2 from __future__
import unicode_literals
17 from ..compat
import (
18 compat_cookiejar_Cookie
,
19 compat_cookies_SimpleCookie
,
21 compat_etree_fromstring
,
28 compat_urllib_parse_unquote
,
29 compat_urllib_parse_urlencode
,
30 compat_urllib_request
,
32 compat_xml_parse_error
,
34 from ..downloader
import FileDownloader
35 from ..downloader
.f4m
import (
37 remove_encrypted_media
,
69 parse_m3u8_attributes
,
95 class InfoExtractor(object):
96 """Information Extractor class.
98 Information extractors are the classes that, given a URL, extract
99 information about the video (or videos) the URL refers to. This
100 information includes the real video URL, the video title, author and
101 others. The information is stored in a dictionary which is then
102 passed to the YoutubeDL. The YoutubeDL processes this
103 information possibly downloading the video to the file system, among
104 other possible outcomes.
106 The type field determines the type of the result.
107 By far the most common value (and the default if _type is missing) is
108 "video", which indicates a single video.
110 For a video, the dictionaries must include the following fields:
112 id: Video identifier.
113 title: Video title, unescaped.
115 Additionally, it must contain either a formats entry or a url one:
117 formats: A list of dictionaries for each format available, ordered
118 from worst to best quality.
121 * url The mandatory URL representing the media:
122 for plain file media - HTTP URL of this file,
124 for HLS - URL of the M3U8 media playlist,
125 for HDS - URL of the F4M manifest,
127 - HTTP URL to plain file media (in case of
129 - URL of the MPD manifest or base URL
130 representing the media if MPD manifest
131 is parsed from a string (in case of
133 for MSS - URL of the ISM manifest.
135 The URL of the manifest file in case of
137 for HLS - URL of the M3U8 master playlist,
138 for HDS - URL of the F4M manifest,
139 for DASH - URL of the MPD manifest,
140 for MSS - URL of the ISM manifest.
141 * ext Will be calculated from URL if missing
142 * format A human-readable description of the format
143 ("mp4 container with h264/opus").
144 Calculated from the format_id, width, height.
145 and format_note fields if missing.
146 * format_id A short description of the format
147 ("mp4_h264_opus" or "19").
148 Technically optional, but strongly recommended.
149 * format_note Additional info about the format
150 ("3D" or "DASH video")
151 * width Width of the video, if known
152 * height Height of the video, if known
153 * resolution Textual description of width and height
154 * dynamic_range The dynamic range of the video. One of:
155 "SDR" (None), "HDR10", "HDR10+, "HDR12", "HLG, "DV"
156 * tbr Average bitrate of audio and video in KBit/s
157 * abr Average audio bitrate in KBit/s
158 * acodec Name of the audio codec in use
159 * asr Audio sampling rate in Hertz
160 * vbr Average video bitrate in KBit/s
162 * vcodec Name of the video codec in use
163 * container Name of the container format
164 * filesize The number of bytes, if known in advance
165 * filesize_approx An estimate for the number of bytes
166 * player_url SWF Player URL (used for rtmpdump).
167 * protocol The protocol that will be used for the actual
168 download, lower-case. One of "http", "https" or
169 one of the protocols defined in downloader.PROTOCOL_MAP
171 Base URL for fragments. Each fragment's path
172 value (if present) will be relative to
174 * fragments A list of fragments of a fragmented media.
175 Each fragment entry must contain either an url
176 or a path. If an url is present it should be
177 considered by a client. Otherwise both path and
178 fragment_base_url must be present. Here is
179 the list of all potential fields:
180 * "url" - fragment's URL
181 * "path" - fragment's path relative to
183 * "duration" (optional, int or float)
184 * "filesize" (optional, int)
185 * is_from_start Is a live format that can be downloaded
186 from the start. Boolean
187 * preference Order number of this format. If this field is
188 present and not None, the formats get sorted
189 by this field, regardless of all other values.
190 -1 for default (order by other properties),
191 -2 or smaller for less than default.
192 < -1000 to hide the format (if there is
193 another one which is strictly better)
194 * language Language code, e.g. "de" or "en-US".
195 * language_preference Is this in the language mentioned in
197 10 if it's what the URL is about,
198 -1 for default (don't know),
199 -10 otherwise, other values reserved for now.
200 * quality Order number of the video quality of this
201 format, irrespective of the file format.
202 -1 for default (order by other properties),
203 -2 or smaller for less than default.
204 * source_preference Order number for this video source
205 (quality takes higher priority)
206 -1 for default (order by other properties),
207 -2 or smaller for less than default.
208 * http_headers A dictionary of additional HTTP headers
209 to add to the request.
210 * stretched_ratio If given and not 1, indicates that the
211 video's pixels are not square.
212 width : height ratio as float.
213 * no_resume The server does not support resuming the
214 (HTTP or RTMP) download. Boolean.
215 * has_drm The format has DRM and cannot be downloaded. Boolean
216 * downloader_options A dictionary of downloader options as
217 described in FileDownloader
218 RTMP formats can also have the additional fields: page_url,
219 app, play_path, tc_url, flash_version, rtmp_live, rtmp_conn,
220 rtmp_protocol, rtmp_real_time
222 url: Final video URL.
223 ext: Video filename extension.
224 format: The video format, defaults to ext (used for --get-format)
225 player_url: SWF Player URL (used for rtmpdump).
227 The following fields are optional:
229 direct: True if a direct video file was given (must only be set by GenericIE)
230 alt_title: A secondary title of the video.
231 display_id An alternative identifier for the video, not necessarily
232 unique, but available before title. Typically, id is
233 something like "4234987", title "Dancing naked mole rats",
234 and display_id "dancing-naked-mole-rats"
235 thumbnails: A list of dictionaries, with the following entries:
236 * "id" (optional, string) - Thumbnail format ID
238 * "preference" (optional, int) - quality of the image
239 * "width" (optional, int)
240 * "height" (optional, int)
241 * "resolution" (optional, string "{width}x{height}",
243 * "filesize" (optional, int)
244 * "http_headers" (dict) - HTTP headers for the request
245 thumbnail: Full URL to a video thumbnail image.
246 description: Full video description.
247 uploader: Full name of the video uploader.
248 license: License name the video is licensed under.
249 creator: The creator of the video.
250 timestamp: UNIX timestamp of the moment the video was uploaded
251 upload_date: Video upload date (YYYYMMDD).
252 If not explicitly set, calculated from timestamp
253 release_timestamp: UNIX timestamp of the moment the video was released.
254 If it is not clear whether to use timestamp or this, use the former
255 release_date: The date (YYYYMMDD) when the video was released.
256 If not explicitly set, calculated from release_timestamp
257 modified_timestamp: UNIX timestamp of the moment the video was last modified.
258 modified_date: The date (YYYYMMDD) when the video was last modified.
259 If not explicitly set, calculated from modified_timestamp
260 uploader_id: Nickname or id of the video uploader.
261 uploader_url: Full URL to a personal webpage of the video uploader.
262 channel: Full name of the channel the video is uploaded on.
263 Note that channel fields may or may not repeat uploader
264 fields. This depends on a particular extractor.
265 channel_id: Id of the channel.
266 channel_url: Full URL to a channel webpage.
267 channel_follower_count: Number of followers of the channel.
268 location: Physical location where the video was filmed.
269 subtitles: The available subtitles as a dictionary in the format
270 {tag: subformats}. "tag" is usually a language code, and
271 "subformats" is a list sorted from lower to higher
272 preference, each element is a dictionary with the "ext"
274 * "data": The subtitles file contents
275 * "url": A URL pointing to the subtitles file
276 It can optionally also have:
277 * "name": Name or description of the subtitles
278 * "http_headers": A dictionary of additional HTTP headers
279 to add to the request.
280 "ext" will be calculated from URL if missing
281 automatic_captions: Like 'subtitles'; contains automatically generated
282 captions instead of normal subtitles
283 duration: Length of the video in seconds, as an integer or float.
284 view_count: How many users have watched the video on the platform.
285 like_count: Number of positive ratings of the video
286 dislike_count: Number of negative ratings of the video
287 repost_count: Number of reposts of the video
288 average_rating: Average rating give by users, the scale used depends on the webpage
289 comment_count: Number of comments on the video
290 comments: A list of comments, each with one or more of the following
291 properties (all but one of text or html optional):
292 * "author" - human-readable name of the comment author
293 * "author_id" - user ID of the comment author
294 * "author_thumbnail" - The thumbnail of the comment author
296 * "html" - Comment as HTML
297 * "text" - Plain text of the comment
298 * "timestamp" - UNIX timestamp of comment
299 * "parent" - ID of the comment this one is replying to.
300 Set to "root" to indicate that this is a
301 comment to the original video.
302 * "like_count" - Number of positive ratings of the comment
303 * "dislike_count" - Number of negative ratings of the comment
304 * "is_favorited" - Whether the comment is marked as
305 favorite by the video uploader
306 * "author_is_uploader" - Whether the comment is made by
308 age_limit: Age restriction for the video, as an integer (years)
309 webpage_url: The URL to the video webpage, if given to yt-dlp it
310 should allow to get the same result again. (It will be set
311 by YoutubeDL if it's missing)
312 categories: A list of categories that the video falls in, for example
314 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
315 cast: A list of the video cast
316 is_live: True, False, or None (=unknown). Whether this video is a
317 live stream that goes on instead of a fixed-length video.
318 was_live: True, False, or None (=unknown). Whether this video was
319 originally a live stream.
320 live_status: 'is_live', 'is_upcoming', 'was_live', 'not_live' or None (=unknown)
321 If absent, automatically set from is_live, was_live
322 start_time: Time in seconds where the reproduction should start, as
323 specified in the URL.
