Each sql file in manage.py is run in the same way. We can abstract this to a function to make sure that we do things consistently.

Also do this in listenbrainz and messybrainz (and critiquebrainz and metabrainz?)