def conduct_triad_analysis(labelled_data, pid_dict):
# get the basic networkx graphing constructs
p, np, edges, ground_edges, polarity_count = create_graph_constructs(labelled_data, pid_dict)
# get a random distribution of polarity
random_edges, random_dist, random_labels = get_random_edge_polarity(polarity_count, edges)
# create graph
graph = creategraph()
graph.addnodes(p, 'P')
graph.addnodes(np, 'NP')
graph.addedges(edges)
deg_dist = get_degree_distribution(graph, p)
# create random graph
graph_rand = creategraph()
graph_rand.addnodes(p, 'P')
graph_rand.addnodes(np, 'NP')
graph_rand.addedges(random_edges)
deg_dist_rand = get_degree_distribution(graph_rand, np)
triad_dict = find_triads(graph, pid_dict, min_participants=2)
triad_dict_rand = find_triads(graph_rand, pid_dict, min_participants=2)
# summarize the types of edges
summary_triad, frac_triad = summarize_triad(triad_dict)
summary_triad_rand, frac_triad_rand = summarize_triad(triad_dict_rand)
return [graph, triad_dict, summary_triad, frac_triad, deg_dist], \
[graph_rand, triad_dict_rand, summary_triad_rand, frac_triad_rand, deg_dist_rand]
def createstaticgraph(pid_dict, data_list):
links, links_tuple = hlp.getlinks(pid_dict, data_list)
graph_obj = creategraph(True)
for p_type in pid_dict.keys():
graph_obj.addnodes(pid_dict[p_type], p_type)
graph_obj.addedges(links_tuple)
return graph_obj
def weekgraphs(week_dict, participant_list, non_participant_list):
graph_objs = {}
for week_no in week_dict.keys():
print 'week no', week_no
graph_obj = creategraph(True)
nodes, link_tuple = getnodesedges(week_dict[week_no])
p_nodes, np_nodes = checknodemembership(nodes, participant_list, non_participant_list)
graph_obj.addnodes(p_nodes, 'P')
graph_obj.addnodes(np_nodes, 'NP')
graph_obj.addedges(link_tuple)
print 'Calculating properties'
toStore = {'object': graph_obj.getgraphobject(),
'triadic_census': graph_obj.gettriadiccensus(),
'closeness': graph_obj.getclosenesscentrality()}
graph_objs[week_no] = toStore
return graph_objs
