`K$2#pX{_0Oj*O1F&%FiS*XJQ|H&$3<$_k<7Jvuv7!z%y0s#j(JUKAy$=Csa zkVZU$ZrFH9H1Xk#hH=K%C~5E_ffqB@6ANDr^_;9~x3}h-6ez2i&m=~B0tvy}*MLq% zB(wbD<-a3fBLD5?VEC*;rI)g5wJXY;-)aUHh1*?xweG@OFFtWTMCBFV_8$31Rzyo2 z6P@>NK}r%$Gczx|ht2B~bxc78lHN(QHmwukXg);%9l>)gIdSZ6>IzKq2`eo3FUYf} z;vK7g_FQqJ$i$AQ(-)(OxiI{R9Js9N6wCSa8oZh5Zwd1e)lX20svbit9)q;NBeu9{ z`XX)RTp*via1xB-Et+7U@fI4)#;<{0sFWFu4P1L0x=5ga(MN0t>N14Yc3+m7qen)} zxLFeHjZ70s!0a?0g2X}6BJoWD%Rfc%IXENa9xJb;xJ_Z_WmZiT1H4)Kgaytc{jK|Qs- zB~7Qas%5M4*ijCZH4ey>9rh{JFc{@LOnn*VoyvsygP}+Iq&A`8BRuXdRZefDcpL#& z_av=?KVeYLf<4zaVyjBrc9^;k%Jfhu)q{&`BSg>*BiZqdNquCeAzY)YC`<$g^|^~1)}5(4{f^ThuJ<&XruQw9jB!LlN4BqB`9b2WyoG>a{4t@izVVd zvDkhVazEWB2IP$qvA9xgP_BxlH6aiL_ChcAPj3 z@qCc%IGw}m-DL(1Op|x|Jh3+v8o0aGS>PiOVSKn*Leutm=E0FkixeMyRD;!F2J#=p zx~Av;bRrg4HbAjmCIUM9xqt412_kwR&Fu2k61>nnSzJ2NeW}O@A9|&t zuS}7DaEYtr+hP0gaYhh zzN+V&-JxG;N>G>pW)q?_0A3&@ORyj3>hC4eZLVm3{_C8vB6ZkRLXkT>SkGlV9KtMH@(=Bb(Lcz1k0S%mon9c(F!mV4~&4G4L|8`x)- z$6JuSlD$goj`h-)NNRbv$j{tH(IH`{5V%X2pQm6??AL(dqa+zJPu=;4AOAkuLbAYL zPps0o+8(iP1|Wr4OEc6jCP*7(5vox`WD_`y?!(mjYJR6lc^o;FRim<)`SM zi8KN2;`lKlPwvNa+2#JsnW7;8E>tDijbjvgbYhV-Xb`PSoCbxPmmA|)@YUUsMblHI zxRJxa6-Je^vt1x`np^jN`{8!Hbtolq$h1`*ca3oLKCC+eMeu`r?u&ny?xA)Nju{3) z`2+nDD!9r>^32`1i)6ZA67E;|@E0J;$U7O`5vp4|o;B4yszofbeF z%6+z26mJP)+YLIp5|ennB)*wNVZz0p!4HniiWBqMZZBXhSl{8P?SdJ+>3 z8$7aPnE%02ze`o&NFD;m6~|zcRJHq?e+YiWL3oTL@$+CaGF?h0n>AjDIAo7-k2~cf zNX6A1NLAJcaZXj-Eab1U*cJB{-wikm6IBY_EWjd*|~eM@O1TO}bP({l}#W zx^xwP=RFwh$)h1T%M|?cd(66 z&R7!p8^H>g)Dkm!(8C7mkkEv?g{`g!447bez?^WxT%5sWp^^A>6h{dc$H+ngn0>d3 zr4s3I)zcoHkJh}ToAu5Lz;3>YQv_&Yx4OwrbnFslue7*7ABFcjxZQTh^mGE_?8fu| z^=q@ONmg`;A6B^Gw7#lA9!q^KbL-Se3<{Q%TS~ZWLVAo`gx;(BczhLN6BPhjOf_S@ zy9`IiD>O<%=MWR@cHFC@69bmZCsh!y7YI(hMG0SZNOKZfuOB4uK@bfw6LPdNxmQV+ zg(=+ehC6>lze2z;k*yFqflmC~mvrlkzvb^#T`7e}!vl90;}zLyl`x=R(*>2-on~x$ zN8<%Zz!_|SH9eGHWKJ;>l}e)B(I#6r>Bq)VH^+|3EsJ`-W;sx`y9W_-I$Iu7m@O8i zuTuA?g)#O(9arlIxwziQZqZ2<$=GqS`jZH9yxStP4+QHRVduS;DMt{4! zJBdy7OpICkN}c^#}Frp3pkTWF-cdDHvQhu@O%*j;^@qaY)%ON>hQWXNV{$ zgHVKZ$H(plF!3(u1S3ezQ2m?36B`GBxjpRUVBy%>2IqTensU8XkA?9FeyEJ1#H$T~ zqORLD;-C}`HfWI?-a_D!_2qHUm#qdOw@-(Uuq6D&R@I#cr0GVwfuG0$s}$R2I*08P z?BlFj5r8D>Va|-kYf_IC(9I|l|K(J{{Mh)?irwYzo8pYO*z3X8^4gzvouH!_tbN4B z-XA#*Jkugj{~(}szZ_$g^^&Yp$uo1->tNGQoOK|dDn4fIQXBr5C}Y*YC0`vQ5Zma& zV)llk^J`KHG5NBbb({LT9Vd(RNpu8%{s(}Ax~?$^_U6eg-MwV0P{o5f@z+p3Lf z<;nM8j7Erz%eXcqjby-ge+}o!=Sb>rzNO)nZXlZW{`VVc@tF-h` zjBid5)!|c8_FYF}L}%2)gA=6S3afXU#euRNmG_f3mhFoFx`zk;aP!Nb!e)o4Kb`{Ei#4;R{AB4o8L4}yhvuX1Y>r&8Jx+tuZ=%D23z97$`c{qvf1SDJ-HHO zSZ3X3$<&7{skukCXB zjv^>HzaP}IqAWT#BOZXPsw8C-yr5UITyp87UB&M5>85~Us!(s+)NR}&YttUqid&G7zY zi3O8+f{LdGRBT^xVL>t)qC(tgJN#;Nl98x?EZSLZIZ~1=qzE~gt)93%d!q}t_`kPa z3&*QxCe62zi^}a0iZfmhI4T)u$2y*q39Lizxs&Hh$FbsWsf^d?kc&(?xqyw;`HNGa6IrmTL)ZL-*Ysn0h*>;6zi_=@bbU#@FqRn zXaSpPc^7`?78}QvXY-^HAE+<^0+SXAOKpt-=8k(GW<{)9P7yhasmLaeSWCfZ{`y9> z>bc>>$Q#T}zB3~fCg-O3?f<%uei%|)lNKJ4#qUl^G6Qc>gL`m?5nT)h_tMNv`HpN) zL`ZyZRFA=4+1&h)tQgW8sJVa3s8C^17*(6wTKMvCMkpDmZLM^^i)GH=GLRfzN^qj> zl-jwG5^H@Oqqe=}O*fKbIv%XFTfhGRc!~^9Q-vok>dF0LCWAc|Es*J~y)PU5Q?gdT zG-4l6FdHxq&2$n6{uE1VJ2jp)R;(|i`)#>8Ba*db$V!uDPn!hllHY0T#1@qS>Y_1t zV&N%YD|W;zc{Lj*ujb)@h5R`i7IMBEgX{P4>CIRw&S`IokxAhT6dPFU@wh)pV)znx zG0lQjSn%|%5kwNr93I{US5kW?F(@u1W~+1)U?lJg*&6J95JjbzzSK>lDer^nFn0@g z>y3r{agaIFs1wN+It7a+eed58E`EA5N@Cr>CU+BTV4)>d`AxlusJ2Fja6Py%e5?;O z2_)mqkRsluqLHH`^~U-7Zo`{8ZIw4B04HaWW~$;Y+uSw!@L(A&tix-xa22m=ueNYL zk0C6A-E6csG}7Qqk+vT63)2b@H*ZS7f9MBXlZQCrcUUGyU7Qw6qlA!M3IITWQt2)V zWzLBvtk4{kqbcFOQuzy8#uR{+ehQ(!CX z6o}BHe90?wgIAmKL;O8SG_xX_B+AG^w%!Mm38Ptf+%Hd0=8t0O2KcB_3f#u1&ncmq?hivgQn{ z(rNhtttzT1D)~}5Vl@M(f~aZ4kiLBKr5C9T8U9=fi~CGur^JVPnxD};-qOO_oajZ!%01nJudt} zs~?JHf+42xt$tevqi zIP86KfxSN&scv*Qu^&XdHn3yGH4qr5I4(EZ**=r^Xt!*1qG&uV83pjc4(h8wEv@bUg7bLNlG;xUQA!aQ0>gmYgZ%efcEy5eyxDZ&AoeDwZtTi4hc$1dvslCNigB zTpt$RnG21ei!n4APkbcn*jz$&%f5u`F?j z*%qLaHBq`rfq*roRS`Nzedb_rdMCx}9cu4MTWKBv^X;O_vsuB$ev+SUh4LB^`c>Sa zH;4ri)oXiLAD^OW%p842tSAeVCpAt8eJUt* z6!xMw4FjvsRnF@_ybXk)PUl+$F;n2khE?(LEPf~wI^f~KBt<^xJVS&oA*g~YzF^vo z5p9G{JQ3Ae?CTSZ>rbk6blx31fdd(p*GaW6CzFavt~KZ^pdm78WV#W5Lyh(f|jfgyx=jG4;W)-s41|8)Sp2mr8?L$`vH zdhF`k#uR9pq;fVL-Bpbo?7~zQ{D>;dn!8H=Ue#6jw-jOD zg>z&nUWG>2EN2tueT9pE9fiYx5*&~pr3=h5$6kans>@j!FRa3KKj}X4B1Iv1?=Pc9 zqNjQhnl&x5R-w?{wQ4#ScluGIGb|1!Hmz(YKeH}4JR6Oa-~Jk$mr_zrs27!l)U4Dm zWT||SSNNl@4P+Iu=B#17P})(~_(^C%HD9m^3_i2!IX2->)&!0de|Dt_;fvyG6Gnq3 z@D9?1CY(dR7Se>K)FvPng)^7w&hY3z<_+MFX5mJI2H=aYH=sZRc%d{v%H=r*ZMba7 zwAkC=`4?5%_BQzINh>a^+M(JMyKlp${K7;F9?q}i;gh*Nd@^wl-(G#PJ(E9T?@4Zk z2lC@4%r`vn-xQrGSh2b=UZ}WJZ}S(zA4`*Fr8Q|fD~or=BOeOB6#j5J5weebDD0Lx z6hB)V5PoJeYLR(wU-f69T!PXWjb~PCDL1X2%v9ZA)xfpJHZgXWdzdziC z^?*M|59Hv{9yE*QS`P|*aS?qR^dPpKCs}BF@p&W!XW63C$6eco*tSBiK-*;F)V58k zt8L>KO;xaO(aNSagYm9r&7`PwUrg(ccIdIw@a@UmTs1#qRs{itNUw_c zSt4^9RSWR!J!JVE1igO{yB!knC5YQ$24}V@4si8$1c+qP*-=i}L|T5Qv}I>bM3-e% zVlk4MwNR+8t9c4D)9c_n8sP?5m2>`UAa>>$so8@Uox>|?g21aLN7#*Nv znBTQG2J2lgLuV-OG^J|RDSqd4Yob*E<9rj|X*-qRUo{m3n9^}gXJO9?Zl%Fh|Abtz-Yf6~%kci#gcd=A%mk+SB+5 zd0JD}wLMKpsp8ZAXit+X{L}s@h?DUy=82yB^}gTkkjmH5<4p_PKkV6h|3>H-;{E>= z@2^+b{r?n^{IVH*?J4{ro)Tt)JtZ6oX-@ncpQ2aTQ+`%pkM(1dKY`<{i3y$F7$?9dkA1oXbtTW+ddxUMzOC{3r8oM0HXMnip@4B1jR z6z7&2*MtB<5_AA$y+w7$lDPDn|J5B2>h%{MTB27{(SgR>7P@jfn{6MwSq~ z?Xc&qdWt_rp$4Kr68I0whsEX7$C`bGtfv(nQiLN1?v6J@csBV0c83$yBWfk9cm|uR z2BHje8z(3zzipNyxI-**yg;!Xcxzf|ygZ#=pE3Fq645A=AtK6?D{DgKHFmzA!h{I) z6K0UAm}+-5NQH*W`|SO8e$n8^n^=<>TdiF7{dRvqKoxWoHCZ8FHCJO z_s>FY2}s!lyE-#!gci)lS`ggr9R-r99f~cY_$G>AAh!_Q%W?wbkJN6nCcQD-Qp#&- z)wM-dF6rNxY1e5LOWc<0&r9lqZJn&zW}pF57awB+!nHfo;TjS_M(nHPHc;=;pR8QX zBpkLL_qP~%#sP6y*R1r(g79B*1mKoA9dGNzq^K2_EHJk_58E(|qfE~)U3?)$_Cibt z%L~x~q)Ng_CZy)pO#?JhyBx?5GV>Xfldy@B2}jE{S+&VJim{PR2-?~%X?4>~=JPqH zffFwoBFY4&NG?k0WgG$X?6%=H&8)?3Ht9B-hz;8gMo+ziKbxpFVVUw8BpIGwjHT;x zNlD{OF+D!zy1XcUQoNFB!k92!$LT-i^tT5_KS!BGFzUD5Uy4mCN%43lQF6PebQH)U zKvP<(*iusZ+#OR@893|W`eWTeLXF7u>;@yoJZfU9U@9yA^Qo8O3>49TH)vWYXGv_F zCW=a<{3NbQxnmMsXP| zGU7MoF~xLpV)pctnYnPjY$b%#_{o*1kVeZl7s0I$D8$KJCwJ4}9?TlvG}uVxSK~Gx z#zNlk(fnY+vx7=V$yULk{223e(Xv3X3PyW>AR_l_cBt1ZnFT*JA$N~CT0}y^D0{1V za$@*UnSB(M*-ImE1)pA+h^q7GKz7#@kK{wmkOESEI0) zL>r^97?H7(TA;XNs_Nz*`8LHz9l`HWD1U62l%I5Nd|8s6EULL}Rd`oVY_!9PN^Vhd zxOT#)2}5sDA5@7;>QGoiW`!?7954du==XkCGv19|CWQeW+r=PV)d<&N(MaJ_4x;Rb zX9Mh9#`kh4aN4{~+V>9*Q~?kyeyjMBs}K;OAg zNpad?o<^dvs3^5d3VyKS^_Vtds`Vun1WE5+c{DaQ74O=+nrMxNQ0$gYbOrAK8;GhBMw{Y_QU`6~8S3Nx-8^-?UtZ@H#+uE7}|Qt+k< z@3uPrtMKl-NED)n`GBFyEzh#OZ-(nr0+hHx;g=`3$IbP<@%=^m@X%1+ITL69f+pvS z&qzdQIXe(D(0de1MC5?u+ady-L~^ey zry)rc1zAOkw@B90PD$kb zTuMWkOD%V-b#^Q#qQK;1W>^3@kq1Z`B7~qw$KTo?e>Xoi#xR*Ue8 z^`PQsg`W%MOgQt7uPfERTPH4_h{(y zCH~l$%JNX8S4-32N1dScjg-Z3znR;9le@Qb@805HPfIK0d&-?!C7pOG#6EAWL!`j_8~z_ytZwI=huc*(H$%Z<-{_(y3c!Nxld{KcQNk zZKs{zOK-=M9I)IHx*7jn8aa8GmV(>EIf)9f%}ugBNxK-sW7l{1CXV>dbTJ1)7EX~$ z>qL`sY43$p9pQ77HNAMnrE#T4*tAXmoz8pZ9!O-sXQkCnQmLVLIs7NPloNXgQYVRusc}RhZy=a$QdXX^gKsWFP4$z>9aMw z7SJ55jvXjSl-pxa5PwTpXbv@vlA4U^A6`$&Pt87zjG@a{JDU`A+(6+6YsrRT11(tx zY~vvkg1AIt2SPCA{fHD){mD)g@$3QeEUd8UF%Bq=;Er|2&vI$H@5V(Ntv(*O$^*iB z9g_-&3!>m2BJZUBP^bIwiq5O-;HtRBN^6XskppyZlXWjWd87naXV)VU?qi9>2*J#Ms4Yn zffa|u^^0ba>|J3Z*NyB9A}8hfcdVfra3ddQaKDfY&uDjsXP=yKcx#zzHeFq=6>iyK zTb@TRZ!vNV|GB-0pT?={RB+)7qToG|WtT=ulDcrqeH7(f*Ec5w?9|Yi$!huG5b20S z%EFflC?_2JWX{CGP2l&=g7|=2AY-Bx$+l6NjMTtfmc-smr1^ za(T`g^5MMYCw8UoU8C%ZZ_L28C^51+9vAWq*JQt{7yAWlC+m zu$xh_Wm=-P?GyVR{;4%?`mdJtZ~CtZ3g1TUq|K)UI4cLq#;?`9S;^Zie~K&&rb)`e z7+)cSG2gH^NOD{&Hq;b@FGtfT|4=%M(e+t6yX9z6(9xpK@4vhl(fu#JJhYXe zS8eO=y?Qj6fToUx_dL+RYRX~kP|9Nq`LjgqP@WYvuxw_OWnH0+Z8F308$M{NX`CBn z*@l0W-;N9UH71#_Y2yt_3T)n1i}0WD0o8#-6)W8LvCzAUG|(+?iNB8L;W%Nm&3%f$ zPADTJbUW%Se}A%LIl2icYsXaG!j4`4GSK^cfV+=UdH>vm(txoKO1Xp%HV1~nP1t9VQ+rItAOhfqmsQ=j%oGyVm2S_(`Lo~vacrJN3Kk80NEn04hauTx;ds-IPHecCE z|B8!c4{q*F;!JO4TP|vhi<|fl?z7ls1@PGTm?SkzJVK<>z&Y4XdKZ#Js@yLf0uFRz zeXkgqv;UZ_kKOWVbLZ0qh2`p+xLs}HPFZR#ITrWSHrKaTt?3|fQv-~978 zTN=l25YWO=6ZYRd3yf92Lt|6$sXp~U4rGgcHp*guMZO7P3qRzEja3{Ga2if5FaYy5 zk8ZYhkQ~y^w|YG~AjXNdd?a;sE-BOIZ~@S8&CPUS|2lJlH`QeJj{RF>smQ{z1>Q6` zt_KUenIHuvP6ZZ>{NjOs>gsN$%|dTAT(i0^S?vo^O<@c$t^~c*v4lvL`H& zkR^9gT38?td<7#aSJe7;fW-321Jd&IS)h+-C(}t7r|t>fmr59Jn7718NZDy0DOa@| z8Y^s@0ox1L33X=w*m)pr_S)C87e9CtKUoPSfCx{ zVyY08?%2b0aoeLYzHWd$^q7tz$wZFey3Haa*}mRpMjX~MA&ckTtGcR_da>a4IoK2@ zEpnjV>bcw(CliAfQj>6f2mKwZXMiwis-c%yhyN9<|GQ91{+<1} z80^QHTRXev)TS}GjpD=f0cH2zT-&`3;1D#e zdw3hTuSMAWYh`Hp#CyV25T+`m;Bo;w+RbZ)nN-I-;Cmi0;P4&Uy&rx)pjiUcGABo}+Y zNF#*MXYZD&-54GyG(2#L*ut4l{xCW7e?~r0(-He@48bC(fx;L=f+|3$vwzk+5D;>*P5Y zrO&T|Tsn8pxx^o%@wWthN;`QkILqj5V{4xJQJpw2mqb?7NCTKD7;a}kbFQ%>tbf8**$fahB|2-t^sy4G1}qG^!f$LQ5% zRY_PP{I?Nqm%6NHl2a8%JkaHplS*B%wpMld?z?Hdkc*~r?I!#O0_wA}kq>7TA`Z<| zBJYJAd2^ef{-jcgaaXKy#RJBdMH2&+^}^E^vi;$|v$M?k^l5|EPm#69r(rFybK9FH zm0OIO8H$hp`>UnmrROn~SC#s(C&|vsV2HvwRFt0LmS?}DrK_~dHW6Kn1hSkB&cG`W z7kqQItSF{fp&B=z&Qb|REY`8C^sR^3?v^i#KN}Rm@Z>7fN{lra!i%{vQ<tDMEt-Y?jl-|DqXKn2dkDO+Co8!@c!z09+uco8-a_Q;kz&QhVkB*lESg` z4cHpK%3y?o84aVhLlAh?AGudk8vZ1@@LMIdx~DM~alK^FxTS2?(mbE-0F#3{GLS#& zcpwNrEfSc?$R8f{wUX`hf*XQE!^)d9G=+qN}ubbLaE?{ zav&|M?gug!IVfk>y=s5=rh+2VN@8562tekrT+ix^3r6mfK_vgYI65T+GG8PV8RImxe&F;|*2Gv(0B{R&zWYVaQzcMP&d0L{x ziV1UGRe~HYWylLBKtO?B`D;CA9AjRl%mNT}FjKb5OEw5%)DdirLa7%cLOq>ym(~>B zF6I!(w%C&v3+XA9!#zALEh{GXUy}oamtfK~cb35r&mRFih4`_7r+an4%Xf0cq&$!L zd$@jW3mWa^`MX(d-XJo{;dK3OG1VmNbeK^dW?Ji^58Z_S74m0&nQ0}F2zY;2Oc;JFQ7Uc0c9Phw=eFzrc zT;1jm9O(Y$$RWnjvUz&_4U$q}xXy2q&j9|Qu}7)d@aHf~pYQ4q-fd=Qx+?$&N?iVK zE?wQ=VfrSTG{K7E#gUo9BaUKHpWr8{6pk0Nu?Bxo^U*f-NPOV7BA@^~MI0 zqKaQn{1m&)N_)=Y8H#-72z)h5uSjFZG)&Acvo@uKB!U~+C}}K`?Sc~00EGc(eX3^x zYFMP{WeMW1Oxv$+@ZV}K{|I-H|JKN!7h-@nlN4D1c1{A1$CHDodrT1O6)l>ZWnm*@ z=k8h*09Zn7x&$dn_sMOhnjqD|OASoTvMripz9zOGn zB7Ff9sQ<~%Me+3IabYHQME9t9CpG|nDsRh>c@h?N7CZo(IDb5gue`Kv4CW)Ek`q~L z6(}f+g5!qdG13#7eXu!|w*?tRANr~a6Z&(>$SZHy!sY!S+n+}rz{|lit{Q3PK8*I^ zeAxFZ6+hat=9L057C*3Hg$Ua<(iLYCO9_fOF`?J8nO~IV#UjKAy<^e{=A&3vY&3Se z+gJ~`xsvCYrU@qXbt(^T*^cM{9L^XO3$twuA+l8z^91SlgN5C?qnkqcGS`F)ZJFhG z_g<{U15odYIc;;JWJ8OjxwJ@_ZTN~s!Z?+NO|ZQoi&90~Q86}OH5OsGxda}~6d$+3 zO?(!~UVBl`VRL-x#H7Bg0f2qToyG4%sZ=nJ@y2Y0B_M*gKEH1NG!@>|*Krq3{qj(Gn+?KTPVt8=Nxx|EU#6NuP)GnIYw0!T>w&+Sav;A=`N#Xy!<66?g*>PR8 zRc*Gycjf(TQO$5*-6OVMNw+lOTgay|K{t8r&ow=VZ{t=3|^`9VxVJ$AV%R z7Zu`oH?`6#sG}bmL0Tk0ul=g@alzZWCe3uNB>Ph(*-v8Zs4;N*u|(5G zrhRP8PU|(YHBkjoYEcSzw@qYYuu4&4RhO%W6yNb3 zG((qGIF<|fvl*&4k$eN8faPB6_Ceewg3e>d1=3mFbU~!(J~B5@7yX2ASQ`99S+DF~ zTd%F!7p7&IG{scENjaf}wu_Z>r?Ck@8zil8Qk2HFnKLNjZFElIX(&E<%P6!X0CO^} z8^mAfu7kF%PNLToA2gB9UPC||v$luYgvR1kWRk~q8)@pJ)sj<>DKjY2k~T4w2+L>! zFb%MuDx%iYixp|^DIE)ye<>DV4t;qz^T23;H#j17sak!%E|SP7N{%Kzka zX3Wt*r5|5U=e=xIcr$hwUM9e~vNW)1S#X0a*~J4&tcpv^Vje{dErWT=PymsYK;ELm z(Hj5{$Ea{gahv4?*s#_zo87Q8RzzC5T_OAn+ ziLuf@Czye0S53!`AO_@?96I4eJ;|xb)wIrP5YlViaUP_C{>T@=n2~TN=p^f4$f=MF zK#ra6GBK7T*HNP@gXc8X{ICKvJoZly7_H}CQ6as=grnDbnUym<$;lBr!!N!vT#+ll zOx8ikfJQ7LNvCnvnD8)~`n=qRnemEs@a1sik)7FuVx|hR!jqzLm(zFjNp0h&{u3%2 z%W);!W+@7+Z9&mQAaWS@jJ1oh^C(Eh#4d-uezN^YAUJUk*yZp(!`WyD#<6JHRX1J^ zpFYYsEJ2*L)dv>xKt9jE;6xn2unta#jDt;glUYQGGO}}RuB#2b)7t&hYcS8JD^{Is zqfK*t7n)`bQ?n(}5Xj@zAgbR9=pdcu8iH9#gS+8CaR(F>GL19rmFB9zn z917yNWKXCS@b)C5yHVnq9i<%(ghlR2G? zrH(G(5RLzBGFu9H4@Yw5CaWc>7vDnx+=H4jeJYly7E#lFoMdZH-I?xS)@!2+fW3Vn zV?#8Yz}XgUi5~HK2=_e6o4O=r1Udc4oT9sM|xU3SZQ(}Uk7qI||HWC|t*NyoR{dNSl#u};V`CiGjSY3@On(VaMAzecrIUz^ zs%bidC$w3R9q=Jzpk9e-)Zd71Vd4uPD|^1NU&uYgA5_{*vG6BW8>)yCXWP30c_}bW z=3AMnk_nzX^5>Gm3}&%s3cl#5Cesjv=Hjw~zk(5URD z#6{BWW?FLM1BMw+p(Q)AO>|zHLh&GtkU7}>wE{BUzE4OSM))=+M5LadmJ0DD(+f$i zW2B4*`Ljsqz4m7!tqKu=SRMNp-3N4++R1?;Q*L?m8~!%ijvuo_2>$dOP9Btknehp4 zZV+JESb(IrgP_P{tskLM7f$FwHv6M|B^LqK*5T;==J8co!ju|O zSIaspDEzQ39QV{4Y?}t+u%ARsjnTJ1y*{1$7s)D2MiG`J5R|d0Vnhhc7C?pmnhvI% z`gPP`2IJgE>}$ZyF2`Hj*@u8H2DBwGL>`mp!;w~6c8-S!JNcG6E%(f?cENdN)w(Tu=O2?j)O=>hn#7@ zm$2j0)AHriX^dGc-gHl#qjillwlR|k;ZBXDp44W$kUw=mHFv-SjPD{K>%2 z5y^n|$=l)`!oWna%+=v|>`s&dMR`KOe0vzE0b`S6Gt>;ADvK%g*F6Kyf0kVicD7B> zWOvfMl&$X3#ln=_kW!9iF{z?#woK071iw$l+Ad-=w%M|!vz8?$djH?w$ilaKs_zfqO}ph~93$57$SH=w#v=-me;$_)>^ zkHvJqIQn=*Yo3>H7ZIFc^@P@Qgq8B2r!M=oXXcSeB}6(=yrIE`O{(r6Ld)7wZ}3>QkWX8CEOxnQvX3PJS%K=oVX8&S&8t=8THLZV@BI$c=_Zd)i(07eWajzq}* zCN>#DW#wop3R{D?&P%UU+E|bXt{_CK2pNaueP==9cgj2yqnOIR@)?`GA*0YO4VniR z68__ky{s0~O0a>mdNE4z9PmeM5`d22l0{Qr$Ci+hMC?SZpj3!tHH)<)O=+;nMPu1q zxsCwIs=X}72`6G35gQZIX?AJY9^&H!Mj@EZLfJY5O^g2k#~CR_P_LEWL-DnDzx-Rc z^8YeU5Y|rR7oJ`}Zjb%?yAR<)@nHTM#`YVcFGYe_)6UwUZvz)L~*Trj6 z^J8^S$@IM4B*))`C;nxH4*lg#;IkeUQ&&y)%ivUmMUvMg)rbbfOY-U^Us_x@v+T;l z)A}zK$oo#{5cINkF)Q3 zj0#XE!}?&ZJ?MX8DHZ%Rd^SWnIpUXCL?!7lCDFNrI|=aUiQUXfuj;x?zq&vfhhK#+ z!k@~nejR6AG^|w{0GE895_G9y1)HnhnvkA3F3iPD%bpt+strc{JJ}9y2_%<2K(A$4 zn!8{v`$nz*MJ3n${k>LJm)h~!0cC2MKOO@Lw24YfemYC>J^91w{qUiQ^b6J}*A9;SR zdv=?(HFX?zeJ)Mtc(~P(1iw=-<1U4|pOA)Sw&Q@i?sex`B;NxIY&AQVEoi-o$NHr6Mfb%J5+%qird{BR@D^&M90JqS?x z%sAEDss@wdnZJxwYCKJJu~l}#E-YEJ<`~!RegXzD7BVfq#_@x+9}|XZw2>{hV|$E z@OlEge!^-;p|Qbx;&QyqzAD*}0!wU(q;Mf)Z-uq+yIMJpz*8LS@|S!WER33{mCPTY zzGotDE*^rDDeU6x5GdVfeBaoCaWsIudwUGDqEDCXIE@EB>|j+!sc6} zz)j{@%}FVZXlzzaoASakLt&Ya2o}O3h$^-4K}(*%d7aS&YMR(9@eK#{o<=SVJ9#)z z)qQkHw>eQ>@iCIICn~{c};Muae*H9rUve3AU^0}Yj?n$>7l@oO?IB~UW8pn!nJ z>Iu>zKX$8do`$iXOw!ovdwv+lo(o26Q8AeHLkc2Jy=mTBy`ZamVogCJc4aJHYpc6d z$gSEgYrPV8TuZ=ZqCu=*Vpx{JbL4JaQr^?K%qme>&d=V21C8^@%KtZh$JBuL*tD=H zT~#WY9udUBr96=k1^!vv{z9ED?TXBA+pcp1o9-VNua!;O9r&e zAi^8t;LRxKHO2>g5u!`lCoPU#+2$0D=ZmrkHtKcd7WGMja;%XhO|!*k*H#E$SMQLf zkLnsMj_Ym}t`;g!eAuAKo$-wCYUQb&l$TuUMGdl+uf*ke=i3uvHz$NSB_$rbei9P1>R;}zEK zSPEB>Tb(~2goxHCg~wo_>=@&a{H9Y26jWm{7P8qH)3d#RR^}Zc56G zENGVZKt|cv^nz(PkCzwvejn_=*3}up)BJGQRsR~gPzCcbr z7lMGU`F?-<`!F2QERqDEie~PRWx86(owSSORI036>&KcbQ1u#oP`1bN8)OV02C>sY%uI-i?XU#l7M$ zWKE}P2tGywrg)6FJX2}xzrrhOnMikcjCYY&rXWbr$nC|il1Rzb*WEgnSro90eCY|q zMt)0TY$ro8tJH=T-_u7CCc&B6iIw@z6;v$>sN4wzZN_F+F@!XVN#b{y0&bPoSDX%3 z<9AsUhQNpjzt}ozj)2?Loma12xgdt{}uA*bQr(5pt4Tz9v4B2AYGa0!JPY6>v7hq zDbF#8;#@veP>*7PzUj12jb{R1u5Dbt$fzw>pE!@Sll_U#Pdb`3d;w1A)vlxFSBy^X zlD2U$cE5azl~Zd?D<3KIZNWYQJB{9ir7*HOb?8^GCJp#?MiZ+)GBE5;(WeY%g!L7R ztg@KOcz03He9PjDciA`bsH_kKAdY<3Zw(2xNkP7_Gk5##c^Afi?#F*9Xi*k~D$oPO z<`rK+5-;r0&Gm)ffSMR3FzdZXe?A*!g?F(hqmKsa%}8lA?h~5dcV9GmgZ+I-T*08D zMcS?QxinTKNNPTEk5MtTA$x}|#e+?WCwULsw}G^%hh@wP)ke?Nw^`)=tc2Tigy9VO zqN#a=5*Ycccq>*{>3fC}>{(r3rR137Y*TE6?WAB-+(Z)5UP(k1Rgp+URpgDjon75@fF`(SgPm1ap-6nqgrbu;-ovu#ngv)^U5eIMyDasFCrF8P(HYIK%>rIpOKNg*BET9(Mx|+Hh~`I%{rGGp^!(@16PZ8};W$O*w1w z3!VEaN4>{&mmcCJc*aAhRpAMb- zG>Guo-y&Df#Q(^Ro#3CPDOd0=MKc9(%6_G8e%(fPzGk-Y7wAas1Qn_Xn1Unq$Bb#O zSr9RsBVdLd_86gpJU&20d6^e*TWF4C&V+{e>VUD#Ytrt`8{$~zt;HJZ2#C^MWLlc@ z-YlF1UUo&Lb({ zApzI|wx99?3epJQ1kX-L?=O=YRssGFtsDi~oR1E4E%x)jzIG?(4tquiBw7LV&q4|` z;~{9?!QkLR@e{<(!Cw522R_TW0GALG`KJ+OCBt^WK$Llb&QhwC7Ogj)`t1<-fh(jS z1bTkNEQvI!VONm)O|8+y&z6ZFRZ&HqLNdgQewc zmjhDN11@q$KRJOvqZ1_d_J??4GBET{C+&{(Q5YN}j=lOh!F2v*w#dPy^<^Mr1em53+^c*wt~#8#H-nSI5;iPrA>2rQ=x!3(bSHCoPv zRSQg{y+3V~W$#3O3E5BNRS#kU#G}1h*hYy&z@RrFRk}LVQv|yGn>nUxT3*M4 z*AaCZUbkA<1Q_#sa)$s3mpM_6X@zGAVM%{YV1fn6A_YBj3o za8}~C&Y-5@9_|)RYKauU+fN2HI`Iu&e?oPGwLn|BDcpq*{_o?M1W{!icf%`2ZIP_n z*YM{@8a75+3=RN)2XC2$>ymL+T%eIvG{*gSXwrh0u&O8F&yBh()Mgp{#%@Q5ws0c(p@Q(ypNX6nvEOer^En zdNQ3CTuidw%V4E2cZI)W{R5M|;YZVQeydwU`1?XE%3#;3Mfr~hizIs4&Is-p&1T_w za&=P|5!=D`DN-VG(QY)jrFASj#F4h#_K_(UbVWOq6qWE|<4*BScZ5_=!WQN?5WP=S zzhdWLp{bSHd6TwHdc4cc8F4CV$??Fri0tRXnPSyM4bHwAbo)?$r zBoD_rRl1Fc>}@J|&p07bWE7(|qmN0{j~2UfRVhxLT$)Bxv(<=p5fSo9J@Bf%cAN0- zTb(#s|Kw#+a=X5!1rUjUVDi~>l$!_Uc<5KnBHP54+rA@b96Vs^)w!dpnO>1I4od-E zoeK|ThgD3mFW}={ovyzs<_ExmjZ2r4vi*1>@rM2~9QT*2qgi+g8 zix`=?ID({@MMCxI5fn6h@BApQhomqzY_qv5o+)A?Ux&|;{w@X_ z{gbjbTgA55C@shxmZE~#W7lzyHQi?JvJvoGwH`*`?}^Pl{0nFT8R>9iYBu{viF6?T z3SR#$pE{e1{Y0W*d%RvK&z1{|8~0ECwJM3;oU~KL6a2JE=%}rrjJG#>OgnVEU9lhf zCMj4wbR%2K8h0jLrumSs`IO)u)T1Md09f@UT#%{tnL|hvSJ(slo7734o|!ONbRFFn zTNL(JZd(~loR)D;gN0B6GoKJtad|Hmx1P!Is=0B2hPe%Ixqj2%=~1~%BnT-XjL7)s zNLh+>BcqCX;9gdX&xd}51kVcjbDsMVR|H>$~=31(OZp@(CwV*2_pT^==z^dWQh%fowT$MkuWJPtRRpXEDiTw15NeXyZ zbTEeL2zn9zym{FlAr)s0f`jDB3vZ>8F+5*Ec(UPWC26XVEu4R$H#8L9os*xT4`CN% zZqyTV6M0drI#8^D6#bYEyuOsAQSahIDw};Q?95}uE}RH^njQe3K=E={k5@ORyDE1b zQhwoTiS0;Tgm+79i+5{m&4Kuxve-iXg#TJuY%x#)q}%ZSp)8i}g-;1+ooM>2rC8u@ z6z=V0V(-NjsDnO+0iQSakbDgGz zlriuw(R#bk^9k5aVO;|`lh|uQ1*B9ec^dg3whf`4ZXjUXF(_{?E6Xm{HCoQWBBWYe3zFmHsUK0%Ifi0)@Ep(_RRc zf*tpX`nOiS(-kH%ESTgDNYgosdTDhCycZe1EHObjuH=yB5iNUj9aT3F{m{ ztBwI*5GKNPsoitE8>~z%8zXXhbefT1GhVqU%Xp=}OI{h5wC2^nkBphX zQHHUtm5hyI<@qM93%*weKHH9`w&1-0#XOsI8~^U>X}!QxwT*iINGtJ*PG=B4o^>>U zlhXzWUo0_`y|{l}PcBzK0zx7qm;5?dyK=f^Q$N_|cHu%t$`*N-9(Yeb^dFSNll}5E zo}E#JhOs$zFgo*7dVxUY`7XgHY;xJqu#;n?B zBi;lrQ68R+xT=5TGi}kZnZFis_%=4v{96px=A!lYl<`w>i)_AW;hrfDJTK2i{i<*Kz<&uhA+k*WE6~zcuS0Z;ZVrbi10C%b4txhN*9IFZ*qJTGb+XRnfw3( z28Pp3d>S!1n#cWdR3U1EB5m*+@P}7P_ZXmxoO`;bTsGh#B5ggSI8`^wM~#>aQhp4v z3(8^lc0P4^R7*23C~gJyC|FUd2EYAvG^<8m0SPG+moY!g)Oa(pR*eUfZHK&HWihn4n+EI{9M|1LQ=g1SF(_| zr`>f&`fdW?)QSvCavDYuP55;*om_uKjsRakL4}pJGV^X-wOPH3yvM!*UqCc3uSa8) z-Fl`39idG=Lh*Jy;gspNxapz}O14w3#mL?sm(R~gzo>!+5ZRKB!l8crWDkBhc6_Sx zwPFUSNv|n9ateWvLquerY2UIa>eJ2omCptZ>anM1pd|Ij?j!>ej&W*Wz^zQ0_DYppnh>?T}t0(+E42noFcVH_~t&7CeD(x5G z^h(_;m3VLi`hP(3>jgqmN3_sASrc=t<{nz0OzE8%A(FG{XjHIWBA<4qE;U_dQxSZ~ zrh_Y4O3v8!%D6>paGbTmSABjJK56K%0K*p!(ly_E?((=tKvL=`lLDm_Ma8J8j?{ga z05YpzMw9sj@)WYcmBZE3wA2dF<0-Lio7Zf6-ck8rGLx7JI`Tet-m+*AuH_@$ajQ}T zy`0jr1qNqEq=HL@d-2S=cKGR2(J^&lo_6d5m`%)Ic)Zw z8@5m3Fj-1AU3P@C`mcg5kV*Ku*C&L5y&o44GGTc9rwy!uz3EhWBWgaWc$qWo)hz2X zzqFA-whPtT+FUhj7lSL+6<5o4bK@7|nXAnR>WjgKEvm%i0pRjLIL4h;%b_PwJNs1^ zMmE*@s{y5Pjl8W)X$8urHf|kLoLC{pbmzjhTM2-MrZPRyZ<0E3|IW!iH+(M7y51>8 z$bj@pFB}bq{-URp!l45hW*s=vrR`sO7Ijs zV=>dKN}lv*OYhiFZ|<6zar z$7^1dg%sll%_tRWs#hgHg;y1N1i*^K`J2+jEi{~q+``rHzAnEgPLYTAsZS!ivE2#N zA~SQ^BtZ3EWLF8SMIj*uc7qu(jO^l0C$g5|O(dZsJ3u=cq-CL+B@i8P(i$!#8%&WH zeOOS_$fDaXMb*q}>Yivog*Df^{NU~~)6mOzaA$TVq-nLtf`Jo=MkPHHxvYg&8b-=R z-giu8JUnRYKB(XT#U^HXWD2Hlp#yCjC?Yv{biZfGqsnn58@OZgfI&171utXrBOX8- zb0>B;-~aPZV)nl7pZZUx<0^Jf7`Mf2gr5fkwf$Z9E(?e+WSU~$(SsYjlM$k;_(7K`E!LhRb@QW7-BjwxC25 zAM&K4D|#@puDXaGTFmcU1d!P9e0J7SFM?-eBaw$j$D1p0DyxxiTlmuYbeck6ZNS&` zX;dt|My{?--WYL~c3Rd0lUEPyeyJ5DurAH$7eaFU<{g{A zbH=_{S^iqJ_ICqQe{;pu*Y@wO%~8E$Hb=Fp+`k|Qb=a%vQVq}W6?E!f4rce3n%Lst z7MsYdpI+8RVZpnBIp0>JM29A|cW}>hMn5ehqsOaamPvs_d04i+!h(@gjL2XU{872z z(1ZwVGEsop$k9)V!`GjJ)nuFbMa{Y7k*AFkkH%Hd%=B|D?*Z{1$J-cnW!z*&OnG70 zlj`D&+$HTvFC#1dsTZ0CZz??=e+c6%pjK(tq(WKN6L!um%o%BhX`VplNv&P>{N&`k zQFWv&ww+F#wDe(yb+?SJbw15#`GQk_R6Yp_PWkvyW@nx(bsJaJ@_1D3v9g# zp^@wWe>ZJ8%n-`fU&x>3MTQWvIE`V=*P7T|N#maS)2>Ux6Gg5~en384xYSSQai)=; z4eG{+w}fac>?e{-nHZ-Ej!4|3={eNGR&#;B1y``95n z$ap|&WG09_q>kI@$sVen1`H%*-W%0Juu|NfGK$?@d>|A@F&K|NLLbEu@mh-bUefX6 zh~(2S_z{JdcAMD2HU{ei2^Dj29LUb!p~v#*DETCF=jAEPR*CFc;1cQcFep#FXAe*I z%Z?#uPqds!@-KorcFw{AqldhX=`}wQgT-2_bp-D-`@wcNXT8+lQia_pyH?~aOSQ|I z3&ES_@V||{W}wNvM;7!R+E&*`d?X_dSS{9`)qV=Fp~#!w=S}Z!PM>-{6MAWJ3kj~N zm~4JmrgMSGPL=o69=Jbgi|c;tRwDWruLzttz4J#IV*iIgFrM zHP|CU{9ZXQ)AqB5sl2=-*TUSJ6|R2hy$6g;4Q{EaCZR@z1b_mqQVI{MliGEQ6F#CS zD@=oBzvpX#mf^>z<^I#%sSDN)JekdvEjtq`bm0&St(S!+kd}x!U|YtB`6R?PJJL3Z zK)1J5Q6!Sn{^o}t|0iN&t0!q#LR-hV?p|$pK66C*8j0BqP14EbdOn3at>T8A%=!Y_ zS29S3Wr@sWS6_+&+^Ao4aO}5tNl?A#w(Lf?H{Ht|i8YMZYw^{C zrki=(naJaw9l%j_3Vv2`k|+vaE%RX>P8D-VgC~d`j@6VZ@pt)Z(zSEZvV|354D|fG z&F`3dhA)!R=NA|++iiMF+?eUWyDX0we#U1}y{sl+rzT)j)>ew=!RG2ZQ~1Z41~-+< z<55O#qz-0Ce||k{0NVJ=mnZdM#GtiBvUL$|71D?-54YGwSHhP4NS_bG)^kyWS<|z& zL9;-^twDATD1?r88PLxH9Rh2tA*v-M9DfWq{9lLtSy-Kcv{!YrXQhElyMGiwLrMBn zBZRqGp)n|{2zmQP6FLZT!g@;I-vbu3&f0)a;ojL&~#KL+_1ouu)C>OGywC_r;!RkwwS${_!m>JRK+higc^_vrgzJhe*TNR7G_=7d7_=z7)z; z_G?rPKDfV*qw^Z?WG6*qhhLJCD=1i_lcdLe#+&SYDxWkYx2_laqcl5JHM-&xfae)CyVpqyH<-Wsd_amwxqrlEHy;Iu8)A?@Tq+jKmW!yitj#1 z6%1+!!<`I+PdzM0Iv_ukd3Jta({6buDOW@tLS$H?I##?~>t_CM;~r2Utn|I7qQdQ- z*ceCoW%ZNgH2;8$ly{9rOSQpPA<;IDsQ~l&0e@)SeB0L9*TO#aL~>|U zMoy4Ubu>!TsR*y8H`#~+m`HxNg7a^t6V(IzHi`tvY*f)Etd~j2dc1g#e=q+-Zo^1} zqWZ=B1K%p^neE4l^8X@>HSCLD#>TK|H^fK@Rd4wp<4)AA|oO=rmKLNP?o7>F`=`MH80Sl~;C#Qfjt;F7PDWH>9IqC3Ttg9rCT z)oEJ|Xnn(W*VcvnSy7`n>>?V8+6{N~*(}?1i#Q=3ZVsz%xzBcVmEh4UtUl3r@ew=0 z(XjQuy>cG}GYa>GI~vwfzW!oGYd9!W)m5Lvm_sbYzp~#l0Px;+OCS8&Zd+=}=Rk+U zafoe{$MrYi`~^X(gy`MKq0) z@y3%fZiG?SM5z`&-8A*p(6QL^m}@=u7^W)d=}16S)|bT0#Lt;oFd4uE=-$L?BasX` z98b~X+W>aERiH69i>N!ezARVjy@RuLNt2>!TYb2%ozT>Y0OS_hEM!ndmO&|+u)X8U zUxm&oPoQj2F{CPOG`$8X-k~dV&(q=;Dlul29ugB_eAxn?T6+Wk>g9M4{0Bs`CF}{Y z*;5Y{mdA^D+vC8=e_(Q{Kz@W8gv5YKYzOmL9F)8810ii@)GR?g?mWp{xG=_i&xd35fm&y4(5;!pfyB#9q5{}Si>INw~Y;P{O^NGU{k%@5f zZY|s=J{Z(=#YJ2B5c7_(vZr$|tCa9UiDD4UvhldPsukRw+$Ybg85m6C^sZ5lDJfS= z`LzAm%_(L#*HR4CMX&l<>pz60g_7PZI1x9E)i zihht9T+^2=G;U)-4MyQ&5(ddwX-3T6Z8)`b&BhI&sL8s?iXrK&mZAJvuv}OYdY?EK z=u2zB;Z}Es9Q-Dn;a1j>cy&&%AJ4}y2FV<>X3#qfjhW!wQ5@O2|4b06518$Acsm`= zh-0_iOP1#8X4S1oK zKXD=$K8@jap}Iu0?3_j2G)1VA83&b&7~13XegSwChwHJ$nCTkn`A`#U%WPv^P`*F{ zTWnXh!-jQXHDEJ0i?X~g|5jDtz84K{8z zlhMZ%=-*_`A%Dm^OpJ0$3%oAF|H*p>E)mm2uV0;+^Q?)kspLA2kdi!w8N!M%n1-QW zJ{TmQ?9PAHgzRIso=RU@0P|i63 zS$Q^1TF>LijK4fl`-+8^;c&FN6xuM^@vM^@na!Pm^1~(}W-5@0^yHF;Y25Jt^%P#v z{}GngKfKE7iYwvRDLm(_$B~smG_}>8aGrMwhhRIMNFL!bt$P|mAU0{6tAYcq*X9t2 zBHQP_m5l~RXuo~!&pmfOWD=mxqbI4>9V6Y>Ae?dRo)Fgk#G^#5fehimh(A*=G`Scc zM8+)~xfU+<`_74pK^Jl?<=!N5VAejV60O-m{i4vp+PX>^)!3gP>& z<3l>#gwr8hWFoa0swPVnA^06*-xu&2m_vI|iV5-OD73#CF`>e4E5%ya>fgSNX#j~8*N;zK9Cxc>!(FfJAIL!3l*}?#@8bPN zZBt7yRUHs83ead8K+Wv75i#cHuiDxVVwYJRBx{>PJJz53=#XvP?@=yL*xX{}h9XYj z`i$b3@XZ18*S4CpICk8&I>VTWd9TS&z*f`gBcZ#~>0WXKHD@l4RzjfT$Uqo%XQ)mq zx|P@c=8U`pEZv3>_wYhr`=|3A7_Eb+FHO6k8^*gf2;Axu<4VOxJMggF2Tq>ah(gsR zxg8oM_d{XqwjLs(1o~OXpLLFe`gHTX1|Pm3Ki#BxAxzk#smM{)GccUiQqVOU2`5$+ z{yCe5#Z(bXCCq6t)BG};hdCH7EkxC}4(sL@uu6n8NPhpT5KM#*!!Tp>o=oD$RrvU4 z0wy1BP6(Ai9faREl_2$g1X2{y3OQ+KlUlzHtrZ1C-`u9Ed~<{SEX|D`V>7`=F?3Nl zk5jtff8LZ+fh9lP{G;_nzLiFo;b%YmXkF1y(bRgucyd$oShKK>y}WdaoMIP5)A3g+XlLVp*2Wk7ysG55{0D2h zemIk@ZemU5htqi*KKzv+653l@Dmn;~q>5P-0VxQ`+Ekze+NjA$R3L?SOW)6=nrb(K zc>~*qy(r|mU&R*7dokkY>_vJ7_Nly>Y7YKi>kPl*`2RFOL$uS2)A2;4M}AU^9w3Ct z@5Z$tX{LzL*qa1eF1*H*VQ{n9)ydu=8xn><3jQ$13Jk$4SdJV5!SnDP$|a%E!AHq^ zA3iUv_p$T5xSGrSTa1N6O8%JJ0O0GQt=lUD0Jv#bSI8hso0meJ@_C*D3 z>~u7rX{(hqj_u~x?P6Bs3x}VFzsg&)cI$^y^cbAG(3ZLm4X@|61@L>(%f8m|<=7GF zs4X)vbadc(AKA*vMq9sFXzNY8Cp;577_X4Ul@Gjrp{pkjy?k8Q?l$4luO)R>%R8uN z_73#tf(Hl#V9$)KKv?~SR5i)WfnU6%+*`+FyX|^<8&MFSwCGAvRn=Vkf+8T>$@Gf0 zoesMaz9=p>0=;~&6X>Vh3b=bAVl0k&rmnW+tFTDC7J|oD!21-^a2Z>-0aSI9n*Snb zpV)zh!Q!?p1|8rR!#yB=ER(F7h3OF{m7OO1?)ok;@f1^W7myrMz6B$Vdpt8ndx#XNxoj1i5iK@{%jlkzWip0{OJCouH}WOXS7!=a-v* z=~iK9IuRM-7W&!~UG0{Lk@K92L16ZIg20JG8#Nvom*DQYbLhjRzq4SdcuQhdm+bWq z5~13kb;Ci&<=XJ8V7LrA&M$_OH+))LPF`Ni$vb@M#w1{tR*@VidsEs5*eX+fQ}ed8 zH{~|?Lam>7+RQI5=4<%4i1RgW*LpSI*3WNyOu@J3F{{D4!C6_DHDpbdosCUhOhZ$p zj95z>q>21C?$4=-{6d=;Td1FMht4-={f;Z@hy1|vf$NPeB8eZvfYT=P_q9m0SMs;u za@FO$GSjT8^xw<8FpOi3-y}P2Hn7w5mOO84zzh3yu7;WpxPHt5|A|$4R;7B$L_}PH zx6SnT(v!N;!LRGmTK-S}^2@(`A6gdduLVCh;}dZU%6Rhi;P=)mn8t;FNqYO?Jc=J< z9@X}O^ka4_R#!ta)c9p~1L`&IQ+z-AbLo}ra_m)8*X-4c}aHUht~R z^{Zc8^dfxh&i;+>0)N3S{CNk-80zd`Vx3-Jm>7$RhipD+JHitt3v2ges_oXeOFf?! z_|PX|mD&iMk)Or|GMU9L73>o4d@+^mW1C9ao$zsCQ|W)QK49&9vKTYl-Ms8jLloAK zK!pbR6q{Oniga*REwqEvdBYBpan`yKE-0{J4OcjSsTO7GfPPFzpbibHTG_VZnV=WY zsQmuepLfsQIh=b^W+bN9EA(X*Hnbhm8^u?|yGRe_jrCw&Y7h2@<9UaS_l%$V1^;q* zitk8%n>MjY?4rIS-a6mJ;-}cesxtN~e4M?adqE%Ey8`*+;jEG4e?WHOIDcS1CHPmTObOL*```b!n!phFxTb7DY#E z{aovjFJ1$`7#?pFsgR64}5|iOEqTRzo#HVaz zfAX_zKA#P{C(r>!@m#Fe-@au34J; z#lRuvr-fSlm7FK}PN_emIbg0aOVhaCVv=@H9+coz$1dS{IBVM)oXghom5&7YZBidD z-47$-Cve+LH7H^X`?@}H;kEUtUktWM`Lu}JWTVqFw}|YP7y_$H$5x5wBD29IiZ~C? z#)q3o_z%J(ROkTk&7SeO(lgm9Id|K+^{I!M8cbMznBfT3A7XFi$8qWo4@SFp+LH=) z+CIihZ(=|B@l}7QFn{Rf$b---uKO5;+iPko+JJID?c=nyK2A#w@~Ikc?0RgY-)X&)!K*1 z>nD564BW*=!UryI=wTpqU_P-qz=%a98{ z8M;@vN9>jHnbO_)Tq8Q|d{SoJh1rfzz1dEJI&@cldab+qPI>2x?&`d$ zqU(iyT;*XG{a9@T>vC%lF)3ok@Br1q(6qz z4%sK}&0Ud1%skf^j#sd>G9~_V7rD-iGN!KX!9EpC^odM~R>O_K=1A*WsAA?4K4oi8 zH!6}K5`&nvi#qPPFVutL7wWAK64qe{)USgnRxPr}PFxTmJa%GH1QBt-rqZ;y=GW1@ z!U3|nd9`Qqb4h4pCH~4&Eo~0Hk&Owt6q2>v{!LNU}`p$MfJRn>m9`&>t zJb*uq5BQ$&wdFSQF(o2MwIW2^E73)2M|qn{@{F~%WigL!*>f=2RlDM1H1&(rS@#rn zSg(-BRO%?pwrtN4Q8Mi=zb`Iloc_>_D9pl#n+VXC&I885M^>#pQY0Gr$hw}!SYI4r zet+$e{9$?|)+5-t!wqmuO6F7RZ1`0EM4l?_e0*wSJ}(tU?XlqkAS!0V$MTo&EG~so zl8JW-Hg(yJe1V6)TD_07WdeF(X-hlRw#1|paK=jbA^b^J+H*2C8Ei@0{o$>w_X z_hBRSK9V_vODc_+n3oJ(v=B+yQ*q5m3WxQnVV2Cp4PyjU8~NKb?SaBYP}Mwnv7>$iE1gCS5h{m%3J0_vT{yR>jH#@DD9;uE+vx9?H?M0n{C>SW zlViAj_mrwC>i5*svQ;`6t5`7i;*Bu4T()DZbfK!(V`Q9rpiDpcnC}g6G?1B%Qa!G!{4AYRQ5#Py(+&D+$y_jtpOs`23TA6T9aZ z)(jNEFP@N*lgYLWhuhQc84?ui5u ze!&r-HC8Upw_LJEqq$MsaFVkE%$tpi63l#Cotz4x~=RNC~&H13LL!R3MrtO z+}b40Q2n$RtX6%t(t|{0f>bfqovbP+_`YFZjolPfJ`}H>8Fk(%R?DQ@oRuDh6_!jk|vrApK&gQA6qa;K)QUyD+ZTg@(+z zB6=3jZBU{pga3Nc%7)D=X)vMzT?7NV9SOc(QMMaZv@6E$G}q$K{*knc~4hO7ez zDrz3P-$3TLl}W-V_EKQxwtB#i2dc$*Oe#|}7Ax3Ptt#P>305g)yJ!S`qQgF7lhBNh z!3_vjl;9;NMcvAfiUl4H=84CtRN^Quf+wJdnl5o!r}NZIdc%Yg-|0lcH+hwy9AYR`;bg3fD*mLK{U+0*H=&1Ydxa zcAZ|2o8`(Hr9Z1Cl8AV=ri2eAtr(s>FKJ@Wg00kEJd@wXXYRUwyTg?Dc=+rXTnCTj zuVCGzNAkn)$in1#W*DU5Bl&H7WZeE=3f!AtFKW3rk1l!YY}Vpt*_*L^tvp?R7TvMX zZ7`>?5#cpow4%`OQ#0bZq+I~81TESB^5G7Xi3HVQ#Kwd zl@nV6e3q0_>VW9?9Fc!s#^*y^$kKf>VlUkn_&f)p= za`5Jn>SnSK)g+# + + + GDT netCDF conventions for climate data, version 1.4 + +

GDT netCDF conventions for climate data, version 1.4

+ +

Jonathan Gregory1, Bob Drach2, Simon Tett1 and +Karl Taylor2
+(1) Hadley Centre, UK Met Office; +(2) PCMDI, LLNL

+ +

15th September 1999

+ +

+

+1  Purposes

+ +

+This standard defines a set of conventions adopted in order to promote the +interchange and sharing of files created with the netCDF Application +Programmer Interface (API). The standard is based upon version 2.4 of +netCDF. Documentation of the netCDF API may be found in the ``NetCDF +Users' Guide'', Version 2.4, February 1996, available from +http://www.unidata.ucar.edu/packages/netcdf/ +or via anonymous ftp at +ftp.unidata.ucar.edu. +The standard is named ``GDT'' from the names of the authors. + +

+This standard is intended for use with climate data, +and was designed with data generated by GCMs particularly in mind. +We recognise that there are limits to what a standard can practically cover; +we restrict ourselves to issues which we believe to be of common and +frequent concern in the design of climate metadata. +Although this is specifically a netCDF standard, +we feel that most of the ideas are of wider application. +Our main purpose is to propose a clear, adequate and flexible definition +of the metadata needed for climate data. +The metadata objects could be contained in file formats other than netCDF. +Interconversion of the metadata between files of different formats +will be facilitated if they are based on similar ideas. + +

+This standard is mostly additional to the conventions sponsored by COARDS +(ftp://ftp.unidata.ucar.edu/pub/netcdf/Conventions/COARDS). +In addition, all Unidata recommendations are supported here +unless noted to the contrary. +Comments indicate the places where there are differences +between the standards. +Comments given in emphasised type +and CDL examples given in slanted typewriter type +are not part of the standard. +Note that examples typically show only the details relevant to the +point under discussion, +and hence may be incomplete with respect to the provisions +of the complete standard. + +

+Successful transmission of data +depends upon the receiver of the file having +software that will correctly interpret it. For this reason, a +strategy that is as conservative as possible in the use of attributes +and encoding techniques will best promote portability of data. + +

+This standard also refers to the udunits standard +supported by Unidata. +The udunits package is available via anonymous +ftp at ftp.unidata.ucar.edu. +See section 11 for details of how the package +is used by this convention to define units for physical quantities. + +

+Useful comments and suggestions from +John Sheldon, Jan Polcher, Bryant McAvaney, Harvey Davies, +John Caron, Steve Hankin and contributors to the netCDF news group +have influenced the development of this standard. +We have made some changes to gain greater compatibility with +the NCAR CSM netCDF standard. + +

+ +This document makes occasional use of some mathematical symbols, +for instance ``s'', which should appear as +a lower-case sigma. Under X, this may require you to add the following +to your X resources:
+Netscape*documentFonts.charset*adobe-fontspecific: iso-8859-1 + + + +

+

+2  Filename

+NetCDF files should have the file name extension .nc. + +

+

+3  Data types

+ + + +

+The netCDF data types +char, short, long, float, and double +are all acceptable. +All numeric types are signed. +The byte data type, which is functionally identical to char, +is not recommended because its signedness is ambiguous in netCDF. +The COARDS convention deprecates char, rather than byte. + +

+NetCDF does not support a character string type, +so these have to be represented as char arrays. +In this standard, we refer to them as type ``string''. +A string array must be implemented as a +two-dimensional character data variable, +serving as a vector of fixed-length strings, +the second dimension of its CDL declaration +(leading dimension in terms of Fortran) +being recorded as a dimension in the netCDF file. + +

+

+4  Attributes

+ + + +

+This standard describes many attributes +(some mandatory, others optional), +but a file may also contain non-standard attributes. +Such attributes do not represent a violation of this standard. +Application programs should ignore attributes that they do not recognise +or which are irrelevant for their purposes. +Conventional attribute names should be used wherever applicable. +Non-standard names should be as meaningful as possible. +Before introducing an attribute, consideration should be given to +whether the information would be better represented as a variable. +In general, if a proposed attribute +requires ancillary data to describe it, +is multidimensional, requires any of the defined netCDF +dimensions to index its values, +or requires a significant amount of storage, +a variable should be used instead. +When this standard defines string attributes which make take +various prescribed values, the possible values are given in lower case. +However, applications programs should not be sensitive to case +in these attributes. +Several string attributes are defined by this standard to +contain ``blank-separated lists''. +Consecutive words in such a list are separated by one or more +adjacent spaces. The list may begin and end with any number of spaces. +See Appendix A for a list of attributes described by this standard. + +

+

+5  Global attributes

+ + + +

+The Unidata-standard attribute Conventions +is recommended to reference this standard, containing the +string "GDT 1.4". +This standard is registered with Unidata under the name ``GDT'' +in the directory ftp://ftp.unidata.ucar.edu/pub/netcdf/Conventions +and is available from +http://www-pcmdi.llnl.gov/drach/GDT_convention.html and + +http://www.met-office.gov.uk/sec5/CR_div/GDT_convention.html. + +

+The float attribute appendices is recommended to record +the version number of the appendices to this standard used by the +application which generated the file (see section 12). +This information could perhaps be recorded in the Conventions +attribute, but having a separate attribute +for it will allow applications to extract the information +without having to parse a string. +The string attribute quantity_table should be used to +record the URL of the quantity table (see section 12). +If this attribute is a null string, it is assumed that Appendix D was used, +at the version specified by appendices. + +

+The string attribute comment may be used to record any +extra information about the file. +Additional attributes to describe the file may be included as required. +For example, GCM output might include an attribute +to name the model integration. + +

+Although not mandatory, +the Unidata-standard attribute history is recommended to +record the evolution of the data contained within a netCDF +file. Applications which process netCDF data can append their +information to the history attribute. +The global history attribute is assumed to apply to all +data variables; +individual data variables may have their own history attributes +supplying additional information (see section 12). + +

+Use of the string attributes institution and +production is recommended. +The attribute +institution specifies who produced or supplied the data. +We prefer this name to ``center'' or ``centre'' because the two +possible spellings could cause confusion. +The attribute +production indicates how the data was produced. +If it was model-generated, production should name the model +and its version, as specifically as could be useful. +If it is observational, production should characterise it +e.g. ßurface observation" or "radiosonde". +The global institution and production attributes are taken to +apply to all data variables which do not have their own such attributes +(see section 12). + +

+The calendar attribute (see section 23) may +be recorded as a global attribute. +The global calendar +attribute is interpreted as a default for all time axes. + +

+

+6  Variable names

+ + + +

+Variable names should begin with a letter +and be composed of letters, +digits, and underscores. +Case is significant in netCDF names, but it is recommended that +names should not be distinguished purely by case +i.e. if case is disregarded, no two names should be the same. +It is also recommended that variable names should be obviously +meaningful, if possible, as this renders the file more effectively +self-describing. +However, nothing in this convention relies on the use +of particular names for variables. + +

+

+7  Data variables

+ + + +

+The netCDF variables which contain the +physical data are referred to as ``data variables'', +also referred to as ``primary variables'' by Unidata. +Apart from the general naming rules for variables (above, +section 6), +the names of data variables are not standardised by these +conventions (since files may in general contain multiple +data variables of the same physical quantity). + +

+

+8  Coordinate variables

+ + + +

+A one-dimensional netCDF variable +associated with a dimension of one or more data variables +is called a ``coordinate variable''. +A coordinate variable whose dimension name is identical to +its own name is referred to as a ``main coordinate variable'' +in this standard, +when it is necessary to distinguish it +from other types of coordinate variable +(sections 17, 18, +19 and 20). +Apart from the general naming rules for variables (above, +section 6), +the names of coordinate variables are not standardised by these +conventions (since files may in general contain multiple +coordinate variables of the same orientation). +The values in a main coordinate variable must be +strictly monotonic +(all values are different and either increasing or decreasing) +because this assumption is frequently made by software. + +

+

+9  Axes and dimensionality of a data variable

+ + + +

+A data variable may have any number of dimensions, including zero, +and the dimensions must all have different names. +COARDS strongly recommends limiting the number to four, +but we wish to allow greater flexibility. +The dimensions of the variable define the axes of the quantity it contains. +Dimensions other than those of space and time may be included. +Several examples can be found in this document. +Components of vector or tensor quantities could be contained +in a single data variable by giving the variable a dimension over components. +While there exist advantages for manipulating such a variable in memory, +we see no strong advantage in introducing this complexity into the netCDF +description, and do not recommend it. +Under certain circumstances, one may need more than one +dimension in a particular quantity +(see section 28 concerning multiple time axes). +For instance, a data variable containing a two-dimensional probability +density function might correlate the temperature at two different +vertical levels, +and hence would have temperature on both axes. + +

+If any or all of the +dimensions of a data variable have the interpretations +of ``date or time'' (T), ``height or depth'' (Z), +``latitude'' (Y), or ``longitude'' (X) then those +dimensions should appear in the relative order T, then Z, +then Y, then X in the CDL definition corresponding to the file. +In terms of Fortran, this means X is the first dimension +of the array. Non-spatiotemporal dimensions +should be placed to the left +of the spatiotemporal dimensions +i.e. as trailing dimensions in terms of Fortran. + +

+The reason for this convention is that these kinds of axes may +have special meanings to particular applications. +For instance, an application might want to plot a longitude-latitude map, +or integrate vertically, or extract a timeseries. +In the COARDS standard, the indication given by the order of dimensions +and information in the attributes of the coordinate variables +have to be used together to identify the required axes. +For compatibility with COARDS, we uphold all these conventions, +but we also introduce a new attribute to make the identification +straightforward and unambiguous, as follows. + +

+If the last four dimensions +do not have the interpretations TZYX +(in CDL order, omitting from the left if there are fewer than four dimensions), +an axis attribute should be attached to the data variable. +In other cases it is optional, but recommended. +This attribute is a char array of size equal to the dimensionality of the +data variable, +having one element for each dimension (in CDL order), +indicating the interpretation of that dimension. +The permitted characters are T Z Y and X, +with the meanings given above, +and - as a placeholder for a dimension which has none of these meanings. +Each permitted letter may appear no more than once in the array. +If a data variable has more than one dimension which could be given a certain +interpretation, +the axis attribute will therefore clarify which should be chosen. +Note that if there are multiple time axes (section 28), +of which only one is not collapsed, +this single ``climatological time'' axis would normally be the +designated T-axis. +If the axis attribute is included, +the dimensions may be put in any order, +but this should be avoided if possible +because applications not able to use this attribute +may not process the data correctly. + +

+ +


Axes for an ordinary time-mean longitude-latitude-height +variable: + +

+  dimensions:
+    lat=18;
+    lon=36;
+    pressure=15;
+    con_time=1;
+  variables:
+    float xwind(con_time,pressure,lat,lon); // order T Z Y X
+      xwind:axis="TZYX";
+    float lon(lon);
+    float lat(lat);
+    float pressure(pressure);
+    float con_time(con_time);
+
+See sections 14, 15, 16, +23 +for details of longitude, latitude, vertical and time axes. + +

+The coordinates of points within the data variable are +the simple ordered tuples formed by associating values from the +coordinate variables (section 8). +If a particular axis does not have a coordinate variable, +the coordinate values are assumed to be equal to their indices along +the axis, numbering from 0. + +

+Dimensions may be of any size, including unity. +When a single value of some physical quantity applies to all the values +in a data variable, +the recommended means of attaching this information to the variable +is by use of a singleton dimension (a dimension of size unity) +with a one-element coordinate variable. +The advantage of this method is that all the attributes of a coordinate +variable (quantity, components, boundaries, etc.) can be used to describe +the single-valued quantity. +Singleton dimensions also result from contractions, described +in section 22. + +

+ +


Longitude-latitude field of temperature +on a pressure level: This would use a +singleton pressure dimension to record the level, thus: + +

+  dimensions:
+    lon=96;
+    lat=72;
+    pressure=1; // single-valued coordinate variable
+  variables:
+    float temperature(pressure,lat,lon); // axes in order Z Y X
+      temperature:axis="ZYX";
+    float pressure(pressure);
+      pressure:long_name="pressure";
+      pressure:units="kPa";
+  data:
+    pressure=50.0; // Pressure level of 50 kPa = 500 mbar
+
+The units and long_name attributes +are described in section 12. + +

+ +


Surface air temperature: Surface meteorological +measurements are made at +a certain defined height e.g. 1.5 m, which can be shown thus: + +

+  variables:
+    float temperature(height,lat,lon);
+      temperature:axis="ZYX";
+      temperature:long_name="atmospheric temperature";
+      temperature:units="K";
+    float height(height);
+      height:long_name="height above the surface";
+      height:units="m";
+  data:
+    height=1.5;
+
+The explicit height should not be given if the surface of measurement +is included in the name of the quantity e.g. screen height. + +

+

+10  Coordinate systems

+ + + +

+If the axis attribute indicates X- and Y-axes, +and these are in degrees of longitude and latitude respectively, +these axes constitute a longitude-latitude grid mapped onto the +Earth's surface, and areas of the XY-boxes may be calculated +on this assumption. + +

+A coordinate system for the Earth's surface which is rectilinear +but +based on a polar axis other than the normal geographical axis is referred +to as a ``rotated grid''. +To describe rotated grids, a two-element float attribute +north_pole is attached to the data variable, +specifying the (longitude,latitude) coordinates of the rotated north pole. +If the attribute is absent and relevant, +it is assumed to have the value (0.,90.) i.e. the geographical north pole. + +

+In some systems, +the axes covering the Earth's surface +do not define a rectilinear grid. +We do not wish necessarily to exclude non-rectilinear systems. +For the moment, this standard is undefined for these systems, +and we invite comments from potential users on the appropriate definition. +The COARDS standard excludes non-rectilinear systems. +In principle, any coordinate system can be handled, +albeit clumsily, by replacing the relevant two or more axes by +by a single axis which indexes the points, +and providing associated coordinate variables to specify the coordinates, +point by point (see section 18). + +

+

+11  Units

+ + + +

+The udunits package includes a +file udunits.dat, which lists collections of unit names. The names +given in the most recent version of this file +and their plural forms will be regarded as acceptable +unit names for this standard, with a few modifications which will be +listed in Appendix C to this standard. +COARDS lists some modifications within the standard, +but we would prefer to put in place a means to allow future +modifications to be made easily. +Users of this standard should not define their own units, +because this would make their files less portable; +requests for new units should be directed to Unidata. + +

+The udunits package also defines a means for linear transformation +of units by a scale factor and an offset. +This convention is allowed when it is natural to express a unit in +such a form +e.g. density of sea-water in +kg m-3 +in excess of 1000 +kg m-3, +which can be specified to udunits as "kg m-3 @ 1000". +COARDS does not permit the use of this facility. +This facility should not be used as a means of data compression, +for which an alternative is provided +(see section 32). + +

+

+12  Physical quantity of a variable

+ + + +

+These conventions standardise three string attributes for +specifying +the physical quantity of data and coordinate variables. + +

+The units attribute is formatted as per the +recommendations in the Unidata udunits package +(see section 11), +with extensions for time +(see section 25). +Case is significant in the units. +This attribute is mandatory unless the quantity is dimensionless +(a pure number), +in which case the units may be given as a pure number. +There are a few defined dimensionless units, +such as percent, but +there is no need for a wide variety of dimensionless units for quantities +like sea-ice concentration, cloud fraction, probability and so on; +this descriptive information is the long_name rather +than the units. +A scale factor and/or offset may be specified +quantity e.g. sea-ice concentration in tenths +may be given as units="0.1f". +A dimensionless quantity with no scaling or offset +may have units="1.0f" or units=ünity". + +

+The long_name is a standard Unidata attribute +containing a descriptive name, +which should not specify the units. +This attribute is optional. + +

+The quantity attribute identifies the quantity by +a description chosen from a defined list, +optionally with additional information enclosed in parentheses +() +if sufficient detail cannot be given by a standardised description. +The purpose of defining a list +is to allow users of data from different sources to decide which +quantities are comparable. +Case is not significant in the quantity. + +

+We refer to the list of +possible quantities as the ``quantity table''. +The quantity table defines the quantity and a permissible unit +for each quantity. +Any legal and physically equivalent unit would be an acceptable +units attribute. +If both present, +the quantity and the units attribute must be consistent +according to the quantity table. +There are two options for a choice of quantity table. +One option is to use Appendix D of this standard, +which will be made available on the web. +In this case, the global quantity_table attribute should be +set to a null string. +Each quantity in Appendix D will be labelled with the version of the +appendices at which it was introduced, +enabling an application to deduce the complete set of +quantities which was available to the application which generated the file. +The other option is to produce a list of all +the possible quantity names, +giving for each one an acceptable unit and the name +of an equivalent quantity in Appendix D. +This list should be made available on the web, +and its URL recorded in the global quantity_table attribute. + +

+The use of standardised quantities is optional. +The presence of the global quantity_table attribute implies that +this option is being followed. +The quantity name may optionally be recorded in the long_name +attribute if preferred (to avoid having identical long_name +and quantity attributes). +Therefore if an application which wishes to make use of quantities finds +that the quantity attribute is absent, +it should obtain the quantity from the long_name attribute. +In the remainder of this document, quantity attributes +do not appear in the examples, but the long_name attributes could +serve to supply the quantity information. + +

+ +


Quantity attributes: + +

+    float tempt(pressure,lat,lon);
+      tempt:long_name="potential temperature";
+      tempt:quantity="atmospheric potential temperature "
+        "(after timestep)";
+      tempt:units="K";
+
+"potential temperature" is the description which might be used +as the title of a plot. +ätmospheric potential temperature" is the standard +quantity, and äfter timestep" is additional information, +which a generic application can disregard. + +

+Whether two physical quantities are different or the same is often +not a question with a well-defined answer. +Certainly if they are the same, they must have the same unit, but +various quantities with the same unit may have to be distinguished +e.g. atmospheric potential temperature and soil temperature. +In practice, +the most specific description applicable should be used. +We intend to expand Appendix D on an ongoing basis +in response to requests by users of this standard, +since we cannot foresee all the possibilities, +and we will err on the side of expansion, rather than restriction, +when it is unclear whether a new quantity is needed. + +

+The subgrid attribute (see section 21) +can be regarded as a modifier of the quantity; +it applies only to data variables, not coordinate variables. +The long_name might repeat information which is standardised +by the subgrid attribute. +For instance, the long_name could be +"maximum temperature". +The subgrid attribute will define precisely in what sense +the temperature is maximised. + +

+A data variable may have a history attribute, +supplying information about the derivation of the quantity that is not +possible to include either as a standardised form or in parentheses in +the quantity and subgrid attributes. +This attribute should be used as a last resort. +The global history attribute, if present, also applies to all +data variables (section 5). +A data variable may also have institution and production +attributes, indicating how the data was originally obtained +(see section 5). +These attributes take precedence over the corresponding global attributes. +The attributes history, institution and production +must not be relied upon to distinguish between data variables in the file, +and generic applications may ignore them. + +

+ +


Optional quantity information: +These optional attributes might be used thus to describe a gridded +observational precipitation climatology. + +

+    float precipitation(lat,lon);
+      precipitation:history="gridded using Thiessen polygon weighting";
+      precipitation:institution="Climatic Research Unit, "
+        "University of East Anglia, UK";
+      precipitation:long_name="rate of precipitation";
+      precipitation:production="surface station observation";
+      precipitation:units="mm day-1";
+
+This would be appropriate if the data variable was in a file with data from +other institutions or methods of production. + +

+The optional modulo attribute of a variable, if present, +records a number which can be added or subtracted +without altering the validity or physical significance of +the quantity. +It should be given in the same units as the variable. +This is most likely to be useful for longitude coordinate axes +(section 14), +with a modulo of 360, +and climatological +axes of seasonal or diurnal phase (sections 25 +and 28). + +

+For variables which have a spatial direction with the +interpretation of height or depth, latitude, or longitude +(Z, Y or X in section 9), +the optional positive attribute may be included to indicate +the sense of positive numbers. +The allowed values are up or down for Z, +north or south for Y, east or west for X. +Components of vectors such as velocities and fluxes might have this +attribute. +The standardised quantity names will also indicate the sign convention +for such quantities; for instance, upward velocity rather than +vertical velocity (positive=üp"), +northward heat flux rather than +meridional heat flux (positive="north"). +See also sections 14, 15 and 16. + +

+We note that the Unidata-standard FORTRAN_format attribute +may be useful for both coordinate and data variables. + +

+In addition, +other locally defined attributes may be included to define +the quantity of a variable. +The Hadley Centre model will give each data variable integer stash +and submodel attributes, for example, which are codes identifying +GCM diagnostic output quantities. + +

+Variables may contain quantities which are discrete rather than +continuous, +meaning they can take only certain prescribed values. +This is more likely for coordinate variables than data variables. +For instance, a data variable containing the results of +a Fourier or spherical harmonic analysis might have a dimension +for harmonic number. +Section 25 describes some time variables which are discrete. + +

+

+13  Topology of an axis

+ + + +

+An axis with ``circular topology'' is one +which can be legitimately +transformed by shifting all the points one place along the axis, +moving the last point to the beginning, any number of times. +The main coordinate variable of an axis with circular topology is distinguished +by the presence of an attribute topology="circular". +A longitude axis which circles the whole globe is an example. +The value linear or the absence of this attribute +indicates an axis with ``linear topology''. +The topology is indicated only by the main coordinate variable, +but since it is the property of the axis it applies to any component, +associated or boundary +coordinate variables as well. + +

+When a circular axis is rotated, the main coordinate values must be +altered in order to remain monotonic. +Therefore the main coordinate variable of a circular axis +requires a modulo (section 12). + +

+Note that the topology and modulo attributes convey +different information. +For instance, +a longitude coordinate variable limited to values in the eastern +hemisphere between the Greenwich meridian and the date-line (e.g. 0E, 25E, +120E, 130E, 180E) does not have circular topology. (This might be from a model +of a limited area of the world.) When +making a contour map of a field with such +a longitude axis, one can interpolate anywhere within +the eastern hemisphere to draw the +contours, but it is not legitimate to interpolate over the western hemisphere +and draw the rest of the world, which is simply missing. The +implication of circular topology would be that one could put any longitude +at all on the left-hand side of the map. However, +this coordinate variable does have a modulo (of 360, +as required in section 14), and the points +can be labelled +in any way which is equivalent under the modulo to the coordinates +in the file. +The coordinate values 0,25,120,130,180 are thus equivalent to +-360,-335,-240,-230,-180. + +

+

+14  Longitude dimension

+ + + +

+Coordinate variables representing +longitudes must always explicitly +include the units attribute; there is no default value. +The units attribute will be a +string formatted as per the recommendations in the Unidata udunits +package. +The recommended unit of longitude is degrees_east (eastward +positive). Also acceptable are degree_east, degree_E, and +degrees_E. The unit degrees_west (westward positive) is not +recommended because it implies a negative conversion factor from +degrees_east. +The positive attribute is not required as the sign convention +is implied by the units. + +

+Longitude axes should have the attribute modulo=360, +indicating that they may be interpreted modulo 360. +Thus, for example, +-180, 180, and 540 are all valid representations of the International Dateline +and 0 and 360 are both valid representations of the Prime Meridian. +COARDS assumes that longitudes may always be treated in this way. +Since we have introduced the modulo attribute, +we require that it should be specified to indicate this. +A global longitude axis should have the attribute topology="circular". +Note that the presence of a modulo attribute +does not mean that the axis necessarily has circular topology +(section 13); +a longitude axis covering only part of the globe cannot have its points +rotated. +The sequence of numerical longitude values stored in the netCDF +file must be monotonic in a non-modulo sense +for a main coordinate variable of longitude. + +

+ +


Global longitude axis: + +

+    float lon(lon);
+      lon:long_name="longitude";
+      lon:modulo=360.0f;
+      lon:topology="circular";
+      lon:units="degrees_east";
+
+Note that if a quantity table is in use, +"longitude" is the name of the quantity in the standard list, +and could be recorded instead in a quantity attribute. + +

+

+15  Latitude dimension

+ + + +

+Coordinate variables representing +latitudes must always explicitly +include the units attribute; there is no default value. +The units attribute will be a +string formatted as per the recommendations in the Unidata udunits +package. +The recommended unit of latitude is degrees_north. Also acceptable +are degree_north, degree_N, and degrees_N. +The positive attribute is not required as the sign convention +(northward positive) is implied by the units. + +

+ +


Latitude axis: + +

+    float lat(lat);
+      lat:long_name="latitude";
+      lat:units="degrees_north";
+
+Note that if a quantity table is in use, +"latitude" is the name of the quantity in the standard list, +and could be recorded instead in a quantity attribute. + +

+

+16  Vertical (height or depth) dimension

+ + + +

+Whereas the two horizontal dimensions are usually longitude +and latitude, whose direction is well defined, +a variety of quantities may be used for the vertical axis, +if there is one. +The axis to be regarded as the vertical axis +must have both a long_name attribute +and a positive attribute (section 12), +the latter having one of the values up or down +to indicate the sense of the direction of positive +(The sense may be useful for applications which display the data.) + +

+ +


Vertical pressure axis: + +

+  dimensions:
+    pressure=15;
+  variables:
+    float pressure(pressure);
+      pressure:long_name="pressure";
+      pressure:positive="down";
+      pressure:units="hPa";
+  data:
+    pressure=850, 700, 500, 300, 200, 150, 100, 50, 30, 20, 10;
+
+ +

+The COARDS standard requires the units of +the vertical axis to be selected from a defined list, +in order that this axis can be recognised by its units. +It gives special status to units of pressure, +for which the direction of positive is defined, +and makes the positive attribute mandatory +for vertical axes with other units. + +

+We have adopted a different approach for a number of reasons. +Firstly, to require units for the vertical axis +means defining dimensionless units for any +dimensionless quantity one might wish to use for the coordinate variable. +This is inconsistent with the treatment of a data variable; +the standard does not require that dimensionless units be invented +for dimensionless physical quantities in data variables. +Secondly, the vertical dimension of a data variable can be identified +from the axis attribute or the +the order of dimensions +(see section 9), +allowing any application which expects such a +dimension to find it without any further help. +Thirdly, the quantity attribute (if in use) is more informative +than the units. + +

+We are not entirely convinced that +the direction of positive should really be recorded as part of the +data structure. +It is mostly an issue for displaying the data, +and is to some extent a matter of personal preference. +If such special treatment is given to the vertical axis, +why is it not also recorded for other axes? +For instance, when latitude is shown on the horizontal axis of a plot, +is north on the left or the right? +This is the same kind of question, +but it strikes us as more a matter for a graphics application to consider. +Nonetheless, we have required the positive attribute for +compatibility with COARDS. + +

+For example, if an oceanographic netCDF file encodes the depth of the +surface as 0 and the depth of 1000 m as 1000 then the axis would +use attributes as follows: units="m", +long_name="depth below the surface", positive="down". +If, on the other hand, the depth of 1000 m were represented as +-1000, +we would have long_name="height above the surface", +positive=üp". + +

+

+17  Component variables

+ + + +

+A continuous physical variable may require +more than one number to specify it at each point. +We refer to these as ``components''. +The values of the components +are recorded in variables +referred to as ``component variables''. +The variable to which the components belong is called the ``head'' +variable of the components. +The names of the component variables are recorded as +a blank-separated list in a component string attribute +of the head variable. +The dimensions of a component variable must be identical with those +of its head variable. +OGDT restricted components to coordinate variables, +but the concept has here been generalised, +since any quantity that is used as a coordinate variable might also +be needed as a data variable. + +

+When a coordinate variable has components, +this standard requires that a main coordinate variable should nonetheless +be supplied which represents a combination of the components that can be +used to order the points on the axis. +As usual, this main coordinate variable must be monotonic, +but the components do not need to be monotonic. +The definition of the main coordinate in terms of its +components may be given in parentheses in the component attribute. +This information is not standardised and generic applications cannot be +expected to make use of it. + +

+ +


Hybrid vertical coordinate: +A vertical coordinate +h � p/p0 + s +is used in some atmospheric GCMs. +Atmospheric model levels are specified in terms of +(p,s) +pairs, +where p is pressure, +p0 +is a constant and +s +is fraction +of surface pressure (which is variable). +The +h +value is a linear combination of the two, +which cannot be uniquely decomposed back into +(p,s). +We would record this coordinate variable thus: + +

+    float eta(eta); // main coordinate variable
+      eta:component="pressure sigma "
+        "(eta=pressure/p0+sigma; p0=100 kPa)";
+    float pressure(eta);
+    float sigma(eta);
+
+A generic application would treat the component and main coordinates +as independent information. +The extra knowledge required to relate them would reside in any +specific application which needed it. +Hybrid vertical coordinates are the only obvious application of +component variables, but the convention could be used for other +similar purposes if they arise. + +

