def __init__(self, output_location, file_identifier):
self.MODEL_FILE = os.path.join(os.path.dirname(__file__),
'complex.RandomForest.model')
self.WEKA = os.path.join(os.path.dirname(__file__),
'tools/weka-3.6.10.jar')
arff_path = output_location + '/' + file_identifier + '_arff_output'
self.feature_gen = Feature_Generator(arff_path, 0)
def __init__(self,output_location,file_identifier):
self.MODEL_FILE = os.path.join(os.path.dirname(__file__),'complex.RandomForest.model')
self.WEKA = os.path.join(os.path.dirname(__file__),'tools/weka-3.6.10.jar')
arff_path = output_location+'/'+file_identifier+'_arff_output'
self.feature_gen = Feature_Generator(arff_path, 0)
class Predict_Input:
CMD_PREFIX = 'java -cp'
MODEL_FILE = 'complex.RandomForest.model'
CLASSIFIER = 'weka.classifiers.trees.RandomForest'
WEKA = 'tools/weka-3.6.10.jar'
ARFF_FILE_POSTFIX = '_complex.arff'
apk_file = ''
output_path = None
feature_gen = None
def __init__(self, output_location, file_identifier):
self.MODEL_FILE = os.path.join(os.path.dirname(__file__),
'complex.RandomForest.model')
self.WEKA = os.path.join(os.path.dirname(__file__),
'tools/weka-3.6.10.jar')
arff_path = output_location + '/' + file_identifier + '_arff_output'
self.feature_gen = Feature_Generator(arff_path, 0)
weka_output = ''
ids = []
def predict(self, apk_filename, output_path, output_location,
file_identifier):
if not os.path.exists(apk_filename):
raise RuntimeError('File does not exist!')
self.output_path = os.path.join(os.path.dirname(__file__), output_path)
self.apk_file = apk_filename
print 'apk decompile path:', output_path
if self.output_path is None:
output_path = 'output'
decompiler = Decompile_Apk(
os.path.dirname(os.path.abspath(self.apk_file)), output_path)
decompiler.decompile()
print 'Parsing xmls ....'
parser = Android_Layout_Parser()
data = parser.process_app(self.output_path, output_location,
file_identifier)
print 'Entries found: ', len(data)
if len(data) == 0:
print 'No data!'
return
os.chdir(os.path.dirname(__file__))
self.ids = self.feature_gen.generate(data)
#print ','.join(self.ids)
print 'Running weka ... '
self.weka_output = self.__run_weka__(output_location, file_identifier)
print self.weka_output
weka_output = open(self.output_path + '_weka_output', 'w')
weka_output.write(self.weka_output)
weka_output.close()
return self.parse_output()
output_types = []
def parse_output(self):
lines = self.weka_output.split('\n')
lineidx = 0
output_type_text = []
for line in lines:
line = line.strip()
if (len(line) == 0) or (
line.startswith('=')) or line.startswith('inst#'):
continue
entries = line.split(' ')
idx = 0
for entry in entries:
if entry:
if idx == 2:
vals = entry.split(':')
value = int(vals[1])
output_type = self.feature_gen.get_input_hex(value)
self.output_types.append(output_type)
output_type_txt = self.feature_gen.get_input_text(
value)
output_type_text.append(output_type_txt)
idx = idx + 1
lineidx = lineidx + 1
output_dict = {}
output_dict_txt = {}
idx = 0
while idx < len(self.output_types):
output_dict[self.ids[idx]] = self.output_types[idx]
output_dict_txt[self.ids[idx]] = output_type_text[idx]
idx = idx + 1
for key, value in output_dict_txt.iteritems():
print key, value
for key, value in output_dict.iteritems():
print key, value
return output_dict
def __run_weka__(self, output_location, file_identifier):
#arff_output_file = os.path.join(os.path.dirname(__file__), ('arff_output' + self.ARFF_FILE_POSTFIX))
os.chdir(os.path.dirname(__file__))
#print 'CWD: ', os.getcwd()
# file_path = output_location+'/'+file_identifier
arff_output_file = output_location + '/' + file_identifier + '_arff_output' + self.ARFF_FILE_POSTFIX