324 end_time: Time in seconds where the reproduction should end, as
325 specified in the URL.
326 chapters: A list of dictionaries, with the following entries:
327 * "start_time" - The start time of the chapter in seconds
328 * "end_time" - The end time of the chapter in seconds
329 * "title" (optional, string)
330 playable_in_embed: Whether this video is allowed to play in embedded
331 players on other sites. Can be True (=always allowed),
332 False (=never allowed), None (=unknown), or a string
333 specifying the criteria for embedability (Eg: 'whitelist')
334 availability: Under what condition the video is available. One of
335 'private', 'premium_only', 'subscriber_only', 'needs_auth',
336 'unlisted' or 'public'. Use 'InfoExtractor._availability'
338 __post_extractor: A function to be called just before the metadata is
339 written to either disk, logger or console. The function
340 must return a dict which will be added to the info_dict.
341 This is usefull for additional information that is
342 time-consuming to extract. Note that the fields thus
343 extracted will not be available to output template and
344 match_filter. So, only "comments" and "comment_count" are
345 currently allowed to be extracted via this method.
347 The following fields should only be used when the video belongs to some logical
350 chapter: Name or title of the chapter the video belongs to.
351 chapter_number: Number of the chapter the video belongs to, as an integer.
352 chapter_id: Id of the chapter the video belongs to, as a unicode string.
354 The following fields should only be used when the video is an episode of some
355 series, programme or podcast:
357 series: Title of the series or programme the video episode belongs to.
358 series_id: Id of the series or programme the video episode belongs to, as a unicode string.
359 season: Title of the season the video episode belongs to.
360 season_number: Number of the season the video episode belongs to, as an integer.
361 season_id: Id of the season the video episode belongs to, as a unicode string.
362 episode: Title of the video episode. Unlike mandatory video title field,
363 this field should denote the exact title of the video episode
364 without any kind of decoration.
365 episode_number: Number of the video episode within a season, as an integer.
366 episode_id: Id of the video episode, as a unicode string.
368 The following fields should only be used when the media is a track or a part of
371 track: Title of the track.
372 track_number: Number of the track within an album or a disc, as an integer.
373 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
375 artist: Artist(s) of the track.
376 genre: Genre(s) of the track.
377 album: Title of the album the track belongs to.
378 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
379 album_artist: List of all artists appeared on the album (e.g.
380 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
382 disc_number: Number of the disc or other physical medium the track belongs to,
384 release_year: Year (YYYY) when the album was released.
385 composer: Composer of the piece
387 Unless mentioned otherwise, the fields should be Unicode strings.
389 Unless mentioned otherwise, None is equivalent to absence of information.
392 _type "playlist" indicates multiple videos.
393 There must be a key "entries", which is a list, an iterable, or a PagedList
394 object, each element of which is a valid dictionary by this specification.
396 Additionally, playlists can have "id", "title", and any other relevent
397 attributes with the same semantics as videos (see above).
399 It can also have the following optional fields:
401 playlist_count: The total number of videos in a playlist. If not given,
402 YoutubeDL tries to calculate it from "entries"
405 _type "multi_video" indicates that there are multiple videos that
406 form a single show, for examples multiple acts of an opera or TV episode.
407 It must have an entries key like a playlist and contain all the keys
408 required for a video at the same time.
411 _type "url" indicates that the video must be extracted from another
412 location, possibly by a different extractor. Its only required key is:
413 "url" - the next URL to extract.
414 The key "ie_key" can be set to the class name (minus the trailing "IE",
415 e.g. "Youtube") if the extractor class is known in advance.
416 Additionally, the dictionary may have any properties of the resolved entity
417 known in advance, for example "title" if the title of the referred video is
421 _type "url_transparent" entities have the same specification as "url", but
422 indicate that the given additional information is more precise than the one
423 associated with the resolved URL.
424 This is useful when a site employs a video service that hosts the video and
425 its technical metadata, but that video service does not embed a useful
426 title, description etc.
429 Subclasses of this should define a _VALID_URL regexp and, re-define the
430 _real_extract() and (optionally) _real_initialize() methods.
431 Probably, they should also be added to the list of extractors.
433 Subclasses may also override suitable() if necessary, but ensure the function
434 signature is preserved and that this function imports everything it needs
435 (except other extractors), so that lazy_extractors works correctly.
437 To support username + password (or netrc) login, the extractor must define a
438 _NETRC_MACHINE and re-define _perform_login(username, password) and
439 (optionally) _initialize_pre_login() methods. The _perform_login method will
440 be called between _initialize_pre_login and _real_initialize if credentials
441 are passed by the user. In cases where it is necessary to have the login
442 process as part of the extraction rather than initialization, _perform_login
443 can be left undefined.
445 _GEO_BYPASS attribute may be set to False in order to disable
446 geo restriction bypass mechanisms for a particular extractor.
447 Though it won't disable explicit geo restriction bypass based on
448 country code provided with geo_bypass_country.
450 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
451 countries for this extractor. One of these countries will be used by
452 geo restriction bypass mechanism right away in order to bypass
453 geo restriction, of course, if the mechanism is not disabled.
455 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
456 IP blocks in CIDR notation for this extractor. One of these IP blocks
457 will be used by geo restriction bypass mechanism similarly
460 The _WORKING attribute should be set to False for broken IEs
461 in order to warn the users and skip the tests.
466 _x_forwarded_for_ip
= None
468 _GEO_COUNTRIES
= None
469 _GEO_IP_BLOCKS
= None
471 _NETRC_MACHINE
= None
475 'any': 'Use --cookies, --cookies-from-browser, --username and --password, or --netrc to provide account credentials',
477 'Use --cookies-from-browser or --cookies for the authentication. '
478 'See https://github.com/ytdl-org/youtube-dl#how-do-i-pass-cookies-to-youtube-dl for how to manually pass cookies'),
479 'password': 'Use --username and --password, or --netrc to provide account credentials',
482 def __init__(self
, downloader
=None):
483 """Constructor. Receives an optional downloader (a YoutubeDL instance).
484 If a downloader is not passed during initialization,
485 it must be set using "set_downloader()" before "extract()" is called"""
487 self
._x
_forwarded
_for
_ip
= None
488 self
._printed
_messages
= set()
489 self
.set_downloader(downloader
)
492 def _match_valid_url(cls
, url
):
493 # This does not use has/getattr intentionally - we want to know whether
494 # we have cached the regexp for *this* class, whereas getattr would also
495 # match the superclass
496 if '_VALID_URL_RE' not in cls
.__dict
__:
497 if '_VALID_URL' not in cls
.__dict
__:
498 cls
._VALID
_URL
= cls
._make
_valid
_url
()
499 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
500 return cls
._VALID
_URL
_RE
.match(url
)
503 def suitable(cls
, url
):
504 """Receives a URL and returns True if suitable for this IE."""
505 # This function must import everything it needs (except other extractors),
506 # so that lazy_extractors works correctly
507 return cls
._match
_valid
_url
(url
) is not None
510 def _match_id(cls
, url
):
511 return cls
._match
_valid
_url
(url
).group('id')
514 def get_temp_id(cls
, url
):
516 return cls
._match
_id
(url
)
517 except (IndexError, AttributeError):
522 """Getter method for _WORKING."""
526 def supports_login(cls
):
527 return bool(cls
._NETRC
_MACHINE
)
529 def initialize(self
):
530 """Initializes an instance (authentication, etc)."""
531 self
._printed
_messages
= set()
532 self
._initialize
_geo
_bypass
({
533 'countries': self
._GEO
_COUNTRIES
,
534 'ip_blocks': self
._GEO
_IP
_BLOCKS
,
537 self
._initialize
_pre
_login
()
538 if self
.supports_login():
539 username
, password
= self
._get
_login
_info
()
541 self
._perform
_login
(username
, password
)
542 elif self
.get_param('username') and False not in (self
.IE_DESC
, self
._NETRC
_MACHINE
):
543 self
.report_warning(f
'Login with password is not supported for this website. {self._LOGIN_HINTS["cookies"]}')
544 self
._real
_initialize
()
547 def _initialize_geo_bypass(self
, geo_bypass_context
):
