Atom labels in CIF file are silently rewritten by CifWriter about pymatgen HOT 6 OPEN

This also struck me as odd. I'm currently working on #3767 which touches this bit of the code and happy to fix this if someone tells me the expected behaviour for magnetic moments.

from pymatgen.

This is because if "magmom" in site.properties is true for this structure at https://github.com/materialsproject/pymatgen/blame/2d008e0dd5c430692e8dcac2505340a6bdff1642/pymatgen/io/cif.py#L1516C21-L1516C51

But I am not writing magnetic moments, and I do not know why that should override labels.

from pymatgen.

I confirm that removing the magnetic moments with structure.remove_site_property("magmom") does fix the issue.

Regarding a PR, my own understanding of what the code tries to do for magnetic moments is insufficient to handle it well. I wouldn't want to break another use case…

from pymatgen.

@fxcoudert thanks for reporting. I assume a pull-request to fix this issue would be very welcome.

from pymatgen.

I confirm that removing the magnetic moments with structure.remove_site_property("magmom") does fix the issue.

Regarding a PR, my own understanding of what the code tries to do for magnetic moments is insufficient to handle it well. I wouldn't want to break another use case…

I see! I also don't know that functionality. Should I leave it open to see if someone else can fix it?

from pymatgen.

@mkhorton , maybe you now more about this.

from pymatgen.