1
0
mirror of https://github.com/ytdl-org/youtube-dl synced 2024-12-22 09:27:01 +00:00
youtube-dl/youtube_dl/extractor/iqiyi.py
2016-06-26 15:45:41 +08:00

461 lines
15 KiB
Python

# coding: utf-8
from __future__ import unicode_literals
import binascii
import hashlib
import itertools
import math
import re
import time
from .common import InfoExtractor
from ..compat import (
compat_str,
compat_urllib_parse_urlencode,
)
from ..utils import (
decode_packed_codes,
ExtractorError,
intlist_to_bytes,
ohdave_rsa_encrypt,
remove_start,
urshift,
)
def md5_text(text):
return hashlib.md5(text.encode('utf-8')).hexdigest()
class IqiyiSDK(object):
def __init__(self, target, ip, timestamp):
self.target = target
self.ip = ip
self.timestamp = timestamp
@staticmethod
def split_sum(data):
return compat_str(sum(map(lambda p: int(p, 16), list(data))))
@staticmethod
def digit_sum(num):
if isinstance(num, int):
num = compat_str(num)
return compat_str(sum(map(int, num)))
def even_odd(self):
even = self.digit_sum(compat_str(self.timestamp)[::2])
odd = self.digit_sum(compat_str(self.timestamp)[1::2])
return even, odd
def preprocess(self, chunksize):
self.target = md5_text(self.target)
chunks = []
for i in range(32 // chunksize):
chunks.append(self.target[chunksize * i:chunksize * (i + 1)])
if 32 % chunksize:
chunks.append(self.target[32 - 32 % chunksize:])
return chunks, list(map(int, self.ip.split('.')))
def mod(self, modulus):
chunks, ip = self.preprocess(32)
self.target = chunks[0] + ''.join(map(lambda p: compat_str(p % modulus), ip))
def split(self, chunksize):
modulus_map = {
4: 256,
5: 10,
8: 100,
}
chunks, ip = self.preprocess(chunksize)
ret = ''
for i in range(len(chunks)):
ip_part = compat_str(ip[i] % modulus_map[chunksize]) if i < 4 else ''
if chunksize == 8:
ret += ip_part + chunks[i]
else:
ret += chunks[i] + ip_part
self.target = ret
def handle_input16(self):
self.target = md5_text(self.target)
self.target = self.split_sum(self.target[:16]) + self.target + self.split_sum(self.target[16:])
def handle_input8(self):
self.target = md5_text(self.target)
ret = ''
for i in range(4):
part = self.target[8 * i:8 * (i + 1)]
ret += self.split_sum(part) + part
self.target = ret
def handleSum(self):
self.target = md5_text(self.target)
self.target = self.split_sum(self.target) + self.target
def date(self, scheme):
self.target = md5_text(self.target)
d = time.localtime(self.timestamp)
strings = {
'y': compat_str(d.tm_year),
'm': '%02d' % d.tm_mon,
'd': '%02d' % d.tm_mday,
}
self.target += ''.join(map(lambda c: strings[c], list(scheme)))
def split_time_even_odd(self):
even, odd = self.even_odd()
self.target = odd + md5_text(self.target) + even
def split_time_odd_even(self):
even, odd = self.even_odd()
self.target = even + md5_text(self.target) + odd
def split_ip_time_sum(self):
chunks, ip = self.preprocess(32)
self.target = compat_str(sum(ip)) + chunks[0] + self.digit_sum(self.timestamp)
def split_time_ip_sum(self):
chunks, ip = self.preprocess(32)
self.target = self.digit_sum(self.timestamp) + chunks[0] + compat_str(sum(ip))
class IqiyiSDKInterpreter(object):
def __init__(self, sdk_code):
self.sdk_code = sdk_code
def run(self, target, ip, timestamp):
self.sdk_code = decode_packed_codes(self.sdk_code)
functions = re.findall(r'input=([a-zA-Z0-9]+)\(input', self.sdk_code)
sdk = IqiyiSDK(target, ip, timestamp)
other_functions = {
'handleSum': sdk.handleSum,
'handleInput8': sdk.handle_input8,
'handleInput16': sdk.handle_input16,
'splitTimeEvenOdd': sdk.split_time_even_odd,
'splitTimeOddEven': sdk.split_time_odd_even,
'splitIpTimeSum': sdk.split_ip_time_sum,
'splitTimeIpSum': sdk.split_time_ip_sum,
}
for function in functions:
if re.match(r'mod\d+', function):
sdk.mod(int(function[3:]))
elif re.match(r'date[ymd]{3}', function):
sdk.date(function[4:])
elif re.match(r'split\d+', function):
sdk.split(int(function[5:]))
elif function in other_functions:
other_functions[function]()
else:
