Histogram improvements

holoviews/operation/element.py

@@ -2,6 +2,7 @@
 Collection of either extremely generic or simple Operation
 examples.
 """
+from __future__ import division
 
 import numpy as np
 
@@ -11,7 +12,8 @@
 from ..core import (Operation, NdOverlay, Overlay, GridMatrix,
                     HoloMap, Dataset, Element, Collator, Dimension)
 from ..core.data import ArrayInterface, DictInterface
-from ..core.util import find_minmax, group_sanitizer, label_sanitizer, pd, basestring
+from ..core.util import (find_minmax, group_sanitizer, label_sanitizer, pd,
+                         basestring, datetime_types)
 from ..element.chart import Histogram, Scatter
 from ..element.raster import Raster, Image, RGB, QuadMesh
 from ..element.path import Contours, Polygons
@@ -482,6 +484,9 @@ class histogram(Operation):
     dimension = param.String(default=None, doc="""
       Along which dimension of the Element to compute the histogram.""")
 
+    frequency_label = param.String(default='{dim} Frequency', doc="""
+      Format string defining the label of the frequency dimension of the Histogram.""")
+
     groupby = param.ClassSelector(default=None, class_=(basestring, Dimension), doc="""
       Defines a dimension to group the Histogram returning an NdOverlay of Histograms.""")
 
@@ -494,8 +499,14 @@ class histogram(Operation):
     mean_weighted = param.Boolean(default=False, doc="""
       Whether the weighted frequencies are averaged.""")
 
-    normed = param.Boolean(default=True, doc="""
-      Whether the histogram frequencies are normalized.""")
+    normed = param.ObjectSelector(default=True,
+                                  objects=[True, False, 'integral', 'height'],
+                                  doc="""
+      Controls normalization behavior.  If `True` or `'integral'`, then
+      `density=True` is passed to np.histogram, and the distribution
+      is normalized such that the integral is unity.  If `False`,
+      then the frequencies will be raw counts. If `'height'`, then the
+      frequencies are normalized such that the max bin height is unity.""")
 
     nonzero = param.Boolean(default=False, doc="""
       Whether to use only nonzero values when computing the histogram""")
@@ -531,16 +542,12 @@ def _process(self, view, key=None):
                 weights = weights[mask]
         else:
             weights = None
-        try:
-            hist_range = find_minmax((np.nanmin(data), np.nanmax(data)), (0, -float('inf')))\
-                         if self.p.bin_range is None else self.p.bin_range
-        except ValueError:
-            hist_range = (0, 1)
 
+        data = data[np.isfinite(data)]
+        hist_range = self.p.bin_range or view.range(selected_dim)
         # Avoids range issues including zero bin range and empty bins
-        if hist_range == (0, 0):
+        if hist_range == (0, 0) or any(not np.isfinite(r) for r in hist_range):
             hist_range = (0, 1)
-        data = data[np.invert(np.isnan(data))]
         if self.p.log:
             bin_min = max([abs(hist_range[0]), data[data>0].min()])
             edges = np.logspace(np.log10(bin_min), np.log10(hist_range[1]),
@@ -549,29 +556,39 @@ def _process(self, view, key=None):
             edges = np.linspace(hist_range[0], hist_range[1], self.p.num_bins + 1)
         normed = False if self.p.mean_weighted and self.p.weight_dimension else self.p.normed
 
-        data = data[np.isfinite(data)]
         if len(data):
-            hist, edges = np.histogram(data, normed=normed, range=hist_range,
-                                       weights=weights, bins=edges)
-            if not normed and self.p.weight_dimension and self.p.mean_weighted:
-                hist_mean, _ = np.histogram(data, normed=normed,
-                                            range=hist_range, bins=self.p.num_bins)
-                hist /= hist_mean
+            if normed:
+                # This covers True, 'height', 'integral'
+                hist, edges = np.histogram(data, density=True, range=hist_range,
+                                           weights=weights, bins=edges)
+                if normed=='height':
+                    hist /= hist.max()
+            else:
+                hist, edges = np.histogram(data, normed=normed, range=hist_range,
+                                           weights=weights, bins=edges)
+                if self.p.weight_dimension and self.p.mean_weighted:
+                    hist_mean, _ = np.histogram(data, density=False, range=hist_range,
+                                                bins=self.p.num_bins)
+                    hist /= hist_mean
         else:
             hist = np.zeros(self.p.num_bins)
         hist[np.isnan(hist)] = 0
 
         params = {}
         if self.p.weight_dimension:
             params['vdims'] = [view.get_dimension(self.p.weight_dimension)]
+        else:
+            label = self.p.frequency_label.format(dim=selected_dim)
+            params['vdims'] = [Dimension('{}_frequency'.format(selected_dim), 
+                                         label=label)]
+
         if view.group != view.__class__.__name__:
             params['group'] = view.group
 
         return Histogram((hist, edges), kdims=[view.get_dimension(selected_dim)],
                          label=view.label, **params)
 
 
-
 class decimate(Operation):
     """
     Decimates any column based Element to a specified number of random
@@ -765,8 +782,10 @@ def _process(self, p, element, ranges={}):
             el_data = element.data
 
