Links from Myspace to Musicbrainz
Kurt J
links to dbtune/musicbrainz [2]
Of course these are only for mo:MusicArtists on myspace and not normal
myspace users.
The linking was done offline using a variant of the graph matching
approach from Yves and Chris Sutton et al. [3]
Let me know if you find any false positives. For now i'll just remove
them one-by-one. Hopefully there won't be very many ;-)
Feedback, suggestions, and criticism welcome.
-kurt j
[1] - http://virtuoso.dbtune.org/sparql
[2] - http://dbtune.org/musicbrainz
[3] - http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.123.9753&rep=rep1&type=pdf
Kurt J
> artists to their MySpace profile pages?
yes, a good point. we cross reference w/ these values (if they exist)
and if they don't we assume we're correct.
musicbrainz currently links to ~14k myspace pages. these links are
already in the dbtune.org/musicbrainz store. however, they should
probably exist in the myspace store as well. consider it on the todo
list :-)
-kurt
Yves Raimond
>
> Can you provide some examples of the data? I was looking for myspace
> links in the MusicBrainz data set earlier in the week but couldn't
> find any.
>
> This query:
>
> PREFIX mo: <http://purl.org/ontology/mo/>
> SELECT ?myspace WHERE {
> ?x mo:myspace ?myspace.
> }
> LIMIT 10
>
> Returns no results from the musicbrainz SPARQL endpoint linked from
> [1] and eventually times out with a Proxy error from [2]. I also tried
> some variants using foaf:isPrimaryTopicOf. I'm probably missing
> something here.
The links to myspace are using owl:sameAs. The mo:myspace property
points to a document (the myspace page), whereas we're asserting
sameAs links between the same artists in Musicbrainz and Myspace. Try
something like:
PREFIX mo: <http://purl.org/ontology/mo/>
SELECT ?myspace WHERE {
?x a mo:MusicArtist;
owl:sameAs ?myspace.
FILTER ( REGEX(str(?myspace), "^http://dbtune.org") )
}
LIMIT 10
Cheers!
y
Yves Raimond
>> The links to myspace are using owl:sameAs. The mo:myspace property
>> points to a document (the myspace page), whereas we're asserting
>> sameAs links between the same artists in Musicbrainz and Myspace. Try
>> something like:
>>
>> PREFIX mo: <http://purl.org/ontology/mo/>
>> SELECT ?myspace WHERE {
>> ?x a mo:MusicArtist;
>> owl:sameAs ?myspace.
>> FILTER ( REGEX(str(?myspace), "^http://dbtune.org") )
>> }
>> LIMIT 10
>
> Ah, OK I misunderstood what you were referring about linking to :)
>
> For my purposes I'm starting from some artist homepages and myspace
> page urls, and want to lookup the uris of the artist. Doesn't look
> like I can do that though?
I am guessing it should be pretty self-forward with a simple regexp
substituting http://myspace.com by http://dbtune.org/myspace.
Otherwise, we can just change the D2R mapping to add mo:myspace links
as well?
y
>
> Cheers,
>
> L.
>
> --
> Leigh Dodds
> Programme Manager, Talis Platform
> Talis
> leigh...@talis.com
> http://www.talis.com
>
Kurt J
> substituting http://myspace.com by http://dbtune.org/myspace.
this is what i'd suggest. should work right now :-)
> Otherwise, we can just change the D2R mapping to add mo:myspace links
> as well?
this would be nice actually i think, even though a bit redundant.
-kurt
Kurt J
>
> can reliably pick apart the uri in this case, but not something to
> generally encourage. After all these relationship between identifier
> and homepage might change in the future, c.f. Facebook profile url
> changes. So a little redundancy might make things a little easier(*)
> and a little more future-proof.
>
> Cheers,
>
> L.
>
> * I can't do that regex trick in pure SPARQL for example, which is how
> I'm doing the basic cross-linking at the moment.
you're right. redundant was the wrong word. also note that the
myspace endpoint contains mo:musicbrainz properties in addition to
owl:sameAs links for artists matched to musicbrainz. i agree
dbtune/musicbrainz should do the same :-)
Yves Raimond
that mapping recently, and there are certainly loads of other bugs
lurking around. That service is *definitely* not meant to be a
persistent place of musicbrainz linked data btw. It will hopefully
live at Musicbrainz itself in the future. Otherwise, for artists, BBC
Music is a stable and well-maintained resource :)
On 7/16/09, Hugh Glaser <h...@ecs.soton.ac.uk> wrote:
> Hi.
> Great stuff.
> I think some more links might help.
> At the moment, I am missing a link that gets me from
> http://sameas.org/html?uri=http://dbtune.org/myspace/uid/4689215
> to
> http://sameas.org/html?uri=http://www.bbc.co.uk/music/artists/65f4f0c5-ef9e-490c-aee3-909e7ae6b2ab
> which I think may be the problem Leigh has, and would be really exciting for
> me to have.
> By the way, I see there are things like
> <http://dbtune.org/musicbrainz/resource/artist/03e76712-dab1-4d7c-b347-1db970f93782>
> <http://www.w3.org/2002/07/owl#sameAs>
> <http://es.wikipedia.org/wiki/Ignacio_Mastretta_Rodr\u00EDguez> .
> I'm not sure it is good practice to do owl:sameAs to html pages, if we are
> living in a Linked Data world?
> Best
> Hugh.
Kurt J
> that mapping recently, and there are certainly loads of other bugs
> lurking around. That service is *definitely* not meant to be a
> persistent place of musicbrainz linked data btw. It will hopefully
> live at Musicbrainz itself in the future. Otherwise, for artists, BBC
> Music is a stable and well-maintained resource :)
this is a good point. the artist links in the myspace endpoint should
probably point to bbc music instead or as well.