Split data into chunks and slap multiprocessing on it to lower memory footprint and execution time

This commit is contained in:
ChaoticByte 2024-06-24 18:07:06 +02:00
parent 87d50d9f8a
commit 8b0766b719
No known key found for this signature in database

View file

@ -1,21 +1,21 @@
# Copyright (c) 2024 Julian Müller (ChaoticByte)
# License: MIT
def obf(data: bytes, key: bytes, decrypt: bool = False, iterations: int = 8) -> bytes:
assert type(data) == bytes
assert type(key) == bytes
assert type(iterations) == int
assert type(decrypt) == bool
data = bytearray(data)
key = bytearray(key)
len_data = len(data)
from multiprocessing import Pool
CHUNKSIZE = 1024 * 4
def _obf(args) -> bytearray:
data_chunk, key, decrypt, iterations = args
len_data = len(data_chunk)
len_key = len(key)
for _ in range(iterations):
# shift (encrypt)
if not decrypt:
for i in range(len_data):
n = key[i % len_key]
data[i] = (data[i] + n) % 256
data_chunk[i] = (data_chunk[i] + n) % 256
# transpose
# list of tuples that stores transposition data (from, to):
swap_indices = [] # (this is extremely memory inefficient lol)
@ -28,16 +28,44 @@ def obf(data: bytes, key: bytes, decrypt: bool = False, iterations: int = 8) ->
swap_indices.reverse()
for a, b in swap_indices:
# swap values
a_ = data[a]
b_ = data[b]
data[a] = b_
data[b] = a_
a_ = data_chunk[a]
b_ = data_chunk[b]
data_chunk[a] = b_
data_chunk[b] = a_
# unshift (decrypt)
if decrypt:
for i in range(len_data):
n = key[i % len_key]
b = data[i] - n
b = data_chunk[i] - n
while b < 0:
b = 256 + b
data[i] = b
return bytes(data)
data_chunk[i] = b
return data_chunk
def obf(data: bytes, key: bytes, decrypt: bool = False, iterations: int = 8, processes: int = 4) -> bytes:
assert type(data) == bytes
assert type(key) == bytes
assert type(iterations) == int
assert type(decrypt) == bool
assert type(processes) == int
data = bytearray(data)
key = bytearray(key)
len_data_complete = len(data)
chunks = []
p = 0
while p < len_data_complete:
p_new = p + CHUNKSIZE
chunk = data[p:p_new]
chunks.append((chunk, key, decrypt, iterations))
p = p_new
del data
pool = Pool(processes=4)
results = pool.map(_obf, chunks)
del chunks
return bytes(b''.join(results))