Test numpy > 2 and resolve any issues

[havok2063]

We should test the DRP with numpy > 2.0. numpy > 2 introduced some breaking changes, and changes in algorithms, which may result in numerical differences. Additionally, not all packages have added support yet for numpy > 2.

To resolve some package dependencies, you need

• numpy > 2.0
• pandas > 2.2.2
• astropy > 6.3
• bottleneck > 1.4
• numexpr > 2.10

When testing, I see some failures between pandas and numpy, with the error ValueError: numpy.dtype size changed, may indicate binary incompatibility. Expected 96 from C header, got 88 from PyObject. Even though pandas == 2.2.2 is installed, there may still be some incompatibilities with either pandas, h5py, or some internal c-code. See, e.g. numpy/numpy#26710 or https://stackoverflow.com/questions/78634235/numpy-dtype-size-changed-may-indicate-binary-incompatibility-expected-96-from

numpy pinned to < 2 in #115. Also see https://numpy.org/doc/2.0/numpy_2_0_migration_guide.html and https://numpy.org/doc/2.0/release/2.0.0-notes.html