def sim_programs(self, prog1, prog2):
headers_sim = self._simimilarity_dicts(prog1.num_of_headers(),
prog2.num_of_headers())
# control types
ctrl_type_sim = self._simimilarity_dicts(prog1.num_of_control_types(),
prog2.num_of_control_types())
# local variables
loc_vars_sim = self._simimilarity_dicts(prog1.num_of_local_variables(),
prog2.num_of_local_variables())
# global variable
glob_vars_sim = self._simimilarity_dicts(
prog1.num_of_global_variables(), prog2.num_of_global_variables())
sim = Config.get('fast.ctrl_type_power') * ctrl_type_sim
sim += Config.get('fast.headers_power') * headers_sim
sim += Config.get('fast.loc_vars_power') * loc_vars_sim
sim += Config.get('fast.glob_vars_power') * glob_vars_sim
self._logger.info("ctrl_type_sim = %f" % (ctrl_type_sim))
self._logger.info("headers_sim = %f" % (headers_sim))
self._logger.info("loc_vars_sim = %f" % (loc_vars_sim))
self._logger.info("glob_vars_sim = %f" % (glob_vars_sim))
self._logger.info("fast.sim_programs(%s, %s) = %f" %
(prog1.name, prog2.name, sim))
return sim
def sim_functions(self, func1, func2):
# control types
ctrl_type_sim = self._simimilarity_dicts(
func1.num_of_control_types(),
func2.num_of_control_types()
)
# local variables
loc_vars_sim = self._simimilarity_dicts(
func1.num_of_local_variables(),
func2.num_of_local_variables()
)
# return type
ret_type_sim = 1 if func1.ret_type.code() == func2.ret_type.code() else 0
sim = Config.get('fast.func_ctrl_type_power') * ctrl_type_sim
sim += Config.get('fast.func_loc_vars_power') * loc_vars_sim
self._logger.info("ctrl_type_sim = %f" % (ctrl_type_sim))
self._logger.info("loc_vars_sim = %f" % (loc_vars_sim))
self._logger.info("fast.sim_functions(%s, %s) = %f" % (func1.name, func2.name, sim))
return sim
def extract_functions(self, target_prog, target, data, out_dir):
out_dir += "/%s/funcs/%s" % (target_prog.name, target.name)
if not os.path.exists(out_dir):
os.makedirs(out_dir)
filename_src = "%s/src.c" % (out_dir)
filename_sim = "%s/sim.c" % (out_dir)
src = SCFileDecorator(filename_src)
src.write_comment(["generated with %s %s" % (Config.get('package'), Config.get('version'))])
src.write_comment(["source code of", "target function=%s" % (target.name)])
src.write(self.__extracted_function(target_prog, target))
src.write_comment(["end of source code"])
del src
src = SCFileDecorator(filename_sim)
src.write_comment(["generated with %s %s" % (Config.get('package'), Config.get('version'))])
src.write_comment(["source code of", "similar functions"])
for prog,func,f,s in data:
src.write_comment([
"program=%s" % (prog.name),
"function=%s" % (func.name),
"f.sim=%f" % (f),
"s.sim=%f" % (s)
])
src.write(self.__extracted_function(prog, func))
src.write_comment(["end of function=%s" % (func.name)])
del src
def code(self):
typ = '_'.join(self.names)
return "%d%d%d%d%d%d%02d" % (
min(self.pointer_count, Config.get('max_pointer_count')),
min(self.func_pointer_count, Config.get('max_func_pointer_count')),
min(self.array_count, Config.get('max_array_count')),
self.is_struct, self.is_union, self.is_enum,
self.TYPES[typ] if self.TYPES.has_key(typ) else self.UNKNOW_TYPE)
def code(self):
typ = '_'.join(self.names)
return "%d%d%d%d%d%d%02d" % (
min(self.pointer_count, Config.get('max_pointer_count')),
min(self.func_pointer_count, Config.get('max_func_pointer_count')),
min(self.array_count, Config.get('max_array_count')),
self.is_struct,
self.is_union,
self.is_enum,
self.TYPES[typ] if self.TYPES.has_key(typ) else self.UNKNOW_TYPE
)
def sim_functions(self, func1, func2):
# control types
ctrl_type_sim = self._simimilarity_dicts(
func1.num_of_control_types(),
func2.num_of_control_types()
)
# argument types
args_vars_sim = self._simimilarity_dicts(
