matplotlib · timhoffm · Sep 16, 2021 · Sep 16, 2021
diff --git a/doc/api/pyplot_summary.rst b/doc/api/pyplot_summary.rst
@@ -29,4 +29,5 @@ For a more in-depth look at colormaps, see the
 
 .. currentmodule:: matplotlib.pyplot
 
-.. autofunction:: colormaps
+.. autodata:: colormaps
+   :no-value:
diff --git a/doc/users/next_whats_new/colormaps.rst b/doc/users/next_whats_new/colormaps.rst
@@ -1,18 +1,22 @@
-Colormap registry
-------------------
+Colormap registry (experimental)
+--------------------------------
 
-Colormaps are now managed via `matplotlib.colormaps`, which is a
-`.ColormapRegistry`.
+Colormaps are now managed via `matplotlib.colormaps` (or `.pyplot.colormaps`),
+which is a `.ColormapRegistry`. While we are confident that the API is final,
+we formally mark it as experimental for 3.5 because we want to keep the option
+to still adapt the API for 3.6 should the need arise.
 
 Colormaps can be obtained using item access::
 
-    import matplotlib as mpl
-    cmap = mpl.colormaps['viridis']
+    import matplotlib.pyplot as plt
+    cmap = plt.colormaps['viridis']
 
 To register new colormaps use::
 
-    mpl.colormaps.register(my_colormap)
+    plt.colormaps.register(my_colormap)
 
-The use of `matplotlib.cm.get_cmap` and `matplotlib.cm.register_cmap` is
-discouraged in favor of the above. Within `.pyplot` the use of
-``plt.get_cmap()`` and ``plt.register_cmap()`` will continue to be supported.
+We recommend to use the new API instead of the `~.cm.get_cmap` and
+`~.cm.register_cmap` functions for new code. `matplotlib.cm.get_cmap` and
+`matplotlib.cm.register_cmap` will eventually be deprecated and removed.
+Within `.pyplot` ``plt.get_cmap()`` and ``plt.register_cmap()`` will continue
+to be supported for backward compatibility.
diff --git a/examples/images_contours_and_fields/contourf_demo.py b/examples/images_contours_and_fields/contourf_demo.py
@@ -6,7 +6,6 @@
 How to use the `.axes.Axes.contourf` method to create filled contour plots.
 """
 import numpy as np
-import matplotlib as mpl
 import matplotlib.pyplot as plt
 
 origin = 'lower'
@@ -87,7 +86,7 @@
 
 # Illustrate all 4 possible "extend" settings:
 extends = ["neither", "both", "min", "max"]
-cmap = mpl.colormaps["winter"].with_extremes(under="magenta", over="yellow")
+cmap = plt.colormaps["winter"].with_extremes(under="magenta", over="yellow")
 # Note: contouring simply excludes masked or nan regions, so
 # instead of using the "bad" colormap value for them, it draws
 # nothing at all in them.  Therefore the following would have

diff --git a/examples/images_contours_and_fields/demo_bboximage.py b/examples/images_contours_and_fields/demo_bboximage.py
@@ -7,7 +7,6 @@
 a bounding box. This demo shows how to show an image inside a `.text.Text`'s
 bounding box as well as how to manually create a bounding box for the image.
 """
-import matplotlib as mpl
 import matplotlib.pyplot as plt
 import numpy as np
 from matplotlib.image import BboxImage
@@ -39,7 +38,7 @@
 a = np.vstack((a, a))
 
 # List of all colormaps; skip reversed colormaps.
-cmap_names = sorted(m for m in mpl.colormaps if not m.endswith("_r"))
+cmap_names = sorted(m for m in plt.colormaps if not m.endswith("_r"))
 
 ncol = 2
 nrow = len(cmap_names) // ncol + 1

diff --git a/examples/images_contours_and_fields/pcolormesh_levels.py b/examples/images_contours_and_fields/pcolormesh_levels.py
@@ -94,7 +94,7 @@
 
 # pick the desired colormap, sensible levels, and define a normalization
 # instance which takes data values and translates those into levels.
-cmap = plt.get_cmap('PiYG')
+cmap = plt.colormaps['PiYG']
 norm = BoundaryNorm(levels, ncolors=cmap.N, clip=True)
 
 fig, (ax0, ax1) = plt.subplots(nrows=2)

diff --git a/examples/images_contours_and_fields/quadmesh_demo.py b/examples/images_contours_and_fields/quadmesh_demo.py
@@ -9,7 +9,6 @@
 This demo illustrates a bug in quadmesh with masked data.
 """
 
-import matplotlib as mpl
 from matplotlib import pyplot as plt
 import numpy as np
 
