Record authors and maintainers for each architecture

[Luthaf]

This way we'll know who to ask for review when someone tries to modify some architecture, and we can include the authors list in the metadata.

I went the the following assignment, but it's only what I saw! If I missed any contributions please let me know and I'll add it to the files!

SOAP BPNN: @frostedoyster is the only author and maintainer
Alchemical: @abmazitov is the only author and maintainer
PET: @serfg, @frostedoyster and @abmazitov are authors (is this right?) and @serfg and @abmazitov are maintainers.

TODO:

• read this data and show it in the docs

---

📚 Documentation preview 📚: https://metatensor-models--145.org.readthedocs.build/en/145/

[frostedoyster]

The tests are also failing on main, we're looking into it.

Are we sure that we need authors? How is this useful for the user?

Regarding the maintainers, perhaps it would be a good idea to only have one maintainer per architecture. If the maintainer needs help, of course they can ask other people. But in principle the user only wants one e-mail address, and having more maintainers IMO makes it too easy for each maintainer to think the model is someone else's responsibility

[Luthaf]

Are we sure that we need authors? How is this useful for the user?

Not extremely useful to the end user, but nice for the authors to get recognition.

I don't mind too much for maintainers, I agree that we should not have 50 different maintainers, but up to 3 seems reasonable to me. We can discuss this in tomorrow's meeting.

[frostedoyster]

Also, the maintainer should at least review the PRs to their model (#131...), otherwise how can they maintain code that they've never seen?

[Luthaf]

Yes, that's what the CODEOWNERS file should do. I'm not sure if it will just ask for review but allow to merge without one from that person, or require a review from that person for merging. There might be a setting for it.

[frostedoyster]

Works for me