         # Get dimensions to plot against each other
+        types = (str, basestring, np.str_, np.object_)+datetime_types
         dims = [d for d in element.dimensions()
-                if _is_number(element.range(d)[0])]
+                if _is_number(element.range(d)[0]) and
+                not issubclass(element.get_dimension_type(d), types)]
         permuted_dims = [(d1, d2) for d1 in dims
                          for d2 in dims[::-1]]
 
@@ -787,10 +806,7 @@ def _process(self, p, element, ranges={}):
                         el = p.diagonal_type(values, kdims=[d1])
                 elif p.diagonal_operation is histogram or isinstance(p.diagonal_operation, histogram):
                     bin_range = ranges.get(d1.name, element.range(d1))
-                    opts = dict(axiswise=True, framewise=True)
-                    el = p.diagonal_operation(element,
-                                              dimension=d1.name,
-                                              bin_range=bin_range).opts(norm=opts)
+                    el = p.diagonal_operation(element, dimension=d1.name, bin_range=bin_range)
                 else:
                     el = p.diagonal_operation(element, dimension=d1.name)
             else:

holoviews/plotting/bokeh/chart.py

@@ -376,7 +376,8 @@ def get_data(self, element, ranges, style):
             data = dict(top=element.values, left=element.edges[:-1],
                         right=element.edges[1:])
 
-        color_dims = self.adjoined.traverse(lambda x: x.handles.get('color_dim'))
+        color_dims = [d for d in self.adjoined.traverse(lambda x: x.handles.get('color_dim'))
+                      if d is not None]
         dim = color_dims[0] if color_dims else None
         cmapper = self._get_colormapper(dim, element, {}, {})
         if cmapper and dim in element.dimensions():

holoviews/plotting/util.py

@@ -254,19 +254,19 @@ def get_sideplot_ranges(plot, element, main, ranges):
     dictionary of ranges.
     """
     key = plot.current_key
-    dims = element.dimensions(label=True)
-    dim = dims[1] if dims[1] != 'Frequency' else dims[0]
+    dims = element.dimensions()
+    dim = dims[0] if 'frequency' in dims[1].name else dims[1]
     range_item = main
     if isinstance(main, HoloMap):
         if issubclass(main.type, CompositeOverlay):
             range_item = [hm for hm in main.split_overlays()[1]
-                          if dim in hm.dimensions('all', label=True)][0]
+                          if dim in hm.dimensions('all')][0]
     else:
         range_item = HoloMap({0: main}, kdims=['Frame'])
         ranges = match_spec(range_item.last, ranges)
 
-    if dim in ranges:
-        main_range = ranges[dim]
+    if dim.name in ranges:
+        main_range = ranges[dim.name]
     else:
         framewise = plot.lookup_options(range_item.last, 'norm').options.get('framewise')
         if framewise and range_item.get(key, False):
@@ -279,7 +279,7 @@ def get_sideplot_ranges(plot, element, main, ranges):
         range_item = range_item.last
     if isinstance(range_item, CompositeOverlay):
         range_item = [ov for ov in range_item
-                      if dim in ov.dimensions('all', label=True)][0]
+                      if dim in ov.dimensions('all')][0]
     return range_item, main_range, dim
 
 

tests/testoperation.py

@@ -74,18 +74,42 @@ def test_image_contours_filled(self):
     def test_points_histogram(self):
         points = Points([float(i) for i in range(10)])
         op_hist = histogram(points, num_bins=3)
+
+        # Make sure that the name and label are as desired
+        op_freq_dim = op_hist.get_dimension('x_frequency')
+        self.assertEqual(op_freq_dim.label, 'x Frequency')
+
+        # Because the operation labels are now different from the 
+        #  default Element label, change back before comparing.
+        op_hist = op_hist.redim(x_frequency='Frequency')
         hist = Histogram(([0.1, 0.1, 0.133333], [0, 3, 6, 9]))
         self.assertEqual(op_hist, hist)
 
     def test_points_histogram_bin_range(self):
         points = Points([float(i) for i in range(10)])
         op_hist = histogram(points, num_bins=3, bin_range=(0, 3))
+
+        # Make sure that the name and label are as desired
+        op_freq_dim = op_hist.get_dimension('x_frequency')
+        self.assertEqual(op_freq_dim.label, 'x Frequency')
+
+        # Because the operation labels are now different from the 
+        #  default Element label, change back before comparing.
+        op_hist = op_hist.redim(x_frequency='Frequency')
         hist = Histogram(([0.25, 0.25, 0.5], [0., 1., 2., 3.]))
         self.assertEqual(op_hist, hist)
 
     def test_points_histogram_not_normed(self):
         points = Points([float(i) for i in range(10)])
         op_hist = histogram(points, num_bins=3, normed=False)
+
+        # Make sure that the name and label are as desired
+        op_freq_dim = op_hist.get_dimension('x_frequency')
+        self.assertEqual(op_freq_dim.label, 'x Frequency')
+
+        # Because the operation labels are now different from the 
+        #  default Element label, change back before comparing.
+        op_hist = op_hist.redim(x_frequency='Frequency')
         hist = Histogram(([3, 3, 4], [0, 3, 6, 9]))
         self.assertEqual(op_hist, hist)