Fix for mouseover image data with a MultiNorm

trygvrad · trygvrad · commit 3e4b469692ea · 2025-02-08T10:50:06.000+01:00
If an image is shown, and the user hovers over the image, the value at that coordinate should be displayed. The number formating is handled by colorizer.Colorizer._format_cursor_data_override(). This is now updated to also handle the case where a MultiNorm is used (i.e. a bivariat or multivariate colormap), in which case multiple values are displayed.
diff --git a/lib/matplotlib/colorizer.py b/lib/matplotlib/colorizer.py
@@ -451,39 +451,63 @@ def colorbar(self):
     def colorbar(self, colorbar):
         self._colorizer.colorbar = colorbar
 
-    def _format_cursor_data_override(self, data):
-        # This function overwrites Artist.format_cursor_data(). We cannot
-        # implement cm.ScalarMappable.format_cursor_data() directly, because
-        # most cm.ScalarMappable subclasses inherit from Artist first and from
-        # cm.ScalarMappable second, so Artist.format_cursor_data would always
-        # have precedence over cm.ScalarMappable.format_cursor_data.
-
-        # Note if cm.ScalarMappable is depreciated, this functionality should be
-        # implemented as format_cursor_data() on ColorizingArtist.
-        n = self.cmap.N
-        if np.ma.getmask(data):
-            return "[]"
-        normed = self.norm(data)
+    @staticmethod
+    def _sig_digits_from_norm(norm, data, n):
+        # Determines the number of significant digits
+        # to use for a number given a norm, and n, where n is the
+        # number of colors in  the colormap.
+        normed = norm(data)
         if np.isfinite(normed):
-            if isinstance(self.norm, colors.BoundaryNorm):
+            if isinstance(norm, colors.BoundaryNorm):
                 # not an invertible normalization mapping
-                cur_idx = np.argmin(np.abs(self.norm.boundaries - data))
+                cur_idx = np.argmin(np.abs(norm.boundaries - data))
                 neigh_idx = max(0, cur_idx - 1)
                 # use max diff to prevent delta == 0
                 delta = np.diff(
-                    self.norm.boundaries[neigh_idx:cur_idx + 2]
+                    norm.boundaries[neigh_idx:cur_idx + 2]
                 ).max()
-            elif self.norm.vmin == self.norm.vmax:
+            elif norm.vmin == norm.vmax:
                 # singular norms, use delta of 10% of only value
-                delta = np.abs(self.norm.vmin * .1)
+                delta = np.abs(norm.vmin * .1)
             else:
                 # Midpoints of neighboring color intervals.
-                neighbors = self.norm.inverse(
+                neighbors = norm.inverse(
                     (int(normed * n) + np.array([0, 1])) / n)
                 delta = abs(neighbors - data).max()
             g_sig_digits = cbook._g_sig_digits(data, delta)
         else:
             g_sig_digits = 3  # Consistent with default below.
+        return g_sig_digits
+
+    def _format_cursor_data_override(self, data):
+        # This function overwrites Artist.format_cursor_data(). We cannot
+        # implement cm.ScalarMappable.format_cursor_data() directly, because
+        # most cm.ScalarMappable subclasses inherit from Artist first and from
+        # cm.ScalarMappable second, so Artist.format_cursor_data would always
+        # have precedence over cm.ScalarMappable.format_cursor_data.
+
+        # Note if cm.ScalarMappable is depreciated, this functionality should be
+        # implemented as format_cursor_data() on ColorizingArtist.
+        if np.ma.getmask(data) or data is None:
+            return "[]"
+        if len(data.dtype.descr) > 1:
+            # We have multivariate data encoded as a data type with multiple fields
+            # NOTE: If any of the fields are masked, "[]" would be returned via
+            # the if statement above.
+            s_sig_digits_list = []
+            if isinstance(self.cmap, colors.BivarColormap):
+                n_s = (self.cmap.N, self.cmap.M)
+            else:
+                n_s = [part.N for part in self.cmap]
+            os = [f"{d:-#.{self._sig_digits_from_norm(no, d, n)}g}"
+                            for no, d, n in zip(self.norm.norms, data, n_s)]
+            return f"[{', '.join(os)}]"
+
+        # scalar data
+        n = self.cmap.N
+        g_sig_digits = self._sig_digits_from_norm(self.norm,
+                                                  data,
+                                                  n)
         return f"[{data:-#.{g_sig_digits}g}]"
 
