Directed graph from adjacency matrix?

[Mars]

Hello,

I am new to python, numpy and networkx.

I would like to make a graph out of an asymmetric adjacency matrix. A
MultiDiGraph seems appropriate, but it looks like from_numpy_matrix
can only return a Graph.

Can this be done? How?

Also: I can't figure out how to list all nodes/edges and their
respective values/weights of a graph (just for verification). How is
this done?

Thanks!

[Aric Hagberg]

You can use the create_using= keyword to specify what kind of graph
you want like this:

In [1]: import numpy as np

In [2]: import networkx as nx

In [3]: A=np.matrix([[1,2],[3,0]])

In [4]: G=nx.from_numpy_matrix(A,create_using=nx.MultiDiGraph())

In [5]: G.edges(data=True)
Out[5]: [(0, 0, {'weight': 1}), (0, 1, {'weight': 2}), (1, 0, {'weight': 3})]

In [6]: A
Out[6]:
matrix([[1, 2],
[3, 0]])

Aric

[Mars]

Perfect - thanks a lot, Aric.

[soumya sarkar]

Thanx dats exactly wat i was looking for.........Plz add this functionality to the official networkx documentation

[Mridul Seth]

It is already present at https://networkx.github.io/documentation/latest/reference/generated/networkx.convert_matrix.from_numpy_matrix.html#networkx.convert_matrix.from_numpy_matrix