func1.num_of_arguments_variables(),
func2.num_of_arguments_variables()
)
# local variables
loc_vars_sim = self._simimilarity_dicts(
func1.num_of_local_variables(),
func2.num_of_local_variables()
)
# control sequence
ctrl_seq_sim = self._simimilarity_lists(
func1.seq_of_control_types(),
func2.seq_of_control_types()
)
# return type
ret_type_sim = 1 if func1.ret_type.code() == func2.ret_type.code() else 0
sim = Config.get('detailed.ctrl_type_power') * ctrl_type_sim
sim += Config.get('detailed.ctrl_seq_power') * ctrl_seq_sim
sim += Config.get('detailed.loc_vars_power') * loc_vars_sim
sim += Config.get('detailed.ret_type_power') * ret_type_sim
sim += Config.get('detailed.args_vars_power') * args_vars_sim
self._logger.info("ctrl_type_sim = %f" % (ctrl_type_sim))
self._logger.info("ctrl_type_sim = %f" % (ctrl_type_sim))
self._logger.info("ctrl_seq_sim = %f" % (ctrl_seq_sim))
self._logger.info("loc_vars_sim = %f" % (loc_vars_sim))
self._logger.info("ret_type_sim = %f" % (ret_type_sim))
self._logger.info("args_vars_sim = %f" % (args_vars_sim))
self._logger.info("detailed.sim_functions(%s, %s) = %f" % (func1.name, func2.name, sim))
return sim
def sim_programs(self, prog1, prog2):
# headers
headers_sim = self._simimilarity_dicts(
prog1.num_of_headers(),
prog2.num_of_headers()
)
# global variable
glob_vars_sim = self._simimilarity_dicts(
prog1.num_of_global_variables(),
prog2.num_of_global_variables()
)
# function similiar
precision = 10000
matrix = []
for i in range(0, len(prog1.functions)):
matrix.append(range(len(prog2.functions)))
for j in range(0, len(prog2.functions)):
matrix[i][j] = self.sim_functions(prog1.functions[i], prog2.functions[j]) * precision
cost_matrix = make_cost_matrix(matrix, lambda cost: sys.maxint - cost)
m = Munkres()
indexes = m.compute(cost_matrix)
score = 1.0 / (prog1.count_of_control_blocks() + prog2.count_of_control_blocks())
funcs_sim = 0
for i,j in indexes:
value = matrix[i][j]
funcs_sim += (value / precision) * \
((len(prog1.functions[i].seq_of_control_types()) + len(prog2.functions[j].seq_of_control_types()))* score)
#funcs_sim /= max(len(prog1.functions), len(prog2.functions))
sim = Config.get('detailed.glob_vars_power') * glob_vars_sim
sim += Config.get('detailed.headers_power') * headers_sim
sim += Config.get('detailed.funcs_power') * funcs_sim
self._logger.info("headers_sim = %f" % headers_sim)
self._logger.info("glob_vars_sim = %f" % glob_vars_sim)
self._logger.info("funcs_sim = %f" % funcs_sim)
self._logger.info("detailed.sim_programs(%s, %s) = %f" % (prog1.name, prog1.name, sim))
return sim
def sim_programs(self, prog1, prog2):
# headers
headers_sim = self._simimilarity_dicts(prog1.num_of_headers(),
prog2.num_of_headers())
# global variable
glob_vars_sim = self._simimilarity_dicts(
prog1.num_of_global_variables(), prog2.num_of_global_variables())
# function similiar
precision = 10000
matrix = []
for i in range(0, len(prog1.functions)):
matrix.append(range(len(prog2.functions)))
for j in range(0, len(prog2.functions)):
matrix[i][j] = self.sim_functions(
prog1.functions[i], prog2.functions[j]) * precision
cost_matrix = make_cost_matrix(matrix, lambda cost: sys.maxint - cost)
m = Munkres()
indexes = m.compute(cost_matrix)
score = 1.0 / (prog1.count_of_control_blocks() +
prog2.count_of_control_blocks())
funcs_sim = 0
for i, j in indexes:
value = matrix[i][j]
funcs_sim += (value / precision) * \
((len(prog1.functions[i].seq_of_control_types()) + len(prog2.functions[j].seq_of_control_types()))* score)
#funcs_sim /= max(len(prog1.functions), len(prog2.functions))
sim = Config.get('detailed.glob_vars_power') * glob_vars_sim
sim += Config.get('detailed.headers_power') * headers_sim
sim += Config.get('detailed.funcs_power') * funcs_sim
