provide compatibility with numpy 2.0

[dimpase]

This will fix #38242

Basically, most of the problems are with numpy switching to a different way to display its data,
showing type. We either use a compatibility version, or add ellipses in doctests.

📝 Checklist

• [x ] The title is concise and informative.
• [x ] The description explains in detail what this PR is about.
• [x ] I have linked a relevant issue or discussion.
• [x ] I have created tests covering the changes.
• I have updated the documentation and checked the documentation preview.

⌛ Dependencies

[github-actions]

Documentation preview for this PR (built with commit 515d20e; changes) is ready! 🎉
This preview will update shortly after each push to this PR.

[dimpase]

Codecov complains about the lines numpy.set_printoptions(legacy="1.25") not covered by tests.
Not sure how to suppress these, it looks bogus to me.

[antonio-rojas]

I'm still getting a test failure:

**********************************************************************
File "/usr/lib/python3.12/site-packages/sage/matrix/matrix1.pyx", line 709, in sage.matrix.matrix1.Matrix.numpy
Failed example:
    sorted(numpy.typecodes.items())                                       # needs numpy
Expected:
    [('All', '?bhilqpBHILQPefdgFDGSUVOMm'), ('AllFloat', 'efdgFDG'),
     ('AllInteger', 'bBhHiIlLqQpP'), ('Character', 'c'), ('Complex', 'FDG'),
     ('Datetime', 'Mm'), ('Float', 'efdg'), ('Integer', 'bhilqp'),
     ('UnsignedInteger', 'BHILQP')]
Got:
    [('All', '?bhilqnpBHILQNPefdgFDGSUVOMm'),
     ('AllFloat', 'efdgFDG'),
     ('AllInteger', 'bBhHiIlLqQnNpP'),
     ('Character', 'c'),
     ('Complex', 'FDG'),
     ('Datetime', 'Mm'),
     ('Float', 'efdg'),
     ('Integer', 'bhilqnp'),
     ('UnsignedInteger', 'BHILQNP')]
**********************************************************************

[dimpase]

that's one I don't quite know how to fix.
I guess we don't actually need to check for values of these things, it is too verbose

[kiwifb]

Why is this even tested? I would understand if it was a numpy test, but a sage test?

[kiwifb]

Nitpicking, the string "# to ensure numpy 2.0 compatibility" is applied inconsistently. I suggest for it to be attached to the "if" statement line. I guess it will be clear if an import numpy statement had to be explicitly added.

[dimpase]

Why is this even tested? I would understand if it was a numpy test, but a sage test?

it's not really a test, it's more like part of a guide. Useful, but testing all its output isn't so important. I added #random tag, and added one more example on how to show a part of it

[antonio-rojas]

WFM now

[dimpase]

Nitpicking, the string "# to ensure numpy 2.0 compatibility" is applied inconsistently. I suggest for it to be attached to the "if" statement line. I guess it will be clear if an import numpy statement had to be explicitly added.

I hope you're fine with undoing all that once numpy 2 becomes standard, see #38275

[dimpase]

WFM now

thanks!

[kiwifb]

Nitpicking, the string "# to ensure numpy 2.0 compatibility" is applied inconsistently. I suggest for it to be attached to the "if" statement line. I guess it will be clear if an import numpy statement had to be explicitly added.

I hope you're fine with undoing all that once numpy 2 becomes standard, see #38275

Well, I did not say to remove it altogether. But I absolutely know what you are saying. In time those compatibility hacks will go.

[dimpase]

it was just a trivial rebase

[tornaria]

This PR works fine for me, except there is a change in src/sage/symbolic/expression.pyx that conflicts with #36641. Is there a chance you can remove that hunk here? Whoever is using sympy 1.13 will need to apply #36641 regardless.

[tornaria]

Note: I've tested this PR using numpy 1.26 and using numpy 2.0 and it works fine on both cases. See void-linux/void-packages#49571 (comment).