+

+18  Associated variables

+ + + +

+An axis of a data variable, or two or more axes in combination, +may have alternative sets of coordinate values. +These alternative sets are recorded in variables +referred to as ``associated'' variables, +having their own units, long_name +and other appropriate attributes to describe them. +The names of the associated variables are recorded as a blank-separated list +in an associate string attribute either of the data variable +or of the main coordinate variable of the axis concerned. +If the association is with the data variable, it applies only for that +data variable, +but if it is with the main coordinate variable, +it applies for any data variable which uses that main coordinate variable. +Association with the main coordinate variable thus may be more convenient, +but is less flexible. +Association with the data variable is the only option when several axes are +involved, and when there is no main coordinate variable. +The examples illustrate these points. + +

+The associate +attribute may alternatively and equivalently be named coordinates. +This possibility is included for compatibility with the CSM standard. +However, in the present standard it is deprecated because of possible confusion +with the normal definition of ``coordinate variable'' (section +8) and because the use of associated variables is wider +than just coordinate variables in the usual sense. + +

+A variable may be associated with more than one data variable +or coordinate variable. +If an associated variable itself has an associate attribute, +variables named by this attribute are also regarded as being associated. + +

+An associated variable must have dimensions which are all dimensions +of any data variable with which it is associated; +the associated variable can be regarded as a function +of the indices along these axes. +The values of an associated variable do not have to be monotonic. + +

+A generic application is not required to make any +use of associated variables. +Associated variables are not indicated in the axis attribute of the +data variable (section 9). +However, to improve legibility of the CDL file, +it is recommended that when +variables named by the associate attribute +of a data variable have interpretations that would be +denoted by T Z Y or X in the axis attribute, +they are listed in that order, with any others appearing in front. + +

+ +


Vertical axis: Many associated variables +will be one-dimensional, +giving alternative sets of values for a single axis. +One example is a vertical axis where one wishes to store both the +physical coordinate and the ordinal model level number: + +

+  dimensions:
+    lat=90;
+    sigma=19;
+  variables:
+    float xwind(sigma,lat); // 2D data variable
+      xwind:axis="ZY";
+    float lat(lat);
+      lat:long_name="latitude";
+      lat:units="degrees_north";
+    float sigma(sigma); // physical height coordinate
+      sigma:associate="model_level";
+      sigma:long_name="sigma";
+      sigma:positive="down";
+    int model_level(sigma); // model level number at each height
+      model_level:long_name="model level number";
+      model_level:positive="up";
+
+As the association of model_level is with sigma, +any data variable with a sigma-axis +will have the association with model_level; +it is not a property of xwind particularly. + +

+ +


Trajectory: The +value of a quantity along a one-dimensional +trajectory. +In such a case, we might have a coordinate variable containing +time of travel and +associated coordinate variables +giving the latitude and longitude of each point: + +

+  dimensions:
+    day=10; // 10 sample times along a trajectory
+  variables:
+    float hice(day); // sea-ice thickness measured as the floe drifts
+      hice:associate="lat lon";
+      hice:axis="T";
+      hice:units="m";
+    float day(day); // time since the beginning of the journey
+      day:long_name="time";
+      day:units="day";
+    float lon(day); // longitude at each time
+      lon:long_name="longitude";
+      lon:units="degrees_east";
+    float lat(day); // latitude at each time
+      lat:long_name="latitude";
+      lat:units="degrees_north";
+
+The main coordinate variable (day) must be +monotonic, but the associated coordinates are not necessarily. +An important application +rather similar to this one +is described in section 19. +Note that lat and lon cannot be +indicated as the X and Y coordinates in the axis +attribute (section 9). +This is reasonable because, even though they have the interpretations +of latitude and longitude, they are not independent dimensions in +the sense normally expected by an application which might look for +such axes. + +

+Since lon and lat are associated with the data variable +hice, other variables with a day-axis will not share these +associations. +If they are wanted, they would have to be indicated on those data +variables as well. +This approach allows the possibility that the same day +variable might occur in combination with various different sets of +associated coordinate variables. +For instance, there might be more than one trajectory, with the +same day coordinates but different longitude-latitude positions. + +

+ +


Transformed coordinates: Associated variables +of more than one dimension can be used to +describe alternative coordinate systems. +For example, vertical profiles of atmospheric humidity might +be available on a regular longitude-latitude grid, +but we might also wish to give the national grid coordinates of +each point. +The national grid x- and y-coordinates are each functions of +both latitude and longitude; +the x-coordinate does not correspond specifically to longitude, +nor the y- to latitude. +The appropriate representation is therefore: + +

+  dimensions:
+    lon=10;
+    lat=20;
+    pressure=15;
+  variables:
+    float humidity(pressure,lat,lon);
+      humidity:associate="y x";
+    float pressure(pressure);
+      pressure:long_name="pressure";
+      pressure:positive="down";
+      pressure:units="kPa";
+    float lon(lon); // 1D main coordinate variable
+      lon:long_name="longitude";
+      lon:modulo=360.0f;
+      lon:units="degrees_east";
+    float lat(lat);
+      lat:long_name="latitude";
+      lat:units="degrees_north";
+    float x(lat,lon); // 2D associated coordinate variable
+      x:long_name="UK national grid eastings";
+    float y(lat,lon);
+      y:long_name="UK national grid northings";
+
+This tells us that humidity[*][10][5] is the vertical profile +of humidity at the point with latitude lat[10] and +longitude lon[5], +which is at national grid x-coordinate x[10][5] and y-coordinate +y[10][5]. +Because the associated variables are multidimensional, they do not +correspond one-to-one with the axes, +so the association must be with the data variable rather than the +main coordinate variables. + +

+ +


No main coordinate variables: +A related situation is when the 2D grid is staggered or transformed +geometrically in some way (other than rotation-see section 10), +so that it is not possible or easy to give +one-dimensional coordinate variables for the axes. +In this case, there would be no main coordinate variables, +and the default of plain indexes would apply. +The physical coordinates are functions of the 2D gridpoint indices, +and would be given in associated variables of the data variable +just as above: + +

+  dimensions:
+    x=90;
+    y=45;
+  variables:
+    float orog(y,x); // 2D variable on a horizontal grid
+      orog:associate="lat lon";
+      orog:axis="--";
+      orog:long_name="height of the surface above sea-level";
+      orog:units="m";
+    float lon(y,x); // 2D coordinate variable on the same grid
+      lon:long_name="longitude";
+      lon:modulo=360.0f;
+      lon:units="degrees_east";
+    float lat(y,x);
+      lat:long_name="latitude";
+      lat:units="degrees_north";
+
+The lat and lon variables are not indicated as +the X and Y coordinates in the axis variable +(section 9). +An application which referred to latitude and longitude coordinates +would not generally expect them to be two-dimensional. +If it can handle this situation, it should identify these axes by +their long_name and units. + +

+ +


3D associated coordinates: +These could be used if one wished to describe +a field with alternative 3D coordinate systems, +for instance on both a regular Cartesian grid, +and in cylindrical or spherical coordinates. +The values of the alternative coordinates would be given on the Cartesian +grid. In the spherical case, for instance + +

+    float temperature(z,y,x); // 3D variable on a Cartesian grid
+      temperature:associate="radius theta phi";
+    float radius(z,y,x);
+    float theta(z,y,x);
+    float phi(z,y,x);
+
+ +

+A particular technical application of a one-dimensional +associated coordinate is to deal with the limitation of netCDF to a +single unlimited dimension. +If several data variables have unlimited axes of different lengths or +physical significance, +they can all share a nominal unlimited dimension, +and each have associated variables specifying the meaning of the axis. + +

+ +


More than one unlimited axis: +Consider a file which contains data variables with unlimited +axes measuring elapsed time with different sampling frequency, +and hence of different lengths. + +

+  dimensions:
+    time_counter=UNLIMITED;
+  variables:
+    float sw(time_counter); // sampled every 3 hours
+      sw:associate="time_3h";
+      sw:axis="T";
+      sw:long_name="vertical component of "
+        "shortwave radiative flux density";
+      sw:units="W m-2";
+    float latent(time_counter); // sampled every 30 minutes
+      latent:associate="time_30min";
+      latent:axis="T";
+      latent:long_name="latent heat flux density";
+      latent:units="W m-2";
+    float time_3h(time_counter)";
+      time_3h:long_name="elapsed time";
+      time_3h:units="h";
+    float time_30min(time_counter);
+      time_30min:long_name="elapsed time";
+      time_30min:units="min";
+
+ +

+

+19  Bundles

+ + + +

+If several data arrays containing the same physical quantity +have one or more identical axes, +but are distinguished by the values of other singleton coordinate variables, +it may be convenient to store them in the same data variable. +The common axes of the separate arrays become axes of the combined variable. +One or more additional axes are introduced to ``bundle up'' the separate +arrays. +Such an axis does not correspond to a continuous physical coordinate. +It acts simply as an index of the bundled-up arrays. + +

+The singleton values of the separate arrays are recorded in +associated coordinate variables for the bundling dimension. +They should not be interpreted as continuous coordinates. + +

+ +


Timeseries: +The Hadley Centre GCM +can generate timeseries of the values of quantities +at individual points. +Typically, timeseries from many different points are produced of the same +quantity at the same sampling times. +It is natural to contain this information +in a data variable with two dimensions. +One dimension is the common time axis, specifying the sampling times, +which are the same for all the points sampled. +The other dimension is not a continuous physical coordinate; +it is simply being used to ``bundle up'' the timeseries, +the points being +irregularly scattered in a space of two or more dimensions. +Thus: + +

+  dimensions:
+    points=15; // measurement locations
+    times=20; // sampling times
+  variables:
+    float snowdepth(times,points);
+      snowdepth:associate="sitename lat lon";
+      snowdepth:axis="T-";
+    float lon(points); // longitude of sites
+      lon:long_name="longitude";
+      lon:modulo=360.0f;
+      lon:units="degrees_east";
+    float lat(points); // latitude of sites
+      lat:long_name="latitude";
+      lat:units="degrees_north";
+    char sitename(points,StringMaxLength); // string array of sitenames
+    double times(times); // times of measurement
+
+See section 23 concerning the time coordinates. +This same form could be used for observed timeseries from stations. +The bundling axis (points) is simply an index. +The long_name and units of the associated coordinates +identify their meanings. + +

+ +


Vertical profiles: +A similar application is that of vertical +profiles at sets of points; +for example, scattered vertical temperature profiles through the ocean, +or data from various radiosonde stations. + +

+  dimensions:
+    station=10; // measurement locations
+    pressure=11; // pressure levels
+  variables:
+    float humidity(pressure,station);
+      humidity:associate="lat lon";
+      humidity:axis="Z-";
+    int station(station); // station numbers
+    float lon(station); // longitude of stations
+    float lat(station); // latitude of stations
+    float pressure(pressure)
+
+ +

+ +


Several parcel trajectories: +Consider a set of Lagrangian parcel trajectories (for instance, ocean +drifters). +Various parameters are evaluated at fixed times from the start of each +trajectory. +The trajectories are identified by their place of origin, +and the position is a function of time and trajectory identity. +The position information is therefore stored in multidimensional +associated coordinate variables. + +

+  dimensions:
+    parcel=15; // number of trajectories
+    times=20;
+    max_len_parcel_name=64; // max length of trajectory name
+  variables:
+    float temperature(parcel,times);
+      temperature:associate="parcel_name lat lon";
+      temperature:axis="-T";
+    float salinity(parcel,times);
+      salinity:associate="parcel_name lat lon";
+      salinity:axis="-T";
+    float times(times);
+      times:units="days";
+    char parcel_name(parcel,max_len_parcel_name);
+    float lon(parcel,times);
+    float lat(parcel,times);
+
+The associations have to be made on the data variables in this case, +because parcel_name does not have a main coordinate variable, +while lon and lat are multidimensional. + +

+This section raises the question of how best to store a single timeseries, +or a single vertical profile. +Following the scheme of this section, +it could be contained in a two-dimensional data variable with the bundling +axis being of size unity. +The associated information such as latitude or longitude would then be +stored in singleton coordinate variables, +all associated with the same dimension. +Alternatively, these values could be recorded as +separate singleton dimensions (following section 9). +We have no recommendation for this. +Either scheme could be appropriate; +which is more natural perhaps depends on how the data was extracted from +the continuous axes. + +

+

+20  Boundary variables

+ + + +

+Along a dimension, +the values might relate to points (at the coordinate values) +or to contiguous or non-contiguous cells. +The boundaries of the cells should be defined +as well as the point coordinate values. +The convention is to define an additional two-dimensional +``boundary variable'' with +a right-hand dimension (leading dimension in Fortran terms) of size two. +The values for which this dimension has index 0 +(numbering from 0 i.e. in C notation) supply the boundaries +with the smaller main coordinate values, +and those with index 1 the large values, +where ``smaller'' and ``larger'' refer simply to numerical comparison, +not to a physical direction. +Supplying upper and lower boundaries separately allows for the possibility +that the cells might not be contiguous; they might even overlap. +If a lower boundary value is equal to the valid_min for the coordinate +variable (section 29), the cell has no lower boundary. +If an upper boundary value is equal to the valid_max, +the cell has no upper boundary. +The name of the boundary variable is recorded in a string attribute +bounds of the main coordinate variable. +We recommend that the it should be named +by the coordinate dimension with the prefix bounds_. +The boundary variable should not have a units attribute; +its units are the same as those of the main coordinate variable. + +

+ +


Boundaries for a one-dimensional latitude coordinate variable: + +

+    float lat(lat);
+      lat:bounds="bounds_lat";
+    float bounds_lat(lat,2);
+
+In C notation, +lat[0] gives the coordinate of the first point, +bounds_lat[0][0] its lower boundary, bounds_lat[0][1] +its upper boundary. +In Fortran notation, the declarations are lat(lat) +and bounds_lat(2,lat), +and the relevant elements are lat(1), bounds_lat(1,1), +bounds_lat(2,1). + +

+ +


Albedo as a function of wavelength and snow cover: +Characteristic values of albedo are given for various wavelength +bands, dependent also on snowdepth. + +

+  dimensions:
+    lambda=4; // number of shortwave frequency bands
+    snowdepth=10; // number of snowdepth categories
+  variables:
+    float albedo(lambda,snowdepth); // no units for albedo
+      albedo:axis="--";
+      albedo:long_name="surface albedo";
+    float lambda(lambda);
+      lambda:bounds="bounds_lambda";
+      lambda:long_name="wavelength";
+      lambda:units="nm";
+    float bounds_lambda(lambda,2);
+    float snowdepth(snowdepth);
+      snowdepth:bounds="bounds_snowdepth";
+      snowdepth:long_name="mass per unit area of lying snow";
+      snowdepth:units="kg m-2";
+      snowdepth:valid_max=1e9;
+    float bounds_snowdepth(snowdepth,2);
+  data:
+    lambda=250, 385, 570, 795;
+    bounds_lambda=175,320, 320,450, 450,690, 690,900;
+    snowdepth=0.05, 0.15, 0.35, 0.75, 1.25, 1.75, ..., 450.0, 1000.0;
+    bounds_snowdepth=0.0,0.1, 0.1,0.2, 0.2,0.5, 0.5,1.0,
+                     1.0,1.5, 1.5,2.0, ..., 400.0,500.0, 500.0,1e9;
+
+A first index of 0, for instance, gives albedo values for +the wavelength range 175-320 nm. +The deepest snowdepth class has no upper bound; any value above 500 +falls into this class. + +

+In some cases such as the previous example the boundary +coordinates are well defined, +but the gridpoint coordinates are arbitrary. +In such circumstances, this standard recommends the mid-point of the boundaries +be used as the gridpoint. +Two advantages of this choice are: firstly, comparison of a gridpoint with a +boundary will always decide which cell the point belongs in; +secondly, it will probably be an appropriate choice for plotting and calculations +involving gridpoints such as differentiation. +However, as shown by the last snowdepth cell of the above example, +which is unbounded upwards, the mid-point is not always a sensible choice. + +

+ +


Probability density function of precipitation amounts: + +

+  dimensions:
+    ppn=10;
+  variables:
+    float pdf(ppn,lat,lon);
+      pdf:axis="-YX";
+      pdf:long_name="probability density of "
+        "depth of water-equivalent precipitation";
+      pdf:units="mm-1";
+    float ppn(ppn);
+      ppn:units="mm";
+      ppn:long_name="depth of water-equivalent precipitation";
+      ppn:bounds="bounds_ppn";
+    float bounds_ppn(ppn,2);
+  data:
+    bounds_ppn=0.0,0.1, 0.1,0.2, 0.2,0.35, 0.35,0.5, 0.5,1.0, ...;
+
+pdf[3][10][12] gives the probability density of precipitation +amounts between 0.35 and 0.5 mm falling at the location +lat[10] lon[12]. + +

+Boundary variables +are recommended if the main coordinate values +are not evenly spaced, +or if the dimension has a size of unity. +If the coordinates are evenly spaced, +and boundaries are not specified, +generic applications may assume that the main coordinates +lie at the centres of their cells. +Boundary variables may be supplied for component and associated coordinate +variables as well as for main coordinate variables. +Their elements are ordered so as to correspond +to the corresponding main boundary variables. +Hence they will not necessarily be monotonic, +and index 0 and 1 (numbering from 0) of the dimension of size two +will not necessarily contain the smaller and larger values. + +

+ +


Boundary values for a hybrid vertical coordinate: +The atmospheric column is here divided into three cells in the vertical; +from the surface to +s = 0.7; +from there to 20 kPa, and finally to the top of the atmosphere, +using the hybrid vertical coordinate introduced in an example +in section 17. + +

+  dimensions:
+    eta=3;
+  variables:
+    float(eta);
+      eta:long_name="pressure-sigma hybrid";
+      eta:component="pressure sigma";
+      eta:bounds="bounds_eta";
+      eta:positive="down";
+    float bounds_eta(eta,2);
+    float pressure(eta);
+      pressure:units="kPa";
+      pressure:long_name="pressure";
+      pressure:bounds="bounds_pressure";
+    float bounds_pressure(eta,2);
+    float sigma(eta);
+      sigma:long_name="sigma";
+      sigma:bounds="bounds_sigma";
+    float bounds_sigma(eta,2);
+  data:
+    eta=0.75, 0.45, 0.05;
+    bounds_eta=0.7,1.0, 0.3,0.7, 0.0,0.3;
+    pressure=0.0, 10.0, 5.0;  // does not need to be monotonic
+    bounds_pressure=0.0,0.0, 20.0,0.0, 0.0,20.0; // note order
+    sigma=0.75, 0.35, 0.0;
+    bounds_sigma=0.7,1.0, 0.1,0.7, 0.0,0.1;
+
+bounds_pressure[1][0] exceeds bounds_pressure[1][1] +because they are ordered to +correspond to bounds_eta. + +

+Boundary variables may be given for associated multidimensional +coordinate variables (section 18). +Each dimension of the main variable requires an extra dimension of size 2 +in the boundary variable. +These extra dimensions are placed on the right +(left in Fortran terms) of the coordinate dimensions, +and in the same order as the coordinate dimensions. + +

+ +


Boundaries for a two-dimensional latitude coordinate variable: + +

+    float lat(y,x);
+      lat:bounds="bounds_lat";
+    float bounds_lat(y,x,2,2);
+
+so bounds[4][5][0][0] contains the latitude of +the lower left (smaller x and y) +corner of gridbox [4][5], bounds[4][5][0][1] the lower right corner, +bounds[4][5][1][0] upper left and bounds[4][5][1][1] upper right. +In Fortran, the indices of the boxes would be (1,1,6,5), (2,1,6,5), +(1,2,6,5), (2,2,6,5) respectively. + +

+

+21  Representation of subgrid variation

+ + + +

+Since a data variable usually +represents a physical quantity which varies +continuously along the axes, +in reality there will generally be variation of the quantity between +adjacent gridpoints. +The data variable can give only one value for each cell, +despite this subgrid variation. +For many purposes, this can be taken as a ``representative'' value, +and it is not necessary to define precisely how it relates to +the subgrid variation. + +

+To be explicit about how each data value reflects subgrid variation +along a particular axis, +use the subgrid attribute of the data variable. +The most important application of this attribute is to +contracted or collapsed axes, described in section 22. +This is a string attribute comprising a list of blank-separated words. +In this list, "name: method" +indicates that subgrid variation along the axis with the dimension whose +``name'' is given +is represented by the specified ``method''. +The method, which may be several words, should be one of the +permitted values detailed in Appendix B, +which include mean, maximum, minimum, +standard deviation, variance, +median, cell, point. +Case and punctuation are not significant in the method. +Like Appendix D, Appendix B will be expanded on request by +users of this standard. +Some methods imply a change of units of the data variable, +and this also is specified by Appendix B. +In the above list, this is true for variance. +The method point indicates that the data values apply exactly +at the coordinate values, and do not at all represent the variation between +adjacent gridpoints along the axis concerned. +The method cell indicates that each value +should be regarded as a property of the whole cell +along the axis concerned e.g. a sum or integral. +The method can be differently specified for the various dimensions. +It must be remembered that the method applies only to the axis indicated. +If a precipitation value in a longitude-latitude gridbox is given the +method maximum for these axes, for instance, it means that it is +the maximum within these spatial cells, and does not imply that +it is also the maximum in time. + +

+The absence of any specification means that generic applications may +regard the data values as representative in whatever way suits their needs. +For quantities calculated at gridpoints by numerical models, +this kind of vagueness is unavoidable. +If a model provides a longitude-latitude field of temperature at gridpoints, +an application used to draw a contour plot of the field will generally +assume that the temperatures apply at points, and will use some interpolation +scheme to compute values between them. +An application which calculates the mean of the field, however, will +probably assume that the temperatures are gridbox means, +and average them by weighting each with its area. +Both of these approaches are valid. +The finite-difference scheme by definition does not have any information +about subgrid variation, +and may itself treat the values in both ways; +it might calculate gradients between them, regarding them as points, +or enforce conservation properties, regarding them as means. +It would be unusual to regard the values as extrema, however, unless +this was explicitly indicated. + +

+Except when the data are point values, +it may also be unclear what the coordinates of the gridpoints should be. +For example, +what time coordinate should be assigned to a time mean value? +In such circumstances, if the boundaries of the cell are well defined, +this standard recommends that the gridpoints should be +defined as the mid-point between the boundaries +(see section 20). + +

+ +


Subgrid time variation in timeseries: +Consider 12-hourly timeseries of pressure, temperature +and precipitation from a number of stations, +where pressure is measured instantaneously, temperature extremes over +the preceding period are recorded by maximum and minimum thermometers, +and precipitation is accumulated in a rain gauge. +For a period of 48 hours from 6 a.m. on 19th April 1998, +the data is structured as follows: + +

+  dimensions:
+    instanttime=5; // 5 instantaneous measurements at 12-hour intervals
+    periodtime=4; // 4 intervening 12-hour periods
+    station=10;
+  variables:
+    float pressure(station,instanttime);
+      pressure:axis="-T";
+      pressure:long_name="pressure";
+      pressure:subgrid="instanttime: point";
+      pressure:units="kPa";
+    float maxtemp(station,periodtime);
+      maxtemp:axis="-T";
+      maxtemp:long_name="temperature";
+      maxtemp:subgrid="periodtime: maximum";
+      maxtemp:units="K";
+    float ppn(station,periodtime);
+      ppn:axis="-T";
+      ppn:long_name="depth of water-equivalent precipitation";
+      ppn:subgrid="periodtime: cell";
+      ppn:units="mm";
+    double instanttime(instanttime);
+      instanttime:long_name="time";
+      instanttime:units="h since 1998-19-4 6:0:0";
+    double periodtime(periodtime);
+      periodtime:bounds="bounds_periodtime";
+      periodtime:long_name="time";
+      periodtime:units="h since 1998-19-4 6:0:0";
+    double bounds_periodtime(periodtime,2);
+  data:
+    instanttime=0., 12., 24., 36., 48.;
+    periodtime=6., 18., 30., 42.;
+    bounds_periodtime=0.,12., 12.,24., 24.,36., 36.,48.;
+
+It is not appropriate give a subgrid method for the station axis, +since this is a bundling axis (section 19) and not +a continuous physical coordinate. +The instantaneous and period measurements have to have different time +axes both because of their different dimension, +and because they do not coincide. +If the pressure measurements were made at times half-way between the +others (noon and midnight), the time axes could be shared. +Since the precipitation is given as an amount, it is a sum over +the interval of time by definition. +It could instead have been expressed as a rate in mm h-1, for +instance, in which case its subgrid method would be mean +rather than cell. + +

+ +


Thickness (geopotential difference): The +``thickness'' is the difference +in geopotential height between two pressure surfaces in the +atmosphere. This quantity is by definition one which relates +to the whole extent of its cell in the vertical dimension. + +

+  variables:
+    float thickness(pressure,lat,lon);
+      thickness:long_name="thickness";
+      thickness:subgrid="pressure: cell";
+      thickness:units="m2 s-2";
+    float pressure(pressure);
+      pressure:bounds="bounds_pressure";
+      pressure:long_name="pressure";
+      pressure:units="hPa";
+    float bounds_pressure(pressure,2);
+
+Here, bounds_pressure[0][0] and bounds_pressure[0][1] +will be the upper and lower pressure bounds of the thickness field +thickness[0][*][*]. + +

+If more than one subgrid method is to be indicated, +they should be arranged in the order they were applied. +The left-most operation is assumed to have been applied first. +Suppose a quantity varies in both longitude and time +(dimensions lon and time) within each +gridbox. +Values which represent the time-average of the zonal maximum +are labelled subgrid="lon: maximum time: mean", +i.e. find the largest value at each instant of time over all longitudes, +then average these maxima over time; +values of the zonal maximum of time-averages are +labelled subgrid="time: mean lon: maximum". +If the methods could have been applied in any order without affecting +the outcome, they may be put in any order in the subgrid attribute. + +

+If a data value is representative of variation over a +combination of axes, +a single method should be prefixed by the names of all the dimensions involved, +whose order is immaterial. +Dimensions should be grouped in this way only if there is an essential difference +from treating them individually. +For instance, the subgrid standard deviation of +topographic height within a longitude-latitude gridbox +would have subgrid="lat: lon: standard deviation". +This is not the same as +subgrid="lon: standard deviation lat: standard deviation", +which would mean finding the standard deviation along each parallel of +latitude within the zonal extent of the gridbox, +and then the standard deviation of these values over latitude. + +

+To indicate more precisely how the subgrid method was +applied, +extra information may be included in parentheses () after the +identification of the method. +This information is not standardised and may be ignored by a generic +application. +A mean over latitude, for instance, may be area-weighted. +This could be indicated as "lat: mean (area-weighted)". + +

+The subgrid attribute cannot be used +to show how a value reflects variation over a coordinate which does +not have a dimension in the data variable. +This should be done in the long_name instead. +It is generally more informative and precise to introduce a singleton dimension +specifically for this purpose, however. +For example, we could describe a quantity in its long_name as +being simply a temporal variance, +but it would be more informative to record it as a subgrid +method, by giving the variable +a singleton time dimension, which could also +be used to should the range of times +it covers and the time-interval of the data from which the variance was +calculated. See also section 22. + +

+

+22  Contracted dimensions

+ + + +

+A contracted axis is one which is formed by aggregating +the values of an axis with a larger dimension into a smaller number of groups. +In the commonest case, the dimension is collapsed completely to +a singleton dimension (i.e. a size of unity, section 9), +where all data points share the entire collapsed axis. +The collapsed dimension +indicates the relationship of the data variable which is +being described to another variable of higher dimensionality. +The boundaries of the cells along a contracted axis +will be the outside boundaries of the groups of cells along the +uncontracted axis, +or the outside coordinates if boundaries were not given. +The main coordinate values of a contracted axis will be values representative +of the coordinate ranges spanned by the groups. +A collapsed dimension has a single representative main +coordinate value and boundary coordinate values supplying the +complete range of the uncollapsed axis. +These boundaries will be +the extreme boundary coordinate values of the uncollapsed axis, +or the extreme main coordinate values if boundaries were not supplied. +A very important application of collapsed axes is to indicate +climatological time. +This is discussed in section 28. + +

+The subgrid attribute +(section 21) of the data variable +with contracted axes +can be used to indicate how the data values +of the variable with uncontracted axes +were aggregated to reduce the dimensions. +The new subgrid information will be appended to the existing attribute, +if any, +indicating the name of the newly contracted dimension. +Any existing references to the uncontracted dimension in the +subgrid attribute should be modified to refer to the contracted +dimension, since the uncontracted dimension will no longer be a dimension +of the data variable. + +

+As explained in section 21, this attribute will indicate +that the data value is the mean, maximum, minimum, etc. +The allowed subgrid ``methods'' +are listed in Appendix B, which will be expanded as need arises. +As foreseen at the moment, the idea is limited to operations which give +a single value representative of each contracted group of values, +without reference to any external constants. +For example, the number which exceeds 20% of the values in the group, +or equivalently the 20th percentile, is a single number +representing the group, but the +procedure of finding it is not treated as a subgrid method +because it requires the constant 0.2 to define it. +Instead, the relationship of this new variable to the old should be shown +by changing its long_name to indicate that it is a percentile value, +and giving it a new singleton percentage axis with value 20, +or cumulative probability with value 0.2. +This kind of transformation is analogous to reducing a variable on three +spatial dimensions (say) to two by extracting its values on a specified +surface. +The contraction or collapse is a special case, because, +in general, the percentile axis need not have a size of unity; +it might be a new multi-valued axis (in cumulative probability) +replacing the old one (in some spatial dimension, for instance). +This is like regridding a vertical axis of height onto pressure. +Having said all this, however, we note that median is in fact a named +instance of this operation-extraction of the 50th percentile-but we +allow it on the grounds that it is a common method +for choosing a single representative value. + +

+Singleton axes are not necessarily the result of collapsing an axis. +In section 9, we recommend singleton axes as the means +of attaching characteristic single physical values to a data variable, +for instance the height or pressure of the surface on which a variable +is supplied. +If no subgrid method is specified, the application knows only +that the single value characterises the data in some way. +All information in the subgrid attribute is entirely optional. +For instance, a time-mean quantity should generally have a singleton time +dimension +to indicate the range of times to which it applies, but +it is not mandatory to indicate in the subgrid attribute that it is +a mean over time. + +

+On the coordinate variable of a contracted axis, +the optional old_interval attribute +specifies the typical spacing between +two adjacent coordinates of the uncontracted axis, where +``typical'' is not well defined. +The old_interval attribute should be given in the same units as the +coordinates. +Further information may be given by the +the optional old_spacing attribute, +which may have value uniform, indicating that the coordinates were +evenly spaced with the old_interval specified (if any) and the cells +contiguous, +or variable, if they were not evenly spaced but still contiguous, +or disjoint, which means there may have been gaps between them. +The coordinates of the uncontracted axis may be explicitly recorded +in separate variables; +if so, the main uncontracted coordinate variable +should be named by the attribute expand of the +main contracted coordinate variable. + +

+ +


Area-averaging a longitude-latitude field to one of +lower resolution: The original resolution was 1 degree, and the field +has been averaged into 10-degree boxes. + +

+  dimensions:
+    con_lat=18; // contracted dimension
+    con_lon=36;
+    lat=180; // original uncontracted dimension
+    lon=360;
+  variables:
+    float sst(con_lat,con_lon);
+      sst:axis="YX";
+      sst:long_name="sea surface temperature";
+      sst:subgrid="con_lat: mean con_lon: mean";
+      sst:units="degC";
+    float con_lat(con_lat); // contracted latitude axis
+      con_lat:bounds="bounds_con_lat";
+      con_lat:expand="lat";
+      con_lat:old_interval=1.0f; // original resolution in latitude
+      con_lat:long_name="latitude";
+      con_lat:units="degree_north";
+    float bounds_con_lat(con_lat,2);
+    float lat(lat); // original uncontracted latitude axis
+      lat:bounds="bounds_lat";
+    float bounds_lat(lat,2);
+  data:
+    con_lat=-85, -75, -65, ...;
+    bounds_con_lat=-90.,-80., -80.,-70., -70.,-60., ..., 80.,90.;
+    lat=-89.5, -88.5, -87.5, ...;
+    bounds_lat=-90,-89, -89,-88, -88,-87, ..., 89,90;
+
+Instead of an area-average, the contracted field might instead have +represented the subgrid spatial variation of SST. +In that case, subgrid="con_lat: con_lon: standard deviation". + +

+ +


Mean over time and longitude: Here, the time-mean +zonal-mean humidity is given as a function of latitude and height. +The means have been formed over the complete time and longitude +intervals of the original data, so these dimensions are collapsed. + +

+  dimensions:
+    con_lon=1; // collapsed longitude dimension
+    con_time=1; // collapsed time dimension
+    lon=72;
+    sigma=6;
+  variables:
+    float humidity(con_time,sigma,lat,con_lon);
+      humidity:axis="TZYX";
+      humidity:long_name="specific humidity";
+      humidity:subgrid="con_time: mean con_lon: mean";
+    double con_time(con_time);
+      con_time:bounds="bounds_con_time";
+      con_time:old_interval=0.125; // originally at intervals of 3 h
+      con_time:units="days as %Y%m%d.%f";
+    float bounds_con_time(con_time,2);
+    float con_lon(con_lon);
+      con_lon:bounds="bounds_con_lon";
+      con_lon:long_name="longitude";
+      con_lon:modulo=360f;
+      con_lon:topology="circular";
+      con_lon:units="degree_east";
+    float bounds_con_lon(con_lon,2);
+    float sigma(sigma);
+      sigma:bounds="bounds_sigma";
+      sigma:long_name="sigma";
+    float bounds_sigma(sigma,2);
+  data:
+    con_time=19960901.0;
+    bounds_con_time=19960301.0, 19970301.0;
+    con_lon=180;
+    bounds_con_lon=0, 360;
+    sigma=0.99, 0.96, 0.92, 0.8, 0.5, 0.1;
+    bounds_sigma=0.98,1.00, 0.94,0.98, 0.86,0.94,
+                 0.65,0.86, 0.30,0.65, 0.05,0.30;
+
+This is a mean over the complete +range of longitude from 1 March 1996 to 1 March 1997 +(see section 25 concerning the time coordinate). +The longitude axis indicates circular topology because +this was the case before it was collapsed; after collapse, the topology +is not really meaningful. +If the humidity was subsequently meaned over the depth of the atmosphere +as well, +subgrid would be suffixed with con_sigma: mean, +and con_sigma would have bounds 0.05 and 1.00. + +

+If the same axis is contracted repeatedly, +the methods may all be +recorded in the subgrid attribute of the data variable, +but only the most recent old_interval +and old_spacing will be shown on the contracted coordinate variable. +But if the axis before contraction is retained in the file (identified by +an expand attribute), and was itself the result of a contraction, +it can record the previous old_interval and old_spacing. + +

+Repeated operations of some methods can be regarded as equivalent to a +single operation. +For instance, meaning longitude cells of 1 degree width to 5 degrees, +and then from 5 to 45 degrees, gives the same result as meaning in one step +from 1 degree to 45 degrees (apart from complications with missing data). +Similarly, meaning a time axis from days into months, then into seasons, +and finally into years could be represented as a single operation of meaning +from days to years. +In that case, the subgrid, +old_interval and old_spacing attributes need not be modified +for successive operations. +The choice of whether to take this approach +is left to the application. + +

+Some methods imply a change in the physical dimensions of the +quantity. For these methods, the quantity name must be modified, +and the units changed as indicated by Appendix B. +The commonest such method is variance, which changes +a variable with quantity="temperature" +and units="K" to one with quantity="variance of temperature" +and units="K2". +Methods of simultaneous combination and contraction +of separate quantities also imply a change in the physical dimensions. +We do not have subgrid attributes for arbitrary combinations +of quantities, but only for those operations which are frequently used +to represent the simultaneous variation of separate quantities, +such as covariance and correlation. + +

+ +


Covariance of temperature and wind: +The covariance method is one which combines two quantities in order +to perform a contraction. + +

+  dimensions:
+    lon=96;
+    lat=72;
+    con_time=1;
+  variables:
+    float eddyut(con_time,lat,lon);
+      eddyut:axis="TYX";
+      eddyut:long_name="covariance of temperature and eastward wind";
+      eddyut:subgrid="con_time: covariance";
+      eddyut:units="K m s-1";
+    float con_time(con_time);
+    float bounds_con_time(con_time,2);
+
+The subgrid attribute indicates that the quantity is a temporal +covariance, +and the time bounds specify the period of time over which the covariance +was calculated. +An old_interval attribute could record the original spacing in time +of the values, hourly for instance. + +

+

+23  Time variables and intervals

+ + + +

+A ``time variable'' is one which represents date and time, +which we will refer to hereafter just as ``time''. +An ``interval of time'' is the difference between two times. + +

+It would be possible to describe +time in terms of six components (year, month, day, hour, minute, second) +in a netCDF file, +using six component variables of various data types. +However, it is more efficient and for many purposes more convenient +to represent a time as a single number, +giving the elapsed interval since a certain reference time, +which may be either implicit or explicit. +We refer to conversion from the components of a time into a single +number as ``encoding'', and the reverse as ``decoding''. +Encoding and decoding are complicated because +year and month are units with lengths that depend on the date +and the calendar in use, +so special provisions are needed for time axes. + +

+A ``calendar'' defines the set of valid dates +(year-month-day combinations). +The standard calendar is the Gregorian +(the calendar of udunits), +but climate models do not always use this. +For instance, in the calendar of +the Hadley Centre GCM, all months have 30 days. +The elapsed interval in units of fixed length +(days, hours, minutes, seconds) +between two times will not necessarily be the same +in two different calendars, +because there may be different numbers of valid dates between them. +For example, +the interval between 1 February 1996 and 1 March 1996 is one month, +and equals +29 days in the standard calendar, but 30 days in the Hadley Centre +model calendar, +since 30 February is a valid date in the latter. +Therefore the encoding of a time into an elapsed interval will +depend on the calendar, +and it is necessary to know the calendar when converting. +This standard permits the use of the standard calendar +(below, section 26) +and of other calendars (section 27). +The calendar attribute, +described in the following sections, +indicates the calendar in use. +If a time coordinate variable has no calendar attribute, +the global calendar attribute (section 5), +if present, applies to it. + +

+This standard permits two different methods, +distinguished by their units, +of encoding a time into a number. +These methods, referred to as ``relative time'' +and ``absolute time'', +are described in the following sections +(24 and 25). +Relative time is a more familiar method, +but absolute time offers important advantages. + +

+Time variables may have +an attribute time_format, to specify +a format for printing the date and time, +according to the conventions of the +Unix (TM) date command. + +

+Time coordinate variables must always explicitly include +the units attribute; there is no default value. + +

+

+24  Relative time

+ + + +

+A time encoded as a relative time gives the elapsed interval since +a specified reference time; +units takes the form "time-unit since +reference-time", +as per the recommendations of +the Unidata udunits package (but see below concerning +the time-unit) +e.g. a unit of seconds since 1992-10-8 15:15:42.5 +indicates seconds since 8 October 1992 at 3 hours, 15 minutes and +42.5 seconds in the afternoon, in Universal Coordinated Time +(time zones can also be handled). +In order to decode the values on a relative time axis, +the application will in general need to know the calendar; +the encoded time values are meaningless without this knowledge. +Furthermore, a given date may result in different time values +when encoded in two different calendars +with the same units. +For instance, 1996-2-1 15:00:00 is 62.625 days since +1995-12-1 0:0:0 in the standard calendar, +and 60.625 days since 1995-12-1 0:0:0 in the 360-day calendar. + +

+The file udunits.dat defines +second, minute, hour and day as units of time. +Units of months and years are disallowed by Appendix C +of this standard, +because they are not well-defined; +since udunits defines a year as a ``tropical year'' +of 31556925.97 s (674.03 s less than 365 days) +and a month as exactly a twelfth of a year, +use of these units will probably not give the expected results. +For example, 1 month since 1995-4-1 0:0:0 +is treated by udunits as +30.4368 days since 1995-4-1 0:0:0, which is approximately +1995-5-1 10:29, +not 1995-5-1 0:0:0. +Also, 1 year since 1995-4-1 0:0:0 is about 1996-3-31 5:49, +not 1996-4-1 0:0:0. +The udunits unit common_year (exactly 365 days) +is permitted, but not recommended. + +

+ +


A relative time axis for instantaneous measurements of a +quantity: Measurements are made +at noon on 2nd-5th June 1996. + +

+  dimensions:
+    time=4;
+  variables:
+    double time(time);
+      time:long_name="time";
+      time:units="days since 1996-1-1 0:0:0";
+  data:
+    time=1.5, 2.5, 3.5, 4.5;
+
+ +

+ +


A relative time axis for monthly means: +Means are calculated for February, March +and April of 1990. + +

+  dimensions:
+    time=3;
+  variables:
+    double time(time);
+      time:bounds="bounds_time";
+      time:long_name="time";
+      time:units="days since 1990-1-1 0:0:0";
+    double bounds_time(time,2);
+  data:
+    time=45.0, 74.5, 105.0;
+    bounds_time=31.0,59.0, 59.0,90.0, 90.0,120.0;
+
+In this example, the main time coordinates are merely representative +values, being the mid-points of their respective months. +Decoded, they are 1990-2-15 0:0:0, 1990-3-16 12:0:0 +and 1990-4-16 0:0:0. + +

+

+25  Absolute time

+ + + +

+This method of encoding time refers to the separate components +of time, rather than to a single unit of fixed length. +It offers two advantages. +Firstly, the encoded times are meaningful and can be decoded +into components of time without knowledge of the +calendar, although to calculate intervals between them +this knowledge is still required. +Secondly, ``partial'' times can be encoded, +which omit the year, or the ``seasonal phase'' (time of year, +time within the seasonal cycle), or the ``diurnal phase'' +(time of day, time within the diurnal cycle). +By contrast, relative times can only be ``complete'' times, +which include information about all three of these. + +

+The units attribute of absolute time takes the form +"time-unit as time-string", +The possibilities with the recommended data types +and their meanings are as follows: + +

+ + + + + + + + + + + + + + +
Format Data type Interpretation
second as %S.%f float Diurnal phase
minute as %M.%f float Diurnal phase
hour as %H.%f float Diurnal phase
day as %Y%m%d.%f double Time
day as %Y%m%d int Year and seasonal phase
day as %m%d.%f double Seasonal phase and diurnal phase
day as %m%d int Seasonal phase
day as .%f float Diurnal phase
calendar_month as %Y%m.%f double Year and seasonal phase
calendar_month as %m.%f float Seasonal phase
calendar_year as %Y.%f double Year and seasonal phase
calendar_year as %Y int Year
calendar_year as .%f float Seasonal phase
+
+Standard abbreviations and plural forms of the unit names are acceptable, +as usual. +The time-units calendar_year and calendar_month +are units of time defined by this standard (Appendix C). + +

+The time-string +codes show how the year, month, day within month +and time within day are encoded into a single number, +after the fashion of the Unix (TM) date +and printf commands, thus: + +

+ + + + + + + + + +
Format letter Interpretation
%Y Year (including century)
%m Two-digit month (01=January)
%d Two-digit day within month
%H Hours since midnight
%M Minutes since midnight
%S Seconds since midnight
%f Floating-point fraction of the specified time-unit
. Position of decimal point
+
+Since an encoded time is an ordinary number, leading zeroes in the +integer part may be omitted. +It is not mandatory to use the suggested data types, which are recommended +on grounds of precision. +If an integer data type is used for an absolute time variable whose format +includes a fraction %f, +a fraction of zero is assumed. +If a floating-point data type is used for a format not including %f, +any fraction is ignored. + +

+In absolute time, +3 p.m. on 5th April 1998 is encoded with value 19980405.625 +and units="day as %Y%m%d.%f". +The advantage of this method of encoding a complete time is that it +can be done without knowledge of the calendar, +whereas if we encoded in relative time units of +days since 1900-1-1, the value would be 35888.625 in the standard +calendar, and 35374.625 in the 360-day calendar. +We also know, without reference to the calendar, that the value +19980605.625, with the same units, is a time exactly two calendar months +later, and 19970405.625 is exactly one calendar year earlier. +But to calculate these intervals in other time units-days, hours, etc.-we +still need to know the calendar. + +

+The only complete form of absolute time +is "day as %Y%m%d.%f". +Note in particular that the forms "calendar_month as %Y%m.%f" +and "calendar_year as %Y.%f" are partial times which imply +no information about the diurnal phase. +This is a very important point. +For instance, 1998.25 calendar_year as %Y.%f means no more than +``a quarter of the way through 1998 as regards the seasonal cycle''. +This meaning is the same in the standard and 360-day calendars. +Because this representation carries no information about diurnal phase, +it is not permitted to decode it to 1998-4-2 3:0:0 +(i.e. 91.25 days from the start of the year) +in the standard calendar or +1998-4-1 0:0:0 in the 360-day calendar. +Similarly, 199804.3 calendar_month as %Y%m.%f means +``30% of the way through April 1998 as regards the seasonal cycle''. +Examples below show the use of such partial times. + +

+Note also +that the only form of partial time which is composed of the +seasonal and diurnal phases is "day as %m%d.%f"; +there is no method of encoding the seasonal phase as a fraction +of a calendar year or month in combination with the diurnal phase. +Should this be required, the application could construct it as a +two-component time variable. +This exclusion seems reasonable because data which resolves both +the seasonal and diurnal cycles must belong to a known calendar +(it will exhibit a certain number of days in a year, for instance) +and so its seasonal cycle can be labelled by month and day. +The calendar-independent representations of the seasonal cycle, +shown in examples below, +are more useful when portions of the seasonal cycle have been averaged, +in which case the seasonal and diurnal cycles, if both present, +will be on separate axes. + +

+The forms of partial time which do not include .%f +in the time-string are discrete variables, rather than continuous. +The interval of time in days or years +covered by a time axis in one of these forms +is reckoned by including the both ends of the axis in the count, +or equivalently by adding one to the difference of the ends. +For instance, an axis with units="calendar_year as %Y" +that runs from 1930 to 1939 covers ten years, not nine, +since both ends are included. +This is a partial time indicating only the year, +not the seasonal phase. +Contrast this with an axis including the seasonal phase, +having units="calendar_year as %Y.%f", +and end points of 1930.0 and 1939.0. +This axis spans nine years, from the beginning of 1930 to the +beginning of 1939, and does not include 1939 itself. +Examples below illustrate this point further. + +

+ +


An absolute +time axis for instantaneous measurements of a quantity: +Measurements are made at noon on 2nd-5th June 1996. + +

+  dimensions:
+    time=4;
+  variables:
+    double time(time);
+      time:long_name="time";
+      time:units="days as %Y%m%d.%f";
+  data:
+    time=19960602.5, 19960603.5, 19960604.5, 19960605.5;
+
+ +

+ +


An absolute time axis for monthly means, encoded in days: + +

+  dimensions:
+    time=3;
+  variables:
+    double time(time);
+      time:bounds="bounds_time";
+      time:long_name="time";
+      time:units="days as %Y%m%d.%f";
+    double bounds_time(time,2);
+  data:
+    time=19900215.0, 19900316.5, 19900416.0;
+    bounds_time=19900201.0,19900301.0, 19900301.0,19900401.0,
+                19900401.0,19900501.0;
+
+As in the relative time version of this example, +the main time coordinates are the mid-points of their respective months. +Although they are encoded straightforwardly, +their values depend on the calendar. +If one was comparing means over these months from data sources which +used different calendars, that might be inconvenient, +and could be avoided as in the next example. + +

+ +


An absolute time axis for monthly means, encoded in months: + +

+  dimensions:
+    time=3;
+  variables:
+    double time(time);
+      time:bounds="bounds_time";
+      time:long_name="year and seasonal phase";
+      time:units="calendar_months as %Y%m.%f";
+    double bounds_time(time,2);
+  data:
+    time=199002.5, 199003.5, 199004.5;
+    bounds_time=199002.0,199003.0, 199003.0,199004.0, 199004.0,199005.0;
+
+This method shows directly that the main coordinates are half-way +through their months. + +

+ +


A partial time defining just the year: +An axis of this kind could be used to record +the number of occurrences of a particular kind of event: + +

+  dimensions:
+    year=3;
+  variables:
+    int year(year);
+      year:long_name="year";
+      year:units="calendar_year as %Y";
+    int count(year);
+  data:
+    year=1991,1992,1993,1994,1995;
+    count=0,2,1,0,1;
+
+As discussed above, this axis spans five years. +No boundaries are supplied, because each element of count +applies to just a single year. +Hence, both the upper and the lower year boundaries would be +equal to their years. + +

+ +


Year and seasonal phase defined in calendar years: +By contrast to the last example, if it was appropriate to indicate that +each count applied to the whole of the continuous period of time of its +respective year, this could be done thus: + +

+  variables:
+    double year(year);
+      year:bounds="bounds_year";
+      year:long_name="year and seasonal phase";
+      year:units="calendar_year as %Y.%f";
+    double bounds_year(year,2);
+    int count(year);
+  data:
+    year=1991.5, 1992.5, 1993.5, 1994.5, 1995.5;
+    bounds_year=1991.0,1992.0, 1992.0,1993.0, 1993.0,1994.0,
+                1994.0,1995.0, 1995.0,1996.0;
+    count=0,2,1,0,1;
+
+The use of floating-point years allows us conveniently to represent +exactly the beginning and ending of a year and a point half-way through. +In the standard calendar, of course, the interval from 1992.0 to 1993.0 +is longer in relative time than all the other years. +But for some purposes, it might be more useful to record that each interval +is a calendar year. +This could be especially helpful when comparing data from different +calendars. + +

+ +


Seasonal phase as a function of year: +Here we show the date within the year of a particular event, +such as the highest daily +maximum temperature, or the onset of the monsoon, as a partial time +within its year. + +

+  dimensions:
+    year=5;
+  variables:
+    int year(year);
+      year:long_name="year";
+      year:units="calendar_year as %Y";
+    int date(year);
+      date:long_name="seasonal phase";
+      date:units="day as %m%d";
+  data:
+    year=2011, 2013, 2027, 2028, 2051;
+    date=629, 627, 626, 703, 710;
+
+The event concerned occurred on 29th June 2011, 27th June 2013, +26th June 2027, 3rd July 2028 and 10th July 2051. +Clearly the date variable could have been encoded as a complete time, +perhaps in relative time units, +but this would have included redundant year information. + +

+A time variable which indicates seasonal phase but not year has a +modulo of one year. If it spans the entire seasonal cycle, it also has +circular topology. Similarly, a time variable indicating diurnal phase +but not seasonal phase has a modulo of one day, and has circular +topology if it spans the entire diurnal cycle. +These kinds of time coordinate are particular useful for +representing climatological time, in conjunction with other contracted +time axes. See section 28. + +

+ +


Average seasonal cycle expressed in months: +Data for solar radiation as 3-monthly averages. + +

+  dimensions:
+    time=4;
+    lat=72;
+    lon=96;
+  variables:
+    float sol(time,lat,lon);
+      sol:axis="TYX";
+      sol:long_name="vertical component of "
+        "solar radiative flux density";
+      sol:units="W m-2";
+    float time(time); 
+      time:bounds="bounds_time";
+      time:long_name="seasonal phase";
+      time:modulo=12.0f;
+      time:topology="circular";
+      time:units="calendar_month as %m.%f";
+    float bounds_time(time,2);
+  data:
+    time=10.5, 13.5, 16.5, 19.5;
+    bounds_time=9.0,12.0, 12.0,15.0, 15.0,18.0, 18.0,21.0;
+
+The first time point applies from the beginning of month 9 to the +beginning of month 12, i.e. September to November inclusive. +A representative main coordinate is given of half-way through October. +The second point runs to the beginning of month 15, which is equivalent +to 3 i.e. March under modulo 12, and thus covers December to February. +The use of the modulo allows the main coordinate to be specified as +monotonic, as is generally required. Because the axis is also circular, +it would be permissible to rotate the values in order to begin with a +different season. + +

+ +


Average seasonal cycle expressed in years: The time +coordinate above could equally well be given in calendar years, thus: + +

+  dimensions:
+    double time(time); 
+      time:bounds="bounds_time";
+      time:long_name="seasonal phase";
+      time:modulo=1.0;
+      time:topology="circular";
+      time:units="calendar_year as .%f";
+    float bounds_time(time,2);
+  data:
+    time=0.7917, 1.0417, 1.2917, 1.5417;
+    bounds_time=0.6667,0.9167, 0.9167,1.1667,
+                1.1667,1.4167, 1.4167,1.6667;
+
+Here, the periods have been constructed as exactly quarters of a +year, beginning two-thirds of the way through the year. +In the 360-day calendar, this is identical to the last example, of +periods of three months starting at the beginning of September, +but in the standard calendar it is slightly different, +since a quarter of a year is not exactly three calendar months. + +

+

+26  Gregorian calendar

+ + + +

+This standard recommends that Gregorian times be given in +units of +days as %Y%m%d.%f with data type double +(section 25), +unless compatibility is essential with applications that cannot process +absolute times. +In that case, Gregorian times may have units of time +formatted as per the recommendations of the Unidata udunits package, +which specify a unit and a reference time, i.e. a relative time +(section 24). +The recommended unit is days, with data type double. + +

+Intervals between two times in the standard +Gregorian calendar can be calculated by the Unidata udunits package. +Udunits implements the mixed Gregorian/Julian calendar system, +as followed in England, in which +dates prior to 1582-10-15 are assumed to use the Julian calendar. +Other software cannot be relied upon to handle the change of calendar +in the same way, so for robustness it is recommended that the reference +date be later than 1582. +If earlier dates must be used, it should be noted that +udunits treats 0 AD as identical to 1 AD. + +

+Data type double gives a precision of about +16 decimal digits, +which means that it can resolve tenths of a second +for years of up to O(1 million) in relative times. +The precision of absolute times is an order of magnitude worse, +since a year looks like 10 000 days, rather than 365. +The larger the year, the worse the absolute precision. +If very large years are needed and the precision is not sufficient, +the reference year will have to be modified to keep the interval +small enough. + +

+If there is no calendar attribute applying to +a time variable, +the values are assumed to be in the normal Gregorian calendar. +This can be made explicit by setting calendar to standard +or gregorian. + +

+

+27  Non-Gregorian calendars

+ + + +

+It is recommended that +times in other calendars should be encoded in units of +days as %Y%m%d.%f with data type double +(section 25). +Relative times are permitted, the recommended units being +days since 1-1-1 (midnight on 1 January of year 1), +with data type double. +Since the Unidata udunits package can process only +the standard calendar, +an extension will be required to process relative times +for other calendars. + +

+Apart from the Gregorian, +calendars recognised by this standard are +julian for the Julian calendar (in which all years divisible +by four are leap years), +noleap for a calendar with 365 days in every year, +and 360 when each month has 30 days in every year. +If any other calendar is used, a suitable description should appear +in the calendar attribute, +but generic applications cannot be expected to be able to encode and +decode relative times or calculate intervals in the calendar concerned. + +

+

+28  Multiple time axes and climatological time

+ + + +

+There is no bar on a data variable having more than one +dimension in +a particular quantity, so long as the dimensions have different names. +A particular use of this is to decompose time +into multiple partial time dimensions (section 25), +of which one or may be collapsed (section 22). +This gives a method of indicating disjoint intervals of time +belonging to corresponding parts of the seasonal or diurnal cycles. +When a variable has two or three time axes, +the first interval of time which they cover is assumed to begin +at the earliest boundary values of all the axes. +If there is an uncollapsed axis in combination with collapsed axes, +it is a ``climatological time'' axis. +There may be more than one-see below for an example. + +

+COARDS recommends use of year 0 to indicate climatological time. +We do not favour this convention. +Firstly, it does not provide any way of recording which years were used +to make the climatology. +Secondly, udunits treats year 0 and year 1 as identical +(which is reasonable because year 0 does not exist-there is +no year between 1 AD and 1 BC). + +

+ +


A mean of a corresponding months in a number of years: +A longitude-latitude precipitation field with time axes +to indicate the mean over the months of January in 1961 to 1990 inclusive: + +

+  dimensions:
+    con_year=1;
+    year=30;
+    month=1;
+  variables:
+    float precipitation(con_year,month,lat,lon);
+      precipitation:axis="-TYX";
+      precipitation:subgrid="month: mean con_year: mean";
+    int con_year(con_year);
+      con_year:bounds="bounds_con_year";
+      con_year:expand="year";
+      con_year:long_name="year";
+      con_year:old_interval=1;
+      con_year:units="calendar_year as %Y";
+    int bounds_con_year(con_year,2);
+    int year(year);
+    float month(month);
+      month:bounds="bounds_month";
+      month:long_name="seasonal phase";
+      month:units="calendar_month as %m.%f";
+    float bounds_month(month,2);
+  data:
+    con_year=1975;
+    bounds_con_year=1961, 1990;
+    year=1961, 1962, 1963, ..., 1990;
+    month=1.5;
+    bounds_month=1.0, 2.0;
+
+The representative year is not likely to be particularly useful in +this case; the important information is the boundaries, which indicate +the range of years used to form the climatological mean. +These years are also given explicitly, and optionally, for reference. +Since the con_year axis has a discrete form of time (%.f does +not appear in the time-string-see section 25), +we have to include both ends of the axis in working out how many years are +involved in the mean: +bounds_con_year tells us that years from 1961 to 1990 were used, +which comes to 30 years, counting both 1961 and 1990. + +

+Suppose we wish to include in the same data variable the mean over Decembers +from 1960 to 1989. +To do this, we give the month variable modulo=12.0f, +which previously was not needed, change the dimension to month=2 and +the data to + +

+    month=0.5, 1.5;
+    bounds_month=0.0,1.0, 1.0,2.0;
+
+According to the convention, the earliest time indicated by the combined axes +is the lower boundary of all of them, which is the start of the 0th month of +1961, equivalent to the start of the 12th month of 1960 under modulo 12. +Alternatively, we could have + +
+    bounds_con_year=1960,1989;
+    month=12.5, 13.5;
+    bounds_month=12.0,13.0, 13.0,14.0;
+
+This is exactly equivalent. The last January used ends at the start of the 14th +month of 1989, which is the start of February 1990, as before. + +

+There is no standardised way to indicate missing months from the mean, say +December 1974 in this case, +except to include con_year:old_spacing="disjoint". +The information could be included as a note in the subgrid attribute, thus +"con_year: mean (December 1974 missing)". + +

+ +


Climatological seasonal means for several decades: +This is an extension of the previous case, +and of the +example of an average seasonal cycle in section 25. +Here, the axes are set up +to indicate climatological means for two of the seasons +in three successive decades. + +

+  dimensions:
+    decade=3;
+    season=2;
+  variables:
+    float precipitation(decade,season,lat,lon);
+      precipitation:axis="-TYX";
+      precipitation:subgrid="season: mean decade: mean";
+    int decade(decade);
+      decade:bounds="bounds_decade";
+      decade:old_interval=1;
+      decade:units="calendar_year as %Y";
+    int bounds_decade(decade,2);
+    int season(season);
+      season:bounds="bounds_season";
+      season:calendar="standard";
+      season:modulo=1200;
+      season:units="day as %m%d";
+    int bounds_season(season,2);
+  data:
+    decade=1966, 1976, 1986;
+    bounds_decade=1961,1970, 1971,1980, 1981,1990;
+    season=115, 415;
+    bounds_season=1,228, 301,531;
+
+Here, precipitation[0][0][*][*] is the data for December-February +(i.e. 1 December to 28 February inclusive) +of the decade 1960-1970 (first December in 1960, last February in 1970), +while [2][1][*][*] is March-May 1981-1990. +The choice has been made to give the seasonal phase in months and days, +rather than months alone; +hence the modulo is 1200 rather than 12. +Under modulo 1200, midnight on 1 December can be expressed equivalently +as 1 or 1201. +If 1201 were specified, it would mean that the first interval of time +began on 1 December 1961 (rather than 1960), +taking the combination of the lower boundaries of both time axes; +the value 1 is a year earlier. +The drawback of this %m%d scheme +is that it is awkward or impossible to give accurate +representative dates for the middle of the periods, especially since +February has variable length. +The absolute time format %m.%f for seasonal phase +is better from this point of view. +The season axis is not shown as having circular topology because no +information is implied about the other two seasons. + +

+ +


Average early June maximum temperatures for several years: +In this example, the dimensions indicate that +maximum daily temperatures (between 9 a.m. on the day +of record and 9 a.m. of the previous day) were recorded for 1-10 June, +and an average maximum found for these ten days in +each of the years 1980-1984. + +

+  dimensions:
+    year=5;
+    con_season=1;
+    con_day=1;
+  variables:
+    float temperature(year,con_season,con_day);
+      temperature:axis="T--";
+      temperature:subgrid="con_day: maximum con_season: mean";
+    int year(year);
+      year:long_name="year";
+      year:units="calendar_year as %Y";
+    int con_season(con_season);
+      con_season:bounds="bounds_con_season";
+      con_season:long_name="seasonal phase";
+      con_season:old_interval=1;
+      con_season:units="day as %m%d";
+    int bounds_con_season(con_season,2);
+    float con_day(con_day);
+      con_day:bounds="bounds_con_day";
+      con_day:long_name="diurnal phase";
+      con_day:modulo=24.0f;
+      con_day:units="hour as %H.%f";
+    float bounds_con_day(con_day,2);
+  data:
+    year=1980, 1981, 1982, 1983, 1984;
+    con_season=605;
+    bounds_con_season=601, 610;
+    con_day=-3.0;
+    bounds_con_day=-15.0, 9.0;
+
+The diurnal phase of -15 h means 15 hours before the beginning +of the day in question, i.e. 9 a.m. on the previous day. +No bounds are given for the year, because it is a discrete quantity, +and there is no further information which could be added. +But if the five years were averaged together, this would collapse the year +axis, and the extreme years of 1980 and 1984 would be recorded as +the boundaries of the collapsed axis. +If, say, 1981 were not used in forming the average, +the collapsed axis would have attribute old_spacing="disjoint". + +

+ +


Daily values as an average of subdaily values: +Instantaneous pressure measurements are made at intervals of 3 hours +(first measurement at midnight) +throughout the days 6 May to 9 June 1937, +and daily means formed from midnight to midnight. + +

+  dimensions:
+    con_subday=1;
+    day=35;
+  variables:
+    float pressure(day,con_subday);
+      pressure:axis="T-";
+      pressure:subgrid="con_subday: point con_subday: mean";
+    float con_subday(con_subday);
+      con_subday:bounds="bounds_con_subday";
+      con_subday:long_name="diurnal phase";
+      con_subday:old_interval=0.125f;
+      con_subday:old_spacing="uniform";
+      con_subday:units="days as .%f";
+    float bounds_con_subday(con_subday,2);
+    int day(day);
+      day:long_name="year and seasonal phase";
+      day:units="days as %Y%m%d";
+  data:
+    con_subday=0.5;
+    bounds_con_subday=0.0, 0.875;
+    day=19370506, 19370507, ..., 19370608, 19370609;
+
+Note that the con_subday axis is shown with two +subgrid methods, referring to subgrid variation before +and after its collapse. +The only point here in having separate axes for day and diurnal phase +is to show when the first and last instantaneous measurements were made +in each day. +If this is not important to record, the two axes could be merged together +thus: + +
+  dimensions:
+    day=35;
+  variables:
+    float pressure(day);
+      pressure:subgrid="day: point day: mean";
+    float day(day);
+      con_subday:bounds="bounds_day";
+      con_subday:old_interval=0.125f;
+      con_subday:long_name="time";
+      con_subday:old_spacing="uniform";
+      con_subday:units="days as %Y%m%d.%f";
+    float bounds_day(day,2);
+  data:
+    day=19370506.5, 19370507.5, ..., 19370608.5, 19370609.5;
+    bounds_day=19370506.0,19370507.0, 19370507.0,19370508.0, ...,
+               19370608.0,19370609.0, 19370609.0,19370610.0;
+
+If the 35 days were then averaged together, +the date axis would collapse with bounds of 19370506.0 and 19370610.0. +The subgrid attribute would not need modification since +it is already shown as a mean over the day axis. + +

+ +


Average diurnal cycle: +The following axes are appropriate for +the average diurnal cycle of +precipitation rate in July 1970-1979 as a function of latitude: + +

+  dimensions:
+    con_year=1;
+    con_month=1;
+    hour=8;
+    lat=45;
+    con_lon=1;
+  variables:
+    float ppnrate(con_year,con_month,hour,lat,con_lon);
+      ppnrate:axis="--TYX";
+      ppnrate:subgrid="con_lon: mean con_month: mean "
+        "con_year: mean";
+      ppnrate:units="kg m-2 s-1";
+    int con_year(con_year);
+      con_year:bounds="bounds_con_year";
+      con_year:old_interval=1;
+      con_year:units="calendar_year as %Y";
+    int bounds_con_year(con_year,2);
+    float con_month(con_month);
+      con_month:bounds="bounds_con_month";
+      con_month:units="calendar_month as %m.%f";
+    float bounds_con_month(con_month,2);
+    float hour(hour);
+      hour:bounds="bounds_hour";
+      hour:modulo=24.0f;
+      hour:topology="circular";
+      hour:units="hour as %H.%f";
+    float bounds_hour(bounds_hour,2);
+  data:
+    con_year=1975;
+    bounds_con_year=1970, 1979;
+    con_month=7.5;
+    bounds_con_month=7.0, 8.0;
+    hour=1.5, 4.5, 7.5, 10.5, 13.5, 16.5, 19.5, 22.5;
+    bounds_hour=0.0,3.0, 3.0,6.0, 6.0,9.0, 9.0,12.0,
+                12.0,15.0, 15.0,18.0, 18.0,21.0, 21.0,24.0;
+
+ +

+

+29  Invalid values in a data variable

+ + + +

+Invalid values are any which fall +outside the valid range or equal the fill value, as indicated by the +Unidata-standard attributes described here. An invalid value indicates bad data +i.e. a software problem, which is a different circumstance from unknown or +missing data (see section 30). +Invalid values are not permitted in a coordinate variable, +but the attributes which define the valid range may be used in boundary +variables (section 20) to indicate unbounded cells. + +

+The attribute valid_min is a scalar specifying the +minimum valid value for a variable. The attribute valid_max specifies +the maximum valid value, while valid_range is a vector of two numbers +specifying the minimum and maximum valid values, in that order, equivalent to +specifying values for both valid_min and valid_max +attributes. Any of these attributes define the valid range. The attribute +valid_range must not be defined if either valid_min or +valid_max is defined. Generic applications should treat values outside the +valid range as invalid. The type of each valid_range, valid_min +and valid_max attribute should match the type of its variable. The +Unidata special treatment of byte type is not included here as we do not +recommend use of that type (see section 3). + +

+A scalar attribute with the name _FillValue and of the +same type as its variable is used as the fill value for the variable. The +netCDF package defines a default fill value for each type of variable, so it is +not necessary to define your own _FillValue attribute if the default is +suitable. The purpose of the fill value is to save the applications programmer +the work of prefilling the data and also to eliminate the duplicate writes that +result from netCDF filling in undefined data with its default fill value, +only to be immediately overwritten by the programmer's preferred value. This +value is considered to be a special value that indicates undefined data, and is +returned when reading values that were not written. The _FillValue +should be outside the range specified by valid_range (if used) for a +variable. +In cases where the data variable is packed using the +scale_factor and add_offset attributes (section 32), the +_FillValue attribute applies the numbers as packed, so they must be +checked against it before unpacking. + +

+If none of valid_min, valid_max or valid_range +is defined then generic applications should define a valid range by using the +fill value (whether defined explicitly or by default); if the fill value is +positive then it defines a valid maximum, otherwise it defines a valid +minimum. For integer types, there should be a difference of 1 between the fill +value and this valid minimum or maximum. For floating point types, the valid +extreme should have a magnitude which is half the magnitude of the fill value. +We recommend a factor of two, rather than a difference of one bit, because +it is easier for applications programmers. There is no special treatment for +byte as we do not recommend that type (see section 3). + +

+

+30  Missing values in a data variable

+ + + +

+Missing values are not permitted in a coordinate variable, so this +section applies only to data variables. The missing_value attribute +indicates a value that is used for data that are unknown or ``missing''. This +attribute is not be treated in any special way by the netCDF API, unlike the +_FillValue attribute (section 29). +The missing_value should be outside the valid range +(section 29), so that generic applications will treat it +appropriately. +The netCDF data type of +the missing_value attribute should match the netCDF data type of the +data variable that it describes. In cases where the data variable is packed via +the scale_factor and add_offset attributes (section 32), +the missing_value attribute +matches the type of and should be compared with the data +after unpacking. +This standard is unlike COARDS in giving +a particular interpretation to the +distinction between missing_value and _FillValue. + +

+

+31  Compression by gathering

+ + + +

+To save space in the netCDF file, +it may be desirable to eliminate points +from data arrays which are invariably missing. +Such a compression can operate over one or more adjacent axes, +and is accomplished with reference to a list of +the points to be stored. +The list is constructed by considering a mask array which has just the +axes to be compressed, +and mapping this array onto one dimension without reordering. +The list is the set of indices +in this one-dimensional mask of the required points. +In the compressed array, +the axes to be compressed are all replaced by a single axis, +whose dimension is the number of wanted points. +The wanted points appear along this dimension in the same order they +appear in the uncompressed array, with the unwanted points skipped over. +Compression and uncompression are executed by looping over the list. + +

+The list is stored as the +coordinate variable for the compressed axis of the data array. +Thus, the list variable and its dimension have the same name. +The list variable has a string attribute compress, +containing a blank-separated list of the dimensions which were affected +by the compression in the order of the CDL declaration of the +uncompressed array. +The presence of this attribute identifies the list variable as such. +The list, the original dimensions and coordinate variables +(including component, associated and boundary variables), +and the compressed data variables with all the attributes +of the uncompressed variables +are written to the archived netCDF file. +The uncompressed data variables can be reconstituted +exactly as they were +using this information, +except that their original variable names are not known. + +

+ +


Horizontal compression of a three-dimensional array: +We eliminate sea points at all depths +in a longitude-latitude-depth array of soil temperatures. +In this case, only the longitude and latitude axes would be affected +by the compression. +We construct a list landpoint(landpoint) +containing the indices of land points. + +

+  dimensions:
+    lat=73;
+    lon=96;
+    landpoint=2381;
+    depth=4;
+  variables:
+    long landpoint(landpoint);
+      landpoint:compress="lat lon";
+    float landsoilt(depth,landpoint);
+      landsoilt:axis="Z-";
+      landsoilt:long_name="soil temperature";
+      landsoilt:units="K";
+    float depth(depth);
+    float lat(lat);
+    float lon(lon);
+  data:
+    landpoint=363, 364, 365, ...;
+
+Since landpoint[0]=363, for instance, +we know that landsoilt[*][0] +maps on to point 363 of the original data with dimensions (lat,lon). +This corresponds to indices [3][75]. + +

+ +


Compression of a three-dimensional field: +We compress a longitude-latitude-depth field of ocean +salinity by eliminating points below the sea-floor. +In this case, all three dimensions are affected by the compression, +since there are successively fewer active ocean points at increasing +depths. + +

+  variables:
+    float salinity(oceanpoint);
+      salinity:axis="-";
+    long oceanpoint(oceanpoint);
+      oceanpoint:compress="depth lat lon";
+    float depth(depth);
+    float lat(lat);
+    float lon(lon);
+
+This information implies that the salinity field should be +uncompressed to an array with dimensions (depth,lat,lon). + +

+

+32  Compression using a scale and offset

+ + + +

+This standard +endorses the use of the optional Unidata-standard attributes +scale_factor and add_offset +for data and coordinate variables. +These attributes +can be used to provide simple number compression (packing), +to store low-resolution +floating-point data as small integers in a netCDF file. +After the data values of the variable have been read in, +they are to be multiplied by the scale_offset, +and have add_offset added to them. +If both scale_factor and add_offset attributes are present, the +data are scaled before the offset is added. +When scaled data are +written, the application should first subtract the offset and then divide by +the scale factor. +This procedure is concerned only with storage. +It does not affect the unit of the quantity. +For instance, a pressure variable with values in the range +900.0-1100.0 Pa +could be converted to short integers in the range +�20000 +by subtracting 1000 and dividing by 0.005 i.e. multiplying by 200. +The units of the compressed variable are still recorded as pascals. + +

+This standard is more restrictive than the netCDF Users' +Guide with respect to the use of the scale_factor and add_offset +attributes; ambiguities and precision problems related to data type conversions +are resolved by these restrictions. If the scale_factor and +add_offset attributes are of the same data type as the associated variable no +restrictions apply; the unpacked data is assumed to be of the same data type as +the packed data. However, if the scale_factor and add_offset +attributes are of a different data type from the variable +(containing the packed data) then in files adhering to this standard the +variable may only be of type short or long. We +exclude byte on grounds discussed in section 3. The +attributes scale_factor and add_offset (which must match in data +type) must be of type float or double. The data type of the +attributes should match the intended type of the unpacked data. (It is not +advised to unpack a long into a float as there is a potential +precision loss.) Users should note that Unidata may provide a built-in +means of packing data in netCDF files in future. + +

+

+A  Attributes

+ +

+ +

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
Attribute T Use Section(s) Description
add_offset N CD 29 32 Additive offset for packing data
appendices S G 5 + Version number of these appendices
associate S CD + 18 19 Identifies variables + containing alternative sets of coordinates
axis S D + 9 16 18 Identifies spatiotemporal dimensions
bounds N C + 20 22 28 + Identifies a bounday variable
calendar S GD 5 23 26 + 27 Calendar used for encoding time axes
comment S G 5 + Additional information about the file
component S CD + 17 20 Identifies variables containing + components of a variable
compress S D 31 Records dimensions which have been compressed by gathering
Conventions S G + 5 Identifies the netCDF standard
coordinates S CD + 18 Synonym for associate
expand S C + 22 28 Coordinates + before contraction
_FillValue N D + 29 Indicator of invalid data
FORTRAN_format S CD + 12 Format for a variable
history S GD 5 12 + Evolution of the data in the file
institution S GD 5 12 + Who made or supplied the data
long_name S CD 12 Long + description of a physical quantity
modulo N CD + 12 14 25 Arithmetic + modulo of a variable
north_pole N D + 10 Long.,lat. of rotated + North Pole
old_interval N C 22 28 The typical separation between points on an axis before contraction
old_spacing S C + 22 28 Indicates the spacing + of points along an axis before contraction
positive S CD 12 14 15 + 16 Sense of positive for a variable referring to a spatial direction
production S GD 5 12 + How the data was produced
quantity S CD 12 Standardised + description of a physical quantity
quantity_table S G 5 12 + URL of the quantity table
scale_factor N CD 29 32 Multiplicative factor for packing data
subgrid S D + 21 22 28 Records how the data values represent subgrid variation
topology S C + 13 14 25 Topology + of an axis (circular or not)
time_format S CD + 23 Format for printing a time and date
units S CD 12 14 15 + 23-27 Units of a physical quantity
valid_max N CD + 20 29 Largest valid value of a + variable
valid_min N CD + 20 29 Smallest valid value of a + variable
valid_range N CD + 29 Smallest and largest valid values + of a variable
+
+ +

+T is S for string, N for numeric. + +

+Use is composed of G for global, +C for coordinate variables (including multidimensional +coordinate variables), D for data variables. + +

+

+B  Methods of representing subgrid variation

+ +

+See section 21. + +

+ +

+ + + + + + + + + + + + + + + + +
Method Units Description
cell u Value is a property of the whole cell (e.g. an integral)
correlation 1 Pearson product-moment correlation coefficient
covariance uv Covariance
maximum u Maximum
median u Median
mid-range u Average of maximum and minimum
minimum u Minimum
mean u Mean (average)
mean square u2 Mean (average) square
mode u Mode (most common)
point u Value applies at gridpoint
range u Absolute difference between maximum and minimum
root mean square u Square-root of mean square
standard deviation u Standard deviation (RMS deviation from +the mean)
variance u2 +Variance (mean-square deviation from the mean)
+
+ +

+Units: u means the units of the quantity whose +subgrid variation is represented by this method, +v the units of the second quantity when two quantities are combined. + +

+

+C  Modifications to udunits.dat

+ +

+See section 11. + +

+The unit unity is defined as a dimensionless constant +equal to one. + +

+The unit degrees is not permitted, because +it creates ambiguities when attempting to +differentiate longitude and latitude coordinate variables. +This unit does not appear in the current version +of the file. + +

+The units calendar_month and calendar_year +are units of time, but cannot be converted into each other +or any other units of time, +except that multiples of 12 calendar months equal integral numbers +of calendar years. The units year and month are not allowed, +because they can cause confusion. + +

+

+D  Long names for quantities

+ +

+See section 12. +This Appendix is not yet available. +As well as existing as part of this standard, +it will be made available on the web. + +

+ +

+ + + + + + + + + + +
Version long_name units
1.0 depth below the surface m
1.0 height above the surface m
1.0 latitude degree_north
1.0 longitude degree_east
1.0 pressure Pa
1.0 soil temperature K
1.0 specific humidity unity
1.0 temperature K
1.0 time s
+
+ +

+ +Version: The version of the appendices at which this +quantity was introduced.
+long_name: Case, spaces and punctuation are not +significant in the long_name. + +

+ +

+

+

+Jonathan Gregory | jmgregory@meto.gov.uk +
+
+Robert Drach | drach@llnl.gov +
+
+Simon Tett | sfbtett@meto.gov.uk +
+
+Karl Taylor | ktaylor@pcmdi.llnl.gov +
+

