Require numpy 2.1 or newer in tests. Adopt SPEC-0 in the package.

[samcunliffe]

We're currently broken because numpy removed the deprecated newshape argument to numpy.reshape in numpy 2.4.

So we either make an upper pin (numpy >= 1.23,<2.4), if you want to support old versions of numpy for any reason. Or (my preference) we drop support for numpy versions older than 2.0. This PR is the latter.

• @dstansby, you spotted this in 2024... Support napari>=0.5 #274 (comment)

• Does this need a changelog entry? And a new version?

---

Whilst I was at it, I spotted a trivial typo in tox.ini and realised we still have a pre-commit.ci badge but we dropped it in #308 (sorry for the unrelated changes!)

[Copilot]

Pull request overview

This PR updates the minimum numpy version requirement from >=1.23 to >=2.1 and modernizes the Python version support to align with the updated numpy dependency. This change addresses a breaking API change where numpy 2.4 removed the deprecated newshape parameter from numpy.reshape().

Key Changes:

• Updated numpy dependency from >=1.23 to >=2.1 in setup.cfg and pre-commit config
• Replaced deprecated newshape parameter with shape in numpy.reshape() call
• Updated Python version support from 3.10-3.12 to 3.11-3.13 in tox.ini
• Removed obsolete pre-commit.ci badge from README (following #308)
• Updated pre-commit hook versions (ruff, mypy)

Reviewed changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
setup.cfg	Updated numpy minimum version requirement from >=1.23 to >=2.1
src/napari_matplotlib/tests/helpers.py	Changed numpy.reshape() parameter from deprecated newshape to shape
tox.ini	Updated Python test environments from py{310,311,312} to py{311,312,313}
.pre-commit-config.yaml	Updated ruff (v0.13.1 → v0.14.10), mypy (v1.18.2 → v1.19.1), and aligned numpy dependency to >=2.1
README.md	Removed obsolete pre-commit.ci badge

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[samcunliffe]

Yet another sidenote: I can't request @K-Meech to review... is that because she needs to be a collaborator on the repo first?

[dstansby]

Thanks for this - I think we should follow SPEC 0 anyway, so this is definitely 👍

Unrelated to this PR:

• Did you enable copilot reviews? I am 👎 on having it enabled, but happy to discuss if anyone feels strongly
• @K-Meech would you like to be a maintainer here? If so I can get you added to the right teams.

[dstansby]

Can we lower the pin to numpy 2.0 though, since SPEC 0 doesn't say we should drop support for numpy 2.0 yet?

[samcunliffe]

Can we lower the pin to numpy 2.0 though

No, unfortunately, we can't. The change is 2.0 → 2.1.

Did you enable copilot reviews? I am 👎 on having it enabled, but happy to discuss if anyone feels strongly

No, it was available to click so I clicked.
I'm easy: but since you're 👎 enough to type 👎, then I'd support a "no-copilot" rule. Do you want to write something into CONTRIBUTING.md?

[K-Meech]

@samcunliffe - all looks good to me! @dstansby - happy to be added as a maintainer, but I can't commit a lot of time to this (I'm still maintaining various packages from my previous jobs which consumes a lot of my extra time 😅 ) Can contribute reviews / the odd bug fix though - especially since we use this package on the napari course, which we intend to teach regularly going forward.

[dstansby]

Yeah, absolutely no expecation of actually doing anything - I thought it might be helpful to self merge-and-release fixes you find when using napari-matplotlib, instead of having to wait for someone else to look over them.

[dstansby]

Since

1. The affected code is only in the tests
2. We should still support NumPy 2.0 for users

Could you put the min pin to numpy>=2.0 for the whole package, and then introduce a numpy>=2.1 pin for tests?

[samcunliffe]

instead of having to wait for someone else to look over them.

I'm still a bit uncomfortable merging my own PRs.
Even in this relatively trivial case, the review was helpful...

[dstansby]

I'm still a bit uncomfortable merging my own PRs.
Even in this relatively trivial case, the review was helpful...

This is fair