BUG: spin-orbit coupling failed with pydantic

[zhenyaofang]

Hi,

I was wondering if atomate2 is currently implemented with VASP calculations with spin-orbit coupling.

When I ran calculations with SOC, I encountered a validation error from the pydantic module; in SOC calculations magnetic moments for each atom are a list of 3 float numbers, but it seems that pydantic only accepts a float number (see attached screenshot).

Is there a recommended way to address or work around this issue? Thanks very much for your help.

[esoteric-ephemera]

Hey @zhenyaofang this is actually an emmet-core issue, would you mind posting there? The document models for VASP calculations are in emmet-core

Since these doc models were written around Materials Project workflows, which generally only use colinear spin calculations, it's not surprising that the NCL input / output would not be supported

[zhenyaofang]

Hi @esoteric-ephemera, thanks very much for the information.

No problem, I will post this issue at emmet and see if they have any clue.

[mkhorton]

Just a note that maybe I haven't actually PR'd this yet but I think it's just a one-line change to the document model to allow for vector magmoms.

We actually don't have SOC workflows yet, but I also added a PR to custodian: materialsproject/custodian#369

And I think that's more or less all you need. There are a few different ways a SOC workflow could reasonable be added to atomate2 and I hadn't figured out the best way to do it yet.

[esoteric-ephemera]

A one-line fix is also possible, but for parquet reasons we're trying to avoid union types