HMS-5947: fix delete-snapshots glitchtip errors

[xbhouse]

Summary

• Adds a step to delete the publication before deleting the repo version as part of the delete-snapshots task. Without this, the task could fail, leaving snapshots stuck in a soft-deleted state
• I was only able to reproduce this with Red Hat repos, which seems like what we're seeing in stage

Testing steps

1. Import and snapshot a few RH repos. You'll need at least 2 snapshots for a repo, and it might take at least a day to get updates (there might be a better way to force an additional snapshot of RH repos?)
2. Once you have at least 2 snapshots for a repo, update the date of the older one to be marked for deletion: UPDATE snapshots SET created_at = (NOW() - CAST('365 days' AS INTERVAL)) WHERE uuid = '<snapshot_uuid>';
3. On the main branch, running go run cmd/external-repos/main.go snapshot-cleanup should fail with this error and snapshots are soft-deleted, but still exist in the content-sources database: [Finished Task] task failed error="error deleting rpm repository versions: 400 Bad Request: [\"The repository version cannot be deleted because it (or its publications) are currently being used to distribute content. Please update the necessary distributions first.\"]"
4. Checkout this PR, and rerun go run cmd/external-repos/main.go snapshot-cleanup. The task should complete and the outdated snapshot should be removed from the database

[jlsherrill]

https://issues.redhat.com/browse/HMS-5947

[rverdile]

would you mind adding an integration test for deleting red hat snapshots? There's an example of creating two red hat snapshots in the TestUpdateLatestSnapshotForRedHatRepo() test