Clean up types and type annotations where possible. Not exhaustive -- there's some stuff that probably needs refactoring to make progress on -- but made some good progress, such that I think we can say the "types" issue is closed enough.
maybe important: by virtue of where json model dumps were happening, this introduced a small change to the order of the keys in some dicts that were getting sha512'd, which meant the ID changed during testing. I set JSON serialization to first sort keys in order so that this doesn't matter in the future, but this does entail a one-time change in digest ID. if that's bad, I can revert and try to figure out a workaround.
Also changed how stuff from GKS/VRS models are imported so that the code is a little more succinct. I think models.Variation style imports were a relic from a Python-JSONschema-objects custom and we don't need to do it anymore.
close #285
models.Variationstyle imports were a relic from a Python-JSONschema-objects custom and we don't need to do it anymore.