Database Migrations

All changes to database schema are handeled with the Alembic, wrapped into [flask-migrate]. The modular concept of beavy makes that a little tricky. In order to keep the database clean on those things you want to have only, we are keeping all migrations that belong to their own modules models inside that model. Which means [alembic runs with multiple heads]( simultanously. That sounds complicated, but really isn’t that much. It just means you need to be careful when creating migrations to place them in the right module with the right dependencies.

Creating new database migrations

Note: make sure your database is on the latest version by running python db migrate heads

If your changes should be part of the main beavy module, create the migration running (within virtualenv):

python db migrate [email protected]

This will create a new version in migrations/versions. Please review it before applying or committing.

Creating migration within a module

In order to ensure consistency between modules and main, you need to specify the current version as a depency for migrations in modules. for that, first find the current revision of the main project:

python db heads | grep beavy\)
=> 242c2fd98af (beavy) (head)

In this example 242c2fd98af is your head revision. Now copy that as a parameter to the migration for your module. In this case we did changes to the 'beavy.likes' modules (you can see all heads with python manager db heads). The command looks as follows:

python db migrate [email protected]

This should create a new revision in beavy_modules/likes/migrations with all changes detected. We now need to add the dependency on main, for that open the newly created file. And add depens_on = "242c2fd98af" in the header as down_revision. And review the rest of the migration.

Now this revision, although only included if the module is activated, will apply those revisions in the proper order. Try it:

python db upgrade heads