print("Performing stochastic block partitioning")
# begin partitioning by finding the best partition with the optimal number of blocks
evaluation = Evaluation(args, graph)
partition = sbp.stochastic_block_partition(graph.num_nodes,
graph.num_edges,
graph.out_neighbors,
graph.in_neighbors)
t_end = timeit.default_timer()
print('\nGraph partition took {} seconds'.format(t_end - t_start))
evaluation.total_runtime(t_start, t_end)
if args.sample_type != "none":
evaluation.evaluate_subgraph_sampling(full_graph, subgraph,
full_graph_partition,
subgraph_partition,
block_mapping, vertex_mapping)
evaluation.num_nodes = full_graph.num_nodes
evaluation.num_edges = full_graph.num_edges
# evaluate output partition against the true partition
evaluate_subgraph_partition(subgraph.true_block_assignment,
subgraph_partition.block_assignment,
evaluation)
evaluate_partition(full_graph.true_block_assignment,
full_graph_partition.block_assignment, evaluation)
else:
evaluate_partition(graph.true_block_assignment,
partition.block_assignment, evaluation)
