Added HiPS and HiPS3D dask array classes

[astrofrog]

This is experimental for now, but these functions allow us to get a single dask array for a HiPS or HiPS3D dataset at a given level. Example:

from astropy.io import fits
from reproject.hips.dask_array import hips_as_dask
import matplotlib.pyplot as plt
array = hips_as_dask("https://alasky.cds.unistra.fr/2MASS/K/", level=0)
plt.imshow(array.compute(), origin='lower', vmin=0, vmax=1)
plt.savefig('2d.png')

and for HiPS3D:

from astropy.io import fits
from reproject.hips.dask_array import hips3d_as_dask
import matplotlib.pyplot as plt
array = hips3d_as_dask("https://alasky.cds.unistra.fr/GALFAHI/GALFAHI-Narrow-DR2-3D", level=0)
plt.imshow(array.compute()[10], origin='lower', vmin=-0.1, vmax=0.1)
plt.savefig('3d.png')

The nice thing about using dask is that all the I/O including remote calls gets multi-threaded, so this is quite fast.

This is something I need for glue and also something I think we can use to then easily reproject from HiPS to regular WCS, so I am putting it in this PR for now, but let's see how things evolve before deciding whether to merge. In any case, this is still quite buggy and very much a draft.

[codecov]

Codecov Report

❌ Patch coverage is 24.30939% with 137 lines in your changes missing coverage. Please review.
✅ Project coverage is 81.24%. Comparing base (dc4b0e2) to head (8097dba).
⚠️ Report is 14 commits behind head on main.

Files with missing lines	Patch %	Lines
reproject/hips/dask_array.py	20.80%	118 Missing ⚠️
reproject/hips/utils.py	40.62%	19 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #514      +/-   ##
==========================================
- Coverage   88.30%   81.24%   -7.07%     
==========================================
  Files          28       29       +1     
  Lines        1411     1578     +167     
==========================================
+ Hits         1246     1282      +36     
- Misses        165      296     +131     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.