Regex or other automated checking/marking of nonexistent torrents

(I've been going back and forth briefly on this with Taiga, and I just wanted to submit it as an official bug fix/feature request for the record)

There are a lot of torrents that link back to dead files, either via Tokyotosho to Nyaa (since that's where the release was listed and where Shanaproject gets its source of releases, I'm guessing) or directly to Tokyotosho. If a user clicks the green arrow link to direct download a torrent, they are directed to the error page that states the torrent does not exist. If a user clicks the yellow arrow link to send to RSS, it is sent but since the torrent no longer exists, nothing ever happens with it.

I propose that a fix be implemented in the form of a regular expression or some other script that runs periodically - daily, weekly, biweekly, whenever works best - that refreshes the links from which Shanaproject goes to find the torrent file, and if it gets a phrase that matches the Tokyotosho or Nyaa error text, the torrent is marked as potentially dead. A volunteer staff could review the marked torrents to confirm that they are dead and blacklist them from public Shanaproject listings.

(I am no developer but I'm 99% certain that Powershell can do this... if it helps I'd look into throwing together a script, but if Shanaproject doesn't run on a Windows backend it'd do very little good)

This would potentially clean up users' searches and allow for a more seamlessly automated experience. If the size of the Shanaproject DB is an issue this would be a great way to reduce it, along with IOPS that go into each search as inbound queries hit the DB. Dunno if that's something you have outsourced and thus get billed for but it'd be a nice way to improve the bottom line with a slight bit of code.

Comments (2)