raise ExtractorError('Unknown funcion %s' % function)
return sdk.target
class IqiyiIE(InfoExtractor):
IE_NAME = 'iqiyi'
IE_DESC = '爱奇艺'
_VALID_URL = r'https?://(?:(?:[^.]+\.)?iqiyi\.com|www\.pps\.tv)/.+\.html'
_NETRC_MACHINE = 'iqiyi'
_TESTS = [{
'url': 'http://www.iqiyi.com/v_19rrojlavg.html',
'md5': '470a6c160618577166db1a7aac5a3606',
'info_dict': {
'id': '9c1fb1b99d192b21c559e5a1a2cb3c73',
'ext': 'mp4',
'title': '美国德州空中惊现奇异云团 酷似UFO',
}
}, {
'url': 'http://www.iqiyi.com/v_19rrhnnclk.html',
'md5': 'f09f0a6a59b2da66a26bf4eda669a4cc',
'info_dict': {
'id': 'e3f585b550a280af23c98b6cb2be19fb',
'ext': 'mp4',
'title': '名侦探柯南 国语版',
},
'skip': 'Geo-restricted to China',
}, {
'url': 'http://www.iqiyi.com/w_19rt6o8t9p.html',
'only_matching': True,
}, {
'url': 'http://www.iqiyi.com/a_19rrhbc6kt.html',
'only_matching': True,
}, {
'url': 'http://yule.iqiyi.com/pcb.html',
'only_matching': True,
}, {
# VIP-only video. The first 2 parts (6 minutes) are available without login
# MD5 sums omitted as values are different on Travis CI and my machine
'url': 'http://www.iqiyi.com/v_19rrny4w8w.html',
'info_dict': {
'id': 'f3cf468b39dddb30d676f89a91200dc1',
'title': '泰坦尼克号',
},
'playlist': [{
'info_dict': {
'id': 'f3cf468b39dddb30d676f89a91200dc1_part1',
'ext': 'f4v',
'title': '泰坦尼克号',
},
}, {
'info_dict': {
'id': 'f3cf468b39dddb30d676f89a91200dc1_part2',
'ext': 'f4v',
'title': '泰坦尼克号',
},
}],
'expected_warnings': ['Needs a VIP account for full video'],
}, {
'url': 'http://www.iqiyi.com/a_19rrhb8ce1.html',
'info_dict': {
'id': '202918101',
'title': '灌篮高手 国语版',
},
'playlist_count': 101,
}, {
'url': 'http://www.pps.tv/w_19rrbav0ph.html',
'only_matching': True,
}]
_FORMATS_MAP = [
('1', 'h6'),
('2', 'h5'),
('3', 'h4'),
('4', 'h3'),
('5', 'h2'),
('10', 'h1'),
]
def _real_initialize(self):
self._login()
@staticmethod
def _rsa_fun(data):
# public key extracted from http://static.iqiyi.com/js/qiyiV2/20160129180840/jobs/i18n/i18nIndex.js
N = 0xab86b6371b5318aaa1d3c9e612a9f1264f372323c8c0f19875b5fc3b3fd3afcc1e5bec527aa94bfa85bffc157e4245aebda05389a5357b75115ac94f074aefcd
e = 65537
return ohdave_rsa_encrypt(data, e, N)
def _login(self):
(username, password) = self._get_login_info()
# No authentication to be performed
if not username:
return True
data = self._download_json(
'http://kylin.iqiyi.com/get_token', None,
note='Get token for logging', errnote='Unable to get token for logging')
sdk = data['sdk']
timestamp = int(time.time())
target = '/apis/reglogin/login.action?lang=zh_TW&area_code=null&email=%s&passwd=%s&agenttype=1&from=undefined&keeplogin=0&piccode=&fromurl=&_pos=1' % (
username, self._rsa_fun(password.encode('utf-8')))
interp = IqiyiSDKInterpreter(sdk)
sign = interp.run(target, data['ip'], timestamp)
validation_params = {
'target': target,
'server': 'BEA3AA1908656AABCCFF76582C4C6660',
'token': data['token'],
'bird_src': 'f8d91d57af224da7893dd397d52d811a',
'sign': sign,
'bird_t': timestamp,
}
validation_result = self._download_json(
'http://kylin.iqiyi.com/validate?' + compat_urllib_parse_urlencode(validation_params), None,
note='Validate credentials', errnote='Unable to validate credentials')
MSG_MAP = {
'P00107': 'please login via the web interface and enter the CAPTCHA code',
'P00117': 'bad username or password',
}
code = validation_result['code']
if code != 'A00000':
msg = MSG_MAP.get(code)
if not msg:
msg = 'error %s' % code
if validation_result.get('msg'):
msg += ': ' + validation_result['msg']
self._downloader.report_warning('unable to log in: ' + msg)
return False
return True
@staticmethod
def _gen_sc(tvid, timestamp):
M = [1732584193, -271733879]
M.extend([~M[0], ~M[1]])
I_table = [7, 12, 17, 22, 5, 9, 14, 20, 4, 11, 16, 23, 6, 10, 15, 21]
C_base = [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 8388608, 432]
def L(n, t):
if t is None:
t = 0
return trunc(((n >> 1) + (t >> 1) << 1) + (n & 1) + (t & 1))
def trunc(n):
n = n % 0x100000000
if n > 0x7fffffff:
n -= 0x100000000
