Support for other sources of music metadata

[GoogleCodeExporter]

Rateyourmusic is an alternative source for information about albums which is 
also an "official" MusicBrainz source where MB pulls data from.

Rateyourmusic has the advantage of having a community that actually moderates 
submissions and changes made which means it is usually less of a mess than 
MusicBrainz and they actually do support multi-disc releases, unlike 
MusicBrainz.

Until now I've found it to be a source of equal if not greater value than MB 
though its database is smaller.

It would be nice if we could use Rateyourmusic as a primary database and 
fallback to MusicBrainz when it fails or something like that.

Using rateyourmusic would also significantly ease the implementation of #42 
when we don't need to fallback to MB.

Currently rym does not have an API yet though there is a ticket about it.

Though Rym doesn't have an API/python library right now it's URLs are easily 
reconstructed, for example Echo Mountain from K's Choice can be found at: 
http://rateyourmusic.com/release/album/ks_choice/echo_mountain/ and a full list 
of K's Choice discography according to Rateyourmusic can be found at: 
http://rateyourmusic.com/artist/ks_choice

Original issue reported on code.google.com by daniele.sluijters on 3 Mar 2011 at 12:53

[GoogleCodeExporter]

Interesting; I had never heard of this site.

As you note, they don't have an API yet, but they do have a ticket about it, 
although it seems to be inactive:
http://rateyourmusic.com/rymzilla/view?id=683

So if beets were to use this as a data source, we would have to write a fast 
and robust scraper, presumably introducing a dependency on lxml (by far the 
best library for parsing HTML efficiently).

And although it has drawbacks, MusicBrainz does have a few advantages too. One 
advantage of the MusicBrainz database is that it's entirely open, meaning that 
one can go to their web site and download a full dump of all their data at any 
time -- we're not dependent on their infrastructure. Also, MB has a very strict 
style guide that they take pretty seriously -- does RYM have something similar?

Anyay, I'm not opposed to adding RYM as an additional data source, but there is 
a significant implementation challenge in implementing a good scraper.

Original comment by adrian.sampson on 3 Mar 2011 at 4:56

• Added labels: Type-Enhancement

[GoogleCodeExporter]

Rateyourmusic is a community. People are free to add artists and releases. 
Other people can report those if something is wrong with them and submit 
changes into the moderation queue which need then be approved by a moderator.

Original comment by daniele.sluijters on 3 Mar 2011 at 6:11

[GoogleCodeExporter]

Yeah, the community looks great. Thanks for bringing it to our attention.

If someone pulls together a good screen-scraper implementation for RYM, let's 
explore how it can be used as a data source.

Original comment by adrian.sampson on 3 Mar 2011 at 6:20

[GoogleCodeExporter]

Original comment by adrian.sampson on 27 Apr 2011 at 11:53

• Changed title: Support for other sources of music metadata
• Changed state: Accepted
• Added labels: Priority-Low, Milestone-1.1, Component-Tagger

[GoogleCodeExporter]

Issue 179 has been merged into this issue.

Original comment by adrian.sampson on 27 Apr 2011 at 11:55

[GoogleCodeExporter]

Last.fm would be a good source, they have releases that musicbrainz does not 
have. It might be best to have support for choosing which source the tags are 
created from.

Original comment by coolke...@gmail.com on 28 Apr 2011 at 1:02

[GoogleCodeExporter]

Discogs has the beginnings of a Python binding for their API: 
https://github.com/discogs/discogs-python-client

Original comment by adrian.sampson on 6 Sep 2011 at 6:59

[GoogleCodeExporter]

Last.fm could be supported through python-lastfm: 
https://code.google.com/p/python-lastfm/

It looks pretty nice and as a bonus Last.fm keeps track of MBid's so even that 
can be requested. The only thing I find semantically incorrect is that single's 
are treated as albums. For all intents and purposes they are but it feels 
wrong, those things should be called releases.

Original comment by daniele.sluijters on 7 Sep 2011 at 8:24

[GoogleCodeExporter]

Original comment by adrian.sampson on 19 Feb 2012 at 10:46

• Removed labels: Milestone-1.1

[GoogleCodeExporter]

I have inherited a music library that was previously very well tagged with 
discogs. Probably more than half of the releases are not in Music Brainz 
database. I'd like to see discogs (and other sources) as fallback to Music 
Brainz, and potentially flag them in the database so I can find those releases 
and manually review and submit to Music Brainz.

Original comment by real.hu...@mrmachine.net on 26 Jan 2013 at 11:48

[GoogleCodeExporter]

as numerous others have said, would be great to have discogs support, also 
last.fm and what.cd would be intersting sources especially for tagging genres

apparently someone already started a project with that in mind here
https://github.com/YetAnotherNerd/whatlastgenre

Original comment by bm.max...@gmail.com on 5 Feb 2013 at 1:39

[GoogleCodeExporter]

discogs python client linked above has moved to 
https://github.com/discogs/discogs_client - doesn't look like it has seen a lot 
of activity... Last commit a year ago... Not sure how useful or complete it 
is...

What about just displaying a link that will open a search on discogs or last.fm 
when there is no good Music Brainz ? And once you have a match on there, isn't 
there an easy way to import (and manually correct) into Music Brainz then 
re-scan?

Beets could skip these, but not add them to incremental history so that they 
will be picked up again on subsequent scans.

Original comment by real.hu...@mrmachine.net on 5 Feb 2013 at 2:06