Exhibit 3 on SPARQL endpoint

[Martin M.]

Hi everyone,

I used the former version of Exhibit to build a prototype on a static
data source last year and it proved to be a very promising tool. I'm
currently looking for the same kind of vizualization/tayloring of data
on a larger source (triple store) via a SPARQL endpoint. My question
is: can Exhibit3 be used on a SPARQL endpoint ? And if not, is it
planned for a future version ?

Thanks

Martin

[mackenzie]

Hi Martin,

I'm not the expert on this, but I'm pretty sure the answer to your question is no, you cannot point Exhibit 3 at SPARQL endpoints outside the 'staged' environment. The internal client/server protocol is 100% restful http calls (doc in progress) so you *can* build alternative servers to ours but they have to support that API. The Exhibit 3 back end uses Sesame so you can have a variety of dbs and triplestores behind a Sail. Maybe you could slot you triplestore into Backstage (she said optimistically)?

MacKenzie

[ma...@coactus.com]

Right, I know we've talked about the possibility of adding other
stores, but there's no requirement to do so for E3 at this time.

FYI, the disk-based database code was removed a few months ago, but is
being added back in - temporarily - on a separate branch.

Mark.

[Martin M.]

Ok, thank you. I will take a look at the staged version and see if I
find a solution :)

[Ryan Lee]

Hi Martin,

Backstage is built on Sesame, which handles a good deal of SPARQL
natively. Here's one of the developers showing it off:

http://jeenbroekstra.blogspot.com/2010/03/accessing-dbpedias-sparql-endpoint-with.html

We haven't tapped into this specific feature, but perhaps you've come
across it in your exploration. I'm not entirely clear on whether you
can build a TupleQuery out of the Sesame Java API and have it issue a
SPARQL query to the endpoint, or whether you would need to modify the
Backstage code to directly build SPARQL queries. You could perhaps ask
on the Sesame users list.

[Martin M.]

The triple store I'm using is actually Sesame+OWLIM and I've been
using the Sesame Java API to work with it. If Backstage is built on
Sesame as well, it shouldn't be too complicated to replace its (in
memory, I guess) repository by an external one ?

" I'm not entirely clear on whether you can build a TupleQuery out of
the Sesame Java API and have it issue a SPARQL query to the endpoint"

I'm not sure I understood this correctly. What type of query does
Backstage use, if not SPARQL ?

On Jan 3, 9:54 pm, Ryan Lee <ryan...@zepheira.com> wrote:
> Hi Martin,
>
> Backstage is built on Sesame, which handles a good deal of SPARQL
> natively.  Here's one of the developers showing it off:
>

> http://jeenbroekstra.blogspot.com/2010/03/accessing-dbpedias-sparql-e...

[Martin M.]

I finally checked the code of Backstage quickly. The class
HostedDatabase (extending Database) seems to be responsible for the
creation of the repository. Would that be imaginable to simply create
another class (e.g. ExternalDatabase) extending Database that would,
instead of creating a native store, connect to an external
repository ? Something like :

_repository = new HTTPRepository(url, storeID);
_repository.initialize();
_sail = ((SailRepository) _repository).getSail();

That seems like the easiest solution to me but I need the opinion of
someone having actually work on this code :)

[Niels Mayer]

On Tue, Jan 3, 2012 at 12:54 PM, Ryan Lee <rya...@zepheira.com> wrote:

Backstage is built on Sesame, which handles a good deal of SPARQL
natively.  Here's one of the developers showing it off:

http://jeenbroekstra.blogspot.com/2010/03/accessing-dbpedias-sparql-endpoint-with.html

Thanks for the interesting link!

I'm also interested in seeing SPARQL better supported for some of my own Exhibit and Sparql-using projects, e.g. 

http://qtzibit.googlecode.com

http://voicetogoog.googlecode.com . 

Apps like these can be deployed on an open, Linux based mobile platform ( http://swipe.nokia.com ) that fundamentally integrates semantic desktop capabilities as part of the platform API: http://harmattan-dev.nokia.com/docs/library/html/libqtsparql/main.html

Although such lower-level functionality is efficiently implemented in Qt ( http://qt-project.org ) C++ it can be made available in JavaScript/JSON and exported to Exhibit via QML (Qt's integrated declarative Javascript): http://harmattan-dev.nokia.com/docs/library/html/libqtsparql/qmlSparqlConnection.html

IMHO, SPARQL queries are just as interesting when they're your own data taken from your own smartphone. For example, a list of all photos/video tagged with location:

n950 ~ # tracker-sparql -q 'SELECT nie:url(?u) ?cty ?cry { ?u slo:location [ slo:postalAddress [ nco:locality ?cty ; nco:country ?cry ]] .}'

Results:

  file:///home/user/MyDocs/Pictures/Objects005.jpg, Jakarta, Indonesia

  file:///home/user/MyDocs/Pictures/Objects008.jpg, Espoo, Finland

[...]

  file:///home/user/MyDocs/Pictures/Places007.jpg, Greater Mumbai, India

  file:///home/user/MyDocs/DCIM/12010001.jpg, Corona del Mar, USA

  file:///home/user/MyDocs/DCIM/12010002.jpg, Corona del Mar, USA

  file:///home/user/MyDocs/DCIM/12010003.jpg, Corona del Mar, USA

All contacts, calendar, email &  social network checkins, photo face-tags, etc can be queried, e.g.:

n950 ~ $ tracker-sparql -q 'select nco:nameFamily(?u) nco:nameGiven(?u)

?num { ?u nco:hasAffiliation [ nco:hasPhoneNumber ?num ] . }'

Results:

[...]

 Mayer Niels, urn:x-maemo-phone:mobile:+1xxxxxxxxxxxx

[...]

-- Niels.

http://nielsmayer.com