+LLNL Disclaimers +

+UCRL-MI-127703 + + +

+ +

File translated from TEX by TTH, version 1.96.
On 15 Sep 1999, 16:47. + diff --git a/GDT/1.4.pdf b/GDT/1.4.pdf new file mode 100644 index 0000000000000000000000000000000000000000..89745255f81eb6c63b52c3559a089486a14f03d5 GIT binary patch literal 291067 zcma&NLy#~^m_=E(ZQHhO+qP}nwr$%s|FUh{Rr6+IcHN84RW5UpnQ>0s`;jV$h|x0A zu|kn9ERU{3F%vKl*c(|v@$f*=%b41kyI2q~u`@Cg{C^9IUd+|FAzfuRh7T3;BpHau{M$2??SgSP9_Tf*#z4mC*c5S+7Q5h2v;XRq6RK(Qst1>;JEP`nYUm_kq2(twFBquO29I;=M^mT|#rXG?dD z|G}(P6V))W3gt}<5rOufDpHl~ly-3`bCQDj=%rQLHD|Kl#tv};XnQZ^8WsF2bh!x^ zO~CMm-Al6YV<~gra(LZ|^6hFT)HMSQZc=C^ogbxS3nX#C>3BxlVel`5FEjUe0c<`N z!&-bfs%M7iq`XMf5Rv3L3qe|V=#-i|6&$gBu6?!%@?d{NysvT97~jJn`#GH_!}BXA zao_J?G~aq)p!^=2qSoe~CzfO-a^*ZbzTIBb=5GJryPqVm1uQJi=zwCOi0Whmy6M)D z0XzjK2HNjd-h-kdiV&@)sAdIG3_t=x3#OgI5D7ecV>!4Vb|uL>&#U-4Jl}PVw$cLF zI{^Cg*X=8-h-8i}5PN{8jQ&GK6G5vTt*n>FllF6xL*y1W7xkiZ?r;PR;GXwUO~DxP2M8{_@T&8jRlGt%Aq0UUaz2VM-nLsy zHx9fJHvOc{@V;l}F&IXM{^8dEU{{|M8}?l-&$=72R_`Gq7FNFqcA&jt4L_7Gj=w9x z$lj`Y7e&yTvF^qmK?o8q_@CCqbV!zUKAN9VQ$#~CdCmh zWO4<0tKk4kPQE<29SXrjltHwC0}AReQaNJP5G~K9t%C?0y+p}^O98V<M^H%_aY_lr5Xr)`NeOc+Gz_e!!0}I$1vNFZik740a~flgB1BeIZ=8T2 z1s<0i*|`g##dwJYj$JKH^*gPrax$R`9W$(d4oN6yReKw&uQ`e=B^BO~h#~rWJ`OOr zvi!>@>jJ&WoH2;sc$aNJ&L&(?zp|}qi@mhny4uMqRRE+|E?E2xy)?{t0Gd~VaKj~0 zd*i&`Yw(ju*MNcS(tM(LDa+jepAf#F_f}ZPD2S%vmFi0Kch{UH?9<%dS7Cj0;Cef; zz0I)oZZSn9@X*$0V)5SFYov7>Ps@>d2guUOTj7|i$Pbq0f^^Ru6gC6*fuHCh8~+?( z=AfWj{Ns`!f2P`6e5a|cHiY1!0lt8Tk#8m|4ICPR3nzfTNFDEs0KxIyz2(EU(F!(c z9j`cAx0=Qtk6(PNT9x_0QrJ=1uo4K!Z4E&*wj^jposn-BI0Joh)4ZSxRWyk~^L@@* zAcV_>a9@VTDqm^Q8s&*usUTHAsjdNVD?bL--P0KP!V_K`ME7)S7``d7tv}C#ss`jr zOt2iAGYnAWuk~3=J#nO?m84~XXGEtYdjc6A(?u+5OJ$JCqnIIqOblVzri~dLic*iK zpG`b)0)Gd{3&DQo(=Y->im|h!$QSjUnb2TLbg=}Ce|7%h`gDqCdKjXG=!7HslLZ{( zbvl+@)ZCU&R+4gB&--j8L=`F3QvjR-mGX^&ZNEkg+SlSZUzQl`SE`&3G9HuMmS9RU z!8AQK+GR2SlU@z@t+cxo%wlFR?mdwT(h&=MzPfH@xauYW_%R(7L<&Kn?p|Ela4v6& zZ6(WF!$Lhl(OK_Spb`nmA7B_NP(j4eRz0>TJ7DBusUw6X5~!M^=QbR#rd)j*a39-e zS{KpcoKJ_*@qB^=r~mc*3O#E`)wx7CB!WY=xLqa^pg!UT>2Ny9J{IcU7SU zADDlUUE^C%Tkq@p5aFS5L3#jd(KPurv2%)CYrpBNyLZDy546O}`$lUfw!THx*4G>37B(xkJ6wQU#w$l>kgy8pcYZaaIy6Uv(-OyAkis&F^iec!1mydjOGxN;2LW=nFSdE*Rwv} zhu5I<@qnbC7*_4fGdTqofUMnR#}u=JDV*^yfp@jzH5I+MHEE0$?9j1lp7p8m$eff^ z6{9e7AflXre^o%riq~kT&)9jX!kRa4e!dYr&}b1Nc9$)!p1Uuyuytn?FEY-n&%)+n zPL{-4J#J1hfNnYE!qF_gZ!!{<-mi0hU}-eI$p(4{O@ebmgKilL*yqwA%o90bQPh|| zrI4@yO0|2bdoVsd9LAd~BNuL}OhIGf-YNT?DGfo}j@#&B@qNTt%e^8uMQ~t&@;-Gc zJOMRtYl_j#4|px<4l&@{bTs_S*=i}Nvc%e_(}3~95@@V$D-Q@a_p$s9+KnQYsAS zz%2OwhAbce{fT*8RBE&eO+Q-eOTNzmKu!xmgWz|%mPv*H{770R4)l0mTZXrdcnKRc zHS|9R|MzIN;tBILT2%u@7u+N7?$L7dA-JIA5PcA1P*!h^5(iB%RmLllsLD;HQi)F2 zT;=b@y&hK>-I=~*PIsq2lH_6cgHh1M=Y7GaW^nk4@N$^i)_Uly z)|BEj7Hi1FRvX+(W>E_!h?|$E>E5E6%l9jMgU}+6z}J2=0%&hCIzyV&(6jwt5|{8z zCv=|;5E{=bb#qGHDjYi#jmygZf@J5;NW%TG9k|2szoDf4lPt8SbKQ1(%+IEF+0j@6 zk7I7)^LbrmA$gMX^dID9*SrsfJ`G_;e}-CLuU}_;U*k`3pMkYRD7#tUh=>p9?ObQX+M<7F8~6~?X+Z>aKK0nFR94yY=D}#b|;axSH-lOcbW=R z>0HDPt2|3DclFG?AC<>zeE0cZczp<*Rwz?DlmDMl|J(h)e9OrCzkn_aEBpTybZa%G zlaD%J`{wJu0RrIxeukgIzvwL;(Z>x*MV#JzdZ?4Six3P|O5EpfcKCyxe~8?_TBv|a`^6eF zN_KCK3N|7s65f~oRd7DFsP#?M_3M?eNmK;=o|{!9QWRIZYQwv^c5FpUf%HpYWyNTqPgt%E*r>*FN9s&--^q-@ZDcnQ6-9sR!q0c!x5Ms;^UBx{&d$m~-9>!vX}rUaZG zih$^=kh{VPEhc$X`fDVO8K#l79Fx22B!L(F)-?Yk7iGx%sV`&0A6v zKHpjym@NcimOeIk1C`4G0O_w!KTlzhRQ}7?u(L`)Xcvha%r!`>xb7%D#UVM&RGR}} zpfkKWDT#KPxHei|oBOC>qKD|f9K5kf5j#ZW|Hr3WFaHPc7pJS=AC9k}f)^lYk0k=h z?Iqj#<8sbc%N{)hVV8Hd0<5pJysF4<>2)4YIlI#ISfFzEIGY;o3G8i(e(*azx&|Xktu7LL_~vJf6+#^@cz!z`;fv*P$K7LZCZK3Y-^3@_`PTHjNz;Mib7!d} z3hY-)byBM`*ya=3z?B+N`jCYqK}D7~Hi=@2;CMhc(Wn(i9B<#=SI`jdlI0~C``^X5cRQjk|7;L{ zF>DZ}<8Gn}onCvs#r_Jo3?L7lBqHW?LSX?qmWXI<&xR5+%bj(OSxHa)s|kaid(M&t zo3ps?72}dcg?j-s$(C8A8kR!3H95UBY#O3fLf>;^j%MPgZ#5mtHlm)xf9@$O zY$j9^V5Ez*?L5Eh4dgv3NzmBQS^uNRh6_Pz=2Hgin24&z37&DE7||sKp^YS$V(nZ| z-68$?1ryWeJzX_*BhR3mzFR_HKsX|UKord(t@dnj!?VWF?%-IWHUb_*;G(p6(dx72 zBY1HdKfxD_JbghxQ}2yT42Q;vA?BGOe1ABvWdE>RoILm0vb}eEe&HAt zxF@Wyp2}V#;uhWdD%f9)@kNaB1X#x9@XMwknc=#A(>~X8d0}72Mx7FXtsux9xa2cw z|9ZHBnU3^sMJUrg7ZjC%fHJQotDV(4kx-g5ol22WgEcd=O@(1XE=h-@mcZvCZAfsc zZPY#&EKaYOB_giEMPqxJ2*3y?iZKqJ@S_3bz?oqztI`rR;HI-f(A@DY3Vr)d(k1G( z{J%`#LStE|avHYz0y&*b7&%ArVcXB*Y8yuuyY z$B4I%z?P_Ire`-tZdNo?q!9aAefK*y#hL{`u?Y!q^8kj?#vGJXcsaNIl8UKg&$4p* ziU=hlX`UyO(!j}Sqb+rl+1sANQ%%-N{$&h=O)FfJyq)r60tn>VI;JSu1>G5ce zj!0HE@F9!!Df{h~LH0&2F!XaWNkSf?c6l|bAU#=p8Z4B@r5U>T;&i{F>|@?I9^8d~ zKE*jn?~*H!(DIM9cq>Z|#A6kwrLGjb*cw_RGRgbtOqz9GDMM?(57P_Rb8zriVg){| z6QIk$7MfovpRUhTXCrTrSE)7P8`!BdDG7n%yjbI8&w599XPv{uazVk1L?KOBLd(8V zQ%{`+X!PSDjp)brE8kxW+|Gc_0zA_6MSrK_h6DtI1|AYPXhl+kdg9s<01%WYL5cmZ z%>n8t8Y4}n2%scE2LECo%LCua1|7surPd*Jl5-5cvV(3SjPYdSH8IFW3g>SdgOp%` zB(Pn}6A^_SuW;eySYVdy*laBf#VjPAA@^J(06^(5Vf8Ehx8fe7Un)=0axvyjCZeVT zxTHf%HQ+e{7~IXoF1v9@Ine_3`EJI+T|P2kpIv*P4P+1-xNq z4il~E%#4dMJcuunOU8k+M}?tGy}MNRGKBD@WI;fo-Wvpn9zh4FVu#WV+|EwEIDFvP z-wX=I{%&409NK=3*YImt(W`TwdhN?Zk-mkgV-%-p&?}q=w!9doMJb^4C3!!=GiQUN zDQb}PGnceAn*6MVf?nXc0Y=-GVVlP@blxw4VqcKe$1U-qBzimvGJd8%-A*s3XliqY z?I%IJOP;Pa$MO03fh|Ue0CEQt@!Ls|iDb~>kGg#UVVpTr)TW&%O2q|^O#!6C@0lga zk{yks?Ssl>3o--GPiN`@9rZ#zU&Qpj_6GR^uTcX*)_)^-1q~7$Z6jjQh8&HrDrp}? zk;aD5OU0dkhJSd!!;l$KPd+F2f7B{Kz z4KfqA@IL;ZliS;mkB`?D-dibERo7Cy?DB5iY*ee2Ek zgW43<{V_KpJ%uIB$Qnp1K2YyV@!H%|(lZ2Y-qdXe_WhUR1ZmQ~!Ey^#|*+18a-z z8tTGjG%cf0;6Mq<)}TzS$fLdERbyk_V|1@sMGM@t*M75@9}l5zjSVKNaOmlrfy1!f z0{%l1jRzrci?;m$zP{);4uX6Bzrz(Xnnf#XHjJ^nj(;+2`nQt>Gf!;Gdm6$-ksLU2v2Ft>lPj0gS%71+QFxQ_q_4ct(o~pqJ!hF; zEV(j^Ck0oS$lh!FTzIzI&g*=18Jc*)x5)cjfNi*tCG|T^H!Vm{{u2DmZ0ykBmV6;$Q+dhO|Nni%TE*^zo}ushAX3C zCk@hN6j@+xH%&$rxr5ysqqds!HGu$wHxveq51Ek|XBK~BpW$0{IPqlG-g+r;BvLwQ z;Xl5O4QRuFMKH7)O1x+j#j%)8lv|bOD1bBRNk>ovhlT+MpYCq(zTo_K1^{Q<)uZ>y zVqw<^w|nPC$}}|@U^aE$G4IC+oO0L3z#wI$DkXXI>B6MZ*8R`mv}C_m6|^`j=|HmR$aQJ#NM zlE+T?^r<6u1@0BKd{OD9q=_W6hV#J?jwOn`CB6)6#4Aulfh#}Q$PpUIWN5jk2Kpex zuPG&!shaK+9-rUQo5jwW7=#Hw+O;Vwh!K2n9qKlBGYkezA?(|O#WC!aE<9WuJa{f` z1{AB=APJH{RFW`+^X(xq1ay&xi)&g=DYeAD3lCtNr7I%lUUpa$`fi+RC?g2 z+=-n)=l~}Z_R63Q*PMCsh)%OS$Mc{ao+WT+2_tYRRIpj_(Yd#bA)QhJe9w(l(9;|{H`cF_o;sTdLW-K+M{V(Q$kVnDGm@xpN7@VaC zcw}&caE+&cb->|y#B@h6g84n=xX1%YQU5y@2=JIJ5_1+uZ^v%kFIFFD@c5YoB4U38 zfYw4hNYvmJ6J+*n_ewMo)Z8Ht#D}^xH|~cPu@$O$5oHbtd&&XP9Lvr_1`v^RLv|wg zfPO>hqB7~0>9e-ng$#HJOa22Nih^i8u&IfcY;Zk&08#6>xx@fKy7`(ZH9{BVyQ(;Y z2_wQL>J=CHA)zrEJ-8=P#zW&Y=tH3Fv?q9I%z2oHLfx{f?pO&E#x=)GiVievozXWf zCT=H^Iezu4@GHwcwz%^n7L4%tZxD}3if7lX5Dh7QQq?FW>&ZKuI2beF&+8P9n*-ew z9K4pYz|G8(`g|y-1ehVw$WxDC5ZNV;slUQ*ylH5bYTFvNyybRiua``{CdVxuGBsIa zf~0ajefye(M}SoshvF-a5~!1_pkyD+mX}1(Vf7Kwc96LT0zb3irF*`}vLx8{IvnaJ zPfkL;hg`o~q9GwuM-?XizK~kWjU8s7+DqElbPr0UK{83F0=n%mfY|%rr0(?^|67nq=DYC|xlFj>^dt*%XF)Uc&SNY_yAfON83Q=U>p#C06MTFnTz+#&8hd7&8ljA!mRU9-T5Bmom`?l@62 z47pz~7i?4fXNbGL0abIECp1JKlA7VAsX2M`69S}|Ovw1*YfAE#Nm|wrggI4;p@GQ8 zLrPA!QMAk@qaSXtt##4lq>0+-VrKG37p^!2`UN)W8ZsPxlwc%EFakb>Jk_SMwuG%G z)p7*jO_VR6iV?cbSq=^CbrH|h&Z)k@MA6gT%3&y$tuYX^OpCHa3!#62yM(z^-C#e! z!naRu>>@W8g>;@#s}XY1PjDZ}-w!78wq!l62TSye}8ZH?A zW88AVN7}%k`w6j{{m7(ZV0k(w>pTXsTtOm~gZBa~dbgp_#=Pd`NjVCMf?~WUp;80d zPQ_?yOZqgdAx9-7H@Ac%0R#5bZ9q+;3)UtOt~1V|IUwB=fvG0$l%Ql@t?%-IX?Ig84Y4^?)dur6V-apGEm6okt8e;?3EO)CBf6{TES zJkE1SWzo1GuOGKCaEjdBRi-hdoH5O+R5gH8h{MFY{Mll}&dRSEl-F}70FVjtw{8#M zT*_lD4iL;S!HN>OzK4?3@q+?&kiyECEFwPkXU0MhVrnS8WbnYtp)|&*OH$n3$6HFf zZzpb*4uSPZ7-kGD6?pXoDwLTeUkUQUCq-iiD4HZoDzY9REpxn@S^TPigr8zEgEM-p z?jqc3iw7h@#_2qA!VK<&xNCy23eQeDcT5O6T=O2Upu%)1S{?ObOx#ySOYBKiv~q8^ zW;cr`Qb5p1-^=OR(J@4c4iGu^vDqB_yrYwsn!u2~5g{n}K%&)g#DAHk%W(eOm zb<9gFEGW2%4U7j&YI){Fq|xX|DeWQBZ`N{pd6hv_gyk?aBO~VC6Uizv0+$?YU`xYs zD)rHwXO3vK`wF40UkG!M^;lRE6HqsoiA^@{Ezrd*2ywav`K?z zb?$d~SiS#`Z1(n$Naj#xhd1Xg>bQ=giGk{0clhb8zM!Xn)PE+z2lrOa0tdFXt|t%1 zC5pzgcP3eoyAofEdH45oPdwCh%fteSGg!|3Kj(0S#|{{7XSTTm z!}jXd7^~m?18!T;$!udnhjFkCp;LME{tb?czQwu^2bYTU=s{3b$d+mMV|#lm=D5vX zr}T&rd*kuRSNUdKdI*X%S95lkZLG3r$0n$cdoR*|u>@k6Lw8%qfcIh|81Kd@qQ#Zu z9YlkO=;7s&)9?=Akx++F{R2~d2VzKoNf`eHZ-Z;z-&|qJud=>bLI$><Q0pl7?R+Y z6Pi4r6J=x05gH~{FnDy_Avf2JfV|-50@$eHfYlaQ_&ZWKJ6d=;0pU60k=`_`! z%`eM+QzM1=Wrgjm?vfRoY35A}kdcl3mKiG=9l^^Sos?irSOYA^AAUOzrOR{YiXpu4 z8Uuql*_(q{ctU=W`c!Rxlyb!kbx;G69e8)m5OSdWtiR`(F|;k~bcWD43){M)s0DWi z=Iu2;==hK&zsU7$wLnP*rED(gfrb;e_rwEukG@CgF*|O3#2G9h5kZ- zLJ?cj#|wZfvlGU(H`gNIV|ckR*gxBiB;1%7=ON_%Z?^?!hHe97pf+APiQf=L03s-< z2pWa`A`o0kAq6YIGdK4`_vNWP40v&Y0|_%eV#YT3v=dv_9u#Mmh^NrR>oKV-uUZxH zjhmU-$@$L4c~TDdV6h(dtC2278^-p4lH4joL-hkD3`YR>Pc@^eGDI;DqZ)waiKXg09Sy zVy5y9rYVl~6v6&FG&XTg6L2Oc4H`deCQ$|>?D7H$KO^52B}Kx4b8l~4?OHKFFtE-( ziK5|8tP5maHnXa#( zA$pn^l6DrZ^E?!TpU6512_5e!&vwrQCZNAu1_fB(+{POXT9FRO`C`PK$EN7WJhFPN zj%a8l0yMzh1A=XJE3VAvzqy1hBZ|sr8 zeNxv`Q47G?Z!clYmAIb~Q|YlJy}QSwZ|1v#j4|e?)z_J?M#B! z523eE6f?-7;e)87ry{joM-eI~rys*@Nfnc}bW@3>Hmb2^MfXrsm=62D;fQ_3tc z&OUzg#-l;7Q6YX+ql7jY17!N+Wf1oj{%Yr~9UzdbkKpPnqy(#$zebOM?FGd-liG-A zlCFQ)fAL%{qt(Cu3b}i_I3SBD-jpL+V-hew9+vVki5mMzwQ11RLmWMe3s^Nwm>)x7 zdyhjY*FX#dF+y8X9_Jm~z>pN{kltuF{8}>Wus0Z-XUQJ*VuW#bl67BSWOCLV(x|PzP zWxh9G$7~lmbP);(pas5mQBS`^R^wbr$o^X`vfZ%tq|DxMVHTt0;fZ-2w zO;XpbsTtuvt~>C1=F~eDdttLzb^O<;xN+JQFJi~>U;O#+89L}}2@+%hM*DxhP^3Nf zAz8yy-k(HG>5JkVr@VJCBEsj93?)T=CBQ{f)wd`V6*C3a$RC&Gc|@MVhv5}Fq7^e# z0VeBG`Hny`x0#aS2%H-H+j7`>cvlToPgb{fgAvXmfM(-`S0-$ZTKfPdF2&L%{nXm2 zkVKVEka`#7>V|cq&HyEr$t5$#KZ^{hMRHA|g{P=b|5lRN=wqv8b-QkAvBPih0R`N8 zI)?_Alyhb7VI!s&-Xd=XOFjR-)UP%9XaloZi|fg7T!*nV?3tqDO|rcN96*_rjLYji z&t+C!IrZab*dM)j@6!WoYF$|MI3m)5dFZOG!Zv1r!m0wWd$>T`({6Ol92C55dxR5Q>uvMAz3zkt7N!xE+-Pgx5VtphM}bOGs3Z-#H{@?eaV% zmeP>~sq`tx)CYU%xxPa zxVEQE2etRiDoZq>30Bc>{(xBzCeYTn&h_6C_)&cT{Vx6zacAY5Q}K#ddT2VTL1;*` zynFSf?HWcP+MYQzq6ui9`WaMHRfQ!FtaJsfbIXDlBK&`PzX1h|5rgUe}5B z_BhyMOAptGHWtiAm5Ma>EL7#-R-hO;7A@mfXoX|jvE&`-_(}L;lr=!Pc*F`tf0G6f z4~T#!rEw`(eX7F@X~?t&pAT|){*HsM1yjot$%rQt8~IRLsZCT?3t6=mpPnQ z^Yi(<;0rtV9{ew^#PUB~iHYI=-bAm}kd8mV+uV=*HMW)myXYrYT;vH<`Yq_ZmE z#*Phw6Kw~*Fq}y7=JI%^Oh1(*-rXGDt1)VnSVN_T+W&|j|6iiPIhU6LHO;ajKmMd# zm;cBETg>Cfr6&Ie1J~`3`v)atMOjkRKrvk&Av)8+Jz|Q7@vDRF+?3u|c|8#0%wGdi z)z!MA@rDbpPyb?B9=$&_3~jFtvLgrTYo584|8YglCr)Dv{B-p}NK_ZhB<7!VcQ|u*)(*nlsaTM4&g(ha!jn^-^@EU9y*PRhp z*~bl8fBT(d>c|WOR_ncmN(G>rOa%Swl*at7j7h$1n|>Iz>0@WY{|ZL+w$cdc2~Hcm z%0~NLElvANI=_FvEH68HY~GdtPX7Y?((+$Vp4bw)oqLu8Su$$>*}-KE5Q$#LNJLVU}RrquHOZFdrdW6Mw^RnYephkK&rK_t3dCbqgh zRv+A*KuG_3VtHN>EB5u6u`C81lTg>k;HT#BczLAj+m71p8&(Wlwi zcpF_JskUuh;vnOW5>T@27u7w$hN#TL)An8ilUSJVlqOHRh+u2eO} zDJ#|bT6O7YSpY;);hBQ-(16rld=&cX2Acxs{`Vv)kgcsBT{aOxS?72d)#Ha_UB~-o zN=XLsoniUxty3>M%R?T9+6Y1*H_zRX*Wbis_QO_WcCt3{MUliC*pW2)Qxk(Lo`=9M z-WTN+=!bE8*?lT|NE(@z4y|?0BI7)&^&P0JO8RX=N(I}nOa#oUH>B%k=5d56lLAb_ z+0d)70}YzC@J)3_Epbt+FbS;?I$_=js461?PdZ z5=w;$y~t&x3W}||E2Lc%kMog6S5NYIi?V6&=#1PA6mkwml_Knz8zi!*0F;Q zad{u9&l>IO{9D8VK@_d1@jD>OPADF=xr^oO#xELB`!awIvBX;LVpF;a2T2b6{wloSx1Ycm|K>wvB zPD1_2=z9bvbO`xy*ZyQMqbEE$sz<7Jq{JynpCWS0m^tf9H|N_&*)crL=1E4jRuAfN zIPYwRa(3_)6vtDoh+F|37J1(S$KN3!?3_{Er=$S7pSqQU z=)KHEA!(adyvFWA4jN2Ha9}ci4Lp(z4MTwGco@C|*^Ugbp|zJ@7@w+bmLodITjLE+pWL+ z^2V_G{R2BRsZ*R_g5}i{0IUlX%V3V5#&rU0Q}~7na~$7w=k%HTyn}8Gh0+~8N%~a_ zeCYKBD_iXcUH*eb{8S&n(a9!t2&HO#P8S#}0PLST1Wi+DhurFp>}vv=AFDr@Pd!7J zq7A|FZk`>V1!I*S7D0W(*|#bPR=-%4+M`L;Nc`}J(`o)LjU0Ej{FB>-34_FK z;83&wUEjj=&5xjjF#|t4%ifLk_q-jSx_jE*&ep%_=c5Z4#V}8Be#jDIEJ#Kv)%SJ# ze*2dWbu;t7z{3Ah@?c_OV*S6CJaRQOosZa%{NCynW~Y@@*~0c^dN{@>9x*?By8qyVwz#q%R!VEsM4o3f#Q_Y9NXyd*2 zhtr)qehkANDW}q-?;Kz2p3XAM+zKA~^w6E-#wh-p`bU!Z-Tt#z@XRAcYkZq&UH5rg zJZESuhcmJT)1q!>U^>_|inBBWQD_Tg)zqS?n_Z~8%$0!BRje2RnU*usM9aFzA_$wZ zJ-R6=mCC1q5H#_+==gSYf1Q1#2dR1&(6ot{su*gEtCh0u_mg&y)=)`@P^l}c7m1o3 zDpgY2jXFNv@0-Ib(0KI*e%+uNB~^mxG*Nu|8-hwzC3)<7Ka1PA-MFi2mU!X@;%{v{ zo12b6&9eQBN!j@uiIG_Q9E5>~9s{&|PeF66ZG|N4sHWpoe%x^HX3aC!TD7x8OR!sy zE!+sc{r(4WtDBd9W-fUZEfm9UK?=Jy$Go-?2N zZD4+N;8DL0Nl7`ppp!di1n&uX!K4)XYyKwux>6pBSm>u-QV zX)ZZ8DeZ2X$LZ|jJkF-U){1GJ&hS8Xe)PGCKrgVpqB#}YKll~d^PoEdBEIkW`wvvn z)f0TUn`TQMc%iVSj*)|IHtjNZxeOr|BfoW?4ZKfmZ!(X9*BtVSBpe4C62gd#Ou|d{ zn@4OU2+DFp=+_bu5}bk74t5X+@x>RuNBL)y>4tc=VUV{r{PN0yZ7lj1N~B!#3qcJA zX@RgO2`dhGmaw-!V}s>b3g%nSV@Fw}@86yw5isRZdEP+BC0eGILr#6duG2Kv&>hju z%oMr%*i4p^cm_T?YV`tnrEEWK8yF$qpc}5WO&&>qo{}Dg0Q*4WsM9>SA!k`us(B?p z-}$AP<~{J>YA##;E)CB z_hz1Gf|O1g`lO~N2Rhyw#^t_x>XIN*rUfbINhb-b@*Q-3Esj4CLOl?S+cH+)S3a>j z3pSl=nZY$=ARLm0$GGvk5LfYL_h0-6dC5cokee`lCfR_PFs32FH9<=0!IJ3;;mgV3 zeDqw91EjIdlhW=bEyB8!5FIk+4(~I-CKcy4L4#v0*~A$GQFbbHBo@XfW!s1Ue+R;xDy;+7ukt!z`C2TEn)oT{#|L%O6vlU4LbSU2~B7N?2$MX90u&5 zrhQp&kg+N%++WXbXeJ_}-Gw4PpZaM7K%?b#f|6 z>1U+_Al9=#&cJVo6Wp9BjDU>`nfsSa1E^@cJo23Cl5>H2KKc?=fa#efltc+rWaM5G z(9wozt3D&M{5(&`(;USm%LcjKj3Edy%!Aw+7zADGD7<3X;ylZu%Mg3dyg;y);39V! z9puMG4dzhPpfm(XOx?Uw2(u{F(hLH))Qxab>K)b1@bk{7jhJsjK!I+?gd-riUWuSQFh>#siJ$|uCT7R>ioYS$Z<;sr4HSoaE_}J z#E+{gTGID5AESX7a(2_sv5515B-8yKGC|b>&K8xIfHaJ}X}FeyYsT(fcllHyMrwp8 z&DJhjIdzK*d&%(z+mEcoGz(0~gOGlEc-@Fa%4K&Zp1XU6tk`1fEfAyVBD&)WoRjO6v`;mRfhAJ2z%Z8Q!kYl)WN zj}(tR5@N@vZ&Y8aN2;L#zxq3mE!&TNS-ePo7hL0#%Bzcx(|1uU9Q&O5qWw#it;lel zM^MoRV9G7Jbi33GO?h-0MEbVr(&a@eqMWI67X4hD)c%IGZJ5D9*lTSNPXQ?ogN9i8 zjaFXxCwO{I;;4EQOseSgO{w@O|3Ww$s~-Xb7#hwnyT*Yxz*pyD^SlG@mm=e7oQh(P z*-Z-0fIriAAp?j`i}*sDBgWoDOK>-j0u~~D1#aYAylk`x=b(2(cEy-#s#HKvuxu+K zOpp5krABBa|0e)3|8L1MEA#((ZogJbI(~Z$vG1-P<9r0Lh2=|>4G|>5hBF`-K<=8= zslj(2u9G{2<2d5jqVAKEORb5@(SCw~fcgN?+z1U4^wA0zk{jFj8 znQmGq$HdHC{njt`^Rn@Lx@vn5tG3uE;qE{ocl@xgvcu2vDpA_jv8rTAc+J_6XH^Jm ztccdIuU&(6QJELdR0vav*YRNa#@-bz^~FB2Dcq>0TdmB4Yc-sgWpbdr%ks5<@5ea( z+AcM=RyOm_8!P_0l+EhWu!OdRu1dSCmpi{xT6;A=^BLEFEY2HhO90YlyaP#sMq5h8 zngSabW?^XWr(=!I>uu>P->}BL<#k}Rkeg6=SFHS}Hma*vdI+wO;|7VVz$8LB-bbZn zc&p)`Rq+t^Y^aedU046UAxB1OlD)g76kT4W`HRBKb~;=1Geek~sE{&(I+N%i{yKn@ zlnobmEM3+Q?Lzi#?`U?^iujBvf$nMsPR2t-&69EB!0g$2+C^Wgx*CL+obncBSJq zoKhbUk8Wn?hxc>FeH}8C)1|I;x?mV)>ecCn6RjMx1-^K`<~Tb`JJu zqQ(+`bE5vysj|R<$@ZdzC#fmdg?`sF>Jtor!n3yuwTymgRkK)Srj0y_wt?6MMgI_? zkQ_qN=E1`5K$9J?47OBEmht9R%ikEK3B9%f;()HjQoE(S)%N){u*i1?GBrlfGR!JH z(FluCpFS}`CKVRs*sxqm_h2@^YUGv|0uv6hGqX`@L<#~z z+CVXTG(fW2#TsRTlzG*_IfEQxRi-jRugZgibKsG`lRz4uV>ySCc4uKQY<+iOZ2olZ z%7o?=7u2+xH+QAUSZz;5%@lja9$P1t@%94~%RW6D%}_!db~@UQmyW@nNZj>dAgwrT zUeh}h085QFRDqc|#O-3GY6en?fiWLXV?KSK{2{*|@x(z1522H73y1MNOn|zYS?zPg zuI7pzn+oD=cZ6uT*KnY`i;bgAd0X?YO!pK&86Ko9Q?eKV=awiLg1kA`0YtX>2*xdY zAo9^Wg%?ANrI|RWX{N8^mPMTgbz{B&X?i^W5%MP(+MfI>IO=vH*PykF63cte%<*hC zd)nQ!kGQvdV=&w%sD!MgZ%yls-9EDyu$cS&u-}kLBkyrNR_0v+){uW;Kc)~Njn!My zSR5=TCY%Y~;PiX4*%93oqEQPFpm$d#HSkx#KJthA`cGs|*t>tps0lnToV~@af`IGl zpvpR%H_W&9dMV^xugX#|7Ka>+8Y^{+n7mvtUzz5rGQnlJ0mN@uy)O8_BgXow`PH=> z;`9gmB#S`Us>~|9k$o^jdY!)RIpB5gf_vVi*^t`c&^B0USV0}uSs+jVQ2G)0t9pJJY>l8dVO5V{*{(RBz#7aM(Mvfj z`oc!khEYmRy%hL&e1#JfgHx#4J(c~)5r0rnmT-0Y7t3P@H<&pwPn&wPrd>=rjc{~w z+2Qe`F%3m%HY~fuDbg|l$0rXF3Q=#vY&l#QW1si^to6IAwkd&|9vbj7a$PwWTAhOqMjAKCS{=dGRx>h}{N?E*yYSPP?ZC z7}vW!tAL5=v_?ALMz=>OGN>prED1`2M&IA*R6T&oxl!&rTTPw?1H~s=<=O;Qd$3`^ z)a3l3;ZyIwBowfbpf@_Ye8{SKl}rgHO{`0=whEnzKp%lok`#{2&_>zF2qMVzbRqO! zbVzuWhU)7-<94S3R>yoWRT7~(+%Q{D3SEK_rJ+w`+|}r!jeK?QnkQpw|6cn!f+~+~ z0UU_pK@(bfj*&=~kB&85MRVP}>wPtO6Os1=oly5Jh+HMT=_LiD!O|;jU&Wn6iW!dC zTm^=w=@aO3HS&Zt8WjblnXY7R$+v=oG$7*%wGcRdTz8_#^|Dj|WiA=)$g)FvjD^qs zz;JFn-a%ZUvW-Hg-lcRcW}utOIY8HqPE7ACTDo^leIUPfX19*!TDf(Ca_EinF^f*P zoBh;ZcmB*i6Mhrfx%2J&`uN}vuw#QVkMXVnTb|ifpTlwIz}fN2PNFW0jmDPe z<3!Gxt&}D09RaY)Q?e9L!aF&UnO)9cJN7Z7i%Hkd(67D8Zm1)~4u>lhUMEALRn9LO z$V_Ze3e&VfVb0SKG2lDZKd5wZ3~l9ioYqcF`GE$PqqCYnkv4UjSyK%pNIy@q$dzHT zi}3KPW!6MzlL z`qx2cb+6|!qH7RIz?b))ZE1T)$42E-(h_3(yb5?pNB71>9#p?MpM@NqW#D{KW@n0T6Y#^iyo94C>uT|1{SaL2X>ex*+Bl0kaw{w{2y zCJZ5e@obBnKsoAdt2yVX8qvFkzqiZ-`Y(7SZYj z(~aHfGtP5CnTV3W!-gYLc$l{CboYEJRr?NN2|3~dG5b~mb6Mn6NzM~UEt8Hx;#odC)GwpFDzRvm3vz%UPNzz( z4M#9cRV}$bDN3FYWD2p4MWReFQQtm=$ve`oyOODMZ==w*EI&4U#ahyE2+g3~)byC} zo1-p34h#YrA2_sWOx-|I)uz2ThGe6ca?Y(O!oWE8Bdinawd~(lJ*%URoHcfc$7iMUhFbj~G0-~>_+l}M@-i2-Pye|m2|u{(=%clp1pFZA48 z5oHh8VBCb{aV;JP1`-0`t69~U5~;F+L|63gTeziLbYSmvLSJVGw?8Uno@jnCMfg0U z@c{f2&#FIDMzXqxqpwDrMe$ShtiipBa^RFBnTuh!3cLDTWm38%JJLBMHDBf~DJQr< z(3TNlz;qAiAc98&?Wsg`%&v&m=;nDsQ1DcC+ccu z0y1p+9bb|j1E3*Hq4;Tu^LMGuCvMBiFpd@!JAq8S`PZbaJv~<9rMP~j9ZSX6o$7I@GYss)ptiR-t{vI}{t@N_;;YbvMeJ2sid4lid(iAWvuvKLm4}K4ADmT1yAau;@_KeXxlop`pdw+QHe=}TNz<#m_O!L z9fx$ss%k%h@u63R-v6NOS5HCmHlGSZE@Sd!l2EcdCiSFXN;7CZsuLz4D?wIwKvio6 zG`}(=i4BLcEgSrvXmoFNj?o8Td!9@9IrU=>7lm(vNovA~dFr}aaZ+{PQc!^N7Cf`Jv*{NZE$8`~}k ztVbC?$}!D~iJnal$Nj6QQBZ!XWL^+rzH%|*xG0fr1Sff{yPn`EwEinK_j?S?+hd=< zMDgQ^*Uw;Gd6J??-*DXtji#=yJu3Ofi3&4n%(L<~fOn?HV1)?&)*&X{NdcRoBLVFB z8wT-B$oIdHLyrHH(peZex&GJhc=11S`2SM+OD*I3A^33gV3>p)W7rxgjENH@x`hEk zXOuw8q76~YVf^9UW6ut?nj%8$dwsg}!x^&XOl{=9B1<%>C&yu>SxRa`QXQr|~n~N`^8xEZdY2|Whr}I{duI5e6=!?V1LH^n? z1TUcuMW&wknv6uXIusrnKSHi=e>KNXThE(gV!u*@lsY+EKXzxsaQNi=(zW};#m!iS z|4^=4gu-?;z2SvJvB>eAdDxPlt#9`m(XIWnBFN_4I+)gMFoPZN2fNlrU6UkR6IS)p z+q1&`nR&m0m_+*U0760A1|^I&PY!%Q80YDW2Z0yb3=+Qw%ZyEcp{uUi6UiTfjDh_2 zpmLaRDwy5B^`Y*u7o-eP_x;BaW*48M7bD)xTUoCu-q&wrLDmR$sx&Awf_Jifn-gUDCEO@d8KbGfAgvzc3gv=w>|i0&!7k68pwxnv)QJHp4zzQh{C6;Z`e5@brdZ-HR?i7ZlSbITCZ zZoONmXxh>$*1O&d1@C6e;UF7b$Iqbb%@b?32Kx;V=s`W}4|V6yeTR9m_x;alKMbqz z<-Qe7brI=K*|>$1qfUZ6Of-%`j4#>==N-F1t1J>Nu|cLR{vV#F-C-=VvR8p}#BY&3 z>2^G7l?aUah528}j26#1g*MIwA&%;}UcWO*>Q?@~f?b8+ z^4C=@pHjf4*&HX$9)B#ZlL}$uNd%mDOaP58@)@`6{&gi!lPQXb^!kaYVVFZ*bq3D4 znN-2 zv5p&f?Ze~w7m)@zv4f*6X)QEz^f*FTh2^cs$nJ^;oC|Z|%mxRt5%@qyF3Hm$xu-)d z0t_(F9_iZ`If)@qe0aD(WD-HmiG%VHLI{W2i)ov;y|@yBTBN8Clfo zp8c6G-U-WJ9b>wZO3{TxVAlA^!#&?#^M1W>)w3^WDr&Zw-T`mq^4zB@Mm=8 z|N4Bpe8Y^W594zOA+Wp*AyEX27B9V9ZL3Xj*^BAbYa@XO#ja67xK;>c9>=_)$j)i@ ze}|@kCx&q;B;X_eVIki7P0y$sIU?$%CKtpW+h@mW+yJYU6LE-^+j+RutD{1kO9A!p zN+L=Ou{qDS#1lb**m>mNQ+J;?4;ZNHwuerbcH=AO8CMr|4mE{nvziHsLAt+G*m;U^ zt(!3AAo_#$tQmcPyCnnXj1~v)-?J;N-$knM|Ad(wv zr%wU8?b5x9fBUO&VEb%lhkX;P%0tDrsBthJhV9xJ5^)Fc+W(4LXx3nnC3_?*l6EUI@s3F?QKLBC~f)3>(BGCKr#yUb%RH6LCol->+GUFJl5+ z-yEmYV;CUvji-a}iS9!Lu)>j+0|C7oc%*oS%`;~Acc!+*qY!S8$)dK!p(DX0Q7*BB z3zO60z&*!;+9!a1$HQ}dbtGAa#y8f(wgtOYdD46`@98++8V0;cIX;y6oDno>b> zO*?p6ezW0hzaFE{!HVei3G2htbg$?-G#`vb?Li%M6i*FlvJP5<23o)2P^6>kH9US|Md7NSpeSQ=^L4fsKk+8VICPn-!$quH zS=sN+h~Um$suA~;@l;1S2=7KaRL$lY;52oE;|QmnWG9^Hp>0654CjgdOnaYtXPuWB z)`AX?B9j6ArYz3hd|%Ke{)64powv7n2AEVBTL2P^i3KwgoL|UO^Zdkow&GCLXj^dw z6Jd;`MxPfPB73yl6{wdpX8i;}&yXrfH}CIdq36UujqFIFgL%*b4Td!zLep&?MKTqd-A+83ZU|uSp6K$6%0{t@LLWhpK=Y{~?HjKcZc7t_O(Rln1E|jtI z^bdw2|CeKymiT34U(O@C)R-6lhB!-V$q%z{Q*s6b-r;S?B<@A>}MVhXM znM#)^U(_nHiBS0+7W5~U56~L%q!{T9Amew*H&C}D9noY&qeT~Ne~0yekP}_^kq4wb zHk+Q`8m?+{r7zqPWct2PZ2k;gUa>B#050?@kN|Ucrhr}#XoiM8Dvl}*z@n~j7-WQu z09jpqY+~2|Y*JeLmK#%uTO5O)CbVEj_#tyM`)5H=$k+_)&^D+iN`-YodJZS*svyPC z_GB}y1=?VN$s!+Yj;bF#gl2*yu;S6w3GJxL$f<*6=6;t~gh5H^ImbD*Cv{YdhKev4 zbZ_H}(`F!oCXriEm)T5hm!W_jJ~u7QWVSsQ>MP@$W#veuzSV$x6d95;G$%G9J3M&% zu0b4(;ZV$rjP(BA!`%>Vv_PWkkx_~2Y+{exQimZ|yUVRE)e3iAqt!b4!xX?UfMhS8 z74Msh@UhgadyJRgenuTVzB=$A+Uz-Fi84OA;Jr>{G9>lqr5u?P1jPQWBkid`YD}a@ zyI6;vlvLcric(3^{ukS-^uQMrL!h8`ARopezqs5&9YT9hZpt4s(D%Dsm|5&Mb?fSD zYNq^h>I8Tf^t?a6PZX3wtb^<{(|x8fW{a(r5Jhc{Kf|}3crtKtf#54fGI9dX$R65W z*%K!9M8AC^MYXCgN!f;fG}3m1W1~jrr0o%X_nU%31opRKRfIJSEg2Z`BOh#)vR)eZ zd4c0t>`KEdf{8A1_uWqN>>zg13_3KFLhCePoss11?#8HdQ%cOkA5@+eJRkCb!CAv@ zN#9;czhhg8A0a?gAK6yqMQ0dGATG1^4kE$^1Acf0hD?qSh;`$;|C3;_GX7VB@xSL9 zH8uagW5su+Hf?wuAK}V9h1$X4GLuodSC5Sl;_aRR0=a)}_a7^A3 zx%}XqHQ>*$Zd-33x2G_Kip@a>jS`E7_6?i;GT6qZ)k6<#vG(;%)lB_BWqrG=u^6)^ z-H(wVyx|%@Pp`i>HmXZYQ#KhzeXOglGgF;G)nWOz?QZ88r9KqdBkR`)Bno2^dL@P&93);)iHcUDPY?b2sno)j{Dy>_=_tMPX1A2L>lCDTeQr#`UD z@Xvkg4p0(ZdA@y(L*un?)_sis*uUTLtQ}OxZg`f;u3Y~tWEg)vcW@a0J?T#5c}z7nqzDHji5rPU(s7VE(vCv zh&)X09v3hk9ePFpJG;ff z{QWz))8>svg`iMlhSim~s=dr$#0o#XBgn;b#C$osk0J~H6!p`yYD74Uw;}-IcKMd3 zG@q(FS|J#%_L2+&E4dfPM3tI?X49cY5dAgR;&8KQG?&vb(uVpl+kOE`LM!{>z7% z7VS_QutN+ll-ute^&uT)z*+ZQ+2qW;AMw{mKiMNEx(PBP7;y|9ocv{(gwj=Qe3Wj~(|EWRfH?+sn{0BZx-8Yh=6x+qeZ zK{Sp06a+I!IvWq{9cgm*yT`3Fqwv%e3{!Zfq|*2OvVERFU)QxU>&?M{Z2ateog@>y zT1G(Fs(}-;M%#8!A3P0ZEM^Ig93sB`3a6T;J+!y_Dq0wf%4Lt?XhZ!X-zy46w0#yX z5+XihH%dhG1bpQ?EB5+#vFq;53*& zeSgxF=*)F_IU1$$tbzRR>#g%%SG1FGIX6^gT;8H zoz$K?en!%Zo0D00;=`^Dh@UYo4bhw|`16}F!6BCFo3Y)zOUpe!{8$x7B`!g3@mUM3 z@4ocx8aUZ|RP_Wem`^ABY#*ybJb*th^=-S2pbV|F*h0~qrX65=6Og7k^U0F>mog&t zQ?YO&!#ij*R(Qt!E^$)IWaA#VEqwUrD&bZH9o$B0WVOYg6rQS%&jW8U%LM}0q@)9f zLW#?3N=B0D6_kG^nZ~Ai>c=mU>c5y1x&`VG)8G&up>rl~{Y5&gn2D=)9)~-~AMC|D zcw6KHn5!2KsmBXDTZE#S+Ds7;&}YKTv|#iAOiC!6{lQ320wYY^zvRW2W$<8(^o@de z(s76wXZ=}v_}AI9mzcKPirHbr{Urq*H!j2;9skj+Y8=jS+Ge^Yil|FhIM5|Qe=xFf zaprtw3_VN9r;Q*^gP?4Eo=m_keXf8r*6B9&IM(?<3f0(XNC%dJ$^b(P#+rBhfKQZN zcH9;*AADO$zrT9nHCABS_DT>ULGd!q6jPu;K_eKbXu)}k&>VCQsx(Ji_k=4o{QifP zp2;-J}5^^Sz5;9Tnd{PtNRcAWb<{ra5GKPfY z7|yh;5hl3vDy=Lxv6wV5GtjW+_wH;K0GM2eb7ripcJh|aa0r>jI_g}&p2tc`8Wp%>cxGz(hgM(Y z2X#%VLTJ}Kefm|{6uZQ{ZDx)_E|5PnQdsh95afmk_|ZsomEV!8)V8e?mTz`z10w$b z7vTW(`PH`0rAE_mriBFQ`eOS8TBQJBFi8$rX2$gnQ`9Ugz(UIxdo}V`(}?{$Dq89_ z8gVf^!X`UF9GhGI;y*Mc(RhBY1Mx)R#H-j08OCHasC+`r*tqL|o^`8^+p20n16$ik zLA|uwn`>)(u=}OouB_Q;*BNl$yb7`Eo*}jjSX-9@nXZ40sd2tg$_$A#h3Cknrgw!+^Z8X4D{EUXdqedrl5(M^Rx@<`MN z!qzr1X6?6cXTB8nQ%8#fDB|%yEf>iS24hT3^Da+ixTNuP`H5KI*dsxAF{&Z_Le$?g zCcD!DH`_`~^YB*T&G9cpIQ}t>!4AcQZdPiH=I@h}JhmT*J@^)E@3t}J=?_CkxZYkW z(g2SLpcl}G{Hiv0NMl&I{5}ZWlF&To{9=(JJY_wva~rw~X{JkHcX*1~K&Lqg3bMR| z@73m_qgfUJa}K!B7spw!R;EUYtDF$Nr>sx-pWyyb%n18iAcnZG z?i@VPO~PD&^uA*0 z1}^*)OpAtnGrtNgPMa^@l!TUbNalE{)Tz#(KLICF&OTx7?3_`tU zG5dJGoXb?H*p}tA824oG!3_=X=`P`m9x3D0!q$#(P7m@)Xi>`zcAT$GfjSR?_o51V zYx7|JtXHrBaa#FY_cuBt!nkK>1ij%^X~iq6E$n|WPK8T6&2ES%ad9U9>g+Mo+m{EG zU>qqz%jm3SSyAt%M}9Vx5cf2M!?5)@{k~UdP|eW%pm@Ah^hNJ0?nkG#|8T@9H#vaC zgl3~T-;W!|!WbTdIk+!-e}jVFnWy|V!){((($vtgKwnkt9;iT13@|j1OpWRU7z}Ff z!t&SBONdIbw4&-K;DG;@Sb@^KmyjKvUg}iSG-TeVR!o= zV<7vVE%ecH*7NrU+o(B1-fb-$G#}kf4*2=2TaGB`7u790dD%Ih-BQonu=gjGmnHj; zg)Mz`qvm*WcJjR-3`o%H1THs^EG?LSa{A@C@&Z>=2x4?Niv{m?Y=30_dW=h$Oc`kc zv;ecKe#rv^ht`NtzvTVuv0U)aY53ftzf2^Oct+7c?uqo-h}Q3(yH!4C*;YSj8xz^A zqV^Hmb!HK{a?Y7X!X6kSu^}Z{N2=la+8cRew%hX7Ry7gb4-LM%wTX9eB9|qSHcN4_ z3hlOc;{YL{WCiAT`%JC2-#uOV((9D*Ba%ibZfwz_(w!<+0;!SLx~FAl>ppxrQmDBK z_J|-Af~EOW6nKHnBT?VGVNdm;XoH4;a;AJXihUS^qC_Dc*$~n9Jj~$=H=3|>4Fm*; z5$b_!3*HO)?L zR@C3>{Bs(9|D_Ut1RG}(B`5__ujwg}(DBvXd&Vch&A zov+_ZqWNan-*@)TeXCf4_ftEz1~@XcqVHJ*K%e0VkL=iNX$d<#H-whI|&1F1!0N8 zfabkE87QdhmA=E>iBOKa6Wjq_wk6w5T#Z?^AQ_2b!$GF?SAp6&69rzx)Su04nzpHC z@m=>T63O0xffW`A_{ONt{`3cFfqZ-I`kP*RF{ei%Z+yHnk*XjlAmj;C=R1-#UYW!M z^|7p7H^MxGcR(CC9?ZKMQT8@aQhFHB4ry8Y#*xt`W4bp{bY`IWHGIxZ_3JmHJ^g1d z6$NoS94`do93v#e=4LtFSb?M6Dx(G@z1uUFIHdkuZUU7^pAB8vx{HVUe zL|X4pjRh6fz^*xUMORCC?@8jI)Va@05o8f1k5{F z?pJ(WFR}d_DA~AwBrmhlTo#hfW(BRMWB3w<^Yv3W5|G2W>qebbvunx>#l)c2P;jP|DhlRVdJjOvU*bl6KRKv&}e2j!@zeY0r*xg;P{&S6lKQQf> z3}K}jV)=*Vh9MI)wz5(d9%jrolpkI}6Cr@)F54O8eoCGyqLk+#xD-{Hyg;h#xs1xE$79(b&hmR+-f zIU-Uvqq=1zlfW0b5*gjFP$TFm8|tZT-XY<@eCqsQs>iOoFwnQXICP(0HX@+x*?(ZZrM^o zz)SuDMSYk1V{>PCxo(_}M?R_CJWZ#Uru5!6Gmtr>CT-KVu@<%9gCks%Qe6DElCQv# zbBWF}5D4=;z?}V$lPSNIn^bYVTo1fBv`~sf4QGGgF<56}dG-e0Q11EqbKFSY)gAIt z=)J$4aVu^MeqXOqunkNH*W?>^XPiQ_Bt3_E zT5Q(uvrAJUq(XOGNRBA4sM4I+YBz5HLO0!D>Ug7-OXfye-qOb!1}kWI*5NiaK5mXKlWGF~#86+Pn zhUnf>xqk;FQ5jjCvQbL*aM!*<}dk<89Vbve$2>L$uS+1OS4 z42742g=Ef9gyKY0nP!P9!Welmn;0c+<1J53h#}^@)t_E0Q$nWDCvcR~GBV8TSo@_T( z`>&8vaEm^e68%T^II|gEH0qqY(y>53H13i_Jp6?(PdER{7nb4Bu!!N768ej*uxSf3 z_<&8lA<8Ce21E1JC&Vyb;fj3Z?EXs}6>PMt%-&4B%aY?-vyAQo8rQGW-lLJ8ZtJORF&! zZKXS7y{bD|h6nxYba$e}S?ZnB%Ar$?SQ;v2>{b};NTb*>OhE5}h=w=ug|ZDHYDf;?qEP>BKaZh;TsX<*K$)nOz`5 zXIejbT7m%iF=&$&H!BuQJnU#5`~$)%Et!o1(|YA07-DxHou6%#@zeWK(?_LUkUWsT zkK|vYdM_lJ-BrE8;;#EAn$)+S=m<^{em`ybgF!RtlC=TD>7xNw%DxyPFnK>dv@4BR zdHAUplD6a+Lik|?KKLbWO6vg8k))!jK~!#X&~(j-nVY`Y&N?co48cc6(jt8&b%-P7 zF{*SRJehn;iCq5d)GTf;6svPXorl@K>)~(GNOCvkUystK+tVFSOVI{f@RvN1AxKg7 z;R*^j=hD@Dcc|Y8LDRWESA?he?CIePDz)qiMMOqd;VF~$1dnqI7ki~LRP4mahv zV`U=NWtRkRqzRhl(~`Kahu=f1T`{}}I+;-At}lU!WSl8?Bn#H2dS7k+Y-TlL1iT+* zb-l5wT?_Mqp{bk;Pg>2^r|S>jBoQ!CfzSo$NqA5-YJ38e{GDp+8oB1`Q?iA06v}VB zU$(o-&n4c!&jT{w-#1r;w`>!5n@afXtqzvSh|qG_8J9RUu4ywVs!}-**JPFbI62JB ztsf>yWIC-h>`t3T)-AI1EUS#r$U=apBoFG@Ts0bHD#|D$C99gWQK4I;&>3~WweyLK zm9CrrdSKf4K){p&V_Yjjs0N$u`fcRPIcO|>I9k(k+;_7Y1G^3$N(h+W|Cx1#@n?eU&%iFc#Ld6+4S%ompSxa1h6HQ_mk- zlJtV6{R#oucA>b_;x4@}t+V*zxd-xU;kjL#ErtnCr zBp%)8oOeLJl;i06Mu$Li?iY%uz2A)xkq?NT2M%NFljr#9dP@Mlvm7IxmzvdKcAZ;w z{mWOAHv=4a~Pe>#^t0tk+g(Z#ww$hfy4&D%{r3Hn{hMzedVs`Xji zQ_dSDlHrTb6UjO5@?GvKVm(5`2$m1kEd&v44ZXp$1@Sq z?aQB$LkMMAd_j)c*KYt%m+Bf<24KOzQ&yfcx-RJc8RRw#1A#nrp6|rM+SJg+hHfw_ zkKz(+DW?hLFXhv&TnmN34KJs30EB_{UxafA45_!Hw8)qTivQHj0Y;UNRe`HM{a3lV zTsqZh8vp!XNEwnA<|pl6;|aZMM=hZ1}-H9 zy!zau20)tvfVq>}|lh5PH-q6r9He`*)S=zVulq0T~Dxtb23qUb(p zWDYox1i0DbtPVe-G_1{m9Z)pP>Bof=sgN?eMT5X#7NdURao~1tf2Jx96o{=Jr#cie zMJXBL4mSvPN1#RHV$_lbQ%GD*AMBc-NKDI**p`)g1T<-senS%h;XtMP=OY;gX@Mz_ zs%Wd9Qi#fA!DSFcpaKazmb9SU3h=y_Pcm~$i}71oa@<9%v8dyxO*?=YmG3)vZM(M` zb|F`=cp8R|Mc0_#w6Q0?W}@=KrPAITE9#tr{_*hvxRd`?a7WG- zW0CP{Mt{)N!OL@{J&Eq6>ym?r@Q0`&H;Wf`eMAMF@g$>zfg?->l{$_eUHD<310hS4 z>eum?^f-P+0-lYm1o=`S)||-c=i09fHN*u9S`$Re~(Ao4!#p82AVA=c9hlRR}Kw94uYUZ;i$J--+sU^MbSp4=K%)i?n+Q^+#X}MX3(b5mhen=CLO^5 z#V@22eGAgng`q`}fwh^_c%;~PAPnmB^a2l56No|PO z(>{sB^@6rn_^_D9rz&d^9?bpyAt4NQ9txFJ&5*RZ(7zR7Rs6uqcq!`4#m=bs%LFGz zm|FZf&w@BKZQS&@t%Vd_gi7h%1d2?p+f;FeJ^vuXX8&Y1W6$v}%=~)+d=K=x8^pC} z&&_rTtsps`9}w3t6|yO+ zRD|!;M5ou@QIbi?M9`wJ7bstpjT?Nj<)RYJ{&=a%KC?6m9r|1VvxN(N zBwwrdKK2=RaEHk^H=>K6G@ehVGf&QX`ot&$!}gHX{`kOR6sv`f3MXh!ZrKL+X@)q= ztndeI+V|q|=e0k6RQ&qX^`X(f!Mf$`mHS7>(Qwc050Mprzs9DWnEnC-evW$)HjFU} zLs%f;dz8h3+bLk{u1`1-4*O~4#YGLpF}ql49jR2k9BN-nObu4nWhypNIHWKLR4yIC zQ;G5p-L={bW!fTBTM$NKuk{2baXk~MdB)@6>p#IT(Dk2-H4yNrKi_s5A)}Eo+Rcj# zHbF)U>GXqU(#8CSXhMz4{!e7|4+H(jJU0t7`~Un7EykWm*l4@^^>3aVMb&B!AbD%J zqB3qZ(kZ9fAjzxxtLjA{VlkWzl`3&DnnFLamQ>(IGdZCQ-zg1U`zs!z|hNWimPSZViq^`HCu4E64;2eY!! z3R1`;Rk&J=x44H(qMHhD7|zcp@)(dC2q*f{b$bNbpMp3M>S$z}`;GHRNa~D(-6X`Z zpD-5x!b(>EzUl#TpDI>)!BnqWPrxk(wsCJ3tb%aD>NYb$vrDcENE4FMq^PDN&i&PI zfhkzd%a$l25d;H*FlOFrVo$|?izE5I>v6&(+Twu1!I#6<5pxw$uZ)*A_tnu#a zLE2oerA-)r$~^G}Gn>Lye^qDrY~TenA~Z1K-VaYpA=OEFCzAI=50GYMS^_GC0186J z%&XJhiK1O!5wk!ar2|H_77!^M)Vy|!@?B))ZwhRkiWy*bK%zm}$)7$%`ZKD~&yvE1 zq#?gQ#_RO z$gd*C#XW1#v+51%_~Hp@KqfW>N%PI77@nahb=$8!*#z+xqjfKIA1V$;s!s~t1bLI2 zMK-1*siTiDqAc0tp4Bz&1i0e_SeqX5lu8p^H<;Kbk0x@0=mSrNz@X@1L*R<)IuO-< zq?qrh!vqsL>=v0q2=4gNee_guNMd^3tVG<=N-RAJ(`R8t^wI2@!HIEYMI~6&Hw6&$ zv)91H!xd?9#s#WdcAoRwupqJ+1_gplN&E)HKg=j^!Jn22JYf?1X9jtp?1rX-);^y* z@{g*y)i0^Kg#po97C`dl6E=>-82I?2V&Blrim#YHkN$z=HCik5AaN3~={>mfBqK?a zcafjtXWzYZXG*=}y^cQ8j+CuX74O}%+)E&mx5z!r(88UrB?BUsQiSlcb zg-RWU1%HE?jU|f>`xG=w6!PJk$#Y1>7*U>yJM_lv8g zRGLYqC8=6w3F!bka(Hih&&sX4^B4Zb3V%LEWQ9@ZCyBTb)l&+l1dZBXdjFQK%&B2h zI%=8-AVTfa736GsgEn=d0Iu#9WPqAVs`adjztL__0ZmlY>m65!%41)QJp?=fNxrYlD7c{W%&oELe>g7|jS_ zp+#d1Ofnu%*7zarYYRi4oOF%(A6Rz1ne+C#=o%Q91;jx&Ikv47X3=&; zD?7T;^Oq#^=Nv4PN@!^$6Qi#F6idd#Ya{}a#~*bP;mO44)iS*|eraHdM0UR}fBO~b zD~yM7zXf=bMF{_$rl9CoNsPivjuM1D-VR_#!YYd);g@66pyvJ(Ax#vRt_|31J>3zk z7C`9(>Eo>H!e3_MM59UstIZ*(R>1m#tUwTY1z&4iS1TONlV){6IIro}Y zEmGb2Cx@ZQDN433(+Y4Id-gPI92$e;E;AoXkoHW%TW=GUt@F$y6= zJs&1_S~855mjp4T)`PYp5a8?VW$@e5bYrl?I-mMz31UWKoO!+F3#z<)IpcJV>78^D;QoG5i4R86k3LUDBo0JA&erly?Y>uTOIbU-ifT8&P+pSKC^!@+y6s62x^QsL>2dx{V!4#oV2Z?Kl;f~eG4 zS!*yEle8x86(>RwQ)tAl4RnerW9);$9q(#8&XumP4*`Dugn?*OG*FP-)IxMd`^x>^ zbcu+FVLP6iZzguwH$kOf*s#HGaoOIqd%sU|d+WuR?&>%X^SFS?Ukk**=aWF@&si6d zpq=MxtKtK%&4|f?L9yQFEXYX^{~Y5THAAA2`Gm#2CpOAQoqe1SL9By~ZZ3nnoYIm) z9T-J7s^K)WKPUE|a)~F3JtX#A!zkXPv;IHI-l<8nwF}lP+qP}nwr$(CZQHh2*|uG) zY}@M95gmPV_QiL8!;F|Y-jO+;Y;GJwnoHUI%QhyAF+or#>Z3KiaVIc;0J6aKQ`Dl_eOah6&f_wN=^Hr0T69VLI+=om}W@D)^QW8YD zkx~`fyo+3YE+(O)^V|6bMdttB9F%TWQF%)sP+=Ff($s;ZAEi2^Y~{jrhi$dlQ~Y(xEvNBs}kTObb>&6I=`Wl z4%7_~U#aAE*|4`W!kg$Vt;Q!h#o-bj?Q~&@(9)eowRn*dh6Glgj0m;TfG~kJ{;Y`m zvr&Ln8C-Ci4P8OwR8cd*%ib&cCudCs&s@z#gQtllp+b_8t@|Mkk-IZi zR1)@mlI=sZZt?`$YFQ-|4uDR;b5Djg(A>?RqRY5%c<@F$^Ym#z!759^s_NtdivOkx z{4Z7ZF(pDgC~#TOw{xWi!dw?;^cfDAA3_E{YL4uMxa5;8J9ws4B`Fm`o(ikC2^{N- zHs|(@qB=79s-Q~ALEuC?cZ~Ml>{Q$bePa?ChVdZA1R^k1@zo%A0H|X)e}s9>(0;Y- zG^O)^ZxfjYm(iwYU!FnbZ6E?kGk~(keRX1Ig;dZuJQ1a{pYNs{YHz238EwyQ4E7gg z51k1HHB%I7=09bgofSa(xQjf#$G+kjO_vYaXxr+tUoacEljosC>^x} zI6mqBL+N)&FQpjhe5ed;rl-kBno2^+6IE@3i+e8b)nm_I^l2EK;9qrxhVDp@nMBI_-bcab? z;#*+G!=|MfbWa$MzlcP)rc+N310TNi`j7xDsR3Y`wWG#eH%P_LiBKZOKB@ZULFswW zRSJfS;kbit1Y}z_M)}{#t27t}PjHio0cA=v(}Js6{}cV;KQPUX zlJFIeKj)ORJqhP?(Rk*fPrwe-?yarjxwxPcraYk+$27#@KuOnb^Yop6lgu~qlum(& zIw<=a$x4WAFx`&VNqqQAY5B#eJ!4#_**^@6{p`&we~2wU47?Fn2$N^R?wTzf?yS>z zA}7z;TKC%nX?%|UzRjAPC+ZI#d^!7X0M1__x`4ZtjNZRG#z637hFzsAxWisqdh!^9 zq+!;Dm!p244<&v@^r4fNn*KU)Z47V?OKflDe*F&1_EO&b z!$p?-(v{p>{ts$Uj3mFvtwEQUO-{oe#)MmVm<1~XfRRJ3Y|}<1&#ZGt;C0^TzDUmK z7On(^mx-CovX7yfb&lG?X#htP9!2$_Z z{Lz(Q!dnyA&F?704bQ^vmVD&oYa#cxB>}U;^ukR%2TgHZX2(-EQ53EqNauJ_yuExd z>Yg=oj8M(-wLU6u(Haq8R)U_R^zI*fn2UTpTmSoU_87X`qIbG)<^nRoDSW-m<~foXITrOCX-@7 zKtgp1lx&h9WKOv+iyRF%IBENV+?rk7pq2BK-8>K$7B2!RCh*K#w3Obs5C_i8FfkeP zhvF8r)%hxEYXvj9XGk^Mw|ii37p+?2&|7lVK_MWUjoZ)%DHeuq22XhEb-C+;jV-{c zY7AjRX9?2gV+9>1k~(O3m%0%Oi0Nj77T$Y633xwWZ+Bb4Kn_ic*)TW84=%;BSeMH- zFXiF-ezI)f8X=dXnx<^(LcF$3?G+B%UGMMv`SAP-(^|Ocl*pGf#9EqK_l>%~RziQv z$>hqom0AcY?}}wNmSDa;p`n^gFhH=!27-nNyjhQ`6lo0vJ!@oJS7!?yz5B^Ng^K5+ zbU3)ChaIK`FqqCfB*Z2xH~_4R++o~$!TG+F!9ui|ffQnMpe91EEg4A%3+9xgE4EK_ zt!JMFSb)n`wAhPaj~~beT7p`WvN@*68G*V?k_iK?k+GR2{{|9FQ0<3ER`J-@PyoW2 z#wRz6IPCe#^(FxduQwkxNzP?+fcGz4JCk2mg78>XtpR0B1h@P<-mt zDohBb;P8i&CWQi3rodtJjf1g%h0O~pCaO_%7_86d{$U!@)QtE86fqK0>|H-uezrns zMs!in(1;ZkleG&!R%1Iopd2mK!-v98t z?gHZq3%(ZkAP(RVCP3WTP^2<2_S&1Y7}Ek_K}~B< zL#*&vAv0#F=*j+i#_Rc5%9lH$NuFriLDSAri7qCHO}Y?3NFLdWt&8JSo^Y1xSh)v1 zqMAB02AboNHFAtI-*;VF%^|rQ!sTr?ZCnllL!JU*NRsJPxLTj^WjHV3N!+0loMDoT zXU%4kVSvB`qFJty=N0zSvnI|!N;%n1L4H&aL|_=!LR7zM<gHoi zKtK3-{&EdUSAG&yi?ADIF%<<|(o{n5(}Un18I14$6ZzVONHVI5AR0d@g8yj^Y>LvW z4>{mx-ETIh1mYB_oPjDBrlCoA7*3DU;k@1PO2%=>zfGezPGuOVgz`3PWMwK^RdyPU z{A$tdP2-~pj@LZ2*zqi5zzcPL%nsFo^WZlfID|@4{~2||opXZNO4?4cC7lC`6RZn` z=VWa_cd(j1?MoU_LKa~j8pM&h*xa2)!;zQN)Bi)gd@Q_fzlzKM_^ZpXFv)`)NPVz* z$+RKeo@3xX5FFi=pkfGN#-JW;mAf{krEVP`c%31#YlZ=nlgBMrb{dt;NuOGR5~$ua zawYnrWHN}Tk1DPZ@+iQReZ=;GxY)nT{Yrgy;&Q44!~0;4|>b46t%QE|l5h0X-D zCL{*2@@n$Ik<68vgdT=-1$l*$SAWP!?n+AL>M}{-;kq>mR+F&meYQXpH-`fWZ+lz4 z>YV;;4C^Plcr;PHMs|2FcebIOH^s&e3|6 zNT6JxM(*>(1l?rmF|H0Dv*G5Fz~pb!%;3Bs8(p@kLiVaR7>bjl7F zenp8|KMpuDrrkoyVi91B^0y`oTA}OdE(MKaI|xPi6xE&=Kzr3eY5a%b#WTpY^UjSg zIXlI$HHsD<56l}-N(;_im>;$}-zGV-J00w3WTq1ZmBY|=vej%$nq2w3ll67+kBSxA z$0`sM?Y`G7Ye6xTuqzE6o&15wsWqX`(GF@R#3zhYN>h&Z>Di$CX(4+0pdfUB?c*S2 zCIyxH0Z~f*qT$_P{?9D1jBNWdSf`2=b>-@QG;G=u>rJSdAR+*-RHutkjla?O$lv4^C8{-x z)AoC&rea z|FQ-fk8`*ylPR{56ft=|2=*8#Z z%K$JfHbwLyPR?KndKWS{)BSo<&s^AFOzkFmY)_u?A_HGdK7@^bZ)Q3 z;k7Y15H@r_Lp%55CORnk(*&03X!2Qk`U`EkDI`igbmL^i3cEv~=FeA+LZsjNv+E%T z(RT_!&rx0G`fndCwF7RQFEvJ{j@@voaoX{vpZ3mA@&kX{NelNM^YMQ`(HS|J|L6Jm zUxSMM76(G_sX7EFr9e$oQiXUFj@A`jgSUo(B7W-@8mKKZqc*ROV$`93A2W_rnhoVW z9~~JhPV!@3yiZ@Z*Q0a>XN)K2X<)ODL({Oa-xUeu>lcYj8un>&^@=g!jpobC$F_vKSO~N7Kz#ByKZLWvf$InN7{F<;Qx11jtOzzFA(6lWajJu~QpC7F#;cflC zGV8V2rt+8bNvjGhL{nn?pOg=4&6OE9r7~w=%&sS}EQVP-?|Ca)r;kmk=k`DTy6UPu zilr7OGbU_3$p3!6pgyl_qJI+K*L3(Me6(;s#ss{DIrfi*cXUfti}M0pm9MALm(@yB zeF#RRtFYNFxsFOLH_)gxm%5vaP%f{>1)RY!b9hu9N|Vd3S^l0`1iFd;O|YI|SVJ>B z7~ye66Fd<06VdKoVvnY{+cu*%9ArZOb{X<@_iZ+ydz1c$Usr*5TT%pKP9TpEO1Ppx ze9-;w?*4Eb3Ao{&hSazSmAtrHXZP%-tK;*%&xNH4Qqs#Po!x0blRG1IglYc!?(jUe z^7)vA2og9}G``RCxL9YFFhG_9XFQ+kdO%ryt;f1WX*)O_4ZP!+osjW2#nTYW#ip&n zJfppcXZ}RxUjegv!eODwp`^4}Z~EOEK3;GCgYW_#D1k5AY+TPE!-MV{tl@eTxg_(U zea_R*jZf?_=O+W6mnK5biEW-2-y9B}BZK+FZZ|G7J!tx#o?=4p$#tUbP7+g9V{uty z4UOP@L$%_sRd`>M9h^umq&q~fV0$(k10}?($KTII8`=x8+AW>s_krR*ksnDfv%EQ@ z@wsv7j+wb**{n_i0+1Q@$f3^4i=pl1%*b0)NBYf04vq*yh{shqL2_m&EWehbkUFOZ z5gJhr+0N&xl;~J0lS)E{hBx9y##N@lHBNFYSW5@ zTgFjujoJ#vkv?|9sF6VsX4uG8sfYt^xI;`5EqTp0AylYUAvmySg4Grv!i1tSg`%AR z@{69qF$`K-A29&Pyp`ePC{pnk6XT4NU3G<93d2LNm^lmu*ztxsSemhhj7G5t^Ak5^ z#wqjj4b!i7K7fq2!~1#fdc++S)q93Z zUWCTrq9glC3OcEeQLsuzSaz5*=8i6^aSkqR8!!oAH^>CAjLSM__|ezq%V)?xyB3ZM zwgLK1*zFK#SvQEHiP{2)Bbprg{(;sC{`=6j44N4S{nAS%mkI`Dl%4=oWxnXb+**QK z+X)2JOBJ+#g6#BCtzrpiE|%?mA2GHA}2uQVz|l}<}*Fuep;bOmb)5e&*yAZC%) z4#2JNpfmdPb_u@8=~NYD>9QBK_@i{`)>W{(5pX+Jk_X{e;C0dfFCYV8_Nk9H3s|)lWsZJlqL=cj%zD=) z(gSznJUJ*nX;{MC11O7)}$Xa*f znUe*NEdUO~BqIWmJS$hglN|(X#=7D=3ju4D7y&#P(QfpCea?kXd~ehdzYm3FN1ASg-~(v!169qzM(8@mEdUsnT%E7Njl zdCic{7J!0ApoJqFuo$xH{^1pX!HgWe=ooHZ$!o;}$%5A{V1_q$HSz&$0EuLE(HX;u z+@06qht!?NDv8vqCmY87e9R#u>&EO}v^bR2;~t)A3nkE`?m_^I|2}Ia1lHDGF-Ia| zlWgSp7{=)V8-OYU=(ank_4Cl0&JU~$G3^Q{)|vh-ZJFE@F>jkbj`!&DWr-&;dL7;s`2tRqm`Fe}fBetT)E)`u>|e4jKaM|+DmwWh}DGptHJ;2PQYFWw*$LvO z(TKVZ2IGB~w^-0F(o`##29ukkapw0mPW+rYZV7!g(s_9QCX06tpY49jA9gPUAup=f zux^9y*bs3z{&?unD2eVqd&u;VJNdKxmVHuc*HlVd(3k> z)7zCwd&}NO{}oDILVg^T!uVmZU(kX@jY~ltc)!!mnjTX8%wHlB|VT!jCo(3nc$OFbGRT9|bDe zq)8SDjFVC<-CS#>fVfJDAw-G7EXbIOs_bgCGfhdM)>}SUuyki`T;mtd$%T^&m?X&H zf!kK7YzxWbgnhMc>!HU)F$1q@DZsv|1#Gp`@Vj7YM?O02OwJqu5U8@x?zTA1v+9Y@ z<4_b$RxLR4MM<0zeGffe#wzR9f2K>iPvNz%<+d7b92iJV3gV^4&IzeAzfw%97U)i@ zO=KuN9A^iY$_?2UF?0(73E-mBp~d;RL@57QHln+H3WW1?(XYjS-YsB^oc0Gd$Z`rS zJb$}#o;7zy6n)U$HLr+EO38!hDswL4Q)cjNP(xXIoVK|f6_ek^&BoZWd4_KOPF0dE zDUG;LPX&1xkQfH!uKGNn1ax}URKPv>59sbV1K}<~9-osQ&Sg#t zpb665xRVfS3tY|;#p%OFAF_5GCA+hzHi#Zdr~)*qPo|K*amC;>#~}@f>)K#nE;7KA z7Wkw17#H)2in%r7wo@+Fl37hJ+iZh*imjrhL?1=O2h!a{{8!u$>$n3f5s=JZI1#;+ zp_`M~Fd|#%r;jSqqSNX|p8=Iu#K*Y(jQ1@W?iy^Ej5*To(#vW%x9742ySNuk?;vQ) zXB|J_ca@V^P5tl3^gPw$Ke!V*e)g=;KEHXr2k0MoE_#a>ja3(PTMoON*MAO#hhtoC z9!>S&hxt^_D=StLdh+^uB1WRhHtJ@X9yyoF-t}T#jaggs*>%7mkS93p>*+*)r4tKK zSUzuIDww5RcyGf4zURelK8kk?80U;*e}01PVC#Wy&Q$wRH(BW2Pq?o)dl&bUP_E-8 z3}-kZ?wjtNhSs%3Ct6Z$Uo_x=L3>n2UAJz8&q6VJ*JgAKWsv67tE>+@l^kD}zH{$x zZ;L)RcI0b#Ixgs+YZ{w~=PbcP#FOD71m0K<+%0VHs5ae1s4?=~HaF|GxybKI(QA{` z6(#8}W6GHBL@?ZhfXS>l%JXD+2)TH?a!axBdO0^4TQ_?I0iJ`7*WF{>WmEEBJ8ba< z6hA8fA!Y)03B?@58G}5~xq~`9N5Ugwy!U7AIaX8s&O#%z%4jnIHmhF1+sUA6w(o6r z_b}h}7&5T4LLj_?;g@6A5eDt>)}3VvaJP*y1-=SpgPVscEgpVG)<8G&%=*%j zc~$f%Mp*KcZ1jyyyQsR_m&GSQ&@;DG=;HH+LMq!&Lqqw@!A|87Im)~|*%qU~@C!US zFi&`7dg>{GjjFJD)yqGjEDyuN47ZpZA47yu+Z0<0%IUAB`PWN0mFT~}aN{(h5F~mb$3KBUQSsksHpnK;& zV+e7S3w<3C`Z?Kz*kX6Cn+$!VIQ2B5BLCPjTTpon5)K@6~q>JswRNnRB*#rzD^6{x)d`g-SCs49FBU|h! z6x+cP5=jAp-1)%JZ{y5=w`GtK--?$*jFMrW)R=HtBfX zByeABGYmJ1hJDYADrWY{YG`-TC$nu9IVHP0w9?ywk>MeOLO+HX?@u0>&=sDMB3po+j(O>E%b}FVn zQSELU9AdvJ>Fd*_o0Axq^&33u-qSMFhWu6e23~8)nzXF%a+`s6-)o!Eg-7?%jOJG4 zCoEQ848Yr(YWyA&kAqLZ4OV?;0W5QyyYq$wv1N1x_+ zGUwT4@S`&?$Qs6`l-eW!5(Y<8O6-eFUnJ=|8%m}X$N}Hp;v4Hmdpn`95emwhKryMO)*WrF^rXU{h-8j9mUiMXIKfi zV;0F{8#IKP=&(;)wKq2W@TU|S>iz`1^#T`oB07RG7SYYn+PyLy?-8*guUS?k$zOKe zSQ!nm#DMHrg~6R>()Ii`@Sy0=?E-Z}OJ$b3v<5X%4SQV&&GlZ*G$-nmu+s-D8Km-s zFEd$<^P@yj6kMuLij>XAihgdjv@mk5A>9Ga`}-dexXUVkB!v_q%nrVJNSBqM)-OGa z&27BIrrel2mgg@TY|_7A6LfrabbdM@059ePG&08Y;IJ4LgH0A{HuSwdkLIKVf`asu z%^N#OcIJucWEu3&SM|~S^gFyZi-uXAl*614QB4`cM$cI?E)hpx9m`OG_?MA3zS%~o zpagaj^zjl#za+$iIFp`vLt&mYun{altH_@lIG=Yx$Ie^Pj><^g{XHWF94EdUy?N93 zAQu8`D3htqnmAnrWNV{W7-j(%PYKixN!CRcDZ;mxC_`^kX zpIgzU0Q>u$NlUjRM;^8!(Dq5|U&SnKXalZXh>+#*n-R668na37SP(QCRJQF@UZns= z(4aMH3!^G72Q4_%BW2Ychxd7H2Kp1*qE}aPFsd-rW>oG`!mzdluB=c^{BcMs+jU;* zk<81)pUkf+}fFmTUlyL#PTt(Aj(3WUZL8j?UCj-0qI38cFPGnEtp$i zwP39SP5&y1oEMwHi6h{&bQTi3!$Qrj{+z$W2eZpUT@5e*2R0znvg=Ye6n=GQ<=q#h zPtO&D^Kq}n9>tZPrKLA54V75noa&P5bL1cfHPyzP}3mEuJ& zEg@cC15Kj;1s9$OY66=RLy2a{RhL3Lvjb5aa+|>d#OOiWkCTe|%pNkt)Bz&5o=FM- z4UB8%1Zr#yg>f=OoW|Zv16ogr6m(9@ceS=$;+NA+7@g=&)3*F@1VWg}+LOw7#pmNZ z2vd`zpj{btQwMOJWI`Y<3>!0$JA(XUMX+=LXA9h@gj{em>e!{tK#}tuz`5TPb*S1I z4=lg~5o&HsS3do0JiPG%U=2A7OSB+F6wUyGph;WsZp66i(pm}`;f{)g%i;;gsAY$u zK$P4bFg+Y&3aEz)-l=R5rm?|_@Ulj=!$tbOtgX_Vy1Jl?6bTfDZBOz*@E%~jc28n80FQD zZZMdnPX@`Q_{}@ik^*MemI!9?ayX^aY`#yR6cj6rWdu^S#V+@E7ooV5<8ju0! zf#6=m+ite*Pj{j9-7vt(fBD^BuPLd>`n`)Ud2L>}Qv9~y4l+Na@TV}IZVQ(rnCnx* z<-G!tZeL%@`lb6M4;^M~XSTQc*0NdTg6)v|eEWW0?>?=NN3?Z1Bi**P?dBfJ)zZ^Qw-%vH z`$8-&Ov!uIycE=)7~*!C`)6Q(vzxfvicSJO5WYhvNL;j~STzhSmP&a&i?_aCCjA>(lCOcpa`s#m_%Rox0wIc=1QyR*yPJdCQ)CMCQ*U8nM4$-I|o zM3*czl=jqQa}jJk+4$Bu`fTd-&#!F##${l>9yhY|r~)dIE9gf#gfp5RTNL>zEvkq9^lGRun^44w&P149C<3)(;| z+Nce3%3}ieZy8L${vIz&qjf zdbNTNb<1<{U}%(4Vl&Mg>Aau3&Uwi8BAat^f`jW^z%M`_pvti*Rbqe9$gtU?iwgEA zQ1dV^+FQA(g7{OD#l{O@gu`0a{q*H9* zf-@i{$7l|{(<0-f5c$ACHFCmmh5*9s0N<}E?B~Hkr%y%-;+j{PX;YjK6ttN7m451F zFc-_N*eUZ+l@CNSDy%mP!kbkx>y%n__l+`&u}cQewR1L1O=6`nBXMnqo-ej*ju=;3 z-4Bxpf(lD30*)p^DfeK`+Fe7$)21)ONLc}@p>4ISH>QS+x39}P)6q3)rX6JT0~@)e z;Ku`$=g0TQ=W4)PTv-FDHkuW58cFVXZv0pJDiMMh~| zOpCycgiup9Sf2qx&;loH8x2V?0HC9>#^P;efnp-H`_-VI?+9ZiSM$FH(TMw zh6zUJ5gii$W&B`u$my@g|1~HQA`_yWeHrLc2r7AFV6v<0=8f+tDuF{0;PDuaN%ppq znNKO{%CgN#d?YTh7>lrbC%6n`0vk2|t0!&B*zAsQK7_@hKbr(=yHN|B_REr)8D1@K z(sSdnoO&^+LfaRjIM}rh8aH1(C@;1gmqr$d^z4$l$V362t zv(4Rr2YOL=t&q*Uj{zXF8m+%r3G5djxA1ju`{+8B{e} z&MjTY<(y}X52-wQ%$yKY2UMzTPy{g)5fJHg@R*odEWHXJd^e|EWpPDOLILnr>JbXt zVHpfu)oo*aLr34<@T0??LRM+NGHlc%%9dQt?HF>XJDe(74j22n!Ph%+IaXrT1S?(H z?ZSNwokWFH=W*{6fp3X`L$`nl1Kt5gWQ7o&yZP}Yq`ZMvPhz-~Y zE(6>QL)!LxGDbcXq>9y9b>&=^y#c@9)0DMs5Q6W9^y6pnIMRRdc=Z$N?JYAO%Q?Jh z#XOrs96FZ?JM8cqIV@9(kCXJF>|elQ2l)N}Q{349Tj1f?F%uimCv?n4yv^5o}0}2^gr8O)?Hq8W))Iq z9r>AT4YqEetqdJKx2mre={SU?6qRzu6(RuG6J@O@Z95WAsWHfy808s${450!7}O6a zk`EQN5IiW@JIy9h1T}&w0iY$B%%iSwM2|}(`0sZXns*fiYtMyVO;*@lt0@;8%u9v* zD2ezd8x(C;11<*IU#*&7T@Zmd9);-WmHDEf>7kYR!Hrv9_xs9wsSOZ~kD`g*zlhe? zko?lh9#!41(S2z4`RejWAbvU|B-EgqC1}Z8nLdsw$n5y>$7!hj=R$-`@Z!S{LPd@{*GZ-s(2?8(D zr0iHXz0^D}CMM+M$?%+EGs%w1MbYdF=|9 z@;y$XYUW)rbOe5JP9qimXs`O4)z>x(mwTD$ho{q;=rjr~7gu?C7r{9*ko=2Dsh5v* zRi&<$d@EckWs_UwsmUOxC@0vfnuC|8HE?D%E2}O6AKc9_nw)V zGEZT-YPPt^bP8Kg@_eK)5(SvYHw&;?Eji&fE_#=M|LxU=Ab4LAJ3al*H~Ai>cm>M zxB+c&+!75Y_YOS%nSg4v#gGV>vNQiYs|kd0!e=_(&I#Sm{v@GtQ3EVP zdOH2ri7mc5ctLR$6D_YMxPO;_eihddA$C6vhjZodmC6hhCawT|2ynfrT;O10JI6&a zVGdsX?s(JI)zp{r$&5t`$6Yl<7!H^o)1AHjQTIHtG3r^J zrps^j%FMWH-g0U8J-uv_rdfXu)znLQ zto+N0TW$}|Y;RWRXmTL=slLf8k9_my=Pc!5=e^Jd7*@-`2Mmiv5Yd?)