def async_trim_funcs(func_list, file_name):
return fhc.trim_funcs(func_list, file_name=file_name)
def main():
parser = argparse.ArgumentParser()
parser.add_argument("Directory")
parser.add_argument("-L",
"--LD_PATH",
default="",
help="Path to libraries to load")
parser.add_argument("-F", "--Function", default="")
parser.add_argument("-V", "--Vuln_Pickle", default="")
args = parser.parse_args()
executables, shared_libs = fhc.get_executable_files(args.Directory)
all_files = executables
#all_files = executables + shared_libs
# all_arg_funcs = async_and_iter(async_get_arg_funcs, all_files)
if args.Vuln_Pickle:
with open(args.Vuln_Pickle, 'rb') as f:
file_vulnerabilities = pickle.load(f)
pass
else:
file_vulnerabilities = vd.process_file_or_folder(
args.Directory, args.LD_PATH)
with open('cluster_pickle', 'wb') as f:
pickle.dump(file_vulnerabilities, f, -1)
print("[+] Getting sparse functions")
all_functions = []
all_trim_funcs = []
for function_list in async_and_iter(async_get_sparse_file_data, all_files):
all_functions.extend(copy.deepcopy(function_list))
all_trim_funcs.extend(
fhc.trim_funcs(function_list, function_list[0]['file_name']))
all_functions = fhc.remove_non_needed_functions(all_functions,
remove_features=False)
all_trim_funcs = fhc.remove_non_needed_functions(all_trim_funcs)
print("[+] Clustering and scoring centroid counts")
all_scores = async_and_iter_clusters(all_trim_funcs, 50)
largest_dif = 200
large_index = 0
for x in range(1, len(all_scores) - 2):
if largest_dif > all_scores[x]['score']:
largest_dif = all_scores[x]['score']
large_index = x
'''
if largest_dif < abs(scores[x] - scores[x+1]):
largest_dif = abs(scores[x] - scores[x+1])
large_index = x+1
'''
print("Largest drop at {} with {}".format(all_scores[large_index]['count'],
largest_dif))
Largest_Score_Drop = all_scores[large_index]
function_distances = fhc.get_cosine_dist(all_trim_funcs)
if args.Function:
bugs = [x for x in file_vulnerabilities if args.Function in x['name']]
else:
bugs = [x for x in file_vulnerabilities if x['result']]
for file_vuln in bugs:
vuln_index = get_function_index(file_vuln, all_functions)
if vuln_index is None:
continue
vuln_cluster = Largest_Score_Drop['labels'][vuln_index]
similar_list = get_functions_on_cluster(all_functions, vuln_cluster,
Largest_Score_Drop['labels'])
reduced_list = []
for func in similar_list:
func_distance = get_func_dist(file_vuln, func, function_distances,
all_functions)
reduced_list.append({
'file_name': func['file_name'],
'func_name': func['name'],
'distance': func_distance
})
reduced_list = sorted(reduced_list, key=lambda x: x['distance'])
file_vuln['Similar_Funcs'] = reduced_list
if bugs:
print_function(bugs[0])