Typing errors in HDF reader

[TomNicholas]

The typing CI is failing but only with errors inside the hdf reader (cc @sharkinsspatial), in lines I didn't touch, which I don't understand, so I'm going to punt on that so I can merge this.

Originally posted by @TomNicholas in #243 (comment)

[sharkinsspatial]

@TomNicholas I can reproduce this locally after removing all cached packages via conda clean --all and rebuilding the virtualizarr-tests conda environment. I'm a bit flummoxed as to why this is occurring as both the h5py and mypy versions (3.12 and 1.13.0 respectively) are unchanged from the successful mypy check when #87 was merged https://github.com/zarr-developers/VirtualiZarr/actions/runs/11900319916/job/33160926453 🤔 . This appears to be another mypy inconsistency similar to #274. I'll make a PR to ignore typing in the hdf reader code, but I'd love to understand why this appears to occur spontaneously 🤷‍♂

[TomNicholas]

Thanks for looking into it. That's annoying though - let's have one PR to ignore the errors so the tests pass then maybe leave this issue open to track the fact that we don't actually know what's causing it?

[TomNicholas]

Wait this probably has something to do with the fact that currently in hdf/hdf5.py we call from xarray import Dataset then a few lines later call from h5py import Dataset. That's definitely not right.