On Mon, May 7, 2012 at 4:42 AM, Peter Neubauer
<
peter.n...@neotechnology.com> wrote:
> very cool approach! I like the flexibility of that. Would be neat to
> have the type info stored in the graph itself, so the whole sytems
> becomes self-sufficient?
Thank you, Peter. Each node and relationship is marked with the id of
the schema type. We are using the same instance of Neo4j for handling
all the different graphs for every user, we hope to be able to sell
custom instances sometime in the future ;-) The inclusion of the
schema data in the graph sounds definitely good, but it adds overhead
when you query the graph or do some algorithms on it, so you have to
exclude the schema types nodes. For the time being, you can export the
schema in JSON and the data in Gephi separately.