549 Initialize geo restriction bypass mechanism.
551 This method is used to initialize geo bypass mechanism based on faking
552 X-Forwarded-For HTTP header. A random country from provided country list
553 is selected and a random IP belonging to this country is generated. This
554 IP will be passed as X-Forwarded-For HTTP header in all subsequent
557 This method will be used for initial geo bypass mechanism initialization
558 during the instance initialization with _GEO_COUNTRIES and
561 You may also manually call it from extractor's code if geo bypass
562 information is not available beforehand (e.g. obtained during
563 extraction) or due to some other reason. In this case you should pass
564 this information in geo bypass context passed as first argument. It may
565 contain following fields:
567 countries: List of geo unrestricted countries (similar
569 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
570 (similar to _GEO_IP_BLOCKS)
573 if not self
._x
_forwarded
_for
_ip
:
575 # Geo bypass mechanism is explicitly disabled by user
576 if not self
.get_param('geo_bypass', True):
579 if not geo_bypass_context
:
580 geo_bypass_context
= {}
582 # Backward compatibility: previously _initialize_geo_bypass
583 # expected a list of countries, some 3rd party code may still use
585 if isinstance(geo_bypass_context
, (list, tuple)):
586 geo_bypass_context
= {
587 'countries': geo_bypass_context
,
590 # The whole point of geo bypass mechanism is to fake IP
591 # as X-Forwarded-For HTTP header based on some IP block or
594 # Path 1: bypassing based on IP block in CIDR notation
596 # Explicit IP block specified by user, use it right away
597 # regardless of whether extractor is geo bypassable or not
598 ip_block
= self
.get_param('geo_bypass_ip_block', None)
600 # Otherwise use random IP block from geo bypass context but only
601 # if extractor is known as geo bypassable
603 ip_blocks
= geo_bypass_context
.get('ip_blocks')
604 if self
._GEO
_BYPASS
and ip_blocks
:
605 ip_block
= random
.choice(ip_blocks
)
608 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(ip_block
)
609 self
._downloader
.write_debug(
610 '[debug] Using fake IP %s as X-Forwarded-For' % self
._x
_forwarded
_for
_ip
)
613 # Path 2: bypassing based on country code
615 # Explicit country code specified by user, use it right away
616 # regardless of whether extractor is geo bypassable or not
617 country
= self
.get_param('geo_bypass_country', None)
619 # Otherwise use random country code from geo bypass context but
620 # only if extractor is known as geo bypassable
622 countries
= geo_bypass_context
.get('countries')
623 if self
._GEO
_BYPASS
and countries
:
624 country
= random
.choice(countries
)
627 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country
)
628 self
._downloader
.write_debug(
629 'Using fake IP %s (%s) as X-Forwarded-For' % (self
._x
_forwarded
_for
_ip
, country
.upper()))
631 def extract(self
, url
):
632 """Extracts URL information and returns it in list of dicts."""
637 self
.write_debug('Extracting URL: %s' % url
)
638 ie_result
= self
._real
_extract
(url
)
639 if ie_result
is None:
641 if self
._x
_forwarded
_for
_ip
:
642 ie_result
['__x_forwarded_for_ip'] = self
._x
_forwarded
_for
_ip
643 subtitles
= ie_result
.get('subtitles')
644 if (subtitles
and 'live_chat' in subtitles
645 and 'no-live-chat' in self
.get_param('compat_opts', [])):
646 del subtitles
['live_chat']
648 except GeoRestrictedError
as e
:
649 if self
.__maybe
_fake
_ip
_and
_retry
(e
.countries
):
652 except UnsupportedError
:
654 except ExtractorError
as e
:
656 'video_id': e
.video_id
or self
.get_temp_id(url
),
658 'tb': e
.traceback
or sys
.exc_info()[2],
659 'expected': e
.expected
,
662 if hasattr(e
, 'countries'):
663 kwargs
['countries'] = e
.countries
664 raise type(e
)(e
.orig_msg
, **kwargs
)
665 except compat_http_client
.IncompleteRead
as e
:
666 raise ExtractorError('A network error has occurred.', cause
=e
, expected
=True, video_id
=self
.get_temp_id(url
))
667 except (KeyError, StopIteration) as e
:
668 raise ExtractorError('An extractor error has occurred.', cause
=e
, video_id
=self
.get_temp_id(url
))
670 def __maybe_fake_ip_and_retry(self
, countries
):
671 if (not self
.get_param('geo_bypass_country', None)
673 and self
.get_param('geo_bypass', True)
674 and not self
._x
_forwarded
_for
_ip
676 country_code
= random
.choice(countries
)
677 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country_code
)
678 if self
._x
_forwarded
_for
_ip
:
680 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
681 % (self
._x
_forwarded
_for
_ip
, country_code
.upper()))
685 def set_downloader(self
, downloader
):
686 """Sets a YoutubeDL instance as the downloader for this IE."""
687 self
._downloader
= downloader
689 def _initialize_pre_login(self
):
690 """ Intialization before login. Redefine in subclasses."""
693 def _perform_login(self
, username
, password
):
694 """ Login with username and password. Redefine in subclasses."""
697 def _real_initialize(self
):
698 """Real initialization process. Redefine in subclasses."""
701 def _real_extract(self
, url
):
702 """Real extraction process. Redefine in subclasses."""
703 raise NotImplementedError('This method must be implemented by subclasses')
707 """A string for getting the InfoExtractor with get_info_extractor"""
708 return cls
.__name
__[:-2]
712 return compat_str(type(self
).__name
__[:-2])
715 def __can_accept_status_code(err
, expected_status
):
716 assert isinstance(err
, compat_urllib_error
.HTTPError
)
717 if expected_status
is None:
719 elif callable(expected_status
):
720 return expected_status(err
.code
) is True
722 return err
.code
in variadic(expected_status
)
724 def _request_webpage(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
, expected_status
=None):
726 Return the response handle.
728 See _download_webpage docstring for arguments specification.
730 if not self
._downloader
._first
_webpage
_request
:
731 sleep_interval
= self
.get_param('sleep_interval_requests') or 0
732 if sleep_interval
> 0:
733 self
.to_screen('Sleeping %s seconds ...' % sleep_interval
)
734 time
.sleep(sleep_interval
)
736 self
._downloader
._first
_webpage
_request
= False
739 self
.report_download_webpage(video_id
)
740 elif note
is not False:
742 self
.to_screen('%s' % (note
,))
744 self
.to_screen('%s: %s' % (video_id
, note
))
746 # Some sites check X-Forwarded-For HTTP header in order to figure out
747 # the origin of the client behind proxy. This allows bypassing geo
748 # restriction by faking this header's value to IP that belongs to some
749 # geo unrestricted country. We will do so once we encounter any
750 # geo restriction error.
751 if self
._x
_forwarded
_for
_ip
:
752 if 'X-Forwarded-For' not in headers
:
753 headers
['X-Forwarded-For'] = self
._x
_forwarded
_for
_ip
755 if isinstance(url_or_request
, compat_urllib_request
.Request
):
756 url_or_request
= update_Request(
757 url_or_request
, data
=data
, headers
=headers
, query
=query
)
760 url_or_request
= update_url_query(url_or_request
, query
)
761 if data
is not None or headers
:
762 url_or_request
= sanitized_Request(url_or_request
, data
, headers
)
764 return self
._downloader
.urlopen(url_or_request
)
765 except network_exceptions
as err
:
766 if isinstance(err
, compat_urllib_error
.HTTPError
):
767 if self
.__can
_accept
_status
_code
(err
, expected_status
):
768 # Retain reference to error to prevent file object from
769 # being closed before it can be read. Works around the
770 # effects of <https://bugs.python.org/issue15002>
771 # introduced in Python 3.4.1.
778 errnote
= 'Unable to download webpage'
780 errmsg
= '%s: %s' % (errnote
, error_to_compat_str(err
))
782 raise ExtractorError(errmsg
, cause
=err
)
784 self
.report_warning(errmsg
)
787 def _download_webpage_handle(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, encoding
=None, data
=None, headers
={}, query={}
, expected_status
=None):
789 Return a tuple (page content as string, URL handle).
791 See _download_webpage docstring for arguments specification.
793 # Strip hashes from the URL (#1038)
794 if isinstance(url_or_request
, (compat_str
, str)):
795 url_or_request
= url_or_request
.partition('#')[0]
797 urlh
= self
._request
_webpage
(url_or_request
, video_id
, note
, errnote
, fatal
, data
=data
, headers
=headers
, query
=query
, expected_status
=expected_status
)
801 content
= self
._webpage
_read
_content
(urlh
, url_or_request
, video_id
, note
, errnote
, fatal
, encoding
=encoding
)
802 return (content
, urlh
)
805 def _guess_encoding_from_content(content_type
, webpage_bytes
):
806 m
= re
.match(r
'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type
)
808 encoding
= m
.group(1)
810 m
= re
.search(br
'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
811 webpage_bytes[:1024])
813 encoding = m.group(1).decode('ascii')
814 elif webpage_bytes.startswith(b'\xff\xfe'):
821 def __check_blocked(self, content):
822 first_block = content[:512]
823 if ('<title>Access to this site is blocked</title>' in content
824 and 'Websense' in first_block):
825 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
826 blocked_iframe = self._html_search_regex(
827 r'<iframe src="([^
"]+)"', content,
828 'Websense information URL
', default=None)
830 msg += ' Visit
%s for more details
' % blocked_iframe
831 raise ExtractorError(msg, expected=True)
832 if '<title
>The URL you requested has been blocked
</title
>' in first_block:
834 'Access to this webpage has been blocked by Indian censorship
. '
835 'Use a VPN
or proxy
server (with --proxy
) to route around it
.')
836 block_msg = self._html_search_regex(
837 r'</h1
><p
>(.*?
)</p
>',
838 content, 'block message
', default=None)
840 msg += ' (Message
: "%s")' % block_msg.replace('\n', ' ')
841 raise ExtractorError(msg, expected=True)
842 if ('<title
>TTK
:: Доступ к ресурсу ограничен
</title
>' in content
843 and 'blocklist
.rkn
.gov
.ru
' in content):
844 raise ExtractorError(
845 'Access to this webpage has been blocked by decision of the Russian government
. '
846 'Visit http
://blocklist
.rkn
.gov
.ru
/ for a block reason
.',
849 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
850 content_type = urlh.headers.get('Content
-Type
', '')
851 webpage_bytes = urlh.read()
852 if prefix is not None:
853 webpage_bytes = prefix + webpage_bytes
855 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
856 if self.get_param('dump_intermediate_pages
', False):
857 self.to_screen('Dumping request to
' + urlh.geturl())
858 dump = base64.b64encode(webpage_bytes).decode('ascii
')
859 self._downloader.to_screen(dump)
860 if self.get_param('write_pages
', False):
861 basen = '%s_%s' % (video_id, urlh.geturl())
862 trim_length = self.get_param('trim_file_name
') or 240
863 if len(basen) > trim_length:
864 h = '___
' + hashlib.md5(basen.encode('utf
-8')).hexdigest()
865 basen = basen[:trim_length - len(h)] + h
866 raw_filename = basen + '.dump
'
867 filename = sanitize_filename(raw_filename, restricted=True)
868 self.to_screen('Saving request to
' + filename)
869 # Working around MAX_PATH limitation on Windows (see
870 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
871 if compat_os_name == 'nt
':
872 absfilepath = os.path.abspath(filename)
873 if len(absfilepath) > 259:
874 filename = '\\\\?
