preliminary test for 0-orbital Atoms

[zerothi]

@tfrederiksen @pfebrer

this is a first stab at 449. I am starting to grow on the idea.

However, as you can see in this PR, there are many places in the code base where it assumes that there exists some orbitals on atoms.

Whether or not these should be defaulted to be present or not is something that needs to be discussed.
However, it seems that one should be explicit on defining orbitals if one wishes to use them.
Many places the code simply fails because it now requires some kind of explicitness.

But I would agree that this could be useful. If anybody cares to expand/fix/add tests for the remaining details, feel free :)

• Closes Atoms without orbitals? #449
• Added tests for new/changed functions?
• Ran isort . and black . [24.2.0] at top-level
• Documentation for functionality in docs/
• Changes documented in CHANGELOG.md

[pfebrer]

I want to push this forward to use it for QM/MM geometries, should I start a new branch and create a separate PR?

[tfrederiksen]

I haven't been looking much into this (sorry), so from my side you can, of course, proceed as you see fit. I'm still very much interested in this development in sisl!

[zerothi]

I want to push this forward to use it for QM/MM geometries, should I start a new branch and create a separate PR?

Yes, please do, branch this off would be ideal imho.

However, I think it would be very beneficial to discuss here, or reiterate the decisions made in #449. Eg how sparse matrices etc should handle things.

[zerothi]

Or perhaps better, finish the discussions in #449, then let implementation details be discussed in the pr!