Only transfer obs columns to pseudobulk when they have one unique value when grouped

[LuisHeinzlmeier]

Description of feature

At the moment, we take the first value of each obs in the grouped DataFrame when creating pseudobulks. This could be misleading if a grouped obs shows more than one value. I would only include the obs that show one unique value when grouped. I'd be happy to revise that if wanted.

pertpy/pertpy/tools/_perturbation_space/_simple.py

Line 152 in b222328

grouped = original_obs.groupby(grouping_cols, observed=False)[missing_cols].first()

[Zethson]

I think this whole thing should ideally just wrap scanpy get https://scanpy.readthedocs.io/en/latest/generated/scanpy.get.aggregate.html if possible. It's old legacy code.

[LuisHeinzlmeier]

Then I try to update this!