eric/gitlab-instance-0a899031_douyinyoutube
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

's'

Browse Source
This commit is contained in:
eric
2025-12-01 15:04:45 +08:00
parent 6d7f3b874f
commit 962b17ea67
40 changed files with 2443 additions and 13052 deletions
Download Patch File Download Diff File Expand all files Collapse all files

210
src/spider.py
View File

@@ -4,7 +4,7 @@
Author: Hmily
GitHub: https://github.com/ihmily
Date: 2023-07-15 23:15:00
Update: 2025-07-19 17:43:00
Update: 2025-10-23 18:28:00
Copyright (c) 2023-2025 by Hmily, All Rights Reserved.
Function: Get live stream data.
"""
@@ -29,6 +29,7 @@ from .utils import trace_error_decorator, generate_random_string
from .logger import script_path
from .room import get_sec_user_id, get_unique_id, UnsupportedUrlError
from .http_clients.async_http import async_req
from .ab_sign import ab_sign
ssl_context = ssl.create_default_context()
@@ -64,10 +65,87 @@ async def get_play_url_list(m3u8: str, proxy: OptionalStr = None, header: Option
return play_url_list
async def get_douyin_web_stream_data(url: str, proxy_addr: OptionalStr = None, cookies: OptionalStr = None):
headers = {
'cookie': 'ttwid=1%7C2iDIYVmjzMcpZ20fcaFde0VghXAA3NaNXE_SLR68IyE%7C1761045455'
'%7Cab35197d5cfb21df6cbb2fa7ef1c9262206b062c315b9d04da746d0b37dfbc7d',
'referer': 'https://live.douyin.com/335354047186',
'user-agent': 'Mozilla/5.0 (Windows NT 10.0; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) '
'Chrome/116.0.5845.97 Safari/537.36 Core/1.116.567.400 QQBrowser/19.7.6764.400',
}
if cookies:
headers['cookie'] = cookies
try:
web_rid = url.split('?')[0].split('live.douyin.com/')[-1]
params = {
"aid": "6383",
"app_name": "douyin_web",
"live_id": "1",
"device_platform": "web",
"language": "zh-CN",
"browser_language": "zh-CN",
"browser_platform": "Win32",
"browser_name": "Chrome",
"browser_version": "116.0.0.0",
"web_rid": web_rid,
'msToken': '',
}
api = f'https://live.douyin.com/webcast/room/web/enter/?{urllib.parse.urlencode(params)}'
a_bogus = ab_sign(urllib.parse.urlparse(api).query, headers['user-agent'])
api += "&a_bogus=" + a_bogus
try:
json_str = await async_req(url=api, proxy_addr=proxy_addr, headers=headers)
if not json_str:
raise Exception("it triggered risk control")
json_data = json.loads(json_str)['data']
if not json_data['data']:
raise Exception(f"{url} VR live is not supported")
room_data = json_data['data'][0]
room_data['anchor_name'] = json_data['user']['nickname']
except Exception as e:
raise Exception(f"Douyin web data fetch error, because {e}.")
if room_data['status'] == 2:
if 'stream_url' not in room_data:
raise RuntimeError(
"The live streaming type or gameplay is not supported on the computer side yet, please use the "
"app to share the link for recording."
