Finally get to resolve scrobbling music from Jellyfin. This may lead to
other issues, in fact now videos seem to sometimes create duplicate
scrobbles. But music can be scrobbled now from Jellyfin web or Finamp
successfully.
This time, we simplify the progress updates, aggressively mark tracks as
100 played if they are marked played_to_completion, and implement a hack
for Jellyfin spamming us with progress updates less than a second apart.
This commit adds a lot of files, but most of them have no impact on any
other code. The thrust here is to start creating chart pages showing
which tracks and artists were most played for various time periods. Lots
still not working, but we're getting there.
Solution is identical to what we were already doing with videos. When
looking for existing scrobbles, don't filter by completion, but just
check if the scrobble was played to completion. This does create
another irritating situation where old scrobbles from days, months or
even years ago that were not played to completion will be resurrected
and made current here. But that's way less annoying than having spam
scrobbles at the end of every track.
We need to have a default value when we pop off a dictionary, and we
need to clear both mopidy and jellyfin status before we look a scrobble
up.
We may also want simplify status so we don't have mopidy and jellyfin cruft.
The issue here was that we update a Jellyfin scrobble to be complete
when it hits a certain threshold of percentage played (90) and then we
stop finding that scrobble while the video finishes playing. This
happens over and over again, so once a video reaches 90 percent played
we get dozens more scrobbles for each update as the video finishes
playing.
This is a crude fix, for the spam, as we'll end up "resuming" videos
that are stopped at 95 percent. So we need some way to mark the scrobble
as complete. I think forcing the percent to 100 on finish might work.
Turns out trying to not resume in-progress Scrobbles is super painful.
Maybe I'll come up with a better idea later, but for now, I'd rather
just resurrect old paused scrobbles of past tracks, rather than
completely mess up all other aspects of scrobbling.
Unlike Jellyfin, Mopidy's webhook only gives us a start and stopped call
to determine when a track should be scrobbled. This means we don't have
continous updating of playback ticks.
This commit adds a fallback when ticks are not there to use the track
duration and time since the scrobble was created. That said, this is not
perfect. If you pause the track and start again, the progress will get
very out of whack. But thankfully, Mopidy only sends us audio, and it's
rare that audio tracks are paused repeatedly and started again before
finishing a scrobble. So hopefully this shouldn't happen very often.
If you send Track data from the Jellyfin Webhook plugin, we'll do the
right thing with it. Lots more to do to clean this up, but it also
involved moduralizing the code for scrobbling so it's a little simpler
to understand what's going on.
