Remove numpy restriction to <2

[PProfizi]

Starting with numpy==2 there is an explicit distinction between np.int32 and np.int64 which results in a difference in string representations of certain numbers.
Solves #1831

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 88.08%. Comparing base (59eb1b2) to head (dec0c8e).
Report is 1 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1864      +/-   ##
==========================================
- Coverage   88.10%   88.08%   -0.03%     
==========================================
  Files          83       83              
  Lines        9951     9951              
==========================================
- Hits         8767     8765       -2     
- Misses       1184     1186       +2     

[RobPasMue]

LGTM!

[BClappe]

Isn't there any other modification to make anywhere else?
Only 4 files?

[PProfizi]

Isn't there any other modification to make anywhere else? Only 4 files?

Apparently, yes, I guess this is because there are only very specific places where we return an np.int32 and test its string representation.