inetref column in record and recordedartwork tables are inadequate

[mythtv@…]

There is code in libs/metadatalookup.cpp that tries to guess what metadata grabber to use for a given inetref. It's complex and fails if you have a TV programme with season/episode set to 0.

I think a better solution would be to add a grabber column so the code does not need heuristics. After all the grabber is known when the inetref is found unless it's an inetref you type manually.

Also this will support more grabbers then just the two embedded in the code.

[stuartm]

An even simpler solution is just to prefix the inetref with a grabber specific identifier. e.g. TMDB_123456, TVDB_45346, IMDB_tt0113243

[mythtv@…]

Kind of agree but IMDB_xxx will need a lookup somewhere that goes from IMDB_ to grabber/imdb.py. Hopefully in a nice simple bit of python.

[Raymond Wagner]

There is no imdb.py, and it wouldn't make sense to require grabbers have the ability grabbers call other grabbers. Either alter the version output of the grabber to specify a tag that gets prepended by the metadata library, or just have the metadata library prepend the filename itself. If no one else has started on this yet, I'll take care of it this weekend.

[dekarl@…]

I just store (pseudo) urls in my guide database like

​http://www.themoviedb.org/movie/35619

​http://thetvdb.com/?tab=episode&seriesid=260426&seasonid=495783&id=4354633&lid=14

There are other suggestions for nicer pseudo urls available e.g. ​https://docs.metabroadcast.com/display/ATLAS/The+TVDB

Using working urls makes it super easy to go upstream to fix bad data, but for MythTVs use case the pseudo urls from the second sample might be nicer, although it uses three urls per episode (episode id, season id, series id).

[Frank Phillips <fphillips81@…>]

Following #12149 (thank you Raymond), the prepended grabber name has been implemented in recorded.inetref column.