Make type annotations for NumPy arrays more specific

[IvanIsCoding]

Related to #1352

This is trying to address some complaints from pyright infering the NumPy array as Unknown.

[coveralls]

Pull Request Test Coverage Report for Build 12776415127

Warning: This coverage report may be inaccurate.

This pull request's base commit is no longer the HEAD commit of its target branch. This means it includes changes from outside the original pull request, including, potentially, unrelated coverage changes.

• For more information on this, see Tracking coverage changes with pull request builds.
• To avoid this issue with future PRs, see these Recommended CI Configurations.
• For a quick fix, rebase this PR at GitHub. Your next report should be accurate.

Details

• 0 of 0 changed or added relevant lines in 0 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage remained the same at 95.83%

---

Totals
Change from base Build 12773634822:	0.0%
Covered Lines:	18384
Relevant Lines:	19184

---

💛 - Coveralls

[barakatzir]

I don't know if you want yo add it to this PR, but pyright also complains about numpy.dtype needing parameter annotations.

[IvanIsCoding]

I don't know if you want yo add it to this PR, but pyright also complains about numpy.dtype needing parameter annotations.

I annotated it with np.dtype[np.generic] but to be honest I find that par counter-intuitive.

[barakatzir]

I don't know if you want yo add it to this PR, but pyright also complains about numpy.dtype needing parameter annotations.

I annotated it with np.dtype[np.generic] but to be honest I find that par counter-intuitive.

You can set it to np.dtype[Any] which matches the new return type npt.NDArray[Any]. It sounds more intuitive to me.

[mtreinish]

Overall this LGTM, just one small inline question about the hint for __array__

[mtreinish]

Do we want to use a generic/typevar here? Since the Any for the dtype argument needs to match the dtype inside the returned array?

[IvanIsCoding]

For the record we ignore the dtype at runtime. However, I cannot annotate the ndarray with _T as that would be incorrect. There is a mapping from _T to numpy types that we’d need to manually add and that is a lot of work.

Also, this method exists to support np.array calls no one should call it directly. I think NumPy type stubs will mask this call anyway