CODEWITHSUNDEEP
pythonigraph
Sait
Sait

Reputation: 19805

Populating a graph from a dictionary using igraph

I have a dictionary which has the follower-followee information. For example:

a --> b,c,d
b --> c,d
c --> d

Here it means, a follows b,c,d. b follows c,d. c follows d. I have the following corresponding dictionary:

{'a': ['b', 'c', 'd'], 'b': ['c', 'd'], 'c': ['d']}

I want to create a directed graph out of this dictionary in igraph.

What I am currently doing is:

import igraph

d = {'a': ['b', 'c', 'd'], 'b': ['c', 'd'], 'c': ['d']}

edges = []
vertices_set = set()
for key in d:
    vertices_set.add( key )
    for item in d[key]:
        edges.append( (key, item) )
        vertices_set.add( item )

# vertices_set: set(['a', 'c', 'b', 'd'])
# edges: [('a', 'b'), ('a', 'c'), ('a', 'd'), ('c', 'd'), ('b', 'c'), ('b', 'd')]

g = igraph.Graph( directed=True )
g.add_vertices( len(vertices_set) )
g.vs['name'] = list(vertices_set)
g.add_edges( edges )

print(len(g.vs))
print(len(g.es))

And this code works as expected. However, it is rather ugly (and most likely slower than it should be) as you can see.

I feel like, there should a faster and/or more pythonic way of doing this. Any comments?

Upvotes: 1

Views: 686

Answers (1)

Tamás
Tamás

Reputation: 48051

You can make use of the Graph.TupleList function:

>>> from igraph import Graph
>>> d = {'a': ['b', 'c', 'd'], 'b': ['c', 'd'], 'c': ['d']}
>>> g = Graph.TupleList([(k, v) for k, vs in d.iteritems() for v in vs])
>>> print g
IGRAPH UN-- 4 6 --
+ attr: name (v)
+ edges (vertex names):
a -- b, c, d   b -- a, c, d   c -- a, b, d   d -- a, b, c

Upvotes: 3

Related Questions