Data from some of the repos is remarkably inconsistent (ids don't match associated images, and various other problems), so trying to organize a mechanical import is rather nightmareish. I've spent most of the day discovering these inconsistencies the hard way, as well as troubleshooting some fragmentary and unlinked entries added to the database by interns in training, which really shouldn't be there. Mostly unproductive time, all of it. Memo to self: resist pressure to start a project too quickly without proper planning. This should be applied immediately to ACVI, which is in danger of falling into all the same traps.