self._logger.info("headers_sim = %f" % headers_sim)
self._logger.info("glob_vars_sim = %f" % glob_vars_sim)
self._logger.info("funcs_sim = %f" % funcs_sim)
self._logger.info("detailed.sim_programs(%s, %s) = %f" %
(prog1.name, prog1.name, sim))
return sim
def sim_functions(self, func1, func2):
# control types
ctrl_type_sim = self._simimilarity_dicts(func1.num_of_control_types(),
func2.num_of_control_types())
# argument types
args_vars_sim = self._simimilarity_dicts(
func1.num_of_arguments_variables(),
func2.num_of_arguments_variables())
# local variables
loc_vars_sim = self._simimilarity_dicts(func1.num_of_local_variables(),
func2.num_of_local_variables())
# control sequence
ctrl_seq_sim = self._simimilarity_lists(func1.seq_of_control_types(),
func2.seq_of_control_types())
# return type
ret_type_sim = 1 if func1.ret_type.code() == func2.ret_type.code(
) else 0
sim = Config.get('detailed.ctrl_type_power') * ctrl_type_sim
sim += Config.get('detailed.ctrl_seq_power') * ctrl_seq_sim
sim += Config.get('detailed.loc_vars_power') * loc_vars_sim
sim += Config.get('detailed.ret_type_power') * ret_type_sim
sim += Config.get('detailed.args_vars_power') * args_vars_sim
self._logger.info("ctrl_type_sim = %f" % (ctrl_type_sim))
self._logger.info("ctrl_type_sim = %f" % (ctrl_type_sim))
self._logger.info("ctrl_seq_sim = %f" % (ctrl_seq_sim))
self._logger.info("loc_vars_sim = %f" % (loc_vars_sim))
self._logger.info("ret_type_sim = %f" % (ret_type_sim))
self._logger.info("args_vars_sim = %f" % (args_vars_sim))
self._logger.info("detailed.sim_functions(%s, %s) = %f" %
(func1.name, func2.name, sim))
return sim
def sim_functions(self, func1, func2):
# control types
ctrl_type_sim = self._simimilarity_dicts(func1.num_of_control_types(),
func2.num_of_control_types())
# local variables
loc_vars_sim = self._simimilarity_dicts(func1.num_of_local_variables(),
func2.num_of_local_variables())
# return type
ret_type_sim = 1 if func1.ret_type.code() == func2.ret_type.code(
) else 0
sim = Config.get('fast.func_ctrl_type_power') * ctrl_type_sim
sim += Config.get('fast.func_loc_vars_power') * loc_vars_sim
self._logger.info("ctrl_type_sim = %f" % (ctrl_type_sim))
self._logger.info("loc_vars_sim = %f" % (loc_vars_sim))
self._logger.info("fast.sim_functions(%s, %s) = %f" %
(func1.name, func2.name, sim))
return sim
def sim_programs(self, prog1, prog2):
headers_sim = self._simimilarity_dicts(
prog1.num_of_headers(),
prog2.num_of_headers()
)
# control types
ctrl_type_sim = self._simimilarity_dicts(
prog1.num_of_control_types(),
prog2.num_of_control_types()
)
# local variables
loc_vars_sim = self._simimilarity_dicts(
prog1.num_of_local_variables(),
prog2.num_of_local_variables()
)
# global variable
glob_vars_sim = self._simimilarity_dicts(
prog1.num_of_global_variables(),
prog2.num_of_global_variables()
)
sim = Config.get('fast.ctrl_type_power') * ctrl_type_sim
sim += Config.get('fast.headers_power') * headers_sim
sim += Config.get('fast.loc_vars_power') * loc_vars_sim
sim += Config.get('fast.glob_vars_power') * glob_vars_sim
self._logger.info("ctrl_type_sim = %f" % (ctrl_type_sim))
self._logger.info("headers_sim = %f" % (headers_sim))
self._logger.info("loc_vars_sim = %f" % (loc_vars_sim))
self._logger.info("glob_vars_sim = %f" % (glob_vars_sim))
self._logger.info("fast.sim_programs(%s, %s) = %f" % (prog1.name, prog2.name, sim))
return sim
def parse_headers(filename):
headers = []
f = open(filename)
content = f.readlines()
for line in content:
if line.strip().startswith('#include'):
s = line.find('<')
if s == -1:
s = line.find('"')
f = line.find('\"', s+1)
else:
f = line.find('>')
header = line[s+1:f]
if header in Config.get('standart_headers'):
headers.append(header)
return headers