Been coming up to speed with SqlAlchemy and Alembic (for migration). I understand that the Alembic revision files get "locked down" as a form of documentation of the migration history. That said, it seems like there ought to be a way to "flatten" all the revisions at run time in a way that yields enough metadata to reduce the duplication/overlap I'm seeing between the Alembic DDL-ish python code and the SqlAlchemy ORM mapping model code. (Especially in cases where you're mapping python classes one to one with tables.)
Has anyone found a good way to do this?