Large Graph Database

[abki]

Héllo,

I'm introspecting the possibility to use graphine as a component for a
graph database.

My first attempt was to use the shelve module from python 3.1. I
modified the graph class this way:

def __init__(self, graph_name)
self._nodes = shelve.open(graph_name+'_nodes')
self._edges = shelve.open(graph_name+'_edges')

but the code fails because of général stackoverflow from python. I
tried the same thing with my own graph model, and it definitly works
even if my classes are far less complete.

I definitly think that this is not the right way to do the job.

The way I see graphine persistence model is that it loads all the data
into memory to do its job, which maybe an issue under certains
circontenses.

How do you use Graphine ? My project is to reverse engineer popular
tools like last.fm proposition algorithm, or ensembli.com. Any tips ?

Regards,

Amirouche.

---
If you are looking for large graph files have a look at
http://code.google.com/p/python-tulip-parser/
Acteurs.tlp file is (may be) owned by David AUBER and you shouldn't
consider them as open-source as it falls under the french law...