@3!6YMqdWa zy=bMmNiM{8gf((MA1kk1c~@x`Ml_nYAULbTZuhHsOxpzDpCic*dq^n%;>)3%n%tfrVV z=*pV!%jTeo1IZ6DGX7L#&AocdDtS+G&N2xy$F;g{H`u(6+DOkNh-{9qZta#4U;fGO zhNdj~AL8?0%tMa<3;c=Zzub-lQtv@!N?FsqM?Xvf>0<4 zD>3o*|6Z+@R4f|n8rK{4Q&D&C>8-N!_3i$=%a%lw3e9x(x^LGm(LB>>l|Vuli`3+? zL*ES=^WC6TS0fcvr_n{R`Dm*2EEw_Aa|=ay`V7MM&yBJpv3C2~W}jP^2xhGVhSFSt z`0OPN8zfF&X!5Hvsi21!Rc|p};O!4+1fu@M7PzqJDW$X zk&tyL!Ogx+Y*Iy}Q?HZ_+D(gu#v%yFtnviQlbKp9D znFKYRV{UPFu4?TzxemoGyWpd@)}kXZomuLzdH7o+(P=@ohW4c7-ar^-T12vb;yYXG zyp#5hg=$g6;X(Jw8=uUuu91I18!Z)dpzssaM8J#hm?2+t+GG1^UCQDllv5Cq*!pmILc1?`Kn`}JK@%XLU) z&78mAZfn3g$zaSbQGp=LS?LQ(iXMW#h%upLH2@r)7S4R!)OW%6fp!)*pvhT68H?mIdL`+);gb&t3k2* zse8i=hVAX@?FjtFu^oQ1`m>3Tm3%*;52c1l0^AO;zcRHs1J1zNdT5%gyCW`*y0jm= zj|~9J^Alp2FNA`s&--e1X_$v!V+_BkNFL*)2)bMFyXkJ)jc0=GNxEa> zdc%2aM5vjTp?oED`dH{U<}~2gIf&tJS$Dh~vy$mx6-nyhQS71vLW-HDCCV6?2S^7m zI^s{v3xx z%ceJ8(iKE)4u7qC+X~+pO{DQq*ui=7hvb-IMj0bcU>jIlH3h{ZE&0M#5}>Q}2UaDQ zR-|br$TEKe@odP- z6D`GIxK8JnVA(+uKw06~{0bC)3J7_XnS*X#HX{XH01&DD00(N$w=3Evm7%DsEs5OC zB|Q#e&1I*Pt{Q3RH1*t*_cRW&5qv+xC?+IhWM>I73FZKXHPe7U2VOZq1qqCaw_D$u zw45XXOcZmZu5&#igs7;>J7W%T=Pu*Jg7I50QhD)%!xF1Bw>7s}xs?GM+2F3=ue zZM3jbA^_9~^0+dc%BI(Nuu{ukqV4Vacz6G8N{o0kD?=5mQ7w9#oDj+Gb)1{t?+7&Q z3}8l~)f7u z2lM(E5Dp2!;_+?{c}>J#h|RHL0V-$HfRR!LaWb(%LT=}_6yk`;KN%Ulk7R6`IAjan zWNwNm7Z5U6FzaJKDY3dHgDlL$(v+w4k@U`NU~^4&5**rJl4i*>M97>az03-v*(X}zF=oL^HYi$r2 z&U>pv%OA^go3m#7d%xB^V>Y37suhcYaZ(WJ9}JQt*a!2nY{P=12x$<@hj&D6GO7gC ziZq>do^>JRvwed73LB5^6=sG?QJ`$*XrY)n_C#*5T4OQdEih|n8=onB?ph?h6#5Bb zNlpB4Y;m-yh4jr1=_ZA2_q9^>cXYCqEWy<}8Ke`xSTI&teq1j7z@DU$e@E_Wp`ISk zM0=cAP>c^OO2mqZ4_JD(y-S@{Lt$f}W$f&bL+I$Ugmzd0(98#YFdTzVio+)XgugIi^~@RSxaWkH+!M;%*H#w0W!$)14N3W_ zJujwU&GsNV+J+FBKR-k5sv6jH%5p2b1ovGgPtpz~9yo|x7Pw37d3)sb{{9{s`v`REB;xy(3z`(ta|wBeotK#*%+k z$y?x(+fk{wU;V_JEu@x>K!norsdoDz1-BdizJhcGQH_(`1qQcHW{6U1r!^~FINuEx zMjBJ!W7t4K)=knlD0OkG%w}IkCeE_(7mL*JIa5qvtfxy>eUw=Pqn57>qJKOIoTT$Y z$52fZ5(D4J1r1o{C+&(uxoa4vmYWfAW2NtcN}kr6lxs3vEJyFuRs=LhZnrDUKUkij z#RV6y*A-Oc>eahnx-cUxOsQm$t4q>2##epcU*;g4s~O3OGM>j3nZ*D^N8}@?I~%DQ zSoWyRa@N8P3c1~o8)^ekm4n)TdfhIIi5`Fv%qF^IN!Lf}mCx)dOTG)|YbnzN2mi2@3dx&ATpsECj$1Egub zDsQPV;D16hNj0&enjgw08nApmV{RrAEFiz-jv3Sv{^loHAc$mqOgeKX!JOsHo+YFJ zgavHmtcJd%m##3VW;HtuQgRC3per2=?)}=xwR*<}_D*5H8_dlg*t#IEp^D7-(F+En z8GQ{jzj55;N53A^>|;_$*cyQ9yL#M@?4#xWf=31*!~A`1WW1SU&7zbW*zJVz7K5J4R3A? z{}YzE3Vzrpg+B&0s$+a+*fYbwC9jk0^VYxn;i5rY#O*2M{-v|?GWT?f7Q=SJ3KVRUv_6P35fW{qZZmgZ5Z-1>}sj6*hox>W|@e zbYbl^7}lqxoBUVSd*ZB>4QCsyjQ`+wApcgq`cwY#)gJi*rEU|eF3Ul(IF;%#;tv0Y zRfFX}Chq_8GFVxe|5xL`RQ><=GQ5)jrAa9#lGyGFLA6?VpBuu_vIz=kH_50Jt4Ox2 zpXS`dgn6$l*s-4sqT@aN9tqx;mcsG<5<3&y3^kZmNP6~Nhy*rep`OzA;{NCEXzMeh@Pl*IH`Qn2pw2}jS6(`~%pf&BB z-y=tCFdPqZj7hS_Y+ac=yF3eS(Y0KmHfY9%SEG5Vg))WRyALDmIpoCL3kFkeU5-Z6 z7V=Xuc#Z0r+R8eQs&3-OIdZ!LM+Ds9#~K*p$H&ZoEOMWjaZF=+Vz0wMYuH0UeQ?4K8P3Cb z5DpzQV1o>n@t-z1jTtgwmv|Q~hM-d7&H80vm;sv?GCW2U!$q@dgI&|My~4V9FlJ)Z zjT$VCD6!h2HZ164KWzf}(quj}=7jM|X!PEQAc@ct>+sY4FraRc4;7c(SQg0C=~;$t z;D_dgQ4KM?KdTBsZ8znzwwcWGoV>w()o2?9f1A1A($$9n-+eA#b20W^#YKY~Nf){W z)hywnG6+q7LPZLBmBE}L??akIk~}+Bs;xfbdcO!58F(9;l+W)rgM}MOx5Ykn4Wx7{ic#X$p8_sCTexIj7n@kL%3MKyGii&psp% z)EWkv2;y)qpR)&RU!1qDMnWwc+?qgZ$c^*@(#(Sjh_X6q5%%E}*jh{n>SwBNji_#} zle3fV8E>Lnu2EEkGTw>)BQ#$?>t)=SG_#_qwxE@4P0TF=Koa67wUth4Q&m)OtYb(- z*I}+A2YL|A04w6Q_Sjl}L+1|{J8QjI0z2sfqF>$izER=Vu%a708bmv=7=JKd6zvUw z2IU0Vu{3j|vRU%~F!oOEp{-pLW^Aw6wr$(CZ6_ zF=~#gyVR~C8WQ%+EDUhG{^{IT=%~CqacEMSWAPIcnsMfUa=!GNH1U#LC|PxLS(7%! zgR+bG`2#*$Tp+mmToe;=HdB_GtZm=V-{ z&3g=oIJlsoI8xMzlv(I)J&71eEiR^}ZOlr_$P$Mun~RE03>JmykGP{r(jd+ldz$7@ zHXxK8W?7KY&1aZyO?OI@hGSwPbtL~7_f=czNP#u9mWj!PjRQ&e2CjNu+=#c8m=+r* zV=k&)0f5s+SoLPXIcL}0)GOs&3xr3|QPS|W@|dg|*p>lBn^R3hV>c|poQ$-lW?@g@ z(0uhFrXVh*B+!Bd6#|q!&GL5bl1R2nAb*(su6gXz@T|Qji9y)wWiVPMFSf%%N$kQ$ z9z&{p6+A8sX*PCg4P*fgx67G6o3qqTA_h`a3|FhR9{J$I@;zxGSk;R~L|h#PpLDQI zpP(#;<>;eDvc*%2JTund`2>gIf+JWvA=LaT4xX={LHaZScbjv5Clv{xDbM4!s?gtj z7WO+i)6jYXC}^~*W!0cxCuNNbnHZ7=5US3{>HEZ$)`SjqDO;R&>M3;5!bR##$FC*r z_`0U1Gb>+$pQDV;So(V0rDwE~L0m_Gv^I}<$HxrdtI%uC-uLC_-RZj|mN;=#cuRQ$ zP*XX*F!rm+SBSgI@-lF8o#F|!qU~;pj|>(K-o!(1w+@5w_7+UoJl``osd#4Rtpqp< zTY1N?jJ5E$(HkfMQut?b5M%DI9KZT|kJ-iP*#%wkD-T_8h?y_NvN+R5z$liBT`+>?I~(jeHhvDfT3bmte}-%ZDVPa>qWi_nqK$ez3K%Cx?@ce5%+2y-&O_&Wvr)3n- zReL|W4%e0Rk^NRHBCwLGk{RpzuZy?mVAGWd z{&G_FH|Vk>4~W1F89ly>7mR2w1_U44Rd>Xqu54BZ_-%v)`eSVwS6#(F4`?HOG4s|J znAI=z#5W0@UC7aL(v%opo!CNI5^4?u=HS1`ixPl9irv9%n<; zlc(K?NE~y96lP)Y+mEEJ{L17K4~G@E#yXR1I0M9?yOs9`hW%y z?3#zW@JPAfiWGw9#zJv0WJRcN64Xy&CYsL90ob8A|3ZqB+dgf(P*WhAmBbv}%GMFu z*+B7*n}q?jJh`mME&K**+rc_RpRIU#!)~)6*M3+b2%NP+C}7(XoS7q37BddMtIe;& zID9s$2oiuYOq-b z-7-R;PSF1ZrS%ehNRns_)X%M)c^?_av#GNsymuVIHKVJ>X6Da4;%{*?-Ng zzJq*2*bujd9~322^u)i$pm>ej{eU)wfNo4Qr*EEKo@bS_TWdZ+`vGV*CRQ2F_s7>G z=qDSRLW^?2boAW$YxZ$~%KEdAL8ee1?L4pbrO}cj_u*E96AbEqK&J>>_vy(?QYxC- zXGj`lUZ#MiUIrd49JJe3(p(*Xc?6VnUF>qUpY&sVO}xF&#-4+ zjH9M!W+VqH`4dt`&mWbOcg0B{e}3P5VoqOdqNdVsHTN_PeB_|VNQ6)7FrUJ~UZnyP!{2&!VpWuqo^+(x{2ZCmxDj%cI85yxgl9p9 zkD7FdMQILRp?&{=Nu@f#Z^=h5L5r)q)?THyyDUbWcm0IdGYvO$VhCvyKS+8__~5=_ zsg~uzY{t9LkPe|({Xh<<=l?6?wFqXa<~4)j?3Y{fr$6cNwttQCELp<-Nu-=E&b(;! zu$sW$JRRgB8yM1l03(c&6*5q5W_L$Z)SZ-;LPdiXsO zfcVbbL#&Q3QiCJ#%v@cHJRYQgqE>HS)?bxy{2TR3nKTbKjMj>bFMsQx2-WQBz)vGP zS9jYjNp4s?>5bU-!`+qi^IDlJLYj6fB+@S@JP&w4<6^#}b}MCqgmO_{@F$$Zq+$Ss zk`eOMA>)(943G&k9-$V(?QONuR46^S8jomX8?TL7(g3Hn<3bWDI= z9x?d88JLV#Tl(T!2k^eH2q^H9pdtyUp`$0EgwV=qyv7bzc?+$+@>Eg!;Und>mQ1H@fdx_kLP>)-FMt>DGL*j;K42%{A6fIWD*>-7{Ld_4yz zeEyFy;FsP+sEJd+#x>d7XBNEb4lUYux@1j<-L(7_;ul&5IySN-{;KZ;#<)4hMxPg> zhbyrJ-)}=y$)NTBw(6%t|C%w->J>uYwkM z;iZRQpUsp8M}+XDeyILtlpK!0vd!pl`>p^`RWi;CVAiUHD~;x_V|P|~5q_)XvLel5 zol3Eec-&gxPXsWhBiB8IPXr2=zC2c3pR)uePvR>oVTZ`dw0+lb=78D$+j#z0+Z2F8 zegn1Gowz0DfL_FjctpbO+NEtzt0}Y}Ln-^p%&lZr_d?vm_R!sZB9J6U>CHr9!?k+d zNWAm8Nq;rSDy9D;6dF~s2FNlgVlk`3Vj@D>!1HGe1c}Ef42PdDQzjlvD6?dDH;U-;*a+M7%~ zg}NH6Ma99vUgC&vXnx1~JL=f%As}s#s=lOgh58!o#FnZ} zRe322gfAJ9ZbirO@?5~lUbJP0v?n^gcB0}de$!^p(x#qh+W6RdYGXJtVfQYdl}HH1 zj|p6X)DG%h5FK+zwYoh#FOkIJaXg`HhOJ~4@(i3_7UkPz0dkTBrY&~mMaoN{olW4z zVM?4^{y(v_ugVxE_@o!JgRr0Ko0NwR0+K9KRIL@@0CZ#i`0qfVN#EAX15T*{Ts`S58 z^k)Jkg#ghwOML;Gmom_ux42qH@GUnop_#L1rsj9){sX+hl%(FkC@J=39$L@~0&JNEjw8f35IkM#Igy4kw1F2E`ahAdrs5_pgb00bZS{6vP_bX@r z91cDxGj8vW4Da%3p5wf=;GNjVUsU zI;-d*#0va77~4U(BoqmUnj_b2!F%Hgy$r5aXH4jR(?T@|J*saNwpjp?-Map3=>s*n z@=8GyvB)bNv$@E&yP)WMf|O-GzGN|gO(`e+c$*^w&p68h&AtNoi+1C9rlXEqByvCP zQcf@bMhu^Snv;wxwwhR*>;sUT&uaL%oOWkm)$D$s-+86T2@BQBr|HkO|c=gMtuIBAu`!~x`zf|m8-;OZZK2hEDN1qd? zK+J_i!wGT)P%ZXNn+g z5$_+)nU<;qbkO#H7RE>QUD4w7)e1Q*o)DOP!GI% zcQnxf#FRe08@D-?J0^-=QOVylZ2=Ix)IJ|EWCEOMab+GObs(ks>H{?CDw6oKnxmXD zs~;mWA5ai$-CiX7eU%~h+AL{pG{qc3PzU&w^H2Z8L}wK!bI5?$Vo$jvhEw?i^_7F@ z%S{KMis{2=cq4ixs|sFaQ38lNDTjzhmHr~A_M1Ow2h63o31J`Pg94vnY7u?{vJJY& zgd7Y$oPo>JbkNzXlpq?;Q!Rz~OK^bAUq z8pGR;5@(rmKhtb4&Q|t#wNGT1cQ`IztQ^yjh~iI{wBj$!R*+(j{2D-qd?(>8mGWus zXVBLj48n~AW1%+?>cGzAaXn49jItLwwpDNHFyQvLaLZfp2mj3jG!kxvm&+d!z8LJ1 z|9vXlT$AA4Fy?xePP_HId9m9iiZFzl=UY$%Qh#4nqRQDvRpfxkzIv0WxTGabo}zl; zaUMnB#6f^2n`7KrA2wQ-Ax$n(T9TIJETqt=ey=)SjR>zzt#*%+3AoeuV>lWW)2{#>Mu0!%oEOcrf@z{mV|{b@V()KNfKV zGy!mToRuV(dqN-&Dx*dwf3~T^5(fBM8S?0m4lBeB9KQ|82*(L2v{c)YK&KY^0e%8t zVE%OnoFJ}ffh0>6NQ!nTP?)An;EzU(AC5Ym7KUYV(KE1%#bs^|0c5Z?g1DvBhJiTt zxoJOX%G=7>hV>UezBKD^L(NDzs~iFn+`ViZ^x0&_`FjtpwnQg6d2Yf!=?y@*GAeka zsep2D0YP_^Lph~uft~G<(_pF%~uWC0JEfwMUq#V5qS6x^X~QC36$GOr>>E|BV~5H87{E} zf=1=)7$7PgwI3QXaxh5x0fhdc8) zmsQ!%fNF;$3CDqvrGZ+~Kh=RkWlY|BFWxIocbJhcPIZIM4f6MMkepL0K}x8Z@jskS z7iS|gXpy}nc@eBV4aC&N3S`ML^}`Z$5n1t}@e_w1%MJ>nQwxubL|JH5Vjx;{>ud`+ zF)SIPB5h=-Qy9Y$vcQUVoqyRAY>w*CPcGrs$^!tH-z-(z1?x3O^iEG+p`#PqC*&8# zRx1FD99SRZ$HT>L%sPFFbZHfd%)2^Zpt@-};T@YHaHgX<_{&>-P{Z11FEFh}7Ig9% zVE^7L%HN55Gqiq_-L1l%b>xn8!7>VSHt4hoer9&7Ls2P|_4Z^S(p^fsNzqKOiBSyS zjw0?E5{L#>^Tayl+7J*4$k>sc4sGtMd2Z-zVdTNIkhXK6i;JZAlJs=R0V|F&QyteL zeNI9~SMhy5-G%-D%xdCj{)Y}@{SQ_KE5rY+!<(CAe0zHdPZNO`d3}AG$U|(} zfNecF^xZdL!O2wr7_?k!xGsFZ5Q)4~xzr5%0g4bS5Q}^ZD?0nGdB47Lh_mLL7%Lpw zxyEJ=?GQ^Gwn@~jp=9z^UlUhUf<5stJIRB@12% z#4m32y7t;^MqdTPtx8;HMKdbjZzkR2cismtE27-%V_!t+1PAW>#juHuN@bBx@S7Y9 zV-P)_z2RrXNOR`~X}Te4iEJNTiczVXb#MGHa; ztb3rg9csV?1hMFeC)<4uc@Qw&E3OD{+u7hZx^#0U0oC}~-X_V-7XpLFJt$S$8q~%? z-4Ynv&`oFIYwm0D1<0Q_I45X~s-nVD#92ZMm4*xDvg_*qBZZBl1DcQ2CVL3%znR$} z7Sy`LpWH=ysGtCC!}Z|J*g@=0OwLPhpI`r!&%yvVw`&48HX0h>UeUx&Mjt?WTIw2% zX_SDTIh&f583E~6k&gddSPwQW{{s|^^AX^mSdwxL?`A`$iDT?}O0{)#j?esl2A1Mg z#4~_S3Zl1FNoZF*j)}jVstvA?!aT^&oXDg*CQduVwgR0@ZaXg1Vs}PzoP0If6U^H0 z4q5|Li4xjTL65GIAHJA=!LMorln|mXe%$YeHK>CBf}Y#PurBP`-ZQH{K9FM}U1ghJ zBs7oV3qpL{g)4024t(uq#(a&kdLu@53GEP z`jS0Ic>xO;XbdchVzp<+Z8UsEsv2#@2z|3Mt_FyG7y%J-iFg(q=G_TC)>^&Z*n@e? z`LC#y_ii?;7Tpe_nq`qNtOA$t19kD)EZ!0xjuXboayuX;qjx0tykhA?58)fsc1QG@~tx?jA=nahnFZx~l3z!a0`_kwE?T!6ZvZ%fTM>h_#Sr~Ts~4(1{Rs1+>``4&!Q5oWpJ^XkO%1pq z>h59+9o!soIaXItuYl|I?^&Ipo?z+VGDiXbRl8B!VuLVk$>RLX&QDO3Q-5MaA> zNw!?z6WYn%&B2}>4Q)dBiRy^BwzddoWGV1u`yr6?B`aOpVKdbb7TeUF`l7$KghfOk zEkfqEk!sbTz9;wbRVogAk4v27%iie5_F4@QV|79kr?(P2t{ubNPW;p-7QSzAda8u0 z1^c6$ZOD^^i~jPa_H7!Q;fE4i9=R@LdM@qy0Sdh;Sfthfk*3{Lepj_6nuWB->1OlAqoGafWjZs47J6$7_k)iN&X=CjF&Ep=1iq=flOqiT_(pn{ZGe4)!>4%#pGDmFi~Cw^kw8)U>oIQ57ZvyE3f`tjm2+#Mt|AGY;4eHXGXJA925t3+85T z$R@awzu3Hs^2XLV3b2d#4_m1j3x+sg$yOUkZXuMdE{DHjo(n4%r z8~PB2S+iO1o@sF#;`#pTic1npnJvH;(`O;cSEn0R+x_-LJ`T?! zTy6IhS0A?`xTN+@>4_y2?PuU)?|Tc^gT3sngn$2II&u8p>BPwX%Xj+!J}5d-3u|W+ zM|?U_YXfH!VG|=eV-qM|UMMGLM-u}ZD7W=$%#E0hwkPkNUy@U4n2JP^E#&>INayKu zxd@oIL;&b$f=zyh8_8OZg>TP^=sBY!D<~8U;r(qKH(O5T!@|NKp0DaSDMS}_#K81| zL1IuQaX2@hLT0JggNYVeYtHHi~G;()tMcZZzFV5smoMX z8Ywul_YH3K?WdaPxK;@BRUg2B^`xoZaqvsq0_WDZqkSj($<*SbswPY~k>IKum8lzP zYCOiFzJ96M?6`917N!y#FtSx~Q2Cx}oiGAAqeUAOBW*{HzuwoFo#uY&o=O4??UKoM z(&z#JdA}!s$T}`!#5Y9tV<>&fMDau^^R1#UV|Jnw9ARXm^9KgVslB=*6?gV%^vAt6 zoGE#@qTa8_?H#?>gcySuTXp|xgZ3*k<%bI})`ENJ4&nPM9ua7=sm%iiVyF#UhY|V9 zf-*DvX`W63P{B9p6 zGI(77Ftg{A0GSE-8K?+M_b;pWmO)Wh-;AHQ&nCVfc~ncO4*n{WULF!%jAA6rfO2{- zPt(>|0HV!=CI!fj?9n&#-o|4JWt|uDJnfVhU3g;q`Umu#&!yxcf~X5amgXdbmn|Vb zdPI_{8_d9vRA00PF7!eb1DdJsDF{IWS3b_}+2DB91RY!#J>+GLw}MKgm>B-yn4x=A z!575@?c&4&<*vQ= zsst1v9ju|$!k=|_bdil(|M|e~C!B%T^xgG7aVa&c9lCxC!;p@vncqM4I4zR$Jv_M|8^Zwc zgO!&lN;xLH-b*kRnTA&`T0)^*uLV%J`gE*|np%`Y9&$D$gZn1CX~FlLx2@R5L`-0s zA#x;E)BDu7>O$joQu5lT56gJLXd+OHh@O(+%{nq_87CE|W)p!6;uU1Cq`wN|O=e#w+eh%<%ppOu_+iOyM^a8K% zK6W>t&a1v2YN1`lJHZV`t)eGETAnLOYh2Qztfr7IGYIaEZArbony;MYwkyJnd~HH* zTwhzC_;SzzrXDajPEs@VM0Th%GAFh7HQTjo8TZ}P&eySaB3+@q+&rJpe%pVv_ef^} zvr?^{u{|}gZJ(b8Q|-P66N?0Nc4BjkJjf&Ho~TNQ=b_R8iT|g{m44j3e^+b zV}f+zI$17Q@o#*)JD5L!X}&4cgEdgcBUz-nNe0{T{c6Ti?p{D=q8ljO1+VZZBBC+5 z{QDswiAae$Y1W|ph{TAcxq0;V3N0xzy4M8shiRn6J=Ki2vx&yclQbn4(oWQuXVD~; zaj9_H!fV=8HmOSQ0m;iobw^zKj-EqC;f?4C!w>D0(9)wsWt8Ob{6!V)FrAn!&E3jY z2vgy|Z?p?SOm%}O>ml90ZAkGa3q;ew3l3M;h#@bFReQj+?7aZDlSzb*ebKmVDyC*s zNV+BUz#Ie&n~M^(D-hAO)YKV((Xr;xE*p~^+C-#vq41j}>0n>4eP_MZRM^+Wz;4g5 zp%|lCy5T^jUztj$LUf=@b#}+{upoU`9jG*4#bnm_Cd74c@m~v%qq$nR&L}!0}iN$Oao_Bag4}32Qf8aN`~QxEa;`)rcPaLkL8gE z>|6XQ46#y4&-icrU<$)6R@kg4LKMMlePBUWgiZ)4UIRCJ#ap=km(bapXWH$EmR%_y zo1m7zua9ufR-!Lk+iyxG2jL1<>8^nGpg>#e9(-W8%0i z4|+B&p(8IPjJZP7aO+N&50>E?BCo$S6 zY%phfw0f5@-#oz;*-tavox(%`BHrrRCR%F7#HgMOJRL2p=~1!}znSCR{fyi>pFU4@$sGR8rS{iWN{?-RcqMfGxV%c-SfmSuR4d9|Tm(XCp!+ zDX~^Tee1a)k|`eLg0k33eRNWQ;p(wD4ia0z!O~ z6^dkflr2X~Rd(U>u1l#f3SLVzeecD}Y^1LY5 zU?!`8uzqI1S<_sA#w}KmjXxlCnV2MRMtVLl658*~JT4<=8D{8jx{H0?SDO-yB9m%Z z95h0ei2p@U7r^J5w!`>ybnK}PCPNA>jQhlUiu$8 zLhDyGo(+W-o_?r^CIU1>IV)}J((?$9;fL}ug7i5e4P$sb}kLwLEEiDy}e z7VYGDb(`7R+kw?WzyEN6G>k%ia>vJ+mwr&fk}QNTaS^*TX;^eP=Vwr$MY^|~i>)P# z(-WEH6h#@6vu_8Zn^`9PQGTQ6$`|lUpp5AMF5euC|9AOjVq*QDQ8vZ@|A~I4mL3Si z9(W-_T!HS88Ozvznr5~A>RQ02(Z;}JoND~)`tx0RNiB(_wWE=r7FELwv;+~ zN3r?xHe87?N6-EH=(v^8R6Fu`-(1T&Q##*TYyoj&Ic@Z!Q9^Wa6xps)GBE>_gZg$)`v314hT;eH=}2W=;QG-6Z=LK3=}h{ z6EkXfZmZ~iKEID*=|cMMLMT*)NtrUa3&&YEE@kz6-e|TjE#hLr6B<^vP0lS6g*C<} zGp1+szQ2yZ3EuX(jbxPN6JI8*V~jS=k>j*(M^l`ygS5Bc3TvN*)0?PU`53mKJ5k<| zO+9a`Zz@aNg%l8uQnO>!USV85 z$ccTXCdpaf5@th3Vh5}bP`lX1UD*STI^S1T+p-O%9w3C6MQ<}TPo(6gjzNc8us+p7 zOgD^1({zqV={6s|i;^bMlBgQSh$Zm1V&UC8l9`Y+Up2G;MPHsyJx~0Vnz*_qhz_)# zE9%B`_5K>-hZhf~3e86Q`+cC;Y{YYKV^~QjyW=3gIT5@(Cy1CZBTU<>us!501rvax0${=-hp0BV%dUg+e8ABw|Z&$(d*6;5O0Fx;& zzaZ`(h?jlO;6dA|mwk~~7q0hjkv85Pc+!_|L(5x>12cglu6#7wtdR~bmSdoz#vmjA zQkBL|=F5uL<4dN~kH%aXHBANR!?ox-(Zwd#w3sAk;s#QeD!`y>qFi-!eJdIcT%(06 zK64#Gw42JU({w0fMN&98tK&>p!epoLVcDIr6cb}}b38`91zwm+-L*>xDzD+pHR`sV ztgRcbESnB8;BqYPyCk-)OZ(F=|40Af#8w2$dkAc6r#oPw1TX;A-G)^(KMl7k&nc?98)2+Z{Q08bR@a5xtwgq|G`CmGGL+c|3R){9I-ys9w zWVn{W2hjgaVw6^fq>o035vVk^#dVOxQP)SP{Iz&-4u}i!qYj@jk1^Q)#}*BzhJ+>m z=YB9SThz24?2cMNGsn4FguGybrZ$LFS-XjuY;l z-%!tl-Ok}tnLs_bR&4g^*SS&c@a%FF8jc(+qGG|fk(j0w%P*|l6!vf#-x8}7;7IZ?|ISh+svq-ha zrvYTl9Zr;N*Nfrg^^!hgk`w1T2vI(z$CJ~WcWOzalI%E?Lb^hWxCtPO+fs{33U>*N z*@#Fy(5u|oK=RM4A;}-}>GnV z|Nb*llAp}}pwBX}<__Xu%ooHtt4CH&b)DhR{|QTg;E`|cPazfzCY=~u z=7jf>A)_No3e&#aCWh2?u}>O4kv8=y%pzEVUuhV0`FX`-$xD0KVvBoJpFI(Qj+-E+ z>8tZwB%XdLC&7=KY^!=QACP>TMHOdwwb`{lI+_4fuyOr`TP3;A&N#ApCJbH+#Uvck zml!VSI`qp9EW%grn`L*!RWe5t?OU=qYRX4e0s*<<^DKL1GDHKgs6&L&ihrrSFsp<%5WP> z7dYCE^~<*KGVD4#+Fd~EM{`XPJo9u2wKDejY1xRqOtggj~kN zN;c`BU4q8!WOO`3M*Flx`dbMB=R5aDs5pTOw%k0A^W_57^>sso5Q0(liAvOGmTMyl z+x=X%()V9G$6dBmNd_J~VyZyTmou?e4Paq)E>xZ;zjy}N%rh6+yzhd=jHX4f_N!`X z%qq-tElM4BJXHQ~)s9-^7tqAS&yODsuHw;yb9H0{TQhcC_z>Cbvz)mrq zMGcXX$iRZA{l9ts_7oP1^4X0hqcTV{@gsLQ`sD0J`A0jkte#ACveE)}cMhtYnZ}2Z zrwAI3qnOE!!#Ct~Gwmrk66s7r$z2MS4={k?;T}5;dyFb^SZqtes|h=J^fX*tdh42e z_F5~tQnKP=X(g!pINvRq0zWP|^twAehV+`&eeWjMKNr>Qp9<#>qFz{8A4<*vduzN44PTOOsOg}ZX-Sgk&M-<)a)S1BbT70EDQa3 z$c>yn7!IHG<+a5B+A_UCKi4T=Ve_`)6FlkDJ@B~i(ivkUANaOkt^1?a(*{IEo_)J- z*AfDlV>ZkyshM(nZ%G`GJ1Jxn<93Mu(r%3_RkZ=M6XO$(FV@~_tY}K3cp`pkVTxHe zoa{UjP;_<|Dqz*hhNVbXNEd&1&C6~j2U2yJnB=#J(iDkrLeU0fur4cd;$~6w_hh{S zLiz-}g!DurD9gD7ifz?xX~q*<=QMOR3X_S9bwfUTMn_s&@*}vgUeQlFqT}NBP&tB# zzjQ|028?z&mjox}QY>lbElc0G=dFi;KlzxxnV7Rz)>db*Wb%0~ct}qmQ@A~aG;vWh zYYG4#s9@a_YnCW(D~LfIPDmpQg9U=`wcx82h-#WfAXp;_e(#1@{IfsHrK(H+d6PHR zOfSy&yqPZ2UT`#ZL9AUg28rU6A9{5X>Jk>!Wgb-YHl&PT!Kxq{D>3}U0v^W17#C7C z*6OFySltAx|9ku29xB0(t_uTzRCaP(q&SDm_FchQvj9L$f|Y3)oD zJTasZdKj9~_)3I?U)(R^_ZkViz+fe>UWf~9wLahB!Dk~eti{Umf>c6x%D2yoa6)&3 z%_L~>cv-=f!W}rm%D_h13t@%T-S6Ehx$*c9+4Ub6c}6A<`v3LjQH(xmd)Rh&U5oHe z47`A;fc%6$7-}6S!kUV&j)!zf4bzdiH7`o@Bx%_ieK}ehp|V)0_3Xw#mM~0s@rCc} z*7o*UD9u9S@pzHy!b%WMRf7Ekow^ASRRdS7$Me(z`T`|k1^S404 zzEk$k*^dri>Qhb}myRu@hFYr?e+9%gn#l`fN&0Q%v2~7&m0G?pYmhEmw?kP+@7L2F zUY`-~kwOLi*!(JV_;BD_N{8uCI(_s@9}7n0rB2b66m| z?5T1!JeY~F0U=~)r|t?C6^kH071z2zzxguuRf|gW^<>HSVAo^Q`fftKt8)(>450EV z8gphAnWT%J>TztTlH7+G=@s)vRyxPFT@@?;bvkhh!s)-Ajm%7??kfm=6dSW*EO)U5 z{dRLW@I#yx@C7tjE6|nB5))w{w(O_Caaf45G~TzU2Lc4Rq{k`GnGQ~WE`kq2*rW-8 z5N>&n&V!lfj2!2**Y{y(ZY%Ia;!{CxhC;Yb*;bP$Nvta8;Vx&g*`iy|fOr$6lnj@u z-YaAVy{TK1{XD<38qKMh?(9B%-0lJN8wsoAPA;z(Hfw@wlX=2usDQJqs2+%>u zHH4rlC_sC84MvQhdRNa%X22BSu1tP5Sw_p{yCYMv(1X$4dUDDiXL@dv>UkPY6l0Kd zN-IsB(84;{J?`PMvwJ56 z+X#%vn0gY6u`2p|PgxWP@0({)qvAi0281}CN znh!JVW2V2~!e>GYN! zXQID zB}ZVJfU4e+ckLwT%O!F_!>we2L`;F>)J^gnMN{Md9&J4M=!TVmU* z*~nXLj^-8A+Q=^hl!Xx_(^k8w$pwn&Q-Mk7$uJG$iQITe13!q!IMj%vQA=|;WptXw zbxLFwE>6v66al-q>3;q6C9oFev{Q{biU(j;%IL+;@`7t<+VQEts`{SHt~`;?(ZsHz zh(j?!s+8Zf&|(fCORDRgNGB>g6ML4{JI8P7#h;iJb%+0ep{UiLQ1}M`(Ic%en@ZxA z+M4Uqkv+D4K})iWVL_{9H0G8#0u+;j#3%e5-_xQgBRWJs}TD(GDADLTTA(mi-`ksRx)F{BHYq zp%Xd|q(SU&@8JPWA`gkd92kc3WHcf6_8}bt+^G`Mdqc?{+fvsDGGw8suKT{gz3hXP zlc+B{5(ts0D7M$Rn(7le7$fAV3)MbNg?x0%YL7wjR7~VbCcc)f{g^Q#amm?TkIS^q zXZ2aGW;?k2AcuPZ(_k$%g)bWbILtkvogc_9S!+7j_}j06>B^?qPsXq)b7+ui_WS*O z)q%7AoHShoSx0V-&dTYOP7$O3As~ksv>)#uz|U-o>x5>ce!+shVAImOt~{{E9U}yV z9imr|=WCnllVTcOk&dx5cO$E1aXn?<*r=-BWvQ?z%ka>9F0Fpk106;u18JpzYz~L^ zirHOV?22rdJ7Yq1&O&OhDVJ6E>!tn?n>ZUi3cO&)_Bsxg(E@60VZ-^J+PnhcvhqoS zz-)O|Xu$JLw0=^iwYo1HmjyesS7`Ih8A+55cmzAxmhSnOSnr+@ZtE%GY_D?fhcn;H z2CFU(kgSW@Whp{R?L3{+{Hu2g9OqD?2{_{pBD58G3FoPH;8c4nY_DvBXOw)DnlL9)fedjLrsOa&`Gb%WK{yGF?kl%$B zt4D0d{s8*0_jS(qqq{}p5RQ(H5Szwut(H9}J?&GAq4{`h&m8w5Tdnb17j3tf`tlXQ z@hNCl-v#q57(;g3s?k96@?~lD?5iWgg5F(@Tf(V;SbnUNnobPxIpV%&S3sgDTS4{_ z>C;zz3npQnt+q!t*0Tuo-^5o;q?yqX+y#; zyIfA#ARDf6XgUDDO{-#pP;K^X&=a*HDym&jG~5&O6AymaIN(3FwEu9XFfcLx?=7Ng zQop`Qg#VXD2nIy%H+(9g292DJBKbo+OAobZQHhO+qP}nW|wW-wr$(!vW?Sc{5^v=e?aEC z2bn8k=d~k}wxfv>%{8xV-SlQ9P?X{uVnS%{*+O4u{B7b~7kF*5J_=C2xyFSXp*VbM zgK1`cmSEQbAMf5Jq&j*JB9*9(el?q+hD%2sb86N0_Us|7__h*$2)uBZ zDmYL(q%^Xz^kIBBA9nP>6(IoU!2XayPFTM5 zf&3whsVCur&acn~(51f+Y7X7VCn(5h81xoKfX zzU=e??LaCHRJLu;VqP%zrKgJIih$upu-$VJAKIM*O)!Zp^o}q`Ec=U6iQ6cn9Pf-E4)#H`QDs-HhrJ!li1#+RtzPyKfSz}610qZRY zy=B)}U&_T@!zZC@4MOeukAkMFoW;t>Tmp_4E}7b_U2RUG=4r|uP>O$=#a{J&R+-e> zB!qpGQ8dyp;0=f?3_7&S6Wu7Va0)_GpN~?zR<*rQ7bo&+7_?i1NzG}YHMs5s_Niue z_re4e2Ic1hyI(dmGpVgCN;lDD;-!-Bh1CSha*R7&TU@?8hm$|b!nSFXwN#@~z#+^Z z8F<98*(?cr;>{bsY1=Us4y@JW`8@l(+>2!GWxm0Yl*`G3FYO5!sJn6QtUPuA@rx@5 zoHjB^(+D(JOQXhswUe_wqU=69R`aIqdUS|Pa8sze$HQvc zE=x%$}MkbetVbpsWu5t+4<9V#4BnGcfGbGYk8?z>kMdV>=f6eNVmtAMb^UG z_-&xK^ft;V3(xJ&8OT3#l7tt3!c%65Cq~lkl9QK7zPYWpE;yjz7zuJ2(RZnz#aSm% z32p8p;q+P=K*#HPR4^5;$M$r3X&Y1fGOhkc+T$UkHpa<8MqSKO=RU&Z9kY?JU}JK< zgHp1cNs2t&i^uaQ1y~+O=+Wonb{8kPN1W388gXfAKKzy)lTTKpr)>?lJ6QJTuhw=8 z2VyLdir>KJ_)9?HDbUNuDK0l#TSnN_;;w^I<|oBEw2@Q?zBz}1zIkbaZOsL5+>taC z%+-zu11mnV2w-3M?pMh^(S)#f!g6X+Obh*Y_$Z(%IuyQaPcs!qhm6I?Q{xtzY@T8! z3XB^q`y6GG{`5BVsXpF0{r3O;{xUNB$KOdNPA1m>)h5?q>DnK)-}CyOn)*w=U07qJ{AyUH2!=ihzxx9U$Wv}6S_#59KqLsw#xRZB4o>6YVPD5eM zFGZR5yZydDzxCB8F4W33D1}MbY4wK2%dsn5Vz6E_ST>hTlv(4qjB)Gin|R{=b{FZm zkn`QfzCYANR25;nb&hY$JonodV(2X%I<8X;?gy$xf0r0G;CtBsEgWiS`OCA>p=B$@0F*xC{yd0lVE) z6zBRczO1A0#wKwrEDLw~iVsU7Ju(`-)M{3@gbM##+QCMzW1@;+bMb0B%l*87jZ{Kx zS#+kIDKw-loV5*W%s;8+skS%lbzuO_OfGvvqmrJ}cs;4N2BZMrKt?PAHR~Ua1gouu zdk{?P(1JIY+T` zSoYSkRyeG6f@js@q*Vp^^w2(%op2$>0MG-mk|M~3L@tE(%Fp}2s><8)=*vurI zA>goMftH-BTQV4ZNOCyPN)4{g>LJWORrMoKAp_zk7^t#(q(H8!$X1>u92#wS28`!e z_+#5QM2s_R!}{gom`d zX*LFCEqW3|xys;ahyCZ+GX21#Aj{^1%-P+EckZ^*R@?bUJ^HpP8Y+yi5@61P(0)EO zK5hp&4rdr`9}czBd*({SX?W_v8%KT}9r^3zfnN2Mbndi1r(P6336ot+Fkc;EYj8du zDo98HzRpqb<05otPS`Ana@&kCD^7G`7|3)1WnwbbDgiDe78>5j5hvGK&75kWirqa03?<*U|5)wP|&n_de)_Eht_)$?517Ww|8ODtjtcKN%b?@!oBRck^g+oKsea>c_8O)fl^)@zTIFJ6JwU;$06{6+es(vhfpNsUm|`XbU5ga#tm zO5g<=&ZNmHIV-oLIX8V?%#3-$lQiD%bAH-4k`0h^S8={d>3K-QYW|e$+|&%>ZJncwa^H` z0D92TF-^zCRmckS;ArY(6@5!8i|m_s43x^*T!dYAJa+{I>WY?dE*C6?J!SBZ-*87eY`%QVs* z%V5~i&b42!9g2w^?BLoLFQ^qZunj&NcoiXG&{?bWJO&McgK!y+BD78Ds(`KJ{izFL z;QHadOtosq~!Ml(6rzbLTe&SQO`% za?PV|Bs>!G0l7VkAPY4Vw2F``u#TY@VGvX638u7u75AU?M^m{?dYxyXUC(ABS`pe~ zb#QFQ79p75q0E(ORoSwW_Jk0ZpSa(z@MX7?0_uw#b?Or?dtt~?W;wnpH|Rl zdTJ zu7lpoAP64OU|+UU7;0w!%7vKFQ_jO3i+p9sH>R*J3KIJ$96ygNT}3*Y&bFoia-myJg$ zlYq?lLHw0icqa&z1}EYAlcTr#I*WRC@}L@}-^di9XFz8>Bl~!V^9wAQ@Ob@S>YepJ z*g{55&i_@tr)X+AZ;vASoz<-`BDY3G8!&Jbk!v^Dlt`UO4XWVL@WTWi&=KLVfJrC3 zKmWC^??-C9qC#^oEn`&I*MIf#cYjQK<=Uy>iBZkGy1{ev?4EjLiE--bs&elGY1qY2 zj#}EU{d@J3|H$I8Y0kx9Y<#1AnQroti7nf=MwNaAA>(ME)%N>*b6X#_Nr{?jiFJgv zaFef^ex9E82FDj$w4by%F&tfvgeHmHH;!#)fddcQV>+0otbBdM!@=m=9AD(xD!zD} zUc#2d@Q4-&yVfq6sU;imczr~4=ge+9RkOgio{+WFubh2lF$0{9Bp*JeD49-ziGS8- z&Q9o!hOsHamv<%G&QHPx+k5GW^B2}0ot`Gt+vYv`AdvH%+K!EB!=+KCJWD=du|BC6 zY4ZCA0LE<)v+rMejA;q9WxLFk&?X-QzpONeFlK3q#5Zs*0rg3ev$nH3+WsuDE7%zp z0-}`u4S1N zxuV^Y6Dv(gejUp-hhl!*_VxAC zjs0y0ERlrP@6HD6cY@|$lJUC_5*huAHe;mSK_?52DFBLwjYXEk3uyWHMIMs;B7RsR zYlRc?PD&ngouKdjMaKe(q`Nk3vFI9#dWV>5t5euRiVbKsys(50uBFErr1yIsFWCAz zvT87+ljqNq;cAlc9wv27=f~t~&i;{?aB(v zPnTA+i&n_Ns}%h-{>m#3%tgKY1oigx-7A%cP~{B{8S+U_iT?0eutXC{1MKykabVfk zbvD#VXzyT!Yi4D^dC^4hUqRS59uhvNvw=&|MNyD^h$8E-+%f#o0%N#NShJmawqw#* z5%WGSsm?@cHhlt2era-T33B4U$0RvHvoY96dSfu>kOxfQo1(D#y#rIz&i9)c=(=I( zxA7z)Cpay-vw-L+rJi>~PDbxwr8|FSFPvST5!l3_B%(1+X9Nci^ZEyu7y4B* z)l0~34Yw0P!Irg~Mqv*FnM_m~W{n4pC1g)22VWe=Iwo6sTU*;J9W!_9LhL=>g5nGV zB#>cFzP#)c?lQE9{U$$wun{e2w4Z=_USo^$zemE-eH>fugV2EJILdV5rOnrhsV}3j~Xu9TS1-$RwST z2eV@m4fZe`n=gT0^kQ1k2-FQn`);m?M4XBJ8 zWSabv?3G$+=c?OWyOwZ#q;QmId2~p3v z1R2Q#4X0b@ELe;5nQ-2M2Va)-w%Ic{je&1!g)hcFxM-?AvMMF8-47KfJ|2(4_in(i zgwljhIaw!o6-1Z6#gf87J*ZOws}14?#om>-!7^0@i)3;FIN*~ zYDK;miH6qjcB~0gQ3XI;Rnu&Jk&g;6(w`5jvEpKQD3>w8hC0DUI#PG1Hjz?*_U+!@ zT;{19KkX};Y=k)xje^m}Cf~i>b0t&L%-&%gW-{jKF$gzz)|kcYBSke|w5p%J4k$L| zQ($n>*8u;Ey%I&KNC0xFV3>D?;fvK^orbweBmbQR_Ol z$5VC`mK8%Kv|MBqOeKH0QpCeLD?pbi&?^6nU@!!{Fg3e^Kw`v~FvE?$q>Rj7vajp6 ztoNgn-TmYQIOp|QYHw+mki#J6Gf5;L4}+)!e6~1}(M8}gA>&w%`3DAVK=Vr(PQ%T> zulyuzcklq{FQ=dIT6##-N#)3YC0`A=Hv@3&t4i}=_z&W%LKG-K_1b&0-KW+o zghEvLNr@-mc~H|GT@r(j7#aZ45JmKqE`Z_7`%#CdI1sX zUpQfIOaaocfPP%uIm5_Ea#DX*3jXF{yEitD_SBg;sODp~(W#MBO($_WF{f^YMZWud zH#Q}lCUZiOFbuGc)ufY&h%hwDWUfru(xUdBq|Mo&^&E|lTa27 z4QDYT4TGcX1D8j&Isi;ZdSQ}pmwppSXGalJ_a}TW;RjYxad2<3){-}2{2XYZ)>wCz z+D@C%x;rKmp2P?;_-4`@=A6OBaE!oV>1$##;^cFjTxRM*S=Vy@=-FLS-!8Z=AV@k3 zLUNNkQ@#_3DmsiML`T|lR#8A5*}Ocvpc!<%AKaX_VXORFl(}h@6ihD;b1e*s9iPrg!WQwYL3Y1hs4UH^{fl*f- zbwy1szP%)Sx9-2Y&8^&HR0qD{Q##&zlFHnbBZ7}NFQ~Pn4eQFGufghslS7Dl9r&Ys zE=wC-xQ2)jIR$Z^E`o`|>o#oasjEr=r{6FfgXjsM&~~O}mfflt9_E%?vgg+65JuvO z%C>iSobFsHus9N-7x%W0S#IVQ3hM=_+TzP_rgo)laJ%6p(d$J5OrKC}s6qj#9NC@V z`CEjUe)`u&de0ANs93HWl`M=oo7^7~We^+jlm1Q9`|^2A+%fQBBwKA3BF|YyXd8}BiFGvdr0W3%F-G1=T`no5eh^D#ns0vuiU}D zH+iA3g7_zisss#M+`)sCz)9FKLozY_T_kZEp@&^jJSuM6ljW)QKI7?x;wpJefv|CS zTeSRgI?Qq(mYVw|F_%s)8?^q;qY4Tt0(aWY&5UiBMs6(IVJIFQZ&3SG=Y-2aB)3_o zzzo8?NvGpaScTz*EI?Kzjj;p;T&T-Z-=IEH_+PUe@IuWB^Oh2u&?uZ=&&zOX!(S1PTnzh6g9qebrz))|I-=~C9pxT zoG`6Acl3#+vbv6?@VBcS_-RsV4>E-uasj?8X_vdWj`qc&f)Hyy4iEBF$7}9#q}hUU ziJ)Itoo9NVgmUvv%fAVQ^GZ0z7m;$JqWD^o2?t@e6*xghBh~!#(JXw;D_;yHCydlD z^g70^hNn0jQcoQK;Izr8-VuFjpnT1`9o80_qtxBhHAJw2U8 z$S-WI3Y2#YUs|vK8*|&&wmc&~KESj-;OVo@5wCh77Vgb9K$JZM)|}PmGp9o;ItWs< zi{PJg#HC5qO7>8J9E~1vKuv+%=$#Q7fJL5^g!le}Tl^*7`Y%7j@qhgc69e=A9#o*A zZGXuD=XYBNG5aq@N9*!bI{MMU%pINqczVdFOWqj_nA{B;I4&nLN3Y)Bk3TB$viQc$ zP5|LjD1k`SulQ+aXJxO~sv}VeWwml;=bK8oil&KZ;_0GR>l-C@rdMxFi+tf|bFZbLZOHET{1yxK1k=4&k$M`Ao$1Ary(>`+f8Bv;%gK4GJ=CYP59(aj| z^=OuU&vQdKQuoW}71%9n*1bxWOeM1hY4@tadURkZ$Fc-Dp(IesBg#fD0Xl0#;NCT7 zqnux;b!Fc4AC1`6s_?W)PkSnDoKr?5yG)fThLOYsuZdz=ZPXE0)0oiAE*Z3!XhE6s z$4jXYGc`nl_JIVRr2?sSD&jV^ahwyyh$T0k*bvhm6*`)CVt0sK1Zc8!TtnAQ85Z8o$X&eeygR`pz@|C#7 zIZGsA0Cy7EdX2_IrUoWuF-^WUq0IsD;VfekR}guY+X5|YoOtF0HT79qA=MZuWd@2E z=w>BZZB8^}1jdE!o~nsdhS=vQ=!|9ccYiH^LcA1miZ!MyYFkxexn8;;=N*A~81}`j zX9SV~EdlHht2jd}zO|joj6$pJts+4@ z4xfP!a%;p(+R#8C_-ujaX+huP*fq-bnC=m#x4R441nGKv=+*;^M6<&{Wl3+^`1M7f)P)-D!cJAGRkWLQvm(TOEc$@nZ=**Ns>r*ze zY^;pAO_z?vP^16#bxogQ14<>1(Tpo*Fl;0%%)R$jU10iy=UP6Q#-d)&(|4MLc@aPnm!;z0|zjy<2&Fg+Y-dwCvXG*9c zZEzv0WuZzTB$WQO>gxOWduF0)h-47*XF^>*&ZTdWh}J^0>*IB$ueR6uw>Im?KTYlQ2d#jWl(t0B@y5Y6F|$6!5OyUd&*hRo+gH zm>okvFyM!uItpn@YQO=u)-Na48z;*}_SaFG?igFWN-Kv9JPJ!W>+ zmZxQ=!Cg8|R%u`yqA#0$@mkxN5^HT_9c% zWrsz-JJ9ef;ZW4DIItZ8Ap?RHk*11cYwVna$wUh+8CWeOIi&AeGSz+cZfXM^+66MGC(IzrG z{4r6SeWteyy_Vk3s3Oab@VZ%#3*iR4s*l5`1&8nP4_aCm=}ZEBnclV`)pi@6s=#Rg zFnb~f0Lpi;Ir%dcoLAJ9QRhOC;k;u-I&IAOa`lVU`QHgY1s}r%mcQ1nZT%eRlP2aq zt~AQ}sVL>2C!y+6&$%qbqIP5eK4-cR>d7*j%zu;+Z66`3+#&$*O3wmF8HgQI z?XTVdM)0)I5qs0!isIONpU&@f)VD1!ZvaW-i!bHlrXPI?PQ!Nsy@{_GdA_&=QTy;8 zg@xZ0;QhYWPB|1NqRaEY&jvu$^t!3JTW9bRt^I%%0rG%$BeN+Uc;MSM750DOE#Rwg z|3g18GX7uvz|6qO_P@?UYqWRbi#U+{^ZJOeDoC|9PrpI~lU@wZUc{QHV_nK_yfv<` zfOG!9X#w>JnAP6a^--k}k~yA}4M`itIHvgbIlWhX9UScJ@k#X_^2iiPRO7W)X;dF_ zKxxvb!)LAB3a%cvxiiHE`R1yMJ~#6wKV_?8OI0^PAyavD_aI9eMB6{O!pq?R7gZ#1F2To8W`|gw# z0gy@Hisf^l#A`0AZgS#1Z3sp)EblB0eyl0`m#{9*WnHWHI*OI5tqAG4_q-fzc)_G6 z{RQ*5ST3lO(=*iPORZVI3km?26D0X1xqNONfg`Z%{F>gbq}<{InB z5ukT9F zAO$DNOn1!@7mAmTE;Ma*Za1o!Vdm^S9Ia*)-Y@E#(4W!g_1qe1yRE|pis)Jf4ZseI zGvJ=&l>q(C+kqymUpUOH>Zxkg^c!JUMc4*hx|oS$QDI5$;wMrpZD+GkyURwB*Nq=s zSHIV}aCiTnqj`eX!zz_9oobf-N>2Tn*IU2e;2O3&sA=wG%f=5gs2g=C1i&`XzfP{d zH+yHQ61b&W@~VKL6tje4FfRueO|;f|hHx(qB?fi8?N0u9V8#q` z%oy?-!#1R|=O&Kj#yv1OIw3L2)Gwl@qiM=){PIm5A#A`|aa?r?o+Uc;8FA~HI z$x{$NAvz(G6rqukt;k3kFf6H(!bMPae6ow}Aw;~<#Ikw?_fj2vXyQg^^BmgY%xFg> zF~}`2v5_oEVoOo7FHVDp_UxZoGb~Bc6wC^oi=})#smMpyEr~rQok93lLMK4JO(VJY zWvUdWjWGUPfO!OD?^y))kg{M8I;v4tmFRkut5ez3&+OUsgGz5EomVs-d-J^J)b68o zQbQ!VWCgRUS;(-e#iHW!RL$@M8$rW%Q#{M6!eYKAexgTiC(O*v2y-~Mq;kxMXu6c^(nZwntKAl@Naid|xMKBES|MVygCvq;TYG1d&Ug4Q6~R~&>CGKqB^r#IAadYHLt0nClWZ)q>`|Xy&OXLr!}1 zO<3TugD5|HR+5kuc^9TM6R6)tA7>trOtzxZ0Sp=mRg>0!7g`f~%>N#CP8&{RN? z>Kyln&%@XPrJ6*gPtkAB)Pnz9h7-DRY>ezzEB%Op=kM`0mzKw?4j$H$hTPcVOAqJ9 zDcdjr^E74CL?u(ySHIRYAeEFx-QyWuFd+!M?|=o9_TfBUBN8wo2{v4iyb)lS@YTy; z0kgcSe=*MjqXEo56{ErZIXAZ$5VB>4kS@$)cH9By1KNJ?yThcO?bq3dp%;%?-!=qb zP)Zz{WrzgIa=}<^en%Dkv|!kTqVP~vgc3HIpdx|xWZ^j&vTLHhp^mBtRy;i(Wk@Jg zh6f4F@=L0P5J1QxA#ffy6kvaOdS32zq+<)^kF@SBKFSai>Z&7o363ajdJN%BdC8fZ zxcSEui`V&w)OI}qCZXIZj;DDZBOm#|sAR>yHr~^K&|Q0*BP(}hq!-#TKCtf{_D6pJ zL-XiBK0E+WNOGvYXb}dLm+BwA7FvlBEeo#<@1k(dI*Jvo)zwd8_ss6$~<883y+Er-NeE3@%zA@I$NE~0?)r)7obsQmmYDo z5-#b*UYIVTllRGK{X6<#uUZ9q4!Q;KWR0!8TPX5oxxo!a-xvb%5$s3&KbZ;PGkQZF zy_I+|#)y>`gb~KJaz>CxGa#dDkk7(bP|pTDFy6l7`ieP0{+Rn6US4;xNOy!*gaIuK z;sqdgWsm`N$rs}5DIXPlENsb$ry$ne%Wd>*s`t&zNy(5L=GG?z%981uB+v&HM$YRJ z9)7v~k5UzWJh~NmmYuHW56UXfh$|jQD;u4j3aFOj9Xb%I>3`uLOR_#8RsR2-S5l;2 zq3PBsWis@^U0bd26cm|p=|QMyWHkC-UhLAI^7SZJ;47)0iS5$*W%g&mRJKBcVy*2O zpBcM%BeU9Y)?3y(`$2W_b$Oxa`UHlrPCX(FXi$RaML?8KhbIM)URz3JD)|YyfaI))NFb_7OJ`l~86l=}QAXiMVm#&jwNf%dt0poU_9iw*`=wqmx=PIqEy2gI9+=I;Bc%Yj-V z!g5|8{H3(#ZVCt&?`$&QW7LNF&H>daE}QNl3N^JL2S7T#0e&0(`2ap$Ai8qlzOJZ% zKh{NUdX46wDIR#!Vvw{^!vXKJ`OCX7An$1Wy#gf>d-Pt>T!V|%Rgi7F(~ipYDi(uf zH!Q%XYikX)O;wz2-_`ki18o-`YQMOL&SWfxWs=P4!8*qWl*VOJA7D}|&!_D)v>Yk< z+9{>)X+<@5Dc*`*){0HO3c)zm55TFph4*5@D233RE*;umB!=0rrrQ+YqvkQtNxBs;(Jb(Zeo7#*Xtyw(st>)fR)# zQW{1}b~jZU*c_0T=-RtYC|l#J4M{))u$lgT-5^N=H9Ti^$S?dxXj9dH8IAt}FPIqq z_d%tWruDG|lHY3`1r}Cie~Ii#LGt$W&rUuRMdh=* zW7o#EdR2a`@BrCCN`uPuC<%dzYEu*0l}{KADoN%r*`B>`_orK_%GQS18oAd(osTGo zC2Unw%UAUuZyqSU4zf5|x;QCwECW;=G0f*5wj$T%s;$-)l;v2u%(gq)=u39RF-vrP zSpSO^KPiIXaBA##QSz?xTE@!`t|+|K&LUeR!hN$Ss?4fn(S=ho-q}bbiMJ z?rG^a9n|LX>~S%;rTc(?nO`>vQ8o{8sDgz@Dy>?Gvm4EZVH8Rp4r}X|ibffvKo%X) zpre*%sCdu1=UQgY`CM^2n%stY8|TTtrs@uoxZ3edc?+fVu5u5&PjGjUVyfp^Tzulj zg8)8p?%?7^Aj{jdAtBFI*4R(}5UJ^Z#jdU3y&G3JYAR&5MywnS=1imIg@+&s>J#jt z$fUK55OxVSlNu%>#UN4m1uP4CQt^$2k`H-ZxFzb(GcnmpEq}$3E~W}YXD~;tL2EQm z(GAdv&UHxva=JPGfejg(b1ck136`Bw0+Q)zmEcF{d^`ZdjpPsw0J!2S{7GRy%mF)@MGPyg(Yv`bi7hyX&t=(6saF>FQ-3VG?-=s zq?0IPZ(#pwq!Dz&2COIQ8);*_$@kJ|aP+@$%6Sn6IwzkFJ*dkDAf;dxMh%g3Z^rDt zf)~7QNDBocRg|$0+nzf~rvh|lFc&OwvPkvFi9#AS=AQY&XRRkA$@&W5mihxAbGLWZ zskO4RklPMBp@>kt_Mj;3JY!>LJHNjJ(I|(4E@B)8bQTYRbOC7K7BTezu?X*-ef$qF zlL=)I104wW_j5e~$uB`#(Q`ZHLdMO|k*6&~U8mr41o3^+HMYOgY6G7ud!A}-mL zqZOh1Y`}63&XHs57y$5YrG%xVy2_XF!(eky7(|YS;yOmr^xyq?t>s*d>!#N;(}Nh z*5;0otD>`;=pLYXItN>9<4w_`-YXpF`A9gJj}X*SJX?HFoDR{aS$6X! z2@TX~9_zNfbI@d5QK~}CiOaLWztgZvcly)y0jAGUbv2z@bd(*kP34FF+Qkv!wR5e#M$uKfm#DdY=Jhn1x{F@>6+=RuFh{;l-J%l95cS2D&t6EF{& z9Xy>_845jqu*9PDJwT#_vkuezql=FY4*!8@o{4?68E z9Qo0#BS48l*GF9i$oKKb$r?YPC=X~1WgT1(f;*^9YAr8L1w=E#6-S2_Pauf0eF;VX zqml9@U9rmm$0H{^T)AV4Jw26(9&L7Wh4-mpbdWb70ygf73y+d}gBtwg>+hYK{99Sr zi1xy4a--w8277aedZL-}r6aJV+ivpvjwL3VzD6+sPKw}QANMoulzhn{b5zr`9JNV4 zut3h(ZP+~|*}(T__mn3j)|R6bDAJ}l6DVm*PiK;+WBf+w*kGs*b5O-ylynK!lay~dkI2=kjs)!+;N0?L)kIj&Uh z5Qx^F0QP-K+bu@*YSO5iQPHal*2K@L$-V^HCfD8-c(?G4|8+2Es69 ztSDxRRSnR56~2yl&}5V(^l6QVM8iQq8p|r-YcyQV5dBC*pF!vjC%Um$K!HB=-HG=Q zs&o}q_vyaxnj>H$a(ze(o53`^lK^@a9cE!Ixcnw8!iH01FQ$gjOahvGW4_0FM+8`P z7*Vt*0e>3b2V%v{YYwH}YbMO<^GG3Lh}$0Wo>x ze}SepHBI-a!D1uAmuphNt%J#1UYi=U(V_UZ?WRBP0R%c6TZG5jB&j(8;1HLDBpB4$ z*?sQuU3*c=3H=cRg06(Ev>FZiZBy0WfOz;xkz1s=-_zVx{S!kfGWpR~)fK&&^Fi_4 zq#v6D03`o06k7nc>y$Si`&xXNaM)moUg=$|&_EDPGGc_frqS4Whxa5vws84kNK;0_ zXvrZj6WdRc^nf$NG%ooB!oHK1l5C4yF-fWwgCtIY|F-95=ESmxF#5s@jDJ%AM<1lq z$~4G+1q31GYU{`)a+NuzH4TwgjwVbEDlKTm3@)9MT(R;-x6S(mD+J{DqNZySv?WC~iWWD45!fpPa-<(&JS#?{%E-vn&;LiW70hSg(fs1nc5lCp1HG8W zSfJ2|h3;!{=%&%TS5xU97Fe_n&v9}4^>zSGk&&IY=Iv(dl#f|o0&$rg&ORhWp<%@;w8iuRJPnyj{cJo$xh-jo9{vd;dZi>tda(MD2eb z+g%g$<{?Wo9|mqJ;1Rf7n$=Lt*5xko7j#8VMDf2I&3~X5|IYnCqvSnmvi4gX2t8-& z7=S?I0@aCy7TT0v=Rn#9uUAurx*#!9@@-<-+nb$)7>^&d^K(`cJx2aLER=u^e0`|F~rbMoRNEMseofl$Fqhac> zx*7bd8wjiK0UUE~ed8`oQ-uXq96vEd`G9 zIp62Ts6tnEecD&7w%&JXobJ}=7!=e#aKSaI#D_rhYem7$eCMk(ZYM^Rz+{@t^4-|I z^-d<9@v6HULKcV+@RpEOUx{T`e(~D_NaIHw$L-d6&L zkxZC*OEp-qw9enGmsYd4p;TIS=()W`OQH%u+2)$Y^tJeceKe@xr7Ik*>mad+LBLdp zpaDmvfs$^Nxf_}`&;d(-1UIS~@OYMA`u`|kFOfOcv)bc!80g=aH`CYn8_@E4gsVHKCFy2 z(k(@*)LkPtGM7Z^HK|V`^Vkz=d3cJCURo5mN-{C9*e_|TqRW-Kv^J!-vPY=9H6Pff zo1Ig&-Vqk*zR2*w&uT-XaVXO_h+p+2kzb`=r)xgBa<%Hgdjon&>UsIw2@7q1#41Xn zao9iVD_yGCLytd{*18ff-$3rrLL5P0i8lm|Fsqk>yL}Cb$Nx&TM~kNlFV%f!N@zeF z?EOqYu($LG?ZJjZqmn#oRB|9>;gMnK;hx7OKllCv|1EEu2y_L8&XpOKOWfk;%AC3+ za-Z=8b9+}VMzj*K!lVE-t3d8Z@`FZ=`D4v1S(`?j6f&1bV#v2|$D^~z;Snk5R$;|$ zKN@<04<4hVjo>4?w_9_Y%Cp{D6U63n$;70>0%B@?-8!S%4+RJt16wZiRQnyy^El)` zEex-%!fS&GGBgW+M135NcILt+tG4@?Qw?^#ro3WJv*O(jDrAw@0Woo0WMamoSBBG= zIdb)q#SkK}6urle{ueMZyVCzZCKuCxn7)i`tW5u_>06>9ZNL4GA@Xl>fdP&8*Gfe# zQ-{LKE)dNWq<q@$P-*MAhk!nlkyf)&@!h{bUjAbtL_ISPw)sgkh zb4?#{bhdBko3O_bcYI!*{)EO!bgr;9vl(6#QOm7y&Aov6 z5q&eh8$Xc$CZ<)KMCr1(4z+9sJc`+}Tm;}>Y{%Q8glmy9ly;w&{aRT!@=#p)vA_0o zMX4$fQA0;*=eWGt0YCA;hgG7T95O3w&f_bVS%}^%T6*7Q_>pLjRKVA@Iwpbixlfx{ z8Ed1JZANAuhRA5H$B!8+H<35K&_*Aai;IozJ_zj^tJjPSvEJ@)0;1TWxOhX39cmqt zOg}=PB&jh*GmMAOy`wtI* zN!6&vK}G;w3# z4*h0EgO#U5<4~*Bdyy`)#K=C@_Cpjgt^>U%#H~u>P*cp%fDM6F&TYC4rw$W|3joq* ziqy&hi?GRf)t$O69cZ9}5bdYK1!^NjU8vJWyBXIyV!)gBOo9DmuToId#JGT@vBx(| zjKXNuY#T;4M-=2QzG}6}vCm9Z;aSQ43~lek?`P7O{CJlREpYLU z%gs(%D@Xg4w3k^kF)tX2*BL(0>^Y&&$m+UKN*%rXBNX2Z%qvgtFGBBbbm+|?$itKK z9Nd0MFx-y>0csO7l$lYagWWs}^@R3bvYZn<0j+ zKo7H7=IN$!;i`GOj52!NgS0Ei-}1s4(wXuR&bbC#B;3JR2^0(SC0L&>-IFazrS}YC!$<6WGQZXcwfSsT9`VjfM ztKnryx24dG3ZN`eP>^4qNhzaJNOUn;yF}14H~W)IJ3S>IG}r1^LO&D3lPqFU z$oKDg$5DMS2f7WLO!oB{5!Z5eG0K6d$VAugvCN8is zt8TGl{C(m}20W21T%RWSOjn){c#!7Z+`RSc&0Scb+qp*6c3gp$BOU4;dBS^S6&dpE z&i9J3#|G_ZO=t)IDq^YuwXLDtx4rVw2lc2A=*`Ls$@kg3OUd+Se(yhZtOCj55HsnX z5n0ag`rFM5LXPdjP%REm_npsili{mxt#XNrTAv66Ga_OanbTyC(7w1_p1C3&=cG%A zBvK(#scpGdFOB}p#kLFI9GNLk@!kkPBalPN5T&e`FNZgF@6S&?ps~qJm&W*;Xv? zor&+3v-h_FLypD%d=DLNRO1{pQ`KJ5J^h%s2Mm56rkE`mZI%@_%10F|u*|&mgQB4e9u;|8I$DrBW!y>%Pwu zz|I57!ZXgJ$s#(CZWvKZ*Wz_u`R{oqr2S)UDm>3ei_Szd?>U4K_igljp6Vjh`RvJ- zpj7=BqDHIYNUuc)ukh*roSI-?aL!!lp!zm!fc2g|f2wRBFuBc+(`+i`a-Uvg<a&mo#p6&Q=oA&s*{B1VRYsR&|;`hDucK zeE0q}JgI5*I}bA)E;#?3ZzjL5sL(wYKL2HL;dd*M5hQmtwHo~E)`S|#vKvj}lE=mb z@}Kj#-;4Lz}QBuvbJ1lblRbkt5`IjsK##QxDhc-+3)TJEU8D4 z#?N{mL%3OK=;h2Xx2e(n2mO!yfFK524kXB4ksOHUAg$@8i@B=1%;i*&rD&|1*ACir zh6up+b&_g^A0oI7D2d7n>$GWEV9Es>HdPepxUVzWS#~9-9$VyHd$#R4_$;{Z=6f7i zN7VReEt3`umj>|s3F`C6G}37&9}$uR$@4tcbPuBQoYAf3wE4h{DC{@h(48+_pZWe9 z&3App;J5CPgp&Bmhxur2Jq-vWIOH(0IR13gS(z+44X3Uok32d9Sp#_vt@paQD_w}o z5(^-?#gnh5i%S0D2l9Ne77V2r$;M%7DegSHszI&wNGxDgkaZHsoKAQT-qFq`FQ?4} zH@vPRMogDNez$?Nxu1)utui#CDwY%mgV3sCGC!NvRwj>~2#H+nsy~ctnCslI`A#~F z9xy{I^F-dOil%*bEbBD&i;~N&(kmYdn@40Bl>T&_1VW^w(79|f02=@jq0g3$$R$sp zu*3UGVgQN>^ebx_3T29l4w&476WK+fa6?M=zZ;bL0AcMnQ}@1%zZ6cUG`7Ecm3Z)8 zC|NinflI79eFDZZPBk!L{U<{3xdzhN0zXbsu^X6dQAUpprL%aWE~}oP&6NPn$m(9F zI>!N?I|F{#e^t%A+!*M)?hPsN9p0{u8m6r-b+LFe{#~J3gXGCnX>M*r;8cnUDoPO| zmc{2ipWhGEx*Oo_{)GRdh`I_M==Gf2we9}>obtLY#1ZYEeNe6;7-ZcC4-NSkgJo`y zE4W%V^QTerskP_oVyh6TNLz|Pr|!k8JP1|iEZosw)$a%GDctgI`^`+1(`HmgU{@6K zppnrJ11G)3$8KLS#P{<3W_>n|kbGD+5Z(*K-aTmbJfB`pakt*; z>)Ve8glJUl*ID!{d^E7HH@pA%JnrT5A+A63df4sJGkt%xh$i<28ti3Q23lkIiL#F`01^}d=? z*VN#^?*$+bYOiV0o(DM|YH_@cM#i>DB=O)j=&{c;!^tn*pIHs3M!rOc5fO|4rfU3k zG|a*h4<-+N1C-k7?3wTwTNIyXK17%oLMf?)+cF0wmO?mMAE$sHLxSvmQeI^rwRTV( zl#$~UN$WuSD4fysPZIF41;=q1cNmi5wf2@SC(blWS#Rdh#Cry z+ikesLQIkDEQ{;j5}+6{<}QB?8Y7E>SPXRl%Q|of6Hy4^e3r=Ygn+S1z|#qx!-d9} z7|uZ>*~xhUSp%8m*{l%Xck!B0C~eTI+p{Jp=Q8dM+~}%s4TAt3oHyq38|O!B!93@%s^cB+CzikjLV(7n4CXMv_D(jG^&(W18Um00#5K)Gi{6#}lK3-5YPjFD3Ag zyt97Gp)hg$CY*?qJDyGGsU*;I9HAd*Fzs9*M_U zfX5JXuD+3pNMm#Pp~R{NeTOrc+9Avg6aZgm^Oyq*h1%o>ojUcz*=EMU*acUfW{`2`gG z>tL|^=l<^RR3^JU7?;`Kx&kUKjLK1Ozh8D~r1pDPn8Y%>B$8?j2^8O#o!^zc8s^K( z|5|kH|M#L}V)@@AKCu|q@kgx>Uq4X9{eW5qH4#WJQi<@Cbh3ARBm!L@3M`@mTaB_g zG#yAS-oJ0T=r8wH-r&G!oH%gdL!N(ME_UAhr6*&R3^Pj~PA}RO6*0{$8z%ly3m#3L zT-Lw29bT1w_BXAy+sd}$_ik+4@U517-_pbX^7w9wWY;VftZu+~T2x;FE-kA%GwkHI zO}fBuvB&{0J+b@vr~GK6hI@PlO*YuY8c*wTvxNW37k`PTh`sy97#5#yC7-+oJD!4T z=~b5^NArsr>Q#Ox&M6ocna0pGeWR%0=~Z-7ikw&4 zB-d(b<7`CpwrIDr;WoG*)gEbsGf`rM+MSUdKHuubk?Sd{Q220tZ&>Au;$w5WnoP5a z$Aj6~WI@B24P1ERnb=2FA71IGrf~B98rQJ-94AOnGSyOxx7QgjH??LTEYL@xe$-|- zUt+thIqwA?M2>j~LmS2pc^OshiWWQkS5sHl)sUpSfv(Yj9RhJ{8!@;^fvFn$=_^8Y zW@PM4gM2Tq&(~JoTnQ@&_ZmM@O2iOb{0v>tKD(!nh5++7B4}p)&DzT6=tNVNIy^oq zD%so7Y)W0N1#gsU#UePMVn7xxit*^$unI-vj$YT*irZS@Um9qAt{fPW zmED*iF`$8*G1-I2H@azC`m4XGcC#%VTA^Tt$N~NzW9z}h4d<|sABl+j{n$fMgnIz* z>pLJIj1Y~0i6nd`TcCT|g3X*|!4#tc2Llb8=MD5i zuLd|+)ltwlYdj8*M(CifPbk6exsgkXs@ za=WkY3om`ZpSsbF+#1@aK?CZk0PsWMHG_p;L>Ap4ow}ACY@U+00Z5+SUTC|x{Mu;& zRxBkI9R02|yB=~i1S2CJn$vOseIND&#W7~L2nLK5brE5K$1_bP!=GJ^Rp1;L z@Pn}NEZ)YC55ko|&IMlQ`@iPW`^>~Ig)H9Q-H__mImDVOF7w-9&wmMwe_d7@E-TIC z$%-ys<0V+(Z8SoeRbb0mx6M%U4VWi|wj3x|%WdC13>f9lCX*2wR0c!OWQ#g$&iQzJ zGpt(}22oeyT{A>O#}4;_z3>9b(xmfeDzk?(H3&v{RHC!HkFn4_BP}P^Xc;C50I@P* zGh>l{1ORmeR+X&uKF7gSUFYB_$J`)!v1iO0PlGc*-HP0PEw9=Dl#@Mo*@b z4{dI<4Y@weGFhXG#BrwGH8&2Xaj_A%+~FqFJ?(ef5s6NV=NU37U524_W1w(rO+Bb+ zg;L|4etoy~1yY8pfmp)N_zD6;qkovptQmexq1=70?)YI@^?NlSeYO2hg(N?<;VpqU z(}QSu>7X3|_J0h#cY5D}5pA%4f46h25V&or#3{6{Gs6ioq$G3+-!42NDp33a<1v_Y z_QTJ_ZU>Tk@#&s~e`*WRVNk}RF5=zkr$%-L=#&9|z_z&bZnlcNG2wIXH?*%P9xG65 z81-z$CxC6*oxo<+Q;2|`VVDdXR2v;$pA9cajHMh3>&nR7GG{CQu%}b9r=hgkV>;Pz z<2lrJV|%SB6c<#lPPFA8%)7QD7|@^8iHl>oyK~%eeum7NJGO_!&lK&zM3Jq;9%r>+ z7vlX{lOb>~iy+a-ge#WiDk75wy21AvE7~cik6Qlsz4sQNCH!r|zW3*yCN7gW5Ng}G1{oh9 z^@sBPO6M)YAdIUdN~vD#s52OoZ9l%S!`w!WqEXQyf(Gtq75)(oqLlk_0yU(45|jW{ zi?W^oM0{ya*{4Qa(oha%EDq-<*I2%7J5-7)DOZkuqd5c7;K2#+7u`}gtRje>HmOTQ z!19qZ*$|Jy5z~$?l-SNmAby6& zcMZv`7S#H|A00Xq$aNk{CCwoGNbNYEjuIR4t8U%g-yg?UUWZX=A89NqA0j4fPv6BnV09Kq@+jJ<&MR^iErXEh8+JuME>*{$@f8qZmx&fQ8cMXPq zA!-y~UAiv#pLOy*h1T&T!$9ch2W#Sf0A{=#BZdGTqH41pXx+A-etm!aIUOHL%0PPF zu#U7+3kzM`?fL0r;*k|%{S+*>=ITOrjVv$ka?FQXnpFEU2z~Nh%$hPKC<Lgq#)J&WRgpnN%I2N#4}05?!{?8xSNIEUGyZopWY1AwLS5@RM?5 zP!z5W+aBAr>wSstjL^7(UTrl&j*OI^z;R10zEjqsM7oF&yl9X3yLz;%?-@#CK(c_b zF}nO4ORj2+xa_81=7uF5%<7~tX8uu_(P0m6tC|$9sJZJJ>U071v&pu_&L#ZA-W&|X?BI4HmjT^xh?{XWPgQsGUD|H5qEHl zsG*ccHo#9Y3B1*dDG)P`EwIgQ5r!Hc+H`Bw;3;%dR-v;oY5-xWvt7c`6s}h_>Y6cT zoH}#Ys1{RZSS{|HqV6~3mdv5j;J0&G^c_$e#M~0PS9up_MDhL9aZmR;zrK_?>7(Y0 ze~#kMhj|GI6DCAoi%%MqD8m^9*vfK4&~GCdU6cautj5uUnq&l~1rw5x}Q$1VEe9&hjGR=lJS`j}y+>E-Rigs$Vx#U8zyrDllF)J>4|B`a9`rq8=0ZjLPZcr#6n(FtGf5D%ZkC08^6l5Oj{!?0*W2VXn0{ z(~R1V$i?~>$HrO$g3xPB#hEs~|rn$GXODzEgG-Mp1P?62(&>>u$TZA-CdvKFw*s zxe-_RP^Mtv#wd-%6J@j`{`XE8wlKYsBv!Zu^J3j)AsZ-lR8%94J<)6IdbKh}Z0`P2 zI7Kwg?Qs9YfYKeVdl`eJxZ{s%*!8HX>0E15#&Gj98}aus*g>=!Qu^lk5-fHCCO^$5 zf$Sg+U5{gI(U7!U#9*<6nH)g@jZ>LAMS2WLK7g%;nAL5vIW#1EM-qr6+%x7-Dw#7y zlZpZMXcqIut{%n>t>?aO#BCDm0~wU^Id&l_h{ps<$sh@ytS27^J(A0y(k@8h31TJW z({Dc)jtlvwcua;>Uc{-@zhdDm^m0y@bQa0r{Jd^Z9Fa#gn3i5jJErk;wF)xG5Nau> zcQ^%*{dg8CMwo~Tuq|sTWF6Jgt(MniXXXBzVC>5%L*JxdPA60%$t!sz(&#~@*dC{4 zi-A7OFu&{G&O7ZcT(2jB6{M2n-ZVzU>mqNqfnAo;(5wK_IIdNu_#rhJGA@fRBtu4HlyDN6^cy;99Xh1NBNgOiBBZqi|@(nKm0y46o zmuzus+E!iw22h7#r=;>qA+s#31y|1Qz|(=1mp0kFb@8*#cko*4?P3b3o92>l6ID83px@w&M<;1s+{EW=zBY`3%mMND~&_2ZgXP4&|otmWmK>!LI?Ycu~dY(3!)Lb(gl-BPiGdN!6>Xm>tpSP*-@Ra zVK{54H-7G`;&GIG6}lqK!1@Leh}sOkVs87dpxSP=wzA|8z@QX+OmaJMf{$jW%J2cR z=W&m5LZK+M{?y#f#7^!TD=FrgQNjA2lLf2ahR{AcgkfXYlY+$gj7)a3z5LB)d4?M_ zLyb9H&}4#nUarm%>N%~tbI^A}3VkP=b)LaB)8yTy8_A01-3%bSp4Cli>uyASS7F3E zun!?ebt1;{yl9c1iNruGAlA|9j`L20Vm`aD>v0dKaHElzz$a>SlnG@%!O}f1A)W;U zC=L&QFD$pMP1hzpy=eR0Eow9A$9N-Aktp1oeXOx*Q2``6wP;J%#${6{Fz`H11)xDt ztOc~YzY)kUZfN>#VWxV_jkx*4*qWkIw-F7jJ}2Ux4O2zGeIov>)5F;CGSM}3$V;>j z5QEel(x02H(ekNunfHo%C^8#=Si!jrgI?%ABW7)#;gjtBbq00)de!Ie7I!TaND0ix z{cR^^>8(8Cy`2DXaHkH2MpT*tvl+&p5g1l@)Gn0%;8P#=uA z<%Zz7VBUTC7(BPlZ1u!L%vN%-pDi5#8GuVj8Hh)-s_!*;j)oP*m!w`C9T!Yf6}iVD zK{MCJ++COnb|N1@+A&8~BRU^H>O^_XZHJ_8&y9t3u6kTwv)IKzs0SvrKKs4e+?lN4oA#NVc8#Bi+;?%0(S?3$%m=s z9E*ikvvw0m!Ycs8v)oL~6q(;C8wi=a#zDWgU5D@g_4ZQ;>3nIoeygqB(?RZ8r&(1r zz(9Pwl4-LoAFJ-Xwo#g&pPlt=N-QR+-Te^&`u%hMFf|%ewl4}-b1?Ixo?ztZJFLE- zuGrQ231K!ss6lU;FlU(?o4mpzcr zAEKyv!S&_G84kytTBo3O;AWQm7c1g@;A=iBJ(rmDHz5LkPz0_h1)A0oIPmtzE;Vn0 zoUz+FeR95#cb_z!I8@ZQ3Sm^mw-lb_>}Wq_HLet2WS|Ur_YWFGyi4(Lw@#F7WTBP7 zc<0Fe^e93uFoDts4>Q1g`+77p&e6@cL+YT5f4-P3ee-=mIYr`tkNZS+RKxYnZY3|Q z{iV1vFEo^>(JsUfqA83gDX42b+sr3=>!5Z&%Sus!m0lMI-KsC8?(5jlNF4;RFXIBS zys3DTBW%>o(HB zqB+hUOA=}2#)@ORJSWeA@kj4&CpFt%g?8NP49acICIw%ko%A@)4ajxLtA+o!mJI%( zO#=T=P0O?f@c%^K;QKA>N-q<)w-7ddJ5^WNuB^=&0N&t8&{~x87Cs1-PO-|Lwfkm~ zO>`yo%C75AV1KGS*c*)mD>2MM%|<5+c@eIy#J{Qg>rrN;QnxrEym#o4dQv%+sBaCe z)b&lM0y-N|8tU|Rg3Jz0M$pIxR;uUM{{DM*z4|;wGPJ=HDVdDX>`Wm~z|#!#hS2$i z4lUBT^Cei`T)Hc~Y&8i$AZ8f61*H-Jt$$q!JgYaN#S}7CE9Q#71~E5FpAotj ze(!+pwSs?!-Xf#7IEqLX(Vf&-lwja|AReO^z(}3*#iiyNWuSzVOm_G)xF+!l=l#iH z%FoU#s?QW{MBtH4C=ZQok805zS2a_{Y-+w0YB0y)x%dKVjOP^LrwC-a z1Tx8iZ3G??IE$YlA)Q{sm}aT&wSoun1=!TB{w$D-l~kUKD?~t*XL`8S z!7Eah6Z&hU&Z#5#Utx}yuYWSBH6Go2v@9--bu8>zD;2yig&bWg1Z|Ck3EENA{E^Um zrb(U}9nZzy=8j&`pOw21e`dsMf5Ugqjk-PFZ zqDxs>Y+Iy6gHl;X63P^!$b{EA12xDD*Wnt8AKL5B8-#(BngK#k(&~bF7P2%~41!v) zefU>qmX6pnFg7{>*8PDzP=zP94;(nBBY?2*Gf~*^@Jh@LCkyL>%YGjV`z270Rv}`9<$Y^b5-woY^H}(Fj4U2 zUr#!`2Q6`?p;Y~iXmCSXjHNS%_7f@x4_|q8(5VpC< z%rT(ga%U;9y`cr?W*9c^(rkiFVso)SU6*k**V#r^CK z0o|=F{#F~V4zt_8S>N7e^jjoP8?ieBs;?^m3m4QowL2Nt1{Xd;Pw>4q6C8J>0&{8F zh$xKB1c0FjRwiQA8giOwQJFltN~}m&{OthMj2us%Sh;Nt74fgZ0;p-Y%~#B&vM8Zf zaTo6cUE9?aR{7plllAXO>4tLJF%3>7KxQ7vlz25Y*Q!tPIhelK0vqmct{j|Id3Kod;dmn*v+d(|mtJIcD8Q

