CODEWITHSUNDEEP
pythonnetworkx
Lucien S.
Lucien S.

Reputation: 5345

Networkx's "generalized_laplacian()" not working anymore?

This example (from Networkx's manual http://networkx.github.io/documentation/latest/examples/advanced/eigenvalues.html):

#!/usr/bin/env python
"""
Create an G{n,m} random graph and compute the eigenvalues.

Requires numpy or LinearAlgebra package from Numeric Python.

Uses optional pylab plotting to produce histogram of eigenvalues.

"""
__author__ = """Aric Hagberg ([email protected])"""
__credits__ = """"""
#    Copyright (C) 2004-2006 by
#    Aric Hagberg <[email protected]>
#    Dan Schult <[email protected]>
#    Pieter Swart <[email protected]>
#    All rights reserved.
#    BSD license.

from networkx import *
try:
    import numpy.linalg
    eigenvalues=numpy.linalg.eigvals
except ImportError:
    raise ImportError("numpy can not be imported.")

try:
    from pylab import *
except:
    pass

n=1000 # 1000 nodes
m=5000 # 5000 edges

G=gnm_random_graph(n,m)

L=generalized_laplacian(G)
e=eigenvalues(L)
print("Largest eigenvalue:", max(e))
print("Smallest eigenvalue:", min(e))
# plot with matplotlib if we have it
# shows "semicircle" distribution of eigenvalues
try:
    hist(e,bins=100) # histogram with 100 bins
    xlim(0,2)  # eigenvalues between 0 and 2
    show()
except:
    pass

Raises the following error with networkx's latest version:

Traceback (most recent call last):
  File "Untitled 2.py", line 36, in <module>
    L=generalized_laplacian(G)
NameError: name 'generalized_laplacian' is not defined

How should I go to make it work?

Upvotes: 4

Views: 908

Answers (2)

Aric
Aric

Reputation: 25299

That example is definitely broken with newer versions of NetworkX. Here is one that works:

import networkx as nx
import numpy.linalg
import matplotlib.pyplot as plt

n = 1000 # 1000 nodes
m = 5000 # 5000 edges
G = nx.gnm_random_graph(n,m)

L = nx.normalized_laplacian_matrix(G)
e = numpy.linalg.eigvals(L.A)
print("Largest eigenvalue:", max(e))
print("Smallest eigenvalue:", min(e))
plt.hist(e,bins=100) # histogram with 100 bins
plt.xlim(0,2)  # eigenvalues between 0 and 2
plt.show()

Upvotes: 5

DSM
DSM

Reputation: 353389

I think that name was removed in this commit:

Deprecate non-"matrix" names in laplacian.

Use laplacian_matrix, directed_laplacian_matrix, and normalized_laplacian_matrix only and deprecate other name aliases.

The lines it changed from networkx/linalg/laplacianmatrix.py include

-combinatorial_laplacian=laplacian_matrix
-generalized_laplacian=normalized_laplacian_matrix
-normalized_laplacian=normalized_laplacian_matrix
-laplacian=laplacian_matrix

So I think you can use normalized_laplacian_matrix instead:

>>> normalized_laplacian_matrix(G)
<1000x1000 sparse matrix of type '<type 'numpy.float64'>'
    with 11000 stored elements in Compressed Sparse Row format>

Upvotes: 1

Related Questions