How convert an embedded Neo4j graph db to GraphML with Python?

Question

I want to store my embedded Neo4j Graph db to GraphML to draw it with a tool like yEd.

I'm managing the graph db with python27.

Does any body knows a way to do that?

score 1 · Answer 1 · answered Nov 23 '11 at 18:01

If you can afford to do it outside of the process -- which means you'll need to shut down your existing process working on the embedded database and restart when finished with the export-- you can use Gremlin to do it. Here's the commands you'll need:

g = new Neo4jGraph("/YOUR/GRAPH/DIRECTORY")
writer = new GraphMLWriter(g)
out = new FileOutputStream("/YOUR/GRAPHML/file.graphml")
writer.outputGraph(out)
writer.setNormalize(true)
out.close()

This will create a nice pretty graphml file that is suitable for reading into a tool like Gephi or Cytoscape.

If you need to export the graph in-process you'll need to use something like jython to run your python scripts and then use the above commands by importing the objects from com.tinkerpop.blueprints.

that seems really nice, but unfortunately when I try to do this, neo4j is initialized, but fails to start :S Does gremlin work well with the python-embedded 1.6 neo4j edition? — chiffa, Apr 03 '13 at 19:57

score 0 · Answer 2 · answered Nov 22 '11 at 14:51

If neo4j.py does not support exporting to GraphML format, or for any reason you cannot do that with Neo4j + Java, you'll have to write a custom exporter.

With some simple hacks you can accomplish that with NetworkX, which supports importing from/exporting to GraphML. Just figure a way to import your Neo4j graph into NetworkX and then export it directly to GraphML.

How convert an embedded Neo4j graph db to GraphML with Python?

2 Answers2