feat: use rfcauthors for nomcom eligibility

[jennifer-richards]

Refactors computation of the authorship path for nomcom eligibility to fix up some stale logic. Uses RfcAuthor records instead of DocumentAuthor for RFCs that have the former.

[rjsparks]

This will work best if the creation of rfcauthors via merged rfc and datatracker data preserves datatracker person links for authors on the last draft that remained authors on the RFC. I believe that merge has that property, but if it doesn't we should look to fix it.

[jennifer-richards]

This will work best if the creation of rfcauthors via merged rfc and datatracker data preserves datatracker person links for authors on the last draft that remained authors on the RFC. I believe that merge has that property, but if it doesn't we should look to fix it.

Assuming you mean the place that DocumentAuthor records are created for the Document(type="rfc"), then yes, those are exact copies of the records that were tied to the final datatracker draft at the time of publication.

When we start creating RfcAuthor data we'll need to be careful to maintain the Person links, but afaik we have not done that yet. If I've missed something, please let me know.

[rjsparks]

When we start creating RfcAuthor data we'll need to be careful to maintain the Person links

Is what I was talking about. I believe the merge we're doing creating purple's database has this property - it is intended to, but we should verify that nothing is perturbing it.