\\' + absfilepath
875 with open(filename, 'wb
') as outf:
876 outf.write(webpage_bytes)
879 content = webpage_bytes.decode(encoding, 'replace
')
881 content = webpage_bytes.decode('utf
-8', 'replace
')
883 self.__check_blocked(content)
887 def _download_webpage(
888 self, url_or_request, video_id, note=None, errnote=None,
889 fatal=True, tries=1, timeout=5, encoding=None, data=None,
890 headers={}, query={}, expected_status=None):
892 Return the data of the page as a string.
895 url_or_request -- plain text URL as a string or
896 a compat_urllib_request.Requestobject
897 video_id -- Video/playlist/item identifier (string)
900 note -- note printed before downloading (string)
901 errnote -- note printed in case of an error (string)
902 fatal -- flag denoting whether error should be considered fatal,
903 i.e. whether it should cause ExtractionError to be raised,
904 otherwise a warning will be reported and extraction continued
905 tries -- number of tries
906 timeout -- sleep interval between tries
907 encoding -- encoding for a page content decoding, guessed automatically
908 when not explicitly specified
909 data -- POST data (bytes)
910 headers -- HTTP headers (dict)
911 query -- URL query (dict)
912 expected_status -- allows to accept failed HTTP requests (non 2xx
913 status code) by explicitly specifying a set of accepted status
914 codes. Can be any of the following entities:
915 - an integer type specifying an exact failed status code to
917 - a list or a tuple of integer types specifying a list of
918 failed status codes to accept
919 - a callable accepting an actual failed status code and
920 returning True if it should be accepted
921 Note that this argument does not affect success status codes (2xx)
922 which are always accepted.
927 while success is False:
929 res = self._download_webpage_handle(
930 url_or_request, video_id, note, errnote, fatal,
931 encoding=encoding, data=data, headers=headers, query=query,
932 expected_status=expected_status)
934 except compat_http_client.IncompleteRead as e:
936 if try_count >= tries:
938 self._sleep(timeout, video_id)
945 def _download_xml_handle(
946 self, url_or_request, video_id, note='Downloading XML
',
947 errnote='Unable to download XML
', transform_source=None,
948 fatal=True, encoding=None, data=None, headers={}, query={},
949 expected_status=None):
951 Return a tuple (xml as an compat_etree_Element, URL handle).
953 See _download_webpage docstring for arguments specification.
955 res = self._download_webpage_handle(
956 url_or_request, video_id, note, errnote, fatal=fatal,
957 encoding=encoding, data=data, headers=headers, query=query,
958 expected_status=expected_status)
961 xml_string, urlh = res
962 return self._parse_xml(
963 xml_string, video_id, transform_source=transform_source,
967 self, url_or_request, video_id,
968 note='Downloading XML
', errnote='Unable to download XML
',
969 transform_source=None, fatal=True, encoding=None,
970 data=None, headers={}, query={}, expected_status=None):
972 Return the xml as an compat_etree_Element.
974 See _download_webpage docstring for arguments specification.
976 res = self._download_xml_handle(
977 url_or_request, video_id, note=note, errnote=errnote,
978 transform_source=transform_source, fatal=fatal, encoding=encoding,
979 data=data, headers=headers, query=query,
980 expected_status=expected_status)
981 return res if res is False else res[0]
983 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
985 xml_string = transform_source(xml_string)
987 return compat_etree_fromstring(xml_string.encode('utf
-8'))
988 except compat_xml_parse_error as ve:
989 errmsg = '%s: Failed to parse XML
' % video_id
991 raise ExtractorError(errmsg, cause=ve)
993 self.report_warning(errmsg + str(ve))
995 def _download_json_handle(
996 self, url_or_request, video_id, note='Downloading JSON metadata
',
997 errnote='Unable to download JSON metadata
', transform_source=None,
998 fatal=True, encoding=None, data=None, headers={}, query={},
999 expected_status=None):
1001 Return a tuple (JSON object, URL handle).
1003 See _download_webpage docstring for arguments specification.
1005 res = self._download_webpage_handle(
1006 url_or_request, video_id, note, errnote, fatal=fatal,
1007 encoding=encoding, data=data, headers=headers, query=query,
1008 expected_status=expected_status)
1011 json_string, urlh = res
1012 return self._parse_json(
1013 json_string, video_id, transform_source=transform_source,
1017 self, url_or_request, video_id, note='Downloading JSON metadata
',
1018 errnote='Unable to download JSON metadata
', transform_source=None,
1019 fatal=True, encoding=None, data=None, headers={}, query={},
1020 expected_status=None):
1022 Return the JSON object as a dict.
1024 See _download_webpage docstring for arguments specification.
1026 res = self._download_json_handle(
1027 url_or_request, video_id, note=note, errnote=errnote,
1028 transform_source=transform_source, fatal=fatal, encoding=encoding,
1029 data=data, headers=headers, query=query,
1030 expected_status=expected_status)
1031 return res if res is False else res[0]
1033 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
1034 if transform_source:
1035 json_string = transform_source(json_string)
1037 return json.loads(json_string, strict=False)
1038 except ValueError as ve:
1039 errmsg = '%s: Failed to parse JSON
' % video_id
1041 raise ExtractorError(errmsg, cause=ve)
1043 self.report_warning(errmsg + str(ve))
1045 def _parse_socket_response_as_json(self, data, video_id, transform_source=None, fatal=True):
1046 return self._parse_json(
1047 data[data.find('{'):data.rfind('}
') + 1],
1048 video_id, transform_source, fatal)
1050 def _download_socket_json_handle(
1051 self, url_or_request, video_id, note='Polling socket
',
1052 errnote='Unable to poll socket
', transform_source=None,
1053 fatal=True, encoding=None, data=None, headers={}, query={},
1054 expected_status=None):
1056 Return a tuple (JSON object, URL handle).
1058 See _download_webpage docstring for arguments specification.
1060 res = self._download_webpage_handle(
1061 url_or_request, video_id, note, errnote, fatal=fatal,
1062 encoding=encoding, data=data, headers=headers, query=query,
1063 expected_status=expected_status)
1067 return self._parse_socket_response_as_json(
1068 webpage, video_id, transform_source=transform_source,
1071 def _download_socket_json(
1072 self, url_or_request, video_id, note='Polling socket
',
1073 errnote='Unable to poll socket
', transform_source=None,
1074 fatal=True, encoding=None, data=None, headers={}, query={},
1075 expected_status=None):
1077 Return the JSON object as a dict.
1079 See _download_webpage docstring for arguments specification.
1081 res = self._download_socket_json_handle(
1082 url_or_request, video_id, note=note, errnote=errnote,
1083 transform_source=transform_source, fatal=fatal, encoding=encoding,
1084 data=data, headers=headers, query=query,
1085 expected_status=expected_status)
1086 return res if res is False else res[0]
1088 def report_warning(self, msg, video_id=None, *args, only_once=False, **kwargs):
1089 idstr = format_field(video_id, template='%s: ')
1090 msg = f'[{self.IE_NAME}
] {idstr}{msg}
'
1092 if f'WARNING
: {msg}
' in self._printed_messages:
1094 self._printed_messages.add(f'WARNING
: {msg}
')
1095 self._downloader.report_warning(msg, *args, **kwargs)
1097 def to_screen(self, msg, *args, **kwargs):
1098 """Print msg to screen, prefixing it with '[ie_name
]'"""
1099 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg), *args, **kwargs)
1101 def write_debug(self, msg, *args, **kwargs):
1102 self._downloader.write_debug('[%s] %s' % (self.IE_NAME, msg), *args, **kwargs)
1104 def get_param(self, name, default=None, *args, **kwargs):
1105 if self._downloader:
1106 return self._downloader.params.get(name, default, *args, **kwargs)
1109 def report_drm(self, video_id, partial=False):
1110 self.raise_no_formats('This video
is DRM protected
', expected=True, video_id=video_id)
1112 def report_extraction(self, id_or_name):
1113 """Report information extraction."""
1114 self.to_screen('%s: Extracting information
' % id_or_name)
1116 def report_download_webpage(self, video_id):
1117 """Report webpage download."""
1118 self.to_screen('%s: Downloading webpage
' % video_id)
1120 def report_age_confirmation(self):
1121 """Report attempt to confirm age."""
1122 self.to_screen('Confirming age
')
1124 def report_login(self):
1125 """Report attempt to log in."""
