Things are moving fast, since my last post. Indeed, Frederick just put online the Musicbrainz RDF dump, with dereferencable URIs, SPARQL end-point, everything. Great job Fred!!

This data set will surely be a sort of hub for music-related data on the Semantic Web, as it gives URIs for a large number of artists, tracks, albums, but also timelines, performances, recordings, etc. Well, almost everything defined in the Music Ontology.

I am happy to announce the first hack using this dataset:-) This is called GNAT (for GNAT is not a tagger). It is just some lines of python code which, from an audio file in your music collection, gives you the corresponding dereferencable URI.

It also puts this URI into the ID3v2 Universal File Identifier (UFID) frame. I am not sure it is the right place to put such an information though, as it is an identifier of the manifestation, not the item iself. Maybe I should use the user-defined link frames in the ID3v2 header...

So it is actually the first step of the application mentioned here!

It is quite easy to use:

$ python trackuri.py 7-don\'t_look_back.mp3

 - ID3 tagsthe evolution of this!

Artist:  Artemis
Title:  Don't Look Back
Album:  Undone


 - Zitgist URI

http://zitgist.com/music/track/2b78923b-c260-44c1-b333-2caa020df172

Then:

$ eyeD3 7-don\'t_look_back.mp3

7-don't_look_back.mp3   [ 3.23 MB ]
--------------------------------------------------------------------------------
Time: 3:31      MPEG1, Layer III        [ 128 kb/s @ 44100 Hz - Stereo ]
--------------------------------------------------------------------------------
ID3 v2.4:
title: Don't Look Back          artist: Artemis
album: Undone           year: 2000
track: 7                genre: Trip-Hop (id 27)
Unique File ID: [http://zitgist.com/music/] http://zitgist.com/music/track/2b78923b-c260-44c1-b333-2caa020df172
Comment: [Description: http] [Lang: ]
//www.magnatune.com/artists/artemis
Comment: [Description: ID3v1 Comment] [Lang: XXX]
From www.magnatune.com

You can also output the corresponding RDF, in RDF/XML or N3:

$ python trackuri.py 1-i\'m_alive.mp3 xml
<?xml version="1.0" encoding="UTF-8"?>
<rdf:RDF
   xmlns:_3="http://purl.org/ontology/mo/"
   xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
>
  <rdf:Description 
rdf:about=
    "http://zitgist.com/music/track/67a1fab6-aea4-47f4-891d-6d42bb856a40">
    <_3:availableAs rdf:resource=""/>
  </rdf:Description>
</rdf:RDF>

$ python trackuri.py 1-i\'m_alive.mp3 n3

@prefix _3: <http://zitgist.com/music/track/67>.
@prefix _4: <http://purl.org/ontology/mo/>.

 _3:a1fab6-aea4-47f4-891d-6d42bb856a40 _4:availableAs <>.

... even though I still have to put the good Item URI, instead of <>.

Get it!

You can download the code here, and it is GPL licensed.

The dependencies are:

  • python-id3
  • python-musicbrainz2
  • RDFLib (easy_install -U rdflib)
  • mutagen (easy_install -U mutagen)