add h5ad assert test

[nilchia]

Current h5 assertions (has_h5_keys and has_h5_attribute) are not enough for an h5ad object. Because the tool developer needs to check the content of each group (like if "cell_type" is in "obs") not the main attribute (which is "{'encoding-type': 'anndata', 'encoding-version': '0.1.0'}" for all h5ad files), for more info please see the anndata documentation.

I had this problem when I wanted to assert those when updating scanpy.

I also wrote an issue.

I wrote a new assert for this and tested that locally and it worked (I added the test to my local planemo ".../planemo/lib64/python3.12/site-packages/galaxy/tool_util/verify/asserts/hdf5.py"

I hope I put it correctly here :)

License

• I agree to license these and all my past contributions to the core galaxy codebase under the MIT license.

[bernt-matthias]

I think to some extent the present assertions can already check groups. See the last entry here

galaxy/test/functional/tools/validation_hdf5.xml

Line 16 in 3e8f65e

<has_h5_keys keys="bins, chroms,indexes,pixels,chroms/length" />

But maybe I'm wrong.

[nilchia]

for anndata file the has_h5_keys works fine and returns: ['X', 'layers', 'obs', 'obsm', 'obsp', 'uns', 'var', 'varm', 'varp']
(Although it returns all potential keys, whether they have any value in the anndata or not)

But has_h5_attribute is not a good assertion for anndata. It only returns the encoding-type and the encoding-version.

We need an assertion that checks the entries of a group. Like if I run sc.pp.pca() does it add X_PCA in obsm or not.

[bernt-matthias]

We need an assertion that checks the entries of a group. Like if I run sc.pp.pca() does it add X_PCA in obsm or not.

Would the following work:

<has_h5_keys keys="obsm/X_PCA" />

[nilchia]

That works :)
Thanks @bernt-matthias