UI: Extending the frontend with MusicBrainz data

[defvs]

I'd love to have an even deeper integration with MusicBrainz, not limited to simply reading the tags; however it is understandable that this is a complex operation, because if done server-side this would require new network code to call upon MusicBrainz API.

BUT! Since all this data only needs to be fetched at view-time, on the frontend, the backend having no real usage for it, this could be the prime candidate for extra javascript/extension codes on the client-side.

Accessing the MBz API client-side could add the following information to pages, which are currently not available in tags:

• Track relationships, like remixer, remixed recording, edits of...
• Artist relationships, like siblings, renamed into, aliases.
• Release relationships, like singles linked to an album or compilation.
• Other interesting relationships and information taken from MBz, which aren't necessarily links to other pages but can still serve as extra information on an album or artist page

And since requests would be user-run, rate-limits would not be an issue, and loads on the API would be pretty minimal per IP.

The only 'useful' information server-side would be artist relationships most probably which could be used for merging artist pages even further, but the usefulness is still debatable. As such, a fully client-side extension would be enough and not necessitate any backend work.

One could start by developping a Userscript which can then be integrated into the backendw and distributed to clients. Trying to see if I'm going on the right track here, and whether this kind of development would benefit others if I were to do it.

[epoupon]

Some years ago, lms had acousticbrainz integration. The idea was to fetch all low level acoustic data for each track, on background, and save data in the database for further analysis for the similarity engine (which now only relies on tags)
Lms also has listenbrainz integration, and some helpers about rate limits etc. (sync are all done in background)
All that to say that I am rather for some persistent sync server side (could be done only on visited artists/releases in order not to hammer musicbrainz, and refreshed if the data is too old)

[epoupon]

And furthermore, as you already discovered, wt is a somewhat server centric library. Tweaking client side to add more information may be trickier than expected.