)
live_core_sdk_data = room_data['stream_url']['live_core_sdk_data']
pull_datas = room_data['stream_url']['pull_datas']
if live_core_sdk_data:
if pull_datas:
key = list(pull_datas.keys())[0]
json_str = pull_datas[key]['stream_data']
else:
json_str = live_core_sdk_data['pull_data']['stream_data']
json_data = json.loads(json_str)
if 'origin' in json_data['data']:
stream_data = live_core_sdk_data['pull_data']['stream_data']
origin_data = json.loads(stream_data)['data']['origin']['main']
sdk_params = json.loads(origin_data['sdk_params'])
origin_hls_codec = sdk_params.get('VCodec') or ''
origin_url_list = json_data['data']['origin']['main']
origin_m3u8 = {'ORIGIN': origin_url_list["hls"] + '&codec=' + origin_hls_codec}
origin_flv = {'ORIGIN': origin_url_list["flv"] + '&codec=' + origin_hls_codec}
hls_pull_url_map = room_data['stream_url']['hls_pull_url_map']
flv_pull_url = room_data['stream_url']['flv_pull_url']
room_data['stream_url']['hls_pull_url_map'] = {**origin_m3u8, **hls_pull_url_map}
room_data['stream_url']['flv_pull_url'] = {**origin_flv, **flv_pull_url}
except Exception as e:
print(f"Error message: {e} Error line: {e.__traceback__.tb_lineno}")
room_data = {'anchor_name': ""}
return room_data
@trace_error_decorator
async def get_douyin_app_stream_data(url: str, proxy_addr: OptionalStr = None, cookies: OptionalStr = None) -> dict:
headers = {
'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:109.0) Gecko/20100101 Firefox/115.0',
'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) '
'Chrome/141.0.0.0 Safari/537.36 Edg/141.0.0.0',
'Accept-Language': 'zh-CN,zh;q=0.8,zh-TW;q=0.7,zh-HK;q=0.5,en-US;q=0.3,en;q=0.2',
'Referer': 'https://live.douyin.com/',
'Cookie': 'ttwid=1%7CB1qls3GdnZhUov9o2NxOMxxYS2ff6OSvEWbv0ytbES4%7C1680522049%7C280d802d6d478e3e78d0c807f7c487e7ffec0ae4e5fdd6a0fe74c3c6af149511; my_rd=1; passport_csrf_token=3ab34460fa656183fccfb904b16ff742; passport_csrf_token_default=3ab34460fa656183fccfb904b16ff742; d_ticket=9f562383ac0547d0b561904513229d76c9c21; n_mh=hvnJEQ4Q5eiH74-84kTFUyv4VK8xtSrpRZG1AhCeFNI; store-region=cn-fj; store-region-src=uid; LOGIN_STATUS=1; __security_server_data_status=1; FORCE_LOGIN=%7B%22videoConsumedRemainSeconds%22%3A180%7D; pwa2=%223%7C0%7C3%7C0%22; download_guide=%223%2F20230729%2F0%22; volume_info=%7B%22isUserMute%22%3Afalse%2C%22isMute%22%3Afalse%2C%22volume%22%3A0.6%7D; strategyABtestKey=%221690824679.923%22; stream_recommend_feed_params=%22%7B%5C%22cookie_enabled%5C%22%3Atrue%2C%5C%22screen_width%5C%22%3A1536%2C%5C%22screen_height%5C%22%3A864%2C%5C%22browser_online%5C%22%3Atrue%2C%5C%22cpu_core_num%5C%22%3A8%2C%5C%22device_memory%5C%22%3A8%2C%5C%22downlink%5C%22%3A10%2C%5C%22effective_type%5C%22%3A%5C%224g%5C%22%2C%5C%22round_trip_time%5C%22%3A150%7D%22; VIDEO_FILTER_MEMO_SELECT=%7B%22expireTime%22%3A1691443863751%2C%22type%22%3Anull%7D; home_can_add_dy_2_desktop=%221%22; __live_version__=%221.1.1.2169%22; device_web_cpu_core=8; device_web_memory_size=8; xgplayer_user_id=346045893336; csrf_session_id=2e00356b5cd8544d17a0e66484946f28; odin_tt=724eb4dd23bc6ffaed9a1571ac4c757ef597768a70c75fef695b95845b7ffcd8b1524278c2ac31c2587996d058e03414595f0a4e856c53bd0d5e5f56dc6d82e24004dc77773e6b83ced6f80f1bb70627; __ac_nonce=064caded4009deafd8b89; __ac_signature=_02B4Z6wo00f01HLUuwwAAIDBh6tRkVLvBQBy9L-AAHiHf7; ttcid=2e9619ebbb8449eaa3d5a42d8ce88ec835; webcast_leading_last_show_time=1691016922379; webcast_leading_total_show_times=1; webcast_local_quality=sd; live_can_add_dy_2_desktop=%221%22; msToken=1JDHnVPw_9yTvzIrwb7cQj8dCMNOoesXbA_IooV8cezcOdpe4pzusZE7NB7tZn9TBXPr0ylxmv-KMs5rqbNUBHP4P7VBFUu0ZAht_BEylqrLpzgt3y5ne_38hXDOX8o=; msToken=jV_yeN1IQKUd9PlNtpL7k5vthGKcHo0dEh_QPUQhr8G3cuYv-Jbb4NnIxGDmhVOkZOCSihNpA2kvYtHiTW25XNNX_yrsv5FN8O6zm3qmCIXcEe0LywLn7oBO2gITEeg=; tt_scid=mYfqpfbDjqXrIGJuQ7q-DlQJfUSG51qG.KUdzztuGP83OjuVLXnQHjsz-BRHRJu4e986'
@@ -77,7 +155,7 @@ async def get_douyin_app_stream_data(url: str, proxy_addr: OptionalStr = None, c
async def get_app_data(room_id: str, sec_uid: str) -> dict:
app_params = {
"verifyFp": "verify_lxj5zv70_7szNlAB7_pxNY_48Vh_ALKF_GA1Uf3yteoOY",
"verifyFp": "verify_hwj52020_7szNlAB7_pxNY_48Vh_ALKF_GA1Uf3yteoOY",
"type_id": "0",
"live_id": "1",
"room_id": room_id,
@@ -86,36 +164,25 @@ async def get_douyin_app_stream_data(url: str, proxy_addr: OptionalStr = None, c
"app_id": "1128"
}
api2 = f'https://webcast.amemv.com/webcast/room/reflow/info/?{urllib.parse.urlencode(app_params)}'
json_str2 = await async_req(url=api2, proxy_addr=proxy_addr, headers=headers)
json_data2 = json.loads(json_str2)['data']
room_data2 = json_data2['room']
room_data2['anchor_name'] = room_data2['owner']['nickname']
return room_data2
a_bogus = ab_sign(urllib.parse.urlparse(api2).query, headers['User-Agent'])
api2 += "&a_bogus=" + a_bogus
try:
json_str2 = await async_req(url=api2, proxy_addr=proxy_addr, headers=headers)
if not json_str2:
raise Exception("it triggered risk control")
json_data2 = json.loads(json_str2)['data']
if not json_data2.get('room'):
raise Exception(f"{url} VR live is not supported")
room_data2 = json_data2['room']
room_data2['anchor_name'] = room_data2['owner']['nickname']
return room_data2
except Exception as e:
raise Exception(f"Douyin app data fetch error, because {e}.")
try:
web_rid = url.split('?')[0].split('live.douyin.com/')
if len(web_rid) > 1:
web_rid = web_rid[1]
params = {
"aid": "6383",
"app_name": "douyin_web",
"live_id": "1",
"device_platform": "web",
"language": "zh-CN",
"browser_language": "zh-CN",
"browser_platform": "Win32",
"browser_name": "Chrome",
"browser_version": "116.0.0.0",
"web_rid": web_rid,
'msToken': '',
'a_bogus': ''
}
api = f'https://live.douyin.com/webcast/room/web/enter/?{urllib.parse.urlencode(params)}'
json_str = await async_req(url=api, proxy_addr=proxy_addr, headers=headers)
json_data = json.loads(json_str)['data']
room_data = json_data['data'][0]
room_data['anchor_name'] = json_data['user']['nickname']
return await get_douyin_web_stream_data(url, proxy_addr, cookies)
else:
try:
data = await get_sec_user_id(url, proxy_addr=proxy_addr)
@@ -218,18 +285,20 @@ async def get_douyin_stream_data(url: str, proxy_addr: OptionalStr = None, cooki
@trace_error_decorator
async def get_tiktok_stream_data(url: str, proxy_addr: OptionalStr = None, cookies: OptionalStr = None) -> dict | None:
headers = {
'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:124.0) Gecko/20100101 Firefox/124.0',
