网易云音乐NCM转换¶
详情参见 GitHub
'''
网易云音乐ncm格式解密
作者：HuangFuSL
依赖包：pycryptodomex
参数格式：
path [path ...] [-o OUTPUT] [-f] [-r REGEX] [-e EXCLUDE] [-q] [-v]
- path：ncm文件或目录
- -o, --output：输出目录
- -f, --force：覆盖已存在的文件
- -r, --regex：匹配文件名的正则表达式
- -e, --exclude：排除文件名的正则表达式
- -q, --quiet：不打印处理信息
- -v, --verbose：打印调试信息
输出格式
解码后的音频文件，mp3或flac格式
'''
from __future__ import annotations
import argparse
import base64
import functools
import io
import json
import logging
import math
import os
import re
from typing import Tuple
from Cryptodome.Cipher import AES
NCM_FILEHEAD = b'CTENFDAM'
NCM_COREKEY = b'hzHRAmso5kInbaxW'
NCM_SIDEKEY = b"#14ljk_!\\]&0U<'("
ARG_HELP = {
('path', ): 'NCM file or directory',
('-o', '--output'): 'Output directory',
('-f', '--force'): 'Overwrite existing files',
('-r', '--regex'): 'Regex to match filename',
('-e', '--exclude'): 'Regex to exclude filename',
('-q', '--quiet'): 'Do not print processing info',
('-v', '--verbose'): 'Print debug info'
}
ARG_TEMPLATE = {
('path', ): {'nargs': '*', 'default': '.'},
('-o', '--output'): {'type': str, 'default': None},
('-f', '--force'): {'action': 'store_true'},
('-r', '--regex'): {'type': str, 'default': None},
('-e', '--exclude'): {'type': str, 'default': None},
('-q', '--quiet'): {'action': 'store_true'},
('-v', '--verbose'): {'action': 'store_true'}
}
UNITS = ['B', 'KB', 'MB', 'GB', 'TB']
def unpad(s: str | bytes):
''' 清除数据后面填充的字节块 '''
return s[0: -s[-1]] if isinstance(s, bytes) else unpad(s.encode('utf-8'))
def aes_decrypt(key: bytes, data: bytes) -> bytes:
''' AES解密 '''
return AES.new(key, AES.MODE_ECB).decrypt(data)
def xor(data: bytes, num: int | bytes):
''' 利用整数运算进行bytes对象的快速异或 '''
if isinstance(num, int):
num = bytes([num] * len(data))
b2i = functools.partial(int.from_bytes, byteorder='little')
return (b2i(data) ^ b2i(num)).to_bytes(len(data), 'little')
def add(a: int, b: int):
''' 包含整型溢出的相加 '''
return (a + b) & 0xff
def build_keybox(decrypted: bytes):
''' 构造解密文件的函数 '''
bytemap = bytearray(range(256))
b, i = 0, 0
for _ in range(256):
b = add(bytemap[_], b + decrypted[i])
i = (i + 1) % len(decrypted)
bytemap[_], bytemap[b] = bytemap[b], bytemap[_]
keybox = bytes(
bytemap[add(bytemap[add(_, i)], _)]
for i, _ in enumerate(bytemap[1:] + bytemap[:1], 1)
)
def apply_keybox(data: bytes):
return xor(data, (keybox * (len(data) // 256 + 1))[:len(data)])
return apply_keybox
def read_chunk(data: io.BytesIO) -> bytes:
''' 分块读取文件，前4字节表示块长度 '''
chunk_size = int.from_bytes(data.read(4), 'little')
return data.read(chunk_size)
def ncm_decrypt(data: io.BytesIO) -> Tuple[str, bytes]:
''' 解密ncm文件 '''
core_key_chunk = read_chunk(data)
decrypted = unpad(aes_decrypt(NCM_COREKEY, xor(core_key_chunk, 100)))[17:]
keybox = build_keybox(decrypted) # 用于文件主体的解密
b64_chunk = xor(read_chunk(data), 99) # 元数据部分
b64_chunk_data = base64.b64decode(b64_chunk.split(b':', 1)[1])
mdc = json.loads(unpad(aes_decrypt(NCM_SIDEKEY, b64_chunk_data)[6:]))
file_name = '{musicName} - {artist[0][0]}.{format}'.format(**mdc)
data.seek(9, 1)
_ = read_chunk(data) # 专辑封面
w = io.BytesIO()
while True:
chunk = data.read(32768)
if not chunk:
break
w.write(keybox(chunk))
return file_name, w.getvalue()
def process_file(src: str, output_dir: str, force: bool = False) -> int:
with open(src, 'rb') as f:
if f.read(8) != NCM_FILEHEAD:
raise Exception('Not a NCM file')
f.seek(2, 1)
fn, data = ncm_decrypt(io.BytesIO(f.read()))
dest = os.path.join(output_dir, fn.translate(
str.maketrans(r'\/:*?"<>|', '_' * 9)))
if force or not os.path.exists(dest):
logging.debug(' Processing %s', dest)
with open(dest, 'wb') as f:
f.write(data)
return len(data)
logging.debug(' Skipping %s', dest)
return 0
if __name__ == "__main__":
parser = argparse.ArgumentParser(description='NCM converter')
for args, msg in ARG_HELP.items():
parser.add_argument(*args, **ARG_TEMPLATE[args], help=msg)
ns = parser.parse_args()
levels = [logging.INFO, logging.ERROR, logging.DEBUG, logging.DEBUG]
logging.basicConfig(
level=levels[ns.quiet + ns.verbose * 2],
format='%(message)s',
)
logging.info('Path:')
for path in ns.path:
abspath = os.path.abspath(path)
if abspath.startswith(path) or path.startswith(abspath):
logging.info(' %s', path)
else:
logging.info(' %s (%s)', path, abspath)
logging.info('Output: %s', ns.output or 'Same as input')
logging.info('Force: %s', ns.force)
logging.info('Regex: %s', ns.regex or 'None')
logging.info('Exclude: %s', ns.exclude or 'None')
logging.info('Logging level: %s', logging.getLevelName(
logging.getLogger().level))
processed, skipped, size = 0, 0, 0
for path in ns.path:
for root, _, files in os.walk(path):
targets = [
_ for _ in files
if all([
_.endswith('.ncm'),
not ns.regex or re.search(ns.regex, _),
not ns.exclude or not re.search(ns.exclude, _)
])
]
if targets:
logging.debug('Entering %s', root)
for file in targets:
try:
output_size = process_file(
os.path.join(root, file), ns.output or root, ns.force
)
size += output_size
if output_size:
processed += 1
else:
skipped += 1
except Exception as e:
logging.error('Error processing %s: %s', file, e)
logging.debug('Leaving %s', root)
if size:
size_unit = min(int(math.log2(size) / 10), len(UNITS) - 1)
size_str = f'{size / 1024 ** size_unit:.2f}{UNITS[size_unit]}'
else:
size_str = '0B'
logging.warning('No files processed')
logging.info(
'Processed %d files, skipped %d files, total size %s',
processed, skipped, size_str
)