I have a graph with reified relations, which hold useful information, but for visualization purpose I need to create a subgraph without these intermediary nodes.
Example :
[A:Person] <--AFFILIATE-- [B:Affiliation] --COMPANY--> [C:Org]
And I want to produce a subgraph like this :
[A:Person] --AFFILIATED_TO--> [C:Org]
Is there any simple way to get that with Gremlin ?
I think that your best option might be to use subgraph() step as you normally might to extract the edge-induced subgraph and then execute some Gremlin on that subgraph to introduce the visualization edges and remove the stuff you don't want.
I can demonstrate with the modern toy graph packaged with TinkerPop:
gremlin> graph = TinkerFactory.createModern()
==>tinkergraph[vertices:6 edges:6]
gremlin> g = graph.traversal()
==>graphtraversalsource[tinkergraph[vertices:6 edges:6], standard]
gremlin> sg = g.V().outE('created').subgraph('sg').cap('sg').next() // subgraph creation
==>tinkergraph[vertices:5 edges:4]
gremlin> g = sg.traversal()
==>graphtraversalsource[tinkergraph[vertices:5 edges:4], standard]
gremlin> g.V().as('a'). // add special subgraph edge
......1> out('created').as('software').
......2> in('created').where(neq('a')).
......3> addE('co-developer').from('a').
......4> property('project',select('software').by('name'))
==>e[0][1-co-developer->4]
==>e[1][1-co-developer->6]
==>e[2][4-co-developer->1]
==>e[3][4-co-developer->6]
==>e[4][6-co-developer->1]
==>e[5][6-co-developer->4]
gremlin> g.V().hasLabel('software').drop() //remove junk from subgraph
gremlin> g.E()
==>e[0][1-co-developer->4]
==>e[1][1-co-developer->6]
==>e[2][4-co-developer->1]
==>e[3][4-co-developer->6]
==>e[4][6-co-developer->1]
==>e[5][6-co-developer->4]
gremlin> g.V().has('name','marko').outE('co-developer').valueMap(true)
==>[label:co-developer,project:lop,id:0]
==>[label:co-developer,project:lop,id:1]