def abstract_code(code, build_dir=None, build_src_path=None):
try:
c_tokenizer = CTokenizer(build_dir=build_dir, build_src_path=build_src_path)
tokens = c_tokenizer.tokenize(code)
except Exception:
c_tokenizer = CTokenizer()
tokens = c_tokenizer.tokenize(code)
return c_tokenizer.detokenize(tokens)
def abstract_code(code, build_dir=None, build_src_path=None):
try:
c_tokenizer = CTokenizer(build_dir=build_dir,
build_src_path=build_src_path)
tokens = c_tokenizer.tokenize(code)
except Exception:
c_tokenizer = CTokenizer()
tokens = c_tokenizer.tokenize(code)
return c_tokenizer.detokenize(tokens)
#INPUT_DIR = "/dev/shm/output/abstracted_decompiled_code"
#OUTPUT_DIR = "/dev/shm/output/abstracted_base64_decompiled_code"
INPUT_DIR = "/dev/shm/output/abstracted_src_code"
OUTPUT_DIR = "/dev/shm/output/abstracted_base64_src_code"
os.system("mkdir -p %s" % OUTPUT_DIR)
for fname in tqdm.tqdm(os.listdir(INPUT_DIR), smoothing=0, dynamic_ncols=True):
fpath = os.path.join(INPUT_DIR, fname)
fid = int(fname.split(".")[0])
with open(fpath) as f:
code = f.read()
c_tokenizer = CTokenizer()
tokens = c_tokenizer.tokenize(code)
encoded_tokens = [base64.b64encode(x.encode()).decode() for x in tokens]
output_fpath = os.path.join(OUTPUT_DIR, fname)
with open(output_fpath, 'w') as f:
f.write(" ".join(encoded_tokens))
default='',
help='The binary to tokenize.')
parser.add_argument('-n',
'--func-name',
default='',
help='The function name in the binary to tokenize.')
args = parser.parse_args()
sanitize_input(args)
# choose a proper tokenizer
if args.input_file:
with io.open(args.input_file, encoding='utf-8') as f:
source = f.read()
tokenizer = CTokenizer(build_dir=args.build_dir,
build_src_path=args.build_src_path)
tokens = tokenizer.tokenize(source)
print(source)
print("------")
elif args.input_bin:
tokenizer = IDATokenizer()
tokens = tokenizer.tokenize(args.input_bin, args.func_name)
else:
raise
print(tokens)
print("------")
new_code = tokenizer.detokenize(tokens)
print(new_code)
#INPUT_DIR = "/dev/shm/output/abstracted_decompiled_code"
#OUTPUT_DIR = "/dev/shm/output/abstracted_base64_decompiled_code"
INPUT_DIR_BASE = "./full_base64"
OUTPUT_DIR_BASE = "./full_base64_decoded"
for sub_dir in ["sinput", "abstracted_base64_src_code", "result"]:
input_dir = os.path.join(INPUT_DIR_BASE, sub_dir)
output_dir = os.path.join(OUTPUT_DIR_BASE, sub_dir)
os.system("mkdir -p %s" % input_dir)
os.system("mkdir -p %s" % output_dir)
for fname in tqdm.tqdm(os.listdir(input_dir), smoothing=0, dynamic_ncols=True):
fpath = os.path.join(input_dir, fname)
fid = int(fname.split(".")[0])
with open(fpath) as f:
code = f.read()
tokens = []
for x in code.split():
try:
y = base64.b64decode(x).decode('utf8', 'ignore')
except Exception:
y = '<unk>'
tokens.append(y)
c_tokenizer = CTokenizer()
output_code = c_tokenizer.detokenize(tokens)
output_fpath = os.path.join(output_dir, fname)
with open(output_fpath, 'w') as f:
f.write(output_code)