4 from .common
import InfoExtractor
15 class YandexVideoIE(InfoExtractor
):
19 yandex\.ru(?:/(?:portal/(?:video|efir)|efir))?/?\?.*?stream_id=|
20 frontend\.vh\.yandex\.ru/player/
22 (?P<id>(?:[\da-f]{32}|[\w-]{12}))
25 'url': 'https://yandex.ru/portal/video?stream_id=4dbb36ec4e0526d58f9f2dc8f0ecf374',
26 'md5': 'e02a05bfaf0d9615ef07ae3a10f4faf4',
28 'id': '4dbb36ec4e0526d58f9f2dc8f0ecf374',
30 'title': 'Русский Вудсток - главный рок-фест в истории СССР / вДудь',
31 'description': 'md5:7d6b8d4bc4a3b9a56499916c1ea5b5fa',
32 'thumbnail': r
're:^https?://',
33 'timestamp': 1549972939,
36 'upload_date': '20190212',
42 'url': 'https://yandex.ru/portal/efir?stream_id=4dbb262b4fe5cf15a215de4f34eee34d&from=morda',
43 'only_matching': True,
45 'url': 'https://yandex.ru/?stream_id=4dbb262b4fe5cf15a215de4f34eee34d',
46 'only_matching': True,
48 'url': 'https://frontend.vh.yandex.ru/player/4dbb262b4fe5cf15a215de4f34eee34d?from=morda',
49 'only_matching': True,
51 # vod-episode, series episode
52 'url': 'https://yandex.ru/portal/video?stream_id=45b11db6e4b68797919c93751a938cee',
53 'only_matching': True,
56 'url': 'https://yandex.ru/?stream_channel=1538487871&stream_id=4132a07f71fb0396be93d74b3477131d',
57 'only_matching': True,
60 'url': 'https://yandex.ru/portal/video?from=morda&stream_id=485a92d94518d73a9d0ff778e13505f8',
61 'only_matching': True,
63 'url': 'https://yandex.ru/efir?stream_active=watching&stream_id=v7a2dZ-v5mSI&from_block=efir_newtab',
64 'only_matching': True,
67 def _real_extract(self
, url
):
68 video_id
= self
._match
_id
(url
)
70 player
= try_get((self
._download
_json
(
71 'https://frontend.vh.yandex.ru/graphql', video_id
, data
=('''{
72 player(content_id: "%s") {
91 }''' % video_id
).encode(), fatal
=False)), lambda x
: x
['player']['content'])
92 if not player
or player
.get('error'):
93 player
= self
._download
_json
(
94 'https://frontend.vh.yandex.ru/v23/player/%s.json' % video_id
,
96 'stream_options': 'hires',
97 'disable_trackings': 1,
99 content
= player
['content']
101 title
= content
.get('title') or content
['computed_title']
104 streams
= content
.get('streams') or []
105 streams
.append({'url': content.get('content_url')}
)
106 for stream
in streams
:
107 content_url
= url_or_none(stream
.get('url'))
110 ext
= determine_ext(content_url
)
114 formats
.extend(self
._extract
_m
3u8_formats
(
115 content_url
, video_id
, 'mp4',
116 'm3u8_native', m3u8_id
='hls', fatal
=False))
118 formats
.extend(self
._extract
_mpd
_formats
(
119 content_url
, video_id
, mpd_id
='dash', fatal
=False))
121 formats
.append({'url': content_url}
)
123 self
._sort
_formats
(formats
)
125 timestamp
= (int_or_none(content
.get('release_date'))
126 or int_or_none(content
.get('release_date_ut'))
127 or int_or_none(content
.get('start_time')))
128 season
= content
.get('season') or {}
133 'description': content
.get('description'),
134 'thumbnail': content
.get('thumbnail'),
135 'timestamp': timestamp
,
136 'duration': int_or_none(content
.get('duration')),
137 'series': content
.get('program_title'),
138 'age_limit': int_or_none(content
.get('restriction_age')),
139 'view_count': int_or_none(content
.get('views_count')),
140 'like_count': int_or_none(content
.get('likes')),
141 'dislike_count': int_or_none(content
.get('dislikes')),
142 'season_number': int_or_none(season
.get('season_number')),
143 'season_id': season
.get('id'),
144 'release_year': int_or_none(content
.get('release_year')),
149 class YandexVideoPreviewIE(InfoExtractor
):
150 _VALID_URL
= r
'https?://(?:www\.)?yandex\.\w{2,3}(?:\.(?:am|ge|il|tr))?/video/preview(?:/?\?.*?filmId=|/)(?P<id>\d+)'
151 _TESTS
= [{ # Odnoklassniki
152 'url': 'https://yandex.ru/video/preview/?filmId=10682852472978372885&text=summer',
154 'id': '1352565459459',
157 'upload_date': '20191202',
160 'thumbnail': 'https://i.mycdn.me/videoPreview?id=544866765315&type=37&idx=13&tkn=TY5qjLYZHxpmcnK8U2LgzYkgmaU&fn=external_8',
161 'uploader_id': '481054701571',
162 'title': 'LOFT - summer, summer, summer HD',
163 'uploader': 'АРТЁМ КУДРОВ',
166 'url': 'https://yandex.ru/video/preview/?filmId=4479424425337895262&source=main_redirect&text=видео&utm_source=main_stripe_big',
167 'only_matching': True,
169 'url': 'https://yandex.ru/video/preview/5275069442094787341',
170 'only_matching': True,
172 'url': 'https://yandex.ru/video/preview/?filmId=16658118429797832897&from=tabbar&p=1&text=%D0%BF%D1%80%D0%BE%D1%81%D0%BC%D0%BE%D1%82%D1%80+%D1%84%D1%80%D0%B0%D0%B3%D0%BC%D0%B5%D0%BD%D1%82%D0%B0+%D0%BC%D0%B0%D0%BB%D0%B5%D0%BD%D1%8C%D0%BA%D0%B8%D0%B9+%D0%BF%D1%80%D0%B8%D0%BD%D1%86+%D0%BC%D1%8B+%D0%B2+%D0%BE%D1%82%D0%B2%D0%B5%D1%82%D0%B5+%D0%B7%D0%B0+%D1%82%D0%B5%D1%85+%D0%BA%D0%BE%D0%B3%D0%BE+%D0%BF%D1%80%D0%B8%D1%80%D1%83%D1%87%D0%B8%D0%BB%D0%B8',
173 'only_matching': True,
175 'url': 'https://yandex.ru/video/preview/?text=Francis%20Lai%20-%20Le%20Bon%20Et%20Les%20MC)chants&path=wizard&parent-reqid=1643208087979310-1481782809207673478-sas3-0931-2f9-sas-l7-balancer-8080-BAL-9380&wiz_type=vital&filmId=12508152936505397283',
176 'only_matching': True,
178 'url': 'https://yandex.com/video/preview/?text=dossier%2051%20film%201978&path=yandex_search&parent-reqid=1664361087754492-8727541069609384458-sas2-0340-sas-l7-balancer-8080-BAL-8045&noreask=1&from_type=vast&filmId=5794987234584444632',
179 'only_matching': True,
182 def _real_extract(self
, url
):
183 id = self
._match
_id
(url
)
184 webpage
= self
._download
_webpage
(url
, id)
185 data_raw
= self
._search
_regex
(r
'window.Ya.__inline_params__\s*=\s*JSON.parse\(\'([^
"]+?\\u0022video\\u0022:[^"]+?
})\'\
);', webpage, 'data_raw
')
186 data_json = self._parse_json(data_raw, id, transform_source=lowercase_escape)
187 return self.url_result(data_json['video
']['url
'])
190 class ZenYandexIE(InfoExtractor):
191 _VALID_URL = r'https?
://zen\
.yandex\
.ru(?
:/video
)?
/(media|watch
)/(?
:(?
:id/[^
/]+/|
[^
/]+/)(?
:[a
-z0
-9-]+)-)?
(?P
<id>[a
-z0
-9-]+)'
193 'url
': 'https
://zen
.yandex
.ru
/media
/popmech
/izverjenie
-vulkana
-iz
-spichek
-zreliscnyi
-opyt
-6002240ff8b1af50bb2da5e3
',
195 'id': '6002240ff8b1af50bb2da5e3
',
197 'title
': 'Извержение вулкана из спичек
: зрелищный опыт
',
198 'description
': 'md5
:053ad3c61b5596d510c9a199dc8ee633
',
199 'thumbnail
': 're
:^https
://avatars
.mds
.yandex
.net
/',
200 'uploader
': 'Популярная механика
',
203 'skip_download
': 'm3u8
',
206 'url
': 'https
://zen
.yandex
.ru
/media
/id/606fd806cc13cb3c58c05cf5
/vot
-eto
-focus
-dedy
-morozy
-na
-gidrociklah
-60c7c443da18892ebfe85ed7
',
208 'id': '60c7c443da18892ebfe85ed7
',
210 'title
': 'ВОТ ЭТО Focus
. Деды Морозы на гидроциклах
',
211 'description
': 'md5
:f3db3d995763b9bbb7b56d4ccdedea89
',
212 'thumbnail
': 're
:^https
://avatars
.mds
.yandex
.net
/',
213 'uploader
': 'AcademeG DailyStream
'
216 'skip_download
': 'm3u8
',
217 'format
': 'bestvideo
',
220 'url
': 'https
://zen
.yandex
.ru
/video
/watch
/6002240ff8b1af50bb2da5e3
',
222 'id': '6002240ff8b1af50bb2da5e3
',
224 'title
': 'Извержение вулкана из спичек
: зрелищный опыт
',
225 'description
': 'md5
:053ad3c61b5596d510c9a199dc8ee633
',
226 'uploader
': 'Популярная механика
',
229 'skip_download
': 'm3u8
',
232 'url
': 'https
://zen
.yandex
.ru
/media
/id/606fd806cc13cb3c58c05cf5
/novyi
-samsung
-fold
-3-moskvich
-barahlit
-612f93b7f8d48e7e945792a2?
from=channel
&rid
=2286618386.482.1630817595976.42360',
233 'only_matching
': True,
236 def _real_extract(self, url):
237 id = self._match_id(url)
238 webpage = self._download_webpage(url, id)
239 data_json = self._parse_json(
240 self._search_regex(r'data\s
*=\s
*({["\']_*serverState_*video.+?}
);', webpage, 'metadata
'), id)
241 serverstate = self._search_regex(r'(_
+serverState_
+video
-site_
[^_
]+_
+)',
242 webpage, 'server state
').replace('State
', 'Settings
')
243 uploader = self._search_regex(r'(<a\s
*class=["\']card-channel-link[^"\']+["\'][^>]+>)',
244 webpage, 'uploader', default='<a>')
245 uploader_name = extract_attributes(uploader).get('aria-label')
246 video_json = try_get(data_json, lambda x: x[serverstate]['exportData']['video'], dict)
247 stream_urls = try_get(video_json, lambda x: x['video']['streams'])
249 for s_url in stream_urls:
250 ext = determine_ext(s_url)
252 formats.extend(self._extract_mpd_formats(s_url, id, mpd_id='dash'))
254 formats.extend(self._extract_m3u8_formats(s_url, id, 'mp4'))
255 self._sort_formats(formats)
258 'title': video_json.get('title') or self._og_search_title(webpage),
260 'duration': int_or_none(video_json.get('duration')),
261 'view_count': int_or_none(video_json.get('views')),
262 'uploader': uploader_name or data_json.get('authorName') or try_get(data_json, lambda x: x['publisher']['name']),
263 'description': self._og_search_description(webpage) or try_get(data_json, lambda x: x['og']['description']),
264 'thumbnail': self._og_search_thumbnail(webpage) or try_get(data_json, lambda x: x['og']['imageUrl']),
268 class ZenYandexChannelIE(InfoExtractor):
269 _VALID_URL = r'https?://zen\.yandex\.ru/(?!media|video)(?:id/)?(?P<id>[a-z0-9-_]+)'
271 'url': 'https://zen.yandex.ru/tok_media',
275 'playlist_mincount': 169,
277 'url': 'https://zen.yandex.ru/id/606fd806cc13cb3c58c05cf5',
279 'id': '606fd806cc13cb3c58c05cf5',
281 'playlist_mincount': 657,
284 def _entries(self, id, url):
285 webpage = self._download_webpage(url, id)
286 data_json = self._parse_json(re.findall(r'var\s?data\s?=\s?({.+?})\s?;', webpage)[-1], id)
287 for key in data_json.keys():
288 if key.startswith('__serverState__'):
289 data_json = data_json[key]
290 items = list(try_get(data_json, lambda x: x['feed']['items'], dict).values())
291 more = try_get(data_json, lambda x: x['links']['more']) or None
292 for page in itertools.count(1):
294 video_id = item.get('publication_id') or item.get('publicationId')
295 video_url = item.get('link')
296 yield self.url_result(video_url, ie=ZenYandexIE.ie_key(), video_id=video_id.split(':')[-1])
299 data_json = self._download_json(more, id, note='Downloading Page %d' % page)
300 items = data_json.get('items', [])
301 more = try_get(data_json, lambda x: x['more']['link']) or None
303 def _real_extract(self, url):
304 id = self._match_id(url)
305 return self.playlist_result(self._entries(id, url), playlist_id=id)