Wimmics / corese

Software platform implementing and extending the standards of the Semantic Web.
https://project.inria.fr/corese/
Other
101 stars 29 forks source link

How to use the extended SPARQL constructs #4

Closed szarnyasg closed 8 years ago

szarnyasg commented 8 years ago

Hi,

I managed to get Corese working with basic SPARQL queries. I have a couple of questions about the extended SPARQL language:

  1. I'd like to use the param[...] construct described in the paper Analysis of a Real Online Social Network Using Semantic Web Frameworks. Does this work in the current version of Corese? Or do I have to "simulate" this by generating the queries for all parameters?
  2. I've found group by any queries in the tests (1, 2), but I couldn't get it working:

    PREFIX base: <http://www.semanticweb.org/ontologies/2015/trainbenchmark#> 
    select ?x ?y 
    where {
     ?x base:follows ?y
    }
    group by any

    This throws an exception:

    ERROR 2015-12-05 19:44:47,855 [562367320@qtp-354432491-19] (SPARQLRestAPI.java:398) - Error while querying the remote KGRAM engine

    (Without the group by any line, it's fine).

  3. Is there a more up-to-date version of this documentation somewhere?

Thanks, Gabor

ocorby commented 8 years ago

Hi,

Unfortunately these features are not available anymore.

There is a new way to perform these types of computing. It consists in associating a plugin to the SPARQL interpreter in such a way that : each time a solution of the query is found or a property path is identified, a plugin callback function is called in order to perform specific processing, e.g. network analysis operation.

The plugin must implement the ResultListener interface (fr.inria.edelweiss.kgram.event ; kgram project)

The main functions to be implemented are :

public boolean process(Path path)

public boolean process(Environment env)

Best regards

Olivier Corby

szarnyasg commented 8 years ago

Thanks, I'll look into the Java API.