def get_max_weighted_clusters(graph, node_name):
maxval = 0
maxclustids = []
cdict = {}
for n, nbrs in graph.graf.adjacency_iter():
if n == node_name:
for nbr, eattr in nbrs.iteritems():
weight = eattr['weight']
#print graph.graf.node[nbr]
cl = graph.graf.node[nbr]['CLID']
if cdict.__contains__(cl):
cdict[cl] += weight
else:
cdict[cl] = weight
for cl, weight in cdict.iteritems():
if maxval < weight:
maxval = weight
maxclustids = [cl]
elif maxval == weight:
maxclustids.append(cl)
clusterlist = []
for clust in clusterglobals.clusters_list:
if clust.label in maxclustids:
clusterlist.append(clusterglobals.get_cluster(clust.label))
#return maxclustids
return clusterlist
def random_classification(graph):
for node in graph.graf.nodes():
clusters_li = get_max_weighted_clusters(graph, node)
best_cluster = random_cluster(clusters_li)
if best_cluster:
old_cluster_id = graph.graf.node[node]['CLID']
tags = graph.graf.node[node]['tags']
old_cluster = get_cluster(old_cluster_id)
st = old_cluster.remove_node(graph, node)
if st == None:
replace_cluster(old_cluster)
else:
replace_cluster(old_cluster, remove = True)
best_cluster.add_node(node, tags)
replace_cluster(best_cluster)
change_node_cluster(graph, node, best_cluster.label)
else:
pass