print "Reference loc " + arff_output_file
#arff_output_file = 'arff_output' + self.ARFF_FILE_POSTFIX
cmd_str = self.CMD_PREFIX + ' ' \
+ os.path.join(os.getcwd(), self.WEKA) + ' ' \
+ self.CLASSIFIER + ' ' \
+ '-T ' + arff_output_file + ' ' \
+ ' -l ' + self.MODEL_FILE \
+ ' -p 0'
return run_command(cmd_str, os.getcwd())
class Predict_Input:
CMD_PREFIX = 'java -cp'
MODEL_FILE = 'complex.RandomForest.model'
CLASSIFIER = 'weka.classifiers.trees.RandomForest'
WEKA = 'tools/weka-3.6.10.jar'
ARFF_FILE_POSTFIX = '_complex.arff'
apk_file = ''
output_path = None
feature_gen = None
def __init__(self,output_location,file_identifier):
self.MODEL_FILE = os.path.join(os.path.dirname(__file__),'complex.RandomForest.model')
self.WEKA = os.path.join(os.path.dirname(__file__),'tools/weka-3.6.10.jar')
arff_path = output_location+'/'+file_identifier+'_arff_output'
self.feature_gen = Feature_Generator(arff_path, 0)
weka_output = ''
ids = []
def predict(self, apk_filename, output_path, output_location, file_identifier):
if not os.path.exists(apk_filename):
raise RuntimeError('File does not exist!')
self.output_path = os.path.join(os.path.dirname(__file__),output_path)
self.apk_file = apk_filename
print 'apk decompile path:', output_path
if self.output_path is None:
output_path = 'output'
decompiler = Decompile_Apk(os.path.dirname(os.path.abspath(self.apk_file)), output_path)
decompiler.decompile()
print 'Parsing xmls ....'
parser = Android_Layout_Parser()
data = parser.process_app(self.output_path,output_location,file_identifier)
print 'Entries found: ', len(data)
if len(data) == 0:
print 'No data!'
return
os.chdir(os.path.dirname(__file__))
self.ids = self.feature_gen.generate(data)
#print ','.join(self.ids)
print 'Running weka ... '
self.weka_output = self.__run_weka__(output_location, file_identifier)
print self.weka_output
weka_output = open(self.output_path + '_weka_output' , 'w')
weka_output.write(self.weka_output)
weka_output.close()
return self.parse_output()
output_types = []
def parse_output(self):
lines = self.weka_output.split('\n')
lineidx = 0
output_type_text = []
for line in lines:
line = line.strip()
if (len(line) == 0) or (line.startswith('=') ) or line.startswith('inst#'):
continue
entries = line.split(' ')
idx = 0
for entry in entries :
if entry:
if idx == 2:
vals = entry.split(':')
value = int(vals[1])
output_type = self.feature_gen.get_input_hex(value)
self.output_types.append(output_type)
output_type_txt = self.feature_gen.get_input_text(value)
output_type_text.append(output_type_txt)
idx = idx + 1
lineidx = lineidx + 1
output_dict = {}
output_dict_txt = {}
idx = 0
while idx < len(self.output_types):
output_dict[self.ids[idx]] = self.output_types[idx]
output_dict_txt[self.ids[idx]] = output_type_text[idx]
idx = idx + 1
for key, value in output_dict_txt.iteritems():
print key, value
for key, value in output_dict.iteritems():
print key, value
return output_dict
def __run_weka__(self,output_location, file_identifier):
#arff_output_file = os.path.join(os.path.dirname(__file__), ('arff_output' + self.ARFF_FILE_POSTFIX))
os.chdir(os.path.dirname(__file__))
#print 'CWD: ', os.getcwd()
# file_path = output_location+'/'+file_identifier
arff_output_file = output_location+'/'+file_identifier+'_arff_output' + self.ARFF_FILE_POSTFIX
print "Reference loc "+ arff_output_file
#arff_output_file = 'arff_output' + self.ARFF_FILE_POSTFIX
cmd_str = self.CMD_PREFIX + ' ' \
+ os.path.join(os.getcwd(), self.WEKA) + ' ' \
+ self.CLASSIFIER + ' ' \
+ '-T ' + arff_output_file + ' ' \
+ ' -l ' + self.MODEL_FILE \
+ ' -p 0'
return run_command(cmd_str, os.getcwd())
