Fast generation of GraphQL schema from migration

GraphQL schema can be used for static analysis.

To get the schema from migration files, Hasura and postgres should be running, and migration tasks have to be completed, and agent(e.g. user or program) should know when this is done(or try polling).

However, I 'guess' it might be theoretically possible to generate only GraphQL schema from migration files without fully running Hasura. If possible, this would save considerable time.

Of course, depending on purely migration files would not reflect possible affection from metadata call, but I think it's anyway meaningful.

Would it be possible?