Re: [igraph] Python Interface to igraph

[Tamas Nepusz]

> Shouldn't g.clusters() return the membership of each node to a
> cluster? Am I doing anything wrong? On my machine, g.clusters()
> returns what looks like a memory address.
g.clusters() returns an object with class VertexClustering (see the  
documentation or type help(Graph.clusters)). When you simply try to  
print this object, Python reports its class and its memory address.  
What you probably wanted based on your previous R experiences is the  
membership vector that's stored in an instance variable called  
membership. E.g.:

g=Graph.Erdos_Renyi(100, 0.1)
cl=g.clusters()
print cl.membership

The VertexClustering class has the following features:

len(cl) returns the number of clusters (if cl is a VertexClustering  
object)
cl[0] returns the members of the first cluster
cl[1] returns the members of the second cluster
...and so on
cl.membership returns the membership vector that you got used to when  
using the R interface
cl.q returns the modularity of the given partition

Cheers,
--
Tamas