'Cookie': 'ttwid=1%7CM-rF193sJugKuNz2RGNt-rh6pAAR9IMceUSzlDnPCNI%7C1683274418%7Cf726d4947f2fc37fecc7aeb0cdaee52892244d04efde6f8a8edd2bb168263269; tiktok_webapp_theme=light; tt_chain_token=VWkygAWDlm1cFg/k8whmOg==; passport_csrf_token=6e422c5a7991f8cec7033a8082921510; passport_csrf_token_default=6e422c5a7991f8cec7033a8082921510; d_ticket=f8c267d4af4523c97be1ccb355e9991e2ae06; odin_tt=320b5f386cdc23f347be018e588873db7f7aea4ea5d1813681c3fbc018ea025dde957b94f74146dbc0e3612426b865ccb95ec8abe4ee36cca65f15dbffec0deff7b0e69e8ea536d46e0f82a4fc37d211; cmpl_token=AgQQAPNSF-RO0rT04baWtZ0T_jUjl4fVP4PZYM2QPw; uid_tt=319b558dbba684bb1557206c92089cd113a875526a89aee30595925d804b81c7; uid_tt_ss=319b558dbba684bb1557206c92089cd113a875526a89aee30595925d804b81c7; sid_tt=ad5e736f4bedb2f6d42ccd849e706b1d; sessionid=ad5e736f4bedb2f6d42ccd849e706b1d; sessionid_ss=ad5e736f4bedb2f6d42ccd849e706b1d; store-idc=useast5; store-country-code=us; store-country-code-src=uid; tt-target-idc=useast5; tt-target-idc-sign=qXNk0bb1pDQ0FbCNF120Pl9WWMLZg9Edv5PkfyCbS4lIk5ieW5tfLP7XWROnN0mEaSlc5hg6Oji1pF-yz_3ZXnUiNMrA9wNMPvI6D9IFKKVmq555aQzwPIGHv0aQC5dNRgKo5Z5LBkgxUMWEojTKclq2_L8lBciw0IGdhFm_XyVJtbqbBKKgybGDLzK8ZyxF4Jl_cYRXaDlshZjc38JdS6wruDueRSHe7YvNbjxCnApEFUv-OwJANSPU_4rvcqpVhq3JI2VCCfw-cs_4MFIPCDOKisk5EhAo2JlHh3VF7_CLuv80FXg_7ZqQ2pJeMOog294rqxwbbQhl3ATvjQV_JsWyUsMd9zwqecpylrPvtySI2u1qfoggx1owLrrUynee1R48QlanLQnTNW_z1WpmZBgVJqgEGLwFoVOmRzJuFFNj8vIqdjM2nDSdWqX8_wX3wplohkzkPSFPfZgjzGnQX28krhgTytLt7BXYty5dpfGtsdb11WOFHM6MZ9R9uLVB; sid_guard=ad5e736f4bedb2f6d42ccd849e706b1d%7C1690990657%7C15525213%7CMon%2C+29-Jan-2024+08%3A11%3A10+GMT; sid_ucp_v1=1.0.0-KGM3YzgwYjZhODgyYWI1NjIwNTA0NjBmOWUxMGRhMjIzYTI2YjMxNDUKGAiqiJ30keKD5WQQwfCppgYYsws4AkDsBxAEGgd1c2Vhc3Q1IiBhZDVlNzM2ZjRiZWRiMmY2ZDQyY2NkODQ5ZTcwNmIxZA; ssid_ucp_v1=1.0.0-KGM3YzgwYjZhODgyYWI1NjIwNTA0NjBmOWUxMGRhMjIzYTI2YjMxNDUKGAiqiJ30keKD5WQQwfCppgYYsws4AkDsBxAEGgd1c2Vhc3Q1IiBhZDVlNzM2ZjRiZWRiMmY2ZDQyY2NkODQ5ZTcwNmIxZA; tt_csrf_token=dD0EIH8q-pe3qDQsCyyD1jLN6KizJDRjOEyk; __tea_cache_tokens_1988={%22_type_%22:%22default%22%2C%22user_unique_id%22:%227229608516049831425%22%2C%22timestamp%22:1683274422659}; ttwid=1%7CM-rF193sJugKuNz2RGNt-rh6pAAR9IMceUSzlDnPCNI%7C1694002151%7Cd89b77afc809b1a610661a9d1c2784d80ebef9efdd166f06de0d28e27f7e4efe; msToken=KfJAVZ7r9D_QVeQlYAUZzDFbc1Yx-nZz6GF33eOxgd8KlqvTg1lF9bMXW7gFV-qW4MCgUwnBIhbiwU9kdaSpgHJCk-PABsHCtTO5J3qC4oCTsrXQ1_E0XtbqiE4OVLZ_jdF1EYWgKNPT2SnwGkQ=; msToken=KfJAVZ7r9D_QVeQlYAUZzDFbc1Yx-nZz6GF33eOxgd8KlqvTg1lF9bMXW7gFV-qW4MCgUwnBIhbiwU9kdaSpgHJCk-PABsHCtTO5J3qC4oCTsrXQ1_E0XtbqiE4OVLZ_jdF1EYWgKNPT2SnwGkQ='
'referer': 'https://www.tiktok.com/',
'user-agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) '
'Chrome/141.0.0.0 Safari/537.36',
'cookie': cookies or '1%7Cz7FKki38aKyy7i-BC9rEDwcrVvjcLcFEL6QIeqldoy4%7C1761302831%7C6c1461e9f1f980cbe0404c5190'
'5177d5d53bbd822e1bf66128887d942c9c3e2f'
}
if cookies:
headers['Cookie'] = cookies
for i in range(3):
html_str = await async_req(url=url, proxy_addr=proxy_addr, headers=headers, abroad=True, http2=False)
time.sleep(1)
if "We regret to inform you that we have discontinued operating TikTok" in html_str:
msg = re.search('<p>\n\\s+(We regret to inform you that we have discontinu.*?)\\.\n\\s+</p>', html_str)
raise ConnectionError(
f"Your proxy node's regional network is blocked from accessing TikTok; please switch to a node in "
"Your proxy node's regional network is blocked from accessing TikTok; please switch to a node in "
f"another region to access. {msg.group(1) if msg else ''}"
