I am looking to importing entity_id as an attribute when opening structures inside of biotite / MolecularNodes. The idea is discussed here. I am able to achieve this without too much trouble from mmCIF / PDBx files, as below.

mol_pdbx = pdbx.get_structure(file_read_pdbx, extra_fields = ['label_entity_id'])

In [67]: mol_pdbx.label_entity_id
Out[67]: array(['1', '1', '1', ..., '129', '129', '130'], dtype='<U3')

Currently inside of MolecularNodes however, when I download from the PDB I download as mmtf files for bandwidth and speed. The entityList can be included but is not required in the mmtf spec. From what I can tell, it isn't included when downloading, if that is the case I can switch to mmCIF for downloading to solve the problem, but I was wondering if it would be possible to change the request to the PDB to maybe include other fields in the mmtf that is downloaded? I haven't dealt at all with the API hands on, so I am unfamiliar.

Kind of a long and rambling question, but I am guessing the best method going forward will likely be switching to mmCIF downloads.