It looks like the Wayback Machine indexed the site quite a few times. Although this does not include the download themselves.
https://web.archive.org/web/20220905221931/https://dffd.bay12games.com/index.php
(last update September 5, 2022)
Where the archive files also stored in the DB? (most likely not)
I'm curious how big the archive of the whole server is? How many GB have been uploaded?
The Wayback Machine archive could be useful for people to recover the file descriptions and some other data for whatever they cover. Technically I might could work up a script to scrape certain info for each file listing and user, assuming they're all stored there, but it would take a ton of work and time. And a lot of it would then require further processing. Like file descriptions needing to be converted back to BBCode from HTML; yuck. And since the user information wouldn't have their email address or password, there would be no reliable way to recover the actual accounts, so no real value to those.
Probably simplest for me to just provide a link the the Wayback Machine for people as a way to recover their listing data manually.
The files aren't stored in the database. As mentioned, all of the up to date user uploaded files are safe and will be maintained. The collective size of all the uploaded files is about 372 GB.
file creation times might help to rebuild the IDs...
No need, the user uploaded files are stored with their file ID as the filename on the server. The actual filename which you end up downloading them as is stored in the database.
And on that note, slightly better news. Since I have the abbreviated JSON file data available for each file, stored as actual static files, I should be able to use the data in those to restore some information to the database for each currently orphaned upload. The correct download filename, the author's name (not their ID, but I can match those at least for user accounts that existed back in 2015, otherwise just show an unlinked username), the last update timestamp, the file version if one was provided, and the DF version the file was for. And the file size, which could alternately be determined by checking the file itself, but no need with that info available here.
It will still be missing their file listing title, description, category ID, original upload date, file homepage info, rating, download and view counts, and a couple other bits of info.
So for orphaned files uploaded by those user accounts that existed back in 2015, I should hopefully for the most part be able to link those file listings back to your accounts.
The work continues.