)}Y*N!OUPj{}Ifh<;Z;2a`50cXh$q5i@odv$l<(al; zeSbVq3`^!}d0`0F@etIYu^A$Uq=uqvOEdsBc6=aIY1RvdZ#dLmstr24rxZVdN@j_8 z{2ruvQ$(=IA@9)QFdvODlTpO=LI&^ZDco3B&OF(W`8V>i&EEu!YH>kIaN&uqwl<4W z+2aUa;hHm}bA)a_sR9Qcyq-(&glSQnMbbtCzqPcgl&ojxzi=7He-$`=0Y2wy){ zkan!<)$ib85TcEb9!%6xL0370Q%;z)#4&HYL1zeupB9fH<10=KM#H<6!m9 z<`~)kf3UtIzdcM#^x5tHd^3)t3AJmYd_>RqC6dq^IfAs_EeJHbWSDoz z(}kAbf{;iEFUlg#0okub*e?k1E-`xF@JKuupPJvPdDv9Wui}jD8K;F6 z2e}V2_)-MX{I_lEuE+lU3)0_tQslbh&rrRUgd+h{88;P9aSD{)<{B)#pR&bxTs39e zvxOf2E?Htxt1NvsI@yqqwR#7Qc=wF;B>`}Dk_A>@Em$jo+c!#nq;mnNn@bDzry4)v zHiKf%8SN_|+M^)laUOdCK({&#Kwe>kJ})Ah?HEr}7*?!IAW|B~bdC^{HV>-<8b$!! zQ~fSaGX_XVo`amY?mxz4Y46ooLS&lO;P?77qTyK*_w$YMRCnlz|4PL_z+cs}8=)L`Oo? zN7vR#M8lSLT|+SfE$Yie;o<3n;2jU%siqOm*C8jrdlkn49edSIm#+#9ia&}+-On1K zZ66kdzbTPC{X5uU;jthCRfAQ1#$oIuFqm`XNOa@$=4++w4${F)0XA|Qw(;RJ_%(Fh zVSZ5>esfzs&dI^oj>)V{+q@|2ethv2`9OCT+Q|q)ifXab>ZOh0s5>J+N?}wMbK{ll z)->sAmkqrM)+mY78xCH~)cK}{4(4|X1@r?a76`L=XHI69DXVgtfu{vowh)62;cl1_ zwCU(?i2n!Ye3#)qM=d`QwnRo=Qil*@Yyep$XTa8+c;vNwI;fknYpq$ za}&f|T4He9l*F}11!lr+@9wM-SH#YRuj<_i&`&Q9_QHX!*rzbGwWmuCv($IUca*tO z$gIS)Gwv`23w}MnnJaLK(B^S~ePN2_U5r&Ni%WIqfiWG#?m9Y}hSg!g*6mEUEQ@mxR0&;$8&XsFVhHs zgt-e|EXkrR1MHgo%Qx*>Sg|W-U;(?_A8~Y500y$ZCumpDb08E2z=GNlekKs4F4H2j zWxYx~RD}Je`b?^2^_61z5Ai)#JxS)wDrT_ERZ&lKamSa z4YcqW73PqLN7ye$rW0l2P^_Q^mTt#=((Ckd7X?7QnLb|aJxh~UsY>;{(=KEJViZ1~1c=|O$gc8uR zn$R&c?AC#S1AXMWZ?sPVdcAH(uMW5#+GSXlY!;C22kHk6JT&l~*rKa+-V1?e$&g!y zXmMd)Q3@{d?4#or1Ay&J1;t7_I-PY?olt_U&pnzv-HXKNCE0s*7j?!FzrTz=fsXetX?t=z_rt=l`)+S^mQyVPs}w`CsD+ zx$3sj|0+)Z-@S^~U9WcH&P}SH*7^*mMKP0pbgPk~A*o~JxkkUS<-dm+d%Yf+Yy(Y* z6)*PmXu*QTOKZ;mb6S>m)yM+FYu(YHlh;TgQ_TT~8SiVIsi}YCPFd+$^;oL6T{EIr zDWfts?&Bat-z+M7^poC|kKaqruN`;AxO~d~&ey=#r;QkUPRYdbPk!rXK~kxi8d~GLFp=JdZNo-y6eFi`k!jV*8@# z;%KR8cGIil0eF;M#RgN^)yO;3ySMRkid74&Fe(UJk7*<&_1%yp!(!wIa%X}zH#{hd z{43!W;1XZ#PVz&HxjkMaIy@wKKD!7^!fl;6RaemN;KN4CAZB*Zt^gJ@yG#?9ga|Jk zb`&~&f*1oSZ5;nhI2b1t0J5KLv!seORY{p zz;0S1o@K}7%o6FprBk6}8@Z#&?37p|PwM0(dSp#~MHa67vj5K0=uEbO!bd4V-b<@7 z5J80+TZ|cj{&}l_zV4qkyr5@Q9NWo<_P%}!Zkb|~V0?22BLBiM4D@$3 zYj0JI&VG7blcDzbCCi3Y?7jtcGBc`qBHO(g$-5pK6QRJZ=#AxBg0o&#axc4e5h*K^ z`o0$-f|7C9?ZZnv$U4g3ez_KmOEm4mgDGTM_PVTRhnHuNPi1=k%$y`qt@5SKEz ziOY3EFzX`t7QDfcO2T02>&KTQE;EY$D|*WBd~Bb}7_{Ya9apzwe;QAxnx82kp3%TeH};P%d|L@-{2LdhOOx9Fff~= z=u_orBC(@j4;Ox4_&T-O$D$hBwYipRoaw>1+EnuffUDRU*zr=1%9b2z6NG&0&u}?s z`LRa(50{&y=X5;T7bgEc(s8tLd$~HEdL5h+OMePM(1T6yX1m4RwRn9uzU$vbQI?nK z*&Iw1&CzsrFp{^!=^c4wB)0zy;+A(}eE*~$c&oO39)yx3;yAji^gQs%Jp4*%CXtZf z-rp*HqkHDK{FjT(!u)@`*#Cy~|NfY(uK$0AwEtWgLm())k!oWN$1X8EuSLO|0>zAQ ze3!AJk-6>VdIhb%-9`FcjF{75wX?RAp=NCH!uQ)Z*X&)PxK8Pd6XQu)d zC7_DkUVKo(a7;L(#?`iX>4yB)hCVZ3**0>XhGgrYK;-ay6aOrq?|~jz`aW6oVv9Rl zwK%=egcYdD$q(O`?dz}W^az!yA`+HnK{RZr<75|yNo13(FEzJQl-hy2sB`Vbxz6RF zmyg}GZJ)NpEq|M-82gwGIxL`>RV6WZfWoUV_ zA+5FS77>0*!1W@QLV4XP@~wM^XCX}f;%tt6$4ac39(pAXlq+J0rA!!cVydrpRTNv9 zkSk|m2AB?`NxKnO1Q)ue?H$zAcV=|^&2#LzR7X+cZFA*;JtD+Cl=EqQ#r zXU)yb2B_g-Fn8s2dFlSXl+i2bO}~7YD#Aw}&bbKsvEpS{i(NA>htj)t^MVt5vleu* zbm@kZxWSqRE|+5=jtGP=-X>i$L?kUoK%@+!e{%RQV@ zsK+G0^W;S!Af)tn17|%^FvzkjtZ74tIEF&F&GZ;5_+74}lp ztR<-996|3iA$!C=OU(pF^Lu^QRo{+MeG#P5;mz~iz>9FTty#>lOpAi01z}WUGZ!YU zGmz%>2forFZb)_ffX@4)nA|Iv%%ikuv8@Aq^In}oACD@R1pw{-m~{^7n@jM%rcYFh zA%9{2E>3LYeR}yPo-C^p=)Yd|2ZU8gSlaM z-2Skmzkp$Yi0MGhBWeZ2KTVcN^3@uFOt7v?bDZWy6QlKbsWIlCPlYBa$=m1+G$$P` z-k{-*TGiLr^?4+odB-8|Tr+Ze+kQkN&ur^HdY*#W(M9dfPa0nP^U^`}`KGKLKXlV2 zyr5#M`E2*#RTnB371+D?bGDOT>)-KZ*mg!Fn&VxoRKg^r(ncJ(vx;u-{wUYp!~V> zz=}ze!O`b8GPO2wP>`MT{QZV{i+GshsN>z{85+#zda~@&CRB3$&a}MF>n8hlfqAU- z7G6R5Y0JoEKre`sKXG&Sw{08fF8V^rB1E*dA!S6PJKvO~uNk3vKsyh+XW}9AfQX=r zu+kvIJ};80hqYoCWu}7=z@Xns`L*!e-m{W7BZ?SVJZDCj6TZnaW2^w-q6((#3PXXE z*py@P5JucH=Kul(Mp*>1LFCaUcJ0OqSH7LF?FlD5XsqMktq&I!q0I#@d&Rz%;rgcY zs1j+v3Cey$&Mm^nOP({wcl@i?-44J6%Kca>uiyWv-nbjZ9oyE34@LEuDDd5*v zhrJcwCq9_Wf<{SYQc#_!bjVSbE;d8m)nULi`=*j7hIo~|Qgv@p6UV~R z(l6*lE|*Wui=lxc@gKeS2F2{*q@dl{&qZ3C@?p0ehph7NN6&#(@jtBcT=UNtVCc(A z8{X&@fofT2q{_3kmotCIe{D?O_Vn0CX}WGP1W;KTTx~q!c`>$N6XmVAzSim zhUD5|xDl0@IVLWr?&N3yU7uGwEB5L*s~d2G zDtRSfMs+Dfs+z6bIkkSDp*K?NAXdtmTRX|Xyf-FUCmz)sKTnrqRd>3!?VSn+F{Yyj zNSOuR@&t>q85ATEf?)st1914CG5-V%4k|ik!|a~Jy*zAW@&m;P0tEE%6~L4s)6&>% z$_vZ%Uo>2)qc6DW-}wpxz!|-0G61*+uVIXIuMe>5qjHST=`|7jM%w`CJ00#N}& zmsLFBIidQCMx@vwRdZ;asKBz00L4Iz^ZrSDxX1^}7F-I16I|Ej0P}P!vE>|fZc*@p z1nl5q!?Elwqo2{peP@&&S!$Ep>3$Jy2&&*MDio2kgfc}D#3H-`=o=ba2%5kR4aSsTa$EQ6?z&sz>VF05PI2pPN znO&x{&aN4(qJSn8gS=YG>2X@+C|R6yC4f_Du#~4fL;wFUc8)Qk1YMsW&mG&gZQHhO zn|Ey6wr$<9ZQHiFvzu(P$@?VxbkZOCLwBVreNO$W&Z+auU_$3nd(O1a>$Rid9$dlh z+TFK;HB^-uv*2q=0@dkF6z%75F~t#6&gGxLrqp|;nP6VcSfIOBj0T^W@I)$^-cyWd za)X0X))%1*kuy*kryGW}fFWjUL=r7dPqt%%3*$mW0c}f4MhgfQ)@tzQcAPLE4*4ys zD?NN}NKrLEp*^8&y4Imwsc@`Hsshu0*?L%jFKlwJ$f}018*-UYuPpvRT5lgMpvA5& zy}t4uc|N2| z=Fovrpb1Wr53ZLVnZ44y`z6(-$gVvGu>&k!v%C#DSn-NRXkX9%{cS{G-+ZNhL`rx( zrVvdbxW={McUHU=A;*Jm=xEjgpJ-zc$;t6;HKd-p9Y-N!dBA zf~F4=EHf~k@_AICP0t#z1+2RcRDlUZ=FjEc{50ebG&R~#gIa{6tt7%Od4s3|250ZFyKXc`>ycW+{D>J~;o zdOYyAGuiI2Q3jarsDvt01-wO5L4~~f+E@dHoF0PxP-=hr`Lx^iE zCZ4{9c4IcG1sBo-0(hu4l>Fc^(uW{S8Ae*W@ z%asvYBeW_PhWr5AOJjxz%#x)veBp)EspC|3yG7Lx^}y9eIot@{+LJI5za%oF_p|5w z5#BBYA|dA3YHcl+zU(d%%=Z(dOb;rL3T8an`2)v7uw^%#brYDkrHPXtey$L~_*d0oZ z8=Dv^Rd}p!oO)`TXR-(KH$_aiMROm~nys<(SULKlfJ}>?T0a?-T0_bD`Dkn2qBXLS z!df^jKbM)FZd3SHHVG5u+vE6Fk%PxaSwjCtgsAODRtHTe<6f-J=RIf4IANAx-yzMa zuc;>`uFhrG2G^tNUv%KgiZ%Xzu0EYaJ6Y|leCL<*V>2e!`dVWcU4&ZdSDgqg0RvNA zGn4`vos0yVpew>fJsh@9VS$hXFBV#6)XqoAjfk@y9GSbI0(7qTjav$!Uk zSty=f+Ni~~3tNa;nQ)@nzFqx+)@lQ;Opq3N_v+N*1T{w+4O4lby6ptDpj0W4WDlQ0 zRO&jzF9y|dYD%C)wp+V;5n`du5gM>4EP4CYbD;}iFSZW>##Uu5MqAC$^q4~zU&qts z)SFwM_#LB2)!kzf&grBko(@U1>o9SlP4sG|2J2r7N=l)L@z6$vB7NLAcRi~qy7HgB zmPHSY6asT*sq9Er^3gcAgZzI_Z>73C4E|6@M{qdaSoZ*gr?4v<6bGPyaDLCN(8!G~ zH+$i&#`%LyJSt(80k}wpsV9?7C-YmTw<%tZ#+qYCX|(ESIaF@YNmMfFKTT4p${n#X zRBE3M`f2x!Q5@#&`vq9}H}2`gx^z0%LQWP&u0Q*Pd&Y5|=o08+(J$=;x!Xh(wGnRn zY7pvL!DsSoG?AjgB03j)1$!nIn4;_i`4q}Gk8^zKIkCTKFpf3^B41HDm(q+?5lC%2 zUesGwc5gd|HrZax1qpit-RtkA^Mc~RMMXmy)e4IR<2}=lqMtKX2hql4t?z5V2Qp(D zw5cx)s$6DNg+jL@du@bWM-RK59_`n|D|!pOU)0CraQ0wA8{|i!+qgIDHzC$q!Q-K6jB3b*;js0*o5u<~3x= zJ8Lp#9Bb%vk9w_2SC>Y={aiC3HUs##HO}8^W<8{nF8slb^buy$_9S3B$mRDeOl_X( z+UgvA07xg6YD*e5n~jYJm=atR#SP~*_dJOv42-A^ULioJHOz(P%3wW6^*0_rMLn4^ z-X^B=GAQmGDy(l+Z6xADp)$|=w(Z=^kx8l)cv&i+tOvRJ2Hmf zF7WW`R1VjrRPBlDU$6;x42_uH(R*QPS>9GTQU?IU#0Nd?-ZqyW2MT$(pO)8JE?lZM zO6bNguim(@==IBERX2nBHxpjBn>E3?f*c#aId^Qi5NZ+&VE#DP^{y>O7-m7m&>d9( zxF_*S&A48S@^9upkev?+C^+3g2cewSGYFi_tB#Bg#ysIAWQ2tSzOm9A^!a!R#Z>?J z{>-tfNJoz0!|(DIiN+7FvvITMcwCC<-;B0>oVJa(BgrK)N7n4^nN+3JC<*bF7PvhF z*S=}rrEuk!UWtNjwwVt4PaqWC9a%WC#we2kT#KJZ{O{|;yrslZ#IT^PKyIHIC?fES(IvB=jDh$QTBGMjoDdJ_fId~p! z3%&Vbj#e*#-5-)R{T=4<8Dk(0X>d@gZO7F@`c-55r9PBwloO&0@Z1-M61v)9aDjby zVX+9EU1Ot*N@GM;6Bw&!+!zfeO|4SJqz%dkkB@=$uvTbqNZ?6@Xfio|#yq#SfoKc! zrj<`qoovcEtTSN&8Wc@Bfi@56n0W} z_6EbVKSNf@7cyVI#I7i}Bxq2(mYR4`KX99rh?JqcxLCR7TYZg^EA&o@!@qIT(V|E% z)jIwW}JFthW+NFSWU#Ueqyz5BG>&aJuqkguEEg zDYYUCL5o-t$AW%TdWfI`YL@ka;30B^#kqg{1@lwh-bqHG?Zo_KuctP2Xim=iH3amk zy8l#Q7%DJ(+weyqdm4Eu={nEBaXd{Rxj6l`YPl3v_Tk&b@ysfrL%YTr^!?bUO)|;> zbDF%A)mfL<)Th9EYCB0i>j^y4h8SOnemtBsi%@u9Z}Yk1^g*wn&ETbi73v*fn1;Ae z)`CpEhO6L_!U#Ozh~?XLg#i?^REIxWIwz&meqIts{o5qQ=vDTnGa2I{RDzQVX1_PV zNEr{C?WjNwiG^;L;+?VwgkW5Cm494|duBmpzaRWzQwoJ{t$WE>C8%5-C?$^?QYvIN zzN-Xmz%7kzUL#?>r_V25gVPK=O_}pfad0*-kyIK?ucxoR zFgPdYBUaPgt$2uhOu;FTQuwWBewf;l2ojKGdMNYK`dXs3;{Yj$g7Hy1P^7I`$A`}m zfpnQA(n;rmF}IGin9vRL9(!)Elb-%Xlt})Zq8wj`8ToAYC5S*|=nC#`KXm z57qjtv(Bt5M*7sL6UG(I$aA%(U$43InjCSPv%J5r?#(VPD3@)QW~ZNf;C(t3N`^Z6i2x;{*OJ z^<_gr`Pt9HKKWWs950~(+U*lL$ns=gW*h{A+wyKXLszC<2lA(QQbY74fL0o-U1ZM? zHDV~evpeCznJv5~pULl{#x+f$4ZdtTMK?yR1s*r{%R*`_+X214aO{K?RL;fpt zK4+p?v@$~-OEFZ3^V49TECBHzRG9mEOOd>)e5pX@U@q^aCoj(mE3Q4ts6T2nTu9rR zA@tw7G)3exBtX9;@PDvP`DId61_s^L*Mus{X@LfN?65avD*D!sOFyzMboHb_w89~l zGxv3H4+mXLq3t&c!oVw#%_WwNTn7`ud&^snym->G7N(IGij9E1fO4@60B4n@U`;Z^ z2Z4}3?0^F1a2fO$QqdWuQdJUeLPO|@okGMlGvH=W06ts6N>jM*9#Q@1{KtgB+zQmF z-?eM$XTSTsp6vj{$wwK$q#25J^FV}uD3x~q0y5+Lz!So>>LoFGP_Mzz2UxSc8V#ip zXF5C{c_w1bn@0bE3zC8dma&5jRdmd z2P+Q|P;lg)x~%q726chnDSu5-q6!@d^2qri4ARCYTyh=*NFSx`hiiNVBWS2Wq7apb zqERl-s?=ZUL^Ptm48uxP6X5PSgzwFi7`!I+m=?TO)mKLi=sPP zXF-!pX!{5RDb-JKpbVAlP&O=6@$$_^si}togwzB5>9zZF%19j*x-3%vM-A~%AV7w2 zxDG$0ecyH8Q%8?)n`T3GNBfn;K|G5xQLRhf2b(7`xVA#NXAkeWdE>Y#5!_G)ZS+Bl zqki@{)E{konb7`3ra4Gu3!+hqtlu3gk;@GY8?O?!? zOu!zHlFeuBO|lJmjPcd*dVOY@MN&&y6090N)se3iP>Up_ycss@Zi>oD^%1REdGt&c zsKV$e#p}=>aGS(4Uzki$ z{hcknaZ6c^Ec?s8w}w#)yUKsS7y{aIakO>bD$)_glw+?OWCQTp6$BidNx1&9u4!@_ zdiB|Hr@Y1{r7DW`6$GUjQYyM8$LC%0wZeQ*Et$el>(PXHHmTnXTAL?*8}d$zwZV&|1MB=}^=oX~z?HWka4T<$ZGAoZ`E#Ay z`Ab!E3&UMA=Lo+=A=WjC^vYf6EC;qz*LOj*#m=5h8vtI^HoLs)HYPOsj2hvGbt^$- zZkw3==D;jxS_pBFDLyQr=m)6c=PU|zi(y7&ky-?&o!<|o3q}+k5Ex8NA`+AyOTVHQ zWfY_KNfpBwk7R6nEN(_Eh>-5&7X>Oa9tDgEegL&D@8_iT4^&M3r7;2wKBqvkh-bfh zVfYtWeMi1Os&%N{22AevZ%7Y*WcPcDAK8W*Jbr!JD-d42(pSsE;-HNO2@`OF=}VYw zw>W=+CR@GPr49LREn8~|;ErmX`hAHCtB^o_U`796CsAcZtG4*r7L+4^#!h z?RxJ0;D#1^*K6lWbH?!|&mBy>a?_)+-WJ0nC!F`p>6IKn5f30Oa`%T$7ukXT_)(qI zHTnsYo0doYXZJ~Bb?VBhFHX!$y3FtFQw%jCC2Hw*(K-q>wpY@I{70A*)9R{gGW*XSf*c@kn;)1o|Abt11DlDD_hVjvBBzrQ+Dd+F z&SZ~)M>42}w)nPv!4SccoDm>*V2zL&1nIQ<&%1r*zkqwoszEk1nvzVU)$~XXc3109 zp$(ZVe6WR)a#vHGZeYJo50V}$`EN6?y9%ywIGArJsiP@S6KbH|C>0*liuaxmu0k5} zcOG-iyMAV~DdXk7w8to>PpH+@O}e8%j+hOawRH8~Wp`Tf5&MbA&jjHU&|KI4n*`!^ zSQ4U^SZbP>#sfGT@-sn*fX(uol3~o8*l#-ReA|PE=?V?NPFYgnfh+n?A>8$*3DK0BN{mey z`N}PUQSrvk8n!y%ps1>~yJz5)0Jup%z+9U*IhiTJKw=EHiAOkBB=9!SdUpjuBZ>4^ zo*Cp#qhSQ5L;XboD&XAsP~~^^YvjI+S@!q`gQApXlh{-d>$J<-(@7&TyT&bdu8mC( zDw?+XFpHmva8kNZPBKGT@uhK@(lnGl-`pQE2gseKg|*UF@Vg<_FwP) zX-GL83?p=3sNJI0Swa})08o~$j*&i7l#d<-u7HODh$IR4f$>vN{=9kd@VGk=SXL72 z&|jvXo#mX3tW@9J+;o3bO^OcYBa7Ny!)2=5?ym+ZRI$Fi9k1|@UCI8j2*Ar&Xs79@ zu|B_&9U^UWHu)MIh~nGgt%Q^-cfw`XBh2q8*!gz-L7GOyBIsX4moO8_%tp*tO^EH_ zOWATIgF$~<&28rRp*-)h=#eTgov!1mUvA-rTS7q2c>6mQf-SCH{ zB|K${8b-2?d-B%e)%KKkfR?g6NvIGG0=wtr(2%L>_!#-x!gB+9G$&%^oHNK6&Q!0Y zpUeC@yI$ETkcnx}WmPDfH-5NO&0(c%qPVJH{rUGXv-NU~fI4zU3}KVNsj3Ra`y6-m zQKHj1l8w4ty#s+`Y2(4e(-%RBn1Rl9nsD^ClCk@5V6_)nG3<+1284J)lT(+F#mn4@ zy~_}R62>M8x*sR+TDJfB^v>LPttPC=?p7wCWNIQqw~@}hZ(;cq-)iC#aIoUE|4LR2 zu|>t%M?`dH^5)7KvdCMNX$5?wX%~;aF6-nYgyN)SZJgbm(&^5OwAC$dajb%+IHo=+z?*HG4Cq~a3`-)wXfXFv&L=hCW*EXg@Vte#qTYE5dDWyB$G zrtd|OpMf2|2lSEiir#3fuD%^uo^na&lQibtK#L$V!MHNo;1fySE_Kd-Mn@DQ){(6q zgdW^Odggm1mXXwW#-GA-BlUkZl&A`&lo?%KDteb24GaEOOD(0IR1hQWg692ul^HRd zY781yUt{YJS_C7Iv$Hld;JWMVlCpG9X5KaF2kF|rr^PA}guVARG+|dX;(x(f_(}k- zDl@C&N@fENZD34N^bWHf#}u#ZQDMU&IE_em7U}jf;9d_8@K5QB1<%kJ;JBL?>_+*_cSn|)xwo8^o4QJiO)3Nc{BD&Hk8$PCOXq;OrC9 z5+_Wo{3Gs2K0tcZ5Dx3Zp*Tr!?jD8Nh+ORgPJzSSYO;fd99^XFt)9V0)Nt9DW5id_ z!J0b$i)4+E-*)|%+sXY%lauQk%`G~YPSo?^l69dWUzU|hYxVZAtKQNE3AY_KJS^G) z$x+a-Xu{Vq>K(5al8Jr%9M@L7f=MwBDvj-CtHvSzf$Ub#E6-Gf_tzt?wB zq6f29WGAqamvrj^6ey*`nQ(~AnEQT^P~WZ}ti&EKa0)}OyGpgqgR3xdLer-zsUxmn zBBSoy!=fPXWvDou>vR#O@5E|`=TN>OB+mwDP~mq zN9`x$T6piF(3?BO$Giw1cjibY#LtYj4#o5H=6&Xl4eKvW zu+Z=wtouTq;Db__Z}tCx8JixXw*i$E>O|3L3$D2`!R0}pO#t)~N4qVe38R$iLx@}s z;|XBMZD&_RXfU1E#$s1fHV@}huz^R&v!6|LP5>3gOeKVAt-;~>lREaiZR#In0zt}N zJFI#`j-r%G!iB=<4n^bT`h>d{g?xKaEI5>SE(go7S%H(XLM&D|#1AyAh>ljzNWquA zz@=PqS;BX@Z9tuD3us*B3_0*L1?&bc_SzkXs7Cv=&-!2ilgmOtM{A>y9#DrxHRPn3j2fsN%mR< zYg*cGgfDi$mG^}vy%3-ne~w9kPj48WGI9`NAc8-DEY+*36_3pi6h#LLZDhF*raY-& zg_V@awAg(LgzcML!EaitdD+CqAcYITG&@*K6*l0$Dn3*k;A!|!YtF2+{%n@a&&0x< zqv(&A{|Pt1fDChVq^Uyt?gK@PE;&wo22j#7Yh=;mcgYqtdNyJ1S$>)`C=fO0UB6Pf zpj=FpQH43Uwvts0WsL|`sVr7TEOEW```W)RfsFXMs+8g-kDp%A!pJCp_H{ptWT{~s zdWr$(8U$#GrAg&=*=8W1>-;hi-tD%pt26shQF1CpM7bU3N!=sY18RoMHFD2Fzga^A zb##P9gIA+v88%5{GfD>RZM2um20C4P8KM-r72wh(jp-p!A1l{+^K3#KyIPJf0E>~~ z69yGgU#4Yg!>B5XD6Lo!o*9{M@2+KNOZ+YHYgk&9E9Xe2H4T)EIKNl70-PL+Uk{7j z;!<_Ca@zkmV6vj_9SC*6t$zCY6H_s!>$GP+?pfPLUz~hLdw;xtF})$9zY-y_S|(wf z`;{+zBe<0GZgliC&pRDZce-R?Xl-O*>$*08TXUv(VK{i+gJHST!qRx)t+o9cV0W51 z%V&Om!Cza1n-}6*&5Wdw8jjjofvRUo*DjF4m%Qb!J&W!|XEQTS;=^p#Yk% zvgAV0?7_;WF)f3)u&ZX5?CR1SoO#$Ajr%-SV}|JGNB!GjT-UvqCMo12N%$5Dm)2zI z#@@G^thbaFnFnPv`(6gYjKavt_=oGoot)~n9Nod!YPq4;JqX|dG*s$_%d2usKyvYO zRE7(5q+$4HMfhMD!_2WDH<3k!M8_MQYQ7#=XVKz76iw@!TXu-;5{NiK z5>4$Jc0N1Z-R&rXmaC$M8!k3)dz&bwnS3`so=;y}l|NmGneg98Sg%;|C3WF}%t)DA zdrLRIA(=d%&kLl`1=L$!=ds+@NkAEkx9#g)zE5&997)hc5#>(3bCmM-llBpIAD&^+ z@B$^x%unJX_=Yei5y06U6{E&g{as*$2>z-NaJs}2c}GdUCkCdyl4%vx4a{+#g6?;v`5a#%I7e}5C{PaGM7I(_nWLH3Mq!L zYx2J(g{gLRt3&WJYUewpfl$VWR`8$Ccn}4EVy@`~yLbK)qxF}&EaRp3#e=Xw?n_4< zJ@bo{1>9mZwfFF$;26{;UIt$f(M$XM|bZy0rV#-Ba!n0(P+uAb<~t~xmQpQs`t zM&LfEk2OW4yUgFIFMc?lSiOlF)Ea*TNAKQ1=rq!2*C0HA3oc<#DlCPSL(n+UXQv?C zw`MBaMm`&7gzM%vBV4}eH+}6LFOzRtwzuOAZtYYPX_xbq!E-{F zbROC>PtH;sfcv+8>uCvd9VLY+n75&$04Xa+#{nN5hD5PV2XP zhKA88mf$k>tim!n<$wT{B5X%WQVoaXo^v$R-iiyaHSgmWI#KZEqi;9#(O@)i{`C+^ zPNh~bp!2uKVl{IUzG2xTZEc=YN^4NJTEW-*u=RQU6p+jN@GM^02`jh@GJA<-<<00h zSIZ8P;0boH1Luf|$APVfUeH7NDqL`fUEYu3$GwQnU*f-9jKQ`%+#{iQ4>QK#CJyZ< zmZ=W-N_~aTZ7EM%6h^Y z73M}tNLl`+T&*yNAY~l;bIKTzvoMyokQzD>6?8-P}Dd7_~F?`(XR{Oxk&J zzzLu?j)2icuAe}bX5`x$cl_92rtUom!P1CXeY&i{M^=14x_zD@jM!KW9N`cF9cn zBcaO%ZF%HPBcn_5Y;oijQqe0TE>uw~n}I`r%^8Ou>MY2b@lWv9>Z32H7GJq8tkLTc z>0baQ5BWzi<&b%z3w{`qjctE_ZEba~@KQ*i7(2}4WzUy$(~+h2$wa$wlEC@bP7G^k zne;f&r<+T==S{PtOFl)amlZnEtH5iRP%EEgg_lTEPTRZL^mb=YD(v~qnygUMZEv#V zy(-^NHdipi2DK2qkKBX6h0q1Lt0nx~&(5Q`Yv8YTCzH2ij0ip-2wy)fS z`1s%jSx`PHo#h_K65@zsjrzGPO%;5anBYNk%K}m8XIiS@Zwaq^ur={MAHTl5S&jzM zlX!u`KA`Y#SZT#fxA%paHpP3TNHL^%lRS9 zXXc@CQm~9DI>F6UAxSptw;JABaW^C&f)0*Z6UL!DH)cua=C~UpVGcyL%EUV-2`=P&^EVY+(C)?DL$asPj#NJrF7v=BH`FPh zUYOaK&Pb3|w^nQKFTH44k=gAC$nqT?kpl52$( zT0v-t$foT*UW;6V&F1A4Aw5n2TkNYoLpHIqpgt^6yprN&Rax1nzNW>8&w%=f47Q{Z z$yCwSP(Q}BUs|9@_}A&W(JwS1^|7gJgD#LOZa90}jUP3A%}dqr9pVl21~yF3_fKab znbfIpYe35v;3vASE-DXlihu?qGVZCDviqmFs5Rl`d;B}iZhm^o1aE|RxaH6OOP`dM zuJ0f7+&61gLaKf`Er#x621X$1XY{_k=x{2Nj#Ijy*2?NH8$LUw_Bi6*QJ;?V70=pP z=IkNNp50&V?xNw|Jb_O$$1HAGBB0475`D~n3K4g-Vr;v@4u&vD?Ha9!6Qtl3tTPp# zV5V^g;>mf;Wv4oRx}3c9)Zr398Y03if3Xaa@B6&qOx??`nXko&w>Y>=30&)*Qkv_?WfQ>Zl)tRMRq}moHWxhG zTHMFjBI$ZJ+3YK%Z>6mTQTSzT>6chP9%5h(YlA4eO9eS5pg>6EKJN0%~1K;&OR zn{PyEg-fc-t5(erX$_cAi<&k-B#tv0dD9k1rzp|jP@9;x{ohYpGYff}tA3NF<{{6) zu;z4hC!0bH)BxMC-?^X|m1joCQ34+b5kh1k2X7)pa|9Y0`s7 zzn)4EU9v?LdvQ)E&2tKC@b}jfM*nVgT1p-nhkhZejj4D0uG}VU>6&#C6QOTVC;hlPsHPq{Iv0nL>|8i9X0m4nFZUJG-UK3X#@w~435;Nw zl4uI-tSS%u(ekPVU&FoD6vj!I7(uG3N|#G``ya)sn7O6h^sLYIXu zYv&|(Mo9)qr*(igMX-;J_m}E`@wnx0V=T}looc^hW0gip)USoDd_n{LFHO+b`bv20 zwL8=6c|G&pLIsjVS3~;3#RXigFR;vR%$ zxU%4fiPdMl9+#^FzirGE;}9c$u07nFS-%>U)x`_SFk+ztZW$A5NW%25-k$DXyhHIH z%n(8Cct8mG$!digDjz5ex8!W2w^CW@;WzW>J4uO9frrcqGDavBC~CN%1|6~FggNn) zGPuqsHbr0g<&#lF6;uDJ_>_M6dBxw(E2eoS7qg#EE24jY$Ehv-EGb1w@xCa7Y^16t z>R=opa-OG7>K##31Kl-S1!kAhw@zJv8y=hd4r5Kf{ zXO^uyF`%!I-NpAV@}1xFQR1=w3}%sVpe+?gxA=Ax?qgYkr}BD*11aI z%Z?~Mj|~j=h7Za6aBP6_UyAYP10L-n;rxsx;K|>z3^Yl&rI>3I?D1tJIctw@n5q=t2HMQD%laYR@CYS5o}-(5d*fLu0{%o zC=E?3JXyW}3ScD51Oe*xs_6Lgm7}2!4WMP9up(lcj3geVndDc5%NirM_}kx1)B$b zl1sqL)~WUjWX&v8)dm&sp#r?Otc+99Tf;%ts`NiZ^md?29&Ei zBAlsjR$O=J(4HNEaNsIIC4ws_1Q8+L4xvQ!&&-6(8*TK?P8n&BYlHn8K!fHLQ=JSH z+jtc4{aC^}$jOsIJz63XeBi>FNy(Irjp9Kew(sGq-L+3JArNKXrSLik!ho%jGQqXy zgmVuvLi>n4LN$bQvI^?G7_R8K6-9mPrzMZzEqfyIB)eyPS&Z_ski`w}{+de0-dPg_G z4+3(fbEamXhL#5#nN9_0{(~%FdwrQ-PRlkEC)1ZGa{;FdRrpe(?mo!JI^4Mi)Hc6 zLzbOpHagwCxHEwBWAjW_`amrML(KvisA3&Wvjly}jb$yLQj#u55+v+Ssq_Q);@1Q0 z&OEv7PfA$ldpC?VU%XuAo!nvFi;B~10`f=zv-Z|@I#d&VXlDO=Ogkk@g*=vg(w~1j zNV*U_Ie9O2KNbPqsuYE!^IGY$KvAZ)8aQn~x}Lo9x{;uyjw8G&6AYGk$aI`(--dEW zIKY6}aoAZ?jC>awiO*;f>5;48Qw!byI~ z*U=4A^PNUdc%m)@t+A_+@2n+}AS4hiOdStmR6nb0#ze5mS3UlH$&6$U$oaB#Hm;6R z>J@#!q%OhE#i-B1QYMsi9u$nms2d;5Ms$DgcD?_?Q-@7;zczmH4tQk|B$9c`kF)Mr zG$ngz9CkxH*~Kq~nj}+$YOZF9vcK zwosj#m@KcwnC7<_m(d5807O1+I1vc;j13M!kmIMtLD>5S*X2f9x!d00v)}kL z^O0JYR^FkVFr;9wSdUNyecazuQo*3Ok~_c0pjUWI=Qx(ZuszzEVwYHCvZ9HrO8EK6 zA|q`ZF>;z_nXQUxTGh*DJNVU5h(7ND#^XY^dg7McV)0lB#Jcgo&%g6VC~3;o;FL~F zA7>-_n>CuU-b;n1OUPLQpWg7Ps&K7jOF;e*9h*F@Z%5RgGo-wvo)J>(W@vy1@=|<# zjY*1Fa7S_*3Vq%0o1#0^GH2dO>NYt^W%uj(<-T~&jc~V)xFhQ}hrLA`LGoatMU8C| z`vjBwr#W|uL50(tB{&2D4S9CLlFJoKto#JLoeurwd91O1jtXg!@)jeN$Y{3Ce|4HDm|4)I#!0?+(`CmpZ zs?%f|Gg*;CUcOPN4N6q!S{HTZV?{PZxbBP-P5q<2sxh#*{BEYcUU&r4_^F+2`XS** zJY%LmfbKndG(SJVj}TJA^rs7!SQCg75(<^)&6CA)_jeLVhJ!|Y@ghr&B}if;FZC&X zJ9tvQ4<|UN5F-xOmr-qfbiW`T_Wu|aagf<=GixeFp_BNm!wnNlrIa|Bur~9cUIUG! zcq=p|K;Nu{HpBDvU0scrDKe)u(ZmWjMw%nd;JKE%2}!(nt}8)p@Ng@zG~7pAORj> zV;#rj(-3Dgy6Q{tcgC8ChVu(jq1*Gu+O^!bUFme~HTx$T^Y7KW`L&@pP7qTdokM`# zBwm_iUr{)7V1MC+d3=x5IJ``CP7290EUd&{X3q=HTm2I&P1rnu*_@3LBIv{AgJYYG z7PBn6U9h~L9)n%VUN0{;oUopPitJ?xQ#>NRz>`t;@eTwszYcr2I$pEX zn6H_E0xtxt$QaXL$}Ay4X)wI||GvN*SUtvpw_pkDTB?$pomn~$GD;onopRfkP=@#Z zSf@-d_5`nuE#;23#X3Gf1&bccNSvNCMy@;%zipKC+or<1daIQC1cg&CSLx%P9Vci& zIN<|(F&wLgN-X-(9w>N};4U~t_855(fxr=uXnGh7R&gv(a!CMOZ<^e(z8mjZH~7wf(f!@z!+x}`De%Vpax_s)mD;})NP)tl7m+DTZ(``>c^wpDK& z?D=^p;!q7qpO8ggyd+Bn2iCmn4Fi?H*W~ckBzHFmrz|AQhg||{wPke*344chT1qN z^aoi=AhXBqOX2QOa%RqZRE+M*sgK`HQSNW3Sa$&wR3b~f-czUVhUvA0N$bsj<1Jou zll}*f#_<0*G-VHa6MQ;3LrW!R8%R1?diR!RX}5P=f>r-;Se2fFvQHB>p5rdBXmd^brI{@`ex;!I(DU=17Gh zL88bc#;;Pu1S+oqkRTN<=#X*wCgMsF&Th4+Q7K4B$8Dk>LH%Q1Y8VPj* zg{1`M=M?Z2#=%U>v!D(12(56c!(h*qqwtHCg32fU02`u0hyXG1qk@6@38z-=D!?lw zqE{#ZmSYM;PsjuE1UQ=nVHQ~dvY5(+8&ChC1vLQqqX0{Yu@_NRQAiT_Z7y)MAOF(3 zC=IOWFYYc)RF?R0w@O?a7-adZD&7#2GdAF_9Y>~_s5LvDpe5h#ejRUHodhme0OKDl zq7ADH-XUIC3^HNsPiJ&5D7YJk695D+)0&qo@_Vba=Bbo3!WZc`_X2Ioqo3pBf>H|j zBbYl$L6F3v3jLA{7@h~))gxk1n;8EWVdoet3=g#FW81cE+qP}nwr$(Cea>@i+qTVp zH=FO*Zn90&v`I6+CY^RNy>3@%E!0qR!D_BuCUjboXc%>SDG5i}LnZO)LIid4q(F&; zqNao^)Cc)nfDPG(IQ|K(>6l0tAyD_*r|13MYpth`! zEn+cF+e+^g;4D(hhvSUZ-}(_T1hMz~22k;kMqo?cWamL)6kR zCWZaSnXnCV6go^iNtv@kDkUt`m6-mQ3;h>z8cBl0N@9hM^mLG)aGnu4EH7EI96Ahg z#1kn=h7lcBW3+Hi`LeL2LKnY||A2=>Lyj~V+qOt#Ys$wXK;{Dz{n!^1%Ko=zaUn39 zmo&AQy*;?lujPz`jaCXB?K#6MYR;L@}WYJx1p6;JEu_0d|TWse02o>^iPLN zo`fX!xFh};?c{{Pyh5rXQ(lgAr$(X4A{0rJZ8nbFi>lD95s8<$5P!JC+=a-+HjT7W zw*Wt@m|MW!q~xMiq&IU+rAs(J^#`6tO-^rBMo&vm5~xO#)OLZ&vgbF~+E11Rp@?M@0Hhw#F zjMV*%q7A!Lk)E%{{HX`IO_THGz$MUe!4JK_6dz0v#rKOhz{!*Md<*_`st=@x#>;lW+5Z{5Ke~RK zh`5e6UbG(W`95;rRR2gqixu+a^kMto$R{#dZC1!v0l$`@d|zM9*)o;%HBycoXw6M@ z#ptcDVeNdYFaEp_W~Ix|xEPl+k@?0~z$)|kQJD0ug293;TqdgNLQf-sST;(DN`EvL zsRL7cv<*i@zKjU=!@od1_&2Cc6^#@oD~r`wuvk=Eji@!otr+k^4ulPV=Zk@MZnG5}`7McA%*gsb%m23j&y;|Xjg9gDz{&pqDbd&$Pz4;VRZuA4!JuyGBe^5I z?VTMgi~j~f2?!AP{~FLBF3_MP#O-f~vpJc2-(S5~eOFcY+H33F?wLX&Wi?}D24_Y9 zDekpyrbedbJAe^X(PoVe02-Q`85$ZJ2o;qpwYv7b|1pG0mVr4sHMn=5zeWUSuwb1% zlVieFc41I3ufXaI+Q1~7!O5ZV$*Hj+U_)cW{qNZPU?w~PoKrg!Fa-m!3GVeE9fXPz zT%6vl8d_SNz31-}x`4G5xWUQEiP?8Lx4oG78FAqs9^ zTtI5(OaH;|&5t?(^_M?qu#7D3HICq`p4uxKfHSo=gMg$`-p0yODk4A=`_~kL$<^Wg zDZCT7=0=drF3hhg7X$%i5fDHx<2O4$y0SU7xH`HSI`vjR2hbnX)2bPyH!~wRH+MsC zHT01FOQbeuuU6+Z+&r?l>a0k5|z!~*-!qcL`89V0QqEct^vr{1l9iPf=}Y>V|qJ*1K#eO_CBBUlXm`ifm{Cc5V!{ZMwMJVvUL~&Ugrv~njM-l z`g|LI{qf%XJHP+3ANz~G`{6$SY9$tDR$ukKHv3_J{htZAskQ6divahz#C^&j< zf`0y3SAhO{yNcl2%>LBLkD$D^KoPy6xBl7|CPfz~k!{LsPOPjx-a7u{pubx$*wkJR zPQuZt`RmdGkp1U$f5qNgbjsQ7SN~3I$^J@#`f2C>%2FEGT^YWtr`9<-0SL^^6VQRC zwkR4Moq&8a_Er~Qu0QH2!2mLEaCP={0PK&y0Z!p&A^dYG(8&QP9@Ee5hd}|Neg(Gy z3fA$5fdQia&>for0+Rd@r~xQmz#D-BnEdCpWj^@NY0HN2pV5>b`XkP0%O3HE!2qg$ z1??GEehJ_J6n@~njtU;Ydry|W=?{IUz(0d`A2YuNdjJY9{1MNxXO{DxGX$JkT_a47_q5NTXrP%&}_oiq50RNPF?{I=?8d=s?|n!DMII zol{#k?_&oyB)=m5-NpfRX|wzOF0Zv5>hR(I96gDyU*JAh$zS}l7c<}Z3mdzKo>IozufD8Hbs@$PuR+9{yVAm)}Q#_^<(fbuOQh(H8flqkKtNt;cmMtQ3zfPH?n>i z#hXh~rN)n%j<-(U0)f(H)e|tgE`5lScvs@udx-_TXnLf+ubfuaA<-7{N*5l_zD!ao zxf@pk8hN9bc~m?|mKa2S^1~%`J$GN%$JLl7Pj#?O;%QWC6S3f;ocK?(#tM_oh7~@x zL_Uu4ru|{ka1WI}X&1#}HGT>cJgRw;t3Y_1HPpzAJiMcsF5GnEUwy{S)EKo@QX%J^Z#=!Ni z-p3r_DzoNa;1IFc9C_7Dqb0K4Po$ANXQ@H7eK8y|m_}M-^80&4tof&vJuF!VI2l>7HpQ_+kZF!~i{~t|ZZf&W9 zAtW_iAHk&Lyol_S*~GI?UIiGvCR#?!R}~KY_A($X$CHx3@V3vJHaqd&%jk5m=G)}+ zK6h*zvw&u6`PJ^86&KPdBEu9$N`a0WgwK23W+@?5FpyH^hZV369abHy{a3)$kT^!j8bteb?;+VffGi`(xWG*gBzV5j7jdE@bBkld5wIylj z`E{1~nkiMz%X;2lo7k|P=e~a67REcFz33fHZl8+>g!buaa zJ?2zso_YFEuWr>^JFE)ONtx`9Ag0a^tTDHqvQF4SL|15DqamhGw}VKequ+P)Oy7G< zg1z&d=JS2-^icwHc77^)ej-;;D!BglzIv}8cg$4frAVmnC>*BrN2<9?9_zVuu7?if z15(4J>V;H<$MqPqKLwoW_$v64p_b3e(&CH6jPH4W(nkIhx%3Fr3~7Uf^Wl@!Ubu#8 zAGrNCRA2MdD+uskB0C3+JhB}9&6PKhMmJfR-RDg1ExgTnmC|R`(rK&O zi4sn>Rf623cJ|Zv%8p(1ELv(DmPR8o^~7y(#J3>k@-|K=Avi-L|C+gX*ggh-3v*1m zwbWtOZ8Lo1THI0m{W=Vx)-PKSXISga03Sl*e~<)_(mPhr_HBySID)yi``nY%^HN2> zDaNDhHl`6hnPxvUZxw?r*1Ua}y0rNI|A4?-wiGzN?}{rwGk;^zmUN>@K@VSx_=*`O zr#)gpll@lT`=roQc2yvTX>ZDm$=S*Wult-yT#=qVg9=#pFijp~3~%pj=kF3gK{1QJ zD`BX`30Og_zvmd(zb^AV9h6F(G8$rqFv?Ge4f^Z1T*%jpwXb*N%^HL`S4mTT`mm6f zYB@4C5V9k_M;mIZp)K4p0i_MV(bW1gCFx>Lt;l44eP^_8e+HBb)+Mss*Qp3~9_dix zQ_>1f0oJ3P6bb!>Gm3CagOQSG{$VAwX@J_)f<)fP>I*wTW&J$%1-ap&Yk5M88o3WG zM$&s+{DR}ux|p#GhH0Z(@3HV0&j1Ys z#=}^G=UQ)H8R11t6AoZwQ*{@>=W=Qv7Tq2ssyoX)vD2`Eqw-Lo7bn9Ft@@5}Jk(nm zI0Q~0z+ZlQKdV^hJiQHBm_^bxF=Ati)_vmz#}&3H7nkt+G7Hh_E2St6)2Lhb5;sJ= z{>KVvgWuPNmW#*)zH@e@2k2PaCE;2oZ(Cn%ny6oKNglTsxp^XV;qYeb1;y=D3Rl16 zP)lEN1xg8tFw=g@Nx7y77q;`1EFQPtwc~pM`2BAM+=e&bJ6< zl8`12`JE1YZ`udPVzgVJ4@tvSMSw7(=Uy*da{=LKJ9&IW6l6j9Y*f0xOl1nFgd*bCr-I@S8!2&$h?Ml#lH+e~-_GD91Q^?Z6D?pLlqZjM=wsiq!wADlB75okYL?6sh&{VgSn$3&C$P&YXu#9l zlby|{T;~Y;xI1~zJJw{m9fn%q>Vj_EjUZI#lhZr;gowMa>YpDo zC&y+(Y0RIfE0O+@`^wEmm_+5Gy4&b4RI7m4|QwY5-z0^Nv<;avjR zAv`u_0#AiYmX#i^lk;&-LO?-2~O40F$baN|G>^sBit zjVDuC$J={l7oeP-hh@+1e)fXlE-4;N z(obblHx~NU!)^Wo7O3XfFH2MyT7mStae^NY>va%TsBjRu%*S-AC1Ej%}`bWf}7k`h6l$?2l0#lx2^`X=pD+VHDMP)j|~ zi2-K!;dZJaj$S^C?0zrmIN@m!g=hdUK5PKmIt@WQ$I`gtdTKwXR}Z$jH~693fytbV znyVUQF^eB893aLG30p$V0*QGk;1k5m*AsC?Vp63Nmji&4Am zi|uzE>$ZmnC3y+`hX%K~?0_=#J`&cCfZzTaE1-Lcy-lNTvy%NDwh~RB+2|O9UJkIM zT1ZfxT90OxDk2@Nl^)Fbfe4Zg&v(*Sn<#BAvr;%!bOx+NwMj0rD8<_Aur#@{$d>1h z5N9%@tCG=7HSHs}*AEikktaQVCT2=D%K!FcHOu47qhD+~*Y8@O|BWJ))F) zTsxzvFY+i^FjFqaTf7Y;4Y7~i-*ipb;oBRh?$LV78$dlHs73ZIn@Ru~T8;VHg+=+l zCDf5%s7-1{-;N^W(XK1FcoulVywDqFvK0Xlv$=nD1LXc#8NY|88qHqhV zfj;I+>l2SxHyLm@>F)ASb0E|FC6`%#XSEXiiTFvrWbNa|5(L3236WAcUQnST6(ewq z-N%x-PjaGz8qQEf+%SI#ox+V!o*pnAX*{QEA=pPr`cX4lC|8ym^$4p^3(pNd9dX zlD&$Ux31H3R=G&EO8zDw{c@*47EEno{x+69b+_zVx_;J?oa=UHb>G_ruFRRm2W!$t znKwVHSNuLKIj~Tr*JuA?e%=kdDY`p@wK$enL-=^36STbw6=F`mxaK!(_u^LOjM#V_ zc=a9dY)ys}qR;cq9Z0pSeASq>c#fsgc$x2dy(un)z?1|yZb{-K&rh@h@-Lkr!queY zW069RqhGj90a{jWjK-cW69eCLo<^%9gdEo|&%=%dim@Tpw3P_*V}8bp=_oc*3ep$4 zkQOsaWXQ$QW61k#O}LUFfp=zs71qo&XPLi|#Sk7NI>9B;n2|ve1ikLP*CTUcpW-Nm zA*ZtCTlm6(^*z)Q50ta^+F&pQnrf%XLL`mJ2BU>ETsh^D(Tv&YVla?B0TaVhYxk>Zv8To-rfO(%b%bXHB|(mq)V@gOao3bt~zUDk+D^455 ze{s&)nFe8>`>?Ow6oH8Ck}qmN#36|_Ta>Y_LuEcSzLmI%eZt>PZ$Lj^Urb(G#yM}Dx6)boRS!lM%5$A?$K7FJgLJ-aw(A+2W2sW#WEuAAY7Hz9*yjva`E3|b3_GwEqP<~vawKg;+Oect^_NVoq_ z&3Wo`WOn0+OZbB-4oLG}nXX5&;7oKZZoyXkq2pWN$^foNUDU%p-N6F%O!tCz^lHbQG&|OJFY~c)j>t2i zia9Oih0%Kt>Waxpzc->Lwi&!1-iOFGNG5}fh!A@<-wi))`Ij8XmtZC+86Sq48s=EO z+I+rywpW40zh3*T`&1l0kh1JLYN70>fcxMR?vV|H1foqRU(6M8wTv54ca2q7X!pWk z0k}G_TY`&hksRWt^eM8D>NCEpZ(CBh!$rE9RJ-K6dhbwMd_vtP68avxSM1lixk`Co zI$o7$43}b4{uAz2B8C=QNaDQP^$lh_E$VgYj46T*!n7(W#5D^6b=J#*<@@LIJ<70y zG_)SGi9*pQL@XGzsH9Z*11c!cNh9%mym^{PhQ0OYPY8%u4egDRWyfqEf!Xa6oh_ET#00iFdudaHbx*rWDph z;zvRuvW$bO!CY0PcfP9 z)t3jzUk-;>!AYE`aF=OjKB1+WufE2+Y{ZDkA()MRVuC@FGX**8zC#J(HSTYqM>77? zY9lp|U)W+811uk-mtTNMwT6$20;?&Pa)KW8FhC3HEbgQDms#%XG)sQ;Mr}kgWoT^W zsg{47eppZD8Yp%(3IhFP#hL^qe5C%4rPPUCR`jT8K=S$TcLA%GA2{f^`4kUgP;kTD z2#)v)nSHGO(51eJSsQCOS@wI!Bne6uILWUG0?s7tS7Juz9SJqjW)u|i`uPG#NEBB% z$GjszUj3s~KL#DZot0@@OK?E80X9rCY>F(ZqxRjCMZa@2+jgh}GUU-lV>1 zQ<#nROmER1hzyn(pu#rs5l}9XVD>b~?W_wZLQ>td5Ww_4<{ zufBCLSP?~gVOwm~%Ju25T&}uBPUJL%$z6(ZF_|MT%vvO)Y{N9=Fn3cUr z^$>ha+8BEX5l)i01U&J0kz~m{v00Y+BrJxyW386b?}E#wc;TTp!L1BjF>uZ{wyKj1 zoq2|}I8wsFe3DQsBnS<2<-8Yt-XDFn5~0n7K-TAsk#D>tW&`BGT18Z8^n0hk0IZkf z2cBGs7M8lvG4_f~LUDgXR${T8I+>}l=7tOZak7f!k-M(%Ef zV`iM! z%tSd1!xGD{qndB`Y}}FYn;iST9rtItfy?WVJAR6h6qNJN^`QbH< z%2RZ7@K^nc7WYUA)5{Oq7<@>Sa^IW^^E5`;c1Z@`d%Ht9qF4qChr0R3e5TcjT$A4l zgZgRKi&rwUU3(NE)iHTdA#uWv<43|=a=*s?a*#OwX&uJ&(e2B!lyp>G^KKlVJhn_Dg(Un7H(c3MLJ^vI`i%D{gQYH5hm8=5e*o z%Dl;RMu-URy947?9&j|Lkmfz$mgHyH9rvG=yshlTFh|F%D5l z_`B&f1>nA_f2i$bwo=tcd368wt^ev{E;%E$D+!jFf_G6^mT=s->1jWny+D(I z0*!ge2n=pEZcNo^dx12=JT>hw#)Gkt&Ue&bw88B|)FL zFQIf-UI{xlh+*u_~@)K{Nw6U7T}3JW>(x7iF1q6wFdlp!zpc`nupi*!=rGgX#>7Tx<7Z1 zgGbx1$2a-76IVD;QV5N1dTR3gD|bHr7lIJqxQrXlS;S=+Vn%@P+Z!%s>gA1%F;0}q zD%Nq1_6>(5fM1!Vt6#t3pI^o^MEHcFTboW{Vl|X*4<}B=*yJA$H4Q;y%=TLaJsSo4 zP8*e+RVwh85@DvYo-JwIq0(MXJIoQAldtH?w+<(p2lO_rA$k8EJ@LHI4U4T?7do=p>DPba%)~HCmY7 z=pBD>&T34*Qd}#?Soax0Jt1{zc2wpG7wM*Ywu^7^9k8EI>!-Z)4~%EP_(I)Rb!>rK(Z#sY%5hKOO27U|Y)ijxJqV9}!D zufU?!B;pf-*ImA;Ys9}9f#;^=idxfsK)bw`jAU&{=6=K4!4kBdoY{SUuHf$G2Ixe_ z@`y?)-zVA;PU!T0K^h8fD>^!N^v?yr;uNmIp)=d39Cy0%aLI^KB`o+rw^d3rx675g zF@2P^B%k_yscDR)QdPhQR8#nzNtvr%MJJDuJB_(tgu%N0(c!>h!`AnZ+6OKJ8|ww< zn2C)nL^f5+i4tA0TR)~3s8JeGlg;1xMg|JXU`m$eKO#SmQZs8&&UrQIQZP|-yn!&Q z#8CcZfW8Op`C{>gSsk%5Ufgy{%dNxJyVvV4T9gNcT8``EteGakn~S@f9-}H$AW7KE zlAeW`E6pXhUqC!cH5Pf13P6Q#sbsujd~noy&fZPXz2t?;VrvU7qVWV>hBvbYsy^*9 zeLT@X5GM}*XFHrS-lX4*!Y8v%XC;7afe9Z>Ja93^ah1Z1n+XbSZ!J}Wg>O<-e67}d z$U>BrnSDou!)0?nX|8vn>_90Qu^>oZ&GiJPy1H!iM_m;VVhvNzPYh50-o=UlqWpj* zO>o&4MH1gq+7i7aQ~X=@8#Wlv)%<-Jp}BWOeMJt%4r|2CJiPfun=qFzd;iK?(}lJP zKe~^+V~Yo3JkRM~l2Ukl<$8ePy*sU|-kgrAUcd%By6ALEKj@~@DmjFb)|h7UbpK*I z%yLD?3BLnnn~(d3;?nmlosQV;5!q&>n*%>6$VKIpZzy~)FGqSI`t8~<)6izqG~WO^ zm?vD%8!HTc`2_}KA_#GI4#y;j6GgE81)z@X zXFP=G>RI9Ze@J{)$;$8H%4oafwYg&@4kYg|a;MIWAECNJKzX74XVyuW1P5t{&_Zy+ z)CLuO4eqjjwVF~cQdPhP4feRQgk1XoGlaz$6O@T+<1mRryjqcCr=}>!ZEE-qF0p&! z!aoShlOA4VG7a8e=`$W^d|T{BBsb5~W*>rmPjdTd`zz?pZm!BvcjTAC@PeeF>lM*H z38IHkO(_1erAV^=p2LL6D-Z-O8C>)^LLt!HN*_>{8z%`eO!+Y?fL~u&f&H=eStZ_aiV{3*Zum&2W7sfi;#iOU4jnO_O(4#$8?nSV@n?0hRQU&Q7%cR>-rhFx~mt_Kl`!D;pcbb zx5cptN~6n;24t--_rUj4gl-VZ+!4w7dUETt#q%^Dbj!=V)PY?*wQ;T=s!NXTjx_WN zr4qpBd~b3SLs$m;r=EX;E*0G-rB)8n>jq@vZj(1isDUJ@44)py5{9dUb$ZZJO@>!2 zf$_$_;3V6Ve#oa^QcCWnQP}F%a@hJu)oo+3qwV zeCN>n4YIY6IC1F?rUh~o`NIYqScglXQjr|G9^#>zS)BqV%r3pGn^fmjkDRoprwcoaTMFPu^RpBf zMF^BZn;WdSMB|9ErZI7yt^~UAkFD%f?<>#BDaKD|unxm3+~b|8aQ;vNJkc{qt>#H8 zK$ndrg?gv@bTgTtXpipf;eux7T*G2jBw|~zE%&;uNdKiGIOCen3X~oahze%+?r1sT zJQ1p=rUWYqgP&a?vtcaw6S}|_;K}Fz!dV`ELYT(`PpZi zaF14gsh}L_A?7>ij+_dHO4IHvlp3Qw#(}J4g%d=n%vxW3MvqAR!<*z}d+EazT-m7k z2Z-mOmI_Sdj8agmW**0FFwxmguSVNpA#sOGU29s`6kBC#rlBfghfZHvX*LzC?JrFDB}4Q7gO*y6uf2I= znt|We1=<3JZhs zN+Q)Mv;9I{E%YjdP(fie6ubTvw9Utne8qG;1-~i7XPifa@6-HL1v;jb&l&-TznV00 z89w>Wwj9HA>YAoRKG*%g>(EXVK=dYWo1$z?Y|cxZ&vcH`1Y_@_^N#(>v78pCa|JP_ zKY>}-{x*n!aH6fKoJ)EOeBROp^XJ(puwN$^Xn@E^b3Q$p)3i7Ym6@!YO&&m4{U*|^ zmI3Q_@}2Q}>eZR-pU^*Zjgt_H&IQ{OLpmf9Ag}IDz11L+^d1lmvi5Pjbw>%rE>+$G zBOyMB}44m_)LF*_HuHBv zR)1O!$QzJ5vpHyWGBqG_Th&TWS1x~O#aWL0Kc~F6lBlV;Kv8{$h|HG)DIh)EP+x1> zKBC!obt9KrUP-NA5ON~RCT`YI1^o#3Uiaa0`R>GgV=@QY5}a>`s8aN%_oE@0eKpz4 zMV_{!qbb20VLb^>{4IZ)47i$YI&@pKY+{Cxu*BWtxaDwG9$yJ|qjY4#AS@J@DJLaQ z$!j{Sg}yzAR@o9ctLq)k*=+S{@!xlZ3o!7K%h)KgVe~8iCKk7Lpax2J={};$^h5Iv zLwo>?FMdJ{T-qnW>I*$DA9KJK1hT4!2NT$GIgfvZ8!M7*N@`8k-P^U(b-E*>CvM*w zk?e#@Slx|lRw8?K^J-izBZ|8&3ktCKa(LfGNEc4#96Ob&?WV|73+#c!J8=Mri&}LE`}O&bwq{*F2AxHsCLzCsLg%)s1f^ z1M#`$EInFn?^7$xFt9~V<2SZJtu9={g%kK{z2)JWyl>W;SwPQb;TvvzE?H+e9O{mOCUi-3H`uVAxZEdCpbD^{TE6S){CWl0K*iQg` zAK#V`PEA_E5kDw_W|g)?twocc?-CSryt&)J7DPc$1hNylK7?I;E$AM?O=CWn*VsH^ z`l_aJ=RlVJs%oelrULHB%I?am%}@A*wM#Rx`BxUOt{PN)s((-wKvZihN~Q;K+SnOr zdn*SEZj~8*g`5|PN5vY$q{D<b0X)A+u^p>iz)oe5fWUG~bbnA1_CGYa!*Pn5@i;q+oLdn2W({oq> zQ``AAM@WUJje?1btLU*e)!quKeeMjh_=28ppH-Nsjwfdm=^7jX5fxf+(_|qvoXc99 z0~_}+C9!7-9igbV_Y*>n{x(CU;nr2HV6%*P_n0*nETN|#!BQlGH-+}7UEjF8N;#*+ z0C@-t`w3=0!1bLd>*%6le6p|2G=wMSV+i-kbY?lc9lcJJ8znHc(LI0XghuBmr8bN{ z>VYf_iIbqYW*(UyDpYAk&47^hd?*Ren=2Sqo$<5LzQkYYcdWXH4`cHm60Yx*aUsu` zC(Y-5w{1%$s>2_O!MK!4p_= zEVb5P6%7YthQ_46HA9}p4RXQf#+bm+O~^?@7GGL8e2?Gfdv{7PgN9cQ0*220Ff{Ib zo}K{6`QEvN6O6yGGSb=y(Z~Hp+?-g{_1W%YWQhZN$#l|hHXpiBx^_I%&WFu`89hna zw4W>hLpZ>>Ry$vYtqtQbh}V!Oz_J;v9k0eKLuKC;w8De&Z=4OPJ&F+xtG7)3-#>a@ z2Ze#?5cCHdMc1#sd%No{4||a-Z52P0s-Vx_ggBd-ac>w%q(%A4!@}%`fnw=i5_9RE zbc1Nf$vZ^JgVW=VO(NbT&bRue1D7Dy-gy!!{z-+S)*c&>H|Llxq4<{=cuRnVt0hE7 z@K|jM%A5S~IQ||o=mukY1|}UoU&2nKK3fLvt6(ouccGbD9r)TMVimP2OuWt&8;{fk zV-wsr!dZ7~oG)C#svm<3bAsw_N~np)3361Yp{@!a?!)L$lY)J=qHoBwU>IlsSz9{; z0NMcJEa=qNMkYy2r@hqB@Q-U;DOG-nHXQ)LwkG#-gL%4nXXCtE=$EskAU>A%__dUL zrZ*EU&}U!d4`PacYZbzzKDybM1hYpGOc@H^nAF*Id?;`cD4Qq;^#3X(_bwp|OXY;= zJ`2iinxg`BR`;9g3h0Go`vLDuV{*o?A3B2QXVQ{=5rzGokwnnx&naG!hph|MyjPkp zzo1EO*~3WDujM;GUwM<}Pw_@7ot#7Zb>b0h{;#0ahd>~~uJ zzVyBFgKQ-3Z9a)%Ze|AZAZ|V30r_+(8UpcNmm@G%ycF+4HPV*!0=9To9fgxKF$_bnazkxQpvz z@NXpfn=*70Ua!7-VtZ=jV{mjnLL0C2($Ts45w7}^IWMfesetjN?(j6S}rSmYAVmQ5W@IWBTJE;x^aq4F;a!zFk z=uRi#O9xX#&aSAqlpYEDQ;7v>L$29HrD#!N1N&1>G8jZ{1C-OjT7Ex}K`H)zw|o{D zJx81c`z>-o>%nadI5aU#4Hnx#M~Ut%-V2NyFBPU>l}Sy8{JfVSZ&cGM9FnUIYw{nj zX_v~Q)t&_A?s>C@tlWF~6JY5#)QsguDr%iT_a>H4X- zi}CW>`1QjW>1%)ryC`I(~09v}|P#4K2skF)7S(yJkG=k^5NzZ&|mg^rQ% zmqI$fA>D@yr_)^Lf%0ZOnOYo>V20wRQyNy^1!BuN<#S}H5i-tK7Ca!CxRR2?*S*^t z))_}L*+|a9);^P4_U5d1$}!mNw;$|jp8m@BrrBHGwRzRfelL02+1eOxV10yOSkG5k z6%MN8=Wi#LlqcB(>$}JkGfjfWt!@#;zLK&J6Li*YW6$otMzA8!w)l*`(@0#L_v>Ds zm~8}dw+68v4LT~xFZ_;3?scb8A{XLQu)X=M_9&`k3390A4fq^!Fz!V~pvP|f0cy$W z!YDAo4S`ZJTRe~_V2wa5p-n}1M=W1P=fd)I3qH3dh!z>P)v~5c{hS>yo^Oe$Yep{> zsOJUdwz7K2dd#PLf|fap@9NGtQk6$$Nd$=4F`1JYw!;$IvqVw9AqeGj&gjj|f(1$+ ze@@%YAL!xLQ>Ub_VlIRFk8hB`T$Ld*Vs5RZrfcqlnQ} zi3#6!8r3GNe|wTQlO~NYM=wN;_ZCvb3tKUT2LMC9bZC46bT{wHD;( zOHAeXo?tzWc<(5~#c5}V^RWUt^u`RYLWmbAu_}|P{`c*B>F7XzzX@@*vw@uhlC?G= zKp;0CNbJ~{5h_wvsjo7^*Y`u7waPR^y^kxM zP!s19fXbQjHMlq_%Z zQVi9+!5NPA1#!uTl~$hk+fV1F*re2XT~N97E-!JKuBU5MZfFPbdT#LB)-8pqxC75X z#R$7Mh8U>7Yn+YCgr0pzS(nzMBG8JQ7@a4UWA0wKFLk}m9*TE7 zHYV;DilopIAiO-MMD1ar7zt5lj&@|g3{z64j9m13P7Bb5{Jcy>4KZpO^UmNVY57Qc z^U|9Ei&6K`QuNC z;j7a8L^@{;io$noysRX+PI>r#N20C_XS8YEt54K=88*~q$KrZA=26fMKQH@+$6=VU zkw=LesNnkhHd|pRnfFH^Xx8r7S5ZLT?L$~q$S-BN_UVH+iZw>8S&j!VW^FAqq=1{( z=xKMZFe@SIkr@S(yG;m=2FrKXd*vZsRCyM6efqPV*`FJJjIz*2V`YK;K=`e%k_Ely zN~?p`CtoxyG0`D0n;BZq#7C(;X^@%MIf^UH@H%40B_#b@!poW7rYOWhl7WmVJOR)* zdS_3y3jPpyeF;TATEeCY{_T38y^9D}KcsqjsnZ}~DD*~caSQX?@r!tdrdF|_$1m?_ zZ)D=H==@9&rqZM}0E_UHlpr_K+D=mrF1t5@U=twS17}5f%;UAFJNqq$1k}#ZAsb2+ zr2+3h0j;(vnuy~knqO;q_oD<6_jv(pB2gq;dZ1OgfffJ$ zqmdR}JKrMjz4^wGYF}1FA649&aw_U5l zGdn^oVYqo@A}Jfuoh@<3a1ywGcFf;9+*3Bf*qJeIQbS1>=N_PEWT!NRq6o9sxj3h`g)-{_x!~)IaPLsg0d;wfFnj2h zWHuP@rlL_$P>o)ydPv^0*AbqHUEzY%-Kz3~z7BPj-Zu}vz>$M~rSg`Y`c4uQ%C53^ z{x8DLDM%C;I>T$*wr$(CZQHiJd+*w|ZQHhOTfLo5A3E(!-;$ikdvfyU`wVL+NU<51 z7BCD6%1GllZE;5ha4@4cB}YjY{^d=HT~~`Ds-?KOeSdWO>7pP;X!i>jeA%^pII_~U z?KhY5@N#&{4fLk;b-HHw?@5^`OfZ6~+h+G#{vAZ!M?`Qx@X!Kx=IQBGFjX%DrcRO? z?QOv7e2b_nYrSNy@^p^hOTR0G4X;Gwqmpg(G*0P z4I{aLf=XcF@6u`B4X-9ev)6^vB_Xz(*Kc z`@bV_jris-GD{qDJzn@ILpkkg4X)daoG=D@$)GUG4i1Y*63^{M_V19scV2u0PopZt znnhrE4B);Btc2mYLe(kCMTiL#b`n6;W%^mRP#9EZ_rAM^UO@j?XDd_B4p|NXP zIV9dG{WVOsOVYWgT;;w#Y$%own5d+*TE^wOc6iz{Q`c2prp4o?(SFk&W*T~7%@V({ zGyB@|Jg+5Z=DP@3Lxx3V3$Tp@IXCH(MkupjMCOrci%MK!61eqqf|c0PPMPR#(?Imo zo=E<2KOb%HktZykoBij>a_G{AoASt6TQM8@!uf#cDL8LaU5 z&3h0jPs`u4efvbY9OEgiZjeeo25!d3T4BrYoPs1D>NN-9o`$y%6*kKNTZBXU zj8s||`%e37)h;@F3X$3=_J=+W9cKe-Wqqcx(VWLC3F33snlX_=@OoUjN7phoW}t^I z&!&=MIW_V-&Y5(a$n@b!c|Yhjtjo_Z7uoo~N(hcb2)O0IVRfob=82ue^xdZIa_;ZZ z&p|0jDe21xz-H=Lu4Rfeg}t1k4IZ0e(FHAOXIZ{mx^z{HT(`2@2G`+A3_o!qosw^f zh~2I2>joAc+8>jD>`#z>ma+mZqGEbf;{$WD8sX{Z**UdWM!$wmB*BN7X$@3uGpt0B zj@d1|ClIDobszX8!))O1uq_rB|D)g2`Dp4C`em0qz9-aoq#>1y)EJH#o!ivRoi^^) zDCaRzMi&rk=xo$!(}s{BDv+Mu6vh!O_Rn8N2oDIGyufT#3D3SGi1m; zwFla>cLR(U3Yk)e6QGI;;A4m`CsCWXdTvdTON3)b$f+c1B1n$7(sjXb4JQOwY@bQQ zVOPOdo$S-}>RG`v5%*KVj<|mfr9Eq+Zs?t=QGE4dz}>zV?)m=_3o)(6F2&l5 zVc45)wQFqLEz@$NEz68tZVUCcn=RFGwvYc(s@=P{S@`BAF($gKl~yaf=F5!_2}l#v z^H3NYXV{S&8r%y^3ycncAtoTp$)p&V05C99E-)}M3`Io~W^pNE{%Qsis}mdGp4UoOe@^0+0qrM<%83+&BaVrY81gdFHSQ^Uf@+4If+c z0<%bYCdLM4M`thmloTE^xVX0MXXm;)I%iK+IA=CjvTqcY0CZ+?B>vd3vGs+~Io*hRiw2fv1}4{cm`5hI7AEj-Tlje-gs=&X ztPMY@jZeKe7689|cmTw>hyIJ-+#j_fW-I>SmzWr98tWIDo#~pH05Q`vu>ej53odar zaV!A5)`Pt`&osC;e233j&r>wdPyDz)bb9avRCxe%pG$A{p>SX?W-W1WE^uZk{^do# z)YGSe(HhuL>l&Jvn%rD{U#huv>D`I_r{kUTzpqiWRkk`+ygzmI>CO4~`bsd_Hs`8U z(^pwBBq6`o&z3>&#>|XPAazU(40VmnKrvSUVl`kTPyV(AS7z2vuLrv)^0<0Au#}9y?#HRlBbQcFt$~++dpl;zVlB9NCyoQ153ZC-@B%Qg4pT-;Izo( z0V$FHyiQ|7Ge8FBrhp#b_GF&CjsEHn@fwiU#@GO-y-+_q<-aJ`U$=63Z?Aa@KtJrM zRnz9+*nkv2EM08C)WQ9y^drB!w!iLUzv`{O$;ZFqgTK2OBH1*wzb|=j@sq!k!m|@o z3)j1Rxx80LZ=CQGAXTz_+?gqfp>cu7`fkkN zip=W7!k*0Hz>EH_Grd<-{?pmT>CK5Lge9|bU&1s1UgP{v>B=AQ zz9)a*g>yMge-X$7^lo~G`T_D%e}!{DDZdEK0J<0OKepuG;eBsPKf?K(t-ew@_}y1O zg>ySEfAi(#gMX1OPcs5zk58MwMh6BSQimrl1W7~Py7s5_B!^*aV#dcAnV^^LDs)=fzFR(Pi|<(4>12?`!_JZX!~#M zFS8+-+)FMI&aWSapWdne+wCm}F|j%^aN02_Ht-F=V@5#BC>5X%`bIY-8QNRm<(xJ} zvo4~|DKIEMVNNOb$yOa{U`_jXT`BEfVzSw7X3!YL1 zk6d}2qU^MnVlHmhUS@6V3_as*%2tU1EIqT5UJ&N^fu=a(u0ty?gP zPcl;3wN6m&v6ni@t^yH9^)?!qJF*R zg5$8PYK)Qo#g?5Nox8Kwf98jtxlbOoj@09;+~8G!b&~8py97|`h91ggv^_g?7r=P3 zztN7)R4=%!82n;F?2VUT!?(WX$&{EG^ci#lN>xpai$+rPMbR5fOUwp4lmbcdh2rR` z_CC*hXgmkoZM%~5@pdi2KDQRJ6o#9J+w(+W4gK`}#@-lz15M>GAByd=ejhC575?2; zp68scA<@vmdQ_c!Zn*udNGBSj?4N{p5=6bb-tk~UZY?n`ggo+RgSKXkJsPKdS1GN*X;gH<+RlC)YNScWO2l=#+t_+`%AWbG`qiDMMtfaXOA z)W4K)6vrE{>fg}|l`iW!if083y%Sn5cqgyVmCVC+2Lc%yxoTY#l|rJ_%uu;PcY*}C zDS{!FzH~s#b=DKBC}lI^94%T)?63@4VmL{ENA+sTkFy-u1?F1SlMPRB4=T_yBrWL5 z^Mtx0=Ecv*_9aw>Nb=(yXorvBx-sS#+NrHjO2wRM*)aB#SqGayx5|4-=7Q9nySR~f zNT5A~%Uf#7*)JvDoyo87ojWhFwA+Hd7tx{97D%FfE@VqCr~$AbHvF9m1O``5bDR~4 z38)D+R@P)D2$PVn|NPBrb1t`VEH)u2<+vu?8< zl=TMc$!WqXHo#Z6UN11#KXHM#g>yj97S(5UC0n6lN8@UrqNnpIBmQvpBS#c<-OP@4 zFDwlo9GCZ}gI&4T2eZ9C%H3Ax_&wOiPfoC#Q<9(D8t9tQ_yT&lpG#d(#U9y`qJ3fhhIVUZ38{gI=~oV1duKnzq%9EKOnAs=HC~OMNOX zdZu}@O9gR3vE`MtizUqE!8dtQT#8u=3-V>?VHTy@E1SPlTfAZu0r%w4eC$-`&p883(M|?c{pH9vlqJuAYj58SML@g7=DdvAN(ei zQvZe6ai*d_Ya(#pG(uYa3YfPaq*6hbXx1DNSbSmQ7SKl3U`8VS&>uZnLeUy#TzixN zp^P}{lV}%y9Hwdk=4+Bam)3NWFr&Q6M&O~EonovJB>+;B_+%Yn#!;*vPyTj%#}S?n zEHlYvZM7Kmi+m8d)ykKAiUfWR2xTtCZ39Cbn5PC=wZxol4)OcyBd#r=Q$h#1#6^2B zBTZ@f68hUjYPkqym+`zg>H3MsCnt4*p*xKVx(@#*)Ka}H+!e68mKZJ!j4J1+U2^aj zQx}znS8oqb>$sbHrb#g(@&^6Qz z=6RTH#o@4tB2A}KQ}xVTP4k_?Y=oj?S!|05cfkP3KZJ?nL?M1eQBGw*y+q}?Hua+P zk1wE?SS+XdfC==0kI2jpJYuQAp?9< z-EugHabGykw4MA`rhg*p<&oE{QS@oQ_M-&m_$cr+`4A>Cw*^&PHO6|%GaPDfSU=fD z@2z_k%w3%j!N-VsV5hK-(*N#VaJIXvbD9WQ?=xoCNlwH?R)~Be%%fpJm=d?`2U*f< zi(t$hzCR978gAxIAsIT-y%WRjdGHsGNAxNN+J>!VvMJ4eLZ5ayiL(?mb>S3$%OuZ3 zmsmx50jXonR@4HP<))U2ZTNjc1`0Po;7KpS;;~ zrcX5jbi9FoaDj#ZSlS{AK^$WjWEYX#+WrMKZXIYqCNVNHRP-fX!2Ki5dv`b)8ug5I zi|VrZ?J>i1gCbbe;yrI!t7|8&t@UEJ+W#H*2w7Ga;>iYSpC;nTY<@b;)?UGdM{$-s z`dGpx=3Cw!g?WvPK+H|gzq+_!7jo<#4J)W0FG@8tc5X*2*mw~dMXi;SJAb|-lY>LU zNtluQi;wn-Ph$n<%1!Wm106qI4t{}l>Y-T|U6Ys#QVK3NzWNu&w+kFFdjURP2le{c2OF<43;ptrG#w`)s191^T#lLyk0;ofadR!9;AE_9 zjU-U-y})sn#9Q@9iJ<)gm1BXwc8x^$j}EX+*9vB9VX+E=*s&wP#~QcMJ8T1~*IF3K zzou|tm-z!tkqWWb{d4zobLe9L8hSJ{U7G0?;*5kz-pZ$(%o!LKG73a-r5rMeT zp4mrC)6vD0bG3at7OY&E%4_;w_wc(xD8xU%<`IcT($EiJ^IBE{MbrJKZVDJXP8F!C zJjNZ{*IXVEIf~5EHh!BvXBFlAJdXe~*m(lMaSyM^64SX*E~VzYjqqP3 zDKD9NE6`97Ymgh5&7Un}$-bSm^e_cCc%(o;U!IiN%RF$Tu z2Cm6PmF_c13ZWt=WwTPq7&us+04h@KK~WRs4)6eo=+AM?T{fDpW>$|`uF6h~n^xGcNAGJ?>@}7^Bb+KS_8jp(tjbvRB?s?N?gyguf2<=m43&J>t^Qt)F8pnHT(>$p- z)$bIe4ICcFqBE7ita}iAgz(kL z{^t7l(3y7)u9A21G#0?eCGCW{eV<(=8iX4Z$=iXIibEwH=SRUf4hYU_8L6MC^%)sz z?PU`^XSUPSW=Ql5CH6)$r?oweucP}i7OBbmeZ_cTP{KK%O++MG)5Yfxo?Q|525c#w z;bpm~WDSC4-cur5GUKhqK2KaOWV{ZKmfc8N&c6l?LB^tsQB)$gAD@2|rdjS)xA!El z=N+l&Y`yKfvqN?Z2f=1xK2^=QH>(+^kE^|WxkM3^*6)m<@|O;ck03pJiTj-_Pe=x9 zAX}&`tEw1`pe48Bs(<3FOIad+XrnNBU>p(zE0 zNMqs}rKjh%Y6g0%o#8RpDG(fQ;Z>WqkgaR^Y*&&oztM!9v(=IE-k8wsr=+8&UAaQt zqPt9gJDY1eQ9{K0AjohbcA2hNF)<^{>bXu@X-k%Wm?5lz zl0FdQgCeuiQ4Ls-D>_2;j?ME1!ngWczNAB0ojv_ucQSY}0010>OhM z32!3Zc@#q1t{g8U>0O28-#)4R6Nx98GU|$w!iUx4TZOZb;(CS_G9^zglVv}*{?UMK zH+A<8ivX%Lcd$A}P3zsk_7{tQV4+pQ;&%I?d z3)$;^=^sO!(_(u)Fe6#52 z{FDBY9ldzj?@B1=7^tx(fSBvdE!;EWc(5hqUv*u>K)iTCvrVGVPIMx65bd@Nn-6u@ zsBNGwpTa2)S<<+S0EGYMGFNGen z*YFL~PAmHga_lFHuV~~paE>#v%*D~UY1g7HRN2Hq9R9p-k?T~hM43wPuv2$qtKJX$<1O98~Ov<*Oq!gOsIE?;4-43uRwqRWm z8`O*P&)%bteHaL_d1lceP~S4SC_}~{jsB<%TOE+kAO~D zN7D?e5G9wz2kBqu9mH-BEXgpG<{$*VD_X6re-rp0FqeR0~3hAL%z;daH^!V-uxyB%gF z@zl5h+kj}lV^UUMswEv>)gF>#A!C8cyW-GW`s&+~93ENg1)&9**$W#AOIhFusWzb| zL`nNll9;L@s7=w*Ljucq7C}T;tWrBHc~{BFld02Hq~dloy1>m#s8VDN$B6Cc z0hyWr$a5-RKZp>?yxxzQm~{F>bn@G@PgGK?v7DuQm;u1bciV0r-}1)WDz{^>lFB`C zWnX8CN#be9y8akrJS3hXS>P@HFVBG~7d4@8zB9 zFk2+YrlG!Zx*$fyxS;e@vab%FzcR52pufA#pUx(Fk=u7EM27E}f-F=dHk2jr!}5Wv z8W)M3V$~-MoGNwMP^+0XSRgMXYm+(hH1q}{ayw6sj#NjY^Lcu@x&FwcCSHsMPnGtQ zNd?)oge)5{Xk}>>6=1TmoyzdP2ZdqKk7XL|#Jf7nI@PobrJ};SPE_q|J z_q5N49h7?55!Gt6cod3#WD6w0gpBq?j#SXjs_hfD4aO#+%Lm?&v2du>E37a35WA3T zy|4^q(^z_y`ABSi7r@t_?Wx1ph1H+!#?BsFD4<^SV40;JmYw0rO(m5Ip|TO8`cT0( zy%7XfGTcMT%5lkQ2bZkFrW8$5?@=}mX^`J?L&p7zRCTy_V>Y`(e^L*W$(RMSxV_Lq z($TE&72?7LwBW!3M^Lb&EqNb5Zbkw-KM|u+BcXi=oCXiTC{?A^gD6S{J7?wPFAVgG z2uIzmEqn3X$ddD8bt}6Fuw%(k1!-2XFH%7{xy1$sVa)0Y8+m6nJ zj~dMOT@RzS&c-yUYD+du5x^R=Kq{8q{{GV1q38aZPNJTirxT z?u{6xMuFW$lyLi);H{kMXV)lG^NC3AZjjPz-w&8ZS|uUqjvS48z68WlGy-HAb-+^y zki|20%1J_E4Vy%N2P)V_4mjq$PBWGy^-)`6z|Rt8jd$c4R07;R`6l^`52j8>v*~*{ z5Ll3|bD((#ffeymgQ~81BeQ&;p)ZOF01&W*uF11_!37*$F?1bnKJhlZI%B)V!xVl7 z{kTrbH%XqFAva-dfmZF-o`R3itnih47pqaDo(BPW5p6;JLmhN()HnSge z0l(4E8bi+;^riZUB&O)(bH* zBIlgUW{}V}Ve=Xx+~S-m_Q)0IM0}g0H!zq24~^3m1AaoKS>(98=f2glN%U*wEdea1 zKxzjK5=R(fwC0E7*x=Mm5tVHNi-AT zkW8`g9!=b`)3QHa{KyVYsFz)+`Q}bJAA`&u2z?p%tU_}~CBN=`cI%OzTyf6ZWMYh* zcbA5Nt5v5wTsFe`<5WrQv_!)at4-Mi9XS)C)lg#yz}g=7;~-C1#X-jrQusz&kKo z6ZcW3ySnFzT$xB5CZf0Lplcsw{#$hda6Q*)hp9L;yEiGiQ)TSjj2AlR>%aZjJ8~?x z1^%8nndE+9-k?8ug#7N|-dK0#Y`T|7Djsh5GYje7?*rj^&3;TOKg4?HK_#2ubP0!xB}C6sDY=i|qgyx7X7dDCb{4{bOtoid zD~=nLtUy|zk3Mw_y-e46%U3aAAsbm2($+8esMOzwP-F24=_fKZuY(AGU7Ei)=rcK) zG?@`0qJsG%AIEr^aYl@)g>Xo!LhrN9%-%;rmBLGCjvAlrk^#a3eC@7dWNmhQ| zuF>K}G(ohvit*yW6bbtR@B)8nYl{l!Cb$d|id&20@OJGPM`ERf z<;SO=hoLK)tD~2&luE}AzVZ`r$qYdC#?dUB`COd}LP8qKh9A3B!P0u_T|Y)#Q3qI2kQQYGZS<+gZN8CnEf@-x!) zMqm*`WD-xFXp1DBXhVXqw6!PdipGUnM|huZr16HeaK8{Kc=~Gn=ln@rV^W8?aq30r zUmy-t?IA^S-B8C@wH>rmQm5T6`wk;-kMtw=3MYyHe=QEU0z#rtA~1DDi%|CD=y*HW z2#-P`t4XpeC<2*;_$9is#$&QB(iZ3Bju+Nbm7)5Y)T|jsuy(_RyyP$;sz53`@Q;lP z4D8nhF|ka^y{VybG$695N9X4^uDYofRk3EN6cIv;qE5UWzo^iYos07kv|AeK-3~4M z-ppjdV)t7q{~f=FY?^AV;5JH(9KLxLtI z&EQesqtEJDrUV8g;+|U?0yFI55J_BRt?&}ppzU$Ir&%qhK`hX0VjJmJM9Bx@h1>pi zi);!*Q0(?5X|@XN0Tj3(PY^@IFt^;o`dV;q+rj4J^WK(4uzI*LQTq<(A)Wn}8e4;t z!u54lvW-H`8U|S`uW)`J;PG_EPFIV+AT=^DV#hPkDqDlBjOZML%E@rV1~7X|lCXUp zcw-1A$sPLBvkNFYWb<;7taNg`2|u`Y7`X_7dmHl%*l(t=4g-BbU^5%n zV1k8cf6m6{N(9j#3KPam41)ccE%sf~b5bQB%?oDQL|Wt~%9DQYWO>hJb@& z2ag^XN^ondO1}8cgDN73U~EEb7dHcD;e&g&A0a-F;YbA(4*a~MfR4o*VW!VDJp`$j z1^2f59Wm+B%K_Z?5kv6s)o!V$dd`zWB{)81G;JJe`mQ-pwglsGB!!LfKu+OpTdHhv z<(yGg{`e7#S7pk>$7vm#dPp>mRQFviBwB?iR_N`uSj_W(OeEs$dTe#(O8K zE6-{z@IMb$_?n#m2nq=Vk58MhJjMm=5n+ zX07ObPC&qDGUb_Z?5xqCY{T*2$#+;~9GY?`oZg`?Ox`;Cx)b6ynf*l<`#b|N@L^C1 z)@6d1*dsHci;D<0ZIzLwhQYZa3x>B@`e(ov*F%BbedE;1>RD_?0!7uKvb)6Ji91B$ zadNJ%+S+Scc^oAOA7Hc(V(S5~_1-kObuk8uoLHYOMc)0l2<0#lP0>PU$xj9_6 zUVG$P@>&aXSt<1UY^ls6(YGGz$Ta0Xvn-zxQ6n4bvp>S5&9j#lV+87iLLKX-+*F^W z7_zk~!8J?qf+^ewZ6fN_i1gh(67Hlz@Xv=h$PkU+<;{4``?aTL7h8;?mU0ruwD%-pZQ zNlAi}*p|}?BjvPI5p%Hs5x?=E12n`H6Sp|Kr_U6vp^*y&A@RU(kg`L7a;S%lwDgZU z5+T6Ug3hD%AlX?=Zb^Hwd}}+2AsSMjAg&Iob5^E=9&F?K{^Xm7!)?DDQmvb;aTB*) z;D&89i%{8aWKva8^3fr^eAMo*fk4d+)Q-c0$>C>6;Oe-6+5WzDAm;(9c5Z-4 zo#-%yArze!;7g+g?2Tzc!Dy!886|7H(*p|Oi|HT_x2om5Eumvaad}{#JONkR8J*uQ z6e(KVVI@Cs_c3x;Ou~t()5nKpO`&dSKVSuin% zTa^Jqw7a9xh8rA~KZu z=0@*##_lBwo6J$onRERk!S@g1MyUhYu|4l3yKwKHRSCGV65niKp?b4gi$-SOsiNDvKN}B$w;d~+O zh_?CGhdtr7K=PNLvGUViC!6Z1IV2gZ%^7Lh*_V1ZywGvF15LXdjqdE#ft;86Xg!Lx zm&7_WzQGU!mjRyb$0dIx_$@h3eUMON0l`Xb{QFA)%p_6RV+37AU^_g(s`+7`!#_`E z$TSpZ{f{J=)GFm^7Ay?Tig(cHndL3KebxMAbeNiL&E?I6Ndu%|eif%lN6zc=ONZuB zpMcO9hWyhEIs;|dq+zCj+C!DhAfUg>hGWgefNjZLKy}z-V zaZ1;5N}`F;(jQ)AMkTCA)Q)F~c)@ca&90?MVsCHX{9%;2-dZ3P?e{wMRs31R9A%=o zedir(K`ex8p){XEyKtgM*Ohil%_=TkaH-b^s) zV{WW4$A2;U_UaSmHUpg;-wi%!lVQ}4v$2}Bf?T8`ba06%NmHTc2K-Id>GL&uB`5rX`cP{`9?Ni%?S+c;!LA0G6L;J-Y~v=M>~GL>y;;t($G0hBlpfN|r&) z#GCSvnc1nma&{2bUO$!KX`gT~yHvKl!!w=-w#{_hfEU4<*XcEm`amaXTDDLvL!y-k zBS5MtSJA`d=f(vR)iAPfWa zJ6)Y%C3&`1k)Uf)%pP@xisq;CAc3=1-68(9qmb-TrX+M3qBL|eICVYGaozeLu)GP9 z-`ECQkBNw4d3=tmwHglDX z${qtWDcLA*tDR$U@}Sz{=R4B8tT0OLlhec`d{ofF&*~5Si%u&#{_Z0B-Dn=eVmRrM z`V7JLs-uaTVX+dR>e}eYE#?d>^4}3|e(ApEccZJ_w$KLMa2us1sonV$BM0?qrH7cZ zcbB97ryko*2Y!UFlDJN6?&MsJeYv7GqmJVh@TaM-IBx1i-)5YY%U^T40nt8$R65LH6@tK_Km!xt5`g_KOwr3&?;g!RyDdOU8 z@A*ecFNgda-`9(?#x9n1zh&nV$qM$aONL1kpajIgDK{>e*y8)}CERAvBP@gY!zUOj zf2uUa8$PDr?3el)=Qn-F()qXxK`E)tzJATdfX@rTc=HqQQiN;X1)N>JC7sKygydVd zO_?}r!$NcXzn1wx!6u1c1{mtQ>OPiaQ;8Q>yT<=OhzQ=U(Mt8g2uO;}Lcrgnex|pI zn@aN<9&UabD6jNE0A(Ba)IAE@P4$qjXiUJE4a(;~6^#k!qtvt_pJ8$wi+PCI!bOWL z=(dr2f42C1!{CZLNkr$cA_!QSJL{u}@1VTDi3llA`Mnk2xjz*w%5^-`Hq^={sidLiI zN@6{E5!%gX)=aRBCj@1NzAv&yQG&X%K84Q(M*_HBdEDU}%I5Yq9covWuuO*l!odD1 z0^FIPacSBONRGjD88}lhDl{ScEOm-@EPQ zvc!1K`WMP#gvAOJ5)6K&*!${FP_vqnQ74V=3d`-9#qp|1Jl&+Coig8z41pQD(A9qJ z#ukCB6)#+T9h}QnPsk6`%p3v=mXrNta1ar3id0l;NnbKDRIkQRr(A5ZZQm^1F% zINf>$Bp-*A>@-llxL?PZ2r%^SiiQ=iBBXUehQmkoy_7#Q+d5tuA{GI&vdkJ|AfY+0 zi9$gddrfsmjE^(DpWA$|fPjOESY5e|wl3{1z`o3S3L)579y-eYMGZMd5PFs>MRZ(0 ze1SgOeuE(N2?`=1mE}m?YFRZ|76Q+R12k6)<(#nn6Bw(M_R~f`Pj#q+@fzk#FWn>dB|vS)$T0UNGT(XEg_{O2fEw z=R^D5kNAka*_NVvi?+j!XzA@#+QV1@qK0619F~R*%tZdxICvuHz`?iEhVa5bn3BT&#*{cs_)%=~8Z4V;Q*3Z%aZtoPW2mYbx`?hYa9Ls=}(ZV#6x zTW-4~uAX3unPejdq-v)yN*W|!$CKyCC-ix}2%S$-jG57pJ~OsiGKwi~?2OOWAN|{_ zSfAQu{-a@}gzwVP28!g#G*G$%KLMQPW(Zurty6s66w#T1_l*6Bub%1h$gL-6MwD8+ z4VDnc#zZbJyn>38LjiGDllqgZE-{?#0}Gj1c>|b8d?4!V(gyjkuHUh7Y^34=f*w5J zc;F;iaXB2Ka^VuKsWX3cn;_XG(Q_ZjkP($>zS7|-OqvkirtTLfeF&XYdwL)#v2Vn7>=B!C?%YODp9b^=rkroks`2ICS~eX~m9ooC=`m%Iz`kp^g;>V!4_0{whu z{fwZ+9wcVAj>c3tGT)}EaLm8Ph7HCknWrs`%s74R>&L#@JNz*uMowK=5G7y*73%C zR=C`5Dwl0pqmw#G-|5s(Vu$Jl@@+oEovYxGCvyqGMtJ5uxndZ-c0_l;t7Z2;GyQ9#65m z-N<}%H&X~>aQr>>HhNCk zR+zuW-`}a{bzlh~8swZf)WnO{{^85)46^oZ_`3%{Nf+PH5toMc&n5bPysCK)YPXOK zRDu^z_MOfw5%Yg@CfXOerx~chAvT{B?0g^x@fe13iW^eEi0cJ`QX-PB%kKWfj~?B$ zxA;h^7Au3FE1JY!loZ>B>b%Tjv)64gsf?=6aP7$S=GCcl-HNXANP5?Q;bvSMrF(D= zfVDIDVW84*8GFoi8- zJgI`7v0HlsS^HQ>`~#9oW*cdn5E8ER*9v+~Up#T|t2;Tl5YlV5KQWB!R-}!kEwsta z8EQvMdFpRrr#~?jJS7o$S?aH}_o{UTe-0to#hOyb4-V-@q1}}Jfs-4)W z)g)y;MIc50XF)Jck>R&n-rv34X`)caGg9Nn@rU=*S9+9iv=6%J@4V?ZkS@9TYt=96 z(-I(SGw+vRlnF=g%n2$eCk9I44SMF6} zSaNXGiuzsctpMf9KSeQlv16#}bJoMIXuOwKU#G+@TDW!AWaz+C5V+=wFRolcOIR3Z z(F8oiCVKPopiS*FO@=mxS54tDnl}{hJJJl`OWvvJ$iiZKcwV3u!{QkwA2Hl|>t^wq~~tWxKJ?dL8B ztbQj#{JkS=t=0$j<;=|kBLnFsT<{-fu@<(*&cinnd>^E0;szk zZUXG2Ba6(ToOzJ%)Dm!%V|tnUJqpfy`%*RW3O>`@jbkAj8U`FhVjLH1N15cRJ-);d z35aX4DpnSbv{FKP-6y3nvv`Wvx&H(434(o1M71?@=?{|5}SPdr8?`yjin>e>&)2&0sI-x0-7I* z1y}PhI(M4Pa{F*EkS|=Dt(_lfXF{Fm1d_K&T_b3V#9>U4YMGW4V(C;A*`qq95a<>w?yHFVi{Y~Y)fyEjUE9)j4` zYRggcYoOz^JFIY2E}tQm#bB2^v!epN=1;2}$mL(`V)RPtoN&Pv64t+rV13u4C*U}w zvmkLdh?`&@EXGUT$ZrB#+oM~Tx<0V$I~KX@Y$t1H(39z%)}s>|t-bw1owbF0Is6}t zy;E#1T-2^zuiCb4+qT`V+O}=CPi@<_ZQEV7?e>53CttD;zGUxtHjZa9k~!uzu3K#H zixQSh79!*WM#P%a%K-kR07GLiLH0-RlrXt{qX6}1h4gU%T!Mi8FO(Vi`s%b*_?eG2;9+STR*JqHKR%R z;7yjyONkX~)bcxL@G76KiS_LoB_SUJoyy6oQxgWarC%|uQCy8fA{wbB2(1s2>|y1} zyN{mJF6I~P_7Y+VbJ!U>zuC2`n)M{wa_*tbYiONcQ8?GnaFsko)GH6B?@&I58oJQ* zAJR4Ut&Gd%K<1A<8=@Eko6ZgGEV#xAXv62?yS>U4U{R4Q~*lS;qAukf9?WGnX-n0lZpK zuWllV!=t&W{e^8H6gNby;c4C8iaBnfC;XFE_H=%Kuh0+cZBgRxL*Ip=0Of{1RczgQ zDH7XT=-(8CD8x-h*J!dV=4abl_gPviCb*K|*z}$@wVt!_2{KA8CvEFxa-?~PNe5qx zTisocLzkbSRzIBXGzgKxvUbJDBp5$+OtSqdeiV*wf6GZbUEf4GLXL^g)csSd%hm+& z{AAY!;7Z3)0`et%`2=SH!;LC~*k;F6wu_Tc08DsegC0nS%N9y?;Zl!v?DVfj zJa$o)u@_YVkEiM-a?(Ergq9sRbJ1qnI)zN#e|E@8f#0bHefylgvH$P^Oq7O9O3W zKF91tmQhNmfx>qWc*(B3!Cmw@r?U{9p)Fm-ZN|^25R#|J0zNKTOYSN5IVQi zcMVu#-zzez4@a@{TqsE?itvhwIMNi? zi;0I}TGS^)u@E$^l2I+($3Cw9i?E7pGW)jrrXdGkwt=MQ4!^r%mg(CB zS*tkC%2*IFg%ej(VrF7xcju;WsGQ-(n@TzvN}Nh}j}!>gD){eN^BgH@0X-IYQtqzV zaws9ir{v>5=M=Fajv(xr)EgnpaP;tqav_WI1n2DK1rs*fQ_Dg)oKUJudX=y?HK+cc z=e|R`%|%Sv9<@LQN;;JB>h0Rq`WpEZRnxKf9r&On|34K^l$k z{`hlaRigixQ4+~)Z_d_TQ?=l~oj?P9bSg>}caAnIA^$7`5E_RQM1q!A~I*PlNA?#e1#|lN8c#@?gmh#7n_uDpkX8Jn8|S4n7+R6^bB< zU%*x0b3)Dn3i%j1c)au5A7mGl=&LN%y2UJvEz5kSunp;6ohsQ@{mJ-ycg@YqL4;d6 zXMxPCuE8ijk5bhMUrv~t;W>^lsiUk+xqHQMGGU|4+UC*-X{xZXoV_?&`!x{-zlEfy8SyV0$8ts(%y;r8+m5DG zN8>Rb@M+(IRP>qOc#0r7#F zE4l|l!y}~L&gh0B2$CuvF&=o6zEDgl&(I@G9Uq%cY8e!gYX%kIDg@@*1b1X&cLAxe zUzh<^J97$4od&St^&( z)V#FJ3^e-@|2m8B%R>$=F0vuEJHI}E5v?Z4%|_W8G5!R7vwi>LG6D$9~k$-Ag1M&w${ia6IA*_GjagLyI8D0p#~+%#6-*!JC z*dhgjFkv7zBcMLt%=|uZ`Cz^QCw&v{ z^vD*R(QSZW?&BOuz)Tn(5Go4F!3O|7!4aZM$VN~x00Qy=$W1)g0@MlQY`})lAmr6Q z#Q@+UNT@sP>B+~(=P5YrfS;b#gUW-^j}qhnbQW}Az=2Rtx6ff3Kt7K3#leQJYZlN8 z3ihKRh+r0e2i6e?EE$e63;`QTqOx7kYQO>X;1ZA<#Vt4m4*LyL{euwzbN6}-6u{f> zpYqA`=}rXsq`*8r3UP7;667V4{{qAlfr6P|Qa&4eC+q+$sJOK!n1F&CHvPvlfKZ;$ znsi^`k}#k!1|BHH4p({;3uaWn(a((n{vkxG z&~6vU?cR_taHyxS?{C8up`uRL9;?lco(Eh$qK12Y)Yl$nG0f{8zc(WY;V=a?F(EZ9 zP!~wQK0ADOAHw{r8}Kg_h<{fi?w7Bg0X+~8DO~-3ZjqR;f#=X*-hQBvES~kfR{_vp z+XIk5Ae{&}5Cgc@a6!jkJ-l;>t#4D3(_4spko0?^<3C_`-#=fM6Cf3gI+)87pL=%# zBqwa@3k!;ir(anQeL;`Xh@mrrreB}C` zMnxLcoNyGf@iR3NKOG=)^;;CihW{lvm?`*xjtAptMXMrf` zQs5LGTkZ5Mp+)0F$ucto)+#9{tz9a&C4}ZX7!*b{&w_sxtilV-9+xGQBm$Kb zqWl=V!3quo`riIwlJ|u<(zhl>@}J~0EA*M`gyWxo&;Xsd@fmx}h8Vo2B3%8Wn?w{B zR5QC2&VG<3dhn1=5z(cGUMA6XiO)@Gdg==Zw^msR*f@aRv{5x|Ef3!nwyp7$*^_fQ zwE&dIyO=n`C{xeqXmZvVB>}IP@?B*5&{Ej#fmk#$que4*TiY9H4~y(J5HtU$c2LEo zo*u%=3%8xO#8Kib?AiQjmZxzMJt45@)l;DW+}f^z$R?Y%PqA__9DQm`N_LWLs~6_& znDxfjm1{Ssyxoh)6%(h}zq*+n6eJA*QG7Uv|D`||wB@N+miBM8e0iKkXp+elZmIiQlB+e2 zAKq*$w3~?>Ag^=i6cg7x=K~McDsK{z@De=#7v>~*x5L|U<^j>d4K+4B8MfzbHBW(i zE(RG-HHD|gp4|P{6S-v#Cuf%a>wW=}{Ij`Xosl~UthmArZ1$nVbSXL#8-7Chbo|ER| zk%$S;+&FFSQS(HIUq5^R?&9VWTw;eI9?J+)rQffgT#wEWlwG7+YUc^-*zpIjL+qKk z`N=g#{DS#3F}4QE7FVL1+#ytWiL=AwyOvMPZuwgXar*QppfNF0exTu6`3iuf3V+Gz zvW4Uwg6z3FcDWS0OXmHF%XB64Gr{yj`K4IoKCwEBO`?k++LU4KDcr+?|A#)7uj(IEU%V{uj>T>;-@^##8F zc24oPWiuz7A*|vAmo?_7qFKD5?ia4Bg%@z5d7uaduu5pmbQYp|1q-)7zI#G-UqG3$ zoeh&OE?m8p^Zg)vOSo|+UEErILMNqsom>Qs^W!Zb+t%Mn! z*0-%>l}~L!)}wg+Sjs&bLXyV)t&-+6O?|xaK|FmX-NL4Yf>i*p;W=B&B16`e$&C7P z0#A`84$9}Y{7T#X2EBr}nw2Fqwr+6m)hEhu2X#WzIEc7UKjx~efBe_dUF~<`0}kdo zhRoEq^D?sNa=BwtU47Jw>S(?VbnusQYTImB7d?)=7*KW@i4pHJLu7w|K4FS&whzRA zD1I8;<`IZ6k{ui)OXievip;R94{l5%wyk4%a@K(^aFDb*t0s0d@bKiu+&6o7;Z%90 zZ?hy*ashm+3Ye9Q+|Lh2PcYepH?CLNi6C8|j?ug3*CQ~S-5iv)Q{1leXGa0gc{8VH zX70DvnNrahclJ~XDiE|xZ;>|PM*FV#+E3ism|ak3q~$UbyJ;*_`cp7lr?i-T7E{}f zlm`|o;TB@tv$FK0C;pQQU^1-L>WF}nKl!`_mXLKGJ;S*Pu5TWXhb7>S-j5|thG(s- z_a8BZ`*vi`Q&LnXb*3(`<;`Z&mdBx<&K>&kBK~D?7JecfPOnFq5x#fV@x7-9pII~b z!-B`p;@^NZc&(TDug}u&!9;cJlaYChOvKvzdW;ak1Ax6c0Ge{E>`{56MbCLR%~shs z{%Hpv5fYKNVbYZOmW&}X5BsW0z*>1{4B=4%2CJFEL2o39{mUe$3A+Zw`1C`+tTFY{ zJJgtXhCxgNjr0~P5!hGewd9BUetT5(>3vBQO+9Xe?6degFWRz{?u4%3S|lQ8ZaK>L zsVCUFBK^z+^|5KNvhWz#IMy6B7eFr8DY*3*d2PDgC6jt=yF$?F@mk?)5RBK z{883?UVScD%DF+;BKMYa)xW`;N}-kPT!=+kG$-|7<)*lK6X%(jVx`Cp|FB*xxw`1; zq(4RAbaL3ySN%0z%87u%X>1IlB(5Y2et9BTQ&}u78OE|_M9b-Y%;hTM_o|MgV2$5- z3hXd1etJfkXiDLeC|UM)WqS^jiEr{m`CTsm)fwWK>v3?qzm*So$l= zxXya3?=32Rr<K?|+IDPbY>TQj1AWbLb$KuTR4=J{UMauGX4d8y z>Xq%0>L_^N4`TdAS!hb%;Be-Y(h!r-$vtvkqa&3rUq+>iHer)hLD z#}giRrvG_J&e0ia-MY4-0e&s=-FSoAOeTY=_bk%P9nO;d_6h>9VtXkM2GS>+_r?TM zau#cN`^qz;Wj3j0kEM)nYdOMtRTDZp3#i&Q#d9AzCy(KETX%biGoyY#3*R3X z@O88t24zO2@KI`w%(YORPHmPZaJ!F7Tuz6F`}7-IPzCf{fqc>jk??qubcM{QL?Bhmz?g;AqWuQ18->vG!FcoT9f)##VOa?Z{}WR3|Cfi zf3_~ATQ#t6?k3+eqg_;3++Y!aH$NxPIe@qPAv(=1B+3)H3WbwhALEG5AuWMHE%5cXAyjp786qach@9`+KG)F# zU*u~~?G1)8IQ#~G4*rjNFs2s$1RADHjVPX0XPV!9Ci-Xy?-JX)Y`QB$_M>T51O3g| z99}I!xAKVek_S45sPJ=Q4YK+X&2Gra@Tt+(p%fa=4uwRCtjN%|D9IwA%u!0#Zc9Sr zxk}{x91Vs-0^;*qV3d%CZOz8cw;|jUo1>vAYvgt+eRWoU;+%eOR$Xd7t&4S=)nxw$NhML%|-U~P8LL#Mk`1cM(j<2x&PP=(pPhp0i)=? zTd_(!>26(7i9@&><403{I}^=ZUYTQviA)s=Q>VD80t((aveJhvlSF&txkm&+F(a=i zB*eElTlr^Drd6VPtxDO)b()(Bi)+|u=+>zpoqGB;p=yxhuF-KBo7Ju^z`+8(KP z9B^oKHF>60%o*Yq!vMuwSEV~af4Sp}NAb^7Dx3K9(X*1;KqY(jS>DlZ+9W(U@@C~a z=)1?*SrbaNMn06UT&zv3KY7e-o%6~a?j_8b=BwfagnRws-rG?W|MBxfVW&)hsjbzMpV4Q}3}-$N+9 z>Z<(?xPHB0zdWWB5~rW~K?_MGFwrif2&#>9X2Fr6aw%DKgx`G)T7VeWje3!Rh<%CxPN*L$Y-a{^3I8>b?#3_FXL!B z)amqJq(Pq*R&795BPXkV;QtN!_LJDUx8irdg;e2?ZVlzrOmz(NvU?)ox1oNhIKF0< zn?*oGyp70LwPfMs>4{Pde@66Y(oTU?HiOZylpkwrt1a+hIWv{;l_GVG*)RE`qD6Y% z5F{Y0?B59Rc1yn=NMGuC;!vKv+a9mqj>nEASM?P;`lf0xU_~JRRNZ8sQ*bhVwz7Op zngLlCUDK~GZku%0Gtd{VZW0L+CcHtUvZ>S<3P_nM5!TX-ISriS$^0=_+J* zjtlO;`uJ%+{lo+~6`oYMLI9OCHgA^N^>P2;N{OwVLDdJmN#xH!O+uJOCyI@bOkEt| zrSvz_EkAr0e1I=5;R#&HP$PahzKfjRr`_FH9e`q^()}5@xf$z_FCz$Gv5K9M0R@q$ z!b9`WEBeze7HePJNybFn{bCl|UPkhQnQ>XU>?U1BWkusSSx;DvFY(a{+SuDmpfuU6 z*K4d`Eg9Z(Hqga<_C@4*rS$fA7WleCbf~6#(>u3}LyM1Yr#<`RX%=KJJ&5@773?`$ zFO2#h&Npq*O>Gpa^xdMw$!>dRnLP<*Z@brFey(JO6RozM?Xapwkl}9Z zb#SE?avvvD2em2nX}<4D?AiS`CYEV9^Ko%^(bIX}q-smMQ!=xPNv|!0ED-`xR{6}V zHkLbn7WG|R$OqW%%mH|3@{U@JY#Hameoe?HPiORxw_4!%FON~?VjPwsuys81$5gxqUp=sSFSF-D1z!VY2 zI=|+YXM!@;Q1p{eBWM;#7pO!WVf*|*;}Om@;a!hBLj%tWCHW|{hEkg2KjveV4AlK_ z#@u8<>Z&b{qKcXrSYKE?of@0DO<#P{&2>MTS3IM@=W zt>CdX;Sz4y4a@*%7Zsp+E*5*H$YBENzOaKZhoUOplwaOiY>NPQ_hBnc@I%VgzGcSi z@6GF-X#GeU%`!A9G70-{N3{-NcSBEAdBF6k)p>^!Yl!tF=TI9bF&G5^88eR1(m_$n z>7uvQ*vm_5HoKm!lFCMKu`#yKKsSNMDFY9+zge?hF7yiRPXq9x{m&0U-;}SqtY*+8 zuyEktZuuk>?EMpbAzx}=6FRNv*lw=>-|~ck>1Jin3~1p-XBHG;;s`*QD6m*{>T`zO zfc}26XTdLOqjk~)(YRQ4B90a!sSEhOL>H&(`+%EsbsNuqH(^}FJ2*i9`REV74CvG7 z*_(95>32Z)6Ag5c_ZRVmwrnjp3H&75x>?$Av#{P*1S0=Tq7L81)H0D)*P<-IIq^+m z4;Is){u?BqU8vq|5A)n_jx6!Nj5ii!4Wv?-uz$V2A7RO#b#8t1l+l|Qec+GP_Y|i2 zyjv5Nto)MBpNdK9fWSzzLb^>ew^rXFFW$03ByC)^&WrBnfQ&1V2Xv(1wyz!ME_?Uv zT^?5Z=pD;k_R~FEhQb_yYu1;#ZpHYd(hGUSB>>cdF-R z5!*M>X=X!;!LZ#FzTmQ0q_>6+%6)ADmE3Y3e-N2+KD#My^8^^Yq|MIK6vvl0^QriS zy4}CzX)4_=$wBstjvhX9cK4A(8J^F1Ib1UNTANB3P5D~pZbdG(Bh_^4PpGA+%#EN$ zq`Sr>BGXD?OldbcH`E@Uyq1X^v9x`T!2WJ~F@ZYRaKT`FOO%h7r3P)9A?rPg1&`3k z8y0=pKR*vIg|66YpaO0V|2F&SQ7f8~3h=V~x5Qnvxo{04Z9a8$?ptDc+?Ou5@({b~ z2$7`Gcvq&!q-OJ-2~Mw$InSs$+yJ-w4ot!ZL`h7?85>-&4-X9eg$<4srb z6@P+mk{&*8tpB23c^{aO6kUlc7hQ6StU9OeyY|Ig&|zh4{ks&uL=*X=IXRSrO9q#G zlL(3%i!IV1+4kmx*5xe|Pg?(@!SYNP`7^)(TNwG_mf!Ut8ChwArU0g|^r30llM7SEuC?ykz=pakGyg~mbf)ie*4)TPX<;?_FoPP~#-aD|9H&`W zh+;>8LSes-u=bx&Hc8F6jfD511n{95h=Lv!guQQq?}b{OBx5U}ZkCiw+-#Y~SRssw zG$@bB`N}Aw>vUXnfm@E^DCxkPDst~c^dV@HPN(=dU5pKjOLb6z8|lT#SDb~ddqmBl zL3vaycWF@=+Eu@d&^tjDL%wH#M*y=S4;$zp0)fD0=3H9!@hPIg(L4+SpmSYigO%> zJJfmqwbos6n5CMAi1oiHw*i6B{DEpBru!LlQs?4oRyi-GLqD}t= z-`U#m>80HOp;M6D=<<<#%{SgI=z)0Vy_^@=?`9P|2p z+@FHKMBe5=OhU7}>V!Ac#j(L6%f}a}n6Af&NjKu;fROeL&FM9?W_uEH**7%^o7f=! zDVaYuKD_B<@Fvo=i3lEV`8(>&c@o^Hy>=d53To`0*MOmdLPAQ_d>f*gqDQ8hj{dim zsn{9%e7$rA6L$oHUJZP_tvki=%dW+8fTcl{rtpc@I#tKW_p;=WFqC0rNFsztr9w}m z`j+TSnje+`O(=%TF|boaA(IqYKMCAvRD};|vJsJ=WeG8rm1kc2H zvU=p8RI)&_3wbL!057uGSQEi0Q$60X=;LXWDFwqvl>q5jNji7oj2)K6*Na7#+x*hs zvn;>1Tak+N8+IWH6X$!MTRnqG-RSQ*nYI52RW9lYhEPw~X#sDonnj1alU9{VvY;I} zQ+6UQi)yg%N2&!AdB&)lTBFpOQ4lqXFhG-};8hA-pozfu9xd+lI;O;!zGBxr+SM9m zmUn-hYD%2bT7l`s9#WPUtPM|Y-SA|BF zX;&l|ziJs1-oDnPl?}&!mXmsh?xBxi7d5MB#MD0{xGg{6g3pJW5Qz)vE2=3IkZI~h z)~Ac(cOPLU*jI@{a*pPn-ugIrvy2(q-zL?EEf>Slc+HteEEIG0CZ_Yji*>HE#`Jci zW9g}3^xaN|0i1pN5SLQR$?3}FTaTleekr~R9z-98eLEligd0x4yY^Rgpvp91hUphI zYmmBlr;`FNh1uyRVqrpBDzu67oyc|q#5f>&pHc^L?{pPihwJfj$m?DuD6@ccxA7c- zOyED{cPmqobbU0wq#0E#^21_Qr%1vB2g8h^CB!$@m@Oh+zWsUh53Z8EBs`tXh5FH7 zrN8Zblk11wHopeI8#o!WeyMH7Yi_IIroHG%G7;?Y$qvbJjmPoitv=L z3o`7obUm7Vsr@c2`*|KW{Q=x3{e>A&$=1X-fDx=o&is+B>3;7ZL*E?B#ye)0L0bnV zMO3sJzqmA-nbr{^-#wy0FkG6goJ>;c_zM4mcJ0FsJC!=N-~$uMkrM;L&YDgVXz;{D z>qhNEZoke(Zx>$6Tnq8JYsVOPaAA-in2T-3aYjsk3oVHT+NAEEf^LH6e<0^t)e+67(6Z4hH zXZIy7X8ewEIV_{_d_fOPer>P$vH$uGtD<`gNyMku9I;pS&eR_=mMX4(Y-paxKTT9P zUP3Qeb*vmqiNpktbM;}k`3l`?dJZ=Biue5MKJuqGrd>tOdXVD%LOOAO*?+7cT%ll3 zbef|eh%OQf9l9xX^Jbzxu%BXl3wKF96PIdl|LC@bhp$U&o%TI7%b*4K)i}?yIZ*Yz zbWC8nBRFR3zJxr|mLV{2tw`}l6Wl&?AE4e?o0?fxK1*<&3Se`(uhdH==l4`L-X|AS zwW!~uEe>DL)Ss#sjhcPY<I+y+Sh+|4x(AVCN30ll*7)l!|C|md{Yd|f!JZlt zBroE%C1A1ph1jw2-TjiEi#rL~9-P37X;0@`aFmC;H(-tiKVXqNPb~2(UJt*<$YrfgE z-ReN}TVG+ATC(ue>Dq@blhm~|Y#iMhujOmeu9mz-$saz>(rYpLc!>S`DlTj8w~KGm zH!P|B^rz4+JeCqP!e6VggKs^HCDK{Wp&G$=?&$Ov*f}A}_kZJW zxc)c(hJ};qe~24qLUwj$*8i#gPyU93jf46BnZI%Q!@7OF%0M(2u`7nvlcrQ8dfW4A zz=(;(60;B~uTNYkN-PZ-6p1YCq9mSNzTnLeiCrlEPCns1`@ZXZtNmK5bty1+<1O&{ z>fOE`BO98mhbWH-2Zl=qhK!y9UU`ufAR1U$csN*CXaq!0&q4_e@vZ~6-xxsXn-AH3 z_3NZ4CfrY1Ee{tn_;F6MFNpUS6Oj}iOll&i*i=j`(5SGG`PWI5&@|A>K%g8(@h@h1 zaDa#pkZvVpognG*3|=VRV;xDtS_T4SQeqQ@g@ZR2E%&T*f7{N}eJ1CIv0!sIMsNp|BPr>a#{&29YK;FVX zvgEM)AekK^zCUMjNVSyyjL0Cv)CP*?klTN74q)5^2?N&;1yyYM@(VD)e#W$J1p|4gO0q8eD>bMF<((38+=X1hXhFpooDMgaR3GNc$USkd6R? zgYW_>VrW27J#;%eAP8q_Oo+*|zjyS`Nm8IsNnl7od9@Zn+_z|8uZ4G078s^w4A*Txkn=O6+8SKB^xXjl6W_Tb2x%rGwe^Gm2j)%SIf8t_k%b8vn@ zQek1?BJ-hO1P3DW{uTWLO^;SS#=kbCCz;qBLe_Pd18|oScwnf|PT_k000V-00*A(V=evdXBCAK=WES42;FmNEVFeOqsT z8#Wy@z!Xmu+SaW%<4&C5W)4jE`eZQ5KM}qm~n!J zA_437!hyArr2A{)WF`T`>tp9>4{b3uBtZ`nU#Y)J;~N-}!%o*6sP%eZGCy)QkfcC| zu<6(moKs@ns$v{Dy(+y&49)Bg_950Y58!!+0&3O4=ej@Rq~BMgf2Wv-HhvIN>>vbn)A998R-f%pcKJyd@Q?2rhAO; z^^<0u@H?Uojt5s|69HcD!mApjL38Gz$IX}p`Xa@j_bggMEK~q6@}GQN z-i|1^Bm}tcddJyF&ANotR?e?xf9wdBYEA6MUnvSq**+oYqKi;b82u+dmDZG$A=cyx z@@{W27~7+(c+5r~PE0G?zEe2zhK*@8z$>s~g`=++QWv1BQotH7N@!)|{caCwljm2j zG_xpgBio{3pEv;%7`rW$EcQOs8TQkyojHWdeOf3BHOOeeiCk_Qy8M2h>La-)G=qmW zebc{*(sKYZmh-djK5f~-zbzk_W){>G+aAu1%&S8Gh8v7a564clvP8`+-EMtK`krS( z`v|plVx!9(r2AA}t9*3Fm%pW22I0fp52Ty@@m9j3)!1-y9}c%?iSxV%Q(E|g#}4MV zRjlv7qm+*0DmKoo9VsH8_I}P2I$h)4B7r{~f{dYBFhi_Y zX-l>y1%(^kTS;c9hF1%}ouHOVtC<5y{0SXBxPeS1_n=mT+K;Hmyuh^D!Nys`?gDtF z6DlvtR5Bf$%+(TQr2hNec|tOm9B~C3)bukj-sfVvuRj_>c-WvNTu4G3jIn+nt;_XZ zX)yrvHYw)zOxHR>;|)ldB4NA zx>JQQ&x{e3r!MNHN3FPo`Jn!JEa{-0l{=H!w4zhhTDg-vb0p-OdRq}ON2l)szwL48 z3JY_E5`hw6-;xbSxei;qec9E}o%hjpmp4GwGw}1yuX8T_oz!k_W|&~mv^ddu;vDKh>29`ZVp#1QZ0da4^LpC^J7@KF4IYSBwN+(x&IG2N4rJ*rzVGC1 zz%vRfBtf5>Drt`8mhZxZ!gBc}Zy#yjUd*NrI+S<$ zz+?vVMM8L%8DV@9dhDgb>ToPl{$jv+-%D_51$5)}roMA3cIQV-tjQI*o)rD^dNi7} z-M&Dpjn-&3KA0dr9#LfX+f4Owx%Nb|tc~=`x6n5k7;N5o_$XG)E7JzJKGt$%e z#@<$25)6ZCaWboCPdq(Sgt7A2(x2Mg^UBoW1VIp3hae#;4bGd}Z)~}uP7=KV(5akx^I?NLy^$WpyvdWBU7JH^VzDFjJkHFDYz}73_`VMA0#v>@p5R+S z)zRZA&pP4#F;CSK(y{ai-o=h{feCwpsPr|KVwA=iUzZ;{RsU|@Dijg$K=%h>K`9W9 zQXDDZCF`CS4`a2U$MpLR3BjIxv%=r%KdbmRdcqhTNYuHJDXWv7!?F$dVXm6U@LkE2s z^#L77#s7lfrvKT_I}rkaAZtU~rJmZOOu#c=oPsiH{|zxBh$W3<@QICkU{SP|nqNy! zJy>21)^{wUR_NT(3K}W~aXfN~NMePxhX3apx!M}@0Rx*yNwf6&F;W^f;~&^_vRls8 zqMucMl7Zddkd3b$eQ4)Phg|^EZItgx1ws}i?|Vmxwj|kfV>9?K1Cu_EBlSPvo~7(iCc}XxMXQ&#^H+v1Z7Tm7T^hBWr6n(KCFQ^w>wsz_G$1N-Soc z>BQ7ADBsN)zM8t-=fjd_!e(3K3lik;X=9EhF`*_nT(+L%)(PtEJJ&2u%PHzpm7@FD zt~^t@8&33LI@di)Wwpn!a(;neE%r03`gZxdy}AVoaaOp+OP|xr zD#d}WgGbnBhw=ErihY%q z_wtQfuc>dSJf2gvWI8bgEzAP|Z-P`3!za1>~OSsQHkA|vwg)!3kPq749br0<;))+iD;3u_XO)|Qhug#%Xr?3~O)fe^L%L=K1^(sS7wZA@I2ly7xiK4p6w2JgOn)(i7)!n@O_0V34&ytp;ssly+OJ%r-(5%6k zI-bsDm&w*$CUz@{svD<=+s;O9PAb1eu~}#}e372gni%>LA0@ZZ?4y)#aoKeC07LnyhJV*JFG5kd(Nw<(>s;eN!NH3-R z-IogTL}ZKBTFePgkjNJ0WnrC$EA{^RvL2~-bsb^DoGtVZaimk5ES87e)*8V?ZQ-3O zHD7YC(w}Z$f|6O9$_*Nhc|jKAE8in2DUD4~mfVkTT&E7)_p)90eLz(&5@mWWHRTF4 zX!4c05B#0{f8E-jDn?OP4ig%rA``lq`4}gV8^2M|%Pu|Y86~=JMQSbYCe{z)=$l`8 zNNRi$=$uI(<^G8|VL4^Ar8{Ie{68a=>}>E+%iG;@n<>NR$cileNVTnibb^8f> z>3;7wKo7}tt{!S6luHe*RqFT8QLQX-NfqesonOy0rOA3e5d^rW^lGB5;-r9`C^0ha z?ngx3(rI+e^=9HvU(o>Wu?#u^gI_bR;-&nbmAAh##I*&7Fy|RMd3!E3_13AyD#^BY z?H8Z0d)THejjVmJJ)E@_HBF<-q?;(=@$wFyPwowtEmvfiR(dEePXdOnoSzAVZCDy? z4v4AU)hefH6542v{;8n-T&^&3>_bHtwU|U1LM2 zL>I8fjHhqwb&Mfob}jt3Ldd|#Wvvm$qmD+9c6e7zN0j&=T?b>jH|zaHa8o_7I@KmR zRlVap*i2TpJMi+}a{2oAMz^~G7#>TxoJNbjQ+L3*#%|YkCsAwX@7Il8y@5^H~tvr*g*x763?d7N?^e&wr6ys1gv(h7UK z&CB7v6){gc%5>;%sTVO5yB~ zGOA*VTRX@?{Ak4|+W&xxU_G(7(jZN3#G|}xp=H_6Or4c63n>}QI!FzuXH@-KbC(c8 z6wK$W4mQu^7AUE?q*tN)ESoBjj z#^A4P4Ip5-8U+XttPBj5<5#$K+hTi4rPwir@1$EbEl11=Y}I&FW@;zsNx%!Q?WNjLxa<@JPt`tCe?pbh6D8%i!;qXfFKOU|M zaSwuwt-R9kZ*`^AL&QCLJP4uA$YaZB0V_9ALUYNuuSjR)E3`Veg)K`<*rj)J|B~tnytq2X|}k3dbaG*5}260Suc zmqt!SWCn&JZ1=Q32dU@rJXf!b;myR~<~ML5WLFBO!EbY%+I7rdyL!_WSdhB*ewT~X zuT1`|W}raQUvGwXJrjmkRAHQOhauSIp_UZ?k)*bH^r7OInhWHgZl71sDWRGEVb!2r z5z!O*<5Ir6{(LxPel7bp3(2&h#;>IFDBTcxpIQq^?+2qG+g-1;OD9Esi~iFxw8FL2 zKU2$WOeMR)Wjn1TIM(FhG;16hobfrC{7^Du8_;?g$G5_ctR$zr!!5-=rhS81n zNOPl6b{-wVV`ouE42wsnBs$`@=)-iWcC$m`3h3>+C*~)erck*t zM_WEJ-o6cScpULbjQt@E+RNiM?ugVBYY|g$u3<$fJ4?xz>39h9MOuJDDyfVKr)8611`w=K)XcE z)~PV%0pJ~T6P0=hhBS>_^NelPd?U#`kaRo!?`}P72N&HaBH%Mj4+fdPfw4@v6iaFN zThT$tMmF2R5k(1QiO;i(e&=Zz{>~7&*~!JUo8&xQr1XA{h&i8@5oVT$oFrhB3Q$5# zr==V*<(;h~m9-Yj`q$I=MPpy6WprTwGzLGUo}KAyqp+aqFRSZSe-cZu*TKo6U|!Gb zTaA@T`@yp;+7K=lE$Zt-Ilqr{8qLE%7sKUNUaj|T4Al>F%xl?IVHBjeq0cbQ^xs$Q z9o?Z2O<6uDwOu3E>zy2^kX$K@q!}trVI*1^jr&#e-7hG#*NDf?$`Dj@pYf5pQqlke*wnkIOUxndZU z<%?3;1JnaeUU`R0$9H(H^l+7~Y!W+A+F1dQ`EAPR=X{Z-FDgpr@XTs+-NWONjarlJ zYwR3L)I7krW>~5aFK9#)O<6KyZLzH7wJqWd_a)XrX2q*jwy{>$sq#gRG&Ls}jJZqo zE4y#>2Mkb}<;`**r>t&tRSvfnZ;H1f81fKOzvVo3>Nr8SQQ?LwiZ2=r63pY@rX=W^ zX9#KHlXkD!fgJtyjnpCZ2gNBhq{Lbb`uljoLN=G;53Sit_RaTP8s9T-{WuZr-*dJO zm{hz=)k*vUs+gS{pZc>#@j#^;Gz%tcmX3G(Pi6+Yw2-hr zMTBDQU&ZRqOmruSuAxf#TAV!%J&51TRXckX#Yy1?$jZ{AyS#rm<`jF)Di*f6Ai$WtjImu*H5k9EMTI}7w z!amkyUUDXSakFb}At{tmis$>=p|=MFaZ<}pFyb%5d1l=vx-H#A$;O>)sDuYR86(d; z!E(RNENApoUINmOrnbO;FI6y4`afr96SMEDiNN*>FOd76B_>R?bT2F;^O~wJckISe zh!@4wln=X$2~a{)vMhi@!%4P)U2Pm`f5?=MM*W2ndQqL2hL5dlS4{ES3Njs(+bOlH z*co@7Ssxo4eR=@;cp@0b|Kc}je^j=?3L#U3Dz^~6#)@Y?0UJx7ZF1FW&b%I}CU9TJ z$T>Im`rHiFB&D6&2S6IA1{{W*))x0ra#oYwC2XP$3nn!8gdp5*;w-i<`5mX%uKT`J`*GI>H=*xA*Su~;2 zV5j{Ry}RM0I&La!QAYcltD$Md4kJ}(lH$zH14Oauwe8%W+9b}%gEhDI;GfefPN=Hon6<9>|Gapsm+b;3a+S_i-STBsAu%6XM$A6|xfab|z~4|WLSXHgDvjNpFqnt)>(Z7^ox zO714TFB@JAw}O*#!=38za2}*Bgcm(OOGUZAyRz}DjMD+apBU*iN+H?#zV2K~qBOE& zAQNKF!R!9DeaOLxzl6#q$@yOt_h|@f$u;I{E8O9Vlod<9iBGMn24*#YEO)SAl1X3m z=?5o}ys$1PH?a;$%`2t{QAy&&+Q|b#_I~UQtjcHDv?-N=tNtfu;dY8&JA9dl@)Ob` z7{IAANy|kdc#>iS2)h!AuOC94!BzFox(Ba0Q*H7TU

