Interoperation between ACSets, FinSets, and FinCats

[epatters]

We have dedicated, efficient data structures for acsets. It is also the case that a FinDomFunctor sending morphisms to FinDomFunctions can be regarded as an acset. These two approaches to the same concept should be made interoperable. The ongoing work on data migration (e.g. #532) makes this very relevant, since data migrations naturally return FinDomFunctors but we ultimately want to get back an acset.

One direction is easy: I have already made a FinDomFunctor subtype that wraps an acset. Doing the other direction well may require some internal refactoring. @olynch's suggestions in the related issue #489 would actually jive very well with these requirements.