954 modified Atoms.index

[tfrederiksen]

• Closes Improve Atoms.index usability #954
• Added tests for changed function
• Documentation for functionality in docs/
• Changes documented in changes/<pr-num>.<type>.rst

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 86.91%. Comparing base (5be4413) to head (e1d040c).

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #955   +/-   ##
=======================================
  Coverage   86.91%   86.91%           
=======================================
  Files         412      412           
  Lines       54333    54349   +16     
=======================================
+ Hits        47221    47237   +16     
  Misses       7112     7112           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[zerothi]

First, thanks!

I've thought about it, and the way it looks is somewhat not coherent with the other routines which only allows the list-like entry.

I can adjust the PR if you'd like? Then I'll ping you for comments? Thanks!

[tfrederiksen]

I can adjust the PR if you'd like? Then I'll ping you for comments? Thanks!

Maybe I can first give it another try taking into account your comments?

By the way, I've always been a bit confused about species vs atom. What's the distinction?

[zerothi]

So the species is an internal reference to the unique species in the Atoms object. So if you have H(szp) and H(dzp) then you would have 2 species. However, the Atoms object will only have 2 unique Atom instances (one for each). And then the species index will be an array of #-of atoms, and each value will be 0 or 1, depending on which atomic species it is.

So internally, sisl finds which atom is corresponding to the requested one, and by that it has the species index, then it just does (atoms.species == species_index).nonzero()[0] and you have all the indices of that atom. Hope that makes sense?

[zerothi]

minor nit-picks... ;)

[tfrederiksen]

Should we also add an example in the docstring how to use this?

[zerothi]

that would be great! :)

[zerothi]

@tfrederiksen do you know why it stopped? I think it's good, so I'll just merge, unless you object?

[tfrederiksen]

@tfrederiksen do you know why it stopped? I think it's good, so I'll just merge, unless you object?

No, I don't know why. From my side it is ready.

[tfrederiksen]

Hang on a sec...

[tfrederiksen]

I think I found an issue, with multiple atoms one would end up with a list of arrays. I think with np.concatenate is should now be OK (now also tested with a separate script manipulating a geometry).

[zerothi]

Thanks!