C(GNE8%Hv^lo^zu`y}ECI1{sV}Zwzf6cu zVAz~NCI@0Ogxug>-2PQjaQ-21eFNi@1H+S3e+*2`PTmQGy~OqW+Y%}PlmY=rgw}SD z&O-T!jc#}LO)RbU5Fzh*c|cYu*g&ud2pHe`H$X_*`9*+;1U!NMv2_p&*hmW_6HtXl z_IfZbuP+R6*4~iDMo@Z`q^vBQqJ(OcqVA%0LO9;O>4kNWd>CgCt_~1c;9gh=3hZ@| z-(gUMP(aW&G_wA#a)byN+{*cXflzVt$^IfJI09{a+tU2D z0`uTB+2TcC@z7jP0h^_=3HT=#^At69GV}J0%{;Xh zB*Y}~j_r;uk1dRhVVc~jpNIiTBBA*smG${eI59aiIorD$IX5)D>k=>8VJoDOUdaq! z+1duU&d5RaKBBZbgJt?}-1I-anp9t3ZC`c${C9?HiCJ$HgNuW~LS5t65lAx9?K0Wvscpzp4Bmn|xX2LVr z{>kw%G$?ylXW$R_Z|(bCsDKb~11rmeKMW9=+8ab)hL8~fW`4$VcX+hhz#UWh+s42Q zzdv86Y4{f1GS}BOJU{8btomf&wu&yY9v!{ZKXGy@OFMXb!+j(01_s9lz<2gepluxK zzP0;@_!l+2{Jg- zWd-~+e)8DA>C=9D!GEV8{5J1@bK;AmVy5-2E_$DS37QkC0WqWe5cxSS4iMM9fxKrh zS3j-g=y$hSsz8}q9cw>!)zOVeXYe9dQ?@^gV$$T2(r)?X4S>do;oZrHR-MDwgmsM# z;FRke8b5wjpg%NIQ@_%X9W*G`b4i2HO$mPG{`rINxADlXWG^MX^wB9nC#@> z{)G&1RD25(Xs7rQ%-2c&32GB;r}!1jKdbOVr0Ccn0E?{ofea|L{`#+!`F~3*|CL_% z`NJWLnLmOAGO&IH38j;M3E|^shISxWE`Ag08h=yQHh%7ZyFq12egpIMKz`iX*KrT@YPzWthv!e2Cogav`m-RPqeuJ;}=2Te46V1T)-edr@DjrgTO zT-0CpN`XG#A}~XaEyFuY{4NApwf-jk8C(D9UCKAY{V5YPfzMwp%7OdO7dgAG`IEr) zvr4fhsLAc0Yw_YYG74`M9266^*!w1!DDU?Iih7+C)y+J&w06=f&O786;}5nASWlFJ^8!C7rVhVjBNgcj(_}?{fXZfqa7(GDne&L_> z;{;GJ#OG+Te?27w==S~Y^nV?I0&WCQf&m_0C_M(18+$LGR!BnZq1_GvSdqNdEZ7jW zq@(a@;;3sTnH5!iaHHA@NBj6q%XrWiIQymx7`2C~o@;e0A}KB8x8_0a%jMl{!NZc- z0TBbn2_3#h?!*u-`#!g6@;jNpEAg~@x(%r9c{**1$&ZY;+aJYSmmkvVV#24dDle5B z$B22-rC5cmg?tdGvr0!;=Ck|VK@Mhwp+&=O_To9+?K%Bp8SZ~mVoHVYI3vuc9~Rb4 zVm#6xb*R`3hkUiQ;(`g=~fCn&Ps$+#|xt( zp4EFbmdiOMqG5HdX0uJN0C#o7E7zqL$7FPu!y}b)@6UbTb7Hyq7!^!vCa~6ko93Yi zMINHX``J9vZ2kumSn+rjm(8V{a2vwy zZ+nnxh2jU9l4`YiAn7E&DiM=88aziQQaC*Bng zpBnYPzNuLas~WP%$YP76;POXNO1U3zGWyUUg}l-pjLVc(jhYnnjZYb5B8Io=N_LPs z+!K8+dcod+-{W&-Zm=4IFtkSBicbfrYN=o96xZAN@?Av zIXykxMC05}X%{;=%D9KJm1tb%>2sg0L&0594oAR^aUi5wTdT2!w36+v=jFIM>r0* ziV^05GTncd&FI>`=CkhwUY!P& zKprNTGU+O)l+rYLp&{G3_5WN9y+aZ#2SGpf^WyJTjQ>6TixVSmBj2FJTfrb+Ku)B^ zyb$Ac+t$8GdEp(Que1%SJ1}P}apt;`(UTk2w|G4E0|L$J-@TXC_q$9Y}sw1|X%}m%CaA=2+Q$OysnpMA2Yz7R6h~(C7ir($JXep9g)3=j+H&vaaSdc1DC|0-2>~L$8G~(GbDN07kJHg9tz2;5$C@#03DyLBxn&AuAU48hWEea`!5SI1zVx zS@wHRpvuIaG$nNs2g32kIpI$p$&|WNG7m<(*O3W$N+p{bj-~j52>njV;R-0rQ3U$> z!ysn$X1WnLQQ++vHp9E_BC|=}q;~d5(XqsX)4{=LXO-`3Y$LoQBZo_C3i}BdEIe%S zdU`~ML`)=MSK9?)UDSG0nl@lDi&L2r(#nWpmxqha$`IqEY zoF!CjTnLRm8_LA?e1qX@eZ#g!)pMj&BWPJN6iD`>vTSPE#TGjgiRG96&Km=91@0+H ze9O48oL^=jG4&wvcY!wy3bl(+l(kq=muUjVVv3`%+t-o*ww+rW21@Wb7W(pnw(YNY zTKomV0E}ybEZw12D5?-x8P(QM?^f~JRK&Ob`v#J1ltHGtt?$}k98yzV4er<=L%Dp-&R1!6 zU89n8=fKQ-UXySUSZk;;o{2p*MMb$}lV3a8@7n2uVfiU=D6FoP?)sj zy!4e=(WLuURei*-X$KTfVxU~lfoh^<$E4vYF~Q+j5RtNq-1~$|S^+>E-4McHI@j%Z z>)tfHNCyX?=2_CZ(r#WSrws)hr4i1^BfZo$S;9L zr%B@{2PdAJ{QGLnlbBVMtK1~PLNin7x|3>oV-iWf#*aav3eovw;HQXP94~>tQZamS zQaVY@GvOP)|LbY+<$de#$*u@zDUSFwS&^ye!;Xb?MfT@5zTppVw3i**Xk!V#_5@${ z>F3!Mx6f82q@QY&kjyKC9UjxS?~)ulbP;%C=6T~jne=9Bm4&OT5ZR&Ba4QN`;FGpQ zdkTI_H~iCNWl^R-6KNE*WA-U@&Itj z2j<{Uv5K4aZWE%}5)-_N@IA>4-+e5&rnGhjVBJoh&-ON}l<$L*^p&V?FzAt^<+#kW zpaQ_h4CNIcI&R;ai{-HWDBn3u2C*lAOhc|>n)s_4+BxfsYhg|%c&}AK6p-gWdAl9c zx93Mu?iaAQWo%LOk2uNJV&VL0ZpZ)jO{_%yxV?_5t8eX0|G__I$N?LetsgOqf=++d zaIcM|+fC(uoTsiWGGgdHp9ArF&bvw~gZRFYJAdoRTz$grbxSh=BkAy`iJ%(A<9(Ju zX0#9X*oU$C%GpWuY}g4#pC@=IYlJr++_90rQ>=|5IQ{guL=U z4_N3YP>O^V)u|)s0qZM0H;jn`!Vw@)UF zS|4_yF9oV_pS(1?Z{NlR*|N<7x5^(|n)+AM$#x=~F!y8b3}D!NpOV?7`jyJFbyP~j zFg^#ZH>tRP^^84#m2AM58;?iew_z{BHJ?R_8l-QSBExeX=?vb(hst#51Hg+|@3)^?{&55c#!O|2ON(bPFyBWnJ6a_6MO^SE~t10V05dNI}L zL4uyv;(S$tb9!V_6d(6an3&WI4FUd~@5ADR3L!`hogxrPQ`bIlhi#j6{B5v6`wvs3 zcdNcUve|VGLIkpFr+)PHd$jL6utAlCn^bn5R(g(&4a-qnEv3yO4^XRFyus4ruN(0 zN8nae2=%*wzL4L`7O-vx)?bZI>aE7~a4*FzJ3k;q_8EkIN1(`&~>G@e<0uYCW^s zLXc^BQZwR9+_hm<3N~>)M-6z6*iM`OSx#a#sms(CK%LW8zv@?Qlgh4!;$K9e)Ii!7WIaP%GqD0e#KIIr8)owik~hjo^p;Mz zf&2qCJqL)(s!9#Y&YuCh*bE94`)FK1;M31bu`8QCdo2m{X z_GQbGd#xL#N9=Sqcezup;12VY;I{|_>b7q;kb10=BGvt^*vo}|V%6EB9VYGYIJcRq z%pLg(X*u>N)o3X$~S0Sa&*f~-n7(30?{OO$MnyD(5=yfm|l=?#hV zCLZ49nEe7FQQrKIh9H@oeH4`&*EslMe%PJR5DD8^bl)sDgQ=^OkNmFFRY*l4w$xFO zfBRU5VMRrMu^@Au%C;mAot~5k5arFgR*5ou8R0^8BB;BiYEX%ONJCLWm5bsUk?Lfw zyx7$Bf-=&rrMs%ncAGlP41@;|PK-LoWn&xDAAERE@GdxGU+JT;CC?gObB}&*B+%ZR z+|@+SUNmTdUDSj4ZYjt<8y`YaBx8?QvJcB{Vbo{3U-BYMG3~lwoWF`MZBd{^>bL-@ zFTs&v=^%?Z^$d!bfir{D4hNG^6w3 z1(W?6&(ItSUGOnZ(zfd0xN448F8ObgW=Yn_#lfx*0CuN?Sf{#*+U+y-4 zkxu#WG^0XLy*r<*xK_?ov`qnZa@8M-$I4lzvvMjwt#pj(WNKeoysa#t8x7~zCT9q} zH1!*4f<@MpML;mOZSK!fW#|`2}KIlC)FdHAs zNVD%{m6s&bbsFx2`%HuBA`-GxWpEssDKO|1wtmMWJWEllUNKc=?^Yb{WYZB17L26x z!|Zz+0){^v4Y-+y!@;sV*kUcc(hG1RC$xG#d!azRsn#cox^lOKwtW(EQIF{Q=c%mU zgi;;x9&=Y%Tiu*-$T zzp6X5_hN08VkDkm=`;hv+_YJf@#UR;!#mQzELX1X?XE#A$DQ7L%S8(adF|h77ZAVt z9pEi<#iofDdV%`Y2&Z->g-w-45oo{PRwD`wM6@DzYl3Av{F^MH{(5wVmKLPp@h`xe zQ7oa#0`_s5_CvO$;#N4|8i!n8K?c05@_`MmL}mGLCYK2x+^RYQU$iRG$zR`QatA&) zMzb@V^~x1LbnnhvZcGidT)xP@tzSg>dIraNycr$+xSt*S>=oe<_GcMO@loE{qwQ$W zq3Sag;+(W_J-0^TG*raXNIb}s%sK1_SXr^U|9P%SkRKnL>J=3#LwiH`_5qzQ7Os0> z4I#cpXH;ji#fmt24x{|CMSl>LW*P@PPx+a9gvO1IH3oL>m+`qmPBd6W!dkYZq(su-DX#wSGl_9t7 z+UoNKz@EMik!KP7vut8U>@vdKts4Wq#iALy; zHs*!Y)E5%+2`9abX6tr?o-n8a&CMEp6j!n4!O)Ihj?AoG2Ashgf6glytcEgs!9(MI zIuS7(s~RKS`?S>kS)!A^$jjc0O1umYR&j_-RC^88zX}ytQq4@jr2kpC@83B86gKbU z^jb^|{+PHPwel=8LSW|#Z>?*(a`O^Kl!-|6sKT(qSku@)Z2ceilV5X}8*roz#b$;X zI&_LDnnt0n{Hdi#p&rVH00&!0%pzgAPBIB811(r=9(K*9!{gl@q7_A7a24BR-ws@K zMTi89Kq?*u6P*iikKSi|FN=PXi~QK;Qu=j0ADomGK3jo)x_kXsb!5rHFpEf0SeC8Q z;TF;P6FjXkIV}jt++0XFhi0FAgPW&On&fm_xJlR~P}q!HvK}p&4DHW3!1TmhDm0kr z7{0J_;=d;b6MDF71C=#+QZeu18MY!E)C%EA)>=NndOc@b-o@d=>CpxQ>qF2OHP>v6 zG=aHMLr**dn7lCyeA@a9(&^ z+-6G7H!`oFmQG#^aNeAZLHo>}DvlybU&k6enu?4;ZsRc(C4qm*yC^w3RG)M+v`OO$ zOPh&5k~0J6o+M($=%<$z2S`v?LtXqVb64Z8CspbJ}ROWgaUM9$T6)z{jW6$!PpQ-7K}L z9CeBW;rB1DKA|mhdSsOuIRI9=H04&YICg1WX6#fzh2bC7^j)fW4wQ!(-Ho${&5_{6 zT`@xlzzwh+Q65nJpq6>~tm|rgF7M!yIk!BM6M?`K%kNXptqQBmT|DlZiOpHp&3VTD zFxQ19=4}85-;A{@n`D?G$=MakvM7u>{=61jCM=PQBz5j50g&vu%H{La**Rr4Us)CqR@e2DzfB;8s6C|996gSGt_Xg$l}_a??^( zJ83@XI3_i(JXaZ1m=S+eo+!5W0+x_5%aU3Ar5`BD32l0I>ws*KZpwL6Z|8z$AQz&* z^D77<8z-17^#b~c4x=u+AJfP#xbR=?)Mb-FMBBo#)9P!5!T!NlhCijMI43Q_8#ZD| zbRuQ4(kq+!9>Vw2tCY08fXee?gnZppVg&yLow0h$j4o#Bu=pejOp5gu``a$ALaO2W z(hdT#Z^EWx3#KFVS22gZu||pyMuOlau`ucYb!M%QpAWjG5C>S+XuU#t9!*wZ?ufnd zLwp577frp6y#sw|Gm?Gey6DA!a1x82$aW&UTrHCVp~~WN|L70jWh_rw*>JXy3(Kp6 zfI+bDZKn=%Zg+?dfrXWP47xE{8EK)0(&F>20(c+$oK1wz|W9#8+hx zDVQ>>U?$xNb`O-ykf!C*rfW}Bksd19*t&JHk0Oq;L+VAY=$!|m<8A-q2ixGln!n(}DeEWxcyL}bkroWJ)1JT7T*3Pt0 z$gX+JAGWb;>FM(f4LZT7xxJEnH;&AZiT&e#*-Uc5d<*+lgt>(<{<7gz|95e8h6@Aj zf4crnFXWr!p73n6Eg4H`>?RGZ>=Q}2D5_`Hp((8=T+>>$il2aE3k-%!S$X9%vs6Xz z;^5v1AEk~S{MuHU21O0(0wvPjbtV-41mIxuA43hm1V>3?b2i9hi2%ie%p--Kz zhmaYY^7;5{k;@luz?ORPKyqwp#X2_(^@M{h*iPZYxV(|X`Y+7nR7DQnpl1nvkSKXb~tG7Cq?qc{+N_`U?}jd1FuZPf08OHQoYjZ1ch|Z<9S0A2d9* zkawnpZQ+rm++2sIk|=*%C;Cv^35{q-2U&m-Wx*9unrss6Vom{ zE5q5fqq29>ZzkWq+pc(e(P5%;Xev|i?~-YsA6(_cY`ERCltgLmEt?l_OSiCFfVw_$ zp=xU*YScxgt-<1ufNE;Tjwoy|=wv`%KV7bhFTrHF|JbTn4fzjOo~tO~XyH)^gW#R$ znug2f=#c95T81tUfi6R_LL(=~z&7Nk2ANUOSV3vsDh*VB;dC=Y+9>HH7gDZom9O4h1iv5vHVsYNhfph52K>*`GmtMGFRF zBy1fv>_|;{prMzi)3;k6qSzl(|GZ8*LzOqKivYLfzlEfo2|BcC+_jx#U1M^bH%whkroZQIx!?ik|K*FmDe^-t1tZ=G`d8b>7MPSie{4satp)R5jwfs zpF_Mw-#ue!VXI39a(aUdCdes1LTA?=mn)dy!2Eu*Aq(Us6yxCJS6}`T0D*cMS<-^& zqPckH`Ujfnj-?n&|A(~P4XS83KOMJr%rvQ3uh*Iu?h!Rn+oH9S1H|3}`lIkrX4&Od z`bCMuOK^!}(_jH*DLV}aO!EpK@mzW_Y{XrA4WVkjyV3mPY5sV2yn&mXfhbbVV8EV; zTb;Q#aK8C?a$c(ILz7+Ocf^}Q_N>z4<#p*pj+L6;l#YEWQj}#>(E3gw8wB?kX8F4g ztHpKbWB~k*uh>q>N%oG&BdM@2$b$w8HP>z;dbCYsh$UsB6?;%A)q{=!fkkGWB&%bl zb1()o#Xa(k&};!ap`K+%rKLcQR}2EGUQKd5gSmw}KQj!&h;C^mi?y z>+Q}u;@xf7nQ)l{_^#EAK&6f}RtT)YD>2Gb_?o)i3~oa|51hh6G(Uo~$_>m#jtF$3wnoM>=VB~sxp$G zHa|lWC6e0BP!H=rIv!+x`BC5lVoZA~=5;0U#`loh*|hwJk#`PL$(1oLgFb->;yLJX z(BL1+PFBXrWma;2I^O5xlg6ui#|}@@%^Z>h9Gj}) zZ9XRTl1W=OThvpvO0RI@`SK_uC!_6VHy4ZMp1ZhNuW zxs|6EV|znJ*Q=1wed?_+=Pp)OY%r_OX1g>riV&I>m z7|~IfFw)M&+Qk}q4&r5$eM1kl7)a!%WLDz$1roC(<)6^EUIrt4o45v+rI%eJ58{6| zf$H@lG^ZhX!G&+IVt>M3eaTa9M^qhOECCrKVY!qve5GHBlzdeTqK>^u>`O{>1LPws zesgYO?hn)_XMKV#@vxdsFM=E-uBjn5;8Z#GSJwbO#*xR)5T!`?8ei7;$!h_iQH5W7HA z0thez#w<5#AVl|}WA#)r?sQeM&KjyV(xjeS*lc%yx`|T_S4ZI(DzgsD<4;y1?+;*t z&`%hDGp}$0?a~MMrvAl~aiubG=Vn_#XApVfT(gpiWQO*-E}AMo>3si1m(j?&rD4@t zUW!j;eHQ(mRuscYlPpdaalW$f3ut!%2)Dl&yA(~R+b7q1xNA4|^i{rH7qWMhc*uKq zxp?~IGwi^Le#z}hKn4itA+=?95k8lTq@nU0n5&U)h^N*m`sz>PWE1JBeqgfLPOuxg zY|`HA_-Hx$Z#xt2jBv+wvovR}kOi`~Mn~4!j6SQR>1mW^%!-=?5Olv97Kqzn@b-Fa zxpxU%#^C!5dZ#7TMV+gS}K3xw|)`Tkm;M?J-6?|1VC)9GN~)1 z%xwx}<1fzrnN&l=XV`AUk;iFB=JJZnc02RYu5Hs74$j%}XY66Yyg2r`7*!C@JMQAe z(EBNhYG>AZg-!pwvy_$P$>B!i$1Y7!$^J9nI}NlYSO?0hBEooACMk9J)7j6cBbwTB zqUg+*_1)Z4Sh>jeuS#IKw#Ia_-a~0QI*CIXRA2BZi34H!g}&fL<9=njGpSS+D%N=B zl?{qSWwp7)Ea8xzApGGrHZ!rQx}hX|;>n&xcZ~)JxHblD9BUA~!C}E#usU(}s$pC` zQ4`d;`Or+$qkEgqoc(7*Yp8dMa#g5?nvnP_+Z?N92O9+JSY36`Y*4nx&Gpy)FQ|MZO!&PAJeGX+%g z{(A1#=M5RXzi@DG-wo33e9EZ;aARD;)?HBjf{m|-a*;wyGC2d*nZ_mdwKPmrrjlFJ zRMH0T7`)cwTvvy@g>v_@HHUDC*tKp>c z4=z#Bw9`VoeMq_8y%~WH8DI5i-k}izv0GDrzWqsRsOpD)v7o9|?a0<_LfmR5XYljL zt|&l))qmKjPy8MDZ3S(Cdj&HiR@}=6N##`R4(zY!XuMW-D|M(^GmV9v=PolHrd54X zqTnpkuym!ogowpjFrwz!KjJVB#Q&!C2w40tMG(tP^@*S+afaG@pY7M8DSC`{g-Fhp z-L`|S`i<5cq38Ue(H)&au}e%lgYk0k&<4xGs7Q5D3!9W5O_#AeM;;ck;lDKGY&Bvm z7837Z9M8q|v;O}~924o}=a%n10ZGJ++~>)J zHC#=7uc~WJ5DVI&MEzDX^pf0O>D{O9rJmvmdN8e9lVW{g|B=s+x_nV>J=Gd~Qcrdc zM-pcSJ3ApU?;zFwTkl~mxNw*qHIi(n7=|ZmMNmJcn4jtS<1J^qVn!Gsn(}tx*jFmP z0C8G6(V*q6Jek#|(n^(Q5ZNLsn!ellYR1W@m-=-ujPnf=Ccey$u5;j29Om5t-yF%1 zZ8?%&eg(*bEHRIf9#}GiF6iYnp$xaB;=?^gT-344D^Ma(2X=BWIk4a+C88S?fn^u^ zM#bTY8}|A{?+vA;+HxQ%9lj(aV;ucXxR=BgiwM4we!jFBF8V%J7}hG>h{#h0jB#&^ zRn$I~IZlSttZDQ$qjA>Z9%`($^MvSZ-J=er+EdQ=oEb}ZSC$0>9$t`{!PajnCRKl8 zgal{2eqcy2!qPc&a9Q(wGQnOXjUCEU>=N}kpmbo$c9^6~#gCy?Vssw<6Ak}%X_95k zg5F7ezIdB~_fqc8PBneF5C1b(d!Lp)-?B_572iGr{~Xe~k7dJr)M$rWcrM}Jp<(ld z5v`f6;>Ggm>E&Sqfsr0!?Wq&$E=}@6pkpR<8DXEy@V=_aPJj^3^xR{A_09h5SsY>2 z5C+w)L;YGk%P_OL~f?4arnjZZaK=wu>y%TeD~@*!uf4+`d8XP1bj@`K6^6|)Pbs;RGo za}H1OTCoc@YeWbJcEk>SEc>L*T52IfqgXZ=X_K|qguqBeU&w1d?0J&QI=X=hFGIZ} ztcCXUgV^!U#vU5NZz+8wId_g(x-z6f!(oh-9X_DQ$+%2K#8k%~xZ1rO_+v<8cn_TN z*Lj*68X_gkk&Rk{hi1GmC!|;p+9>yi3Pi9Vlb%(~w4?c`8NXT@x4h^*EiU-PFBbgN zK=Q7LOoaQJw!|x-UG*LGZx=<-Uif`fv~(PBc*QoD6_CnxIMJBmr7G6YqMcE}foh^D zi0S#2)p6z|QY{V^VtT60v~b5*+@odPEI%BQ@l1UB))wHOI6;)YltyrZXLA%NVSV3v zX|guZ*VgJ?%0j>}R%zP>b*Eb8kq$KMjYKey4)}7l$qu81UTKXv*l632nKb4)%55*psB`r08#%Xy z2{`TkE{ud`BK%JKWitP2#B`8^aZ3EIa^C7J%3POLdcr)42)^Zgp<+qiI?=G6Jc*D# zWi^!sCx@UQol`GUHyY%c4O#34L=4galhwelR!280UfwsW!RHA*N$j%(zV z2d7HR*&TG*2iIl;1c4?>^f+{3Dnd$-@KQdrno1tVV946kIML5192V|}{LHb~I?)Q& zB}lk+d38X=%_1m-$gQezQBYlTsFt^*?Kol1GbbvIgm<<*Z($|$+uAJ|{t;9Wd>_0_Iy>ZQFbUt_M`CUuxmc7jK*&)+-Ayi1VRM}#w-xPxBntb_48G=B`uH%-+|H#2D> z;IDQ(1&pt)OiuvbKr6EQ-(!+C`|eiOt-MbQqV0u8_u2`q>&z;p0@aYjo6*+idFFV= zm~uHNeNEKOpi`gfrh^!(O?ZRUGKD9@A{FmT$u9n@HqKUfTEzQ-e6vbYCQOuMxRT0| zy_3j^Xr9?`{%kxS#QLXaaG})4px$q#`FnF+&ft^&WF9n zDuUv|o3Tto2n>n^MmEk+rd|hS*nWfe<4u1qp=e7Hi&3|l`XJ`9~VUpb-= z!b=ZH?Z^ZUMN($8!baSonI%1t@PDzNXV~|kODj;?i-bntC^w8IHyP22S-(p+rDKz#RL_}lk< zk(U#_NVY^|ts~DcQEZp5@?Rkz$Vf+&UE{s}(yP^u)AlG+?$!+)D_MKuOdSw+(k9UZ z1#SGtt8#Ct#%0I0fxfOb&iLWE^!iLNQrU5q*I0U%QFN^Y(iaw&1+q0}dqB(ST?vAr zfdzufV6=ybyVfpC{_5yShcb)MB*^K!#%$f|V43i0~W8y3fqxI2RnB2g$ z2GNZOEvh(s4rkM<8vR*|n@Bz z+HshPQIF%_lGl1E4pr;Iev@vZ*kkeN=H)0eWmos&u&vEKVPk4r4J+25EENpw{O8W& z-9iS*uKdPPFG2)75-?_D+EcGkrp%lsaNtK7M!0PK8b`sVXoa1nX{Y0#D-myG#@BFA*d?%gY=~VG@-N8TxV&>1?al zMFt3L*;m`s8(QTrJGB@>Y)C&UN03Xo;16? z>-=z>(o}N}Ffd@+m=#}F-p(DXon|pDjh{sG#-I7FM>Lti0Hw5xx1gp5to{o=^L8Ay z$R}l;XH~Ny;GG*5-RSJ)E+tu&T#@heP_;zYr&9{oAniTj}sn*FOG z6XDv2LFoRe0mc1dS6KU}C0NdIQHt%>ekX0D%jPSbLo)$pwx0${Vl+YO!H%JvC2BxR`@V_QwN9>~6gpY2&K4v7c)2tQUAxrX zN7@L>X1;IszPhHZz~g$guu?>n#Fg`QIS0HpLJjLqa+~WS-8o(J0!EBL$o54{Wp7G2 zVv4OJbU{AIRz_HxV#*t{1fpa~1B{H(bRK6QX(%m~eTs^qrdg&$t4?yhbB#Hr?3zH_ zg5iOw5o@p#*!`u-x1ztxcU@z#*G~C9H*+mgyJd}tb!$5|6-9RlWfZja>>?n*aoPe+ zE;cLD6|XzdyJ~r8RJ{_pzu}cOiaCMsT46z%$*APTb{lumDq`lVO3@!p1{C-gf~5H8 zsvPR;fW`kpZHg4v#0cARsn+3}{MWCta$r(x^@~9?_P+RblF!m>)ZSt}^$H8B0BPxU zLnF#`GE~!2VsDAPeNI1xixOvELVik9xa;lkK#If*k{z0Yki+hg`?6wW{8K=KUp`sz8zb_H<0wOz!w(3A%W#Q1fv;+BQQp|f8lkT|pj zrS+zmm^nxDpU~qX@t-=fl7#H{!O4vhSxk3n*G00ane_x`xsQ(=hAt+vc+XdzOz1`o&Gy!F z1>rd%RMBSEo;R7P#ah7YK2fC`C(>MmCye zH3-&jJOb0rs~##ZHqEasQkyb=d^IfvPMX&TuP#@m)@moqsA3{=k2Y8(pMO-!&=@m($aP&NNQ-Y09e=lSBlceM4(yqE`~zF70h$g z>S}w;@yWW32SHoc6t`-J=7U_&5!ct=esC4JgAWrYCjU6BHMqdV_q5{75S4FQ$kxKe zdpHvVYp<`d4u5B@4)h8G7%`TaeVX`6fn(SQzv!e`8)%K?n2A>;6L>T`_Tn!}F_vhV z51G1zX=yEfLLjkYqkXe8)5jL0M?G_5`~RwX+MRxQ_BgY=9dvVHTUK0dT*{AIr7si2 zkmp}*oqPVG)G_6;zZ^5ww})HXODu0ETBA`7q&=DL<_Btn0~o0-Oe(4Swn*co1cjyI z7=gpgOEwmbm)8yMQ1Tx8_X&zb#zSQt=pt29+p>-&qb=RsFnufwrG{@q ze|3IwF`KW9D0BVUDPoF4!}r3(yo~{JNB> zavdyxY9d5~)rjAN8J%Mlg9FX>aW2cG+lvv^3%l#~$Ih7XUTye+u~vOW;Xt%1^bkc~ zZNafnb4V&|MRlN;ySa8ePYw9_CJm|L<$8U{2oI1uzhgqOjE`9X-GQNND|V(j4*$p2 zTv={TyCcC$iLqQT)Ui*A+OZLx=7%-j6XecGl-9 z-uU##z6+S2?hPn1`H!aQc{BdVfo}6`dN$4b&fB71calBROY5p{-D{6WO5j4(4kedF zrGuhq{G~khb`^U4JAQ@v^Y=QaFmJ@poISDYB+Vf4@}3E7y~zc?=V@2uN0FU4v8|3Q zHPoGdMf&~xjX_t{n6yEapRmwQEhBL{c0#&Q$vx>jI}ztV8RxvKsy$P>3PyJ3qkqie z%d3;aB&*7WY>T4Fwb6e~8EFe_oHWQUheUtv1CPO)Kp9c5y@kS|yC^JSMvD*_> z%f2WZw5x+2U!{S<3~`~v!4X&oz5bZB^#nG8(tG(9airoYoEevraA@F@pBJv`cmRY%H;ZP{w)UJ5)PxgnXz2%#ke9S{KyDiH zR~S#Ex)hueJ~v{#kaq;(b2-+OxlYSFaW(Ds@qjAecO>U$3^hIrJ6^_`?vk`tMPnv?N!2cU^AdV=qgm2&2QsfUJqjrC$F`Z?FUJoL<-PRsN0iag@)rQyl zX(3MQ)PnMwQZ-(okWPo_#xOCb6e)*DgUOwlppErIVxi9M-PFU+z$xsI7eUM5mj8}Q zbDEO+i(g{p(iKybQP^DBu^qH0 z@44(&?EL0UViW=$NmO@`0WIiIZg7Bw!eP~r0V8W5jSL{+fJijFoWZ5Oup3+bl98!c zv{B_0I-n+{2ENWzdQBr0yq)D~W${;{TRo$iT-gEn;eanrKf#L$K8sLDfOp|WI*nK#<8WMQHLHkdsI z2GwLCs_NE2ymVi!uaUM0&)lvXHL7A@TC7`zPk(XMd{>e{84(o@2z+20(m~iF&l;K9 z(o@(?hZJv8>`Ek51IqQF%Q$(hj<-~$E?%3~0v<&~&ubWFxVwNz} zBfHGF?^NK8ZJTJ^4=S%fF*?HZR)2>gw5TwMn`*RYc?b{(l0xtCKfySZZN7Q2SvJjV z7!j4|SceuA)F>uHNaI1_aLy)%K%_t` z13x>lr_h4d3AWkmH_D6l5@FCmI4Ig#uGJe;c5aD-OAa0k_ZV{doTI?S5&q)aml|!SMZUPUU>&M;E)cW07HUggVpT*JZ)D}(3)xjiq|=6`F_7f7 zbCJ$08elp2RJxFr&!&m#LQPrhqo@gGuwNWO|G@{?Ns5xOZp;V}N=fHli!DPbTcoa| zvCiLVnp*VAE#7B$hZqAKZVCJ3>Gqz)*>ksj{nz1^`FlHY6;!g{PiIeMVr7s~>plG) zxmV!8rH6guCf-Exvoub*mnujeb7g9M2mXfe$}6H)Zr^UnwQYkL1I`94O-ElYqw-No zEaps}G=OIScYIIWa?Av)xcpt9=Mu!d6MkTjqn+C#?y}voN#r%r>D?8@m^!vghnCSp zJ2~E}ylC3_!yMZ&9U(shZTrG6#gYc61We-8M~{tLrwRJC?*kdgMzc$1q5ovAR;ip^ zQ~Dx}kZl7W@bt7Q|8#_qA|*$@4t>WZiw7f{cT4)d$OOkLtig-9&LG<@?eaFXe)(*C zabg@)LTT3*#NqR?P=hLE)KSxe z3MH>w2yXIcs+w|O2I3+u^By;V-SSJPy8rrCqap|?A8rZys8hZ5kUW%XT3S;NT4))! zW$;Hy*0RQYj(&-pLJ%rZnmCD6kKdO)+y0k}lv%p<7^WjFkttLbN4#Kd{9G!Isd&M$ z$v=EW?WMsNV6%!pnw4$i9Cfv=WRP`*v(hC-^tFlld#ype|J{BsBy_FBKOc6G3wG^T z=V-YEM{w^o`#WGQTZbZ=5Q?sr40`2Zdl%-QQXy2|J%@hD=td&+K1ZKk%6(c$CL|Rb z4>2>s(W4ylDiU3hV#NFOLI9pS&YCbdpkdt3uKJQ(0GMw`br>rGKVRA(YPx0d^ za?tWi09}(Rw|}Lw8YZF-Xuhq%Atw8IW7yW};|sITZ@I4K@MsnVvgPeh8hvh88bd}t zNNl}1**9F=>RCO@0|rR`2ON5?L=rt5HzJAWK0eDRrBALQLJaI=py!uL4 zPMbpnaC8$>#1jZbAe=tfeE^UzhMoh1B~I+m^7Dkv;D<+AsHF{*vRr7T=BQ&Ti%l zYH#@kf$6(cI$T+6N-6LEME8)Jv#|)Lv+UTlRM|EK4HYr0J&P*hVmHM3GGj35j_Meq zobWAJHx$JuqR#s2rYjAf9N9wXrCrlWVW&WW#8kZT$9Ce!8>8e1kbeRiO`|S%-fYal-fquLe2qQ?cQaccDYB-KCuBOJD<_}45{eVnBj!0wbvMfMb0U- zc-RT3To^PJ>vRZS<2^^beWJ=@P@PAB^Q{ia$>P}KERxq-X{MMy?#!j6P+SW+Q)Z1v zqZ*6%w%<_B{!hQfGqXX)DQt+9JE!|i!3JsJ=w3~oO`sA0G0A%YtCJmuFZ(+|LEhC= zalharfK0tc0cE%`t}q7bbu%_EO33*C;7RZtN3roG>>(PAHHFy(gvn8r*en9|K0njv z{<#f#q8c|KCG_|_*72%tgg~I=tF(TDmVbSBTICxtu=Lb9q!4zsp(F{^!BsA9&iTw0 zwa<{YE%APQJLupaun5pE29mD|jIW!Qf^a;zFb#9_dN_(9)5E1!q;Lr%js!7HL7Tj= z`?oh(k@I+7`Z%?2GritbCnWq#&uHC;{PsNpvLl-Ioho4Yyj52N_fcH0i-1b_)xZhI zSp*pPZ2ueB&PDWHoHrW{H)tfUm)JpsnXh%*bc|L?JRv!)k>x4#S(slb;S&CEI@?&ts7U#~7Vb`g7i*V&O*A*Rn zQ%g077OZTpC#;xVQep=cKg7z$M}eWauSgODW$3(l8myE)r)d9W(tuLZ6Gp3`d=sO@ z^|Wb+7>9K>B*c!VMBc|ZY79W94nJ-j6M3@Q^ul-*lb5J|WdyYuX0A>hlY5(ovAdp{ zp=EPKVxrh&<7Anq&cA>`L zW;4>sqRO1~_zY-sPl^4Aw&8DQ^IT3q1tY%#AkYuaTjFqI9m8}`Zwf-f_{h7!8uQ;i z!ITD9<lW4tE8_Vb?QwFHifM1Hs>~O&nIsg3XJ|;Y}gEUM-{Q^dzA}^;&SvKCf_W z4{f4)X(fi+(dP38j>EOU)3i$r`Gu{!h428*e3>Q<)fr?{vAa~Y zAtP10A6*HmMIpV-VH(?etgpbEP@^DM_tB8Y0l^ab;Hi79JG{pro&BPOcMoHqws~T( zJqJ`j*DA*mZ6*AI8?|X zEoD9DS|H_$<87b)1#LE-db@>tAV!%9r0(qmgWKR6Y*{n>hJDbqm9`4%B0DX`h)1QL za=f9`X-!!Id=|sBq=so`rJxV>%Q9knjb+T}68Eo)8Vzv`J(?Lc`re;3R@A<<=rpV_ zQ9RN%51Blquo#*i4c!yl_oL4Q=}6k{eE;b?5al-qfHr7egl))DJ?ic|Uz)>rH;%TA zk=xx4emi}0=Lmgg;8>L#1Qb!-zO9CmmWuN_?EYsKX!~#pv2%&NWB2GlfVICH$Plzr z1#bllSp=BBk9adLtKwqbf2Buh^x!9vy^Pd%UAaK<*_xQK+%prY#p==UP?M?%m6m*s&4BL?N% zb6bag5^((vcP|05HJW%!QR7sJmvv?w0%MOY6}OxUUk-l?75CZ|mFmk^$9hO%Z8bo& zqSh95cviTZ>(eo3H3XcF{$$-!<)Uqm+5su?m({2!s}(CrE5>WChF1`@fu$%*4zbfd6fnWFZEy|DP1${~aG> zX5!#v{{Qq%#fpkmmIM~`>jrgNgU7lYTwlXLtQhpeTwh-k>Y!~> z{Ao8ao0-aa&|B(WQr-Gl@9DL_()C|ZqN!ww&JdLSlcW%V#)=1*fJFSeeqaOv1Pq`N z@W%~ZON+fB8R3+X8MYQ`tRDqF^!97=&lIT1IuK1Nh{FgWTByIy?jYJ4+QTW4`(V{JjI6qK39#=m-n<{q@U0c&@SO&uK_P1qaUO@9dJ zAaF83>I^8Q17OEc4tK#=z(4J<3k+=F-kVu5(X|4yC|2*}|IsXNjxVi(LU>@V2FMB` zkD4QyN8pTs-sa#J&^!_rFCL$AiaTycB(MjY@H=k=2)a2dg^18-y=Tf!CLu84TVxH3I{2+j9ir(0c^w@#Bt8VSKBv3i9Ka8kivr+U1Y=;Uo1!y?Rvt zQT}7iferFQAPxF1qlFysO%!(l3J5Hl)wlVXmH8Pw`ibA|E&kl$`tF64RavaUnPKil_7VNlMA5Sc%aG(-2C2U3P^_wT?2B}`(iFg4--3dfs^;x}hdVScu zbwS(w0DrENmoD{h0W7Eb2?Tvs-<*e^u>~`TKl9FD8vQ>(v+Nt6aQu{QU%#68%7gwJ z%Y(j^*PE{%jJ|(4gKAvm0bst=9c^ZherN$3&Q7EL{pLvJ{{1IfI^O|5u99Ulh={Y+ z=kd&5J^eqs0&b+xhfpk{nmNH)J#bA`uvY(Skx3j)w{g%i%C>+qW(LnXE*7rek%1|F zx9ofd(|cOvh7EAmp3*_zQXY}}z=wiHDC(B8WT%Jjm0-h~i97b7Z$<=v>-G_>tEQIxt6Wc}6ekCJUGI|3D(~pt*pR6!svEzfBw=5wJfk?Z z`1uFMjKXP-@eGO*As3_H!e&DN*TI~B9eE?;xc)^6vA)6kzSz@=Hno?fD1ig0h)y)6o z$nOR(o+4K43KEw0w(tb(MGlpb#xWal565(;4Q#2~QH)o#<>6vd-ipGi9w+nO1*H&<} za*TXu&9^87Eh^2J(}mbeQE%F2&|2npN(PA~hK-%jPm6E?ZQ`pS~!yIy69`JJXuQ|p%H<1#`V0DgqotfgLaVi!YV=W z*SDyhdiItx*ZSR0KYr7vt(JhLre}_pG?@{0Afhi-oPctU`L#;6aPV%YwAGKvTQNLf zQur3!#fUrK@4mp_)cZ+;TZ-K6iV<>#D;D%X!GSB+ z4;s+l3&x2JT)g>JfNg0!SaqjO?_cKDif?}@x^FzjWrd!Er9Mr=-4SDl#SI_CAjiPO z*30g&g#M5vyQy>YvV^qDtLOKuD-JwNg5!TY zLjuGT`9*OyCu*ieYxIcDjlat^W}zzsvkN15WI)qJ&H4ph^muc=f8i7DL6t}-*PwqX z`fBd85A0lw+G3ejF$r8>Xg>*Ft~+d>&d64;6;3_|5*gBR#%g((rSoI5VU=c=xpE-;{s#bKjb6 zy9`VdI14^jA6NTA&WyU7)Jx)FxEvp27FVCR^RN`c7ZuZcE2O4bov^`D+^nd+-$oFG zdI%DU=xJL!40iyQS{dEet4Uyrf8y7*EOow8QJO2Egc7UXsBiT5hyp^OJQ=ebc|DV2 zEJvjgj%eVq03r#oC{DvWr5E-%A}=8*&{`)$73$tFI>)BM2m+7wT;3iz3Z(QcsD?IX z_oi>HP%-#^Y@`%}W0NBHk(hzxeNw}D=cFyx@ITjzu%T=XN z7IgV)onG?9(}IV=GV_n(Z2rbek0`$2)?I@0LuKpBk0%D;8<=(W=bzyn^4}EH#q(0) z%E-NF)4ywd$tAWI+9KC)*Eq+QjJ@2ueoHi2<&i8Z1C{bonbNlnD#eo)V0`MFAEjnJ z_`gQ|{WYU$wLH<7ap|I}g-pb4*SdltVjFjQsLk6ky_OjKJ-bg!83-G*lvwb_xx9u} zJ3ZC`rH5u4X_|X~dVZ|PlFV~dttDblG2E-A5~!+K39~M0F*|=^OehMd=XOVw z@Vor=7S<4@$`!oTD?a$rYh!)fq(#bP8Hm&~uWI%Q{n-0TG!J0yJfLct9F;a+CMn%L zOz;?@ULv#NAQbBhdZ^HhVV3M{dfC*{lABAwfCJ!hAH}_)DK46;?bzM^Q~Qr!O0rCF z&aG@e+$i@{C}v#p&Fz=xMRDUof9n@x@B;TQj*Z&`cKh|NdgudPFMGm06MbFWOQ4kI zI1#`lH;x~g z>sZ%FD-&ORzz$)bbUH&{lx_{O+(8i4p&N$m1%;JhLT6>6qM$!N)`DJLWSXh0>(kkM z>TE5s9^D&Vqq3f9XXOvV2C-zi-1Tm4d2!$uGruvtdD1H6H&Sy-6Q|{;mOWD@TCs(> z3?nyBWr#(8=Vwrh=u3Y${_j`g0d0KPL*xPF%o!%c@QNT^-U_4y`5dnyXs?ou+DX2a z0al^l?pzw~6TP`?Bi!?^yOv3V|_L(#?Brl{3t> zt!BAfxf6Zy=Rawy?m1dD$e%N*a4IKpaaJ=HqjDy+!)6`oGp1cT?mp1!OSZHhmelDT z&tJ-W+V3xXCZl&Briqg6nskt*DiE7}blK#zv3Ct+l{5)ce90RMf~Y_8 zvF!SNe3p?djujmbX*jP+Qz*x~$|&ETqUHu|-+Hm3xtsk;%eoPfill%Upqp+;1GdG2 zmNVQDd)^v;xe?0ApjhMH*r6qfLL#7KqLU&}y32Hey?EW|!qnn}I@hcqHA@>?FW_v> zRWnTLp*b3X`MYGp3@FEQccx9Q`U)@@|3yDwcB!2?>b-f$+~Yk8wy=x_Gn7jnsOOH( zh@DF*TEuyAM79xQRFI=0d@dtseq-XXZK^jJOR9Hdk~4Ke@C}aBGasPgos->xk*8PY2# z&or|dUFDJny~mr^*iv}a5{_1TV>xr>g7UVERj!M<9-opm`(Yz3x8WX z6-K=FZjG{i?!XBX;U1Qg`qgn1zR_7>P;>l!lM)nSp|o0LBU2buM=Dl%vJiJJg@%kM zLxkV8jw*fdQ`)I#Iuca>CPgDY+;xUcsw|+~Dd=UIc_<#9<>2Y6HLl{(;){~tPOwqg zU9BK`c*~dR(`It~tGG^m-A7|El;urfn2(xPc0$j`r&k^T%?`RqVsAFj=C5ptjw8{3#=5{2&6BGpz2dzvnfVpS z|9w^vx9kW=nA3=uL93VZTm<=SVDPr(jC=Xt)uB2vE2E1rkR&qo_j( z9;V)8Zp-q`aC!q)22C9-f_lN<>r@p_XS~I8zm4uYZMjG#RVzdA#~P}3n>DBlc%t8E z`C=GzPw%bZIoO0L|D(9Q;x_L3CF^)z5H|nJ>k0umK7qREO2F-aWzuq@-4@q;GB;xkxt&yXrssjQm`N6wtiG+q!vX6LZvg@8b zg439K$cV?J4k-2zWbVZ;*cmp$di?$+;PdWv>Bq@g znp35$S)LaijXpXpj?&){USd?oOG>=r#p3ZWo2Qv1vq z9fZQ~A;~;c5MgO0UuGZGJa@S+vtOu&|K|f?U6B$aFe;tRmL(=*L|k`|>E;L+4Ei#Ii14-|78ELUTz4wSxY8!)rKAt<22>h2(Lncdb zNcxiF$H!lWx~2PBV~5ze zCeexr(C#OeEK2N3fHacvg=LIl^D#`ZbhS8& zQRB1?@|GKn)Led2v`UL&Jxc%J#pS+av%^}RWG!Nv?j_~O^zxa~r?rxAAYLTAcY%i& zdIGK1Vy*Bdda~7=|5{8Pb;9XCa69qjiWar zR8g~1BILTt#Xz!5gMTAFcx|WkIMmUC@wg+L0b{8Fl#_?`7VS2rC7K@`eha0gHZjQ}@*R{WR4gKUt z_SHX2C44|gq>NH#G-5FYu?z1WKgaN2!s>C8Ovyceu==u|0Q3rYxHyc;ZdQ!wtu1xU zF*K?+y=o8YGP8Hvv@j9myojXtHBtdefrB`Migwv=$-Shsi_g*0uaIS`9 ztp4k6XYDnJ%L!x_+?)DzR~Oh}A95wgqWLuJBPL(c0;;w>6*@$>f_W`V?P4yBo*ceQ zZ?pJ#;|funy12_eNX1i|l?Zpa`8$^8kAqE7dU)N{u+r7N>`hx#E-1vztg*DSufLU&62gz%D9+ilD;1z!aY0Ronkqny zj*jYkrbT^&M&A^ANFb;Q$7XRIKVGN7hkm{3Z0_S632UO&G%h-U|7|(bYO$kzcz~I! z;+R$yq$Sw6^R^1$myz?$lI@ak#CUjed9^SbS)3!(4A6e&3v|I_MibJ+GK3J;Blz-5 zt7|Ijl`-s$utl)`d&lY^#P1UUB zGxTKW>9tZ+tGWw}(oZvlsz=v*+Sj>)fU;It8Lie2@>62miH4%zLJboxdY0LUMzs)E zty`(x{C+!gRubjsmT0HD4pICmq10{Sos_##yl7AyOj@e4-j}tcNJe;(e1B78yPWbX zIF)Vcbgcsb`%sPh66KH8y`_%SDdnDdL8Yb%X3Q(pn`p0#KlBfvKxC-Hl|ku5N0AH; zuJcoiu?XxW|4j zt8fIXF)mbsB3e?M45MqenxBTr$dODWSFHl%`+CYg;1qL=@%VKv!a%(Zmz<)|90N%} zig&|#RFbWP*<}$v%MXb;yYTzp-O#={dGdxX@6fn10pCF7#Qr*3-o}Wv8hJUKvOL4k z_dLwA!T=hzSQygy`q@H#HR!Dy3`c>^+2J#`=>Q+Wy|zf%fgY}+ zy$W(daU=Uu3c^`;;lUmEeL1I&={MpzJ+h57`|^tx2YtMy*G-*RRlTBmptq+bYcEZA zUw2xWe_-?j?e}}Fp$4WFL=zoWlUpE;{BJ6$QvJFase6zEJMt`4D*rUX4Bqca){x=2 zy=hS%mKp>FZca`<5iZI6L&x_cM^-wUmo)&U%Z7S3F}x89PD_|-W7VVuSx)CK27&^V z5`b(uhO`(P$v8@rE5V)jNnT83FcYkg-BlTuOOy^RK^5f$Uy*T?@+b(K37IMjplJ>p zi0vD!)o0xm5oK}gFTORk-jQlzpW>FNg*XOv{=dIS|0+G| zl2(PC`5S6Bzg+}({7+ifG($&lbxu&djL;XlNqBE5}gAY=(Vu`YV zzdOPK2hSaYHBl?Mmaekrkf$B1O-V)#H*GXv``Yf6K1=2})HAtG2Kv5L|AfNJL05DX zu}2M&NZP#$WudwmmjQHh2p1($5Q~3b*!)H=0W4L%7%(17v$MC2Z=~;ZRHDbs4(V}G+~He*C$7M- zD<7;ApcP7O!!R?~$)k!X(KiL{4v;PtYta15%HfMp;jxc2NIaq6dg!&zlfz)JOK#dk zDouQ+VLkf>lH7c~8%V1MaS16Ij$H7?l%bNDZPN|u7jN1aEb{^wk4t)jH{@;B?Hg_xeSdE^$krrpP$+UZpt-|7;^$E3uwg$zFp5VV8W~st? zX2trb`l(or6Xvg*xCbMuucv-3rdMhZlsupKLZy+$q}7p-cne$bc!{L-;|b>Tq@a3^ zOBC|D+pJ^k2o=?WH?KV-JK_dMOX9+gsn$G2g=*LUI~wh7xtN%|!El)1;uNi>GR;2U zLI;*?uTyz@+GRQpc!X${VbQ%skZHriDl<8wtg`!4oCF@C0qnw20Z<$}@iftkJ28OG zlck>Y)d;Y~YAqKI@kJ=8ZIu<%DkXvfz7TivQgpw?AiiyL-XZ^9k9%5^( z;^zC^-JyB(7w&#!xPc)FY^kb^*4w8(zu4)`Ka@sxWgi__8*75j71{<{U7Mabm%lWi zVNO%Yihf&5YH%3XWT;0aSC>VtjDizpw)5~w`aWk>IdOFz%B=GjxlqXXc*3IsJ6BZ} zjA!CoRz_ncrVbcP(MoM1{`}Psq;ZVeiIN>Zr%jm@Kp;}NyTim~JpMM|$({?U^}MpA zPP{yJ&z9f!I(wPa8v40+;pUpuvq6TNH1M+Jp;F=)(N9DRyLAzydgQAy^>*R1#jgOC z*jz^a{(?GB6&(DARrO5R%+}M@hLHcOrnH`B)$j1Hli@|%$vsEy9}xk9JGBVXw0aWF zKG@yx?d?&CvV+A^t3~pZyK_^VQL5s=a4SFT9;)5E;61$9p@?Y2kdbT<;$1`|zVl2p z@xR(^FSB}Gt^I1xAwWjw&E$!uY{e~YVR2RK6~0|t`_U#|&NNE}Y<2kG?kktdlcJD# zPF1$iZLVfI(=R<#KFg${MIC1EyLK6eG!Tf37=Gol#P8!23UrzrrsmN2tQXQz#?FIR z%wbZau{5E*pfsR7+OfwV_S=9b5*Fv=h~fH2!lL;tH8x>A7@WP1_e`CiDH_~Z&pwaz z8>C6ojBBYORPrY4k4hD;*GnV#?E}SK9?iJVGjlzc#q+x5AHl@8)J#1SI#qT=l(&%Q zlJDW8ZNx5A8R@Ml6cn8C>kIX@PX>-M3#D`5U7eDhhTj!%(V48CDraUk5jY4J+(~q^ zfgl~pE%|24Hqs`!kX5wA*>ThB^`HSEGpuF!=mW=x3yj%`1AX~P5{W>axLwU?mV_@i zs`4|>_>D@)U`gMg(c&l&R5Qy>9Hi1}ACw=uANt)9H#I5sOsm#+em3of} z&gDOkOpUGC5h?%47A=brvEgl_8e@G#I9STA>4A=l*$n~VNpy6_L}^S7ML$Ul^YVGV#qjinjI(YzS~iY?FP&4 z{h)yHkXJ_a6jra|D3T!ZR&$JjW7`zd%T~O6aXV1iiB9(4eb53=miXs;t|2X)^xizYtJ*!@mCRK{iena|3(`tPN6c7I`h_{A ze0+QRYe?g3*{_4p{GzUV$nb?j-pljE4w~Z;L;r~%6?ji_W#n8JsT;x!=+wIIq_S5* zr+=yfM^P8bB6F@J^Rb4_f>}|5Bi;X2HMHz?mnz%Jg7O{oFp-lzcO;q-PO@8p4O$Bv zq&>-AVcPQZIC)rc>}<#J$B|p_r|t%t{7W4~CN$Jzo>Chzic~MbA09l%guGLitb7uZ z*1|&uzZ*|;@t+4geCW)AKwE8lZi9b1N69(YQ`}S!xhQp&pzLbHiL0HC8jN)CtnXqM zK948P3iXX(qVg(gxL1XhpuHGC;&CRE0=H^htZvPcZTm_V=gh$}yf7ehwL>19^M*R7 zs4n9XNKAh;>#`k)X zT#bsPi=(>H1fEBDc-M?$F;ecfEZ8--F1o0M4lPfA>wXCOC=#49O_RXsgdGw0gAwP8nt?DrhJ zB)vq5N_PE6@mP)WWt1OF-(ZbLR8v29ds$k(^JyKcnLf=Gv>Q)5^(5Wa)ydb<_foXIZz+>sMtj)OteCV1(%0ZjIM-SBQOW6aNi$2WRFF}QK-U~4L+ePF zmnEVOc|xM#g&MV7?9~;sJ-@g)V?lCS5!!$`n%%$ryVa`x?|PJ65QUuV8WKd&CB~<~95UzeVk-@Wx-2NdqZv^1pR`Y? zmNXy7WF7H_?<)^k+(j9yk-3Ydz*>>FcW_b)qL~p*uzt^lPLI1z7K54&0hdjJe66KI1Xp`k6b`f{QJ!!QVSp#r( zd{{&ssy>j?E$^Mtz8M+#_VL=>SE=UV!PYt%#?%VSPff?5H7 z7_nNa!)yb>TOWtngQ6^z>0eAf#ruUW4T(i0IHcL2!ep$)LtKc^N=2){Nas>Ms2+5% zCCe@8-m-s6t({+_b%U(8K}a{5&0|9a-^_p|d$+L48%k}olL$LGUGdhGjk%So4=x37pm<7ROMyC0)6mH^3?aBLf zlLAhQPY?KTK~JHm>6T`XzNnx^Nb1`R{Ijz#tMXkR=b8!u_KG6ffC-RlJ1(Qj>5#$x z0L)r!Knz+nO zY^{7Qm6;F3V0~P4)+0ir+rHLhTkgUm7(1Zk0%?=zQ6wEOeTpgMQ-1@h}u%1 zrKT+1ezGu+A_%!x zMrSMM#SnF3HQlzFv7ez*v{SJCg0KwtqZNP6_Ni~$rF{Ak=$AZ-o6NY?mv;0;m*dN= zG*PlYeoDhC&nRa%AfUPp8^*2WN_HF~q8{Pxzbj5?R4!GhU$WY;c-W9Zow7P>07euA zWIBb{2|7O({VH+4*@DmcGE%)Oy&;}7Fw{?SvdR5>`tYLj%`Y37aWePiL8g>3UGh5? z9&B~k$gjy!#6t2fi>HfstAmT*o+}z{<64v%UqE~Or5o1E7xB2g8>cevC#P%8W$&mt zyPNl#2L}8aCS+h4a--p2hM)kxH6Gz#o7(VQEy?FK1@$Z^c^_8v$ISbg(LJQft+8yP zp#3Q1E%@E*$UWwP%1=74(3Woz4@{&{r9CK3RL}6`_aSy(7X>rJNv!NHga+O9*6`np zGBHA4C0h{SGyv&fq+gOPTa|=7WKezP7|8+Je43r8tHn7zPJQx1#M(b0yV+BfcRlW0 zRTrzt7mkCh6F#BzZ@vI`HuqZp%_hq7-)y2x9E|@ljWQCjGSM^t=k~vCqHK&TZ2#jD zC4iz6HMepyb|9b=wbFMo7B)7tH8O_c<%M!|axm7nhH~48b^=ve+Ww&ZLxG#XNU$wb z>gU8pbPfOv1H(iQvrV-9hwzVh;87?u5zY2Ow4Wb2%`^Tp*XdWU-O!Vl?tggu#UTC8;cqyoX`W%I|vCNl*Y#a03aYCFCd_Rmx_uI1QOIw;6Wo+AfgQd zyYz$aIx2!50#YzVKpZ(64LeNU4IC&yFaW=TL|$5v9{|5RKmUPlj9m~3KjIN!J*e3| z;L<*Q9X4+z(AF?EV7)7Uf$aM#QlI`HB!J|<@#PmB_!w6pA_6u7OdhBpr(d=L*@j;p z0B{>DNYvAh7!;g2ILU%YLhkP3Vj}#})d;xQ8sEtgpqpSE9DsKnHpV_+3+R^xdLFC; z*so0#d@icq8G!vKFm~T{Tn+&Uh96KK0yg~=Eb1OyBPcOIStB5rMFs#aI{j5_;|V@E zUf-@NFu()gZ}!dYnQk~i-EKbtHg;PhG{gz4|2ps$fjm9Gn0g}odGHaSpI%EZj(|KH zJ0;!`U;w|s3VG;ma3H|Mg9iXYMQ?AXW1|4SR-BEO4Z>M1+`DhFfOZ4YlnUU@6-)@> z_MPu?DTIKZwS4Q%^si@?EPWf<)a&d998hEP=W;-GCv-0&%)tem%HOX|V9@IBR3y+ls0auEKOh4j0EVT3-&}Pkfw2&t`4;Z*Q9*qTwgGVwC+>do8_7f5?fT00)DY|g7{g?;_HEn?fqtuS>Sz%gM(mvLH`6o0U?6E4+;acjp=qODtx~5C(7Zzi8r+657|?)&AJ_aHkad(b#bWI9#7nEP%8L(Rf7ILVkqC9ZPq?G^T4=0 zb%i)=`mojbjP%Q<$(-R{RJqsO90+&3AED=LQ|dVL7()x`GiiEkDi>gZtQsSC)r&Hm z^bh~7vaMMy2vYq8WLcG zZ>DA@U5-gFHy$#w?v;DhrRYn4sj1i&fFYS#Vu7vWC<`d)%I?+D^h}4)$w$4eCiaKJ ztn8hEVyVDTB||weD>6NyF1RrkZCSNz$tR3Oz4QKzMBv6L&!AaYWqb7H(D~fjra*DC z9N&2{pkeU>o9a8>NavXwe~EjkFsLIp|M<$6vbFEW-}g_W8-n~wUz))CA5CXmz+}bY zM4W>=$))2}+&0Pf#i_i{c{UtNyZn(GlWyr%MS5(6G||xOf9>jdf@HCEq0stUjiOK6 z+f&ntM6>;^C^}PVzL3FuSuqerJ^Bch1>866tx^uWP1aTg)x0A1j2V7A-k6VjRmSoN z4)0I}?k3Tn85?iz*vbGIJj;#b55o11S(oHFq=s3=b+;w21b}?dc)wZ2t%n^~y4qlJ zN5Mbe_&dsw)){+bJH$qF4{d$a`pWaX7P;5%eM|~VBlve-)DKzj&%7lcJP9DU0)+SP zAP}qsA*+Hi$HCJknb~V6O7in_@crxh8$<6vzCr-=*Qq)K3$*Qq+^lqxP+%3woQ(G1 zU#UrsJui74h##oac|GQGF96h-*(Jy_7TBNA+s_$1vBKe4$Nj6kd=%;uI!W4bJs_Xk z!e>SIavJy{O7lUjzC9b8Oe0kAD+F3{4OSLcaLBn(w(%!R6<-Kg$Ss7?k^5Z%sG3A9 zPshqc-bnVvy)iU%150{A2a`gk+X$Wd`xoUVvLQ~s)nlt2BEVcP0fx=;dEQrZ9S`SY zHl*{d$6n|I%!W;`KYRvW*b&9k;uPm11Q(pM6L|1)k$MT~!tv=Xi{SNFxir8c2|eNY zwP4H&hgWGQdHQpyGo>+jD2TnsEKtfJ&Og4w~v^05Cyxy$%cZq$(Fp?Oy@r4`-}wcbJ7=A8`IoA zc^id{R7_VH+@s_FL|Yxqm-^dt^_Pug_1_wtn#RWlbiqZp9faeqEyNvJvK8DqKxX@dTA_VBK`P4+TiI}63&u_mqMZ&VEn8%dO`yb zEdK#k(iAnQLG6| zZRLw+Pu#%&g=n;tGCj`LcH31b{cUsRk%UJ!48ovrWn5dO*+!}HE9))M;d!E1+mTkm zftNWY-~;|ty_Qfsq>L^Yrvwdd$=&$7ol3}aleMc~$BV27hg<2wvLw(tH^d5WF^O1zqKa`wKh+Jk+= zt{pU=&kCxgsrV;QZ{7W}I+>Tb?BBlN?mI;J*5-}9&VrZ6Xzhc`C6aG5wytJwSNR|@ z@+f_mMqRaY5t_*ci;cDHF34&QQGqm1iL`s)hzPR2S_9Bq35FE7oTt)`fv(3E4gcjh zunWBBDN)6=Fha)uPfc5h7~V__6qDxhipI84!V8Bbc((SwsoVAlr!EBn@uqIV>0}tb z3hx8$@3GLf|aKGrfxm0O*E-=eL+pvJzaCyI{dip&ePCr zEj{q3hkH!sb{E)>A``jzpvQK$vw(@$sFnL3dCQ-q;-~RJ3@+k8V5;4@2l>M&wp|I< zlh3}^sd17X779(>nR6EOEiyQGyY^rzmxJx-JBT)mNmu$65?@X-bV&)Ox{mHL{exjT zEqnwvcb=*10aCA%G=X-BMY|Hib*L%|W~ciwWte7Zsl2nB5&fufEduU)1Bu^DzfCxg zhV$ek@kbNQszynZY%nu|kF*;If%m_Wv|Hc=T^&Sv_F`!&^i~Oo74l8oZDZm+tKnqg z$ZoZ*>l4?qL)!HN4`iKJBTzfh#B0eH2H>TSr{?$FpUc%aNAPx=v}B2=Pf1(XYTdjy zcI-@t1%-Tsdv&61ypF~8Vyzeys>XGV-U|WOC+_J01wEJm1vM zo_B0KWN>^F^|30a{$QYA-+`wGc}9nmJ}NG6UEq!JdN_ZFed>_|1H*ExzV0&_XRgbX zJXxRFs6(TMQHliq269(=L&t!jC0XVAeNNOE3Tb_NYC!{&rfb>X$1$3{!2@$$qdYL{ zs{ZxpqI8jN?@`vQg2qV7pMMQeFVwsz!T+uyL$(#gXxDqA&Uy0}KFJR6tpirORHrE6Oi;=XvND^c)Qt~W=R<{A3gz}u9)G#BsE0j z4S2Eg1?d>sKm}ILzl@X#gpsCo`OanJ5Ee+d1{ODVqWaatHPpEf%E;qL$do=7PQSgh zhf&ljX{V{DOXYej5Len`*mLwMZc`V*b5HT^JG1{-I^F?fuwd5aVZI!~?_>-Yv|A1! zLHmRGSdk<~yu0|&$a`vELRlMlbd_j@T~*TwIKR+@Clni++JX$_lTjmgs;GafNx0lN zNumA8$Z5YM=de@6-u}vWU*kQbt)dO3w7m-A-TYI82`;`KF*Wvu=g_e%b+?pJzY*}4 zF2!kO9xeo#@9Y3Ux*@n~Dh`W5PA!T9Id$vsKDw+rxxHVBOa%&&+4DoTeqFU*%&I8Y!h3Xn3}g?LGyC_PgMDI-}d-spn8#sz&yBbiCcV( zcACqcmodn7aSlG=rnH3PZ|L~z^v4pbpsZD5wc~}m`)hEv7)0zeuP?pSjq4qDz1|oR z`nne8K&rg$S**3knlC30Q#hTx1ao)1<*#94$s$zl4sy%Ne&=Qd+3yb?Ok-P<#4xYA z%(HS7XbjT6G0VI?5qI4v`{&KKihN}qi5UA$-h0UYrfC$r%>u5odvs-OM^mPe>V31r z=Pf?MO@pd$gT>x7v(cb~3jM2{x?vkRvU%A=T}HkFF-iBKW9llN*Byh-+cvx6jYsJm z50mfbdt)+1#ZQXdgR7;$XkCEKYI6kq{jE|+Pd6&|ak*X1@R~g3jwMR6TNbCgHc<(d zC{-@2%r+PvU-{EEhG!}|l_fVvnRee{fu~7Y7Yy$BpSs4`c*Ra@&2#DGY10`sH%1Po z@-rS4PhCgdONO2WB)Ql+nR)5bU0U91Um^vN`EHu@T-^!^tdMTVI)>6^I}MW*HjIw1%uAS|Xplv<DKj+roSj@d1arIWQc+yz5of3Nb))1pZzgXEiyH>#t|dQk7h7`RKRN)I42$?b zght-#&z&4-thEt6jMMC3Ny8H#jS6vUnO!orPT9faBG5d2B(IWAJ(%IWU5u#Nw2GXF zleK@>4pNrHvo<27Yt%cnpyXSI7h33m4=SRVZ5wm+N%LW6!BUI9ntSOxy4D*JfSz<} zJ?pxHwPg_*P_EWX0J>f4v{e)mnU0{8Zk8G+feor(nuJRw(@K3d6F=8JKA?$5zR<+v zK)eHurAfPyfLGMEVJ!_|N4>e6bI0;wCqqGMFW3KGTG%th>>ytNUx0=l4)W7!mVXF7 zz~{yzP6ws4_HeW_6vLTCI_BaNC$zWZmr1^Vq4?Am$!3I7F|=ItpE4b^i#5*n44VE?<+p;2fs$J4 z|3`)lz5!n<#j>8AtN`EU8`;jp2iJSZHE#cX0-+!@bkqQh&65`Rp^HJDfLyV0$e1&% z5yIsQ4K)^#M{RWUf5vTb3paen=(tQ(aRUMvl#ScKbfk;|}o3i+Lb1 zE9zRpvvAUt`<8ZlnY+6~VG}3;=~iDt(#wkX`AjKB<2o6O;vJ{ecRbQF`CXRdU~0uA zYwB~`H5 zS;&LGa#POFy9jZHj;A+h*JNS)x12!xq|yUT)!o|KWvQr#Mkh>1Q;RE1X-%qVDH@-mIc2L@Bds)n6*E1t)-a?K_eD z$>r&#MaTrKW^ofq8XpSIIwmW=rjl_g>NGSRr)Z7F7&)dEk4m^0upI@AQA%ZLtO41M zqAFfB%uWm@v#GS^`gDvyorwjau&MHT!>FHWc91Dy>f=L1QqLBmRcqa3%<;DX5~KC6 zEAq}y(TL>$!;NLD37bs}l~eJ=k`IAvn#zL!Kg%T=R~NTW*lepRUwE8lfO368h-p#H zwwZ0Bm27J(Q&!CwBA1A|LzxXle+(M*?9_SY?PFYN;A-QvW{^_>LCgFZxKgLfm?pf5 zF{Spstq+RCRvIthM*lVR7NZ;xadVmYEYG#9-3)%DR`qtsXc%Kv68OZdX+^nMI@> zU5EQtFOLUBoAd2j>Lm$bt3RaJr~|)Is(i&+;F;vDQtO7EfpiFyZEfln!^qJ# zBb-3fig*`>^@m$9$EE7|7HHgcU&`wH7BB<)@gu#W>+kZ_R+_6ALukyLN<6V$x0%qsm_lLM96b_PQz&?ep{+Su>ww3FgNBne<}4su0t*(WT<9q>YE6Ea|~XS`C;w$?#vRS0J6S zEnYLh$7OQtbSr}~0=kEul@l2>d>3O6-aKTBQ!%%?ocfobKiJWW2Rdk&E3#t6>e|Me zF;QcOGLLe>Ec)MPau2s>#Ht{{;Ay=Sd>)-u@yg}yr9;m~tL;fX^N9q$;~}Pb8~$Xf zAMx6e=$ACFgQzly`&~-6mD_OM%v)<>_U5BtHqrL^TZZgpF~(Ka-QSi>s*POQ5UbT% z&lgvFF+m&X%R|}ynZmg<<}G$J%CCz%pnSSRq%3z*JjUP(jN$3o(`gN$_dNG=aS(_v zuoyK4s*>&i+uT$d>K$H0_PR?yhA2~hY16-Y;_8ph5I|?P4T;-CR4f%6s?7s=m4n43 zQx+^7QK(g?xDqqq5GPa z_ifA#M4UD6EXHd?E{U&ze}j?$lD#sCS+a3 znC}_Q;|k2LKU?FgL0&6i6$<~!7_-iM`4gRF*LAP=_$*B!M2)r1|Gv%GPqr%c>JUI0 zh}`RfImy}gEYhVrL2K-Gz$CEl*c5rWVequ&iJ@zU z`Z{e50GqjG*UhR(NkV%2A_(SKl(9$d^Mmr!+<1uTnl5@faJlzJeq4iRR3^vv>xu7O zn%5@x@{GWdh%vFaLLmK%`N>hmf71u1@SYajY`>Nl%)`xnA%Aqv$XpC;ghM-^aP_R7^8Ewy= z%sfZI)~zml-VyZu(=3vsgCd6tfemY{Z7AE}jkF z1pjB3J^J|XViy`*&?1iW#SWSq85>48lWH`)yosGMxiOHi@DB2^|d0sdsy!owfy)x3zDqpV1k!^{jdh~h=-31s) z$k9T|oa*NI269yG)un4FpFi1WN zePkKlYUC6(cQ$5sPtGpM7X5+;+0nC^ya{&U0_ISjL>Z2g`S|&>j4b;p2~9GfOQfpl z_9*I^FC#@(ZefRnMgNmYy>2wm4%kY=v79Sy8BCAfzB94LlD-;4Z$pj>AL*L`#g`uw zZAL-*j3xXVQDy${HCSwbT?8uyR2S%LP%&oHM(;@XNafW1dfinS1r&c>93n=I_adhBKRpXZ))2?DygJoQA5A1%aBFi8KdxbVvc8U+}w@k@h&6j4Z#WK1U4UiF}W^Z~yYvvg=rBo1so&8xN2UVi)v{m+)LAM;eP8Vnen zW+a9Kvr|ynk^NmxT_2@iObkuyGT6!WSqfhf_N_VEJ zH(>lB^&FPR-lv%X*?~VS7QJ3i+0;CD(;#sA{HhW83kC`o&j#l8?MPW5=@hIefuv55 zF7G2WgAm+f#J~8$34d(=o1n?|Uj$7?7FLe`NSKWOpMt}}@}KMfRx}ydnVFdXkKmgq z7h|RD^#+<7ge^jB=o`5H?d?BS|F#5f;QaaRph@`e_WogS{yheEZEv5tINW&h?dZYi zrOZ?)H?N3^lvIt78(7%FBs8@;>Khpw>;s2Ul#N$bHZ}mLsja2|J8@-YOElV6{JkvV zN*M%Xrc$0D$}e>3%IU(l4?n0ZlLfu2SgGew1UF zNB>|SpkKW*FsyHx+z12|1wa4+mS1wgGo-R9aI`bhGbyaT z%HvD<8hz>A zpt7iW06eJo2#rANzMda%eZlrPn3`J~@4u{nUImyU%`Yb=$`*b#pL#766g=Vm*od*A z{oDm;22iLFp0{I_vdGbNOT&YDo z?*IY;n0rsPbAuCyLB1lNe?2dLUEY0RKXny;_)flk9TS|I>)$s6PJE?*=%Dkei=$6{ zt@?B4(#-K`Rsk?JJ2ijrYG9%hxX}n-P4yQ}k4+BsKgX^us4px4 z8&z1H8yMWbrhO5wyjfmbQD5^aB=|h#H;lQ4UddN-kE^Xl~Rly-nxU)Gp(<8`e_5~L3sn3yvU0C z;?kg#`Kce#&GUs}bIbQ4!uzRzu^oWYOMOMQsRK+U^de&WseiB;fYVQYiSEHT^dh4B zssFG6o&J4`?t4f172)24ndn7i54^AinaO*>wg*UC_9eFCjoyXGl6(ADxA^g|uIb;C zH|H;rj-AW)F2oi17rMP)J3w~7p~1P;cPEb?5`XG0-rm|sFCrBN2cS%qz+UO-6+G38 z@9{77R4)`gW}+`58GBlIH>{iJPjTOU>d?vyTX}Ct==u{pge>F-F5ni}oUifTht%yE z)K1qf?Ow-}Pp4mdx_ZYq;vrpgcjY2JoF4KcA3<8uux-SiW8Ixo9l-jxK46Oa4`6>) z)sNpEt>!Oazw+-m+!oU2E`rxG!