Kgram: a Knowledge Graph Abstract Machine

Contact: Olivier Corby, INRIA

Participants: Catherine Faron-Zucker (I3S, CNRS), Corentin Follenfant (PolyTech'Nice)

KGRAM / KGRAPH

Download

KGRAM is an abstract machine that unifies graph match (graph homomorphism) and SPARQL query processing. It enables to compute SPARQL queries not only on RDF but also on other kinds of graph such as conceptual graphs. It can also perform mashup with data coming from XML using XPath and relational data using SQL. KGRAM is ported on Corese and an experiment is currently done to port it on another RDF platform. KGRAM can query several RDF stores for one SPARQL query.

KGRAM Goodies
KGRAM Proxies

KGRAM implements statements of SPARQL 1.1 Query Language: select expression, aggregate, property path, exists, minus and nested query.

Use Kgram interpreter with Corese

import fr.inria.edelweiss.kgramenv.util.QueryExec;

QueryExec exec = QueryExec.create(engine);

String query = "select * where {?x ?p ?y}";

IResults res = exec.SPARQLQuery(query);

Mashup with several Corese engines

import fr.inria.edelweiss.kgramenv.util.QueryExec;

QueryExec exec = QueryExec.create();

exec.add(engine1);
exec.add(engine2);
exec.add(engine3);

String query = "select * where {?x ?p ?y}";

IResults res = exec.SPARQLQuery(query);

Bibliography

Olivier Corby and Catherine Faron-Zucker, The KGRAM Abstract Machine for Knowledge Graph Querying, IEEE/WIC/ACM International Conference, September 2010, Toronto, Canada.
Olivier Corby and Catherine Faron-Zucker, La machine abstraite KGRAM et son langage GRAAL pour l'interrogation de graphes de connaissances, Proc. LMO, Langages et modèles à objets, Pau, March 2010, PDF
Olivier Corby and Catherine Faron-Zucker, KGRAM: une machine abstraite de graphes de connaissances, Proc. EGC, Extraction et Gestion des Connaissances, January 2010, Hammamet, Tunisia

KGRAM Logo : Colin Japiot