)
if 'UNEXPECTED_EOF_WHILE_READING' not in html_str:
@@ -935,6 +1004,76 @@ async def get_sooplive_tk(url: str, rtype: str, proxy_addr: OptionalStr = None,
return f"{bj_name}-{bj_id}", json_data['CHANNEL']['BNO']
def get_soop_headers(cookies):
headers = {
'client-id': str(uuid.uuid4()),
'user-agent': 'Mozilla/5.0 (iPhone; CPU iPhone OS 18_5 like Mac OS X) AppleWebKit/605.1.15 (KHTML, '
'like Gecko) Version/18.5 Mobile/15E148 Safari/604.1 Edg/141.0.0.0',
}
if cookies:
headers['cookie'] = cookies
return headers
async def _get_soop_channel_info_global(bj_id, proxy_addr: OptionalStr = None, cookies: OptionalStr = None) -> str:
headers = get_soop_headers(cookies)
api = 'https://api.sooplive.com/v2/channel/info/' + str(bj_id)
json_str = await async_req(api, proxy_addr=proxy_addr, headers=headers)
json_data = json.loads(json_str)
nickname = json_data['data']['streamerChannelInfo']['nickname']
channelId = json_data['data']['streamerChannelInfo']['channelId']
anchor_name = f"{nickname}-{channelId}"
return anchor_name
async def _get_soop_stream_info_global(bj_id, proxy_addr: OptionalStr = None, cookies: OptionalStr = None) -> tuple:
headers = get_soop_headers(cookies)
api = 'https://api.sooplive.com/v2/stream/info/' + str(bj_id)
json_str = await async_req(api, proxy_addr=proxy_addr, headers=headers)
json_data = json.loads(json_str)
status = json_data['data']['isStream']
title = json_data['data']['title']
return status, title
async def _fetch_web_stream_data_global(url: str, proxy_addr: OptionalStr = None, cookies: OptionalStr = None) -> dict:
split_url = url.split('/')
bj_id = split_url[3] if len(split_url) < 6 else split_url[5]
anchor_name = await _get_soop_channel_info_global(bj_id)
result = {"anchor_name": anchor_name or '', "is_live": False, "live_url": url}
status, title = await _get_soop_stream_info_global(bj_id)
if not status:
return result
else:
async def _get_url_list(m3u8: str) -> list[str]:
headers = {
'user-agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) '
'Chrome/141.0.0.0 Safari/537.36 Edg/141.0.0.0',
}
if cookies:
headers['cookie'] = cookies
resp = await async_req(url=m3u8, proxy_addr=proxy_addr, headers=headers)
play_url_list = []
url_prefix = '/'.join(m3u8.split('/')[0:3])
for i in resp.split('\n'):
if not i.startswith('#') and i.strip():
play_url_list.append(url_prefix + i.strip())
bandwidth_pattern = re.compile(r'BANDWIDTH=(\d+)')
bandwidth_list = bandwidth_pattern.findall(resp)
url_to_bandwidth = {purl: int(bandwidth) for bandwidth, purl in zip(bandwidth_list, play_url_list)}
play_url_list = sorted(play_url_list, key=lambda purl: url_to_bandwidth[purl], reverse=True)
return play_url_list
m3u8_url = 'https://global-media.sooplive.com/live/' + str(bj_id) + '/master.m3u8'
result |= {
'is_live': True,
'title': title,
'm3u8_url': m3u8_url,
'play_url_list': await _get_url_list(m3u8_url)
}
return result
@trace_error_decorator
async def get_sooplive_stream_data(
url: str, proxy_addr: OptionalStr = None, cookies: OptionalStr = None,
@@ -949,6 +1088,9 @@ async def get_sooplive_stream_data(
if cookies:
headers['Cookie'] = cookies
if "sooplive.com" in url:
return await _fetch_web_stream_data_global(url, proxy_addr, cookies)
split_url = url.split('/')
bj_id = split_url[3] if len(split_url) < 6 else split_url[5]