*9$Kq)Rg}-4wouo%c)cxL@}5uM{lp$bKsp&J^Gm z{l@p|xn13l*vGEvAl%F`fCJD!!hV{=H%(*HuQw3R-U<6!8|QC4GPV5|F4z{0n5rjw zT=_k|azj9|uTvW8R^2Wld%rS!pcT!pCmQ>217Q32ys4}mg6W;q105>8p(%D=<1oH| ze~|AYxbEB(Et+f|9~|};BD(7*A5I^GmhUVX`{^z|+zjc9tZxgdmx+xpKfV6zF2WZY zgXb4uzdFya-yZS($22C+IP}pIeBf`#gS1w5pXtHF>m^7J{4e&RMXJF|_h_O1hgca#5?LBNILP%ACb!C3dc!~~KmUJC;CYM_c= z2+XH2$$k9hc7|$GG9(@{6$-K>XTn`t`A0Vq{safm^bn{No4yx}&qOLP!z}MPQ$vC# zE~P-df^;*zCysA|=XSTZ*dKPHg#=sUFv(BHrmnVFfl|o@Rv$_)JElPE18Nn4eTuP$ zYv)AxrG}0%=TTQR%wcRShQ+~0qZzin0Ez^ zB~MkD_dwUkhdD?elq4Y8Ey}iO1%TC>At3wl_C$r2672L?NievZfI}uHLN7=1a+NTX z8-KBPw1v`oo%Jmdu|Pnr#RXZGa?##V*vaI47TQ}=m`Xg=G^65Jx`h|c zoT_^=XKYh82qX$(H}et{aej1$G&R31;txHDQHSdG9jVS`mNZq?JE^i9!?IS z^X!z6C+nr_AU7H`%=~*Re?)3ARtXA`0`v{aIj3We2LNp@_TW)7 zt1OYsGP7`Fw8Iy}je^d@4Mb$n8hGcgak(+Car`;c_QSSfh%&Qg>9@MO&cAA9IiVN4 zMySPj9~9m!wwi&M=MUo+<>9`fZ%bEsj|Hv;`LL9agov_F!X2?ddm)Bxb@eqO#>3)U z@gXuzB@x5ihu5Sh5K+NNVZiLGOXv)^ zEVWF@^LP`+FmTtP33D=>%rXQ4IOf!JX4=NZr9#Y! zZJ-7)7+hDnr|PxVIc7D+ra2&R-;kM+!m>JmrEMx}zFZY2_Io^q91n<7R`K}+=%t=X z@6Rt3EBPug!YJzUkKG21OO6J~jg1_*)J}XTUGPNn%mo7*Z4d}aik0;Qnc8lvRF)aC zd`z296k^;>ajFcd8W7X-HYeIT3EJE|C)I^08z!1M0tPe*LZ#-gs)$|dKKK5Sgsj`(| z3(SlEjw~{deSt)y$XZ>5N4T5?Qe&8Ji3=-Fvq=7F5hk5%yB#yP@Hg#rRw3Js)Tu?$ z#MTUB3ck>Z;Nn1pA~MLLP|OBNb0mD8*02_%E||}H5-pWcJ-N==i4xY+R#sjuYxIKI zH_qdN^p+y2gmGerMi5v3i133ymM)$KFm<`(*gI4CbL8-V*5EhS!&{bpZZPX1luVDk zLsgSVa~-#QE9|6n>d{GAA*ygJL&StBdj9unxu+cFGIQ-4!_igy3_fy#h%y>%jL1A7 z`%=aaP?xO8xBr(8Oq^R+P{X#uz3`zvCh z-9rhvY-g}G9PEwUZck-$K2FJ1J0HIR~`ntZjfQANR#dRAJ4$=k0yi|9N6?xr!t*p z{oz4dC(D%f3M7=ui&~GjFl^K+{hb4<$w-Q&z((fH2m!ILJ2{h(p4wq#NrUDFr z2PAsk7?O>AlMz$=8M3M@KIgQ~6s5)#D3&KqzFLmWnbdwnu4E`~PX~$NN)m64OFv27 ziL#Tr5_Odhw+jwbNi?`*5W$idfCDw{dc7)b1sc;%JmeUUz_*O-ahg96B;~2z?m(6= z11|{%&_qPaY*R~+NF};UhwsB1)z^ctz<;q0pwEf*OT3QGXlDQyXYqY@ybz2dId|!Z z?Z5>k;y37g{%yU1Kj(rfG{jMNfu!(dlShawNN=9V@I~?SI~A)oxM_ps9LsEwipjWK zglnkq5|*DbD6=p>?t6`)Z~UI++!JiSyzmPWCUTYRQ)X72ShdJ4nWRz_JErvH6vCQ3 z;jX+}R%#+^j$U~dX&mIDj10kGATbTow@!gns7=?-h)l-W^6J&yR(y?G}sE!*JDR={7C@nO{IyYrgmt=euvsy4id zP4TvGNaV!@Ypb}+a#nddP1q$cpbCjP+2)UKe{&M=$sScMSy04q3;FFj6tRZ+LXDbFVmZr;@_u`a0srRWw>cL68 z$cNs~gz|1263QP~xOFvBLzM%UYP@`|V(;`Vh^KU|7jNK|LZb$1Uu)D#)J|!Qffhr4 zEwy$#truFStJkXr9EDph8~`0#U4M>rVZlHn`5R&5O8&8{Ko1$}n*4-Rd;Zq~0fYXe z8qLTka`1XkTgp9+PwCG9CC5jGB4In*I?OjSKmmtH^!l)BhF675L~-V+jiE)%A57@w zvQmGgc^1q88$i$o)MG<`hsBQQ`PGWb*5%4a?ZC0ojt%Ju!8(gHLABnWNUxRN!wi~P zhxN{c)#)hNaZW!asNZ^gN%M38tiP)B-MqEKe#sE8H{BlvS&=}?)I4ZVUX~;);-9zA zwc;@h&?sfAbd2+LPF*diU59GaJ6O%G^~&8jZt_nD_cvMZFE(tZtd+t%0J5n&KIUJt zL^|(URnx3r0Vi<#G~2&~-f((jy?p>I2HEYND|*jcXT_yIw(;0H5p(5S+f)9Ho#bc5 z2QyPFKa|~=Ce=FsBM@7_1A7DqSE7NhEi1`GEVXmmD#$9$K4w|>cgIaleWCGMqQVS* z%yh5`KlYmxqtPbL>f1Unvju+s47rHUE-Ix-@Ku#Tjd~9m)>9n7Ubpa>8iwzTOA$Nu zsF{pSl2oS_!FvM_=kzUz5SmUTq~l1bNxLA+M5}1R1$oFfNqJ#CP9%2hp4F3q&B8?l zK9D5}bN8Y~c>MIw&=Ebc{3GFdEzx#Hy>@_vM zWDy7nUvfd1g9y2Fv<=AF6L1PaR+tkFK_c==6$+V4gri$=?7$gB<8@}_Q97$q(Kax{ zxdgt=j0Iyg@0N5UE%DNMlL6(thXQwUqs(v{qR*vEPXwdO5Q-Voi=pTHMsNWmIBMZ9 zDhy1rX7TTkY7~{bTbFbeK>M^MDtQ&oqYmE&dp24~rAsw|6(a+aL}TKePzxDK5pp?> zX9BGU`q{~{vM550K6OZ!zp?+#eg39iw{Tz(mTJO)w`I+gz7HlxzGIx zlKH~jtXRv>Pi6Zll8qFpEy)3&y_7UAw9CdNH7F26_UAY3;}-XXDks&cWy8#9iOj^ zm2PV=wnJgYxkiv#m4m-`dQq4ZrN z-UfGmDUMYEO@)Pqp2-MEGMSHjoF}zwwF9$>A(jHCzgOa_66^#5)NZ8Imt1Noym>ZW zH6zzpICm~M5ERfs{PJB*!1khEo2FJP6?|}rAh6@kxD}1G+(njp51z0a@+ml#k%i@#;)P+78f3?mef}Wv(MG})>~r>dQ4)&LNbPolUnqE zFd4po$*LeupBi!6^YON8;17e%4J;za^C*hN{gM5PM zEWEy@xQPYNw}K0`s%!gEYE1OPi1l$-z4Y%@EsB1#^qea9WL&!s00p8&92qtK3D72v z$KzA+XC;L{Ldn=Lt=T8~R#yi8SKbEQxHNC&?2bLQEWVPnDnUET)mFenF%WzjM66ue zdNDOx%f3==9;&vxPPZ2A%ssz_c$t|zhSkkRDO8fDO>hom{@J*~Fq?~5q8bVOT{iit zd=Y_b(<99bbs6lWe6E1Sh7SmXPTB)yvnCz(4BhYBJkHrEz>qU6J``#9;rw#XgHpI` zF{A&YFPC$4{(ddjUdbTfTwZNG5cEX#g8wob4rCZmC$7wg`q|Q)hUei-9Hxe`3m_{o zGPjoJzh2OBxG}M?(is|vd>Z8AQaf1#K(#jPBl~4bhnmRs^a07>d>%-S$*?WtNL^92lrqIuD$B|rDT|h9Ylz&97892 z)qZMStKlQ)o#ogvS}37r849Q6bNXzZ17U#fO*N{SEBhI+dQYj9-h)2+9XFt0lq1aN z*zbrk`Z7wjjngi0qfkoPNKVaDCJ`5gd?rFhrZX)>5ysya4XYru_oShnZF$08*%TU1 zd@q5&k|+ha*-o?X=50TwxqI(rLftPBrO>)SlWB^7eZa&MKR}a|xi#vknNNj?z1nop z9E7wmu@Ng9!wf}fjQ|YpXYFsrniw$#pS_0U_!fGu#ZFY>6UBjLo%#}+w&a@&SWt$lx@hT9ZQ-E%UxVCOPI=` ze^*#}S9`xXkndIsq&yS-154eYkE}4*LL*}O;}7b-qZZAUKe=IoTGJXyY?TyS_9`7g zhgUHm^Ir{+zgkXT)c1WF0W;)YQqH-pY2}v}(q25eHy$((IDDfyD@mh*a&W0Hs=Dh& z6UU=hBM@nQAxQ|Ti=U_z0{60`NA-9%4;uPz7?Imr9aeEESWP7dr)v;clJkCaxrm#o zGH^K4)-$=nhdavuL$3;xyu*>;xz5LZxjV0O0P17zhceeaD=L_B8UdpiM6T!>&gM{I zV!Z?vQx8`(f|M2ZD*iGRS2Jn)CIbcJFU zxVpi>d~w`(5y0z(C1=z{$P6k&b8rmoZ!ST~u`2h4Tz}6pRSUq}8hJaq-Xt{P)n?mM zc&77@PcR5|+=GxB#fTKO15Ph*kFjB~UPu@i-7wj|w7o4TbD@G~>r{4guUC!7kIPdO z?LQdPZPR}k*>B0#^a&DQpK7MANtoh!o+nPn`ZG~b7lz{ahZ&xB=0H9S`x*LrMy$Pl zM4gr1^UzO`p@U(o@@0LSr%m$14%v5amfr2;J}tzr_Lh&#)X>1wF&MlP*cKnBwJ?fl zb>;=vWI!F7EbmrRW#r1Z^oXcqRq=C&>hd8_;^J2~tx6KRhy^%YNo}V@y?owHCvR7t zeotzj0J$LKuqJ)pku`H4H2^s^d-*|J)4sSq_t+!-xS7eg9wn^d+ZkPLY=1y*O!-@9 z$4hid8Si)Pa3^r|#CA{GQ*9}cR|#Z=1jgHV`)59g#$XYz@6}Ew=p}C}_){CaI--vG zH9L%SyHJ0LzpCl5^bPG|7clgE~rLx66&>1STx_!EFumxBcOF5fI zuK|eX=@`+2P9Pip_7SLvbKhBOP**19+5Y%%aRzgK@(ch27p4Ysw!-rd4~r&OMVD@J zPt2rG{V=^e8dMD-bE!LVZR`oghW@ZVuLik~7ny%Ys6hb=96B(dkK3TKkLZf@eBNW?cVw$%>StjfE- z8*xc}x!hcCP|s4`NF+SSGe?zBwpzic+cRnE%Ba8HRSyU2M_9W+xX8mTBi96iHa3M_ zC{~3qbWNpnkL2d^+V*fW9%LvCaWZM}LguIP3#><;NT`f~W;uM0Mj3n>K+yAj}M28{z2hK%~ z%?Km!jTpL-Or<5u%SXlciMV}$1z`EOB}olOO`XWafwOL(KxPtU!JCDVB?iaxX&N%} zI$-}d);(y<-YboIb_}*!hZ7xd%q_#>vM?Mzrc**~v5GfXv8z7fWW9w-pG#1HOl)wg8f=tyI?7|2+zwd!Q~0~OPanW760wQlQ9 zG5j=A^XNEd9Jrh{fxvS#JJvMRI-E($$w*4litMehCL-mcp^bhR^G=VI~B7@{p+=id&Y|ZKq96`5&SvHvxN!9h1 zq%<$##vqg;_Z(xY=|jwxiPqB53GrGzj4ZW84)n4G_E0N}Q%*tpFs&60K0%FgyR)Ze zy%_9YH;n_5blIF=kBi!mACUVd<;FcfVE%pXnUGu92oFi0d`L%}YC46RFh}2}z0>XE zUX&04olDQD!7khd^tCol6m*LRA2B1No&SbPpm>IZIkjSSORUdYnh7V=wCRW+-kn(b z#M;gcd;8sr!WTGobe@+%&i=qW%(Sv7J$ns(d~BQ-$O~;~4bYrQ6A_-3kCS&?ZMYxv zuj_7&i-=rTOK;A{TVU3A2T!fkCXg~$i)W5?|5>qHIxI!^5ZH=O=dz+cw2gXT?;+$H zjFCBm**MV)%{fGH_ur^%HbZdyw?S$;4A(4ig_L1WPXK**)?#OD*_qlJzN@`ooQtd4=yB?=kolmYZ`zr>?c7cuC!vjMF6WDaqBZTapRmQVi(0`5^zpLzT@ zd^S7`!Z9mraqM}?pH8`J5_wK^7O(Ua&+RXZU)5AvTreicLpGh>0GoIRkRPyDvAlHV zjh&?ZpPzaACR!!R?mRv>@HFA4P6&HRGQ-q;OUeOWDP5ZPq*}I2>IeZfM&nxS9g+l; zGCHg+2okD<#zznB46rbVC!#l_!bf>B!{{DORE0sT-mW0W9Qr681)nCBjG->LW0>xA zUyltB8;gw=Wb~2q$_{>`@0{RUCnU>aZ)uJ1*ocS(fIoh3p7zSE`Xx%bl^L|r*!nC> z_S6qz3e>V@aAhy2o;tcmU%iG@%IH_L`q10OwO&B;G)S?vx>;}ZkQM1Vz}pCx`sAg7 zE!I^~Y^G{2r3GH9Ub$z)XT+u#n3(rsjVH|FLc*2YUO`D@o_uf@E=-;W@aO?^0Ar7S zw|S%arlEh*F&B)K^!b)GP10RL(rnPSHz??k&;^qhRj9=3T0Cp-oVBo0%e7}u|F-29rtSAHYd(F$|L&Vz}qn8$_J z6%i|)clREZk-aB1&+{afEiLcwCatW3ga54T0(CCcz_yVaVF^t|HDtm4=4toLj|2Es^CJWrzV&gq{fBe{?5<}#P)Ih0k{^l<$dde$fgFg$S#hx`GQ>#vQoxj!bGS-&T zC#tl@H@mIldBYXB4f;@LZFQx+^ryKUixzmY$kjt52H4v`bU@7u8yM=ff?kYRY~wS0 za_gan!i;0S4`(>YVwQ+A`aVM)xe^EaTai%04MU(%HBw^2_ef!4uh#2W!m> zj*mugQDP0!l>veOl1kc;#ok=*9)Y325^F`I_iwyq%%Hfk_T*WV6TZ@fs}|h?L}BSC z#|?RYE$;N9C|jId4&F<`SJRop!pBdMIaA>AfjmCrduR-7ZSSrQj`*t{VDqR*bee@8 zlPyTMrhWLfCTKw(Icjr7V%8+DE(bQDVL2sd(@KO45=$dq>~x}oSn#x(!vtW=5y!?R zS{|7!Nr2~(OBiK93`8skGvL2e*dCyV%5OEV{LW|ItgKJzs% zenR)e_aifO2Eb(T@$yN_b7xiJQ=21Ie{$4c46=Mjcbiq`3yH2wxWK?WUm~fL7@I8F zHEFrxif9CUYEGRRL)lDz6ZHf`a+3&}rlV;THS^BVm_}1-MnKf708O_GsXN~)tnxbY zgvz>$ym2NmmzX^02NHMBsHuL2=Vy>1EY<$K7pK8%U>b^htR%QSOUgPF^9Vav;>E%{ z9S?LYKUctr&dN3P%k*IzHQNp!{LQuO7FT@F>PJ+If>Zb@XcBQpM%}&!-RQGH?Jj`X zXqUTcanJn}5tA}?V~jW({3w0e?MS1?{?jytP_)Dd<@@B#Sxqg~G3v%arqmst$04|HTV z`X`Xj_-vN(LaF3w9nKCWyL@C%ey;9G6YSo3xUe85E;h{?c?5=BzA#I9>f4UQybH^l zVm>F-DQZ=fS!9jRgT`i8z@1Wnbyl{t@?i5GXcPg4Vt9(P7N8R*%90;HX_hrE26S>#t zl~xgX5bFx;0V!ja>+ajX@dUS9Ov>u2SxaJjE(Kgu^E}(XdmSV}o=g$%Kj0s2MCe5a zwBSv>X32c7Zo??`UmXG$bS?*;dsr->I_8h94>(V2O3+x~V4i`9&&E5Db^;g`R46`$ zw>fMv>ETzD1ohXIY~?q{YK#u6i>^BoGxduUNvlCRvFnqHqu9gvg!psJD_Vcl;$h6f^%7dAoTVSNJHN`lrTVjIVm`UugJ!$;Do~|bDGGU|LHN{sa zfcy%gD{qSel9u;`L~aMTxm? zsQ9EL?GO`W}Ne z+3PynSK%fHKU1Zi7XGv{D|p;p`;lqrnxGZQi}r}qJ`is_c^*7N=b|ePWQq5cL=Ri} zH^Tw&-ETsy$-ceJnl}UtAW|U$EUjjPK+LOb{-tVNib%a2Ngp06$*BPRDn5a;^V|2} zPyoQ2#$Ur$l(g{$_nL2k$%JPMp15byD+u%5Y!lNKNvYGfS@uwRy{Z3oMB>)7SbwIy z<&7CnFX!aQ{%cUrnc6xMJ#vBUS*`%SOaimbUB<1BEeG?q*pj_(`NNqO+jpk*ii_!7h6NvZIWJID2sNOC z_BQ%*-pcfW=`kNp`)`8MrPE0_15OS&=}||}V&glkG1>UYd9KYG)_Gfug*CeSygrTI zXAHQoAIK)}SF`W1d}ynjG6*(0;Y3)4+hr>EO-S)Q)oA7?cP%J5z7bREiODsK&-0E5 zs5Z;7Z?wdcz9qTzb}grkWzVfEv*Xadf_Fs8qZ_WOvqGq=Mkf9BqZ(68Bqm-@>D4yi z6EPbY@w;+jL-n>AG^i3my&gfBjzYYo+7bz9V_+{HN?hw>G!F1XZGMZbPw^g_>wjiM zbow<51Os|W35a}k$0LXlfl-|oj$5SnmztMJ$K(+1b*$6u9~(-7^M$jVCKM@BJ_Ts? zGB21ILUC-D^)DMd%K9qx>gD5%pc;VigQ-%15d{+p5LGynsg@8r-RfG1BIh%?mEqsHrfqAx=MahJ+kSW8p`03+?a3a&PG7!x~`hQT}woRMXCODo8Nf z;MjbPN#Vt7)R6khH`z%h^!1 zKznq9bJO|c1Ixaux&7YDOJwQ~u%HJjCQfC~JF zfDGhXu;=`z|1J@QM=soUOFo9$$1DTWwk!DBm2y|poFfkPzMfWf0u*l7|G_0UvbwCN zB`S%%=zDdOl(jF=Q5oGSUm=thcQN>sCrj{!@H*fyvbbqb#4XRj^mXK*b8?rmQFqsw zJWAMkYs{NJp6|K}t$j}VHFae-(V?~r3t&N^Ck477!O-vdCfC5r(xL=ZUTlFUdD-#! zXCZG+srQ79Uiiew8g>Yfej(=>(9!sNk~92X{9vUNcg)N8>}L-1VxSiXo#<~WlV^SGUnTFmw)mpnP_qiR zDu1#lwpdndz~hVdo}U(vT9UZP^tpL#9{S~At{l>E>dmAx%C;UD{GlSb^}|8H=VCp4 zJB~S$S`)FbNn5xRTr9tanjSMih%j@q=l8Ghyrk>TB}DbB4jPWGJ;x>$EfxlbC}t#+ z!|fhAb-|8taV+VY#BZ9)b5C?Czol}J&-#Y-#0kj3E`19WAzmQPTs`dz9dsSwQ7DyK zHq7pD5%^D){d{B-2LNw&f|p+({U1@^efl-KCfjp-JW$9XOayYiZuU%1|~_w6p(9K||4*UyE~PaJA@ zZu5(a;6vq%YNm;ce(e!s>fH*pf<{o_-b<75#~k3uw2pVcMA$**r}U8<;~1lS^Q-=+W}n>;5&lIABhWaIq0;LSb|Sy zW*qhVp@J|)R_k8ZNN^Bq!k6>`d+jMhaFWH*;3z3E)gNJ^Y+FjviJTXwAIUgB4!7CJ zVo@2=ZS@wFTl8Kqi&^-)R)A5 zAuMhT;fANb4Pg)A`d{69oA2d~Ctu!D!UDCJ*5xlY0XO9M2)Es}tA*pUbT}5+I=C`u zn?UfDIk)UH_Vq2}Pb1{+hPwQK;?fyTdrs86-(5C2&+$wtWA)X}xZ11(O8t%40Db#i zQ^Pt@IG^+g`Sfi)Q&Vml`Z+X}6Xy30&$Wvj4-9F(={B)}o(bPz6$p-CQ;T&#MfHed zdIuM3H^@+$>qu!)b}~+qeT>5)O*#SFlkdV%loWqlpKKcvDu6^k?DD^umjNao-myevcUmr`hf_1j-7#T_pt^qrJ9VX?tX$R@ir znL!EoG!;PHc`-uE()EumjS+I7c?^}3S6e~k2OcHYmi}LMw4vr^nG9xWM3llR!mOz~8S0Tf^H+^LLX$26AU?Fiedw;>|Agq3lhK+oodwAJ00XaO(vj7(+ zrZq-W(v)z#JpYpVOS;yD&u&0+vd*lBffU}zH4!S&?G0|tNbM@t*mczVcs|=sXJvqG zgDnM-rP)MLUF!@7iHotxllO8_`^gpxQpNJ{ohuLzM%Li><#H5lXNoOkWyceWna-*% zZJDOVf!g^++&6F_AJue381Gj^UeU#pyQwo#bXg7DrF2_}y+C# zk2SI5vGw?tq{tpOOQ^mdzi_ckCd(J)?l0M;U=bOUfauWEyBqY*?7GB2++BZxT$`Av zb@c(EL?|${Q1$Y8b5G}vtQYaki@5~3{#@lPZ&weTTOfXJv4g31<~AXTQ}}lqDvZ8o zigm|p<$blG7kpI{gu>g-g*>zYd;?bS9N?&f(!{qBSCbB$?Fye^bqIe>MwLo70Q<#} z8kF);mZO|x8^~fo0!^Ryb+;3cwf(nSTP1jfS$u)DB0qVm0VWaWPMJ0JKrgs1Kv8Xp2gVWwH)En{zAF zIDl88?8A4^RAwBY9;($Yq+{0b$1QnemQAhe)Lx4lhMu2cu<@~Cl*gA;3w5P% zr!_oSOD|vbRBgZU9}4dpOisDm$b%v1l`{e*V8wf0Uf!7w&lAE6-Yr#f9MuXw=~3$o zrN8j$@1f$_+}H(PQ|erf?T5}`jb+jey9Q9{s~qpCx&s?fwv03uq6d@#`xI_?vmf*mmxw$-M zLaL{=u)_`&Iok88=Q1NJDRxkgqa9}-50@uJ)R?v1=L!clRB1oaks1Wrc*+vxQ!Q%` zc1&OOmEt%%6bo&%Q(KYt46)WD$8I5JcXL?9um{_EuKdaYOAb;myaR%16-cJI0%}lp zwnsS0eju)GE@{}g4(N<{Vh%}Jtc1D1*xMJYLNd*bMl?`>1Vu4oS#1%BzstoI=*KJ$ESzFc7=J1HUlaeam1J6m zGZ0e45#CEsZ`)}3v{}u;jQIFbjFo^q>QsHY44(}LD5q$6q#U?wX!^^Eo}uh3=s9i;WZ^bgXl~x#ER? z3)M(FN^8;~g@k^Jk}%&2<>n;feIh?buuU|D_?WRl+Q_`*7277G(+${ay(P$bvy_4tT9y?S~3 ztoe#Ckz7_DeO{MTOVRh6IgClEVr0>zldOA^;V|Om+=f5WMnu%K>%|8+tIZc=el*Zn zCZZ$Op{E*}SKGiYLoN*9(?K6t%uUQ6W2GkGZjyYl2v`A>hz51#?~6umRL=#9s1==# zDCBIIpqgd?E|7I zc~3gSyY=@UN39_jqm<(u+4M~Mp&{pKV}c=Xs+i{W1LQRWrJ7Mg4N{&dyC&!CULAP? z;1}wsS6lk)nk23KsE9cO=`EPZ%u(jdp$!);drj zmqKLJ{*CHfH;()h0xjsV_J(&FlBkG? ztXM?IKIAnU*r!F*X(mZpH_QWFi=JYbEQqPhdF}WOjjtzmQO%ahbe1^3PW}m!EA-#> zA?>vgaI6S%83RYsfyuPSEG@o&xx{GOibLd;-=H|XRx|zgxp5kXW1IG4*B)-3SQID5 zT^}?zS;Gc0za)|UGPUR{25wP0e{%c1#LEtlN)#eJtRkUU^c~3f^l0)5)ciF~m;DsP zfGHB!WEc@Qu%iul*PzX!o>8ABZ9g7DkkyrW@LE;UZ@ewA* zZukQ{4csPlAFqBfQBS>9r;?_d{yODRO6QIz&ePCvTKFaZtAVc%$Gec8 z!G=p72d{;B5e7JNxlN5?;7H+7z&28|xe^{tC%qGtH-WQ$g3vp~X+%3!W?eVaI(@tU zmd#>))*?w-JXE(zluUCe#{=Qfx>XMbx>Ha+mHU*r!h* zm4;!u5eqT;HxkA%2B1p(d{Hx?j_JP_XQSqvM8$$o0feJ6bZwi5Z^6~iLU_!GNhe6&z+98 za?G+_XL$9|;cFG0&Q-vM4xs%6QNjR~?Oc3soFSvA}bF%M6qEnI>Z56sd3Oa4=o zUW&|8q<1;t&aQiWe9D~oOCqhNeZwryWZjNkv2)kcToSxLVrbd>HKwB<+}T(x zqIJ|X?qA=>E^&Hb$f8XFPF=-1ajyJq8p;HmK1`v7%)8&w z^>I#55oS5rXzC&<)NF3d%XqE&Fl>6}5oF8H*_DzC!ISqMzQJf>(ElWl3+eE1^O9Iz z6Bs`Ha{%7Ms<>Q&{j1l4@bx5m7i~4qFF_^4673%ei&BkaV0>kmAr@b263qhw%l<7( zq^-OD7xU%zM>?ak3hg-Y{`x!MFh_hR&i8@$-+f7DOpNPRz`F%rq^M}zVKk0UvVVIf zy(n}(Hxn@@k5tl&{>LKlM7g_nlPnWuJVlwfa8g0z#Anu38*r#6fN;qv%lRJ{qdF%> z4P$yiMb&s6sPSJ)lb6Ut&~A2~V}+xSg`?@>^yvc-+4-4nbd1dFqvkULC@T`>S@`S_ zaY1xRe?`bxyWT)gwR1#EKIL4U@n-xo1Ej%HaD3>v_qf9q-;0a%ee+^2diDdnNHzC5 zYK~Cetjh>?xAUwqi&GC~MX@hJKFMVR8Q6M*fLt4$aCCunxL~U1FN7Ul*gVg6+eDg5U!znhR#f8?5QPrhh(x8ldgB_y@)Bqburd0`MbbDh)L4Yi5=`-icux|e$Kv)v??ysLCrtYo$k{F=VvoE`tHUFVD!Zsyqg zrlZAME#^7&9dFHrD-T`XRIcobu;8rZxfo*ng`Vf1EqXdXVc*ZERNqbS7mJ_Gdx$HH zQ5^TYpV{8RVX`%H3Pq^#X%qbw)Q~)tAA^C{Bc3vhQ@(;H1H_#(O>QnmBK&jN! z%{zVbk}IV(!`$$)pM>ZN%-7{BxuoMWFk2v}g2F1u5BR4moajkD<{f7C<2cnCdlHHE zU~%&rn;bWK^XcLGpbAkXcQ($iP03{4d=4-SeB3!YTOkC4MW-$#FEZ(o7AhLcNMl8f zxYi%$2EzIdEtJ>JOWa&uu`AnSJr4N4;8~e1RIRA9CbACwMfF&mdBR;s#Mu1Ix-mD> z6<{spfcN%3h{-4BYPX$xLs1+(C|do*bh_r))}+V_e*VoKIKhOBvX*Bj{e)qsmHc(C zVp&walyGf8Fb#T#p-~%_quH=j$`|bGR;z&HZ-u&KKFgRK{8^>hiOsz{VW zT^&H_ab@SfttEeao*!^eR-Wv4d{Nf$x!`>vNW#qIVm4DUemrD;9BCy1s~VdZ6Pq*t zRYZIQX!D~Q-lk1O&sj*q5jOTI8qX7i+y+;>UNpMn1nX-hvW`3FG>Ln4tSK&3Cnp-b zgiqUWUL|fym045tG|bmC(Is(>3+gvq;;8TS6LOaL-(h6>b9+$J>2|CqNhMP6EuPmH3B4apwrmg6Od9-4 zafe|c71)Ox7gyIyVFSs*ij%&LX+bbTfcXk-)QM~JLj`f5&b&Gi5A_DRzwDPWi(jy~ zrTq_>b57p$z@FS-G*Ss>)YF&pCUi1+p-(o&>xWo zxg1$*B|?%o3mQtiar8vjM`#C~%?`WL+YFiu+?OpzK`)u}3HMr!4SNk2ivSkAmmVn85_M5K1J=HJOB&W)W>}weiG#QBxb%T8z`^0 zwvvhQh?%jzarQSsp027k2NPaK5e_&&kITEvF1(QefvK%tENBOwDp)zDvI9jKbmF6n zv2zai-?Qc3SX=mpNjUoPb0qv>L#MObo=(y5Af)hWj*Xn(#gdz4@R`K~4Qz7~r9{zH zPu$;>@2IlM;<($cCga-pFAD_&yjvISHZztLD$r3ZX4fa6l%Nx@Ax6gIVNwb;n>sEl z7euw!@&%!v^L9eygy6|bSS^-lK|q}?sE88x_vm;?0vJ=$rp z4@hJURL8i*kPS)f0XaZ5>>(>XBO!?RA}q6}(y<;+UnL5jsl`0UP`ud~!he9CapuEp z_fW~=Ycp(xEtT!sJFG(IsMCB?=LzDRmHxrAA`=8ZDpER1={4P*97r(04lm4>e6YFa z1rCR+V=6@Q71jBizD?BcgwaV5GpM^hxmB#MW*PhE@z=ObkW6XcHK|Tte_NsKizS__ z_;IyO7f%gSrp^K=)oeN*-IbprQ++0AM2#VEX&BRN3b`wmD>_jPEc5({`R<4VXZ@|x z_8B+BHTo*AfHm=qu_BqNM-1nKphTegQxzyda#JgZ4xEZ0w_`Jy>E~#niC#(@yb}eu z31k9a26P$V&UDV@cLb)~T@4)pARDzF)9-bbAQoH&0tONf@@Ci}D;kS;EcipsA6ojt zPWm%#uAhzVY5_WHUSHi^G-vw{oWjrW zx@QC$6-6MM+ZGs$>_R~RGddq`>`=XWpOmQ1ZZ+^E2lqXg$}N=!02vdsE6FRDT4h;O z)~nOl-wM8=z1(S}v!opt)~}t}i;i zZ|Mt!4XsO1um2d%2$vb!l|S2Wd4*;nWsrzcD~Vg6fBP-)J9D$evGck4VdcX z*i4HMZ2=9gnj=syc-%6uYn0u+bqqX@3cNpRGS$HE$+Y=hZRsprBPF@J-Z(_6(%R%6 ze6wwN%rto);*r%mD|-ZtErS(0StsiaW`G`9@~JlHr}_*hkwvET2y#Ss^LQFx9l!ZT zex4<~>~sPYUABW-*u~fo3Z*|T(mJ7MiebbUl*D#Mz{XGxg^#!;kM*E%Qo zkyEg7o|6Xv!h8-dR02&%cy*EcPC7|QB9BCcFXa6MrayGG=srE4ppbcI_B66ktup1$ z-K=A0TvWm)=VsZb6U{VT7v*S7{UkIJ%kYz@g4@$SQjIba?MyXxyUv#T-~E9eZsM*K z?^-YqU%-v5m2u9LWZJn%w+P0j?l8{|1kHtQASj_5ueqSm)6o!U$MZ<)ViFq@(tzKH zhjWcObrC`MR5v+=90jRmpvzJDmLdC=E!a%{s(nMBej&Hp8mlzK?lNc{j?+2e5{F{- zRQ(Xylo~1Eod7h{7cu;>wK#-$jhO=6Ah$rmP8th!Z~+g+zRigNUmvTL!-#T>9Pet{ zg+GtUpZAsWEsHiqH1lwQ?y`Z{#8$#{-p=)P7nwZKn}Ut3@%E5@mf2x*BMXfa5X9ph$c_y6zr_)ppchYx`Xt z_yQlr-)wuYnK?DAN5~3ipVp$w0V`uLU-K^_tUWp}Mr10i1b@j)Ej{XFqk68K&UchG0S>0$#L)u5#@HK7xC3q`qXDG15XI5TxgN3O zo*OeH;ySfv!DLJcgW=_Nz)AiWdWiFXpobV)nEszV zCy-fPxb6{)1t84!g%fZzk8sh)hz@~80GXVSn1})_&<%j#Ute`z2RqG}8^1TY$Iu$OFCbq` z4iHHT{NEu7)#DfcXDOry33z=C1OfjoJIJF}Vq^?phKUB_5Y#b%So9Yk8$>wQx3>A~ zG5iJ4CQ$P;0@$xx>&v~ahjJDp+E4di@^8~v-dvs(8`HR#pZH^{tR(sY3>-XU0x-m+ zBoL61k%1zhrGr;OHDFW=}2OflfH>VLMQR_ee z*y{`NgQtK5G2R>gopbuTdHhrUsH6C+n*ZY_mc{4zZp*o&AN(^0_w3(K=ZC3w@||C& zWDqbi2YlF5p&r*SvIcMr{qXp1Qw4_|lna6y1g)|#Y}g{|-||0TNG6-cKeT}n4$}Qo zn$d3`sq+jPLO?JkfqK0P9q@lJ{I>mPV1AzWWkt|0ai%{+Mn?$> z;OQO2P`s{ZtezMkz<8~~CH(Clg-Jm5PoS`Y4!|a60DxjK>sj4+q!_RefgV4d9s*9# zKs$`}cQg<{5gua@{Docsir^CRP4EZ%TO7*(-`@UO=64mEZLDvtj~)PG7$1mCH(U7M z3DHijxSP&$Mg2QuA3m{;_FbfO>qT6O$kQ}Gq{|BVw*n$BEBY|e z{e#i=-)>cR{ff*D5KFAsU+HCn!@>Mok~$rNn>k6-RPb|AyX>VrC7DffZ0mb#x22kM zn&Kia($*Xw%HO$=?Ow_TxCemLTsa4m<~+RWW7n%A{c+LjN;rlVw&q@AvD?Rd?X8IMSG?S5J!)>jH~RGhipq2XRG^j0>e_C%pm( zPi3OH6)NULwhSrW0#B7k>|)iPc_i4TG8m~(rVh-?U@i6zzU`}}fYdDI@mn=s6PL|N zravmFxs7CuGis&xaLJ_ZRkSl(_Dh8tkluWh{2UyC&AHaEB|{JmHGkW}n68x^K_@<; z;h$mR!{CFVTx5E?ZYa?A?0#2SrP8Q_+ZOyjj`?wF{CHZ~ zPd(LDmSy--PU;%&5HF*+L~Wdn%aY;5Y7Xa@uQw&V3(4HBf;N1%>w6PN2wY zby^a5Pfj=H*6*h5-t2vxMRsV52{*o@ZTk4a1Y% ziq^yxYZ#~|?muH+Ut4Z!dNh{@U^T%FY;9|ka)sLEQ8tNY;!)Kxtg+Aj;r3kNZF+T< zH3C-VAc>KT@AAwGa5oQb?(YnsEyNS_v%Q*_yC%#t3ljWvywi5FnFG8(vq`-fpnVk| zNxypGe5|z&dTA;ZFEGL}c@f}I3=v~k-2_bR}rkU6IyEhPg z;M9#kp{qxo{Hhz^8EbwG)GrWwL|?$v&??j|Sy*z$3!HfmKBk}UwxacAD(Z0Nnbn*& zc8c{JUu#ilIGYPESLpkX$@A4faAMOIT%E@7`!YkyJ+N{rZA+AP zm756H`Ws9-xI9H`SJ`B+4I^wkny3Hk(3{6B8DVz+Nsw=PX%vaV< zqj5qhmxO`zm0O}ikiL14UDcJyA~8+D3%&(E6L==&-mstHbMo(PQFEY;y|*VrZTd!V zGL_782yXz~`R+SJg8$x3Uf&Dqftw7WjsKHD zDNQp|&|pUu{9|vSSW1q9RVKZOBVjt6-ecbmY!wAgXCy0OM640}PKY^P5xB^JU)n}d zBuUUR8QD{k1x)A6R4>3r3eTX<4qKr%8qTmI|HVc zh^-cUn@IMk~jnK>MTHeSx?BsB4hs zNIs-MzQ}G_X2f%nz7lrI4rpm76K3R@)F@t>;I?A#oCDOnoXpWN6hvSz{mtQa$vjq! z7n_`pb?4ygu|W0J^aT5-U|u*{6Ow~uzI(a*wnH;mGh5 zhB3|~R%msn$MlAJp}d{p;o13)z~i*Y5?HJ|ToXIDvPS2*RNCg`wU|AZ_YyKrt?%H6 zX%jSYj&Ds*Io=~Cb!5Phc|WA6(GdvYF_d@@4t;02EAn%!3A$PBQ=Y!4d12ebipvM9 zA1o`Pve(g7pqex@B|HQAfxEIFI<5N_dZJgzx+L6NvH|J+0BT1?P&;`N2OKxja5aM0|I}xGBqGrAQ!us7lQ$dO>n@~=$S1Vbq5JLE2D%5E(?}odl zR&{cH*5Xm~Xw~l0j+Z=2 z5@7kW+m#?#<9NBY4sHjn?57KEh2zCG#=OV+wdz#J)Q+bti?em_m19{TYD(SnqDqdn ztIA*dQ#6k@#xsr2YNs%aW{N{T{|*9FAwdUQwnK|jLKsBaPVAdQ?xKEQ!5_v#0A z2GmHa6T+(YqD3{(X}=u~&9TwTJdeMuTX_Dx1Z$l8#8BNfv1oihROCul*QkizVpy7- zlq?TTzvWar0fNxn#~!Du4U2dqyk{?cF&9ZjA>0Uw!0 zPHlrsh)9Ifh2VlzmF4}LQLRpnUR(P&aZHyhvhLZ`zyRdA>9{>X&2PEXAqIiShk)FZ zFF_FxCyg*4-~BSqdFssgAMRvcu0|>03Q9hSq(8)%No}995!O7kpY}JV{@|^Cs3EZd z>-Zt3*pHVVIjs$z_X%Y9+LB^W^z_geKkn^%E$Kpgbcf&f=jm}UC>sq9i@fcqlo#QE z##P3y&cVa%5oi=^u>0iWv6?@@H$5C8C6YP$j?JhUEvYz%KiGVLxR>wY_^-iMfc1BKb6<@z-e&^4n+qw>p`p;0?izF)1 zzk88rcjzI*ZH4GmTY*mQly2<___T3VQ#0H2ikE&3}Ww0J4NQY{Kv{Dspi1e9X2C4K5yN zQ+9}Hxhsf9as=CqwtdtX>Ve5iP<#gyv#T;j7+;{xjHd^l|8k5V8`av1y-Dbg3&*B% z23!Kg7DK_-jhbm2PtTkyNs;2RSl)H1z-PZ4q(aLR$9!D=YM_$RVYkRr(Ugq)(9Kxo z2Df$xIMs+Hx8xBWJznKHyNBf4iO-*LW6VNc@YtY5-npnSW< z&}d?>M*F*SC7L@bZWo3hSP=ANyjfErJijdyI->0K0X!sRQ<9nIB(B`gCk~f` z)C?CI-;G0B*69_i?`FC95{bYqeZJSlnc4@iLyP~+*Cv(7MXq0p`I0j-E4a|HLJbAXa4jBl?8~>N{|U*D*6#BV z^-egqrwix~>+(*le^RvWG9!hez~e4i`!sYgoLjGA#Koc2TxGUg1dA<`15|x5&4(Gx(Nb}K z#5mF2i89pL-)&rilaTjg=X!mTowG}}e)E|mt7*61uO~WV*s_u|w3fuWUJVSQ^oR2+ zdS+DMkV6nIkm+U~*)ujTqsCIFrEmtXX2^M~88}+M zsH{gw#HDY%nK>h=qWZmrV+oXA%%~)FGFW93!yh_hvy}kDJRNBSjz6JMegJnfLUvZ0 z!c{^juimS!_?1Yp^+&_J!m#pD8KGJhfU*qbtUPqzZYH=FerTw6UqT#qPosVpqofKYRvY-)`FntuG+ z>0`#b2?S*ym8P~#hyK;aNjW7`Q=TdU$t&n8DZBGL{6x2DrD#LDKjPmcpsylfz4z)L zIU5V4iQ3cxj`Pc-5c1)lZpLiAWbMalF^XZ*7Bh@_DTI^o$X0U_F}Bdj-4zXj;U7e(mZcNMN>jqzPBcv?E543P5_@PFQphkvJ2Tkp9U zy{dvjhg!HmQzgJkrj9=Lo66N_$DooU(v^{LWo|V}@t!}2m6e1D zwJ!v-Wtgk8|HS9Alyz<3Y+(70#+9(9#786h+KnbNVUMw9chu=DsvO~yN59cy`Tu9j zQf^_FWU7=%df%kY_tGsS90Xs89eXj7H4iYlGY}9 zCB1{iA}IKfg02G#do8`ntZh)nZMQqv7)j&8bxrzSxBWlH&M8I|pv%^6+qP}nwr$(y zY1_7K+jjS9+qQMj+~iLtnR%F(N_KYYp^{2H)VKCpyH!qW$Z2_<7*jY1!)06oO)n~c z*X52@#+c60^i(k@(Vel2&;HciBo;|$q4_`(7nqi)q7&ilgc;FtnzC`?Nk*@djWGyW%c9uPUjKVV&E{5W-r3O)e6Y6 za1O2w^arG*yq0ToOL2?h{v*Ntj2ElW^6N2AS&4S+uThvoUsoaC(}CI~z~FYK<<|3D zyG&w?5*s#ukzAEgzp{-jiQ;kC8Qo2PF&lFhFg3*iIKsvKrk+^zDU9 zcE}QNkKB@+v%Y@0dQG>1?=OLn+>Htq;S_QQQL)f;h<(~!f(TzJA8I| z>N|!#G;Eq%p&3YPeaxizkDb1eSzmakEB`ylhw2DW%iKcD<%8ed`40dvco?W zcciOMshw;#Wkm-F|L!|GP8-8A(r}h<#wqSI>BqEtyM@l4W;J6AKj8ewEpL#p{XSbO zjwXE@uT~XN%;+F)zW1l8k#Cla@`=x#5w3Y|CWmv0Dvt}&u8%3QRXbqyqwg@~fqN&p z(XUo)rKWbwPGtCr3M_N#wD8ItoVH~e@4R6fx}z_Tp!D4}Z97*yti2&$dI*!Qdo;M{Gs16KHneFmKsSBwq*d z@kAM{q28JB$Ej@Y?+$NvF5?z_wkF|eHy?gj@EHe68fh;uHa^pn8&6A~Z|dwge-?< zlM6%BE6WL_Ud+{qr&n4Rc6^TlnaWMzUAGTWLIUV5Zme)tj{n)A88r9NF5)u_ZnuW6 zkLJV##mYqaG%|}s zxEGX{68qv!1~c|t-EO`OWTgZVn9FL*QSGiUVUB?@bYQVKB-CExJTAmkIKkKu&o6Q4fR3&}iZ0XHIg`K8DX|B-sX{n~% zV~X}Ji;_RX?y1l9z_RsXU#LN{T9$hi02l-8nf*ZSmL#5`2O>CxT0mYv@*q#z%*qwmbpP>SM84117S<*` z${d$cNQ7lfp(wks;hkznWQSHEEn{1y32^dtjsDu3h8V8RkFH!kUL%|KVVg)<7@8r>pq zeVf2F=qfFhQFyolBzJO}vOXLrol(~~l;!@l7XtbF2s8JmtghYjt=c3Lf<9VD>i&fD z+2t_74`a(h2a3vd8ZNXYp`4}&Knp?^z579JdAIi<4*U)QLSYiJPUD1uhc}2BtM{fM zJq?W0O;r;nyBj*m1xs89*FksA(_d9Rk#MI#i~61FmP?w^RiK@sO*1o36JiGM<7M&h zTL{(3dm~rb{$`$}UjaY0eCu)$32EQ0$)4E@aL5Pw?wHXumTI9M=!|}F#%62f_>^Nb z+G`xkbzTRfypL&WVHW7qV{|qu_$^;S!O4jem!iN}QCn@8_d0{kOWM3fr0zZRAjYje zv8JJ}tbH|B%`?=??higX%!14R1dL1!jQ<;8WM*gl4^m_zVC7_B{!jJaqfOX2SUCRA zfYJ3IVBBt{T^xp36pTdjLKHev^`c)Cgve)vM34fKJnK;hOn@sy0ZgDGAd=;hpjQ$> z0Hq}AQMac#&Hl`L9rds`i9XtNo@sf}@}7GS$=aS5k-&`g3z!!uau6aSsHag--5plK zf&hyO>^D#ZU~i*DI*opV25z50#GpZnN&MCaQH24D95S|2z{xF(69aVJ$_XSg5LZv(j&KIRBCdys1_y%sVr964TL!&1XF>?Twzvow{uz*+cjJ(!00SHbo|V7=3K=nP zFWQ5ELFnV07tw~d*ozeSCcOM6gaG=@g;f9{dAN1*Z}LYKD)b{4He_gyqDDE65m^Vm zj-Q3d4-CbFI1NDt5mLO<7b#~(iaJ*`K5$n!LXXbt-(5w$ zhH4&y54;oKAiU!T(5G1j1{^lrLj39d<-VIk6%YY|5Fd2t2Z!(`fx64P@W3#9HIM1@ zfmwux06c6V2Lk#1`1+gxqo?Pju z1sZ?FPtXn~?PvH(8~<%v_AmR|Mfs_k{FOfX-Hs3KYI?k7*u8!GMGWmE$m#h-x+`=Y z+lS4M8{GtW-c^Af#$UQ1-ji@o`;E0a%=h>=dvSGdgocC&{!f3fCs7Ty2o1En!8F6sCsSaWM$nDG(EkL2MTUOzMpb=Zu9PvNe6|U2eo8;H5Rsarh{TaQ zVcrYpmB!2}zO8aIG*$arf+^P%j*)R^jJU>u_0(uL1D7kk;uT4(aI2%Rjx>4AyJOoT zzV)Xl<2tN(u@JqS9UpT$zZ;$~^p3AJodY_}3R#;@pNhkNlumo18c;VN1YG5+f~Gxe zl5@uTPebaJj=sY!ZB502B}2T6yRvOlf%Bk-$yb4tG0sxMD-a3a;rEq|yqmgt=9*lt zop*?dCI~HZP8^)KoW2q}@^_I<4i^nzT;y=zMO3l~Tc95kA)@qTHp(J@z7FU%|OYv?pz)a%3 zU(dk@=zUDrpXP@*fUhX%qdIRNu(CNV8`} zJN~D;$_?_8>f)GiRhHf;hM4m4q7y^QH-H=&>Kr@A>I-wCo)g+Od(h6NQ`OHzt`#07 z`JI)S5Lf&Y%(lf?2BMdfH>_>HMe0_zA+dTCd7C}@zylV@MmuD~@LjL{>XUyy2S{ED zDjS*0EsMjCSDOOr3D4KVLA`l1~*R_;zqBy@@yx4zaS zj&^4)xEb<@P|Dx7&D(u1tkw>3qvc07N;gZq2M6mhJtIFRinN{Q4k^HIJO${JX^c<# zBcWmJ~m?UhR@tzwUq31@nPA3Y`BE)*981<^e~&K_B531GIx;%#FLK z+gTVwNrnj(o0SFiyoA&1WnBE9zO`%2yDEKE^So~xRPk#YRm6R$HvoFrFMC9J!l|&J zPp1bJ$K>(p=gAQ&>lV5L+U*N+PYd0q!*E?7Ooy8&zvhS`n8S6FJQ7*n{o03VBu(X$ z1cY?fZ0*U%D60poi~w*5+c<0ezV5Q`w1VI`tvJHRUptgH%VAygkH zteyc(STm}#8rGNp=8kka-o|Qj3G4 za!Pv!Uw2gYflpaxJO{+)cvpb7D-RqLwKa+mbW`P~{#kr6I$~>m9h(4ODi7XaPU^6t zbn*;SyT!%?a}94;&l_OmG~`EuwV#wSyZD{9>7YUW5o{UCrd4=uQZ?YgI11>vDW?^T za#Mr!%G?l^KY%Xgf$-egXUgq3OB>U`B5m0?1bcoCB_0j6#f5t&g;UJb(10zoT(F)u zE*0w2g;1{Czp^`X3hp=&S|!bl=gEm%R(;5w8rY=BumKeRCZJ&y`Pq2B#ARb^SGD0& ztWEfAxJAbOO&l`;>|iNbl3$BT)77U7J5P(3N{4NSMxqcQ-lP}23Z6;E`5;9gxcu}9zIsRi+JRxCQtqhhk^Af)akpnf>TuL#g~w?) z?td9Ki;T2}Lh9|lYcE7irX-3NSj!B5cf;dEAH{wM8E=(zG8!+muO|&L%Ng8QG~Ha( zxbJWmChHV&ZlQsd86cPnoS7jIc3Q(&uEjw>FU-TgTqS&YTE6_+rbv`jtTGI+Ab{L&2@ppEbLgm&-@UN35lRAo2DQgLAG`S{4iF zf2vtfV<^9=lWq5& zd@d{ksd7cFh*l%|lGS@N$89yICj<4+?PPG93FnK5%g#F8ee<$}rXdPsP_e>WqM+v? zheTky!_K#pQJHw)DuKqdAzfWZM#VYk)v3{cr5iEpL7aumH;#uXK2y*NkX+=tzqLMN zb4Y42UatVkpyh4sX$Qmo{HdJ+LS=nJH_Y^x+m8!Q@2i0WKf%lgVOmYG_j%MbihO^i ziY}{VTqs%8V*J@o%8ldL^u!b0J zKwkzfi+@p|=kZXusK@p zi_{cD$byLLf;BVph9+BG40@Ng-7`t+Wy$7*-URZ1o*|<}s(=>2vzd3q?D^GE4MmbX z6Ki4w;<^^1OgFfg$||mD^(N0YSsq_bi656wy~keYoeY$@ZjhPHXhlG&r5&9^{3CXo zb42%)Cc-b$cQenIr)Q?krta(G!bn<*^zF{uvXw*qEQR@m zI66ZNr%uJjs?)bXd()w;_$NT#LY|d++u@QZ`$tvbOq{~|2v|8T?}+VaYA4o8)JQec zJ8syKco7ONDHhK5a5zcqcKFpn{Z0>+ZRt%qvuxyZFI`yGISXjh!PZVVQB-7+c{h+%aytbm3K2DdTGq zA!HZ!6Sb6eyK9JjVWf8dGKuj$CGe0Tui8>S(+yh5Dg5dY;gIMBC8wL$0lk1)h6Mga zq-01ZZ9JPdYPz}Fc7n5+cgWQyqer7Mp#Z7cVRtYX-x{lmiu!NM%O{fkg z$iuSk@r+r^3^T7MOglS8{kEy;=nF%Q3t=bUFpdc3%n5prrgxKa`KtE4_-CKu=5a~( zh^@hZIx!5uw@u9K6HNusn%2nx1u%L_7N?83>qobn*asm2=x^n>*Pp42j9AdY`%XMV*=sfxc$vt9+VXAlAA5#_*D2CuP^7iWgcs1D zjO-O7Nqn*FJBKI%oZfX`@N)Ai@jwZZ|UHlWCxv03GJd}z*);7Ax zfiFHk)7LP>PkRPO(EJ_^ACjB!JMEx>#Y)(ppT0j*b1(xWkb^5dyEPxxM-v(m z-?LCgTRqL*_`4tZj_fc)YEK+(AG~*)DPmQ}>p%R|8f$CGlPOb)WCao`Rw_C#HL;b5TXLae;Bsdn9>yd2YL#HChQ zr0${nuAij&gN1uV9Ti(8mvzZ8S8bow+hF4#2U~WK+KSVfNL^`dib!~+T9X^})6&U9 zl|%GfS6@*|R_igmpF^>^C~eCpzhS zB%AaFp0fdamvd^b{MFcX({`^OIeKayP*_{SC?8G%5X5LOCzE<|rrw;=uWK>bHoKrJX6m33hWg z1Zi#KyIQq0g9*+XD8|CH_pn=B*44A(hESNG$l}JM)6kD8dBw5wJ{&zml#xew4ahcNG4{qfVq`0*B11yl#0D#w6Nbf!5J1uU=+#xV)TB zcOCV0BBUhnhSJqR;so_LZa zlmd%uhhUVKJbmuj1=reOBIL`aVRci*IkKIVjQW zT5ea36c%Ti6`z2e&F!V6hY7Qwop0e=?N z!!Fvqi=phW)|Pp5xl{bd_6j6ru#)A5i*{19)-eoR5lxUlp!^WNih8YdA@hkiZ%A={ zx>65m*eNZk%6H|(w6?x%KVx^M!WKXy?2>KvZ^p`8KVL;+=x0A}h_TI%0v^i@O4vS$ z`RWD698jG1t)zDzlhW%Sfvu}U+HK3%z1JtKA%$EpCj;x#!(S$sTj)6_rvZ~YhyU6A z${+-TRWWpu6CwbXJgv+t{doRj!E5hMUk(R+&9TyrxS4|+2E#U6h{m{ej^ zRi|h@Ek#j= zLZUClIB;4i=Xv2$u#Ouq2W%IS0ueS4o*LnI_^i#4>Fp zE<*(kjw)++r|AtLiQI0gh0XmNRD8utEkmj(lJOHGI*2g`60MP^=KOD(_gMn|q{q4$ zPk!fNL#jX@3_;6CX_g9Wr!$c}#r4w6J^E;HY&`sGPEZcene=*62&aVFSW+xqgh)qY z{Mbm&ttkKc1?y>Vv+y^P1JAn8oItdTqd!tLY5q(2mG#1ocRWc6JUQiGMFWq}oI3O6 zJ{r361m&vxWnw;))iM6wpg(3beM*;-k?;LGqVr4xjF7dDqo0U^2%og46*h+nx$*)C zT5Bp?6t=93XQz}x1K@*p-*0{&LCr}OR)$LR@S6dr`LUxHTfz8Yq8JLKG>O)vq+Wvm z#15xd6Cw?V9ZudlrR!=^!0LQmiTQnu;tZx8_ER)Hb@|Hb?Rs4n!ZW3_fgQae6p$mN zIiklP+$dZYu&3! z+CXC6J;?$9bH9O1UIn~v)NZNhVyxmIPhu-wnUm+ue=VY+lJbE6$eFFQCLBExi{GN1 zz|o7UHA8OV%)P$LVi~Spz5y>x{9s0PfVJ^?ZYZoUF3Rb<)=~JX=@4gw$E!n;HB((@ z=&Gw>ZY*PzpN}1^L8YdEe@|d|^|F&sp^W3rD1QB!OHs+*lU;JN>t*h!IHH8I29N1w zTV*x8FYM*EP*e3b#mHa1TWT`S{Av2trr7p6@qP-E-F|w7yTynDk-r#wmRTJqrQd`r z@4enW)E4{WZqBneT!PTEtL0lTewm0q9o5g7R37E&)_PVlCKVAf_^joUF`VkbR}iT$C=Ke%M7kyzetL7~ug~h=ONZ4j!p+np*hysdcF^?p{Ohm1 zycC%|<*tSDHrLM}8nYqh44!cZH|OUE;N#Kl4I-tW=qjlK_biV2URBn*ogBA_3=FMl z`Rl24?$#M4!IPC&jioBeDd2*|TN?Qde14n`Ef3wy*mgPvalg&B>m#o;jhh+IC19~R4=3T9fjK%L`HKL5}Y}dC`g})i_{M|H@ z=FN%KGD)GkQBM=9-b*dp*$N)RnaR*u3fkALmJbYc=?YLR6`^#=q(f`y@NW z4h0LyL^vn%OB8%w%h1zAX%zewi4VqqtRsk2KaY-ef6_meIpzwGgxR49SO{P za%nYP83kXb5wD@q8Rgd(9f5XIafy#>{|OIfwDYh@La8;*r+8WH1EKt^Si3SWmMU!5 z>^d+;!Zpvw8v9OW$?b_m6!1@b@ohywA4PpMq466g5iW`@FG^wYi6fxdk<2?#NRad; zLWUm@2AHO&P<*J@n$n()Nsl zoI77%e~q`enMNnQ^myq{b&iH3sa%9gUq%HSy}8l3$h^et0>ZN^QIMGf5)wjk6B00E zvvpP$N5J0nIdM9;<|oz{cVa*B(TL;(h|}CN$`hu#3bK1}1vXZ|`iFrGj84rC_RLH` z7?_!yzE|gWBCrXpj?65f;)}o~Ha7ut_Qyr9aCvfQW@vSGJbyk>1x$rt4GsFxVp zII}r6x;QwNy0k<#$GThptH8G<$w6%b4bZ`uP&&n?^LRviRxXXwB!mMu0?rhI~ z_Ki<3&MsfMW0uy!!C1Ma#?}f@F8r=_NA3NLPY2Qgu!)I@$(hjs%p(FkF*F#w$>v|2 z1AIwOJZXRF1=qC&_W<`_^}yuUm0;d|<7c-uC14}$9UMVDJ^z#)>_WswKpI$C9sn_b zWomB>d`JAyfn|Qj?4I75oCD4q`snkM0UX};{rWC>?=mQRb7Rl(pZLe53(5%#>B^lP z`b+)fDQT?j0Po3;4Z`Icn;8Q!Ffuj*aI9ef_Ws5eS(v=%@nHPENv&tB1GoRkz4ple z%C2AYa|=A@0T+OMai?~Ff1UsVs`#VnrX|EI9DN6$|2}s99zXorPx*_#`|CdaW+k*& z#mwk?RrbOD_&J`_>pSq) zaOMY4(_eVXU&Ld~p8$Kx{_KSuzH}ETi~nQyd;TPN|?w1_>@9U112bTuG z^^f5_-6uHzp?87WZ@?{Y>kqh(een$co_Ctf13Z1!(2@A;5-&a8m;ls`Se zragB4CO_H03*PYiH|t~H*RHiw>-#DdV^7*s*vt3*%%0Ic$k$kE@9Xr)_76VXEu+pa z;O^Mz3;bJd#wWkp?e^v&{vB`k>A%zHcmI&Q;!QxGJ+ZI!<@%D<%H1A1zF(3)`$7GV z_KozC^ z@j87e%41WMPoLV+s%B-xta)blSuE;cOv~aToQZa~@GVt*&y~+M)y;S{_3~0HLqx}z^!{ZX^JKBpwzO)tE8pNr_7Q+N9-pvE-Gl?GAe&eWOx)HKTz$dpCq3d2$X?bx?BvZXJif6hV$vp`h5RmUcq@8N9mt?6P;C9ER3 zJw#2^M}dO9*J1q!@L35R9yh(z1?;QOQD~dQRBLnG31iv3`t!0-47Z|lDr*&jgVrMh zwH)3qjluy}peOiVW>S4f=;3A{I3ng}4rM~K?qjt=l^gRX*MyK`Txb#YH(~`EWf21C zQmR$c5E{jt!o}*Pr97Z`S~7Swcl~HA6;f_FR035^Vl|#~8&F7TG^kQxiMY^!`4Y+R z10tR;?gFpy11knnz2xsjeuRNUf;ER^aDIVo1Im?&L+(eVP{~<}8u@njguw&P#1yan zq_LQ22M!?*`B*pQsUHCj?<*53M{b|5SQPZN0>-@%;E-j-X zPiv%Kz+9aFWPE`Fdy~L44Z4v@st@Dm2-POG4^+-G!aRwy_#iBQ(WOQA(l{NomkN&! z$KE_Lm2#PKa+G)Z2+&@yhQx&I>My*)RS7480f&+7kh`sMp-ZJQS50(Saq5dVVLlA|w7wNzlbmpKYYnd-s$p8Y>mXoGnKS zauHXGbm2xT2fC0}gf)3w<6RkmjLyUn3rL7QD-?F7)T$<8$%C(0f8aB!yQ<~7mB+DE z`n0w(xjy9n&_iHqi6o#pVwxf9b=YEq6wT3KNDeZsCTs~d;pf6MqTZ)i5>Zp!D`D1L7bLut;qxmRO z9;`7X)Z)jH`isMI|4oKR z#w&s&Z=pA=Wn5E6@ch)5`eI~8R7YB0HNu|X1JIfXwVyQLH(cDHkr=}WEefJF47Uwp zgL-w-q*7Q?2F_>10(12Xld1+RNd@*d_`aE zumYlJd_`_tjX@A^WG@4beTVu{ka0JknLI{hsKXt)@1ixI;G$E$zeEsX^`Hwk=U*(N z0MG1`_Cil*ULt#Gt$gL1#O(1jN$sh`9|FBYSgi}q5zy|`SVbnPl`ULTEdA;#Nsctg zh+j1-N?ja&2v$7EySj!X?FuMFzUG{Na6w6RNg~pjwQk$jv;TrWS z6>V1DN(PYEOIAWr%S#kqEw!c^#YMx6TAV*#4P1u zy)r#?JhRZG3LVi23WguwA>5_d0zbXrL1+UP6Qkr)U85Bekg%uN>=d}Q3opEyCLHnX zJo#)TOttiS2}#})iqFxmSfxV5`Ru>)N1b&5x1n!s6VUIh?GJbW6Ej}ry3K!sESs0{ zgW<%94g1t;?aQ>+D2MJk0#(5_ySg>X#F*oP8!pQ*{gU(mbB0{5!hAfWq9sj-IBAlu zBS`8JGR_uKv5XiEP8rKKgUFCZ8o>Pmp)hv@marS*d-y4?Qq>}iTo*eF`55EDWw-IPChIMAd~&_FUaSC z5y(fMk3ag3COF4aEXjlu6S;%~gp;vOTRq?!Bi(l>IV$_--Gok>0*+Ogk-!`IHKAfzEr?}n{)GZkKA#@59cCA4q;3g#4`b1Ua0X9>%}aoq$ozCvI1K&0wa zNYU~vEzIdo^<0RlKC2#r(y2`aD5$r<7ysqYPB-gdWI0v?T(>@~whZU-?dAa*eX3TVT z+L)ViKYfs;RBjZWPvUT2oO#6)a0CTRyuR>=VewO~xBKKJzW)wTznWRA0J#cK1?Xrm z6^mir6mLZI8$(=wjp#`8{jN+@PYmO_uiiSls@FS&07U;AtwjM1nR-1S<^KLVxn&WS zahgQdOr(fVoOghtjd;1Mk;~IurP$dWHE-B1~#5JHpklw5u1x1OhXs9 z#)?v*rZbNNW(&Rg=B6G+CluivS^L__CdKG&oQ{WDiwMGBoeSab%N%N|YbxD<*@8b$ z)QwD>j?TcD!YSC0&zKT0)eluyQg?^M8?faZVPFCNS9iG2xGbB>48eK`PAq`N?<3!I z)nfHPwsT0U5am;_=Wa5ZAD`j z9b)>p!^wT?;RtiW`QPIyVT>4j*OHwOHy>OGb)_P!1A~!8D0#nHiygC2eg+1LboxjY z&+WF+?-VmxNWe_4 zD41eUbtuWdl$w$%KjcLwQkz$cbRW>W3kr0Dr?@pAg+(jNlni|=T+;iir{zc7nQ`Q$ z|9p=FGc_g&jo*~;n+W_ zSy0eeCnodh@lZ@01g;!O(KuZwY3WBos#6TzUNC&$blgZCXVW3EDDyjrPkZ|XWrKon z9=a8$cw7AX*v7ycS`td21l7#nW&Ny$x~1Ku;3?2P=xV>$;ZPoc;0F?@)9ux^#qz!j zW~Ou*q$IQj^W@#~osOL*)Ju&>drbnEuL-~!du~~uhbYI}#NGz{-o$!Q9WjDlg3EgM zpOIjvVs)Ir=Nn0Io%)IE=5;tsCILl3_LK6j!9AYFYTs7~@N$QB8CHvWti-sU>~Oy1 zNX1j(5CIeS8??+v1P$91nyXe|4eYubejathcZEn@y{u~49G)pt+-Mg1w8`;ilS6>Z2)p@Rp?TfYRGER`bCAthb@Cb)ZkqNP(qhc|DJyha77#A~-=%>O&*Y8SK4 zIuRajpNp~t6%sh7PB&#nBT@zmHA?Y#^R#?zoAK_S=#urQO8d;wExLN$Grw+;#q<@n zt{p|3gti+p3nJ>nM$yUapLA++ui(+AOQN$FC^raRH1r>vvWRi3Xr)wejbU42@R4$%|}PbX#cGUOhRA zLSK7gleS#D!Sx(tUR3kDyQsRCq}*QaVdM~%MH{zU#Uj=QXAIpm_6 zJzB!595WUs*6$N-e`KF=4Os!KAoF~+L9lp#syy(IWI?7|o-Su(vs{$sfw~PN=LIY& z2}~33L0#Qc+82QQkK$2VHk~w+?pyIva+j&ww62ZrS7KNkJvfd0HHVGJ;dlpZM)($N zE#Q?sHS`(li5N}Nw^P+dO*Q>>JSeWS2B@P-2z{WPB<)Rq?u#$w$tN=Mpk0xYNogh3 zr8H5B>sE9N^>w|oJ^R@6`zY8guhzEu_gMYeDgwNo5kk8c(BfNsc%z$dyp-nu@IyB( zi!CN@g>WW%luq%Dz}x?>gf}YQ^v%g^t3C~yoT)xOhkAKHWc^aE%K$VW7-D0@PKTAa zjyQZF9QgIBNOEmCF@EBXem@Zgkk@kDV0mKT6LZhwW$I3dO=rY#5^NiamyzQS90gr1 zV>A+}ITkuHbcUB~zDHa>fiG=k(69fqi%C6@IXWVAD!n%s`iy(bBlC+G8|Lm>n7M(# zyKJoORh`^K28nGb>yO%cX%Egk>NNvV0KpnugT9MLVjSVOXl%}P-EZ!*DK7f{XYRLQ3=txai zMhNa(iO|0zhMEad3M&3YTmy6JpmM%yeRbQXB4_kyP!QGcJNBZm^{Ziu z6zY$K&?LO0$R%`|zEt??FhVgi6k?M$*fhzXP~7NNtxYH{UC-#5cy^C2EQ1 z@jQk0wkh-zRSJbphZ~)Kr<2cS@WRTGLlbQ?sC-Fk+pH*1MH0_w54YHfKq$tkpV~@~ zEh6}kc*7>u35b6t{zf#%xnQ&)Nh}1OUw;x@;Guf_>TFa%Fg5EzwG=t032+N=zf~Qb zymS_lunnnAEgoDS)>u9E1MJ)g;-KuBQxi`OiABkz0^D-qJM?x`b$ zl@%QB-G?bDi&W;GqSlzR&zZ7sgoqUer-bHhXBMs^0f`$VWx`HCo-z1b=P9NUVrK$7 z=l&FzxvEJAdk;o}EG>=6lD`}AIN_H@rmf%^rY$@XkA)3e^LA;M|5pU^Wz!!hB{*p# zeYIx|9~pt8p07Ci(#&Sr9gpgu0Tip_r{f_R3^o+bNL{A8Dri-yoz|%2x_z=Z1i=;# z?W5Y+b~z1;E9L8ASN36y4p6~ID?5iLahIS4T&w~Rb!1qh(%5US{0q`N+T^P98JX3~ z!>gGb4IT#@HDkRZ{9Nq=Ib6mZ@GSOKi5Q+Z0p(n3nz>XB!wcJOQ_|y-T-mwDQo_j;~^Mj1ilCw3vf2Yb!O9ig;17lGUEl zL^g8yn`-{(v(jf%iAJkGAT~kcPVLMwDABLrSYft6P+qt757uYyCgBh|)E9+pm!E)( zx43%F)5P61q^LoShsHKx4Mu!AM1oa0>4wd(2skw2oTDpknOn3X_#X$g5HR_R9}vb6 zDDl2n@yy&Vw-u$}dnH$wr?zgCP}@$n6FH{wcHx+D-Kw$!Zf-g_m`f_(4Eigbx}HS* z?+(ZbOdB@xvEI`8^zNnbBeqQse-q@|&Z|w-Yea0KFQ;IOwWoI(_`8F_h1RN z?U~~J3L0G7FDG!59v&7CbBeu+?i6!ymt?C*so81ZNx;f_wN--5pTechRnl9D#kHmL zAq>L7sWF);Y)juO<_b0@DyFGaC1lY|5Rv`E*DF}RMj!XEMd=Gh=(B)kM8Z(P(?T|! z%8#sDeT^*W%xude&1C%$miiY3lG0H-G%@oCVhYGhxHkvya~OLoh9!Io>JH6)idFgi zRE?C-Lt@I*wf5Q>sX{&jL2kPBL}uLDhC-REay`L6VQ)ospAzvQV) zX0@!K7KI0|U@^wp;y)oSa;NH0d!XDKg&dQrjK=aFQdN> z$dOn}Hwnp+JnkUCYvWY@>3#dr;rN-wXE4mXvSlQx>c*4m)QCt#?$Vk0TQ+A{&F6;h z4xdqtT8B29wDQ<4!OkcfUgY zQ}`$v|2SO7zeSZHD9Elj24~YBskB0#igAmGtxn49EhE(t4)LX)L~TqUY$StypG7#~ zT1f;eV*amI)|nUn9x06^t3cM23nQbUf1*v9 z(QYf)qHkyl`pzo79@(UbvDNSdvTBnrgy}STeeLkxac&~wU4!K^{~9F1!zDvEa|P`9 zk?n)&v^UWn)~Ftj;TQ{;D9M!p+q%n)mKT z(;pUv+aAa+#I@{Ylm}bJz5+r+L+o@9e#rLtPhqJR^jobm=_{DtP+`u<2R~8CPYG!S zO{;Z^+Q~R7s0Ey;c0Y=50H*3l=KTGGbz8K7bn8#f^A$27!jd+9D5+#pBxj;qqdV~6 z`ownJbFf0s{>)QcHg*K<30{CGxE8W!sw4q6hiewWq%5^dI_9%2FI6E{il;0jG7xqi z1xJECj}R=!=`@Z)Pp5KYlM59xs0+98X`eZy1Q4TIu2(-X7 z{0Dl@8FfQxDchq;*O=gW92(?e5F5^nbSkcqEka+ny`HxLYfwm2Q7uE#ggetpCpLCT zkNNeEkK>xZeFUk3<@1J?4BbvD8>bQ#t=XO1x%tV4|o%vk$GcE5KgBIxlLNEd1`?PA(y&Os53xM)?~7ZmIJj3Bm&I+KHfT zu&_f~^K_sq$pn%t2HB}BPhroU#`>V$X^?3ljH6^Hm6UosbkvZIKUHpb@OiXG(U0z$ z*6j>riFF@g!muo^Rk|&h@9B{y?H8C&=EXDxCW{;s7Zv}Hv2%tNMO(J*vTa*?