"aka" alias processing

[brendano]

tracking the state of "aka" alias processing from the FE database.

Last I heard, Brian added in "aka" clauses into the FE spreadsheet. I gave him some loose guidelines how to use it. I asked him to not bother with crossproducting out alternate first names and such; the semantics are super duper basic, just a list of full names. Katie added in processing when creating the fe-all.json files (which are extracted from the FE google sheets CSV), which now should have an alias set for each one, as well as defining a single canonical name. (I think that canonical name is just the name field we've been using all along...)

The end-user interface (and/or evaluation scripts, debatably, but longer story) are where aliases need to be used, for

1. when making the "in FE?" determination for an entity.
2. for grouping together mentions for a single entity (when different mentions use alternate names). this must be used for both (a) display, and (b) the noisyOR entity prediction calculation.

Status questions/notes

• What's the status of alias processing in the new webapp interface?
• The old HTML report does NOT have alias processing. I plan to never add it, assuming this interface will soon be deprecated in favor of the webapp.
• The evaluation script does NOT have alias processing either. Definitely we should stay away from that for now, but just FYI for the future if we return to it.