return n
def transform(string, mod):
num = int(string, 16)
return (num >> 8 * (i % 4) & 255 ^ i % mod) << ((a & 3) << 3)
C = list(C_base)
o = list(M)
k = str(timestamp - 7)
for i in range(13):
a = i
C[a >> 2] |= ord(k[a]) << 8 * (a % 4)
for i in range(16):
a = i + 13
start = (i >> 2) * 8
r = '03967743b643f66763d623d637e30733'
C[a >> 2] |= transform(''.join(reversed(r[start:start + 8])), 7)
for i in range(16):
a = i + 29
start = (i >> 2) * 8
r = '7038766939776a32776a32706b337139'
C[a >> 2] |= transform(r[start:start + 8], 1)
for i in range(9):
a = i + 45
if i < len(tvid):
C[a >> 2] |= ord(tvid[i]) << 8 * (a % 4)
for a in range(64):
i = a
I = i >> 4
C_index = [i, 5 * i + 1, 3 * i + 5, 7 * i][I] % 16 + urshift(a, 6)
m = L(L(o[0], [
trunc(o[1] & o[2]) | trunc(~o[1] & o[3]),
trunc(o[3] & o[1]) | trunc(~o[3] & o[2]),
o[1] ^ o[2] ^ o[3],
o[2] ^ trunc(o[1] | ~o[3])
][I]), L(
trunc(int(abs(math.sin(i + 1)) * 4294967296)),
C[C_index] if C_index < len(C) else None))
I = I_table[4 * I + i % 4]
o = [o[3],
L(o[1], trunc(trunc(m << I) | urshift(m, 32 - I))),
o[1],
o[2]]
new_M = [L(o[0], M[0]), L(o[1], M[1]), L(o[2], M[2]), L(o[3], M[3])]
s = [new_M[a >> 3] >> (1 ^ a & 7) * 4 & 15 for a in range(32)]
return binascii.hexlify(intlist_to_bytes(s))[1::2].decode('ascii')
def get_raw_data(self, tvid, video_id):
tm = int(time.time() * 1000)
sc = self._gen_sc(tvid, tm)
params = {
'platForm': 'h5',
'rate': 1,
'tvid': tvid,
'vid': video_id,
'cupid': 'qc_100001_100186',
'type': 'mp4',
'nolimit': 0,
'agenttype': 13,
'src': 'd846d0c32d664d32b6b54ea48997a589',
'sc': sc,
't': tm - 7,
'__jsT': None,
}
headers = {}
cn_verification_proxy = self._downloader.params.get('cn_verification_proxy')
if cn_verification_proxy:
headers['Ytdl-request-proxy'] = cn_verification_proxy
return self._download_json(
'http://cache.m.iqiyi.com/jp/tmts/%s/%s/' % (tvid, video_id),
video_id, transform_source=lambda s: remove_start(s, 'var tvInfoJs='),
query=params, headers=headers)
def _extract_playlist(self, webpage):
PAGE_SIZE = 50
links = re.findall(
r'<a[^>]+class="site-piclist_pic_link"[^>]+href="(http://www\.iqiyi\.com/.+\.html)"',
webpage)
if not links:
return
album_id = self._search_regex(
r'albumId\s*:\s*(\d+),', webpage, 'album ID')
album_title = self._search_regex(
r'data-share-title="([^"]+)"', webpage, 'album title', fatal=False)
entries = list(map(self.url_result, links))
# Start from 2 because links in the first page are already on webpage
for page_num in itertools.count(2):
pagelist_page = self._download_webpage(
'http://cache.video.qiyi.com/jp/avlist/%s/%d/%d/' % (album_id, page_num, PAGE_SIZE),
album_id,
note='Download playlist page %d' % page_num,
errnote='Failed to download playlist page %d' % page_num)
pagelist = self._parse_json(
remove_start(pagelist_page, 'var tvInfoJs='), album_id)
vlist = pagelist['data']['vlist']
for item in vlist:
entries.append(self.url_result(item['vurl']))
if len(vlist) < PAGE_SIZE:
break
return self.playlist_result(entries, album_id, album_title)
def _real_extract(self, url):
webpage = self._download_webpage(
url, 'temp_id', note='download video page')
# There's no simple way to determine whether an URL is a playlist or not
# So detect it
playlist_result = self._extract_playlist(webpage)
if playlist_result:
return playlist_result
tvid = self._search_regex(
r'data-player-tvid\s*=\s*[\'"](\d+)', webpage, 'tvid')
video_id = self._search_regex(
r'data-player-videoid\s*=\s*[\'"]([a-f\d]+)', webpage, 'video_id')
for _ in range(5):
raw_data = self.get_raw_data(tvid, video_id)
if raw_data['code'] != 'A00000':
if raw_data['code'] == 'A00111':
self.raise_geo_restricted()
raise ExtractorError('Unable to load data. Error code: ' + raw_data['code'])
data = raw_data['data']
# iQiYi sometimes returns Ads
if not isinstance(data['playInfo'], dict):
self._sleep(5, video_id)
continue
title = data['playInfo']['an']
break
return {
'id': video_id,
'title': title,
'url': data['m3u'],
}