Fixes for NumPy 2.0 compatiblity

[WilliamJamieson]

NumPy 2.0 has removed a couple of type aliases referenced by deepdiff as part of NEP 52 in numpy/numpy#24376. Namely,

• np.float_, which is an alias of np.double according to the NumPy docs.
• np.complex_, which is an alias of np.cdouble according to the NumPy docs.

These are blocking my ability to test code using NumPy 2.0 (see spacetelescope/romancal#886).

When I ran the test suite with NumPy 2.0 installed there was one additional deprecation of np.in1d in favor or np.isin.

In any case, the deepdiff tests pass locally for me if I install the published dev wheel of NumPy via:

pip install --extra-index-url https://pypi.anaconda.org/scientific-python-nightly-wheels/simple "numpy>=0.0.dev0

to override the current requirements-dev.txt pin for NumPy. I have not touched the GitHub CI for deepdiff in order to run NumPy 2.0 tests at this time, please let me know if/how you would like me to do this if you would like that added.

[seperman]

Hi @WilliamJamieson
Thanks for the PR. We should bump the Numpy version in requirements-dev.txt to Numpy 2.0
We should still run the test suite with numpy==1.24.4 as well to ensure we have backward compatibility. That can be achieved by modifying the matrix section in GitHub actions in .github/workflows/main.yaml. We should add The 2 different Numpy versions there and cache pip for each one of these configurations. What do you think?

[WilliamJamieson]

Since NumPy 2.0 is still in development you probably do not want to cache it because you will need to pull the nightly wheel.

[WilliamJamieson]

Per NEP 29, NumPy is only supported on Python 3.9+.

[seperman]

LGTM!

[codecov-commenter]

Codecov Report

Merging #422 (c9232d2) into dev (2f62074) will increase coverage by 0.03%.
The diff coverage is 100.00%.

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

@@            Coverage Diff             @@
##              dev     #422      +/-   ##
==========================================
+ Coverage   98.60%   98.63%   +0.03%     
==========================================
  Files          14       14              
  Lines        3299     3299              
==========================================
+ Hits         3253     3254       +1     
+ Misses         46       45       -1     

Files	Coverage Δ
deepdiff/helper.py	97.37% <100.00%> (+0.29%)	⬆️