Add imdb, tvdb and tvmaze to feed base release

[Celsiusss]

Description

MTV uses the BaseFeedIndexer. So now imdb, tvdb and tvmaze IDs gets added to MTV releases.

[ilike2burnthing]

Thanks for the contribution.

Anime Tosho (the other indexer which uses BaseNewznabIndexer) returns the attribute imdb, rather than imdbid (although I'm not sure they actually use it).

The Newznab docs specify that it's imdb, rather than imdbid. As an example, Jackett returns both, imdb for the 'short' (numerical) ID, and imdbid for the full ID (not sure if others do the same). We'd need to be able to handle both.

The only other ID attribute the docs specify is for TVRage, so not TVDB or TVmaze. Either we'd need to stick to the specs here, and then add overrides to the indexers, or we'd need to add all ID attributes here. I'd be leaning towards the former.

@garfield69 / @mynameisbogdan thoughts?

[garfield69]

IMO The basenewsnabindexer should stick to the newsnab specs.
if MTV has added other ids to their service then they are now in custom territory, so the MTV indexer should apply overrides to support these non-newsnab features.

[Celsiusss]

@ilike2burnthing @garfield69
Thanks for the feedback.
I moved tvdbid and tvmazeid into the MTV sub class, and kept imdb in the base class.

[mynameisbogdan]

https://torznab.github.io/spec-1.3-draft/torznab/Specification-v1.3.html#predefined-attributes has them.

This is how Prowlarr treats them: https://github.com/Prowlarr/Prowlarr/blob/6449b89eb647e5ce7ff98385952dc17fe19c0c30/src/NzbDrone.Core/Indexers/Definitions/Newznab/NewznabRssParser.cs#L109-L114

[ilike2burnthing]

Should we rename BaseNewznabIndexer to BaseTorznabIndexer?

[mynameisbogdan]

Not really, except for the download protocol they are pretty much the same.

[ilike2burnthing]

Is there a reason Prowlarr only accepts rageid and not also rage?

Should traktid/trakt be included, as it's not part of the spec?

[mynameisbogdan]

Seems like that was a fix for some indexer.

Prowlarr/Prowlarr#1261

[garfield69]

Apologies to Celsiusss from me, It seems I did not read the current specifications document properly. I jumped to the external references section that pointed to an old newsnab spec https://torznab.github.io/spec-1.3-draft/external/newznab/api.html#predefined-attributes

[garfield69]

As for Traktid, doubanid and tmdbid, I added them to our torznab output along with some categories that were not in the 1.3 specs. I figured since it was unlikely anyone was ever going to go beyond the 1.3 draft specs I might as well add some new ids and cats that were becoming prevalent at some sites, and as long as they did not clash with existing specs they should be be ok.

[ilike2burnthing]

So going forward, do we build on the specs and include Trakt, TMDB, and Douban in BaseNewznabIndexer (and NewznabRssParser), as well as accommodating non-standard attribute names such as tvdb and rage?

[garfield69]

My approach up to now has been to only include a new id only when I discover that there are torznab/newsnab indexers in our stable whose web sites either allow searching for, or present ids in their results that we reprocess into torznab/newsnab output.
What I had not appreciated up to now, was that the torznab 1.3 specs also included an upgrade to the newsnab specs to 1.3 too ;-)

[garfield69]

as for whether to add ids without the id, up to now I had thought that the only variance was imdb, which also had imdbid, one being the number only and the other the ttnumber. All the other ids had the id suffix, and I did not think we should have added the same information under a different name.

[ilike2burnthing]

Prowlarr/Prowlarr#1261 and Prowlarr/Prowlarr#656 were both for newznab indexers, so yea, doesn't really affect us then.

So just rageid, tvdbid, tvmazeid, imdb (short), and imdbid (full) added to BaseNewznabIndexer then?

[garfield69]

yeah

[Celsiusss]

Alright great!

[mynameisbogdan]

I did a rebase, so do a git pull.

[garfield69]

Thank you.

[garfield69]

v0.21.2387