Close open file handle if there aren't loadable_vars

[maxrjones]

There's a ton of warnings like the following when running the tests:

/Users/max/miniforge3/envs/virtualizarr-tests/lib/python3.11/json/encoder.py:357: RuntimeWarning: deallocating CachingFileManager(<class 'h5netcdf.core.File'>, '/private/var/folders/70/hc_nynms54d8lp67z4rsfctc0000gp/T/pytest-of-max/pytest-22/test_cf_fill_value_roundtrip_c5/cf_fill_value.nc', mode='r', kwargs={'invalid_netcdf': None, 'decode_vlen_strings': True, 'driver': None}, manager_id='9996df4a-48c7-4871-b282-9d4619d8f32e'), but file is not already closed. This may indicate a bug.

I think we need to be more careful about open file handles in the readers, this handles one case in which I think it can be closed.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 88.58%. Comparing base (bf63593) to head (8534ec7).
Report is 10 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #455   +/-   ##
=======================================
  Coverage   88.58%   88.58%           
=======================================
  Files          28       28           
  Lines        1743     1744    +1     
=======================================
+ Hits         1544     1545    +1     
  Misses        199      199           

Files with missing lines	Coverage Δ
virtualizarr/readers/common.py	92.72% <100.00%> (+0.13%)	⬆️

🚀 New features to boost your workflow:

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

[chuckwondo]

@maxrjones, see #390

[maxrjones]

@maxrjones, see #390

You called it 😅

[TomNicholas]

Oh nice catch. Yeah the whole loadable_vars implementation is a bit hacky atm, but it will be significantly cleaner after #353 and #427.

[maxrjones]

Oh nice catch. Yeah the whole loadable_vars implementation is a bit hacky atm, but it will be significantly cleaner after #353 and #427.

Unfortunately this didn't actually fix the issue (test is still hanging). Thanks for the pointers to those PRs - good to know that they exist to not spend much time cleaning up any file handling in the existing code.