Fixing Hubbard formulation

[t-reents]

Addresses #1220

@bastonero I applied the fix as discussed in the issue. Moreover, I also added "ALPHA" to the supported values. Please have a look whether my understanding/logic for the dudarev formulation was correct.

One additional aspect:

I was wondering, do we actually need this check here:

aiida-quantumespresso/src/aiida_quantumespresso/utils/hubbard.py

Lines 65 to 66 in 0f4318b

	if hubbard.formulation not in ['dudarev', 'liechtenstein']:
	raise ValueError(f'Hubbard formulation {hubbard.formulation} is not implemented.')

Isn't this automatically validated in:

aiida-quantumespresso/src/aiida_quantumespresso/common/hubbard.py

Lines 151 to 152 in 0f4318b

	formulation: Literal['dudarev', 'liechtenstein'] = 'dudarev'
	"""Hubbard formulation used. Allowed values are: 'dudarev', `liechtenstein`."""

[bastonero]

thanks a lot @t-reents. would you be a lamb and add a little test when actually using the liechtenstein "keyword"?

[bastonero]

could you also add J0? https://www.quantum-espresso.org/Doc/INPUT_PW.html#id329
this should be the one used in liechtensteain

[bastonero]

re the ValueError: yeah, in principle it can be removed the check. it was put there in case we wanted to have other formulations (don't even know if there are so many others...), so that we didn't need to modify. but agreed it is currently unnecessary

[t-reents]

Ciao @bastonero !

Apologies, it obviously took my half a year to look into it again..

I should have addressed all your comments in ea518ed.
Could I kindly ask you for a review?

P.S.: In general, since I'm now trying to take care of the maintenance (since @mbercx left 😢), are you still available for Hubbard related questions/reviews, or are you no longer working on any of these tools? Just to know :)

[bastonero]

Thanks a lot @t-reents ! I think all in good shape. Just a comment, but not nothing really needed. I think we can merge this as is.

Re maintainance/hubbard related things: I now left Bremen and my new position is not particularly devoted to the maintainance of the plugin, although i am very happy to keep it going with some reviews, especially for the Hubbard side. Just ping me, and I'll let you know if in that specific moment i have time or not to do the review.

[bastonero]

I am not entirely sure what's more consistent with the Dudarev formulation, since I have never effectively used J (but only Ueff = U-J). I suppose we can leave this "open" to the user, and worst case QE will throw some error. Somehow as with pw.x, we do not want to do input checks on our side (at least, just the basics). So i guess for the time being this is fine.

[t-reents]

Thanks. I wasn't a 100% myself, as I'm also not a heavy user of the Hubbard part, but converged to this solution with Claude. I will indeed leave it as is and we can always come back. Thanks!

[t-reents]

Re maintainance/hubbard related things: I now left Bremen and my new position is not particularly devoted to the maintainance of the plugin, although i am very happy to keep it going with some reviews, especially for the Hubbard side. Just ping me, and I'll let you know if in that specific moment i have time or not to do the review.

OK, great – thank you! Indeed, my main idea was to potentially ping you for the Hubbard related aspects, since you have a lot of expertise there, and potentially ask you for a review. Other than that, I think that the maintenance work should be fine on our side.

Btw, all the best of course for the next chapter!!