1126 self.to_screen('Logging
in')
1128 def raise_login_required(
1129 self, msg='This video
is only available
for registered users
',
1130 metadata_available=False, method=NO_DEFAULT):
1131 if metadata_available and (
1132 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1133 self.report_warning(msg)
1135 if method is NO_DEFAULT:
1136 method = 'any
' if self.supports_login() else 'cookies
'
1137 if method is not None:
1138 assert method in self._LOGIN_HINTS, 'Invalid login method
'
1139 msg = '%s. %s' % (msg, self._LOGIN_HINTS[method])
1140 raise ExtractorError(msg, expected=True)
1142 def raise_geo_restricted(
1143 self, msg='This video
is not available
from your location due to geo restriction
',
1144 countries=None, metadata_available=False):
1145 if metadata_available and (
1146 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1147 self.report_warning(msg)
1149 raise GeoRestrictedError(msg, countries=countries)
1151 def raise_no_formats(self, msg, expected=False, video_id=None):
1153 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1154 self.report_warning(msg, video_id)
1155 elif isinstance(msg, ExtractorError):
1158 raise ExtractorError(msg, expected=expected, video_id=video_id)
1160 # Methods for following #608
1162 def url_result(url, ie=None, video_id=None, video_title=None, *, url_transparent=False, **kwargs):
1163 """Returns a URL that points to a page that should be processed"""
1165 kwargs['ie_key
'] = ie if isinstance(ie, str) else ie.ie_key()
1166 if video_id is not None:
1167 kwargs['id'] = video_id
1168 if video_title is not None:
1169 kwargs['title
'] = video_title
1172 '_type
': 'url_transparent
' if url_transparent else 'url
',
1176 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None, video_kwargs=None, **kwargs):
1177 urls = (self.url_result(self._proto_relative_url(m), ie, **(video_kwargs or {}))
1178 for m in orderedSet(map(getter, matches) if getter else matches))
1179 return self.playlist_result(urls, playlist_id, playlist_title, **kwargs)
1182 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None, *, multi_video=False, **kwargs):
1183 """Returns a playlist"""
1185 kwargs['id'] = playlist_id
1187 kwargs['title
'] = playlist_title
1188 if playlist_description is not None:
1189 kwargs['description
'] = playlist_description
1192 '_type
': 'multi_video
' if multi_video else 'playlist
',
1196 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1198 Perform a regex search on the given string, using a single or a list of
1199 patterns returning the first matching group.
1200 In case of failure return a default value or raise a WARNING or a
1201 RegexNotFoundError, depending on fatal, specifying the field name.
1203 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
1204 mobj = re.search(pattern, string, flags)
1207 mobj = re.search(p, string, flags)
1211 _name = self._downloader._format_err(name, self._downloader.Styles.EMPHASIS)
1215 # return the first matching group
1216 return next(g for g in mobj.groups() if g is not None)
1217 elif isinstance(group, (list, tuple)):
1218 return tuple(mobj.group(g) for g in group)
1220 return mobj.group(group)
1221 elif default is not NO_DEFAULT:
1224 raise RegexNotFoundError('Unable to extract
%s' % _name)
1226 self.report_warning('unable to extract
%s' % _name + bug_reports_message())
1229 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1231 Like _search_regex, but strips HTML tags and unescapes entities.
1233 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1235 return clean_html(res).strip()
1239 def _get_netrc_login_info(self, netrc_machine=None):
1242 netrc_machine = netrc_machine or self._NETRC_MACHINE
1244 if self.get_param('usenetrc
', False):
1246 netrc_file = compat_expanduser(self.get_param('netrc_location
') or '~
')
1247 if os.path.isdir(netrc_file):
1248 netrc_file = os.path.join(netrc_file, '.netrc
')
1249 info = netrc.netrc(file=netrc_file).authenticators(netrc_machine)
1250 if info is not None:
1254 raise netrc.NetrcParseError(
1255 'No authenticators
for %s' % netrc_machine)
1256 except (IOError, netrc.NetrcParseError) as err:
1257 self.report_warning(
1258 'parsing
.netrc
: %s' % error_to_compat_str(err))
1260 return username, password
1262 def _get_login_info(self, username_option='username
', password_option='password
', netrc_machine=None):
1264 Get the login info as (username, password)
1265 First look for the manually specified credentials using username_option
1266 and password_option as keys in params dictionary. If no such credentials
1267 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1269 If there's no info available
, return (None, None)
1272 # Attempt to use provided username and password or .netrc data
1273 username = self.get_param(username_option)
1274 if username is not None:
1275 password = self.get_param(password_option)
1277 username, password = self._get_netrc_login_info(netrc_machine)
1279 return username, password
1281 def _get_tfa_info(self, note='two-factor verification code'):
1283 Get the two
-factor authentication info
1284 TODO
- asking the user will be required
for sms
/phone verify
1285 currently just uses the command line option
1286 If there
's no info available, return None
1289 tfa = self.get_param('twofactor
')
1293 return compat_getpass('Type
%s and press
[Return
]: ' % note)
1295 # Helper functions for extracting OpenGraph info
1297 def _og_regexes(prop):
1298 content_re = r'content
=(?
:"([^"]+?
)"|\'([^\']+?)\'|\s*([^\s"\'=<>`
]+?
))'
1299 property_re = (r'(?
:name|
property)=(?
:\'og
[:-]%(prop)s\'|
"og[:-]%(prop)s"|\s
*og
[:-]%(prop)s\b)'
1300 % {'prop': re.escape(prop)})
1301 template = r'<meta
[^
>]+?
%s[^
>]+?
%s'
1303 template % (property_re, content_re),
1304 template % (content_re, property_re),
1308 def _meta_regex(prop):
1309 return r'''(?isx)<meta
1310 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1311 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1313 def _og_search_property(self, prop, html, name=None, **kargs):
1314 prop = variadic(prop)
1316 name = 'OpenGraph
%s' % prop[0]
1319 og_regexes.extend(self._og_regexes(p))
1320 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1323 return unescapeHTML(escaped)
1325 def _og_search_thumbnail(self, html, **kargs):
1326 return self._og_search_property('image
', html, 'thumbnail URL
', fatal=False, **kargs)
1328 def _og_search_description(self, html, **kargs):
1329 return self._og_search_property('description
', html, fatal=False, **kargs)
1331 def _og_search_title(self, html, **kargs):
1332 kargs.setdefault('fatal
', False)
1333 return self._og_search_property('title
', html, **kargs)
1335 def _og_search_video_url(self, html, name='video url
', secure=True, **kargs):
1336 regexes = self._og_regexes('video
') + self._og_regexes('video
:url
')
1338 regexes = self._og_regexes('video
:secure_url
') + regexes
1339 return self._html_search_regex(regexes, html, name, **kargs)
1341 def _og_search_url(self, html, **kargs):
1342 return self._og_search_property('url
', html, **kargs)
1344 def _html_extract_title(self, html, name, **kwargs):
1345 return self._html_search_regex(
1346 r'(?s
)<title
>(.*?
)</title
>', html, name, **kwargs)
1348 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1349 name = variadic(name)
1350 if display_name is None:
1351 display_name = name[0]
1352 return self._html_search_regex(
1353 [self._meta_regex(n) for n in name],
1354 html, display_name, fatal=fatal, group='content
', **kwargs)
1356 def _dc_search_uploader(self, html):
1357 return self._html_search_meta('dc
.creator
', html, 'uploader
')
1359 def _rta_search(self, html):
1360 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1361 if re.search(r'(?ix
)<meta\s
+name
="rating"\s
+'
1362 r' content
="RTA-5042-1996-1400-1577-RTA"',
1367 def _media_rating_search(self, html):
1368 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1369 rating = self._html_search_meta('rating
', html)
1381 return RATING_TABLE.get(rating.lower())
1383 def _family_friendly_search(self, html):
1384 # See http://schema.org/VideoObject
1385 family_friendly = self._html_search_meta(
1386 'isFamilyFriendly
', html, default=None)
1388 if not family_friendly:
1397 return RATING_TABLE.get(family_friendly.lower())
1399 def _twitter_search_player(self, html):
1400 return self._html_search_meta('twitter
:player
', html,
1401 'twitter card player
')
1403 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
1404 json_ld_list = list(re.finditer(JSON_LD_RE, html))
1405 default = kwargs.get('default
', NO_DEFAULT)
1406 # JSON-LD may be malformed and thus `fatal` should be respected.
1407 # At the same time `default` may be passed that assumes `fatal=False`
1408 # for _search_regex. Let's simulate the same behavior here
as well
.
