René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"imshow padding around NaN values","articleBody":"There seems to be a discrepancy in how matplotlib's `imshow` displays rasters which contain NaN areas. The NaN areas of the raster appear larger than they really are.\r\n\r\nI assume the problem might be that the raw raster gets resampled/interpolated during display (but _before_ converting to RGBA) and the NaN values propagate during that interpolation (since the ordinary mean of a list containing at least one nonfinite value is also not finite). If so, a fix would be to ensure conversion to RGBA always precedes any image resampling/rescaling.\r\n\r\nHere is an example. The first subplot shows a float32 raster with values of zero and one. The second subplot exactly replaces the nonzero pixels with NaN, however the resulting shapes look different (making narrow non-NaN features invisible). The third subplot combines both, showing that the non-NaN area (red) is _eroded_, resulting in a black outline. \r\n\r\nThe expected correct behaviour would be that the red layer perfectly covers the black layer, leaving no border visible.\r\n\r\n```python\r\nimport matplotlib.pyplot as plt, numpy as np\r\n\r\nraster = np.full(shape=[200]*2, fill_value=np.nan, dtype=np.float32) # example image\r\nraster[100:150, 100:150] = 0 \r\nfor i in range(10):\r\n    raster[20*i:20*i+i, :] = 0 # lines of varying thickness\r\n    \r\nplt.figure(figsize=(8,8)) # small figure\r\n\r\nplt.subplot(1,3,1).imshow(np.isnan(raster).astype(np.float32)) # 0 vs 1\r\nplt.subplot(1,3,2).imshow(raster)                              # 0 vs NaN\r\n \r\nax = plt.subplot(1,3,3)\r\nax.imshow(np.isnan(raster), cmap='gray') # black border/underlayer\r\nax.imshow(raster, cmap='autumn')         # red interior\r\n```\r\n\r\n![image](https://user-images.githubusercontent.com/12852539/96075301-7a357700-0ef6-11eb-9869-fed7622c9da6.png)\r\n\r\n\r\nTested on matplotlib 3.2.1, in a jupyter 1.0.0 sandbox using `ipykernel.pylab.backend_inline` backend, ipykernel 5.3.4, python 3.6.9, Linux 4.14 amzn2 x86_64.\r\n","author":{"url":"https://github.com/benjimin","@type":"Person","name":"benjimin"},"datePublished":"2020-10-15T04:00:01.000Z","interactionStatistic":{"@type":"InteractionCounter","interactionType":"https://schema.org/CommentAction","userInteractionCount":25},"url":"https://github.com/18735/matplotlib/issues/18735"}