On 25/07/13 15:13, Luis Daniel Ibáñez González wrote:
Hi,
I'm working on an application that uses as backend a modified version of
Corese, a Java-based GraphStore [0], the HTTP protocol implemented there is
somewhat old and I thought it would be a better idea to reuse something
newer than try to embed a Jetty from scratch.
Can Fuseki be used as a component to achieve that? I imagine something very
simple:
1) Fuseki server runs and accept HTTP GSP requests
Do you mean GSP = SPARQL Graph Store HTTP protocol? not the SPARQL
Protocol for sending SPARQL Queries and SPARQL Updates?
http://www.w3.org/TR/sparql11-http-rdf-update/
vs
http://www.w3.org/TR/sparql11-protocol/
2) A custom handler receives the query string, pass it to a Corese instance
(my thing is in-memory right now) that process it and returns a response
already in RDF/XML
3) This response is passed back to Fuseki, who serves it to the user.
I think this is very related to a recent question [1] about Fuseki as an
API. I've seen in the documentation that I should look at 'Dataset Accesor'
and 'DatasetAdapter', but I saw that they work with models, which leads me
to think they are very coupled to Jena.
They are the the client support for the SPARQL Graph Store HTTP
Protocol. GET/PUT of whole graphs. Useful for data management.
Thanks in advance.
If what you are looking for is an engine to do the SPARQL protocol
handling then Fuseki may be of use.
You can plug in your own storage+query engine and let Fuseki+Jetty deal
with the SPARQL protocols (both of them), SPARQL result set formats and
HTTP.
TDB does this; other than the fact it is in the server combined jar, it
is no different to any other plugin storage+query.
If you want to modify the results formats, or have a different query
language, then Fuseki may help but you will need to modify the code.
Andy
[0] http://wimmics.inria.fr/corese
[1] http://mail-archives.apache.org/mod_mbox/jena-dev/201307.mbox/browser
https://issues.apache.org/jira/browse/JENA-483
