def test_ra_clustering_zero():
G = nx.Graph()
assert_equal(bipartite.robins_alexander_clustering(G), 0)
G.add_nodes_from(range(4))
assert_equal(bipartite.robins_alexander_clustering(G), 0)
G.add_edges_from([(0, 1), (2, 3), (3, 4)])
assert_equal(bipartite.robins_alexander_clustering(G), 0)
G.add_edge(1, 2)
assert_equal(bipartite.robins_alexander_clustering(G), 0)
def test_ra_clustering_zero():
G = nx.Graph()
assert_equal(bipartite.robins_alexander_clustering(G), 0)
G.add_nodes_from(range(4))
assert_equal(bipartite.robins_alexander_clustering(G), 0)
G.add_edges_from([(0, 1), (2, 3), (3, 4)])
assert_equal(bipartite.robins_alexander_clustering(G), 0)
G.add_edge(1, 2)
assert_equal(bipartite.robins_alexander_clustering(G), 0)
def compute_bipartite_small_world_index(G):
Gc = max(nx.connected_component_subgraphs(G), key=len)
actual_cc = bp.robins_alexander_clustering(G)
actual_apl = nx.average_shortest_path_length(Gc)
random_cc, random_apl = compute_cc_apl_random_networks(G)
result = {}
result['n'] = G.order()
result['m'] = G.size()
if 'python' in G.graph['name'].lower():
devs = len({n for n, d in G.nodes(data=True) if d['bipartite'] == 1})
result['developers'] = devs
result['files'] = G.order() - devs
else:
devs = len({n for n, d in G.nodes(data=True) if d['bipartite'] == 0})
result['developers'] = devs
result['packages'] = G.order() - devs
result['actual_cc'] = actual_cc
result['actual_apl'] = actual_apl
result['random_cc'] = random_cc
result['random_apl'] = random_apl
try:
result['swi'] = (actual_cc / random_cc) / (actual_apl / random_apl)
except ZeroDivisionError:
result['swi'] = float('nan')
return result
def compute_cc_apl_random_networks(G):
ccs = []
apls = []
for Gr in get_random_networks_2mode(G, n=10, r=10):
ccs.append(bp.robins_alexander_clustering(Gr))
if not nx.is_connected(Gr):
Gr = max(nx.connected_component_subgraphs(Gr), key=len)
apls.append(nx.average_shortest_path_length(Gr))
return mean(ccs), mean(apls)
import networkx
from networkx.algorithms import bipartite
structures_dir = '/home/ucbtepi/code/network/data'
tissue_c = []
random_c = []
gd_range = range(1,21)
for gd in gd_range:
print ' '
print gd
tissue_model = networkx.read_graphml(structures_dir+'/gd{0}/cr0/gd{0}_cr0.graphml'.format(gd))
random_bipartite_model = networkx.read_graphml(structures_dir+'/gd{0}/cr1/gd{0}_cr1.graphml'.format(gd))
for k,g in enumerate([tissue_model, random_bipartite_model]):
mf_nodes = list(set(n for n,d in g.nodes(data=True) if d['bipartite']==0))
grc_nodes = list(set(n for n,d in g.nodes(data=True) if d['bipartite']==1))
print(['tissue', 'random'][k])
c = bipartite.robins_alexander_clustering(g)
print("robins-alexander clustering: {}".format(c))
if k==0:
tissue_c.append(c)
else:
random_c.append(c)
fig, ax = plt.subplots()
ax.plot(gd_range, tissue_c, linewidth=1.5, color='k', label='tissue model')
ax.plot(gd_range, random_c, linewidth=1.5, color='r', label='random bipartite graph')
ax.set_xlabel('dendrites')
ax.set_ylabel('Robins-Alexander clustering')
ax.legend(loc='best')
plt.show()
def robins_alexander_clustering(self):
self.robins_alexander_clustering_dict = bi.robins_alexander_clustering(self.G)
def test_ra_clustering_square():
G = nx.path_graph(4)
G.add_edge(0, 3)
assert_equal(bipartite.robins_alexander_clustering(G), 1.0)
def test_ra_clustering_davis():
G = nx.davis_southern_women_graph()
cc4 = round(bipartite.robins_alexander_clustering(G), 3)
assert_equal(cc4, 0.468)
def test_ra_clustering_square():
G = nx.path_graph(4)
G.add_edge(0, 3)
assert_equal(bipartite.robins_alexander_clustering(G), 1.0)
def test_ra_clustering_davis():
G = nx.davis_southern_women_graph()
cc4 = round(bipartite.robins_alexander_clustering(G), 3)
assert_equal(cc4, 0.468)
import networkx as nx
import matplotlib.pyplot as plt
from networkx.algorithms import bipartite
from networkx.algorithms import centrality
B = nx.Graph()
B.add_nodes_from([1, 2, 3, 4], bipartite=0)
B.add_nodes_from(['a', 'b', 'c', 'd'], bipartite=1)
edges_list = [(1, 'a'), (1, 'b'), (2, 'a'), (2, 'b'), (2, 'c'), (3, 'c'),
(3, 'd'), (4, 'c'), (4, 'd')]
B.add_edges_from(edges_list)
pos = nx.spring_layout(B)
sets = bipartite.sets(B)
colors = ['r', 'b']
c = bipartite.robins_alexander_clustering(B)
print centrality.degree_centrality(B)
for i in range(len(sets)):
nx.draw_networkx_nodes(B,
pos,
nodelist=list(sets[i]),
node_color=colors[i])
nx.draw_networkx_edges(B, pos)
plt.show()
