RandomNinjaAtk
commented
1 year ago If its a bash script, it can be easily integrated, if you want to develop it.
The problem with multiple sources is your likely to end up with lots of duplicates because each source names them differently.... And the reason I never opted to just pull from the Artists linked channel is because there is no moderation to know if the video is what you want... So using something like IMVDB lets you filter out the noise and hopefully only grab the stuff you should want...
You can write a script within the container /config/extended directory and test from within the container using the CLI. It's how I write/test scripts...
For lots of code samples, just look through the existing scripts because if you take the Tidal script as a base, a lot of the work is likely done for you... Since it already processes each artist, you just need to change the source and etc.
I will welcome a PR for a script or if you can provide code samples, I can possibly look at writing something. But with 3 different video scripts, probably would need to make them all optional and let the user choose using the config file which ones to use if any at all...
Right now I think it does a decent job at using both at the same time... And we're likely to never have a perfect solution because it's a mess to start with...
Hello Mr. Ninja,
While using the Arr scripts more and more, I've notived some issues. Like many music videos are not being downloaded due to geoblocking, a deleted user account, a private video. In case a video is downloaded, sometimes IMVDB only has the link to a poor quality version uploaded by a private user 15 years ago, while a hires/HD remaster is residing on the artist's channel. IMVDB seems to take forever with upates (I wonder for how long Massive Attacks 'Angel' is misspelled as 'Angle'), so I quiestion that IMVDB as a reliable source. Also some artists do release videos, while no one keeps track on IMVDB. Adding those videos to Musicbrainz takes forever, and can't replaced the outdated videos yet.
I was using the Tital video script recently and noticed its behaviour. With some artists like Weezer, 4:3 videos have been uploaded in 16:9 format with black bars, which overwrote my good quality YT version, as the file size was bigger :-( Also it takes all videos into account, also live versions or TV broadcasts or one hour long interviews (see Interpol, video "Interpol") I don't want. As Tidal is a mess considering music videos, I'll skip it for now.
So here's the idea: Use Youtube Music to get the official music videos. I've written a python script which is able to get all official music videos of an artist's channel by the channel slug, custom_url or LidarrArtistTag. I'm currently including it into the VideoService, and I'm not sure what's the best way to so. I'm thinking of the following possibilites:
I think I'll get some kind of cache working first like ImvdbCache. See below the log of my current test script for the artist Air. I'm also testing a relevancce system to filter out unpopular videos. Do you have any ideas regarding the best integration approach?