[MRG+1] Fix imshow masked interpolation by tacaswell · Pull Request #8024 · matplotlib/matplotlib · GitHub

The code changes are in the first commit, the image changes are in the second.

This probably needs some more tests and docs. I suspect there is at least one more 🐉 down here...

So, defiantly 🐲 in up-sampling when mixing masking + over/under with in the kernel foot print.

import matplotlib.pyplot as plt
import matplotlib as mpl
import matplotlib.colors as mcolors
import matplotlib.image as mimage
import matplotlib.cm as mcm
import numpy as np
import copy

cm = copy.copy(mcm.get_cmap('viridis'))
cm.set_over('r')
cm.set_under('b')
cm.set_bad('k')

n = mcolors.Normalize(vmin=0, vmax=100)

data = np.arange(100, dtype='float').reshape(10, 10)

data[5, 5] = -1

data[7, 7] = 101

data[3, 3] = np.nan

data[5, 3] = np.inf

mask = np.zeros_like(data).astype('bool')
mask[3, 5] = True

data = np.ma.masked_array(data, mask)


fig, ax_grid = plt.subplots(3, 6)
for interp, ax in zip(mimage._interpd_, ax_grid.ravel()):
    ax.set_title(interp)
    im = ax.imshow(data, norm=n, cmap=cm, interpolation=interp)

Right, forget about the topo failure, that is due to the resample vs not resample change and I have not looked into it yet.

On the bright side I have managed to fix this without having to change classic style and only update 1 test image, on the down side, one test is still broken.

and I have fixed the checker-board effect in that png above, that was due to Agg's aggressive clipping.

Of course this passes locally...

When determining which pixels to mask in the resampled image, if _any_ contribution to final value comes from a masked pixel, mask the result. Due to Agg special-casing the meaning of the alpha channel, the interpolation for the mask channel needs to be done separately. This is probably a template for doing the over/under separately. print out exact hash on travis

I'm not concerned about this, but would like to make sure that you intended to have this here.

I did, was a bit worried that gh/travis were having caching / timeout issues and not running the code I thought it was running (turns out the problem is I had accidentally depended on dictionary ordering in 3.6). I think this is a good idea to keep around so that we can verify exactly what code travis runs for testing locally (it should be running on the merge into the target branch, when it re-sets what that merge is is not something I fully understand yet).

The CI failure appears non-transient.

What happened to Appveyor?

It doesn't run on the v2.0.x branch.

How did I not know that?

Looks reasonable to me. Only a minor question.

Would & work here instead of *? Would make more sense given the boolean masks.

Are there performance rather than semantic advantages? I think ~(a & b) would also work and have one less temprorary.

I don't think there's any point adding a title if remove_text is True in the decorator. Otherwise this patch looks good to me.

The title is there for human debuggers later (I often copy-past tests into another buffer and just run them)

👍

Same as above, prseumably pointless with remove_text == True

This makes for an ugly change in Cartopy:
v2.0.0:

vs v2.0.x:

and diff:

You might need to view the full resolution image, but it's pretty evident on the lower-most blue one. It used to match the outline, but now there's a very obvious not very straight white bit along the edge.

sigh

The build also failed matplotlib.tests.test_image.test_rotate_image.test on 2.7 and 3.4 when this got merged, though I'm not sure why. This wasn't a troublesome test before.

Even weirder, #8144 is failling on that test on 3.5 but not other versions, even after a rebuild...

The rotate image test did give me trouble when working on this.

I have sorted out the reason (but not the cause) of the rotate_image failures, for some reason the set_bad from the test added here is leaking out to other tests.

Maybe it should use deepcopy instead of copy?

