Mostly column, property and table names had to be updated to follow the
schema. A few other identifiers have been renamed as well for
consistency, although this was not strictly necessary.

It is never used and never passed, but causes a TS error.

The error message is already specific enough, no need to make it longer.

In the end we of course want to reuse `bbid` and `data_id` of accepted
entities instead of duplicating the data and deleting the original.

Also make `ImportMetadataT` more versatile/useful.

It is not needed and it is causing a type error since the last commit.

The rule to disallow trailing commas often forces you to pointlessly
touch neighboring lines when you add or remove items.

If we want to enforce consistent dangling commas, we should rather force
them to always be present. But since our whole codebase follows this
rule, it would be better to have a grace period to gradually add commas.

metabrainz/bookbrainz-site@2fc11a2

Run the kysely-codegen task to regenerate the types.
Don't forget to reformat the generated code to follow our linting rules.

Since this is using DB introspection, you have to specify a DB
connection string via the --url param or the DATABASE_URL env var:
https://github.com/RobinBlomberg/kysely-codegen

Rather than collecting all properties of the pending entity data and
creating a new entity (with new BBID), we simply create an entity header
and a revision which link to the existing data.
This is a complete rewrite using Kysely instead of Bookshelf/Knex.

The final loading of the entity model (for search indexing) is out of
scope for this function and has been moved into bookbrainz-site.

From an older schema, there are still editions without edition group.

The `pending_entity_bbid` column is no longer set to NULL on approval,
we have to additionally check for NULL in accepted_entity_bbid to find
pending imports.