1409 fatal
= kwargs
.get('fatal', True) if default
is NO_DEFAULT
else False
1411 for mobj
in json_ld_list
:
1412 json_ld_item
= self
._parse
_json
(
1413 mobj
.group('json_ld'), video_id
, fatal
=fatal
)
1414 if not json_ld_item
:
1416 if isinstance(json_ld_item
, dict):
1417 json_ld
.append(json_ld_item
)
1418 elif isinstance(json_ld_item
, (list, tuple)):
1419 json_ld
.extend(json_ld_item
)
1421 json_ld
= self
._json
_ld
(json_ld
, video_id
, fatal
=fatal
, expected_type
=expected_type
)
1424 if default
is not NO_DEFAULT
:
1427 raise RegexNotFoundError('Unable to extract JSON-LD')
1429 self
.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1432 def _json_ld(self
, json_ld
, video_id
, fatal
=True, expected_type
=None):
1433 if isinstance(json_ld
, compat_str
):
1434 json_ld
= self
._parse
_json
(json_ld
, video_id
, fatal
=fatal
)
1438 if not isinstance(json_ld
, (list, tuple, dict)):
1440 if isinstance(json_ld
, dict):
1443 INTERACTION_TYPE_MAP
= {
1444 'CommentAction': 'comment',
1445 'AgreeAction': 'like',
1446 'DisagreeAction': 'dislike',
1447 'LikeAction': 'like',
1448 'DislikeAction': 'dislike',
1449 'ListenAction': 'view',
1450 'WatchAction': 'view',
1451 'ViewAction': 'view',
1454 def extract_interaction_type(e
):
1455 interaction_type
= e
.get('interactionType')
1456 if isinstance(interaction_type
, dict):
1457 interaction_type
= interaction_type
.get('@type')
1458 return str_or_none(interaction_type
)
1460 def extract_interaction_statistic(e
):
1461 interaction_statistic
= e
.get('interactionStatistic')
1462 if isinstance(interaction_statistic
, dict):
1463 interaction_statistic
= [interaction_statistic
]
1464 if not isinstance(interaction_statistic
, list):
1466 for is_e
in interaction_statistic
:
1467 if not isinstance(is_e
, dict):
1469 if is_e
.get('@type') != 'InteractionCounter':
1471 interaction_type
= extract_interaction_type(is_e
)
1472 if not interaction_type
:
1474 # For interaction count some sites provide string instead of
1475 # an integer (as per spec) with non digit characters (e.g. ",")
1476 # so extracting count with more relaxed str_to_int
1477 interaction_count
= str_to_int(is_e
.get('userInteractionCount'))
1478 if interaction_count
is None:
1480 count_kind
= INTERACTION_TYPE_MAP
.get(interaction_type
.split('/')[-1])
1483 count_key
= '%s_count' % count_kind
1484 if info
.get(count_key
) is not None:
1486 info
[count_key
] = interaction_count
1488 def extract_chapter_information(e
):
1490 'title': part
.get('name'),
1491 'start_time': part
.get('startOffset'),
1492 'end_time': part
.get('endOffset'),
1493 } for part
in variadic(e
.get('hasPart') or []) if part
.get('@type') == 'Clip']
1494 for idx
, (last_c
, current_c
, next_c
) in enumerate(zip(
1495 [{'end_time': 0}
] + chapters
, chapters
, chapters
[1:])):
1496 current_c
['end_time'] = current_c
['end_time'] or next_c
['start_time']
1497 current_c
['start_time'] = current_c
['start_time'] or last_c
['end_time']
1498 if None in current_c
.values():
1499 self
.report_warning(f
'Chapter {idx} contains broken data. Not extracting chapters')
1502 chapters
[-1]['end_time'] = chapters
[-1]['end_time'] or info
['duration']
1503 info
['chapters'] = chapters
1505 def extract_video_object(e
):
1506 assert e
['@type'] == 'VideoObject'
1507 author
= e
.get('author')
1509 'url': url_or_none(e
.get('contentUrl')),
1510 'title': unescapeHTML(e
.get('name')),
1511 'description': unescapeHTML(e
.get('description')),
1512 'thumbnails': [{'url': url_or_none(url)}
1513 for url
in variadic(traverse_obj(e
, 'thumbnailUrl', 'thumbnailURL'))],
1514 'duration': parse_duration(e
.get('duration')),
1515 'timestamp': unified_timestamp(e
.get('uploadDate')),
1516 # author can be an instance of 'Organization' or 'Person' types.
1517 # both types can have 'name' property(inherited from 'Thing' type). [1]
1518 # however some websites are using 'Text' type instead.
1519 # 1. https://schema.org/VideoObject
1520 'uploader': author
.get('name') if isinstance(author
, dict) else author
if isinstance(author
, compat_str
) else None,
1521 'filesize': float_or_none(e
.get('contentSize')),
1522 'tbr': int_or_none(e
.get('bitrate')),
1523 'width': int_or_none(e
.get('width')),
1524 'height': int_or_none(e
.get('height')),
1525 'view_count': int_or_none(e
.get('interactionCount')),
1527 extract_interaction_statistic(e
)
1528 extract_chapter_information(e
)
1530 def traverse_json_ld(json_ld
, at_top_level
=True):
1532 if at_top_level
and '@context' not in e
:
1534 if at_top_level
and set(e
.keys()) == {'@context', '@graph'}
:
1535 traverse_json_ld(variadic(e
['@graph'], allowed_types
=(dict,)), at_top_level
=False)
1537 item_type
= e
.get('@type')
1538 if expected_type
is not None and expected_type
!= item_type
:
1540 rating
= traverse_obj(e
, ('aggregateRating', 'ratingValue'), expected_type
=float_or_none
)
1541 if rating
is not None:
1542 info
['average_rating'] = rating
1543 if item_type
in ('TVEpisode', 'Episode'):
1544 episode_name
= unescapeHTML(e
.get('name'))
1546 'episode': episode_name
,
1547 'episode_number': int_or_none(e
.get('episodeNumber')),
1548 'description': unescapeHTML(e
.get('description')),
1550 if not info
.get('title') and episode_name
:
1551 info
['title'] = episode_name
1552 part_of_season
= e
.get('partOfSeason')
1553 if isinstance(part_of_season
, dict) and part_of_season
.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1555 'season': unescapeHTML(part_of_season
.get('name')),
1556 'season_number': int_or_none(part_of_season
.get('seasonNumber')),
1558 part_of_series
= e
.get('partOfSeries') or e
.get('partOfTVSeries')
1559 if isinstance(part_of_series
, dict) and part_of_series
.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1560 info
['series'] = unescapeHTML(part_of_series
.get('name'))
1561 elif item_type
== 'Movie':
1563 'title': unescapeHTML(e
.get('name')),
1564 'description': unescapeHTML(e
.get('description')),
1565 'duration': parse_duration(e
.get('duration')),
1566 'timestamp': unified_timestamp(e
.get('dateCreated')),
1568 elif item_type
in ('Article', 'NewsArticle'):
1570 'timestamp': parse_iso8601(e
.get('datePublished')),
1571 'title': unescapeHTML(e
.get('headline')),
1572 'description': unescapeHTML(e
.get('articleBody') or e
.get('description')),
1574 if traverse_obj(e
, ('video', 0, '@type')) == 'VideoObject':
1575 extract_video_object(e
['video'][0])
1576 elif item_type
== 'VideoObject':
1577 extract_video_object(e
)
1578 if expected_type
is None:
1582 video
= e
.get('video')
1583 if isinstance(video
, dict) and video
.get('@type') == 'VideoObject':
1584 extract_video_object(video
)
1585 if expected_type
is None:
1589 traverse_json_ld(json_ld
)
1591 return dict((k
, v
) for k
, v
in info
.items() if v
is not None)
1593 def _search_nextjs_data(self
, webpage
, video_id
, *, transform_source
=None, fatal
=True, **kw
):
1594 return self
._parse
_json
(
1596 r
'(?s)<script[^>]+id=[\'"]__NEXT_DATA__[\'"][^
>]*>([^
<]+)</script
>',
1597 webpage, 'next
.js data
', fatal=fatal, **kw),
1598 video_id, transform_source=transform_source, fatal=fatal)
1600 def _search_nuxt_data(self, webpage, video_id, context_name='__NUXT__
'):
1601 ''' Parses Nuxt.js metadata. This works as long as the function __NUXT__ invokes is a pure function. '''
1602 # not all website do this, but it can be changed
1603 # https://stackoverflow.com/questions/67463109/how-to-change-or-hide-nuxt-and-nuxt-keyword-in-page-source
1604 rectx = re.escape(context_name)
1605 js, arg_keys, arg_vals = self._search_regex(
1606 (r'<script
>window\
.%s=\
(function\
((?P
<arg_keys
>.*?
)\
)\{return\s(?P<js>\{.*?\}
)\
}\
((?P
<arg_vals
>.+?
)\
)\
);?
</script
>' % rectx,
1607 r'%s\
(.*?\
(function\
((?P
<arg_keys
>.*?
)\
)\{return\s(?P<js>\{.*?\}
)\
}\
((?P
<arg_vals
>.*?
)\
)' % rectx),
1608 webpage, context_name, group=['js
', 'arg_keys
', 'arg_vals
'])
1610 args = dict(zip(arg_keys.split(','), arg_vals.split(',')))
1612 for key, val in args.items():
1613 if val in ('undefined
', 'void
0'):
1616 return self._parse_json(js_to_json(js, args), video_id)['data
'][0]
1619 def _hidden_inputs(html):
1620 html = re.sub(r'<!--(?
:(?
!<!--).)*-->', '', html)
1622 for input in re.findall(r'(?i
)(<input[^
>]+>)', html):
1623 attrs = extract_attributes(input)
1626 if attrs.get('type') not in ('hidden
', 'submit
'):
1628 name = attrs.get('name
') or attrs.get('id')
1629 value = attrs.get('value
')
1630 if name and value is not None:
1631 hidden_inputs[name] = value
1632 return hidden_inputs
1634 def _form_hidden_inputs(self, form_id, html):
1635 form = self._search_regex(
1636 r'(?
is)<form
[^
>]+?
