3

I'm loading RDF data into a JUNG graph to do some analysis. So I create a new graph with:

DirectedGraph g = new DirectedSparseGraph<String,GraphLink>();

I created a support class for specifying the link:

public class GraphLink {
  String uri;
  Float weight;
}

Then I populate it like this:

for each rdf triple <s,p,o>{
  g.addVertex( s )
  g.addVertex( o )

  GraphLink link = new GraphLink()
  link.uri = pred
  link.weight = some weight;
  g.addEdge( link, s, o )
}

Is this an efficient way of doing it or there are better ways? The representation of the edges is very counterintuitive, but if I do:

g.addEdge( p, s, o )

I get an exception of duplicated edge.

Any hints?

UPDATE: this code seems to work well:

DirectedGraph<RDFNode,Statement> g = new DirectedSparseGraph<RDFNode,Statement>()
// list all statements
// TODO: pagination for very large graphs.
assert m.size() < 10000000,"graph is too large."
m.listStatements().each{ stm->
    RDFNode sub = stm.getSubject()
    RDFNode obj = stm.getObject()
    g.addVertex( sub )
    if ( includeLiterals || !obj.isLiteral() ){
        g.addVertex( obj )
        g.addEdge( stm, sub, obj, EdgeType.DIRECTED )
    }
}

Mulone

Mulone
  • 3,603
  • 9
  • 47
  • 69

1 Answers1

4

This may not be what you want at all, but you could try JenaJung, which presents a jena model as a Jung graph.

From the README file:

Model model = FileManager.get().loadModel("http://example.com/data.rdf");
Graph<RDFNode, Statement> g = new JenaJungGraph(model);

Layout<RDFNode, Statement> layout = new FRLayout(g);
layout.setSize(new Dimension(300, 300));
BasicVisualizationServer<RDFNode, Statement> viz =
    new BasicVisualizationServer<RDFNode, Statement>(layout);
user205512
  • 8,798
  • 29
  • 28