DQP Distributed Query Process

Introduction

The purpose of this extension if to allow to split a SPARQL request and execute its subparts on various SPARQL endpoints.

Comparison between a local request and a distributed request

Local Request

import fr.inria.edelweiss.kgram.core.Mappings;
import fr.inria.edelweiss.kgraph.core.Graph;
import fr.inria.edelweiss.kgraph.query.QueryProcess;
import fr.inria.edelweiss.kgtool.load.Load;

Collection<String> queries = ...;
Graph graph = Graph.create();
QueryProcess exec = QueryProcessDQP.create(graph);
Load ld = Load.create(graph);
ld.load( data_uri );

for (String q : queries) {
  Mappings results = exec.query(q);
}

Distributed Request

The main difference is on the exec object, which is of QueryProcessDQP type instead of QueryProcess. Then the user adds remote endpoints that should be used with the method addRemote.

import fr.inria.acacia.corese.exceptions.EngineException;
import fr.inria.edelweiss.kgdqp.core.Messages;
import fr.inria.edelweiss.kgdqp.core.ProviderImplCostMonitoring;
import fr.inria.edelweiss.kgdqp.core.QueryProcessDQP;
import fr.inria.edelweiss.kgdqp.core.Util;
import fr.inria.edelweiss.kgdqp.core.WSImplem;
import fr.inria.edelweiss.kgram.core.Mappings;
import fr.inria.edelweiss.kgraph.core.Graph;
import fr.inria.edelweiss.kgraph.query.QueryProcess;
import fr.inria.edelweiss.kgtool.load.Load;
import java.io.IOException;
import java.net.MalformedURLException;
import java.net.URL;
import java.util.ArrayList;
import java.util.logging.Level;
import org.apache.commons.lang.time.StopWatch;
import org.apache.log4j.Logger;
	
Graph graph = Graph.create(false);
QueryProcessDQP execDQP = QueryProcessDQP.create(graph, sProv, true);
execDQP.setGroupingEnabled(true); // @ToBeDocumented
execDQP.addRemote(new URL(...), WSImplem.REST);
for (String query : queries) {
  Mappings map = execDQP.query(query);
}

REST API of Corese Server

/dqp/reset