A library for graph-based learning in Python.

Provides several types of graph container objects, with a unified API for visualization, analysis, transformation, and embedding.

from graphs.generators import random_graph

G = random_graph([2,3,1,3,2,1,2])

print G.num_vertices()  # 7
print G.num_edges()     # 14

G.symmetrize(method='max')
X = G.isomap(num_dims=2)

G.plot(X, title='isomap embedding')()

Requires recent versions of:

• numpy
• scipy
• scikit-learn
• matplotlib
• Cython

Optional dependencies:

• python-igraph
• graphtool
• networkx

Testing requires:

• nose
• nose-cov

Run the test suite:

./run_tests.sh