0
0
Fork 0
mirror of https://github.com/yt-dlp/yt-dlp.git synced 2024-11-27 03:03:01 +00:00
yt-dlp/yt_dlp/extractor/stacommu.py
2023-06-30 18:27:07 +00:00

148 lines
5.6 KiB
Python
Raw Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

import time
from .wrestleuniverse import WrestleUniverseBaseIE
from ..utils import (
int_or_none,
traverse_obj,
url_or_none,
)
class StacommuBaseIE(WrestleUniverseBaseIE):
_NETRC_MACHINE = 'stacommu'
_API_HOST = 'api.stacommu.jp'
_LOGIN_QUERY = {'key': 'AIzaSyCR9czxhH2eWuijEhTNWBZ5MCcOYEUTAhg'}
_LOGIN_HEADERS = {
'Accept': '*/*',
'Content-Type': 'application/json',
'X-Client-Version': 'Chrome/JsCore/9.9.4/FirebaseCore-web',
'Referer': 'https://www.stacommu.jp/',
'Origin': 'https://www.stacommu.jp',
}
@WrestleUniverseBaseIE._TOKEN.getter
def _TOKEN(self):
if self._REAL_TOKEN and self._TOKEN_EXPIRY <= int(time.time()):
self._refresh_token()
return self._REAL_TOKEN
def _get_formats(self, data, path, video_id=None):
if not traverse_obj(data, path) and not data.get('canWatch') and not self._TOKEN:
self.raise_login_required(method='password')
return super()._get_formats(data, path, video_id)
def _extract_hls_key(self, data, path, decrypt):
encryption_data = traverse_obj(data, path)
if traverse_obj(encryption_data, ('encryptType', {int})) == 0:
return None
return traverse_obj(encryption_data, {'key': ('key', {decrypt}), 'iv': ('iv', {decrypt})})
class StacommuVODIE(StacommuBaseIE):
_VALID_URL = r'https?://www\.stacommu\.jp/videos/episodes/(?P<id>[\da-zA-Z]+)'
_TESTS = [{
# not encrypted
'url': 'https://www.stacommu.jp/videos/episodes/aXcVKjHyAENEjard61soZZ',
'info_dict': {
'id': 'aXcVKjHyAENEjard61soZZ',
'ext': 'mp4',
'title': 'スタコミュAWARDの裏側、ほぼ全部見せます〜晴れ舞台の直前ドキドキ編〜',
'description': 'md5:6400275c57ae75c06da36b06f96beb1c',
'timestamp': 1679652000,
'upload_date': '20230324',
'thumbnail': 'https://image.stacommu.jp/6eLobQan8PFtBoU4RL4uGg/6eLobQan8PFtBoU4RL4uGg',
'cast': 'count:11',
'duration': 250,
},
'params': {
'skip_download': 'm3u8',
},
}, {
# encrypted; requires a premium account
'url': 'https://www.stacommu.jp/videos/episodes/3hybMByUvzMEqndSeu5LpD',
'info_dict': {
'id': '3hybMByUvzMEqndSeu5LpD',
'ext': 'mp4',
'title': 'スタプラフェス2023〜裏側ほぼ全部見せます〜10',
'description': 'md5:85494488ccf1dfa1934accdeadd7b340',
'timestamp': 1682506800,
'upload_date': '20230426',
'thumbnail': 'https://image.stacommu.jp/eMdXtEefR4kEyJJMpAFi7x/eMdXtEefR4kEyJJMpAFi7x',
'cast': 'count:55',
'duration': 312,
'hls_aes': {
'key': '6bbaf241b8e1fd9f59ecf546a70e4ae7',
'iv': '1fc9002a23166c3bb1d240b953d09de9',
},
},
'params': {
'skip_download': 'm3u8',
},
}]
_API_PATH = 'videoEpisodes'
def _real_extract(self, url):
video_id = self._match_id(url)
video_info = self._download_metadata(
url, video_id, 'ja', ('dehydratedState', 'queries', 0, 'state', 'data'))
hls_info, decrypt = self._call_encrypted_api(
video_id, ':watch', 'stream information', data={'method': 1})
return {
'id': video_id,
'formats': self._get_formats(hls_info, ('protocolHls', 'url', {url_or_none}), video_id),
'hls_aes': self._extract_hls_key(hls_info, 'protocolHls', decrypt),
**traverse_obj(video_info, {
'title': ('displayName', {str}),
'description': ('description', {str}),
'timestamp': ('watchStartTime', {int_or_none}),
'thumbnail': ('keyVisualUrl', {url_or_none}),
'cast': ('casts', ..., 'displayName', {str}),
'duration': ('duration', {int}),
}),
}
class StacommuLiveIE(StacommuBaseIE):
_VALID_URL = r'https?://www\.stacommu\.jp/live/(?P<id>[\da-zA-Z]+)'
_TESTS = [{
'url': 'https://www.stacommu.jp/live/d2FJ3zLnndegZJCAEzGM3m',
'info_dict': {
'id': 'd2FJ3zLnndegZJCAEzGM3m',
'ext': 'mp4',
'title': '仲村悠菜 2023/05/04',
'timestamp': 1683195647,
'upload_date': '20230504',
'thumbnail': 'https://image.stacommu.jp/pHGF57SPEHE2ke83FS92FN/pHGF57SPEHE2ke83FS92FN',
'duration': 5322,
'hls_aes': {
'key': 'efbb3ec0b8246f61adf1764c5a51213a',
'iv': '80621d19a1f19167b64cedb415b05d1c',
},
},
'params': {
'skip_download': 'm3u8',
},
}]
_API_PATH = 'events'
def _real_extract(self, url):
video_id = self._match_id(url)
video_info = self._call_api(video_id, msg='video information', query={'al': 'ja'}, auth=False)
hls_info, decrypt = self._call_encrypted_api(
video_id, ':watchArchive', 'stream information', data={'method': 1})
return {
'id': video_id,
'formats': self._get_formats(hls_info, ('hls', 'urls', ..., {url_or_none}), video_id),
'hls_aes': self._extract_hls_key(hls_info, 'hls', decrypt),
**traverse_obj(video_info, {
'title': ('displayName', {str}),
'timestamp': ('startTime', {int_or_none}),
'thumbnail': ('keyVisualUrl', {url_or_none}),
'duration': ('duration', {int_or_none}),
}),
}