Benjamin Good wrote:
> Cassio,
>
> The short answer to your question (as I understood it) is that you could not
> issue such a query to the dbpedia sparql endpoint by itself. Somehow you
> would need to get access to an endpoint that contains both the freebase data
> as RDF and the mappings that Paul discusses here in order to run your query.
>
> Please correct me if I am wrong!
>
> -Ben
>
Ok, actually it is a bit more complex thant this. Both my and the
mapping files contain fbase URIs like
http://rdf.freebase.com/ns/guid.9202a8c04000641f8000000000ac9819
If you go to ~that~ URL you get redirected to
http://www.freebase.com/view/en/uab_hospital
Now, if you follow a few links you'll eventually find
http://rdf.freebase.com/rdf/en.uab_hospital
which contains the facts that you probably want about this subject
in NT format. So far as I know, however, fbase doesn't offer a dump
file with all of the NT assertions about fbase, so anybody who wants to
fill an RDF store/SPARQL query system with fbase assertions really has
two choices:
(i) run a crawler against fbase to harvest said assertions,
(ii) or derive these assertions from the "link export" file that can be
found here...
http://download.freebase.com/datadumps/
I think (ii) would be a straightforward project, particularly if
you used GUIDs on the LHS of your assertions; there's some talk about
this here:
http://www.freebase.com/view/m/04xw5db
Practically there are all sorts of funky details, such as expanding
namespaced keys into blank nodes,
<http://rdf.freebase.com/ns/en.uab_hospital>
<http://rdf.freebase.com/ns/type.object.key> _:AyiJNCGP1602
_:AyiJNCGP1602 <http://rdf.freebase.com/ns/type.key.namespace>
<http://rdf.freebase.com/ns/wikipedia.en_id> .
_:AyiJNCGP1602 <http://rdf.freebase.com/ns/type.value.value> "4267124" .
Having messed with the "link export" file quite a bit I'd say the
main issue is that it's an awfully big file and the scripts I run
against take a frustratingly long time... For a long time I've taken
the route of "efficient data structures & algorithms" but if I had do a
lot more of this I'd be looking at parallelization.
Although @fbase is officially trying to submerge GUIDs, both the
"simple topic dump" and "link export file" are highly dependent on GUIDs
and as of the last dump files, the "mid" identifies that @fbase wants
us to use are nowhere to be found.
------------------------------------------------------------------------------
ThinkGeek and WIRED's GeekDad team up for the Ultimate
GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the
lucky parental unit. See the prize list and enter to win:
http://p.sf.net/sfu/thinkgeek-promo
_______________________________________________
Dbpedia-discussion mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/dbpedia-discussion
