added units flag scatter

.github/workflows/full_test.yml

@@ -16,7 +16,7 @@ jobs:
     strategy:
       matrix:
         os: [ubuntu-latest, windows-latest]
-        python-version: ["3.7", "3.8", "3.9", "3.10"]      
+        python-version: ["3.8", "3.9", "3.10"]      
 
     steps:
     - uses: actions/checkout@v2

fmskill/comparison.py

@@ -839,7 +839,7 @@ def scatter(
         df: pd.DataFrame = None,
         binsize: float = None,
         nbins: int = None,
-        skill_table: bool = False,
+        skill_table: Union[str, List[str], bool] = None,
         **kwargs,
     ):
         """Scatter plot showing compared data: observation vs modelled
@@ -902,9 +902,10 @@ def scatter(
             by default None
         df : pd.dataframe, optional
             show user-provided data instead of the comparers own data, by default None
-        skill_table : bool, optional
-            calculates the main skills (bias, rmse, si, r2, etc) and adds a box at
-            the right of the scatter plot, by default False
+        skill_table : str, List[str], bool, optional
+            list of fmskill.metrics or boolean, if True then by default [bias, rmse, urmse, mae, cc, si, r2].
+            This kword adds a box at the right of the scatter plot, 
+            by default False 
         kwargs
 
         Examples
@@ -952,6 +953,27 @@ def scatter(
 
         if title is None:
             title = f"{self.mod_names[mod_id]} vs {self.name}"
+
+        if skill_table != None:
+            # Calculate Skill if it was requested to add as table on the right of plot
+            if skill_table==True:
+                skill_df = self.skill(df=df,model=model,observation=observation,variable=variable)  
+            elif type(skill_table)==list:
+                skill_df = self.skill(df=df,metrics=skill_table,model=model,observation=observation,variable=variable)  
+            stats_with_units=["bias", "rmse", "urmse", "mae","max_error"]
+            # Check for units
+            try:
+                units=unit_text.split('[')[1].split(']')[0]
+            except:
+                #     Dimensionless
+                units=''
+            if skill_table==False:
+                skill_df=None
+                units=None
+        else:
+            # skill_table is None
+            skill_df=None
+            units=None
 
         scatter(
             x=x,
@@ -969,41 +991,12 @@ def scatter(
             title=title,
             xlabel=xlabel,
             ylabel=ylabel,
+            skill_df= skill_df,
+            units=units,
             binsize=binsize,
             nbins=nbins,
             **kwargs,
         )
-        if skill_table:
-            # Calculate Skill if it was requested to add as table on the right of plot
-            skill_df = self.skill(
-                metrics=["bias", "rmse", "urmse", "mae", "cc", "si", "r2"], df=df
-            )  # df is filtered to matching subset
-            lines = []
-
-            max_str_len = skill_df.df.columns.str.len().max()
-
-            for col in skill_df.df.columns:
-                if col == "model":
-                    continue
-                lines.append(
-                    f"{col.ljust(max_str_len)} {np.round(skill_df.df[col].values[0],3)}"
-                )
-
-            text_ = "\n".join(lines)
-
-            plt.gcf().text(
-                0.97,
-                0.6,
-                text_,
-                bbox={
-                    "facecolor": "blue",
-                    "edgecolor": "k",
-                    "boxstyle": "round",
-                    "alpha": 0.05,
-                },
-                fontsize=12,
-                family="monospace",
-            )
 
     def taylor(
         self,

fmskill/observation.py

@@ -45,7 +45,7 @@ class Observation:
     # matplotlib: red=#d62728
 
     def __init__(
-        self, name: str = None, df=None, itemInfo=None, variable_name: str = None
+        self, name: str = None, df=None, itemInfo=None, variable_name: str = None, override_units: str = None,
     ):
         self.color = "#d62728"
 
@@ -67,7 +67,7 @@ def __init__(
         if variable_name is None:
             variable_name = self.itemInfo.type.name
         self.variable_name = variable_name
-
+        self.override_units= override_units
     @property
     def time(self) -> pd.DatetimeIndex:
         "Time index"
@@ -99,12 +99,17 @@ def filename(self):
         return self._filename
 
     def _unit_text(self):
+        override_units=self.override_units
         if self.itemInfo is None:
             return ""
         txt = f"{self.itemInfo.type.display_name}"
         if self.itemInfo.type != mikeio.EUMType.Undefined:
-            unit = self.itemInfo.unit.display_name
-            txt = f"{txt} [{unit_display_name(unit)}]"
+            if override_units==None:
+                unit = self.itemInfo.unit.display_name
+                txt = f"{txt} [{unit_display_name(unit)}]"
+            else:
+                unit = override_units
+                txt = f"{txt} [{override_units}]"
         return txt
 
     def hist(self, bins=100, title=None, color=None, **kwargs):
@@ -162,6 +167,8 @@ class PointObservation(Observation):
         user-defined name for easy identification in plots etc, by default file basename
     variable_name : str, optional
         user-defined variable name in case of multiple variables, by default eumType name
+    units : str, optional
+        user-defined units name in case user wants to override eumUnits 
 
