[igraph] Problems with constraint calculation in igraph python library

[Eytan Bakshy]

Hello,

I am using the latest version of igraph for python 2.5 (v0.5) under  
mac os x 10.5, and the Burt constraint calculation only seems to yield  
0's or 1's for each node's constraint, for both synthetic and  
empirical networks.

For example, in python, if I take some graph, say,
        In[66]: g = Graph.Barabasi(100,3)
and take the union (set) of the list returned by g.constraint(), the  
only values returned are 0 and 1:
        In[67]: set(g.constraint())
        Out[67]:        set([0, 1]))

But in R, if I do a similar sequence of commands,
        > g <- barabasi.game(100,1,m=3)
        > unique(constraint(g))

I get a good 70-80 unique values.


Is there something wrong with the way I am invoking constraint in  
python, or are there some known issues with the method?

Thanks,
Eytan