beetbox
Add initial version of the Listenbrainz plugin
Description
Fixes #1719
I added a basic version of the LB plugin that imports the play history. It uses the same process used in the Lastimport plugin to update playcounts.
There are few additional functions that I added here to import playlists created by Listenbrainz. These playlists can then be imported in other programs such as Plex.
To Do
- [x] Documentation. (If you've added a new command-line flag, for example, find the appropriate page under
docs/to describe it.) - [x] Changelog. (Add an entry to
docs/changelog.rstto the bottom of one of the lists near the top of the document.)
Awesome! This already looks fantastic! No objection to merging right away if you think this is at a usable initial state.
I also noticed this contains a few changes to the last.fm plugin. Was that intentional? I didn't review them carefully, although they seem credible. If they fix bugs in the plugin, maybe that deserves a changelog entry of its own.
@sampsyo Yes, I have tested it, and it is ready to be merged.
I made minor changes to the process_tracks function in lastimport to improve error handling and reuse the same function. I am not entirely happy with the process_tracks function as it does a poor job of matching (e.g., right now, it only matches 3 of the 25 tracks even though they are all in my library). It may be due to the fact that I am using Spotify as the tag source. I do want to create a better matching function, but that can be done in a separate PR.
Sounds good! Would you mind summarizing the changes to lastimport in their own changelog entry? Then this is clearly good to go!
Correct me if I'm wrong, but this seems to be fetching the whole ListenBrainz history on each run - could we leverage min_ts somehow to avoid that maybe?
Yes, all the elements could (and should) be configured. We still need to carefully think about how to use them (which is why I called this PR initial version 😉). I honestly don't use the play_count information, but I am sure someone will find it useful and update the code to use those variables (heck, even I may do it at some point if no one gets to it).
The current code has been reviewed and approved by @sampsyo, so I would not change it. Once this PR is merged, we can create a feature request for additional configurations and discuss the best approach to implement the same.
I'll merge this then, since it's been approved. I should note that I have attempted my own tool for listenbrainz here, though it is implemented in Rust instead of Python. I still have to reach out to the listenbrainz team to try and get the song matching working as well as it can. Right now I can get it to about 90%+ but still not 100%.