     Examples
     --------
@@ -191,6 +198,7 @@ def __init__(
         z: float = None,
         name: str = None,
         variable_name: str = None,
+        units: str = None,
     ):
 
         self.x = x
@@ -254,7 +262,7 @@ def __init__(
             )
 
         super().__init__(
-            name=name, df=df, itemInfo=itemInfo, variable_name=variable_name
+            name=name, df=df, itemInfo=itemInfo, variable_name=variable_name, override_units=units, 
         )
 
     def __repr__(self):
@@ -333,6 +341,8 @@ class TrackObservation(Observation):
         item name or index of y-coordinate, by default 1
     offset_duplicates : float, optional
         in case of duplicate timestamps, add this many seconds to consecutive duplicate entries, by default 0.001
+    units : str, optional
+        user-defined units name in case user wants to override eumUnits 
 
 
     Examples
@@ -408,6 +418,7 @@ def __init__(
         x_item=0,
         y_item=1,
         offset_duplicates: float = 0.001,
+        units: str = None,
     ):
 
         self._filename = None
@@ -446,7 +457,7 @@ def __init__(
             df.index = make_unique_index(df.index, offset_duplicates=offset_duplicates)
 
         super().__init__(
-            name=name, df=df, itemInfo=itemInfo, variable_name=variable_name
+            name=name, df=df, itemInfo=itemInfo, variable_name=variable_name,override_units=units,
         )
 
     @staticmethod
@@ -494,6 +505,6 @@ def unit_display_name(name: str) -> str:
     m
     """
 
-    res = name.replace("meter", "m").replace("_per_", "/").replace("sec", "s")
+    res = name.replace("meter", "m").replace("_per_", "/").replace("second", "s").replace("sec", "s")
 
     return res

fmskill/plot.py

@@ -30,6 +30,8 @@ def scatter(
     title: str = "",
     xlabel: str = "",
     ylabel: str = "",
+    skill_df: object = None,
+    units: str = "",
     binsize: float = None,
     nbins: int = None,
     **kwargs,
@@ -82,6 +84,10 @@ def scatter(
         x-label text on plot, by default None
     ylabel : str, optional
         y-label text on plot, by default None
+    skill_df : dataframe, optional
+        dataframe with skill (stats) results to be added to plot, by default None
+    units : str, optional
+        user default units to override default units, eg 'metre', by default None
     kwargs
     """
     if show_hist==None and show_density==None:
@@ -268,10 +274,18 @@ def scatter(
         plt.grid(
             which="both", axis="both", linestyle=":", linewidth="0.2", color="grey"
         )
+        max_cbar=None
         if (show_hist or (show_density and show_points)):
             cbar = plt.colorbar(fraction=0.046, pad=0.04)
-            cbar.set_label("# points")
+            ticks = cbar.ax.get_yticks()
+            max_cbar=ticks[-1]
+            cbar.set_label("# points")        
+
         plt.title(title)
+        # Add skill table
+        if skill_df != None:
+            _plot_summary_table(skill_df,units,max_cbar=max_cbar)
+
 
     elif backend == "plotly":  # pragma: no cover
         import plotly.graph_objects as go
@@ -496,3 +510,59 @@ def _scatter_density(x, y, binsize: float = 0.1, method: str = "linear"):
     Z_grid[(Z_grid < 0)] = 0
 