id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1637 html, '%s form' % form_id, group='form')
1638 return self._hidden_inputs(form)
1641 regex = r' *((?P<reverse>\+)?(?P<field>[a-zA-Z0-9_]+)((?P<separator>[~:])(?P<limit>.*?))?)? *$'
1643 default = ('hidden', 'aud_or_vid', 'hasvid', 'ie_pref', 'lang', 'quality',
1644 'res', 'fps', 'hdr:12', 'codec:vp9.2', 'size', 'br', 'asr',
1645 'proto', 'ext', 'hasaud', 'source', 'id') # These must not be aliases
1646 ytdl_default = ('hasaud', 'lang', 'quality', 'tbr', 'filesize', 'vbr',
1647 'height', 'width', 'proto', 'vext', 'abr', 'aext',
1648 'fps', 'fs_approx', 'source', 'id')
1651 'vcodec': {'type': 'ordered', 'regex': True,
1652 'order': ['av0?1', 'vp0?9.2', 'vp0?9', '[hx]265|he?vc?', '[hx]264|avc', 'vp0?8', 'mp4v|h263', 'theora', '', None, 'none']},
1653 'acodec': {'type': 'ordered', 'regex': True,
1654 'order': ['[af]lac', 'wav|aiff', 'opus', 'vorbis|ogg', 'aac', 'mp?4a?', 'mp3', 'e-?a?c-?3', 'ac-?3', 'dts', '', None, 'none']},
1655 'hdr': {'type': 'ordered', 'regex': True, 'field': 'dynamic_range',
1656 'order': ['dv', '(hdr)?12', r'(hdr)?10\+', '(hdr)?10', 'hlg', '', 'sdr', None]},
1657 'proto': {'type': 'ordered', 'regex': True, 'field': 'protocol',
1658 'order': ['(ht|f)tps', '(ht|f)tp$', 'm3u8.*', '.*dash', 'websocket_frag', 'rtmpe?', '', 'mms|rtsp', 'ws|websocket', 'f4']},
1659 'vext': {'type': 'ordered', 'field': 'video_ext',
1660 'order': ('mp4', 'webm', 'flv', '', 'none'),
1661 'order_free': ('webm', 'mp4', 'flv', '', 'none')},
1662 'aext': {'type': 'ordered', 'field': 'audio_ext',
1663 'order': ('m4a', 'aac', 'mp3', 'ogg', 'opus', 'webm', '', 'none'),
1664 'order_free': ('opus', 'ogg', 'webm', 'm4a', 'mp3', 'aac', '', 'none')},
1665 'hidden': {'visible': False, 'forced': True, 'type': 'extractor', 'max': -1000},
1666 'aud_or_vid': {'visible': False, 'forced': True, 'type': 'multiple',
1667 'field': ('vcodec', 'acodec'),
1668 'function': lambda it: int(any(v != 'none' for v in it))},
1669 'ie_pref': {'priority': True, 'type': 'extractor'},
1670 'hasvid': {'priority': True, 'field': 'vcodec', 'type': 'boolean', 'not_in_list': ('none',)},
1671 'hasaud': {'field': 'acodec', 'type': 'boolean', 'not_in_list': ('none',)},
1672 'lang': {'convert': 'float', 'field': 'language_preference', 'default': -1},
1673 'quality': {'convert': 'float', 'default': -1},
1674 'filesize': {'convert': 'bytes'},
1675 'fs_approx': {'convert': 'bytes', 'field': 'filesize_approx'},
1676 'id': {'convert': 'string', 'field': 'format_id'},
1677 'height': {'convert': 'float_none'},
1678 'width': {'convert': 'float_none'},
1679 'fps': {'convert': 'float_none'},
1680 'tbr': {'convert': 'float_none'},
1681 'vbr': {'convert': 'float_none'},
1682 'abr': {'convert': 'float_none'},
1683 'asr': {'convert': 'float_none'},
1684 'source': {'convert': 'float', 'field': 'source_preference', 'default': -1},
1686 'codec': {'type': 'combined', 'field': ('vcodec', 'acodec')},
1687 'br': {'type': 'combined', 'field': ('tbr', 'vbr', 'abr'), 'same_limit': True},
1688 'size': {'type': 'combined', 'same_limit': True, 'field': ('filesize', 'fs_approx')},
1689 'ext': {'type': 'combined', 'field': ('vext', 'aext')},
1690 'res': {'type': 'multiple', 'field': ('height', 'width'),
1691 'function': lambda it: (lambda l: min(l) if l else 0)(tuple(filter(None, it)))},
1693 # For compatibility with youtube-dl
1694 'format_id': {'type': 'alias', 'field': 'id'},
1695 'preference': {'type': 'alias', 'field': 'ie_pref'},
1696 'language_preference': {'type': 'alias', 'field': 'lang'},
1697 'source_preference': {'type': 'alias', 'field': 'source'},
1698 'protocol': {'type': 'alias', 'field': 'proto'},
1699 'filesize_approx': {'type': 'alias', 'field': 'fs_approx'},
1702 'dimension': {'type': 'alias', 'field': 'res', 'deprecated': True},
1703 'resolution': {'type': 'alias', 'field': 'res', 'deprecated': True},
1704 'extension': {'type': 'alias', 'field': 'ext', 'deprecated': True},
1705 'bitrate': {'type': 'alias', 'field': 'br', 'deprecated': True},
1706 'total_bitrate': {'type': 'alias', 'field': 'tbr', 'deprecated': True},
1707 'video_bitrate': {'type': 'alias', 'field': 'vbr', 'deprecated': True},
1708 'audio_bitrate': {'type': 'alias', 'field': 'abr', 'deprecated': True},
1709 'framerate': {'type': 'alias', 'field': 'fps', 'deprecated': True},
1710 'filesize_estimate': {'type': 'alias', 'field': 'size', 'deprecated': True},
1711 'samplerate': {'type': 'alias', 'field': 'asr', 'deprecated': True},
1712 'video_ext': {'type': 'alias', 'field': 'vext', 'deprecated': True},
1713 'audio_ext': {'type': 'alias', 'field': 'aext', 'deprecated': True},
1714 'video_codec': {'type': 'alias', 'field': 'vcodec', 'deprecated': True},
1715 'audio_codec': {'type': 'alias', 'field': 'acodec', 'deprecated': True},
1716 'video': {'type': 'alias', 'field': 'hasvid', 'deprecated': True},
1717 'has_video': {'type': 'alias', 'field': 'hasvid', 'deprecated': True},
1718 'audio': {'type': 'alias', 'field': 'hasaud', 'deprecated': True},
1719 'has_audio': {'type': 'alias', 'field': 'hasaud', 'deprecated': True},
1720 'extractor': {'type': 'alias', 'field': 'ie_pref', 'deprecated': True},
1721 'extractor_preference': {'type': 'alias', 'field': 'ie_pref', 'deprecated': True},
1724 def __init__(self, ie, field_preference):
1726 self.ydl = ie._downloader
1727 self.evaluate_params(self.ydl.params, field_preference)
1728 if ie.get_param('verbose'):
1729 self.print_verbose_info(self.ydl.write_debug)
1731 def _get_field_setting(self, field, key):
1732 if field not in self.settings:
1733 if key in ('forced', 'priority'):
1735 self.ydl.deprecation_warning(
1736 f'Using arbitrary fields ({field}) for format sorting is deprecated '
1737 'and may be removed in a future version')
1738 self.settings[field] = {}
1739 propObj = self.settings[field]
1740 if key not in propObj:
1741 type = propObj.get('type')
1743 default = 'preference' if type == 'extractor' else (field,) if type in ('combined', 'multiple') else field
1744 elif key == 'convert':
1745 default = 'order' if type == 'ordered' else 'float_string' if field else 'ignore'
1747 default = {'type': 'field', 'visible': True, 'order': [], 'not_in_list': (None,)}.get(key, None)
1748 propObj[key] = default
1751 def _resolve_field_value(self, field, value, convertNone=False):
1756 value = value.lower()
1757 conversion = self._get_field_setting(field, 'convert')
1758 if conversion == 'ignore':
1760 if conversion == 'string':
1762 elif conversion == 'float_none':
1763 return float_or_none(value)
1764 elif conversion == 'bytes':
1765 return FileDownloader.parse_bytes(value)
1766 elif conversion == 'order':
1767 order_list = (self._use_free_order and self._get_field_setting(field, 'order_free')) or self._get_field_setting(field, 'order')
1768 use_regex = self._get_field_setting(field, 'regex')
1769 list_length = len(order_list)
1770 empty_pos = order_list.index('') if '' in order_list else list_length + 1
1771 if use_regex and value is not None:
1772 for i, regex in enumerate(order_list):
1773 if regex and re.match(regex, value):
1774 return list_length - i
1775 return list_length - empty_pos # not in list
1776 else: # not regex or value = None
1777 return list_length - (order_list.index(value) if value in order_list else empty_pos)
1779 if value.isnumeric():
1782 self.settings[field]['convert'] = 'string'
1785 def evaluate_params(self, params, sort_extractor):
1786 self._use_free_order = params.get('prefer_free_formats', False)
1787 self._sort_user = params.get('format_sort', [])
1788 self._sort_extractor = sort_extractor
1790 def add_item(field, reverse, closest, limit_text):
1791 field = field.lower()
1792 if field in self._order:
1794 self._order.append(field)
1795 limit = self._resolve_field_value(field, limit_text)
1798 'closest': False if limit is None else closest,
1799 'limit_text': limit_text,
1801 if field in self.settings:
1802 self.settings[field].update(data)
1804 self.settings[field] = data
1807 tuple(field for field in self.default if self._get_field_setting(field, 'forced'))
1808 + (tuple() if params.get('format_sort_force', False)
1809 else tuple(field for field in self.default if self._get_field_setting(field, 'priority')))
1810 + tuple(self._sort_user) + tuple(sort_extractor) + self.default)
1812 for item in sort_list:
1813 match = re.match(self.regex, item)
1815 raise ExtractorError('Invalid format sort string "%s" given by extractor' % item)
1816 field = match.group('field')
1819 if self._get_field_setting(field, 'type') == 'alias':
1820 alias, field = field, self._get_field_setting(field, 'field')
1821 if self._get_field_setting(alias, 'deprecated'):
1822 self.ydl.deprecation_warning(
1823 f'Format sorting alias {alias} is deprecated '
1824 f'and may be removed in a future version. Please use {field} instead')
1825 reverse = match.group('reverse') is not None
1826 closest = match.group('separator') == '~'
1827 limit_text = match.group('limit')
1829 has_limit = limit_text is not None
1830 has_multiple_fields = self._get_field_setting(field, 'type') == 'combined'
1831 has_multiple_limits = has_limit and has_multiple_fields and not self._get_field_setting(field, 'same_limit')
1833 fields = self._get_field_setting(field, 'field') if has_multiple_fields else (field,)
1834 limits = limit_text.split(':') if has_multiple_limits else (limit_text,) if has_limit else tuple()
1835 limit_count = len(limits)
1836 for (i, f) in enumerate(fields):
1837 add_item(f, reverse, closest,
1838 limits[i] if i < limit_count
1839 else limits[0] if has_limit and not has_multiple_limits
1842 def print_verbose_info(self, write_debug):
1844 write_debug('Sort order given by user: %s' % ', '.join(self._sort_user))
