supporting AtomsBase

[rkurchin]

👋 Hi there!

As you may be aware, some of us over at JuliaMolSim have been developing AtomsBase, an interface package for specifying atomic geometries. I think Fermi.jl would be a great candidate to support AtomsBase and aid interoperability with other atomistic tools in Julia!

I thought I'd put this on your radar, and would be happy to discuss the best way to go about implementing the interface, and/or contribute to a PR to do so!

[mfherbst]

Quick comment: One thing I realised recently when thinking about the current state of AtomsBase is that for molecular systems a potential nuisance could be that we don't yet have custom system-level properties.
Point is that charge/multiplicity are usually stored on the level of a full system for molecular structures (in contrast to solid-state systems where this is more commonly dealt with by atom-specific magnetic moments / charges). Not sure this is a blocker, but something one should probably fix in AtomsBase first if there is interest on the maintainer side of Fermi.jl to integrate with AtomsBase.

[gustavojra]

Hi there,

Sorry about the delay. Fermi currently uses Molecules.jl to handle description of atoms, more specifically handling XYZ type data. Do you guys have some minimal examples of AtomsBase usage? I could make it fit into Molecules.jl.

[rkurchin]

Ah, I didn't notice that, my apologies. Probably Molecules.jl would be the right place then; I should have filed this issue over there. For more details on AtomsBase you can see the docs. At a quick glance, I think the way to do it would be to make your Molecule type a subtype of AbstractSystem and dispatch the interface functions on that by pretty trivial analogy with our FlexibleSystem example.