BUG: avoid deprecated shape mutation APIs in time by neutrinoceros · Pull Request #19182 · astropy/astropy

neutrinoceros · 2026-01-19T12:29:34Z

Description

ref #19126

opening as a draft: what I have seem necessary but not sufficient. Will provide more details later

By checking this box, the PR author has requested that maintainers do NOT use the "Squash and Merge" button. Maintainers should respect this when possible; however, the final decision is at the discretion of the maintainer that merges the PR.

github-actions · 2026-01-19T12:29:45Z

Thank you for your contribution to Astropy! 🌌 This checklist is meant to remind the package maintainers who will review this pull request of some common things to look for.

Do the proposed changes actually accomplish desired goals?
Do the proposed changes follow the Astropy coding guidelines?
Are tests added/updated as required? If so, do they follow the Astropy testing guidelines?
Are docs added/updated as required? If so, do they follow the Astropy documentation guidelines?
Is rebase and/or squash necessary? If so, please provide the author with appropriate instructions. Also see instructions for rebase and squash.
Did the CI pass? If no, are the failures related? If you need to run daily and weekly cron jobs as part of the PR, please apply the "Extra CI" label. Codestyle issues can be fixed by the bot.
Is a change log needed? If yes, did the change log check pass? If no, add the "no-changelog-entry-needed" label. If this is a manual backport, use the "skip-changelog-checks" label unless special changelog handling is necessary.
Is this a big PR that makes a "What's new?" entry worthwhile and if so, is (1) a "what's new" entry included in this PR and (2) the "whatsnew-needed" label applied?
At the time of adding the milestone, if the milestone set requires a backport to release branch(es), apply the appropriate "backport-X.Y.x" label(s) before merge.

neutrinoceros · 2026-01-19T17:48:02Z

ok so I added an experimental commit that should help bringing more context; The problem I'm currently facing is that the remaining failing test contains several fragile parts that I don't immediately know what to do with

there's a with pytest.raises(AttributeError) which I believe is expecting an exception that we don't really control (I amended it in my latest commit)
later on, we test what happens for in place mutations that expect automatic broadcasting: this isn't replicable with np.reshape, ndarray.reshape or ndarray._set_shape, as far as I understand. The only path forward that I'm seeing would be to skip this part of the test for numpy 2.5+ and somehow deprecate/drop support for this pattern later if we can.

Any insight here, @mhvk ?

neutrinoceros · 2026-01-19T17:48:25Z

(Also I think image test failures are unrelated)

neutrinoceros · 2026-01-20T11:06:20Z

note that the problem I'm facing here is similar to #19133

mhvk

@neutrinoceros - this basically looks good. Let me know when you take it out of draft.

neutrinoceros · 2026-01-21T23:28:04Z

I was actually waiting for early feedback from you. The change may be sensible but it isn't stable: there's still one test failing, and the failure mode is new in this PR

mhvk

@neutrinoceros - I think I found the problem...

mhvk · 2026-01-22T16:08:01Z

+            if val is None or val.size <= 1:
+                continue
+            try:
+                if NUMPY_LT_2_5:


The problem is here: you assume that all val will be ndarray subclasses where _set_shape will just work, but for Masked subclasses this does not change the shape of the mask. I think we need to adjust MaskedNDArray -- see comment below.

p.s. I don't quite recall the failure modes for other parts, but possibly the assumption that every attribute is a plain ndarray goes wrong there too.

mhvk · 2026-01-22T16:50:08Z

@@ -744,13 +745,22 @@ def shape(self):
    @shape.setter
    def shape(self, shape):


As noted above, for not NUMPY_LT_2_5we also need to override _set_shape since it does not go the correct thing. Possibly easiest is just to call this _set_shape and let shape.setter refer to it.