*|u%l zwr$(CZQHhO+r~TL4{q@D%4b!ZWAwh05tA2obEe|FXPnw*I+6Ik^zjiJh&q{Ue}3{) zM`hvqqZds>l$lognd&%xM?mYF6P*-uc#shQtodaMp*+rAE`Kd~o(PxMa3O*gerV=- zpX}dwtFU#nMXzijDUS8xbW+PR9fNjLesnwUg(Xz&av&G$d43(q3WHZjsE=lH4%y5@ zZDLnu{XT5f3V?(UFP zJ%kCR0qD_ERI_XJND9-40Rl-;*`E$kFhT*&QAm@i{!vMNFJ_Lcp|g_dCG;Z{gK?x4 z6e`J_$4H~ZevUMa3ju^I7GS-~t1IUjS+`NtvG-*-bZ>hjzsjxwoASxTlITz=bK^Cq z$}W38@RsX%`BUUouM7JMwpfu9<~O+12; z%n9uo!edE+T1#e%LGF^Wa>|r-AAQyHo1mq$b3@T`U22G>T7$mdPb3^5NLQ9aG;fg23mTqEtvK(~>o7QI9L+^QM68=WN&6RE z@BrmB6&sI7o2Y}|lvY_BWkepaFTg9`!$8tsw6gL2J2e-O6*#Nab+jP!wbuXOUdJGt z;Z*Zf0-N*ve)>yue7%{<>eYb0$ znQ(_uuoFBb(u`Mp-1RKyRYQzfY%)^f-h25e5suWRds8UHa-o1dN%3=@zQSj6xu299 zLKISoq`(qB!}k(0utA)#jBkD_yngq`chU5r4G~h}VbJ(db6h-*2i0dT0&i6 zvzGS504g)=)Nj3+;v#Xa_LSa&!Mr_yc?7HX1Vl3MH&OhCp;dH>(pe=XCFU?tvQaLu~4$~1((KWr?a3@Yy252IlH(mrNhlQ3{dS^&2r*&y+bAR7=m%- zWb;zG@))S25S^$Jl4py%QR0h>b91c7 zbYf0m4H(XnR|OPbYI|?xCvifAZHVle2U%0MD%1cQ1^XeYwo+LUE25)Wi@b z%wfvSO{h@z4(Q;(+NNos?j?+0t|*!7UHYSu=SF();3HlA5;3jQIYbV$-n0HUKGD?< zE)E+oo4_F1g%j?7??WsA#l{>b^Nobfemzyy{5d_efa6L033BEBc$QbicN}M}x>Syt zAQZREMHjUS^*fI4a%{K*JuVXBaBRW(-ulUkc$-f*1E1wP)Yo2{w^Bw`%ltx1KXV4J zDQch3C{S*4Ou@v0wyO&kMVkE1Q^3HM`%4G3Gj9JeD`cRXnoIK&MKrRt5}K1m4~^l0 z;M?g!j)AV1m-^TlF!sGsoC#WJ+*j96{F?Uwa{Wqw4O?^({B+PT{8jK>2V*^HBIMgY zuOBk9a))3nr8*t~oeT#dWmoe&89oiAi&Q(?W)Vn2~233ieDJEU#8w3<&g4?1gjVRFEnp5jnX{0 z$$w||aQ`5x``LeKk+aeXuNkj^uu6FgOZjRqJSWky+DZgmITfM3NxO74#=SYJgwa*0 z`76YU5cIb%l{}?XbzV>cDK&*n`&bi&9{8i!GqZ(CWl3bbAYO@>tNcnw-Dn@C1JU~l zmbQyqdC&cvQj!Mu%;rS97J=#4$9x^9!75smzNoq3II`Hy?$r6z67~=Z+Qoe!K03Z3 zjp>*34Z)q7c*5~QNzH61sWAuA$+?#8lxRa1^H`Lk@e8?#1z5FAjGb+8H0PmDv^J-3 zQ-wugMt97Wtu)oRS5t02pi>N_d_LD^Pj_NTRnh~YUyNfxQ4qAfSs?Z59SwysIx*3f zBcwU-33KY2x-p64aRodHi`mg^)##6@;`EnSsH}%$qMg_aS{JeOL}7_|aur9x4^QXJ z(JM+|jS(xg&N}8~)3oP-RJbC&3dQJR0?YK^OYItpBc(hlJ3GT?@a6E)N~|8Z1j5d9 zJxKTxfgC6sMZ=$F>!Y+j+lIgew8nTV3UqC;9xDk`QlZmAd0sco1A=@j{uox7jN$R4 z-FBHSrwe;WPrtge!mFd~AFmKM=H=C!-8gM{PnI~eFwOtV2QxY&zbtH#MF((NuBon# z{g}Uq%taZHGW9Uz77FPeDzy~RV@4%j5r@9R*M)?Nx(1!Jj}I1Wr4Iyy9)GOxf6JU4R_Rf1z+|^fHb$ zW?vQ>UKY9^tCnYBG)Yr*iej7v%hGr^d>YOMAPV96_*N*j1aeRZE~Ac0Be-^#6B9Wx z0Kc4c!K-5=j;nYi>__SvcLA8RV~v#Fd=F;W`4Y7rFg#de)JrbNFBtOapLdL|0<>`a zZAC(;J>Ngq#Viuy_jodxZ927;C&++?;1dmD)Dt*E$goFo#K(Kbvg`r3Q~reJFpMS< zF1Zqlo#+~yVhhy6!m(^R%K{~|Po56To!0ayT|A@ZS`f}g(-K!^Ik5ec&&8Xdc4kCV zfjgG*G?+CK)-&`HJ$pu=jejw-_Zj0b+DWq}G(q|oy$?OEK$lLj!5itPA=u3AVyq1k zFq}veGHa<@0F^F?;JDHqMBCaYp0yl~%@zET^5F|K;x9rVBx*q$ezy4m|IoLf1R{qJS*p$>*tc)3#H{{!1fXh3O7+-$zyJ=a(wVg5){tmt-FfV z&MIT$UKw3mGDZwzVddT65IR}A;VJuAbF(7!VZiY%WNrZo2U?ubJxBNW>^Pk!Xwrjy z5+MQ>h;wq=#Na@*NcAAVz2%!aYRb`k2S}&D7z&ba-&Lc2wphmlex=|NbsaZsG8L^A zgL`-j&~`d+!=wH*YWBht>a-9hzgm7#hod0wC=ze2coKbfo?)IyI>3H;f5px_sK{^)5veugH<+3ej(!)J(#P>>C9^!P+_|-@89w+jO4Oy!v0dK-sMUDO_w04m)TriuM`HXMMjxzN|zWoxJ@S%5_d7(@V(aS zX2zTML|x9)5QR^iz^BnFJHwUSxTJSXe*Xb?R}!=*@qx&x7x!9dSY#^mkU)s+#{v44 zs;h2tPmFOdBN0VC;ljERWLdu54WXRwG;)LvoFH^;g9rQ9b~BR&1SIJCNI^s4Edl#) z^$-ecK)rPr%iY?15QH3d4hg(6&;y1lIZx*{C#sMEGoQm5c%d(zVv@DhgKD83_A6Ew zD`^t=YUH+-OZ@nF4W|os%{F?;mh#l5l^VDhm**?Q5k^*z03dHP3JsPnBX=|C3I%rt1&{=pnLxWCL#y zp_g~-M#8`tTm;q*arQxgR*f+k3VjGN1V=>0`+Dyl+NhgtF}={SDgL)vq&(tR`Ww<2 z*G?^2T%?7ucwA(0YWWQiF-p7Fq7~Rj?k;?3kbxH~(_VM8@4pqbJ_nzL3*Hy4$36M)% zxWPHrO2fW^W@@aZ!j%j=p`=Pjj6_@MXVOfGd^v!W_0!3Uq2-OdYq%Bl9;7Q20yG1L zaSycg25>tVD3plt$kk+o%HQk|hmR0F33;p2MQBb7!2%I{u|#4v5j-M-@G1&(NTEcR z<4_m9K9tERbt4$W)Bu3?_nRs%U6A#-5uRu47<*&%JyG)ZxL6jJ8%6rN(m&hkmSW`E zh29m-+VmC!s)BYv?i=<+8*mZdy@$oPt-j3qAV3|9_H+SisPVFfaPk4jsbvkbh9q6G&`$>=)nGA?QHXMHm81vV<`a;xoZkbQxi!WHiNYiEzan9W{)jqWmRT6 z<6?JGA_I6up=aC;CuUC}EJob50c_V$5%-KW?e>VOXiufbJ9!fi5B)tz_4Wx$A%v*Q zD4GwKW2A<}^3-b(+MT&{*kQe=&h9*puV2#Cu{a%!rxDYD%wZ^%AHxpfzg`P951~9G z+WPXTtT_RwtZ7;QQv zCNtxfIS}_gDezzMhZ2EeQ%(|?t;Z6^a|81w{|sC{>|9=-9b;lo-N-29l^cc(gf=8x zc22nE$EMry{SF3Tu&LLJ#>_g@^4fk6$ujV`Mg;k~{GRysK|ODgFYS2p2}(Rx$Btav z++1%^($gEGO4KKn@i$;~cnuc)+OjHqZZS4zrjqygmPQ{wO)@lt5Ao7iBO>9A5!49P zL-8hWILFu0{J@)n6xt%y5Q61^sPu(Vc?hZtd4BuU;!$Hw3CCB`D2Tf5m(QYB(tiOm zS32QsMS;Q{4SDFwrG3d7=s>6~^J|QLIz$aT2F>(5LeOzB9*K1p6XavkIk~2&-;*|q ziK1Lwb%AyY15BL0k1n-}r6uNpX}5SJ5clhewna%(DZdLm6&=xt$*`duZ>3HEHsDy)7) z{CIHDHWfP;k9D6OOFU1I-XO4~DN}M=j!M>%;Ei0iE8G)8#acxMq~0QNL5|l-&lsh9 zFl>fO%Ib4IAwO}n+u5KxN;tBTgrCk*2qe|LxP+XtVu|`^Q31#D%u?6}wV%rIM84qF zRW8@l%{&_E0RiQl(YU1u_-y*FxIf>@_ELywWEs3WjPyB{3Cw|fySw#FDNtd7JYcFW z=@z1q-mfpvUs(DW3-bPAWc<}2SlY?O$ZiGFU;)9N)a#t-tS7%+AZd~ezOJ2?G&`2f zR1G5r+}2+8z`8Z$#QGFN1|<38AYV4==~fca=5I3=V_p*k(6u`KJUCk95yof506#^} zzJ+#87LnJQF{-CaV^ugZyJ*6QjxYN0B+1)7FxhD~?@kk7&PQ3Y?_uwtz7AjYLL~yp z`3aE5U~0_YU)tKD7(U;4tS zfOJRb7T;*U1zjZ(k9>8MV01D-SE=yBax;QND53)d0N_fpRnqiqYJ2-yJ6mK`; zr~2AEngRl+u!jsY-j@bXY3Fgf&~m* z2PsNj;4k*tu&<^ONNtvXulXEHf!ww|855GhcdFJX35?lheQlaoWShS{3)9qBnmaiK zrYhV3#0W+&&_{>3%na3=H*|oc4kNvtwPGMR99t&YVtoG#uk6Fb(HMlqYdIIns@U<- zkp=2Pp>52k#}n_x-gGDe>q)aOkflU$+$WN7284UgyxunDAu1 zov8b>Ap7`bRIl8*{mo=b4Yb}GSs*zel)-+}RXsc)n>c4B$snnkRIYvzpC`b_tvs9( zA&Z0B`Zq8GMA#ETHPZ}*Tlb3|rx`CIDYZs6j}v_Rj*9V%|7O(s9_0ui>^*&m2K$*q zg?*r_qbdyMi(L0xcC{aV9`}&@@Wb6y3*zvDAeHheaV-QM$wg_TkzoUpU7_#L{Y;92 z_&&dZgimE>e}q(HeKDp(rtbu$V%3`S&vTvojloH~@IJSv0)Ea}%Qz~rX)k=7TCnx& z5F#lUc@u$m89XI|C#90jd;HRLe#V1SC&5?Pa-|u&*P1y{1Yx8m@76JABs8KC21Zag z^jUSjy{)Y_d)kAVUkcIBxxD!T&!e#W%e|6QcGsU&5TG>3Nhng6`SiaWTivs`k6^zd zh701klkd6%o7@7YFi;}Vw{>Ym`|4VrJ~4KKDC`;v2t$1BDUn8%C>m za}O*&n-|aqRx#+G=A}-%DW5bg0QFQ z^YlNwF9W;0(O0u)@t>Ne0k$ioK5jzsJWZ=`>p0zWq?^+mzVq^ z9^kmC)my56ViX85g4wJTQ&pPUoCZG@ub;8YW*%Jme^Uy80<79X2rBKa%P81)PCKoF z^#kha>krbS-6KkJ;f`Wct9 zXLE-a9NVC*Lx*0Voh9x9QsJri+H;$9XEQWg^5?1Vjuckaf?x|&9FV6{F4{d==o!~S zFlZZc(y;@L0P~+tio=DY+eZeTW*YZVEPy)uhNbxnwp4Gb*-{+p-B{@?RD=~mwOxrs zskdNxYH}OZAiURxUhT${n^SKd>XEE1m4{6~XpYn*@qz~-{OeOB=FyfNWbGynLZG|Y zlm9{VCym-TBS3vKufspR|l~LVww(JKH}st7J5*^8Ai~-A0_)bDiv$^R#3B z6f!O4;%JCjy5r2QR7f|4~*}jXMP)0wa1OE-= zlT06RDtRO^Dz=Gn7v+8}5#K`xwW4TkQ4LeRVWL#X!aTCV?Y z?9ihG@j817Eevn%!F2?L0O)g&cGSFFT9>8=P&tNXSp1TLl0&UIP8_#m2@#x77=#J@ zTar52rl%IHh)@>Sxg1i;7ezX$cqJ~xhM-5NPd}Rlh}0Dc2a-`<09={@zS>4cQ(6fm zm7o#*%W{*N0+TX~1Vj$-GL)5?R8#&(p{R`R2O-8G5Vbp<%(XA8-T(u9k{y{^UW7q* zz$za%zHfRs()H@whE2VN)P0y@-O!nZE2e80Tla2e5~ zxbg3WZSv|MUJ9*x8l%XF&%GE7iskwl_jTDRl-pyow~3{~bqot(C{=3xFZ_6EIjoe7pAwxm6*r)4n$m%z9gtW?k*CXkM`@vvN@C;LmyiBh0yF+n* z{`L(#0sVTW)3|@K7nD(wRKey)Rn3WY(?;VW_#Ycd8j!ajS%7nSv+&049_4e zyq@1$9<*a=oI&0VGQGB%E+hXY(O_GA-w3S5M>iAcH}_+Rs|y`4%{Mk@D2M zJlyV0)`Q%}NT+Py!$ozF9dDZ$N&WoS_*L2yKUCX9hgNxKw1{WmGI&fF>at&5sZg6M z{2hNbs=BwUUKL3Nc*g(;9#l``lv$uZjpKn1w1iM7h=V|$2zd9YcGM61c%#aw4J_e0 z1;W(s=b1*uIf<8fB|C5<4P4s{&j>xuikYKD<|UV8`J-UZ-4t5)#5K(GhGw>l)eW&0 z8{+YT2-gU+En^hF6Scv@yqKQJn9~SD3*08XTy&lzd1R9wj=Kuoth$ru+M?F!Sfio& z*nY;lmIs;Ev#@5i3IkSh2HvTw=j&G!tDSvR_mHXvn?pNw&lY8xuFoXV?p%NqdsUVZ zK*E(J;)M-!uln=Th9J=}@No+tpvh-aS?|UjVEDT_RKs+d4oJa;34K>@P@cxU*t?^k zr_xGKRtw#Dtc&iR7{NPrwI7O7b5IY1@CfJ`1*et%7!^+JV4}JOF)38jpRd>XGjF{F z&fqf3Z>XEp~y3SHYf2UvLh=`7ju6vWv zPZKV)=JQJe*X4%tjo=srt7~wbn&se{u+B@B^DwOv1;fMh=be`25+eJf+wiUZw36+tD$N8@kJDcgyi-e#L|d>@zx-mAh}^v<$^qifSBHW(t~u z;^oy_w~hS@{_n%;M^YDEuNT#Y6x^3yeE`eseU(=(Qs1Hn*-&HRg7%#}>P_F}=u?d@ zRdvC|?(>XnD8z~>DqH@B`A=6H&L27BajBuPK31^jGi}W!N{qf26ACOgMs=SUu14sR z#i}JY#NKI(0XEPH)a;i&zv*_22Yiyl@|SZVXAC=GApF1{PqVu7bbAnJD#rYf;r#Mg zgrb5RWbNJ>yX>oE2|+GZ8Gn7aZgJi=2YOJ(LLALK*lFYXCg0KdE2Np;9~|EoG3^3%6F+M2-Ul?s4Y~NAjh`yrc)1NZ;pIh-;h#8V1v+DI zB9iYZ>ul1{XSPRUpW&EX(VQia(@8Qr)(5Z8pw zXg<3R?)gK(eWQpdt*BFARQa*t0caY{nq;*cn95x#mFlntlA+(p*-p;TnOy+t;NhF; z{2cyg_p#QuLbLM1RJV5+fDRow&<6|f`4Nioh;}Y{P5omiPD7dZK8+E-(=d!$lbEd=V^NB~?1Uu$biT70r;QT*KcrO%<&MoXy{mzfsK-L{XLYV0ucVo-3&( z=+~eiAV5M4_;g<%l36cxq;OFcGL%B{Zg<19=9k7;%N=4R`_o+&q1ZIek_osKl;=s4 z7ucuaJV%|AkOUoR=cWq2(tCbuB0#2`h`z7-$I&GbdPsBUkD>qA|+!ErbRYqanIqbXL z9vK1bvXDf3nQh8aubQhjjf@?bM@m<|xm>~Ik&A}{`iuZV*zA=l=t$h>y*NB=(#&RK z6V8Lju_+xLllkT1_9Rp^^JWvLuUN%|E!JDidoxGWz8Wm_+p7q|CLG%$!xExj~ z!X1MeJbrqsUE|_d%Xhv_oqD7Do}vL@Fn`A`g*dsVFV=12n7{8qX(XY1j?I%9i-l;H z%-;3S_-AiAZGS4~-Ts7*Rm`E@3 z5nFc*{gEQ%pUv&cJ{)2*xlZJ*`OLslhhU~q9q(Xp-GbIA5_*M*QvIj%>OfZ#qe42{ z2$Qi+a+i@qBE#cVwHj|bpcFdM3=+QxeXZx#V)~Yw+s>vc3Htocr34N?hOV*)LE~i~ znP;4~8vr5^0w%3CUMj5j@bq$j^PftARQO&HSd^8r?qb4sBc)usEPMU!R-s~K*F4zz z-=OIp{wB|(%1EFxhA!$#;o_2dhM=N_DBL}!*9Fw6Z&V5`Lgc`{bgAM;1Xzp+;whvv z&w2Ecy+l!TVuFY8^8tsU%Xa|GsNL#fo=_a9k@&IvLOu14W!YRdO7hH&Y|C>lx7$@U zGE~^B1765{pnY#^_<(j6v7)(Q#whjhp6F~AmM67OZ`#=zK4>Z3GOKpmqIpnDUzK-iaYU}%m{@BRSL57Pyv{zb)^-G{1pgS#((W$^7H=pX|rVJyB_)r`j&Og z3U>$Xr1K5n0U3-KKALm5i#6Q>q)nXcT<}p*V3oAAR;K;R_jQhnOBkM-RY>vJzePek z2m>DEOlTg+(2mQCYZ3$Z)fP@R?VqDp{yPMFSt&j%VGTFXk-10}em#koc+d0se=;Pt zT?m5-K1En_>^JUgwt%|HL#T!{a_w3`P*$89t@gzsL03%as5Zwx#bumy*mVp;D)zq=MtA zMHH4wVEY~Xy;zADWSj*&YPGV8~Mh|7>FjaoRyt>g_kVeQoj+LARv zuIG5cq{w1PwGQ3_FMyl70V+a zAC3*0`b6eqfbe`l(Zwn<=%)x7EOS2jRHps+WAx1Ic*%0ixpByPvUr;HTA*{K$QP17 zHrs0jdONDUD4-gUNihBzt-@LDpZdB8Jx+I@nFgw!n>kF{nf3Y)m}JBV?!k^&5;Q!4 zH)I^bAH^jufJ&HGkTm|i;p&Fd$&eIRsEf@)`ttHX{2tTL{>fX@&ndn4Q}dVnOOC1m z0)wG6iEFI(DHC^Y@$0cFzoK8^CS=lO%R4Mpd*WIiWQrU-J=;;MgwDDN6z3TY_y=SN z-&O*(KBK4!Df0@DuNa3Bc`KJ18^#!GJ~^Nr^7lx>e6I6So>}Jbg(WmN3`%ca*bQv+ zUr8R6g_+Kr&?!&Vs3Y32tJS-ZwbE=64)Q$*EEIdaJ)89C^)02?_B7!lPh(FI?h{j}&h(_pM0RQ>k)Y#;f<-=IECaB8BfklSOWoAia z?1X5S=C06cff)WP;>!~^|k3|J`rs*s{+LeZGa#>4iV%k_e>Cf?VJOb zTDHZ4n*7B6Jq7u=y{tA-XRxOoP=s`=#t;xxJ$_^#^4Z7USIPc7I_uU9Uprlegow9h zgonOoEX=f$D?uqiyXc$GWEGqROM_-ev+=If^yNAL(%9S|5EfQ zq5AJe5X9~50yl@`F_g|~1oJd%0_hWtruow?m9PSWK=`(_syh$w zdWNjhCD;*AWt8q@@N9$v)WRb z(4}0FGV=S<2SO`kI@irQa2~b0OD4uvBNM|$?)H6YL~hR^Js2!iBTg9zdzcc;o&(aq z)a4?1fVU8UJmTPCBsOlc!S?pTij`3^7vgzr4B6J<9wX$&wWH%xLAc&b|4MZ|U6eW< zaW2MriG)!x1sFbOe_>iqRaYWslTmMk=6-PTJ_|&D{F{qNz30Io#aekv(J=U^2XmRi zZasu3JM-m3s|gayZo*t(dY&zhdc~-;mM?c=td-2^#3}KiFR_fxCUz@ie<>t*s5oFi zq3E1jbHv`v=(hOMz84=NV+$3LIzH7DqV(`-;2!WJ=B7*;iXGP4(L5WJCqlmr0l$Jo zanZn%S?PZK)8=gc@Kp0Dis;~uTwKUsJPO==L3ezQSnJPqP>bMEsN$ppg>oNqbBEl@Y1E3u9!J%W@lM{Woie5+rgy&&`hBL9+xn9Cb5*;$UQY9b(P9`3w0zy4EBH3G zx9>qO-42JFZCWsLl}i#>nzSRpOLeeL7KC07tq7)VAG5U`z4>)KR@)dJS-K8Vzde>DPFO) z$lKFa-}#*K5?6Gg%ygc4N>Cmn${tw+S$LPS^hOkHM;%|&VjX_jm#M(VbMXSgF`%A& zA+$O$F6Xs0*27X_0vQ3sy^*k#Tk*V1thqu#@C!`Ee&%(128*X^6FXIJJW4i3Wchdz zL4~A~{}5X7l_yamyf4Uek3JVt**D$Mj3_7S(Q?t<0GL^LWU)I9*jm~evd99@>rcz| zErL749{uCPkR9|YZa)jKw-5PZ1q@2XMOv1L_hYLu`%_o{n6*noVg#_UUW++LRPY3}AWGOjFWe2- zge3)miJ`O94%)s zuVvG}OBOhC4ls9qt;IgL?Xt)}e6Zc{|D-?PLelKbkdWfM(?xU}7HR;_F5y;y$69pp zz0NKBdd<~!#zbH)e-PU^^922D9`C$#JWz~q*B1ZGIbJvS^(JzXOC9vqQHyXWpHjI+(qQ`a1Gjl_=R&8j(3cfeFTeIodqA-@m|-q_-=K z<_!9ZeRinpryZ9dmjWO=x*W!{Wrt=3M*m<3xL?vV;OCG>8WtSLqS)g=+1&{@4=Acg za$a3x^cUexf8z+#g`K)f`7y@k$pLfB@ctnQyh9BbX2J3670IlG^#q@AVB@kL)CWIb zmj9viRjgHiirv% zPym)pqzR`sZy+sgy9;3lgW&K_@X#`_w3$b5fyRn8Ly1!VhPztUF*>_K<*ef#ky5L z7xQnZkRAL`OuSW@7+S;JFmOy0lz${OLVlBj8n=rC8k7x*R9g45`w51&L?MN{zHftu~Jz?qhn~&|0Wo*nc&DitA4y69x zNG#6x?MhPK5b!sL_QzOPB51)!50FWYzd`qLq~Y|fSi1cpE`RBoh{%CbEq6ya=I=Jo zfb4!$@1)IQQfs*d@*&fq4wH?PPh?DYF@-oKnY-YL&b8=Ap9k|SjSshX?}uXDYcve^ zG-qQfZH$fH(Q4p@Ls&@=qJpdIazm=mRJY_<4(ZF}U&oH7F*M2~^Hd@j+@CqxO48A7 zFs6qQ5h%-*Ay!C4hr8auCH_p@X}dGdEKO0(?D2tgf&rk%7YMVdI~iyF=bCs9Z*fYk z<(NaMk@!?|4EPVYI?IJKNBDy&CY}_4X{T4a6xKp%rX#*yb;T2{oZE$R@z&|OuL#SJ z$BD!tnIUZyFoZ*ha&_ZwL?yTLJ&kCphHJ9rQxcoKk04Gkm2KP*pM+@6k_4;%Hcpey z0y0a#1|(6^|HXi~a*z_`YFmP(o(EI45d*RA|4`p#kidVK9CplWYEWuvD!p|R0Y zsGm^HZVl3sN%Ln7Q)M(BnykrnMTsLiBEW=o%o1fHEi3ayy#!2>N2QE$AA*B9t`9lG zpz0P^h~I8celDP&a~34nTXX(%f1wU91->-$;LrhOxW^Lno=yodfIEQEiurpya_vUkkU%a1 zB>qla`j3bQ9Zmf_{AWqC*<2*7ic&nehD7)-eV#o~@E=u7eNcuOA1soM7tR^@_QHrkJm$UFebv%b;{9k2cC0Y zg2-bnqShDb0M1fBoH#W~{Ou8hD{L6z<6ivvW^ZrU-CHX~Xl@^qkkJM!pc1vsXlC=W z>=j;P%c~Mg2p`eEU!hSRIK@D8rF_Ymuu=7NeM6=hC0U1+N4GnoC8r|d630y4xfk>1 z)V03oToqTxk+0c`oKwmMIcQ<48r!C`8ZTIo7uDPCG1X+I`u=k{VLbgg5U}wapj2|q zsvR!pboW*QjlAow8F=ICS6HG)7lK7YNh8ELczIU!YqbJ3N6&&}$2 z77LO`A0sh3bW2?1Ln#WYh-P7HwZ8WYE4(HY@b|7*UE)hAjp`qf^;SaPOoqHwS#`+0 zn1WrrN32p)G=xT(L|ZucFF;R=YAd|i=MFw#0XB^*2w&MUvNvH!lP&`IL}eKPjNFOK z=U-s!tvJa`w^3lQ|D{kn{1F7B&8Xo zxBV4~Oe5lMBwL1Yr^h64=`AM8_2xnde(?yrA_-RarzEtO5r0h^TAJANVmhGAVqC(g zzIQu0EV){e_|sVo^3j+YCTq#^W{X9eYNq}x+UgZn(tEv2pKkSEp=<#9W*M4Z5Q3#f znF@p2y75Y4gUeGQ?@NZF(XF*e(VtdZxy>xe+*1x)#shzVpd*EIqv>}5 zKuYOnAslW*m2u5$EeIz!91ghQ6ZZNm-UQi_-xaK%l7Hg5WAlzafQ!hT#L&z!%h}Wg ze5rnmOP)^uNlxK*nuaH`qyGx2wryd*t!w4Yr@WfqF7cL-aW!lZCaAXVUKvf&4`HDY zVo>HjR-rjc3A^grocAP&CZMWO#X}EXCa#Q!)^iygE5DIA$5=?&HN#6RW8w=9YtjFn z2it(Ng%4k#W7CT}7e30KQZ2!7dmJ>8Y-jT%1GYJVM&pf;VK*}loxSHOOX=xNO`?3s z!c+qP{d&yHedGJXGTj%YVXxuP;}e!om2;ds8JRKA_etkakAuU2@s0&}w{-@cDmX=t z*DCg{PxnepL+~9unHBTghr&|1x=*GnToO~$dr=CJm{=O7obVjjy-NUUeOAD7*dI@? z{0U}d#`GRdY&&MA`pk3q+tG!_{>q5&PzpnqBFwzIri9sh( zM#!N5{==}x2|Ns$GJ0r5h=|D16=r|#P6TD(?b=Ve(MKhOFrk+D(({YJNUhUUD?K%}By|A_>S zX$4hAxkO`yp;#n9l7tXqgW1`MNe;}1q3<832WE11b4x2OW)2Te7H<=mASM+rduGdZ zx*hQP^Q*bjYC>b0yZ!0*>Z6l(=JeEVOtp8aN5VNBADMzh0Uo!yyn1>B^yl}rC5VSl zi;%!SgaG{$i=8rt5G5eAPhZDJdxW4LMC!tN^-rj`;Y-n{5KKYoF1E>PU_E%F@PJg0v^RLFm z0Bisl0LUZ|!>-4rECPjqnh(w@jSd?d841BQ4m%Ddt}a<;2c$(1 z!}h1efD~m9q5#8Qw*lhzaKR zdmDe^f{8R0NiX5vo`r*P5-a}(8_mrJNh`Yw=u9sPapG(M+GspcSqG|%;0s;x(&xhdS1GokN=+a!j|CY_aGll<^n)uR+z8shu!n1|1 z7ij93qf=TA))+L9Z_|3}i%lQnSJM58U>M`0`KwA!3hDylZg&UN_WlL} z5a0#$k1r?azv?SG_b2#M2JC)O4$&9@Ec8QzV_?*{+oHfqxf56_}hz2&d#6jd#?7w{`Pw&;DT7b!G~Vn=Hw`J zp|ewzW>KLiE(6ARO%h1OcW2@sAs)}y5VZ*Ao&BAN!W{qUN22MGZn00BA< zg%*}U%F8F{4?=H=N!4HL3&rRkf(_y-)CMTO?FLY_2f4i?MnwX-AOA)Bi3tYeeryv5 z=`XU65(eas{39|T@E5y4djN7zwnN1C7dt^KU@YGwQU;sR3ZBe&iIT@`{u3c8hv964 zKvjA}djNKSwoL%=7yCnl1ayb7L!=yP`fpJGnGQ+BguYj#e3HKn;>651O5PXkk4PYS z76?uRcf*gQPcC-s5kCA7*n4W*?sz{K>sAX`nM<_|P~k>hAlr@y#INc=Hk=Pfo@~{R zUwMl4PY@7ggw^(yb)Sz)aDu>(ABC#!*T3A<=Fh)8_O}{trFMtrScK(G%y{`_1h{V| zrcs2)-mhGtWoxfeq4%|%4Gauai(^;P<1gXYhv!>tI<9gOFPguH)Grxjh$f-m z@9=*ctqm>VO(WoZACY#uTfD^Jo4xAy(oqP+f2wQ27AC#Z7aWiLkKg2qLlJf|JuNHi z(fFnoGQC^*TSty0PboKqWkbK0?x&%RAWlLf79BIRXi`6DY`mlL${?%6h={)4(fy_y z)2C1!mK$|OG?mfgz!?QM!DEuxW)?X=-hq2vrBWOVgi;pDSy>>jmtq-NiWPjDArvb! zq5ygss+*|zI>U{h_$;8G4VlE?&W6w+6gAKAyMNU7i>2_-Ld@1{)#WgKG)G!^DZGno zJ4kTGZN7MIlQPUTgkd@h@3;eMaky1OnLvS_SW=L`@AGCkJD06C)X*J*!%=@8N3fL{Yn2i1C{VHeGMl`B zx6BlM(G+yKoim4D&FT5oO}eW}pfY}&gN+fAY>e&fe5gyW0%v18r1J1cE5;4o! z={|!eq(4APjfNH01?y3xi94)nF~hI!>foYc62+RS<e+O}=mw(WM+?y7Cuwr%tD*)Ousx1Z$fFX#JnPTpi@CY>{xzj|d{ z1NRt3hZBn7Z5)n;*g53~NESOlwpx}=JS6tEc^BGA z7m~LtQhAYAHf|p?;conrg4a?`N2|_tnq{*dFH;#8Et2G|qO+7=jlyD&I>~9}Tn^ut zZ$A+osVb}6#9fLP8;Q$M-YD-bK1iJEkd!nsKMcH?#!!>M=WC#G+a9>36!WY3Csii> zPswYI0#}dFV(mn$RHd;#$$S)xeyS8?0SgJMA-}06qdJS zGA_n2kT_b+b;xKn#>A98L85_VEGL0y3Kb@KspC;kaNT&Q!txQqh{{(Ch9jiyyZ3P+ zryN)vYx{>#)pqXQnq5FiqtwjKQ~Wc{&Tax%dcNZ@>;Ab zm_q2#?K1keWs2Qe7$W-@)ZRyUWG~bfF;p+bMv8=ux1Q?p{Hqc~?C}qdbR?__%h&Ec z`IeSFDXB(C9y>}NAls6mR$OG ziu;WY#B$nkwLXv3gCxbsCHDJ!MZFLs2#`!$$3w^Wl_4}BrJvgBn>rv<2c+<_M(RLPC0h3)z1YV+NK9M%a7r%URi)+ z7@G&0{V^|S@^L3KD#S-=r;jg!C$8Z4uQ)WQA_k#H{b!y1Ac8&MnFrwdh(7%f+a&P09&9jvBfo-V$X}wHN>n*2B%= ztzA_3MF#w|jNTiG>bX`>$}t_n)h(G#|0xtFAz8IH`OYavm7u&naE>DY!+D{9Wt~lk zmNFn{$}bX#m*pJ)@uqD?Q7LIRwqWaGqmhWkO`XSJR$$#UxN`$A5_{QMcYOM`efNN9+%nCk2lY zOct`A0x2om{+uL=+U}lHcLXaHTIt91`XXRq8BAsYGr0R1X7URBKo&V*BR>?6ItrL1 zNKF~-uUG+-D@;Hk;Ea>v>=*$Arjih;3bbN~L*Km*6b4(5Oh={5YRj@gI-W@;B7DBG zog#tVK? zcu6S7{`_4^g0+-&-$2OgXb-FU;ojk%dpv^p?m(cgpgIY{h?dGDMaQ+^G_ADi#3s$#{jaDX+)lAT3ir zusT7Pkd9GX0|UKlPWkV+$c#o?N!!47r)_jrx3_F5@8fPpt<@K!KOFUj z)zEggP*T+*L;LxH5zFyy!;!DJR}W!mH$~ubm)_a0XDYXDH7`8VEqb%s&mH`yBI2$_{XRN#giZu_#>#&7HjhSI6AZ}!vc(kKaNw?Q zz$tZ+N~a(=`U~smH@&&{mG<;FN!wv*_DMO%2;AqVUmef5=hXG?n zmuL*DQdMj}3}X~buYx}Dh>6qhf#mQRpu6iUzCUTEskZ9O7(Dx8!*g*=8&o(%!8fRs z#AW~(&?YJt@f_3`C)JVcQl=s&bl^n{Eji1e6Qy4}sI1}E?P%R`S4dN6!Wo!uOu;20 zpfQJv4@y9w8r4nQ;$q5ssa5+l9P?vUmLomih3`Mz-i`j`tBkz9;MTD;Hg$bq7@XPO zTw*aZ`VGHsYz&jV!0aBLUENzD=DMT;9Y+uR!GVaak*5<_iwz<;)ur_r+)F|)u_zN!R%~R zlJfzdg^|7C zpYcHio|;bmRF$ITvqzsWOgo4_%S;3#^oecLp=L?Z(n)!Uq3P8(#z2t%eM}HWdoE7e z2^Up9RTx&Z$t?9OA<{*RmNbCTQ7KP>pqz=|Y@L@Yw`R{vI_*j=Efc?=1f%;sFC-(0 z1{>ki-zuNNNYdcGEUYjCeJst+0@vKCnMpT}3%+$?XmWGV1Tad0=55PmxlBWBDd>{P zO0)f;S3XvLLWnS1WBw@jKE(!@T!9M~(?Y>q%F}xfJ(Myxa!gpshQ{Wak9>snvJ~G! zbv>@XDs|ZfDPHUTd|x5HF5CHgrD;9ldln7x^?Oyq0WRu-#F6_6r`t`#4ehHNy>8J~ z#i2nUjn9r2ksIDg3Ep0vTh;JHKBIZ39AoK^L-6v36WRvPgfHP8d%LIDaEKM9Sx=uZ z(z|Xx7`k4y)mh1+OeDQZhx4`2h zTl}ZtQmBV^5bjV0yE%1ff-z!whgGFNd7mgP1R*!XDhWBV9z8WSReT$ojy{I8!P zWek$AEK(?7rw3gq0E3C)@siNg@m)7?;0h{u7(@zSu$LqPhsIYixdajw*f;yKlA0i+ zKkfNROg?cF1vWwhq;8z$Tr@p)XmkjSMqd|9OZiag=#JO9gbd!IB?r!$nwzN+Xw+TC zBQm&)Zy^MjQJImd${#*Pd-6U;0I#{J>)@}JS^2AU9uGuUMCDBekG++{A#EL0?pK1( zp?PsQ!V@Ubo?IJy)KBg%&u5TYamr_vT38glYlS z2UjMfX%6m>iZ#(N6)4t+-1%5#UO!9U{xn0qy!OEHq{0(VzjCYV{Vp7a!zBeb>~*4Q z4Xx#d?xX+<*^XtrdRiY#=iTXaH2(a&))aQ`Bz`e^bE+ZTBmt_^e_8Kenp!6K!Rb^N zMy_6sH>a{eF7|ps6nkFgP zs`Q}Qy3b*6P5{MHxHA6vl?-OjzocDOqwu697-F9jE3&yZf1|fzCB^|>s*p_SoHiAN zjr?;~fu1H(_*XM?=mWTbsi8ae#sJ4;plfkb#8WjbF%s@{@Z>8)mKP+xAmbD3NFydQ zB+}ugv+Fvd#LSQnd+8l>c0*yqPc;(~RXOS^d1Z-NIuGgh=63o+X?E%S9^g3NoxSx- zAYeB!=2s4rf@IrBIKPit;YorM+(-UV6b5st9Cgz z@p@q#B6C;O2pZbzbC*A+)62I0PidG9CtPkER%o|$@`q!Y;35f^s39zFP>P#QB(05}3`zAY`wJBL6~{%%VAD9ND}36hpPM8UrED&`_=p zO}DTrkVxMJKn+-mq4@pAtdDOdzvi7}_scn9`3_m9PQ(Fn2ToWSvp=Ay&3x25)%yG3 zViSRTWG5d)Q7A4Cf+xhcy|EHCV({fUu5>Ze1EbtdOcS8(nee081Od zkgb_&{YiWHu{87u2{X~%S{P^pZJVHKmXo$Y>?nDmkbB{tTs=anuO=dxsWKK5*Z44> zPhpJJ0X&KlE%)6MY_cT%9do|PkY@=V(BQ?aY+$ON&lvto8M=ZP-3VgyJfcc;Zq)X|_m zp%`)|t-!#oh9x4)HO<(O)N+VPQ-qj4S@Biq#bvF2^R5f`=UC3Iz5#W0GwtHaSnW0? z=GFY}R1e=uWo9wKMZGJai8jugYdsv^?}T5`BylD{P}0VOcDnJ4L1L!>*+NT3?SnB1 zD!sIjScnk3{PYn^qn%@zwKsF84NjOBW0|}Pss~iYBCW+Dm%^-z$B<3*h#F8`Fk)Mz z!~^}3@^iB<>&o!j!4NV62x%OdN5bG!J?Fxi&y6!apjn@P5Bc zQHu)KQEU=Q@P~}K9Mh#o`O`66|G-qLY$RkIF6!tb<|lsw0w(E5jYr?`n!EjhK67mV zgF9J=3?0Xw&Ipn2h1>|6rLh5PApWC$g-PCAT=>fR*twQ1n|rZtv_Z65v1&^&fqjx~ zYkX`bsQ^~K*Jy^MoU)G`gHphm=i!J|uBrLmOgjPjn$fkD6w%j7+woAwZK2a6#6xp4 zz6Y@kve-Q*X*-n&L;AO{thSQIQ{`y)mX5|(MrDeHJ?UVb6XVUnM6t$l^p@3@=pQl~ zf+ay5hB*3JKE_mZA)pa!#xM*FOsMFDgSEU7p2ka3I-ZIYT1Q=yyVY*#oPN7j=3V}K z0m&iB)3I_m$_Hfq!(5ReqpQ}+D|7aF?e@J$Mvg`}Vk+oFw`jdHi^1n)Q6@!9s(UwT z^G~U!yw?#$nIn5PXEO}|*kVnGt&bOxtwB!W(B91pJoH1a#$l6@gicH)}pTZ0Bu^hfk zsKLm2)Pl1M#hsxiZ3cC{Zwi-2(M90pznNdMXn@K?Bx3s2C+~c`39^m0xL3VEVp0pO zw~&UYn!xb}DzqJRvMy3=ypXr&hmF4%j1$FQOjGzYqKos2{OBBVVk=qvW8CfM%`o@# zmYj1&xoZ6JCMm|XQjf^Fj?{*=sB=L{`TIO_(p$vHb8R2b3*tbq*+WS$jBdJQ5dc9) z!INwC?||xf`=|=hL>2lQ3@D@6*ePK)R1q&_rz4)PW*D8=Nk{`d`E<7_QSTC+M-SI! z%(DWq0AG6sKDXnIfgMJA{mz;a85qlLhpTh|rKWVL3L^D-6?d?8VXR~Biv=FQ00q-f z-L$=Nkq;#lhBY|Ucu!o}Q^)xTN*=a6t?ESy}?+N`<2va%%x z(8Bk*c0Cy0{r0s#I_}hZN^D376rAI58XR0;d2O6T-KF-_OF7&R*_{R>d8~gN9pgJU z=cR3ijuCt{W6SoWBsZCK8HyJm=F#oEfb~)GVqk6o+#s2H9a5IN73Svhp%2%z(ml@} zYfDTnCr^(Zh9Ll(7)-%bp?uwg6!OdJBb@(~4ECIj6}-P#ge+fSeQkZ{X+!3)R7ilk zVpz@M_KM#7v{=fzF&hnhP+;{l;qEB~eG8f>%7@7o0~H@}ExSlRGvm)`r64MeIZPet z@Q6=sh4B!0pub*4AN+tVtD9LWosXv?B3IhSli)UCy$FnA#rJ+U6BiG@?MQ4{^|U zvA-2!#cuBQTH+&XQL?k~!t508b?eX?^-y6Qj2x4~iI{-{zzo)MSai#vWNw%vP z>|@;@|c3( z;khXNtnW@BQTZXwn`pJBXHVrF@Yx-i#uA5D790yCb^W;@Zp>M8jB{Gzo&YA{Zl#0N zrbOc^h53(2-_vpJ_ky~gOT5MLUgm6$bBfw+A>~*3Ys4#D@WKIiWyz+TR?i{5*&(s1 z$BT2Sk&$-Vh4F3tn%_@P2Mc0^)=iK=Fqagm&)ym0dFPG(yjBESA^`KuCrNx3x%GjHjxTCU!uslk{|kY`I)AWRK7bsJV* z{30nou9xnQ;qLCscAYwcE~}`Mwcb>CD=XQJ+k$K+kIU8WvCQX4wQtTZjZf2w1H$6L zrErAx5T>3jndh;#@`J?c)MKG?P00Uc|Fr!`a$rtwi$H13OZb5R%SdZrF8 z5UbtS&k)w;H0%i6_iM7zZLQuC;^2|0mrN;EV5K~mQxwT@L}#3SOFLJ7YN%KDqArB7 zjo&Ma53Ln-8OP89`kE$N)Tp{bN3wVpJE9=wkS<)3EamnB-Z96l?w~G$Q#?&)0s2He zPuaL_(w(-zyJn^8DMK0N@e`AKyTS|mn#N$@8Gttpq~t{hF1sKVqQz8-kRQBEj=i{| ze(Fs@fQ&P)W4C60A3qZn(H%g}!c<}_v$26j%tkXGe6|PaYKc{;@7%RS_G*Ags^WsQ zjv&7qU!ce`G;{B|pzPVf=&$|iCVQJ{!Rr@nG!tgqwcIkzG!ef!z)V{Ds{MKpYE!A@6n7P|FS=Tw9o4{9ZA((qx`)Dh0WAv`xkMTlRlv-Pg}}`kf505B^a?xe!olh>n?v`$ zlg7R|JlLY3ffBhfL-iG;{V>7)s+FX|N!gjQZN=%X5T%1cwlz)(O6auN1bI_msErkN3;x#0chiBuS@t!>WL$ zOEV&6H->y+=MuFmzin{GpNYJIh+MxInYT`?WX)5xSsznV7Nr4Jp3|K60p55H%3tS9mHj;-9&p;i=20 zJKOwTna!-G{@JSb>iYV>dW5*9xWuImj?I_oLYdv9hah>=m+r*{W){Yf4U1eH@iK}v zqmxTCurO=eU};HQb-A z{Z?ot$OwfTczF&Hl9fCOl@VJ!vLVKnEVC=KEF(rGnAJc{nb(Yk^o)U$>TRJ|P^dfP z3(%2CF384F21BEIA2uQi5lP1LkC!3MAo)?xT;I1u1h({5cAnr_dpx=kzcm){ogK+h zJyc?#j6+GSfr3p(O(0*xSL?J!SF_S9lLPV)tHouV;P2~9*Q0DAAohKO)Jn)eMO-WX zI+y;tgky~LPm}T$TBP-Z4a%MKcagg8Wu;%CjHIcXqr*TCMVJ^#uS+bG^?MhZL&>_U zJBo>z793JoAg1=Q5*y_G>tP>_x_);%!(X!f3l~cA?%Em}g=FzUhs1h}C6jGDX&23M zM-hTY>KA)&c5tctW+Swf+#Q~Ix$&%XPjy`<9i^dl_<)BxQuY1`bZ5sxj#>8tx4&FZ z6xlRY*>Yq{NqO-TIpbL|qwbW?ayX)`N3PmpY&%{I*-lMc)7)Z^Uq|Q1FF=!9dd7by z8zEY-tCL5yHKpG6`$qrV+uV>D7L#V3VgCqRU>3N|Svz(St&7`1bDg*{?6|lbqyYOB zB?+Ir4YRO$cKhZkyd)C=Iv+f)i??)eH!6Nnr8kDlh1$`)Y8PWID*9>4KW(Q@PSwra zxr@-SH0bXWW9k!1RZK^DHvw|Ark{T_$q*UVK^4&AM{k2_FWv0LDEaC^&fkqDLd$X| zh9~+dt0zLbz9I|LreLVWS172ULTiEPezT1GY&1;*iOKWHUIfnoKG>u}_R8@?5B2%l zJe0@Fi_emZ3yH1}pS`Qr_@$-)>riAl9pf#<3!}kTr4vEshs-acJ=;=|>UBaiS4z;@ z#?Z^vtCLS?O=_j}z(&%|cjRi_kcMwWn1WLZb#u z2!DRy!Kk7M;V{a6QneRgW7LFVc|=3um>o2z%*x>y=t{esJ?<|&F?a6YbSM3uVE0&&+O9tPel7>rOBf9v1>be+{L}9B zPWNo4S@DD;4 z;SqxyYiaR%=I=KS@tcRZG`8v_a}Q~lQnGPyfzmxQY1B#tR@_GcH^KvvoRH!-j)Sag zojHp455M%ordOpMQLd0rS6R2;m=)Oe0n5e^N?p4JE@)rfSu zW^`*l3qazP!4SJTD4*rVncH> zjY(nihIzy$N0zc0Rl80G`GOfqFuNLK|71jTh&$VEgXn&kIu;B=1yw8?ZIHGXzxlHx zj6~ZClCqF_pg~81U_mmZ74PHh8Tq45Z1PZ|WoFqVr`I7YF)u5g=H#LrJKLC&ciTJ&<4iP{5(~0l zI)6*P(1xq{mQ*<~O#sVn_1FO@qs2l-SD|d9_G9ISM{MQIjx>T^L~KUD?Oty&V4*Vnei=Kj;5N@}B(0Woz%qr~Bj^oR$gT`A}UsH<0sz|RCV(QjyTWm~ZJ?-t;(#=vbQ7~(GQ=Z;7|ly%|KwSmRQ)1_9Ud zZGySTrGjwy2yvx6g9d8BWzl&ik>VrK$dX7qc=dFpBB1JX-SNKn*JS*|xk2=v8#m9c zaSHA+1PuH#3$ziPK0sASjtop8vXY=;X~+NqSYhAf+yYR$qPc9Gjcc3{MK*fyme2fo z(Q_)LE9I2)tNAJTawe`3@k55ofeK6<1ci2gkl{~)5Kx0Uzsk7)(0+WH%>q)Cbtr}u zwC6Yt6_k17z;}e`{2R7zy^V({FJp&VydRmJbcTgHd7zzN%N8@zgAN^|G4IibL>R`- zQ^zKjtnP)=Ee77^wPB9>jz1W)EDKm(Xur_7(-iCOB%(r(0)SE^uui^vSyMeqT3W*p zt_TZY=INu+&MEt-KD`&oRpwF=oSPLeC)Ztl=9bIe3x_K@4joz8fe*(IE*LS@4E|^6 zUC@dStcUE-nu&tHYLkwRHuvqB(bwx-tw)b2@*s}W8snk8bKm3Y^fEk0J%gS)e0!1y zY5NPpjI=2k)R*Z3@Z**4m(XZ6HS`CYlT_M(I*ze<0lCGkAEJ^lk8aBBA2MEn&wz|V zSR4B0Y4GmCqAsWHO|Nz`fw2DYUJkV0dShE)w=j|W9Z*}(D+HMTBJm)> z0G4nJ?`dTf8LHmLHEvl9CTxPFb04f#i{kk~zscKaLnk-!(X65P(#H6kaH108kCM`|Y5DSgO`@d4e1VU>#jq3!i=dBdu{u~Nk8DK~ zt2ukZ zl*$&g@#!~2V{_pPF1mDLbo^7S;Bs&7p4|p<$2Xc%CVnYhOiK2OOlnWGbT}5K6>K#X zZ(q__jE&7OeZUz|C`8pAkv`5g91Px{^Jtlz6 zc99&|J&6pDT*~9Uwx-#}Zk>MWis?);Y8W;N9io0PZM0ZtkdolCA2r6ZDy`z`?O{uT zlUs6xuPsR}z_>5No&w>kDW!bS*>4@pWP_vH@Jx2FS7mr%QKeYEnv+g>J?${@fXp3L zF=%bH*S%?&Ih#AMXTgFopu;IsDyMZ^n0|aM#0_FBxz`E4NUL9TyQgxUyn3J=y^pW5bjfG0+urBHPB- zgW)2~1Qw|W29Jhlf+WK9LK)#^AUzKR}L`bVFOC^rfl{sVLm#7QL6{VzcfTM`620^8r2 z?*f8~v9$M*>^FJ%BPFoj0`JgXh_fe`@^-UfcZ$77!eTjCD6i`!X;V?*p4N)z4*7^V zScRU}$VNi^F_`U~-iLPuw`*LbyDT#4TZdv1Q*N`j4%v>!cLgRqG90&Vvwx-%q&>6m z*N?<#J+te-KUh5yEB^lK&TY13DuKo`TmJGc5i+M(@s+Ev?kuZCr}M3F&#G6#-Rp5w zVrd0!hP&x9uMbIu_3)J-XBA1gCB|cP=nFP7J7SyG-MUQw@2Fw>8)}#t|7~iV{y`0V z$A3|SN8|^Izw#2xW~q^6-k>5X$%;rMhxivYQ_wwy8@Qp_LpFmp=fyjRLwCl%$Ba!# zPqXXL)>w);gvTbv_l>2&SI`!z#jC1;+k$hIFs?%Tn z!aOEhpbVYTm_$sA;EGK6Ejh&3Rh(9`l>XqDLI)ZeDM25hct_;7p@?p_KLv1Cop~Ys z#mz3uTb_qVmP^llDW#Rlg3bA39a~cid&5PKKm1_8LU@w%m3ga&>pCuVhWFe_(8KGd zI*A>%sK!iUqfJxLMdL#TOmb%RGQ*ULU*(vL2ffkxBO$Z1B2P%n><)$M=(HXNzpn&z zcM$_}YMbvU4lMb$ZB1lg-=9I>z8;3)2{u0gO)|7;SwUKwVm7UOe|9_+6raWoHBJ>M z(LGsJaa_P;WN=VT1QuVDK*nzTtZqP&z(12OhU1`F@}d=DQoj)1b+euyfEG)2NNXl} zu0^E1ae%YCaRdF^;HQC8D8Q(~pk6^9yC%uXBR{Cf2_Yp^=?qXhN+Q$WS z+#s{xgJIkg)UYB5eEGkQ#>sFK508Wr0xICSRw3W7=IV&M4%BZ{Xo}@h`z;@&~?tc^*T#;6GD24on^5dUJB?dDi8hWg!=NG>Gv^PVAan$7(E>JhAN` zTyTvm^&vFsG%NQLIN2ee2~0e{=2ALqy*9jS-l}KUXCCc2aHtLGa4VH7YQa@zCcfG& zs4_m;_M+}S-D)Lg+pb$jt7?zts_H+gK5Ay-?@%;Q2PkT-9xf#9b82}b*XQkCu;B{# zl7OLinE{2__;S-jxnlwN!mr6mz-s_Q)vqydj1vRGfj}=cA<{KLY$KKK zE9N1AH6LA$ZP<6p^s8>iKzvQQp*2OQc%Nl>E>yafnDW=*fb~8h zdM=#XPFd`s%>wU=g!WDox~4>8?K&R&O$+g5d7ig6OGfK;^$0;tsg9%y?_^=~c0QjK zE5xfFxOX_PCV}t)p8pmw`8OcY)BP(5tc>jcJ_!FRU`m08M}oUG|14mzA5UjDgsv-( z0GPZnIVR^l>mROKGKw2k6GWc-=M$N&4mxRKj}8vdmzTFy9C(tPTv!Sj?wQ=&-yoDP z8h?X8q0=yNt^0k&X*I`!v9^=c>~5w$+uv*XqwR&e6t?60F|w8wI4RDZ?M~LdP5aIa zFdxMV&+w8uudDY`Z#|bbi7ROjFRtrYCAwUyG-_7P zjd>5B?d<3*ACucjx1VlIlC6J!Dk-UC%-JcVdCz&L{ygv;nXT0XE2{j;kT4x^W|c>h zIB@U`qEH7bg{DrXhg{d!=r$}VrE3bRsb~A`KfsR(6@hf{tzl;VBSeCPG}D54A?lhg ze_Dk>M?4i3KUVJ1F1R#xF+>c=-Jq#TVL`cwD*Q|Wp|(Svx4Y@}mX2!cG@!o2$Wnr2 zKRpk+iFbQ9+tVKO-Y(wfFu9GzHpCHsAR81Q+#?Z^u9)UA(j@Q_qwW#Ear7S7Ft$3V zRzNR)MBpi;tF#5qklS9%-ZT#KH9;mTLQmIEox;6@;&D63T~XcJoyR19YpmD80*i4! zq~2>Bc^Ob?QEow70S-bK?iperwuDeN1U5N-?>u~p5j6L~%;)LYeMN9yeD7!;h!aPb zQnr(RH;NsH{DP?{NH6R8@grd&9#)FS7P+vgD0v>%@Y>(GB3j@2b$M41{;glP|EXWq z-}+_o|D<1knc0)uY|yuUd1W(v>lZ{$vE(b4(Ld=Ir@PlfnZwo^_zZXBU;4FFfB1@v z*GBl4ehm+G!9;40Zqwe}mB{`bGxYzzU?%o|nE4-O{ySy)KgA4XY>E2cF~jgT%rMdY zzlbx?{}N|@iLu>pRp;RCP_NE6wc!n!$!3|zgNYBE*jnGGd(6?^5v21upKHbE&ob%<7>$_^7LseEq|xWTdf@x0ckn#f--y z%BE9SK(^>-RrYm=T(yixQf9B%8CWeVa`rPWua6-Z$YmKtmw`!DwN*fs4O_2g=Vw?@ zpZJ@9AOCM%`NHW?sN=l65N$0NyH*lVb9Qoyuk+UkXWS}XjF3V3TlAWKj_)*IAK2B8 z4IFz17TY#*`Kjgne92#$yx=N~WEY)>PTH*9-%fZBdGLkzJxpPJ(U-Ix9HCe3$Mv?f zJYe!Ls8_GYwlUzcM{MblD;xuGB~UegA3ju&6O3z^1^1B0UfR(%&i=NLs|^9r7myc9 zc1*VX)JPm5U_M+|HL~m4cfQg-2wfup-6?op>W8F11OpD}6`=q4jOS8#Pld-DHTRV!j%s`DsMU$~|6s?;Iwald3 z=hIVS-XBM}Oihu=aCA*1&{*FYK3io+KBtosb{&ocCX&|Wa;1gd8tRWG$bJ7&z?Ft_ zFk(8IK%YpOAfGKM?8d-;FoK*|oT!j3&+jHqd_bGAaQ zb!d^-@w$H@s-tsAYpsXk>jv)Sz6mfth2R;g&7s8M@1SA)8)*KI$k|$%VtI#verw!M zethiLtED{=+y9%Ksr|Q{E&QvTy)zWCJg|AV>(hFvR8DkX)+u$GXYBOEt5`MgE*Tqo z$;=*hm~n%?)x5FZnaJb3z8HE~KMKxCvz6FP)HEg97^hMHnIb+!eP{V7Jz4#a+L&y; z(;|-UHnJow!LnmwgT-oV7nz>MF=U{jlLAai+*=jwBT@J5e1_wz0N6!99&M4iz?*Edrr)lnn|B^Gz z|CF=AAs^_-?6GaU9)RM2zhj2!UzuU~&;1MkPDR-%k>4ht4nFAW9r7wq-b2(N13(N# zs1`<)-BN>j!MqxRO{<{)=g+@V`$$1!B`dfv+`hV_m7KpKO*=_b#G*?(~eHZSZ53x?X&h{y8bdOOk}c!7P*S$0t1%u z*PX>rmK*|HES0Av#d81*LVUFi#k~iJukyMs6xH@OY+;YhRs*ccCCkr49aw{}F!t|x zUkRP>-^4rrv|#(PcR6WI2>KZ*k^xhGF>`ftwd3Z~X!$uIW>fVvHOUY*MPOLIx(d6| zS^Kf+wQM@^5?ncEkEEvJuFP{=pEPMI^&NCL8kMhFuTV@uL+WD-agj5tHfC z;Udxk<(!=8vQ)16Cjl#eJ};cf2w#W+*?l0RxsQ#7AG;8=6Cis_WDl(&8;c>l-8C(r zfwb9l0J{;a+plKPAP?ibDn8mi^?P0K?0k_tnx9?MsrY{5y}BL0e1>hN6+mJ@aBdBB zAYKjxdetGkh@Y&uIB}U=Yqy6yG(PuKxYpq^X|auGtvnS|I-e3i!O^0j(2=9+ znpPGYa4LH7Ui=>C0(P$R)cxR|mE1Z=OldE}v+|JE3r$ zn`T}atxV>v#*p31stuPTEo_-BgL4BPEpBfV-eRvObeY{wH&eX9oqKj--pGx$mYAiA zm;?b&oWV?gp50IKPtG&vAs1qksiU0>rbx!`^@%(8$?-gzXb{Lg{{Ac!g$)~RZ78cmXVopG0|3@@A| zR}S`F1tJ4f1X)J(1>FX4?E>5e1ljxzn44|x76D!ZR)gOSI6V)L)S!+IFyGCv4Is-+ zP^4~%K2N9F7a&q9ubJ1BNKTWC!ay-OU808{sMHSHPZeA@zb1NJldDS!W)3jjM+4Bz z3;-@Y`G@a#x8`_SS%MmA8_+d}_&#-=5iXrLm%m1B&euI}G97IG2ag4&EMY}m5#R%6 z8nYFhO1@2cQD0}ad1Y1c`bSe1P4KjOUbG^SDgbRMxi$GAp$c+Jpbgn!f=#!)`X&?x zk;HFXrub;}Mr!?C%}V;B^&0Vgh(lQdk()TxdASOIGk97y@K%cvZQovXyGRaohscg} z_C&+n^h$vjsA%#n2gw)eD^S*sLs!4lDRIAyyv|wOYQ5}{4szqd5u)IRctHBJW=O-C zaDy^Ibqv85u6z*IK1T6TlRNiu>*r6=@9F9bl$~)kKcA6w^nJz5YQ3QFtLi~26rDEA zOzNBzV3{*EFxD|IoYOT4oca3o6iLe#8;P6rTgp+BO7Zbsnob~a5p_4llMd zCbs7IE0Fg6WfjL4tqBfUZ{3)kNbH+G$@c2eb~M%#^jbqOL)!{phZwY_uv@Z8A6HI~ zU92l#o#UTRN{?MlD_@&NUvC_5yu;^TxwT&blb>)TZ@h^sU-*V!AtY~I^lr&T?GvI+ znbypoEqmmx9atM~>9Yv}S129d+q#44H?!;eK7F)b(vcbgU>a0cF1OmVXZy3XT(J^5kR?CuRy2U6eE z!28hrkQutpzz5#=?C9Q$t!soXzTqGBFrLl=f3wG{H7(@37_jz2t`5B4)AyCBN=;XL zH#HZ#^*A~Pe-iD9lQSmDLJhu_(w*B6UJlYViZxdX@G8+3%2X+`^o@R$Q2QeC3Hee3 zssh0JK4Agi@@-RN0-Tm6H@N~>0AK-O14zNhs+9);^1)C6!vZW2>X!)x^TYHpbGar` z44}A6r0wdf;pgu{M`5B0s19iGjq9^XbVPx`K|wa^=Yj!np9A@70NZZwV?zbfz(&AY zHOw;r;`%-hBDWPKicAv>|>V11q!Fsj0Xj#&O$B1jR!lM`pW5&`l^%|@Cjk%e5jRdMH~f5^odi9;Z9*n@N=HSEerYFBWiDHd44P-e*rs$oyh$i)PFbc!2G`#ssES06UBeb zJ2X;!%V}(?HIe9VqQ5daM#1aa3qMmXZ$|GaZ_G>Yc59PX^C#ybTy~E-L61ubP6?mO z)_=@9+>Nm8)ioql7H#Z=bv#({ z)!i~J3QYc29pw4D4g!0B!Q&w2oMEVXKlUiN!SauKyB2{>Sv|L3Bz~$+NzKczOIcPW zpPjzPZW{Q%8c*8F&(t!PJh`rK?2(W4X0TSVMBi@867uKkS2SGDdttSu4(qp@Ca15h zX186M`ho5BxKB`c*y3b;9&L=&zBFN0r~X4tE%@6<8*2rdD6P6Icr0UWJB`KUDuCA% zzb*z8|L{sqFY~`3@`#80RR)PL0)e2R2h13i)jvTjv$Ye2Km!*e(!{oqY-SVHUyf8T z?GLhtfZ@JBbr0WTrdPcUlY22)(rI>R@MnOHD%?ltkH78qr3{DZhYEKH73A;}#FmTn zq9HY)zD^X-mqA1jx(;wUU7}Pq@y_by_yXtE3tz+D1~G?+ycF;SqfvtXLqv6rRN|Kb z+F!z76v_|709^H-ZQEPSUE&1-u!b9G`g<^TQ-l(!z+9c!DxjMJav@1+bO`#BZim9$ zC9qHVSw>@cBohek1OH|{Y$UKt!5KA7bCj&YT;-V!jBO0<3P2YWL=q6m)H!Z|rkbA} z(~P@;;MR7+ZrwwHQ@Jvd?>AA*q}jGp>aM(ICm#h#PYK$tyeHFdVj`O|9FN|!t8Zcn zAxzgZXcwVV?Ng54XN+3wujAJmIT$w!Cc@qiX>yw;#yeid0^coLx16-)w{fbDsqybt zuIUw{!9TZb(|X^i4mM3nuX!X2pVV5LJ29j|zBfwN>Go<5>3vs3Sjp|6_W%NCR9GK5 zO4}L(DBgEMb3PSxALK!N0Ew)sVgC-FzZraBVP*LDtB8s{*5Aop#~sS1VBurzAPZnT zMYthWto?GM`S-XGl5MMK7(XBE$3Jqrwo|}+^sY3wJ<`L=p)>2E(??yC8$uHsXH!}) zBa_3U&WP`BRby*rC00&lr4@G;tju|XO?ej1O2^8f#!@6)cDvKUiTqmF1Gm^(%pd~j zYt5U>n}>4NklgUf&TQu3-~?OjR>sY~*j6?po}o?vbb|Max2gJ5e?Q^TIyj_C#v`-= z^vpX&&;+U7(BQ&{5*(**tEi|l>w@b?uZz>;6W~bOlwSQ&C1>M|`^2Hmr!D%M%=|O_ z#CNmmG3rhCNvH57f1~hB8uhztYsFI0D55hn_0`U+!Sdxih;{T=5plQaM{8MQqhM3D zWtR&ci@HOn-dXMX_nw-sEMojU+su-*;#=Z&{UpoZ4am4-$Uy4`c)ooF(0+w*1>M7t z&;{M!_wmX#V78pfa&Ag>9!$1$l5Sj4-kAQ%>j1~#LteRHRe`lZGy$Xqz+u^7(9bkE z4XXUp1@xt&TtN|AAkbgjQi*}F2W5vL{v)m95Q4S?`W*X+t zVgTnR&@AXpR|}==9*d2@o}Y5qC+i12M#E$#%2h*;Oc{oIg|QFsj`I!hKBwgG(i9y| znpUyj8uSu3pE5d3TloD72*$Qe zrN4H#l|$Ljrx+2L7!jc!7NH&y!AXd5=jZvq8oTbGCX{Y3pb)AcHl&4Kf&>Byh)9vp zLq`aq2_%pJ1}=maniK))P1;3JKt+T|6$>befJiUW5$PQ%p#%cH_}zDB?#%mU?%Ub@ zRV2pX04LM|; zei7aGUk*IvpD%!>K3-@)7aAQ-(i`n+i9ueKmsOOLg+QcOz*;mu9Knm$LsS1P;=J&1 z8fy+H30Hy0K@{bbJh&qu@Dt`eInXrk2`B zSsg!5PiH4Q{x=p&EY255tAAmV0a{^vd}&Jq6=dZV<)POgFsQ;+h%D@Hu>XMwMBw}} zKpH9t7>)4={r`6kn5=@V;%^hF->g3j9_{7q&LXe$&rPum z^2PwcCeH2_ggY!?W8m)`oH{rkUjk6!+O@yaaRMPQ+MxfGf9DjgqilS+_0NS3zbtK= z9MSM|#^JGPObB9WwSlE`XDO*Oj2 z8X>pLZo%P@g<}A>H8eT&lvL6I0?QLL;7{SFg#k|+sx|b0is#~S74{~K&9$*QjBSav zEOAYV7xLOAbGCHoALra>Jb*Wso5@M2xY{>8Zt{$~xX0CrLxa!z8kZHEi+S_twj154 zy3|gEpOJ2No0%$d?-c+*e5~d{t~RC)kmkrVj;5;H{GLGM<|i^6nMvc>1y(Nq6d414 zDy|d*;=pHf;2tTko5d77TttowCAd`rQ=1C(Rw_*bcw; z=zOw!Z<+b)Pj@%YOKaqC5(Ue%LPshl!`4CpUw7FX8tX#a@|8jvtolRIo(T0^>6?Yd zvnoi};L*_0Yu)(x#EK-!ArRc4JqDW6eO7t*8hp;B zAv#PVL!e2?o?Lo)+Du(h3`zJ@%UPO=m~P@$O=65t7n7 zmieJ#vI=72o4bR|Fegh@x>0?lND>O_ukBHFNLtJ-A2qcl*@5~ZpS}`M9a)7~&#s$~ z2fn_Y8X+$<7gqnl)>B?cLk_Dq2)ki+6gIgsXE%mb?YAgfa*w)yoKH)JlZsTI|9o@S zv?rQCpnI9`4e`3C7uk8e-?PY?7Eyw=Bt|x;zvZ6}9qX*p$P$%|m!gW>geC?GXXIC+ z78Pf;RhpypYh0wwozqjhy_xROXW4mQ{HiL{6m+HPoKZYnWhtlIM~2z5EE$n^Y5Bl9 zx8$Pqh?x7tvq2rHG^mu&ylCO~a~); zX$Wnk53n_R!J{%7Kh_xxjk!noIO{fcgmtSE5*w7`*S<^q}z98)|cKlU(d%TJr?rLK!pnNLfYb`yW$!_ z1E|shtwWt}&n|S>C|dE-O`mJLBJvt;VXI89oCW#9x)Jze8_vrcRxPH6S;Zi(Tcm%= zBKF#)nqL*CV@gXVEt~F2KeCCATxvbkSj}$iTx!eV;M?ZTM zKHn9}t*gFwa!}mHq)w7uw!!%2+bj?k~qCX{7@>cdS1hpo@i;R}V+Q-G7u7YB>WyNQK6tYKTTD@A#m zT!K4>YSS5gB{LRk_x9o)+1HuO(@V` ztsiy-sfF&aW2NQAc0ftGrou;u-4p7L!{PVwsbIChO!AG9>>1aj7jai&-7SJ-vbtVf z2?ur=ES?xYRG$x~+6++~AMIxAueMR`wN{@EbaV(Uhxx$r9c3O&nNq?j+2wo7>tiIp z#KYr16(=lTPQn`{j=(7e92$+fNA?@#7%P(L21mfqz9p&V9==Y-&v2~lqSoL|D6dgZ zQ%(0{4>g_1h(ROVv3URcso-BTlod`bNzS32K~HvQdQqH=d%y3Z|R z>semNkaWmKH)^zk)H>`N_x$DXfVSTrF5mT=;I$u{33^|Bvqk799&Y)-Np6jjO7d$7 zN+~HRQ=-j>9L?@NgV-ONu&wVgb^I}PZxzcW*=+^E402;Qr}7?jY#+o+>|utFlm*lU ziqAYaYvNzn7#t%F;J@PSS4!M63>klCJmtU>G+D_zKR%hU-?Qf!VfpebQ&jgXTL$Ol z%i(VW`YuH3k8-l9XGcmsK`|?R7h%3(*D9m4_L2RgZ7?ZH@+pJQo&Xx7|NUcq8--)*5&TKSSDj;B(J0r#;E%) z#XmVc$6lVLfbZ%_$SKVI{Vi*~QT`_$+mcmgLpO*=fT zS{wz>pHQ9ZjT-k~WE%_!)OHdZI^R|tjfZ9|JrxI8EavgE@Rr*=nx!(Z+l8Y_Ox8~K z1;p&I=njGV*rVjBwyM>0COS0~`U|b5(8F_-(wM4r6%^6BDmbZujJJt#Gcm5BQ~J!kQ*YN1^XwSF6?b@w|2Gj`9jj zSvxMMCb{(vmvcp4Df=SH@(v3;jKA;V^S7aK^;~bRtCo^iDwCID;K$B^4qekr1r9Dk zy-`1T1da}-iu<>t`bP6Z^T?v-#j~~pnhShdvWj)=KeRnJ%zS`>QFtiY`5n$uiJy93C zHqku!+|8ao(?~4UHzQ9$#lS_zwXrwfwrbhVwGrh5glez6f$gmisufMp?@%W&y zU?02d%4@?F zv~(bf+L|;;3!(=6cN5w=2n^nhfCa)Jf91CS|HM|s1`PFXnO$z#@_HBy*)1IoaCdii zb@vbc)8^?DOm4jlsKZ*VOcNPBSOO^4)PqOB041E!!qXZ_r3P-5Qv90Q2Yv^5QhoO=&LwvzQ!%eU*YMAJ_r*UPiYfp5@a9{?GmNOHo(XQ^mp za1_(QZkCRK! zgAfo{5H>9pr0hWFS#9uK>1yPS4pWkFOXYK}J1^~w3~s-)yDBB}r1FizS7-?zY^vDK zQ2Msu2heaDU~ScCp0QeB4^*;lnDgfsoslUJs`{QMqcdx!t-e^QGHF7m?Cu(V-rt-( z6P^i0GjYCow7xr2hn6rsb};l$Q+c#?zq(+JG4JR>upUD%s?1X~IAZVod|36_rC;Ui zhPh07+;WwHduIJ?b_JFK!b}k!$=h-`I=rgYt8|ejZ_A)WzB4)(dmI)gbdnin7i;H4 zuhFT#EnWkzHw}1wmF;KlY_KR0o@`3s!GM?P76|EK*J@J83hem8Ns6mGYCJFf2bEeZm1x>M!`%+JC5tuC z#4Z&*wTy0EfIbmYyzoqlAFpg;!8oyE_;Gaz!8mvfD0Kv2WmvH*EOE)FYdaZhjhSjE z*_V|0dv82uKSUsETu$0h#Xo8?Pz@*xEJ(5@2Xnw-2cj73aL J+eC-ue*mHRm_q;n literal 0 HcmV?d00001