Adding modality to MuData.mod

[racng]

Is your feature request related to a problem? Please describe. After loading a CITE-seq 10x h5 file with muon and 10x vdj file with scirpy, I tried adding the AIRR modality to the existing mdata by adding it to mdata.mod. It seemed to work, since mdata shows that it has 3 modalities. However, when I tried to write the mdata, I get an error:

SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

However, there is no error if i initialize a new mdata object with all three modalities and save that.

Describe the solution you'd like Would be nice if we can add a modality with mdata.mod['airr'] = adata.copy() without causing warnings.

Describe alternatives you've considered We currently need to create a new object whenever we want to add a new modality

new_mdata = mu.MuData({
    'rna': mdata.mod['rna'].copy(),
    'new': adata.copy(),
    'prot': mdata.mod['prot'].copy()
})

Additional context

mdata = mu.read_10x_h5(gex_path)
adata = ir.io.read_10x_vdj(vdj_path)
mdata.mod['airr'] = adata.copy()
mdata.write('test.h5mu')

Full error message when saving the mdata after adding adata to mdata.mod

/users/rng/mambaforge/envs/compbio/lib/python3.10/site-packages/anndata/_core/anndata.py:1230: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  df[key] = c
/users/rng/mambaforge/envs/compbio/lib/python3.10/site-packages/anndata/_core/anndata.py:1230: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  df[key] = c
/users/rng/mambaforge/envs/compbio/lib/python3.10/site-packages/anndata/_core/anndata.py:1230: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  df[key] = c

[gtca]

Thank you, @racng!

As this seems to be a warning, it shouldn't generally stand in the way of adding modalities.

So far I see that this warning might be related to a scenario when the feature names are duplicated across modalities. In that case .varmap also looks fragmented (e.g. array([1, 0, 2, 0, 3, 0, ...]) and array([0, 1, 0, 2, 0, 3, ...]) for two modalities with the same var_names). This is not the case when creating a MuData object from modalities with these duplicated feature names straight away.

With no name duplicates, there should be no problem like this and no warning!

---

Version 0.3 of mudata will come with a fix to this warning — together with improved name duplicates handling so that varmap looks better and the behaviour is more intuitive when adding modalities. 🎉

[gtca]

This should be warning-free in v0.3 but please feel free to open a new issue if there's something else we can improve.