THANK YOU ! this is exactly correct, and does resolve the remaining failure.

mhvk · 2026-01-22T16:51:32Z

+            if NUMPY_LT_2_5:
+                super(MaskedNDArray, type(self)).shape.__set__(self, shape)
+            else:
+                self._set_shape(shape)


Since we need to define our own _set_shape, this needs to be super()._set_shape(shape).

mhvk

Great that it worked! Let's get it in.

pllim · 2026-01-22T18:52:10Z

And the remaining 33 requires jplephem release or something? What is the wait time on that? Should we put in a ignore filter for now?

mhvk · 2026-01-22T19:01:36Z

Yes, ignoring the jplephem warnings (or, really, anything outside of astropy and numpy, if we can do that...) would make a lot of sense.

neutrinoceros · 2026-01-22T19:05:58Z

And the remaining 33 requires jplephem release or something?

yes

What is the wait time on that?

No idea.

ignoring the jplephem warnings (or, really, anything outside of astropy and numpy, if we can do that...)

I don't think pytest (really, warnings) supports this, but we could temporarily use my fork of jplephem in CI (as I did in #19184) to achieve the same effect.

pllim · 2026-01-22T19:23:48Z

I mean, just ignore the entire "Setting the shape on a NumPy array has been deprecated" in pyproject.toml here:

astropy/pyproject.toml

Line 247 in 143c9c1

filterwarnings = [

with a follow-up issue to remove the ignore before next v7.2.x release by setting a milestone to the issue.

lumberbot-app · 2026-01-22T19:23:56Z

Owee, I'm MrMeeseeks, Look at me.

There seem to be a conflict, please backport manually. Here are approximate instructions:

Checkout backport branch and update it.

git checkout v7.2.x
git pull

Cherry pick the first parent branch of the this PR on top of the older branch:

git cherry-pick -x -m1 081b25b00ffc41cd4ee728f8011dcaef5cd56d2e

You will likely have some merge/cherry-pick conflict here, fix them and commit:

git commit -am 'Backport PR #19182: BUG: avoid deprecated shape mutation APIs in time'

Push to a named branch:

git push YOURFORK v7.2.x:auto-backport-of-pr-19182-on-v7.2.x

Create a PR against branch v7.2.x, I would have named this PR:

"Backport PR #19182 on branch v7.2.x (BUG: avoid deprecated shape mutation APIs in time)"

And apply the correct labels and milestones.

Congratulations — you did some good work! Hopefully your backport PR will be tested by the continuous integration and merged soon!

Remember to remove the Still Needs Manual Backport label once the PR gets merged.

If these instructions are inaccurate, feel free to suggest an improvement.

pllim · 2026-01-22T19:24:16Z

Also can you please manually backport? Thanks! 🙏

neutrinoceros · 2026-01-22T20:15:51Z

I mean, just ignore the entire "Setting the shape on a NumPy array has been deprecated" in pyproject.toml here:

I know it's closer to our usual workflow, but I also think it's riskier in this instance: we could easily silence a real warning from ourselves or another third party library.

Anyway, best case scenario is still that we don't need a workaround at all.
Will manually backport this now

BUG: avoid deprecated shape mutation APIs in time (cherry picked from commit 081b25b)

Backport PR #19182 to branch v7.2.x (BUG: avoid deprecated shape mutation APIs in time)

PR: astropy#19210 Issue: astropy#19182 Base commit: 7ad6f3d Changed lines: 66

PR: astropy#19182 Issue: astropy#19126 Base commit: 143c9c1 Changed lines: 64

neutrinoceros added this to the v7.2.1 milestone Jan 19, 2026

neutrinoceros added Bug no-changelog-entry-needed numpy-dev backport-v7.2.x on-merge: backport to v7.2.x labels Jan 19, 2026

github-actions Bot added time utils.masked labels Jan 19, 2026

neutrinoceros mentioned this pull request Jan 19, 2026

BUG: incoming deprecation warnings from numpy 2.5 (dev): direct mutations to array shape attributes are deprecated #19126

