René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"[Bug]: broken 'proportional' colorbar when using contourf+cmap+norm+extend","articleBody":"### Bug summary\r\n\r\nI get a **broken colorbar** (see _Outcome_ below) when using `contourf` **and** specifiying colors with `cmap` **and** `norm` instead of `colors`, **and** requesting a _proportional_ colorbar, **and** using `extend`\r\n\r\n### Code for reproduction\r\n\r\n```python\r\nimport numpy as np\r\nimport matplotlib.pyplot as plt\r\nimport matplotlib as mpl\r\n\r\n# Data definition and plot based on the Official contourf demo\r\n# https://matplotlib.org/stable/gallery/images_contours_and_fields/contourf_demo.html\r\n\r\norigin = 'lower'\r\n\r\ndelta = 0.025\r\n\r\nx = y = np.arange(-3.0, 3.01, delta)\r\nX, Y = np.meshgrid(x, y)\r\nZ1 = np.exp(-X**2 - Y**2)\r\nZ2 = np.exp(-(X - 1)**2 - (Y - 1)**2)\r\nZ = (Z1 - Z2) * 2\r\n\r\nnr, nc = Z.shape\r\n\r\nfig2, ax2 = plt.subplots()\r\n\r\n# Original demo levels\r\n#  levels = [-1.5, -1, -0.5, 0, 0.5, 1]\r\n\r\n# Irregular levels because we want to use a colorbar with\r\n#   spacing='proportional'\r\nlevels = [-1.25, -0.5, -0.125, 0.125, 0.5, 1.25]\r\n\r\n# Define a colormap and norm instead of using 'colors=' in the\r\n# contourf call, BECAUSE we want to use the same cmap and norm for\r\n# scatter and pcolormesh plots\r\ncmap = mpl.colors.ListedColormap(['r', 'g', 'white', 'g', 'r'])\r\ncmap.set_under('yellow')\r\ncmap.set_over('cyan')\r\nnorm = mpl.colors.BoundaryNorm(levels, cmap.N)\r\n\r\nCS3 = ax2.contourf(X, Y, Z, levels,\r\n                   \r\n                   # NO PROBLEM with spacing='proportional' if we\r\n                   # define the colors with 'colors='\r\n                   #\r\n                   # colors=('r', 'g', 'b'),\r\n                   \r\n                   # The problem appears when we use cmap and norm\r\n                   cmap=cmap,\r\n                   norm=norm,\r\n                   origin=origin,\r\n                   \r\n                   # NO PROBLEM with spacing='proportional' if we dont\r\n                   # use extend\r\n                   # extend='min'\r\n                   extend='both'\r\n)\r\n\r\n# Our data range extends outside the range of levels; make\r\n# data below the lowest contour level yellow, and above the\r\n# highest level cyan:\r\n# CS3.cmap.set_under('yellow')\r\n# CS3.cmap.set_over('cyan')\r\n\r\nax2.set_title('Listed colors')\r\n\r\n# Notice that the colorbar gets all the information it\r\n# needs from the ContourSet object, CS3.\r\nfig2.colorbar(CS3,\r\n              # BROKEN colorbar if we use proportional spacing\r\n              spacing='proportional',\r\n              # spacing='uniform',\r\n)\r\n\r\nfig2.savefig('test_cb.png')\r\n\r\nplt.show()\r\n```\r\n\r\n\r\n### Actual outcome\r\n\r\n I get the following warning\r\n```\r\n/home/share/unix_files/cdat/miniconda3_21-02/envs/cdatm_py3/lib/python3.8/site-packages/matplotlib/colorbar.py:1083: RuntimeWarning: invalid value encountered in double_scalars\r\n  automin = (y[2] - y[1]) / clen\r\n/home/share/unix_files/cdat/miniconda3_21-02/envs/cdatm_py3/lib/python3.8/site-packages/matplotlib/colorbar.py:1084: RuntimeWarning: invalid value encountered in double_scalars\r\n  automax = (y[-2] - y[-3]) / clen\r\n```\r\nand a **broken colorbar** looking like\r\n![test_cb_proportional](https://user-images.githubusercontent.com/2785573/131685123-56fa26b8-6f2a-4083-bb11-8ce6d108ddda.png)\r\n\r\n\r\n\r\n### Expected outcome\r\n\r\nThe colorbar is OK if I do use not all the _ands_ specified above. Example below with `spacing='uniform'` instead of _proportional_\r\n![test_cb_uniform](https://user-images.githubusercontent.com/2785573/131685421-a7417fcd-3182-4031-9259-f5e429d5a739.png)\r\n\r\n\r\n### Operating system\r\n\r\nCentOS Linux release 7.9.2009 (Core)\r\n\r\n### Matplotlib Version\r\n\r\n3.3.4\r\n\r\n### Matplotlib Backend\r\n\r\n'Qt5Agg' and 'Agg'\r\n\r\n### Python version\r\n\r\n3.8.8\r\n\r\n### Jupyter version\r\n\r\n_No response_\r\n\r\n### Other libraries\r\n\r\n_No response_\r\n\r\n### Installation\r\n\r\nconda\r\n\r\n### Conda channel\r\n\r\nconda-forge","author":{"url":"https://github.com/jypeter","@type":"Person","name":"jypeter"},"datePublished":"2021-09-01T14:14:52.000Z","interactionStatistic":{"@type":"InteractionCounter","interactionType":"https://schema.org/CommentAction","userInteractionCount":5},"url":"https://github.com/20963/matplotlib/issues/20963"}