def test_modularity_dir_bug71():
"""Regression test for bug described in issue #71"""
fpath = os.path.join(TEST_DIR, "failing_cases",
"modularity_dir_example.csv")
x = np.loadtxt(fpath, int, delimiter=',')
bct.modularity_dir(x)
def community(file):
"""
*******
The function uses Girvan-Newman algorithm to get communities from a given network.
Packages bct (Brain Connectivity Toolbox) and networkx are required. Slow (>48 hrs
for one book) but accurate than community_fast() function.
*******
input: file
.txt file that contains the network.
format: u,v,weight
*******
output:
communities:
N*1 np.ndarray
Q-metric:
the measure of modularity for networks
"""
net = generate_netx(file)
n_edges = nx.number_of_edges(net)
Q_best = 0
while n_edges > 1:
#Calculate the betweenness for each edge. The output is a dictionary.
edge_betweenness = nx.edge_betweenness_centrality(net,weight='weight')
#get the edge(s) with max edge betweenness and remove it(them)
max_values = max(edge_betweenness.values())
for key,value in edge_betweenness.items():
if float(value) == max_values:
net.remove_edge(key[0],key[1])
n_edges = nx.number_of_edges(net)
#transform the network into a numpy array
AM = nx.to_numpy_matrix(net)
#utilize bct.modularity_dir() function to get community structure and calculate the Q-metric
#for current network
[ci,Q] = bct.modularity_dir(AM)
#update the best Q value if needed
if Q>Q_best:
Q_best = Q
ci_best = ci
else:
pass
words = print_com('unique.txt')
ci_best = [[words[i] for i in ci] for ci in ci_best]
return ci_best,Q_best
# -*- coding: utf-8 -*-
from dyconnmap.graphs import mutual_information
import numpy as np
import scipy
from scipy import io
import bct
if __name__ == '__main__':
rng = np.random.RandomState(0)
a = rng.rand(30, 30)
np.fill_diagonal(a, 0.0)
b = rng.rand(30, 30)
np.fill_diagonal(b, 0.0)
# In case you get NaNs you have to modify the `gamma` paramter
Ca, _ = bct.modularity_dir(a, gamma=1.0)
Cb, _ = bct.modularity_dir(b, gamma=1.0)
mi, nmi = mutual_information(Ca, Cb)
print(mi)
print(nmi)
def test_modularity_dir():
x = load_directed_sample()
_, q = bct.modularity_dir(x)
assert np.allclose(q, .31446049)
def test_modularity_dir_low_modularity():
x = load_directed_low_modularity_sample(thres=.67)
_, q = bct.modularity_dir(x)
assert np.allclose(q, .06450290)
def test_modularity_dir():
x = load_directed_sample()
_, q = bct.modularity_dir(x)
print(q, .32742787)
assert np.allclose(q, .32742787)
def test_modularity_dir():
x = load_directed_sample()
_, q = bct.modularity_dir(x)
print(q, .32742787)
assert np.allclose(q, .32742787)
def test_modularity_dir_low_modularity():
x = load_directed_low_modularity_sample(thres=.67)
_, q = bct.modularity_dir(x)
assert np.allclose(q, .06450290)
def test_modularity_dir(): x = load_directed_sample() _,q = bct.modularity_dir(x) assert np.allclose(q, .31446049)
