def call_gspan_from_library(file_path):
""" Runs the gSpan algorithm to find frequent subgraphs """
from gspan_mining.config import parser
from gspan_mining.gspan import gSpan
# args_str = ' -s ' + str(s) + ' -l ' + str(l) + ' -u 4 -v False ' + '-p ' + str(plot) + ' ' + file_path
old_stdout = sys.stdout
redirected_output = sys.stdout = StringIO()
args_str = file_path + ' -l 3 ' + '-s 5000'
FLAGS, _ = parser.parse_known_args(args=args_str.split())
gs = gSpan(
database_file_name=FLAGS.database_file_name,
min_support=FLAGS.min_support,
min_num_vertices=FLAGS.lower_bound_of_num_vertices,
max_num_vertices=FLAGS.upper_bound_of_num_vertices,
max_ngraphs=FLAGS.num_graphs,
is_undirected=(not FLAGS.directed),
verbose=FLAGS.verbose,
visualize=FLAGS.plot,
where=FLAGS.where
)
gs.run()
name = file_path.split('/')[-1]
with open(SUGRAPHS_PATH + name.replace('.txt', '_frequent_subgraphs.txt'), 'w') as f:
f.write(redirected_output.getvalue())
sys.stdout = old_stdout
def test(ds_name, minSup, params='', random=False, graphlet=False, cork=False):
if len(helper.tests_run) == 0:
if random:
helper.tests_run += "random"
if graphlet:
helper.tests_run += "graphlet"
if cork:
helper.tests_run += "cork"
print_info('\n-----[BEGIN]-----\n')
dataset = TUDataset(root='./tmp/{}'.format(ds_name), name=ds_name)
print_info('Starting Tests with dataset: {}, containing {} Graphs'.format(
ds_name, len(dataset)))
is_directed = dataset[0].is_directed()
# create the run arguments for gSpan-python
cwd = os.getcwd()
f_name = '{}.data.txt'.format(ds_name)
f_path = path.join(cwd, 'graphs', f_name)
args_str = '--min_support {min} --directed {directed} {params} --verbose FALSE --where TRUE {ds}'.format(
ds=f_path,
min=int(len(dataset) * minSup),
directed=is_directed,
params=params)
print_info('Running with params: {}'.format(args_str))
FLAGS, _ = parser.parse_known_args(args=args_str.split())
# mine with gSpan
print_info(
"Starting mining with gSpan-Algorithm and minSup sigma = {}%".format(
minSup * 100))
gs = main(FLAGS)
_report = gs._report_df
print_info("\nFinished mining. Found {} freq. subgraphs.".format(
len(_report)))
_freq = []
sum_edges = 0
sum_vertices = 0
for dfs in _report['dfs']:
_freq.append(dfs)
sum_edges += len(dfs)
sum_vertices += dfs.get_num_vertices()
print_info("Durchschnitt Knoten: {}, Durchschnitt Kanten: {}\n".format(
sum_vertices / len(_freq), sum_edges / len(_freq)))
_desc = [desc for desc in _report['description']]
# get info needed for testing
ds_graph_classes = dataset.data.y.tolist() # graph classes
isomorph_graphs = [gids for gids in _report['isomorph_graphs']]
# perform test
_tests = tests.Tests(_freq, isomorph_graphs, ds_graph_classes, _desc)
_tests.run(random, graphlet, cork)
print_info('\n-----[END]-----\n')
def generate_subgraphs_one_core(gspan_file_name,
s,
gspan_data_folder,
l=3,
plot=False):
filepath = gspan_data_folder + gspan_file_name + '.txt'
args_str = ' -s ' + str(s) + ' -l ' + str(
l) + ' -u 4 -v False ' + '-p ' + str(plot) + ' ' + filepath
FLAGS, _ = parser.parse_known_args(args=args_str.split())
gs = gSpan(database_file_name=FLAGS.database_file_name,
min_support=FLAGS.min_support,
min_num_vertices=FLAGS.lower_bound_of_num_vertices,
max_num_vertices=FLAGS.upper_bound_of_num_vertices,
max_ngraphs=FLAGS.num_graphs,
is_undirected=(not FLAGS.directed),
verbose=FLAGS.verbose,
visualize=FLAGS.plot,
where=FLAGS.where)
gs.run()
report = gs._report_df
gs = None
return report
else:
print("ENABLEMENT:")
print(pm,'\n')
l_sol = solutions(l_sol,pm)
stc = stc + 1
s = s + 1
# il file viene chiuso esternamente alla funzione per evitare sovrascritture dei dati di input
sub_graph.write("t # -1")
sub_graph.write("\n")
sub_graph.close()
# richiamo gSpan e visualizzazione grafi, per ogni enablement del doc considerato
args_str = '-d True /Users/filippo/Desktop/UNIMI/Tesi/inputgSpan_sentence.data'
FLAGS, _ = parser.parse_known_args(args=args_str.split())
gs = main(FLAGS)
for g in gs.graphs.values():
g.plot()
print('\n')
match_caus(l_sol)