 
diff --git a/lib/matplotlib/tests/test_image.py b/lib/matplotlib/tests/test_image.py
@@ -14,7 +14,7 @@
 
 import matplotlib as mpl
 from matplotlib import (
-    colors, image as mimage, patches, pyplot as plt, style, rcParams)
+    cbook, colors, image as mimage, patches, pyplot as plt, style, rcParams)
 from matplotlib.image import (AxesImage, BboxImage, FigureImage,
                               NonUniformImage, PcolorImage)
 from matplotlib.testing.decorators import check_figures_equal, image_comparison
@@ -1130,8 +1130,14 @@ def test_image_cursor_formatting():
     data = np.ma.masked_array([0], mask=[False])
     assert im.format_cursor_data(data) == '[0]'
 
-    data = np.nan
-    assert im.format_cursor_data(data) == '[nan]'
+    # This used to test
+    # > data = np.nan
+    # > assert im.format_cursor_data(data) == '[nan]'
+    # However, a value of nan will be masked by `cbook.safe_masked_invalid(data)`
+    # called by `image._ImageBase._normalize_image_array(data)`
+    # The test is therefore changed to:
+    data = cbook.safe_masked_invalid(np.nan)
+    assert im.format_cursor_data(data) == '[]'
 
 
 @check_figures_equal()
diff --git a/lib/matplotlib/tests/test_multivariate_axes.py b/lib/matplotlib/tests/test_multivariate_axes.py
@@ -629,6 +629,54 @@ def test_cmap_error():
         mpl.collections.PatchCollection([], cmap='not_a_cmap')
 
 
+def test_artist_format_cursor_data_multivar():
+
+    X = np.zeros((4, 3))
+    X[0, 0] = 0.9
+    X[0, 1] = 0.99
+    X[0, 2] = 0.999
+    X[1, 0] = -1
+    X[1, 1] = 0
+    X[1, 2] = 1
+    X[2, 0] = 0.09
+    X[2, 1] = 0.009
+    X[2, 2] = 0.0009
+    X[3, 0] = np.nan
+
+    Y = np.arange(np.prod(X.shape)).reshape(X.shape)
+
+    labels_list = [
+        "[0.9, 0.00]",
+        "[1., 1.00]",
+        "[1., 2.00]",
+        "[-1.0, 3.00]",
+        "[0.0, 4.00]",
+        "[1.0, 5.00]",
+        "[0.09, 6.00]",
+        "[0.009, 7.00]",
+        "[0.0009, 8.00]",
+        "[]",
+    ]
+
+    pos = [[0, 0], [1, 0], [2, 0],
+           [0, 1], [1, 1], [2, 1],
+           [0, 2], [1, 2], [2, 2],
+           [3, 0]]
+
+    from matplotlib.backend_bases import MouseEvent
+
+    for cmap in ['BiOrangeBlue', '2VarAddA']:
+        fig, ax = plt.subplots()
+        norm = mpl.colors.BoundaryNorm(np.linspace(-1, 1, 20), 256)
+        data = (X, Y)
+        im = ax.imshow(data, cmap=cmap, norm=(norm, None))
+
+        for v, text in zip(pos, labels_list):
+            xdisp, ydisp = ax.transData.transform(v)
+            event = MouseEvent('motion_notify_event', fig.canvas, xdisp, ydisp)
+            assert im.format_cursor_data(im.get_cursor_data(event)) == text
+
+
 def test_multivariate_safe_masked_invalid():
     dt = np.dtype('float32, float32').newbyteorder('>')
     x = np.zeros(2, dtype=dt)