Closed

neutrinoceros force-pushed the time/19126 branch from 5d4735d to 826f83f Compare January 19, 2026 15:58

neutrinoceros force-pushed the time/19126 branch from ce8ee77 to 409225f Compare January 20, 2026 10:53

neutrinoceros mentioned this pull request Jan 20, 2026

EXP: try patched jplephem (address numpy-dev deprecations) #19184

Closed

1 task

mhvk reviewed Jan 21, 2026

View reviewed changes

mhvk reviewed Jan 22, 2026

View reviewed changes

BUG: avoid deprecated shape mutation APIs in time

f04b9b5

neutrinoceros force-pushed the time/19126 branch from 409225f to f04b9b5 Compare January 22, 2026 18:26

neutrinoceros mentioned this pull request Jan 22, 2026

MAINT: shape setting should be deprecated in astropy container classes. #19168

Open

neutrinoceros marked this pull request as ready for review January 22, 2026 18:42

neutrinoceros requested a review from taldcroft as a code owner January 22, 2026 18:42

mhvk approved these changes Jan 22, 2026

View reviewed changes

mhvk enabled auto-merge January 22, 2026 18:46

mhvk merged commit 081b25b into astropy:main Jan 22, 2026
31 of 33 checks passed

lumberbot-app Bot added the Still Needs Manual Backport label Jan 22, 2026

neutrinoceros deleted the time/19126 branch January 22, 2026 20:15

neutrinoceros mentioned this pull request Jan 22, 2026

Backport PR #19182 to branch v7.2.x (BUG: avoid deprecated shape mutation APIs in time) #19210

Merged

1 task

neutrinoceros pushed a commit to neutrinoceros/astropy that referenced this pull request Jan 22, 2026

Merge pull request astropy#19182 from neutrinoceros/time/19126

b19e0e5

BUG: avoid deprecated shape mutation APIs in time (cherry picked from commit 081b25b)

neutrinoceros removed the Still Needs Manual Backport label Jan 22, 2026

pllim added a commit that referenced this pull request Jan 22, 2026

Merge pull request #19210 from neutrinoceros/bp/19182

2ee21ee

Backport PR #19182 to branch v7.2.x (BUG: avoid deprecated shape mutation APIs in time)

tonybaloney pushed a commit to tonybaloney/swe-complex-astropy that referenced this pull request Mar 19, 2026

Human gold patch for astropy__astropy-19210

9ffd64a

PR: astropy#19210 Issue: astropy#19182 Base commit: 7ad6f3d Changed lines: 66

tonybaloney pushed a commit to tonybaloney/swe-complex-astropy that referenced this pull request Mar 19, 2026

Human gold patch for astropy__astropy-19182

c4cab58

PR: astropy#19182 Issue: astropy#19126 Base commit: 143c9c1 Changed lines: 64

		@@ -744,13 +745,22 @@ def shape(self):
		@shape.setter
		def shape(self, shape):

Conversation

neutrinoceros commented Jan 19, 2026

Description

Uh oh!

github-actions Bot commented Jan 19, 2026

Uh oh!

neutrinoceros commented Jan 19, 2026

Uh oh!

neutrinoceros commented Jan 19, 2026

Uh oh!

neutrinoceros commented Jan 20, 2026

Uh oh!

mhvk left a comment

Choose a reason for hiding this comment

Uh oh!

neutrinoceros commented Jan 21, 2026

Uh oh!

mhvk left a comment

Choose a reason for hiding this comment

Uh oh!

mhvk Jan 22, 2026

Choose a reason for hiding this comment

Uh oh!

mhvk Jan 22, 2026

Choose a reason for hiding this comment

Uh oh!

neutrinoceros Jan 22, 2026

Choose a reason for hiding this comment

Uh oh!

mhvk Jan 22, 2026

Choose a reason for hiding this comment

Uh oh!

mhvk left a comment

Choose a reason for hiding this comment

Uh oh!

pllim commented Jan 22, 2026

Uh oh!

mhvk commented Jan 22, 2026

Uh oh!

neutrinoceros commented Jan 22, 2026

Uh oh!

Uh oh!

pllim commented Jan 22, 2026

Uh oh!

lumberbot-app Bot commented Jan 22, 2026

Uh oh!

pllim commented Jan 22, 2026

Uh oh!

neutrinoceros commented Jan 22, 2026

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants