If anyone wants to take a stab at this, I recommend setting up a local development instance (following the instructions provided in the lemmy docs) to test against. You don’t need to learn Rust to do this, but you will need to have Rust, PostgreSQL, Node and Yarn at your disposal. Alternately, you can use the docker method, but I find it much easier to introspect what the hell is going on with the database when it isn’t inside of a container.

After starting up the dev instance, you can use something like pgAdmin to poke around at your leisure and figure out the schema. You can also look through the Diesel migrations (written in SQL, used to initialize / upgrade / downgrade the database), however these are split up into dozens of individual updates chronologically and it is easy to overlook important bits of context (i.e. a table is created in one migration, a column is removed in another, two more columns are added in a third one, etc).

Edit: Speaking for myself, having no affiliation with Lemmygrad, I think creating dummy users would be excessive and complicated. It would probably double the site’s user count just to import the archive. What might be better is to use a single account named “GenZhouArchive” or something and add a string to the start of each comment indicating the original reddit username. Along the same lines, it might be a good idea to do the same with post/comment scores. Applying the Reddit comment scores to Lemmy posts/comments will flood the top sort heavily.