Replace file/name with hdf5 group/dataset when decoding

[iguinn]

When calling read, the recursive functions will now pass the current group/dataset rather than a file/name pair. The file/name pair required re-looking up objects, which is quite slow because it involves re-parsing the hdf5 metadata.

[codecov]

Codecov Report

Attention: Patch coverage is 59.67742% with 25 lines in your changes missing coverage. Please review.

Project coverage is 74.28%. Comparing base (1317008) to head (b954b88).
Report is 67 commits behind head on main.

Files with missing lines	Patch %	Lines
src/lgdo/lh5/_serializers/read/composite.py	53.84%	6 Missing ⚠️
src/lgdo/lh5/core.py	40.00%	6 Missing ⚠️
...rc/lgdo/lh5/_serializers/read/vector_of_vectors.py	20.00%	4 Missing ⚠️
src/lgdo/lh5/_serializers/read/encoded.py	50.00%	2 Missing ⚠️
src/lgdo/lh5/_serializers/read/ndarray.py	75.00%	2 Missing ⚠️
src/lgdo/lh5/exceptions.py	33.33%	2 Missing ⚠️
src/lgdo/lh5/_serializers/read/array.py	88.88%	1 Missing ⚠️
src/lgdo/lh5/_serializers/read/scalar.py	83.33%	1 Missing ⚠️
src/lgdo/lh5/_serializers/read/utils.py	50.00%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main      #97      +/-   ##
==========================================
- Coverage   74.41%   74.28%   -0.14%     
==========================================
  Files          45       45              
  Lines        2806     2815       +9     
==========================================
+ Hits         2088     2091       +3     
- Misses        718      724       +6     

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

[gipert]

Wonderful Ian, this is exactly what I wanted to do. I'll marge and release a new version.

[MoritzNeuberger]

After this merge, I ran into issues loading the data by providing it with a list of file names.
I adjusted the code like such, and it seems to work again:

src/lgdo/lh5/core.py l. 117-120:

        for h5f in lh5_file:
            if isinstance(h5f, str):
                h5f = h5py.File(h5f, mode="r")  # noqa: PLW2901
            lh5_obj.append(h5f[name])

The issue seems to have been that the isinstance was checking lh5_file again instead of h5f.
Also, instead of += (extending the list), I used an append.

[iguinn]

This is fixed (and there's a test now too): #101