read_hamiltonian() does not work with lanthanides?

[novidad21]

Describe the bug

Reproducable code

import sisl
fdf=sisl.get_sile("PATH_TO_FDF")
ham=fdf.read_hamiltonian()

Freshly installed siesta from master brench yesterday. After running a calculation with SpinOrbit true using psml pseudopotentials downloaded from dojo-pseudo.org the hamiltonian cannot be read:

ValueError: AtomicOrbital does not implement shell h and above!

Version details
3.9.12 (main, Apr 5 2022, 06:56:58)
[GCC 7.5.0]
0.0.0.dev+eacb8351ff416ad9f7f93732777c3d7931c5f711
lase.zip

[zerothi]

Yes, currently sisl does not implement the h-shell.

[zerothi]

Could you please try the latest commit? It should work now :)
Thanks!

[novidad21]

Thanks @zerothi it seems the issue resolved. However I still don't really understand the issue here, because in the produced pseudopotential files I see no indication of h shells, it does not go above l=4. But anyway, no it works! Thank you.

[zerothi]

But if one of the l=4 shells gets polarized, it will be an l=5. Probably that is why. Weird anyway, I haven't had a look yet. Thanks!

[novidad21]

Yes, I assumed the same, so I switched to DZ, and still I ran into the same problem.

[zerothi]

Hm. Sounds like this really needs to be checked!

[zerothi]

I'll just reopen to make me remember

[zerothi]

This was related to #462. So fully fixed. Can be closed now.