It seems that there is a bug to write the lobsterin file

[RedStar-Iron]

pymatgen/pymatgen/io/lobster/inputs.py

Lines 271 to 272 in 15dc77d

	if key.lower() in [element.lower() for element in Lobsterin.FLOAT_KEYWORDS]:
	f.write(f"{key} {self.get(key)}\n")

key in self.get(key) should be lower, then it can be found in Lobsterin.FLOAT_KEYWORDS.

And the cohpsteps might need to be a positive integer.

So I just revised it to:

if key.lower() == "cohpsteps":
    f.write(f"{key} {int(self.get(key.lower()))}\n")
else:
    f.write(f"{key} {self.get(key.lower())}\n")

[JaGeo]

Thanks for looking at the Lobster part of the code. @naik-aakash and I will check your suggestion carefully.

[JaGeo]

I think the bug is coming from the fact that the overwriting of the dict. We will look into it more carefully to not break other parts of the code!

[naik-aakash]

Hi @RedStar-Iron, Thanks for the suggestion. Can you also please post the steps to reproduce the issue? It would help me address it properly

[RedStar-Iron]

Hi @naik-aakash ,

When I run the code below, it gives me None for the FLOAT_KEYWORDS.

    structure = Poscar.from_file("POSCAR").structure
    potcar = Potcar.from_file("POTCAR")
    lobsterin = Lobsterin.from_file(file)
    basisfunctions = lobsterin.get_basis(structure, potcar.symbols)
    lobsterin['basisfunctions'] = basisfunctions
    lobsterin.write_lobsterin("lobsterin")

[JaGeo]

Yup, I have also already reproduced the issue and have started with a fix. Will start the WIP pull request in a minute.

[JaGeo]

@RedStar-Iron we have worked on a fix for both of your issues. However, we investigated that it does not matter if cohpsteps is a float or int (see here #3439 (comment))