Closed xbc5 closed 7 months ago
Invidious is used to get all video metadata except the transcripts. This includes the thumbnail, duration and video description.
Timeouts occur often.
If timeouts occur, elfeed-tube should automatically switch to a different Invidious instance and retry. Multiple Invidous instances would have to be down simultaneously before there's a failure in fetching metadata.
could we put invidious behind a feature toggle? MPV supports direct YT links via YT-DLP
Invidious and mpv (as used by elfeed-tube) are not connected in any way. I'm not sure what you mean here.
the API has a monitor field that we could easily map, reduce, and sort at boot.
Yup, this could be done. If you would like to submit a PR, please go ahead. You could start by looking at the functions elfeed-tube--get-invidious-servers
and elfeed-tube--nrotate-invidious-servers
. In my experience simply trying instances at random has been working fine.
Invidious and mpv (as used by elfeed-tube) are not connected in any way. I'm not sure what you mean here.
I mean elfeed-tube-mpv
fails to launch MPV if the invidious instance times out -- no feedback (except the log buffer).
Yup, this could be done. If you would like to submit a PR, please go ahead...
I probably will, but not immediately.
Timeouts occur often. Do we use invidious for:
If 1: could we put invidious behind a feature toggle? MPV supports direct YT links via YT-DLP.
Another good addition would be a scoring system: the API has a monitor field that we could easily map, reduce, and sort at boot.
Could you clarify what purpose invidious serves within this package?
Thanks.