Sesame and Exhibit

17 views
Skip to first unread message

Batla

unread,
Mar 19, 2009, 7:57:36 PM3/19/09
to SIMILE Widgets
Hi all,

Please forgive me if this post is a little naive. I am fairly
confident in my knowledge of RDF, but I am new to Exhibit and Sesame.
Up to now I have built a very simple exhibit and a very simple
implementation of MemoryStore in Sesame.

Here's my question. Is there a "converter" than someone has written
that I could use to convert Statements from Sesame to Exhibit JSON?
Here's what I'm thinking. I like the sesame API's to the RDF Store -
querying, inserting, etc. and I love the faceted browse in Exhibit. I
would like to leverage both. Does that even make sense?

Any pointers for this newbie would be helpful and appreciated.

Thanks in advance!

-Batla

David Huynh

unread,
Mar 21, 2009, 9:35:59 PM3/21/09
to simile-...@googlegroups.com
Batla,

Babel
http://service.simile-widgets.org/babel/
can convert from RDF/XML to Exhibit/JSON. It only processes resources
that have rdf:type's.

You can even feed Exhibit an RDF/XML URL directly, and Exhibit will
route it through Babel, e.g.,
http://people.csail.mit.edu/dfhuynh/projects/data-theft/data-theft.html

David

Batla

unread,
Mar 25, 2009, 1:32:12 PM3/25/09
to SIMILE Widgets
HI David,

Thanks for your response.

I guess I'm thinking that I'm looking for something Exhibit may not
be...at least "out of the box". I ran into a problem with a large
volume of data with an exhibit I built (about 500k triples). I was
hoping that perhaps I could query something like a sesame server and
try to keep the json objects in manageable chunks. Pagination I'm
thinking.

I may be making some pretty big assumptions here. I'm sure someone on
this list has already done something like this w/o any need for a
physical triple store - I'll check the historical posts again. But in
the meanwhile, if you can shed any light on this or point me to a
relevant thread, I would appreciate it.

Thanks,
Batla

David Huynh

unread,
Mar 28, 2009, 8:55:00 PM3/28/09
to simile-...@googlegroups.com
Batla,

Yes, Exhibit isn't meant to scale that much... A few thousand items
(records) are pushing it. I started a hybrid client/server version of
Exhibit called Backstage, but haven't had much time to follow through.

David

harshal

unread,
Mar 29, 2009, 3:52:57 AM3/29/09
to SIMILE Widgets
Hi,

Is it a good idea to consider Google Gears or similar for locally
storing the items ?
I think some scalability can be achieved with this approach.

I would appreciate any comments on this.

Thanks for your time.

Regards

./harshal

Tomas

unread,
Apr 1, 2009, 10:32:53 AM4/1/09
to SIMILE Widgets
Hi,

I have exactly the same idea, and the same problem... I have rdf
stored in Sesame and I would like to use Exhibit to display this
knowledge base in a user-friendly way.

I think the performance problem would be solved using SPARQL and
converting dynamically the results to a json file, but I don't find
any tool that does it at runtime. Do you know about such a tool? I see
that Ivan Hermann (http://www.ivan-herman.net/Misc/PythonStuff/SPARQL/
Doc-SPARQL/SPARQL-module.html) had a similar approach using Python,
but we are using Java.

Any ideas?

Cheers

Tomás

David Huynh

unread,
Apr 8, 2009, 1:09:32 AM4/8/09
to simile-...@googlegroups.com
Tomas wrote:
> Hi,
>
> I have exactly the same idea, and the same problem... I have rdf
> stored in Sesame and I would like to use Exhibit to display this
> knowledge base in a user-friendly way.
>
> I think the performance problem would be solved using SPARQL and
> converting dynamically the results to a json file, but I don't find
> any tool that does it at runtime. Do you know about such a tool? I see
> that Ivan Hermann (http://www.ivan-herman.net/Misc/PythonStuff/SPARQL/
> Doc-SPARQL/SPARQL-module.html) had a similar approach using Python,
> but we are using Java.
>
> Any ideas?
>
First, you'd need an extension to SPARQL that actually supports
aggregate functions like COUNT. Actually, Sesame's SeRQL does have
COUNT, and MIN, MAX, GROUP BY. So, SeRQL is probably the better choice.

Backstage's code is here
http://simile.mit.edu/repository/backstage/trunk/
It might not work out of the box but you could get some hints for how to
architecture the faceted browsing engine.

David

Reply all
Reply to author
Forward
0 new messages