That would fix it, but we should also fix ColorMap so that copy works as expected, see #8299

tacaswell added the Release critical For bugs that make the library unusable (segfaults, incorrect plots, etc) and major regressions. label Feb 5, 2017

tacaswell added this to the 2.0.1 (next bug fix release) milestone Feb 5, 2017

tacaswell force-pushed the fix_ishow_masked_interpolation branch 2 times, most recently from 8786852 to 236b48f Compare February 6, 2017 01:02

tacaswell force-pushed the fix_ishow_masked_interpolation branch from 236b48f to 5e83b93 Compare February 6, 2017 02:05

tacaswell closed this Feb 6, 2017

tacaswell reopened this Feb 6, 2017

tacaswell force-pushed the fix_ishow_masked_interpolation branch 2 times, most recently from a1fbee5 to 73091ac Compare February 6, 2017 13:23

tacaswell force-pushed the fix_ishow_masked_interpolation branch from 73091ac to 1a7c4ef Compare February 6, 2017 20:22

phobson reviewed Feb 6, 2017

View reviewed changes

QuLogic changed the title Fix ishow masked interpolation Fix imshow masked interpolation Feb 6, 2017

dopplershift approved these changes Feb 24, 2017

View reviewed changes

dstansby approved these changes Feb 24, 2017

View reviewed changes

dstansby changed the title Fix imshow masked interpolation [MRG+1] Fix imshow masked interpolation Feb 24, 2017

NelleV merged commit 067f9b6 into matplotlib:v2.0.x Feb 24, 2017

tacaswell deleted the fix_ishow_masked_interpolation branch February 24, 2017 22:30

tacaswell mentioned this pull request Feb 24, 2017

check imshow performance #8143

Closed

tacaswell mentioned this pull request Mar 2, 2017

Allow array alpha for imshow #8183

Closed

QuLogic mentioned this pull request Mar 13, 2017

imshow interpolation uses masked values #8012

Closed

tacaswell mentioned this pull request Apr 3, 2017

Fix imshow edges #8300

Closed

WeatherGod mentioned this pull request Jun 6, 2017

imshow() pixelization in matplotlib 2.0.2 but not in 2.0.0 #8723

Closed

tacaswell mentioned this pull request Feb 23, 2018

imshow in 2.1: color resolution depends on outliers? #10567

Closed

+                              # 'unshare' the mask array to
+                              # needed to suppress numpy warning
+                              del out_mask
+                              invalid_mask = ~output.mask * ~np.isnan(output.data)

               install:
                 - ccache -s
+                - git describe

[MRG+1] Fix imshow masked interpolation by tacaswell · Pull Request #8024 · matplotlib/matplotlib · GitHub

Repository navigation

[MRG+1] Fix imshow masked interpolation - #8024

[MRG+1] Fix imshow masked interpolation#8024NelleV merged 1 commit intomatplotlib:v2.0.xmatplotlib/matplotlib:v2.0.xfrom tacaswell:fix_ishow_masked_interpolationtacaswell/matplotlib:fix_ishow_masked_interpolationCopy head branch name to clipboard

Conversation

tacaswell commented Feb 5, 2017

Uh oh!

tacaswell commented Feb 5, 2017

Uh oh!

tacaswell commented Feb 5, 2017

Uh oh!

tacaswell commented Feb 6, 2017

Uh oh!

tacaswell commented Feb 6, 2017

Uh oh!

tacaswell commented Feb 6, 2017

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

QuLogic commented Feb 6, 2017

Uh oh!

dopplershift commented Feb 6, 2017

Uh oh!

QuLogic commented Feb 6, 2017

Uh oh!

dopplershift commented Feb 6, 2017

Uh oh!

dopplershift left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

QuLogic commented Feb 24, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

dopplershift commented Feb 25, 2017

Uh oh!

QuLogic commented Feb 25, 2017

Uh oh!

QuLogic commented Feb 25, 2017

Uh oh!

tacaswell commented Feb 26, 2017

Uh oh!

tacaswell commented Mar 15, 2017

Uh oh!

QuLogic commented Mar 15, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

tacaswell commented Mar 15, 2017

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

[MRG+1] Fix imshow masked interpolation#8024
NelleV merged 1 commit into
matplotlib:v2.0.xfrom
tacaswell:fix_ishow_masked_interpolation

QuLogic commented Feb 24, 2017 •
edited
Loading

QuLogic commented Mar 15, 2017 •
edited
Loading