@@ -30,7 +29,7 @@
 axs[0].set_title('Without masked values')
 
 # You can control the color of the masked region.
-cmap = mpl.colormaps[mpl.rcParams['image.cmap']].with_extremes(bad='y')
+cmap = plt.colormaps[plt.rcParams['image.cmap']].with_extremes(bad='y')
 axs[1].pcolormesh(Qx, Qz, Zm, shading='gouraud', cmap=cmap)
 axs[1].set_title('With masked values')
 

diff --git a/examples/lines_bars_and_markers/horizontal_barchart_distribution.py b/examples/lines_bars_and_markers/horizontal_barchart_distribution.py
@@ -43,7 +43,7 @@ def survey(results, category_names):
     labels = list(results.keys())
     data = np.array(list(results.values()))
     data_cum = data.cumsum(axis=1)
-    category_colors = plt.get_cmap('RdYlGn')(
+    category_colors = plt.colormaps['RdYlGn'](
         np.linspace(0.15, 0.85, data.shape[1]))
 
     fig, ax = plt.subplots(figsize=(9.2, 5))

diff --git a/examples/mplot3d/polys3d.py b/examples/mplot3d/polys3d.py
@@ -32,7 +32,7 @@ def polygon_under_graph(x, y):
 
 # verts[i] is a list of (x, y) pairs defining polygon i.
 verts = [polygon_under_graph(x, poisson.pmf(l, x)) for l in lambdas]
-facecolors = plt.get_cmap('viridis_r')(np.linspace(0, 1, len(verts)))
+facecolors = plt.colormaps['viridis_r'](np.linspace(0, 1, len(verts)))
 
 poly = PolyCollection(verts, facecolors=facecolors, alpha=.7)
 ax.add_collection3d(poly, zs=lambdas, zdir='y')

diff --git a/examples/pie_and_polar_charts/nested_pie.py b/examples/pie_and_polar_charts/nested_pie.py
@@ -8,7 +8,6 @@
 
 """
 
-import matplotlib as mpl
 import matplotlib.pyplot as plt
 import numpy as np
 
@@ -31,7 +30,7 @@
 size = 0.3
 vals = np.array([[60., 32.], [37., 40.], [29., 10.]])
 
-cmap = mpl.colormaps["tab20c"]
+cmap = plt.colormaps["tab20c"]
 outer_colors = cmap(np.arange(3)*4)
 inner_colors = cmap([1, 2, 5, 6, 9, 10])
 
@@ -62,7 +61,7 @@
 # Obtain the ordinates of the bar edges
 valsleft = np.cumsum(np.append(0, valsnorm.flatten()[:-1])).reshape(vals.shape)
 
-cmap = mpl.colormaps["tab20c"]
+cmap = plt.colormaps["tab20c"]
 outer_colors = cmap(np.arange(3)*4)
 inner_colors = cmap([1, 2, 5, 6, 9, 10])
 

diff --git a/examples/userdemo/colormap_interactive_adjustment.py b/examples/userdemo/colormap_interactive_adjustment.py
@@ -54,7 +54,7 @@ def adjust_colorbar(mouseevent):
 Z2 = np.exp(-(X - 1)**2 - (Y - 1)**2)
 Z = (0.9*Z1 - 0.5*Z2) * 2
 
-cmap = plt.get_cmap('viridis').with_extremes(
+cmap = plt.colormaps['viridis'].with_extremes(
     over='xkcd:orange', under='xkcd:dark red')
 axesimage = plt.imshow(Z, cmap=cmap)
 colorbar = plt.colorbar(axesimage, ax=ax, use_gridspec=True)

diff --git a/lib/matplotlib/cm.py b/lib/matplotlib/cm.py
@@ -102,6 +102,12 @@ class ColormapRegistry(Mapping):
     r"""
     Container for colormaps that are known to Matplotlib by name.
 
+    .. admonition:: Experimental
+
+       While we expect the API to be final, we formally mark it as
+       experimental for 3.5 because we want to keep the option to still adapt
+       the API for 3.6 should the need arise.
+
     The universal registry instance is `matplotlib.colormaps`. There should be
     no need for users to instantiate `.ColormapRegistry` themselves.
 
@@ -136,6 +142,16 @@ def __str__(self):
         return ('ColormapRegistry; available colormaps:\n' +
                 ', '.join(f"'{name}'" for name in self))
 
+    def __call__(self):
+        """
+        Return a list of the registered colormap names.
+
+        This exists only for backward-compatibilty in `.pyplot` which had a
+        ``plt.colormaps()`` method. The recommended way to get this list is
+        now ``list(colormaps)``.
+        """
+        return list(self)
+
     def register(self, cmap, *, name=None, force=False):
         """
         Register a new colormap.