colorbar(boundaries=...) doesn't work so well with nonlinear norms

[anntzer]

Bug report

Bug summary

Code for reproduction

from pylab import *
imshow(np.random.rand(100, 100), norm=mpl.colors.LogNorm())
colorbar(boundaries=[1e-4, 1e-3, 1e-2, 1])

Actual outcome

Expected outcome

Either properly positioned boundaries, or (better) disallow passing boundaries in such cases.

Note that the docstrings in colorbar.py say

    The following will probably be useful only in the context of
    indexed colors (that is, when the mappable has norm=NoNorm()),
    or other unusual circumstances.

    ============   ===================================================
    Property       Description
    ============   ===================================================
    *boundaries*   None or a sequence
    *values*       None or a sequence which must be of length 1 less
                   than the sequence of *boundaries*. For each region
                   delimited by adjacent entries in *boundaries*, the
                   color mapped to the corresponding value in values
                   will be used.
    ============   ===================================================

but that still ends up part of the public API, which makes my intended fix for #7320 more difficult...

Matplotlib version

• Operating system: linux
• Matplotlib version: master
• Matplotlib backend (print(matplotlib.get_backend())): qt5agg
• Python version: 3.7
• Jupyter version (if applicable):
• Other libraries:

[ImportanceOfBeingErnest]

You get "properly" positionned ticks when in addition specfiying spacing="proportional".

import matplotlib.pyplot as plt
import matplotlib.colors
import numpy as np

plt.imshow(np.random.rand(100, 100), norm=matplotlib.colors.LogNorm())
plt.colorbar(boundaries=[1e-4, 1e-3, 1e-2, 1], spacing="proportional")

plt.show()

[anntzer]

Thanks for pointing this out. I don't think this should be an option, though... (it sound be the default and only behavior... no?)

[jklymak]

When overhauling colorbar I didn’t really look at boundary colorbars. Maybe they need a similar overhaul

[jklymak]

Oops. Didn’t see the above. I’d be in favour of proportional=true as a default.

[ImportanceOfBeingErnest]

I would have though the "usual" (if you can speak of "usual" for something that declared "unusual") use case would be something like

import matplotlib.pyplot as plt
import matplotlib.colors
import numpy as np; np.random.seed(42)

data = np.random.choice([0,5,250], size=(12,12))
cmap = matplotlib.colors.ListedColormap(["crimson","gold","indigo"])
bounds = [0,1,10,1000]
norm = matplotlib.colors.BoundaryNorm(bounds, len(bounds)-1)

fig, axes = plt.subplots(ncols=2)

for ax, spacing in zip(axes, ["uniform", "proportional"]):

    ax.set_title("spacing="+spacing)
    im = ax.imshow(data, norm=norm, cmap=cmap)
    fig.colorbar(im, ax=ax, boundaries=bounds, values =[0,5,250], spacing=spacing, 
                 orientation="horizontal")

plt.show()

Here the use of "proportional" is clearly undesired.

The consequence of such change is then the usual "My plot looks different, what can I do?". Hence my vote here would be: Don't break the API.

[anntzer]

As the docstring quote above says:

    The following will probably be useful only in the context of
    indexed colors (that is, when the mappable has norm=NoNorm()),
    or other unusual circumstances.

I guess we can add BoundaryNorm to the list of reasonable use cases. But it would be nice if NoNorm and BoundaryNorm could somehow signal that they need this behavior to the colorbar machinery, rather than the user having to specify it.

[github-actions]

This issue has been marked "inactive" because it has been 365 days since the last comment. If this issue is still present in recent Matplotlib releases, or the feature request is still wanted, please leave a comment and this label will be removed. If there are no updates in another 30 days, this issue will be automatically closed, but you are free to re-open or create a new issue if needed. We value issue reports, and this procedure is meant to help us resurface and prioritize issues that have not been addressed yet, not make them disappear. Thanks for your help!