Add more lints to Ruff for better code quality.

[hameerabbasi]

This PR also adds lints that check for NumPy 2.0 compatibility.

[codecov]

Codecov Report

Merging #621 (f403a0f) into master (d6d1b19) will increase coverage by 0.11%.
The diff coverage is 79.05%.

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #621      +/-   ##
==========================================
+ Coverage   90.36%   90.48%   +0.11%     
==========================================
  Files          20       20              
  Lines        3603     3573      -30     
==========================================
- Hits         3256     3233      -23     
+ Misses        347      340       -7     

[hameerabbasi]

@mtsokol Do you primarily work on Windows? If so, would it be possible for you to look into why this fails only on Windows?

[mtsokol]

@mtsokol Do you primarily work on Windows? If so, would it be possible for you to look into why this fails only on Windows?

I only have macos machine unfortunately.

[mtsokol]

@hameerabbasi I have and idea what might be causing the failure. In test_elemwise.py (where tests are failing) there's a number of .astype(np.int_). np.int_ is an informal "default integer", which I think causes trouble here - on Windows it's 32bit but here for the sparse operation it stays as 64bit.

While working on NEP 52 there was even a discussion to remove np.int_ and np.uint to enforce users to use sized aliases explicitly (e.g. np.int64) or c-like names (np.long). np.int_ will get removed in a few releases: numpy/numpy#24743

Maybe a solution would be to replace .astype(np.int_) in test_elemwise.py with .astype(np.int64) or np.int32? (it's only bitwise and bitshift tests that are failing)

[hameerabbasi]

Maybe a solution would be to replace .astype(np.int_) in test_elemwise.py with .astype(np.int64) or np.int32? (it's only bitwise and bitshift tests that are failing)

That worked, thanks! I see you self-requested a review, I'll wait for that before merging.

[mtsokol]

LGTM!