def compute_greedy_newman(n_vertex, edge_list):
from agglomcluster import NewmanGreedy
graph = tf.compute_networkx_form(n_vertex, edge_list)
t = time.time()
clst = NewmanGreedy(graph)
clusters = clst.get_clusters()
exectime = time.time() - t
labels = tf.compute_labels_from_clusters(n_vertex, clusters)
return labels, clusters, exectime
def extract_biggest_component(filename):
from load_data import download_graph as dd
n_vertex, edge_list = dd(filename)
import networkx as nx
from transform_functions import compute_networkx_form
graph = compute_networkx_form(n_vertex, edge_list)
components = list(nx.connected_component_subgraphs(graph))
if len(components) > 1:
components.sort(key = len, reverse = True)
newgraph = components[0]
n_vertex_new = len(newgraph.nodes())
new_vertex_indices = {}
for i in xrange(n_vertex_new):
new_vertex_indices[newgraph.nodes()[i]] = i+1
f = open(filename[:-4]+'_new.txt', "w")
f.write(str(n_vertex_new)+' '+str(len(newgraph.edges()))+'\n')
for i in xrange(len(edge_list)):
if edge_list[i][0] in newgraph.nodes():
if len(edge_list[0]) == 2:
f.write(str(new_vertex_indices[edge_list[i][0]])+' '+str(new_vertex_indices[edge_list[i][1]])+'\n')
else:
f.write(str(new_vertex_indices[edge_list[i][0]])+' '+str(new_vertex_indices[edge_list[i][1]])+' '+str(edge_list[i][2])+'\n')
f.close()
import os
if os.path.isfile(filename[:-4]+'_labels.txt'):
f = open(filename[:-4]+'_labels.txt', "r")
f1 = open(filename[:-4]+'new_labels.txt', "w")
for i in xrange(n_vertex):
s = f.readline()
if i in newgraph.nodes():
f1.write(s)
f.close()
f1.close()
