Renaming global obs_names renames and reorders modality specific obs_names

[galenxing]

Describe the bug
Renaming global mdata.obs_names changes mdata['airr'].obs_names in an unexpected way

To Reproduce

import numpy as np
import pandas as pd
import anndata as ad
import mudata as md


# gex has 4 cells
gex = ad.AnnData(
    X=np.random.rand(4, 3),
    obs=pd.DataFrame(index=["cellA-1", "cellB-1", "cellC-1", "cellD-1"]),
    var=pd.DataFrame(index=["G1", "G2", "G3"]),
)

# airr has only 2 of those cells (subset)
airr = ad.AnnData(
    X=np.empty((2, 0)),
    obs=pd.DataFrame(
        {"VJ_1_cdr3_aa": ["CASSL", "CASRG"]},
        index=["cellA-1", "cellC-1"],
    ),
)

mdata = md.MuData({"gex": gex, "airr": airr})

old_airr_obs = list(mdata["airr"].obs_names)

m2 = mdata.copy()
m2.obs_names = [s.split("-", 1)[0] for s in m2.obs_names]  

new_airr_obs = list(m2["airr"].obs_names)

print('Pre renaming')
print("GLOBAL:", list(mdata.obs_names))
print("GEX:   ", list(mdata["gex"].obs_names))
print("AIRR:  ", list(mdata["airr"].obs_names))

print('Post renaming')
print("GLOBAL:", list(m2.obs_names))
print("GEX:   ", list(m2["gex"].obs_names))
print("AIRR:  ", list(m2["airr"].obs_names))

Output:

Pre renaming
GLOBAL: ['cellA-1', 'cellB-1', 'cellC-1', 'cellD-1']
GEX:    ['cellA-1', 'cellB-1', 'cellC-1', 'cellD-1']
AIRR:   ['cellA-1', 'cellC-1']

Post renaming
GLOBAL: ['cellA', 'cellB', 'cellC', 'cellD'] # expected
GEX:    ['cellA', 'cellB', 'cellC', 'cellD'] # expected
AIRR:   ['cellA', 'cellB'] # UNEXPECTED

Expected behaviour
Either throw a warning or error about renaming the airr obs_names

System
mudata version '0.3.2'