Add Numpy 2 compatibility

[cmeyer]

We would like to use py4DSTEM in Python environments with numpy 2.x.

The core packages seem to all be available to be used with numpy 2.x. Tested with this command:

conda create -n trial_dependencies -c conda-forge numpy scipy hdf5 hdf5plugin ncempy matplotlib scikit-image scikit-learn scikit-optimize tqdm dill gdown dask distributed emdfile mpire threadpoolctl pylops colorspacious

Some of the extensions dependencies may not yet available on conda-forge, but are available on PyPI (tensorflow requires 2.18 for numpy 2 compatibility, 2.18 isn't available using conda-forge on macOS arm64 yet). I'm on a macOS machine right now so I couldn't try everything (i.e. cupy).

[sezelt]

When we restricted it <2.0 in #677, the reason was an incompatibility in emdfile, but that seems to have been addressed: py4dstem/emdfile@3cdf83e now. Some incompatibilities remain in the main repo, but hopefully these are only the trivial ones from deprecated aliases.

[ericpre]

The conda-forge feedstock still have the numpy<2 pin, can it be removed?