I have several hundred items with files that don't have any file information. Most importantly for my purposes, the mime types are missing from the file record, which causes a number of issues on the front-end. Is there a way to get Omeka to re-process files and fix this issue?
Thanks -- E
[PS: I have no idea how they got like that... this is an older site that I'm bringing up to date]
Not sure if it will do that, but I'd try using the Derivative Images plugin to see if it does the trick.
Ugh. Turns out some volunteer developers manually entered 0/null/empty/placeholder values in a bunch of rows in the files table during the initial migration of the data into Omeka (the original site dates back to the 90s), so there's probably no easy or cool way to fix this (though I did try the very excellent Derivative Images plugin to no avail). In case you're curious, I think they populated the database with thumbnails and entered falsey values for all the other sizes. Philistines.
Out of curiosity, in the files table, for the offending images, is there anything in either the
No values in the
metadata columns (or
authentication, etc). Some have legitimate (?) filenames, and some say nopic.jpg. The ones that have a "real" filename seem to be missing the derivatives, though they're roughly thumbnail sized. It's odd. Does any of this info sound like a potential clue?
For what it's worth, the site underwent an aborted upgrade some time ago, but –from what I can gather– continued to be used with a mixed codebase and partially-upgraded database.
double ugh. That's a clue of sorts, I guess. It sounds like the migration didn't use Omeka's code to ingest files, and didn't quite manage to reproduce the structures it would have produced, both in the database and in the file directories.
I'm doubtful, too, that there would be a reasonable way to get the originals, match them up with the items, and generate the correct data. It's certainly beyond my ken to think of an approach.