     return Z_grid
+
+def _plot_summary_table(skill_df,units,max_cbar):
+    stats_with_units=["bias", "rmse", "urmse", "mae"]
+    max_str_len = skill_df.columns.str.len().max()
+    lines = []
+    if len(skill_df)>1:
+        raise Exception('''`skill_table` kword can only be used for comparisons between 1 model and 1 measurement. 
+        Please add `model`, `variable` and `observation` kwords where required''')
+
+    for col in skill_df.columns:
+        if col == "model" or col == "variable":
+            continue
+        if col in stats_with_units:
+            #if statistic has dimensions, then add units
+            item_unit=units
+        else:
+            #else, add empty space (for fomatting)
+            item_unit=' '
+        if col=="n":
+            # Number of samples, integer, else, 2 decimals
+            decimals=f'.{0}f'
+        else:
+            decimals=f'.{2}f'
+        lines.append(
+            f"{(col.ljust(max_str_len)).upper()} {np.round(skill_df[col].values[0],2):{decimals}} {item_unit}"
+                )
+
+    text_ = "\n".join(lines)
+
+    if max_cbar==None:
+        x=0.93
+    elif max_cbar<1e3:
+        x=0.99
+    elif max_cbar<1e4:
+        x=1.01
+    elif max_cbar<1e5:
+        x=1.03
+    elif max_cbar<1e6:
+        x=1.05
+    else:
+        #When more than 1e6 samples, matplotlib changes to scientific notation
+        x=0.97
+
+    plt.gcf().text(
+                x,
+                0.6,
+                text_,
+                bbox={
+                    "facecolor": "blue",
+                    "edgecolor": "k",
+                    "boxstyle": "round",
+                    "alpha": 0.05,
+                },
+                fontsize=12,
+                family="monospace",
+            )

tests/test_eum.py

@@ -5,3 +5,4 @@ def test_units_display_name():
 
     assert unit_display_name("meter") == "m"
     assert unit_display_name("meter_per_sec") == "m/s"
+    assert unit_display_name("second") == "s"

tests/test_multimodelcompare.py

@@ -275,6 +275,7 @@ def test_mm_scatter(cc):
     cc.scatter(model="SW_2", show_points=0.75)
     cc.scatter(model="SW_2", show_density=True)
     cc.scatter(model="SW_2", show_points=0.75, show_density=True)
+    cc.scatter(model="SW_2", observation='HKNA',skill_table=True)
     # cc.scatter(model="SW_2", binsize=0.5, backend="plotly")
     assert True
     plt.close("all")

tests/test_multivariable_compare.py

@@ -119,6 +119,7 @@ def test_mv_mm_mean_skill(cc):
 def test_mv_mm_scatter(cc):
     cc.scatter(model="SW_1", variable="Wind_speed")
     cc.scatter(model="SW_1", variable="Wind_speed", show_density=True)
+    cc.scatter(model="SW_1", variable="Wind_speed", observation = 'F16_wind', skill_table=True)
     assert True
     plt.close("all")
 

tests/test_pointobservation.py

@@ -1,6 +1,7 @@
 import pandas as pd
 import pytest
 import sys
+import mikeio
 
 from fmskill.observation import PointObservation
 
@@ -17,6 +18,18 @@ def test_from_dfs0(klagshamn):
     o2 = PointObservation(klagshamn, item="Water Level", x=366844, y=6154291)
     assert o1.n_points == o2.n_points
 
+    o3 = PointObservation(klagshamn, item="Water Level", x=366844, y=6154291, units='meter')
+    assert o3.override_units == o2.itemInfo.unit.name
+
+    o4 = PointObservation(klagshamn, item="Water Level", x=366844, y=6154291)
+    assert o4.override_units == None
+
+    o5 = PointObservation(klagshamn, item="Water Level", x=366844, y=6154291)
+    assert o5._unit_text() == 'Water Level [m]'
+
+    o6 = PointObservation(klagshamn, item="Water Level", x=366844, y=6154291, units='inches')
+    assert o6._unit_text() == "Water Level [inches]"
+
 
 def test_from_df(klagshamn):
     o1 = PointObservation(klagshamn, item=0, x=366844, y=6154291, name="Klagshamn1")
@@ -35,6 +48,13 @@ def test_from_df(klagshamn):
     o3 = PointObservation(s, x=366844, y=6154291, name="Klagshamn3")
     assert o1.n_points == o3.n_points
 
+    o4 = PointObservation(df, item="Water Level", x=366844, y=6154291, units='metre')
+    assert o4.override_units == 'metre'
+
+    o5 = PointObservation(df, item="Water Level", x=366844, y=6154291, units='inches')
+    o5.itemInfo = mikeio.ItemInfo(mikeio.EUMType.Water_Level) 
+    assert o5._unit_text() == "Water Level [inches]"
+
 
 @pytest.mark.skipif("shapely" not in sys.modules, reason="requires the shapely")
 def test_coordinates(klagshamn):

tests/test_trackobservation.py

@@ -16,6 +16,11 @@ def test_read(c2):
     assert len(o1.x) == o1.n_points
     assert o1.name == "c2"
     assert pytest.approx(o1.values.max()) == 17.67
+    assert o1.override_units == None
+    o2 = TrackObservation(c2, item=2, name="c2",units='inches/hour')
+    assert o2.override_units == 'inches/hour'
+    assert o2._unit_text() == "Wind speed [inches/hour]"
+
 
 
 def test_from_df():