3 from .common
import InfoExtractor
16 class YandexVideoIE(InfoExtractor
):
20 yandex\.ru(?:/(?:portal/(?:video|efir)|efir))?/?\?.*?stream_id=|
21 frontend\.vh\.yandex\.ru/player/
23 (?P<id>(?:[\da-f]{32}|[\w-]{12}))
26 'url': 'https://yandex.ru/portal/video?stream_id=4dbb36ec4e0526d58f9f2dc8f0ecf374',
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',
41 'params': {'skip_download': 'm3u8'}
,
43 'url': 'https://yandex.ru/portal/efir?stream_id=4dbb262b4fe5cf15a215de4f34eee34d&from=morda',
44 'only_matching': True,
46 'url': 'https://yandex.ru/?stream_id=4dbb262b4fe5cf15a215de4f34eee34d',
47 'only_matching': True,
49 'url': 'https://frontend.vh.yandex.ru/player/4dbb262b4fe5cf15a215de4f34eee34d?from=morda',
50 'only_matching': True,
52 # vod-episode, series episode
53 'url': 'https://yandex.ru/portal/video?stream_id=45b11db6e4b68797919c93751a938cee',
54 'only_matching': True,
57 'url': 'https://yandex.ru/?stream_channel=1538487871&stream_id=4132a07f71fb0396be93d74b3477131d',
58 'only_matching': True,
61 'url': 'https://yandex.ru/portal/video?from=morda&stream_id=485a92d94518d73a9d0ff778e13505f8',
62 'only_matching': True,
64 'url': 'https://yandex.ru/efir?stream_active=watching&stream_id=v7a2dZ-v5mSI&from_block=efir_newtab',
65 'only_matching': True,
68 def _real_extract(self
, url
):
69 video_id
= self
._match
_id
(url
)
71 player
= try_get((self
._download
_json
(
72 'https://frontend.vh.yandex.ru/graphql', video_id
, data
=('''{
73 player(content_id: "%s") {
92 }''' % video_id
).encode(), fatal
=False)), lambda x
: x
['player']['content'])
93 if not player
or player
.get('error'):
94 player
= self
._download
_json
(
95 'https://frontend.vh.yandex.ru/v23/player/%s.json' % video_id
,
97 'stream_options': 'hires',
98 'disable_trackings': 1,
100 content
= player
['content']
102 title
= content
.get('title') or content
['computed_title']
105 streams
= content
.get('streams') or []
106 streams
.append({'url': content.get('content_url')}
)
107 for stream
in streams
:
108 content_url
= url_or_none(stream
.get('url'))
111 ext
= determine_ext(content_url
)
115 formats
.extend(self
._extract
_m
3u8_formats
(
116 content_url
, video_id
, 'mp4',
117 'm3u8_native', m3u8_id
='hls', fatal
=False))
119 formats
.extend(self
._extract
_mpd
_formats
(
120 content_url
, video_id
, mpd_id
='dash', fatal
=False))
122 formats
.append({'url': content_url}
)
124 self
._sort
_formats
(formats
)
126 timestamp
= (int_or_none(content
.get('release_date'))
127 or int_or_none(content
.get('release_date_ut'))
128 or int_or_none(content
.get('start_time')))
129 season
= content
.get('season') or {}
134 'description': content
.get('description'),
135 'thumbnail': content
.get('thumbnail'),
136 'timestamp': timestamp
,
137 'duration': int_or_none(content
.get('duration')),
138 'series': content
.get('program_title'),
139 'age_limit': int_or_none(content
.get('restriction_age')),
140 'view_count': int_or_none(content
.get('views_count')),
141 'like_count': int_or_none(content
.get('likes')),
142 'dislike_count': int_or_none(content
.get('dislikes')),
143 'season_number': int_or_none(season
.get('season_number')),
144 'season_id': season
.get('id'),
145 'release_year': int_or_none(content
.get('release_year')),
150 class YandexVideoPreviewIE(InfoExtractor
):
151 _VALID_URL
= r
'https?://(?:www\.)?yandex\.\w{2,3}(?:\.(?:am|ge|il|tr))?/video/preview(?:/?\?.*?filmId=|/)(?P<id>\d+)'
152 _TESTS
= [{ # Odnoklassniki
153 'url': 'https://yandex.ru/video/preview/?filmId=10682852472978372885&text=summer',
155 'id': '1352565459459',
158 'upload_date': '20191202',
161 'thumbnail': 'https://i.mycdn.me/videoPreview?id=544866765315&type=37&idx=13&tkn=TY5qjLYZHxpmcnK8U2LgzYkgmaU&fn=external_8',
162 'uploader_id': '481054701571',
163 'title': 'LOFT - summer, summer, summer HD',
164 'uploader': 'АРТЁМ КУДРОВ',
167 'url': 'https://yandex.ru/video/preview/?filmId=4479424425337895262&source=main_redirect&text=видео&utm_source=main_stripe_big',
168 'only_matching': True,
170 'url': 'https://yandex.ru/video/preview/5275069442094787341',
171 'only_matching': True,
173 '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',
174 'only_matching': True,
176 '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',
177 'only_matching': True,
179 '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',
180 'only_matching': True,
183 def _real_extract(self
, url
):
184 id = self
._match
_id
(url
)
185 webpage
= self
._download
_webpage
(url
, id)
186 data_raw
= self
._search
_regex
(r
'window.Ya.__inline_params__\s*=\s*JSON.parse\(\'([^
"]+?\\u0022video\\u0022:[^"]+?
})\'\
);', webpage, 'data_raw
')
187 data_json = self._parse_json(data_raw, id, transform_source=lowercase_escape)
188 return self.url_result(data_json['video
']['url
'])
191 class ZenYandexIE(InfoExtractor):
192 _VALID_URL = r'https?
://(zen\
.yandex|dzen
)\
.ru(?
:/video
)?
/(media|watch
)/(?
:(?
:id/[^
/]+/|
[^
/]+/)(?
:[a
-z0
-9-]+)-)?
(?P
<id>[a
-z0
-9-]+)'
194 'url
': 'https
://zen
.yandex
.ru
/media
/id/606fd806cc13cb3c58c05cf5
/vot
-eto
-focus
-dedy
-morozy
-na
-gidrociklah
-60c7c443da18892ebfe85ed7
',
196 'id': '60c7c443da18892ebfe85ed7
',
198 'title
': 'ВОТ ЭТО Focus
. Деды Морозы на гидроциклах
',
199 'description
': 'md5
:f3db3d995763b9bbb7b56d4ccdedea89
',
200 'thumbnail
': 're
:^https
://avatars
.dzeninfra
.ru
/',
201 'uploader
': 'AcademeG DailyStream
'
204 'skip_download
': 'm3u8
',
205 'format
': 'bestvideo
',
207 'skip
': 'The page does
not exist
',
209 'url
': 'https
://dzen
.ru
/media
/id/606fd806cc13cb3c58c05cf5
/vot
-eto
-focus
-dedy
-morozy
-na
-gidrociklah
-60c7c443da18892ebfe85ed7
',
211 'id': '60c7c443da18892ebfe85ed7
',
213 'title
': 'ВОТ ЭТО Focus
. Деды Морозы на гидроциклах
',
214 'description
': 'md5
:f3db3d995763b9bbb7b56d4ccdedea89
',
215 'thumbnail
': r're
:^https
://avatars\
.dzeninfra\
.ru
/',
216 'uploader
': 'AcademeG DailyStream
',
217 'upload_date
': '20191111',
218 'timestamp
': 1573465585,
220 'params
': {'skip_download': 'm3u8'},
222 'url
': 'https
://zen
.yandex
.ru
/video
/watch
/6002240ff8b1af50bb2da5e3
',
224 'id': '6002240ff8b1af50bb2da5e3
',
226 'title
': 'Извержение вулкана из спичек
: зрелищный опыт
',
227 'description
': 'md5
:053ad3c61b5596d510c9a199dc8ee633
',
228 'thumbnail
': r're
:^https
://avatars\
.dzeninfra\
.ru
/',
229 'uploader
': 'TechInsider
',
230 'timestamp
': 1611378221,
231 'upload_date
': '20210123',
233 'params
': {'skip_download': 'm3u8'},
235 'url
': 'https
://dzen
.ru
/video
/watch
/6002240ff8b1af50bb2da5e3
',
237 'id': '6002240ff8b1af50bb2da5e3
',
239 'title
': 'Извержение вулкана из спичек
: зрелищный опыт
',
240 'description
': 'md5
:053ad3c61b5596d510c9a199dc8ee633
',
241 'thumbnail
': 're
:^https
://avatars
.dzeninfra
.ru
/',
242 'uploader
': 'TechInsider
',
243 'upload_date
': '20210123',
244 'timestamp
': 1611378221,
246 'params
': {'skip_download': 'm3u8'},
248 'url
': 'https
://zen
.yandex
.ru
/media
/id/606fd806cc13cb3c58c05cf5
/novyi
-samsung
-fold
-3-moskvich
-barahlit
-612f93b7f8d48e7e945792a2?
from=channel
&rid
=2286618386.482.1630817595976.42360',
249 'only_matching
': True,
251 'url
': 'https
://dzen
.ru
/media
/id/606fd806cc13cb3c58c05cf5
/novyi
-samsung
-fold
-3-moskvich
-barahlit
-612f93b7f8d48e7e945792a2?
from=channel
&rid
=2286618386.482.1630817595976.42360',
252 'only_matching
': True,
255 def _real_extract(self, url):
256 video_id = self._match_id(url)
257 webpage = self._download_webpage(url, video_id)
258 redirect = self._search_json(r'var it\s
*=\s
*', webpage, 'redirect
', id, default={}).get('retpath
')
260 video_id = self._match_id(redirect)
261 webpage = self._download_webpage(redirect, video_id, note='Redirecting
')
262 data_json = self._search_json(
263 r'data\s
*=', webpage, 'metadata
', video_id, contains_pattern=r'{["\']_*serverState_*video.+}
')
264 serverstate = self._search_regex(r'(_
+serverState_
+video
-site_
[^_
]+_
+)',
265 webpage, 'server state
').replace('State
', 'Settings
')
266 uploader = self._search_regex(r'(<a\s
*class=["\']card-channel-link[^"\']+["\'][^>]+>)',
267 webpage, 'uploader', default='<a>')
268 uploader_name = extract_attributes(uploader).get('aria-label')
269 video_json = try_get(data_json, lambda x: x[serverstate]['exportData']['video'], dict)
270 stream_urls = try_get(video_json, lambda x: x['video']['streams'])
272 for s_url in stream_urls:
273 ext = determine_ext(s_url)
275 formats.extend(self._extract_mpd_formats(s_url, id, mpd_id='dash'))
277 formats.extend(self._extract_m3u8_formats(s_url, id, 'mp4'))
278 self._sort_formats(formats)
281 'title': video_json.get('title') or self._og_search_title(webpage),
283 'duration': int_or_none(video_json.get('duration')),
284 'view_count': int_or_none(video_json.get('views')),
285 'timestamp': int_or_none(video_json.get('publicationDate')),
286 'uploader': uploader_name or data_json.get('authorName') or try_get(data_json, lambda x: x['publisher']['name']),
287 'description': self._og_search_description(webpage) or try_get(data_json, lambda x: x['og']['description']),
288 'thumbnail': self._og_search_thumbnail(webpage) or try_get(data_json, lambda x: x['og']['imageUrl']),
292 class ZenYandexChannelIE(InfoExtractor):
293 _VALID_URL = r'https?://(zen\.yandex|dzen)\.ru/(?!media|video)(?:id/)?(?P<id>[a-z0-9-_]+)'
295 'url': 'https://zen.yandex.ru/tok_media',
299 'description': 'md5:a9e5b3c247b7fe29fd21371a428bcf56',
301 'playlist_mincount': 169,
303 'url': 'https://dzen.ru/tok_media',
307 'description': 'md5:a9e5b3c247b7fe29fd21371a428bcf56',
309 'playlist_mincount': 169,
311 'url': 'https://zen.yandex.ru/id/606fd806cc13cb3c58c05cf5',
313 'id': '606fd806cc13cb3c58c05cf5',
314 'description': 'md5:517b7c97d8ca92e940f5af65448fd928',
315 'title': 'AcademeG DailyStream',
317 'playlist_mincount': 657,
319 # Test that the playlist extractor finishes extracting when the
320 # channel has less than one page
321 'url': 'https://zen.yandex.ru/jony_me',
324 'description': 'md5:a2c62b4ef5cf3e3efb13d25f61f739e1',
327 'playlist_count': 20,
329 # Test that the playlist extractor finishes extracting when the
330 # channel has more than one page of entries
331 'url': 'https://zen.yandex.ru/tatyanareva',
334 'description': 'md5:296b588d60841c3756c9105f237b70c6',
335 'title': 'Татьяна Рева',
336 'entries': 'maxcount:200',
338 'playlist_count': 46,
340 'url': 'https://dzen.ru/id/606fd806cc13cb3c58c05cf5',
342 'id': '606fd806cc13cb3c58c05cf5',
343 'title': 'AcademeG DailyStream',
344 'description': 'md5:517b7c97d8ca92e940f5af65448fd928',
346 'playlist_mincount': 657,
349 def _entries(self, item_id, server_state_json, server_settings_json):
350 items = (traverse_obj(server_state_json, ('feed', 'items', ...))
351 or traverse_obj(server_settings_json, ('exportData', 'items', ...)))
353 more = (traverse_obj(server_state_json, ('links', 'more'))
354 or traverse_obj(server_settings_json, ('exportData', 'more', 'link')))
357 for page in itertools.count(1):
358 for item in items or []:
359 if item.get('type') != 'gif':
361 video_id = traverse_obj(item, 'publication_id', 'publicationId') or ''
362 yield self.url_result(item['link'], ZenYandexIE, video_id.split(':')[-1])
364 current_page_id = next_page_id
365 next_page_id = traverse_obj(parse_qs(more), ('next_page_id', -1))
366 if not all((more, items, next_page_id, next_page_id != current_page_id)):
369 data = self._download_json(more, item_id, note=f'Downloading Page {page}')
370 items, more = data.get('items'), traverse_obj(data, ('more', 'link'))
372 def _real_extract(self, url):
373 item_id = self._match_id(url)
374 webpage = self._download_webpage(url, item_id)
375 redirect = self._search_json(
376 r'var it\s*=\s*', webpage, 'redirect', item_id, default={}).get('retpath')
378 item_id = self._match_id(redirect)
379 webpage = self._download_webpage(redirect, item_id, note='Redirecting')
380 data = self._search_json(
381 r'var\s+data\s*=', webpage, 'channel data', item_id, contains_pattern=r'{\"__serverState__.+}')
382 server_state_json = traverse_obj(data, lambda k, _: k.startswith('__serverState__'), get_all=False)
383 server_settings_json = traverse_obj(data, lambda k, _: k.startswith('__serverSettings__'), get_all=False)
385 return self.playlist_result(
386 self._entries(item_id, server_state_json, server_settings_json),
387 item_id, traverse_obj(server_state_json, ('channel', 'source', 'title')),
388 traverse_obj(server_state_json, ('channel', 'source', 'description')))