I’m getting a lot of garbage results when using Usenet Crawler as an index. I think it’s just that indexer, but not totally sure. In some cases, the search results are rejected, but today a LOST miniseries pack started downloading in response to a single episode search (!!!).
See this screenshot for some example search results. I tried doing the exact same search directly at that indexer and those results don’t show up.
If you enable debug logs you can see the URL drone built to search, most searches use a tvrage id, but if they have bad results mapped to it they will return that garbage, title searches aren’t as reliable, but never seen it THAT bad.
If you out the URL from the logs and put your api key back in, into your browser you can see what they returned.
Wow so usenet crawler is returning garbage itself. Exact URL is:
It’s a title search but the same title search via the website does not return those results. So… I played with the web search to get the same “bad” results. The website has an option for “index” that defaults to “release names & files” but has another option “all” which includes searching NFO files. Changing it to that returned some of those garbage nzbs.
This begs the question, does the API allow you to specifically target the release name and file names? Perhaps that API invocation is searching NFO files too, and with a generic title like “last leg” is matching random scene text in those files?
Is there a way to get NZBDrone to use the TVRage ID in this case? The entry exists (http://www.tvrage.com/shows/id-32497) but NZBDrone isn’t using it so may not have linked it to the entry in my DB. That may also fix my other issue where the title doesn’t match because the scene releases the show as “The Last Leg with Adam Hills”.
How should I best give you info for when the bad shows are actually downloaded. I did a full series search for “The Last Leg” and the following got sent to my clients:
I then edited the DB to insert the TvRageID but still go the same results. Even if the indexer is returning bad results, how did those 4 items match the title series close enough?
I don’t see an option to change how it searches via the API, but maybe there is one.
You can make a request to trakt to have it added on their site: http://support.trakt.tv - once its added drone will pull it in the next refresh.
drone won’t grab an episode via a search unless its the one it searched for, if you’re seeing different behavior, please provide debug logs of the search and subsequent download.