Configuration of JanusGraph

Here is a guide of configuring JanusGraph, an implementation of Tinkerpop Gremlin Server.

maxParameters: max number of bound variables

If you try to insert FoundNode with a lot of node or link attributes, the request can fail because it contains too many variable bindings. In that case, you have to increase the max number of bound variables.

To do that, edit the config file of Gremlin Server (e.g. gremlin-server.yaml) like the following

processors:
  - { className: org.apache.tinkerpop.gremlin.server.op.standard.StandardOpProcessor, config: { maxParameters: 128 }}

Pass big enough maxParameters field to StandardOpProcessor.

Index of node ID

net-spider often searches the graph database by node ID. In the graph database, the node ID is stored as a vertex attribute named @node_id by default.

We recommend making an index for the @node_id attribute. With JanusGraph, insert the following code to your Gremlin Server script (usually located under scripts directory).

graph.tx().rollback();
mg = graph.openManagement();
if(mg.getGraphIndex("byNodeId") == null) {
    println("--- get fnode");
    fnode = mg.getVertexLabel("found_node");
    if(fnode == null) {
        fnode = mg.makeVertexLabel("found_node");
    }
    println("--- get node_id");
    node_id = mg.getPropertyKey("@node_id");
    if(node_id == null) {
        node_id = mg.makePropertyKey("@node_id").dataType(String.class).make();
    }
    println("--- make index");
    mg.buildIndex("byNodeId", Vertex.class).addKey(node_id).buildCompositeIndex();
    println("--- commit");
    mg.commit();
}

Make sure to specify the script file in your gremlin-server.yaml (at scriptEngines.gremlin-groovy.scripts field).

If you want to use a different attribute name for the node ID, you can customize it with nodeIdKey config field in NetSpider.Spider.Config.