Re: [igraph] Python igraph Documentation and Examples

[Tamas Nepusz]

Dear Lorenzo,

> (1) Generate a graph with a given distribution of degrees P(k)
This should be something like:

degrees = [your list of degrees for each vertex]
g = Graph.Degree_Sequence(degrees, method="vl")

Note that degdist contains the actual degrees, not the distribution,  
so you will have to generate the appropriate list of degrees (and hope  
that it's realizable). method="vl" generates simple graphs with the  
method of Viger and Latapy, but it takes more time. method="simple" is  
a much simpler method that does not guarantee that you won't get  
multiple edges.

> (2) See what happens if for instance I disconnect certain components  
> of a graph
You can remove edges and nodes by g.delete_edges() and  
g.delete_vertices() if g is a Graph object. Just remove them and  
recalculate your measures, whatever they are.

> (3) Calculate the minimum distance between some edges
Well, if those edges are A-B and C-D, calculate the minimum distance  
between A-C, A-D, B-C and B-D and take the minimum of them. Distance  
calculations can be done by Graph.shortest_paths(vs), where vs is a  
list of start vertices and it will return a list for each start vertex  
you gave, containing the shortest paths from that vertex to ALL other  
vertices.

> Is there any place online where I can find them? In a sense, I am
> looking for a Python igraph cookbook.
Once I came up with the idea of collecting igraph-related code  
snippets in snipplr.com, then I abandoned it thinking that there will  
be no real demand for that. So currently there's only one such snippet  
there:

http://snipplr.com/view/9914/calculating-assortativity-coefficient-in-igraph/

Anyway, seeing that the demand grows, I might be reconsidering the  
idea ;)

> Last question: it seems to be that the documentation for the R igraph
> package is more substantial (and also has more examples) than the one
> for the python binding. Is there a reason for this?
Yes, sheer laziness on my side :( More precisely it's a general lack  
of spare time rather than laziness. The good news is that I plan to  
release a proper tutorial for igraph's Python interface after the  
Christmas holiday (assuming that I will have enough time to finish it  
during the holiday).

--
Tamas