def main():
action = sys.argv[1]
if action == "create":
indexfile = sys.argv[2]
filelist = load_listfile(sys.argv[3])
create_graph(indexfile, filelist)
if action == "create_feature":
indexfile = sys.argv[2]
level_filelist = load_listfile(sys.argv[3])
feature_filelist = load_listfile(sys.argv[4])
create_graph_feature(indexfile, level_filelist, feature_filelist)
if action == "create2":
indexfile = sys.argv[2]
filelist = load_listfile(sys.argv[3])
create_graph2(indexfile, filelist)
if action == "create_correlation":
indexfile = sys.argv[2]
filelist = load_listfile(sys.argv[3])
create_graph_correlation(indexfile, filelist)
if action == "print":
print_graph(sys.argv[2])
if action == "correlate":
find_duplicates(sys.argv[2], float(sys.argv[3]))
if action == "dot":
print_dot(sys.argv[2])
if action == "rebase_log":
graph, filenames = load_graph(sys.argv[2])
n = float(sys.argv[3])
for s_index in xrange(len(graph)):
for e_i in xrange(len(graph[s_index])):
e_index, correlation = graph[s_index][e_i]
graph[s_index][e_i] = (e_index, rebase_value_log(n, correlation))
print pickle.dumps((graph, filenames))
if action == "rebase_exp":
graph, filenames = load_graph(sys.argv[2])
n = float(sys.argv[3])
for s_index in xrange(len(graph)):
for e_i in xrange(len(graph[s_index])):
e_index, correlation = graph[s_index][e_i]
graph[s_index][e_i] = (e_index, rebase_value_exp(n, correlation))
print pickle.dumps((graph, filenames))
if action == "adjust":
graph, filenames = load_graph(sys.argv[2])
values = [float(v) for v in open(sys.argv[3], "r").readlines()]
value_index = 0
for i in xrange(len(filenames)):
for j in xrange(len(filenames)):
if i != j:
edges = graph[i]
for k, (e_index, correlation) in enumerate(edges):
if e_index == j:
edges[k] = (e_index, values[value_index])
break
value_index += 1
print pickle.dumps((graph, filenames))
if action == "extend":
graph, filenames = load_graph(sys.argv[2])
graph = fill_graph(graph, float(sys.argv[3]))
print pickle.dumps((graph, filenames))
if action == "level_update":
graph, filenames = load_graph(sys.argv[2])
level_signatures = pickle.loads(open(sys.argv[3], "r").read())
c_value = float(sys.argv[4])
l_value = float(sys.argv[5])
import signature_make
result_graph = []
for s_index, edges in enumerate(graph):
result_edges = []
for e_index, correlation in edges:
l_correlation = signature_make.signature_matrix_compare(level_signatures[s_index], level_signatures[e_index])
if l_correlation >= l_value and correlation >= c_value:
result_edges.append((e_index, l_correlation))
result_graph.append(result_edges)
print pickle.dumps((result_graph, filenames))
if action == "clique":
graph, filenames = load_graph(sys.argv[2])
r_sets = find_all_cliques([[e_index for e_index, correlation in edges] for edges in graph])
r_sets.sort(key = lambda x: -len(x))
for r_set in r_sets:
print str(list(r_set))
print str([filenames[i] for i in r_set])
if action == "auto_clique":
graph, filenames = load_graph(sys.argv[2])
weak_threshold = float(sys.argv[3])
strong_threshold = float(sys.argv[4])
weak_graph = [[e_index for e_index, correlation in edges if correlation >= weak_threshold] for edges in graph]
strong_graph = [[(e_index, correlation) for e_index, correlation in edges if correlation >= strong_threshold] for edges in graph]
print pickle.dumps((strong_graph, filenames))
def main():
action = sys.argv[1]
if action == "create":
indexfile = sys.argv[2]
filelist = load_listfile(sys.argv[3])
create_graph(indexfile, filelist)
if action == "create_feature":
indexfile = sys.argv[2]
level_filelist = load_listfile(sys.argv[3])
feature_filelist = load_listfile(sys.argv[4])
create_graph_feature(indexfile, level_filelist, feature_filelist)
if action == "create2":
indexfile = sys.argv[2]
filelist = load_listfile(sys.argv[3])
create_graph2(indexfile, filelist)
if action == "create_correlation":
indexfile = sys.argv[2]
filelist = load_listfile(sys.argv[3])
create_graph_correlation(indexfile, filelist)
if action == "print":
print_graph(sys.argv[2])
if action == "correlate":
find_duplicates(sys.argv[2], float(sys.argv[3]))
if action == "dot":
print_dot(sys.argv[2])
if action == "rebase_log":
graph, filenames = load_graph(sys.argv[2])
n = float(sys.argv[3])
for s_index in xrange(len(graph)):
for e_i in xrange(len(graph[s_index])):
e_index, correlation = graph[s_index][e_i]
graph[s_index][e_i] = (e_index, rebase_value_log(n, correlation))
print pickle.dumps((graph, filenames))
if action == "rebase_exp":
graph, filenames = load_graph(sys.argv[2])
n = float(sys.argv[3])
for s_index in xrange(len(graph)):
for e_i in xrange(len(graph[s_index])):
e_index, correlation = graph[s_index][e_i]
graph[s_index][e_i] = (e_index, rebase_value_exp(n, correlation))
print pickle.dumps((graph, filenames))
if action == "adjust":
graph, filenames = load_graph(sys.argv[2])
values = [float(v) for v in open(sys.argv[3], "r").readlines()]
value_index = 0
for i in xrange(len(filenames)):
for j in xrange(len(filenames)):
if i != j:
edges = graph[i]
for k, (e_index, correlation) in enumerate(edges):
if e_index == j:
edges[k] = (e_index, values[value_index])
break
value_index += 1
print pickle.dumps((graph, filenames))
if action == "pagerank":
import graph_reweight as rw
graph, filenames = load_graph(sys.argv[2])
# graph = rw.make_agraph(rw.reweight(rw.make_graph(threshold_graph(graph, 0.1))))
graph = rw.make_agraph(rw.reweight(rw.make_graph(graph)))
print pickle.dumps((graph, filenames))
if action == "extend":
graph, filenames = load_graph(sys.argv[2])
graph = fill_graph(graph, float(sys.argv[3]))
print pickle.dumps((graph, filenames))
if action == "level_update":
graph, filenames = load_graph(sys.argv[2])
level_signatures = pickle.loads(open(sys.argv[3], "r").read())
c_value = float(sys.argv[4])
l_value = float(sys.argv[5])
import signature_make
result_graph = []
for s_index, edges in enumerate(graph):
result_edges = []
for e_index, correlation in edges:
l_correlation = signature_make.signature_matrix_compare(level_signatures[s_index], level_signatures[e_index])
if l_correlation >= l_value and correlation >= c_value:
result_edges.append((e_index, l_correlation))
result_graph.append(result_edges)
print pickle.dumps((result_graph, filenames))