1845 if self._sort_extractor:
1846 write_debug('Sort order given by extractor: %s' % ', '.join(self._sort_extractor))
1847 write_debug('Formats sorted by: %s' % ', '.join(['%s%s%s' % (
1848 '+' if self._get_field_setting(field, 'reverse') else '', field,
1849 '%s%s(%s)' % ('~' if self._get_field_setting(field, 'closest') else ':',
1850 self._get_field_setting(field, 'limit_text'),
1851 self._get_field_setting(field, 'limit'))
1852 if self._get_field_setting(field, 'limit_text') is not None else '')
1853 for field in self._order if self._get_field_setting(field, 'visible')]))
1855 def _calculate_field_preference_from_value(self, format, field, type, value):
1856 reverse = self._get_field_setting(field, 'reverse')
1857 closest = self._get_field_setting(field, 'closest')
1858 limit = self._get_field_setting(field, 'limit')
1860 if type == 'extractor':
1861 maximum = self._get_field_setting(field, 'max')
1862 if value is None or (maximum is not None and value >= maximum):
1864 elif type == 'boolean':
1865 in_list = self._get_field_setting(field, 'in_list')
1866 not_in_list = self._get_field_setting(field, 'not_in_list')
1867 value = 0 if ((in_list is None or value in in_list) and (not_in_list is None or value not in not_in_list)) else -1
1868 elif type == 'ordered':
1869 value = self._resolve_field_value(field, value, True)
1871 # try to convert to number
1872 val_num = float_or_none(value, default=self._get_field_setting(field, 'default'))
1873 is_num = self._get_field_setting(field, 'convert') != 'string' and val_num is not None
1877 return ((-10, 0) if value is None
1878 else (1, value, 0) if not is_num # if a field has mixed strings and numbers, strings are sorted higher
1879 else (0, -abs(value - limit), value - limit if reverse else limit - value) if closest
1880 else (0, value, 0) if not reverse and (limit is None or value <= limit)
1881 else (0, -value, 0) if limit is None or (reverse and value == limit) or value > limit
1882 else (-1, value, 0))
1884 def _calculate_field_preference(self, format, field):
1885 type = self._get_field_setting(field, 'type') # extractor, boolean, ordered, field, multiple
1886 get_value = lambda f: format.get(self._get_field_setting(f, 'field'))
1887 if type == 'multiple':
1888 type = 'field' # Only 'field' is allowed in multiple for now
1889 actual_fields = self._get_field_setting(field, 'field')
1891 value = self._get_field_setting(field, 'function')(get_value(f) for f in actual_fields)
1893 value = get_value(field)
1894 return self._calculate_field_preference_from_value(format, field, type, value)
1896 def calculate_preference(self, format):
1897 # Determine missing protocol
1898 if not format.get('protocol'):
1899 format['protocol'] = determine_protocol(format)
1901 # Determine missing ext
1902 if not format.get('ext') and 'url' in format:
1903 format['ext'] = determine_ext(format['url'])
1904 if format.get('vcodec') == 'none':
1905 format['audio_ext'] = format['ext'] if format.get('acodec') != 'none' else 'none'
1906 format['video_ext'] = 'none'
1908 format['video_ext'] = format['ext']
1909 format['audio_ext'] = 'none'
1910 # if format.get('preference') is None and format.get('ext') in ('f4f', 'f4m'): # Not supported?
1911 # format['preference'] = -1000
1913 # Determine missing bitrates
1914 if format.get('tbr') is None:
1915 if format.get('vbr') is not None and format.get('abr') is not None:
1916 format['tbr'] = format.get('vbr', 0) + format.get('abr', 0)
1918 if format.get('vcodec') != 'none' and format.get('vbr') is None:
1919 format['vbr'] = format.get('tbr') - format.get('abr', 0)
1920 if format.get('acodec') != 'none' and format.get('abr') is None:
1921 format['abr'] = format.get('tbr') - format.get('vbr', 0)
1923 return tuple(self._calculate_field_preference(format, field) for field in self._order)
1925 def _sort_formats(self, formats, field_preference=[]):
1928 format_sort = self.FormatSort(self, field_preference)
1929 formats.sort(key=lambda f: format_sort.calculate_preference(f))
1931 def _check_formats(self, formats, video_id):
1933 formats[:] = filter(
1934 lambda f: self._is_valid_url(
1936 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1940 def _remove_duplicate_formats(formats):
1944 if f['url'] not in format_urls:
1945 format_urls.add(f['url'])
1946 unique_formats.append(f)
1947 formats[:] = unique_formats
1949 def _is_valid_url(self, url, video_id, item='video', headers={}):
1950 url = self._proto_relative_url(url, scheme='http:')
1951 # For now assume non HTTP(S) URLs always valid
1952 if not (url.startswith('http://') or url.startswith('https://')):
1955 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1957 except ExtractorError as e:
1959 '%s: %s URL is invalid, skipping: %s'
1960 % (video_id, item, error_to_compat_str(e.cause)))
1963 def http_scheme(self):
1964 """ Either "http
:" or "https
:", depending on the user's preferences """
1967 if self.get_param('prefer_insecure', False)
1970 def _proto_relative_url(self, url, scheme=None):
1973 if url.startswith('//'):
1975 scheme = self.http_scheme()
1980 def _sleep(self, timeout, video_id, msg_template=None):
1981 if msg_template is None:
1982 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1983 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1987 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1988 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1989 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
1990 manifest = self._download_xml(
1991 manifest_url, video_id, 'Downloading f4m manifest',
1992 'Unable to download f4m manifest',
1993 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1994 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
1995 transform_source=transform_source,
1996 fatal=fatal, data=data, headers=headers, query=query)
1998 if manifest is False:
2001 return self._parse_f4m_formats(
2002 manifest, manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
2003 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
2005 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
2006 transform_source=lambda s: fix_xml_ampersands(s).strip(),
2007 fatal=True, m3u8_id=None):
2008 if not isinstance(manifest, compat_etree_Element) and not fatal:
2011 # currently yt-dlp cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
2012 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
2013 if akamai_pv is not None and ';' in akamai_pv.text:
2014 playerVerificationChallenge = akamai_pv.text.split(';')[0]
2015 if playerVerificationChallenge.strip() != '':
2019 manifest_version = '1.0'
2020 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
2022 manifest_version = '2.0'
2023 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
2024 # Remove unsupported DRM protected media from final formats
2025 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
2026 media_nodes = remove_encrypted_media(media_nodes)
2030 manifest_base_url = get_base_url(manifest)
2032 bootstrap_info = xpath_element(
2033 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
2034 'bootstrap info', default=None)
2037 mime_type = xpath_text(
2038 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
2039 'base URL', default=None)
2040 if mime_type and mime_type.startswith('audio/'):
2043 for i, media_el in enumerate(media_nodes):
2044 tbr = int_or_none(media_el.attrib.get('bitrate'))
2045 width = int_or_none(media_el.attrib.get('width'))
2046 height = int_or_none(media_el.attrib.get('height'))
2047 format_id = join_nonempty(f4m_id, tbr or i)
2048 # If <bootstrapInfo> is present, the specified f4m is a
2049 # stream-level manifest, and only set-level manifests may refer to
2050 # external resources. See section 11.4 and section 4 of F4M spec
2051 if bootstrap_info is None:
2053 # @href is introduced in 2.0, see section 11.6 of F4M spec
2054 if manifest_version == '2.0':
2055 media_url = media_el.attrib.get('href')
2056 if media_url is None:
2057 media_url = media_el.attrib.get('url')
2061 media_url if media_url.startswith('http://') or media_url.startswith('https://')
2062 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
2063 # If media_url is itself a f4m manifest do the recursive extraction
2064 # since bitrates in parent manifest (this one) and media_url manifest
2065 # may differ leading to inability to resolve the format by requested
2066 # bitrate in f4m downloader
2067 ext = determine_ext(manifest_url)
2069 f4m_formats = self._extract_f4m_formats(
2070 manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
2071 transform_source=transform_source, fatal=fatal)
2072 # Sometimes stream-level manifest contains single media entry that
2073 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
2074 # At the same time parent's media entry in set-level manifest may
2075 # contain it. We will copy it from parent in such cases.
2076 if len(f4m_formats) == 1:
2079 'tbr': f.get('tbr') or tbr,
2080 'width': f.get('width') or width,
2081 'height': f.get('height') or height,
2082 'format_id': f.get('format_id') if not tbr else format_id,
2085 formats.extend(f4m_formats)
2088 formats.extend(self._extract_m3u8_formats(
2089 manifest_url, video_id, 'mp4', preference=preference,
2090 quality=quality, m3u8_id=m3u8_id, fatal=fatal))
2093 'format_id': format_id,
2094 'url': manifest_url,
2095 'manifest_url': manifest_url,
2096 'ext': 'flv' if bootstrap_info is not None else None,
2102 'preference': preference,
2107 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, quality=None, m3u8_id=None):
2109 'format_id': join_nonempty(m3u8_id, 'meta'),
2113 'preference': preference - 100 if preference else -100,
2115 'resolution': 'multiple',
2116 'format_note': 'Quality selection URL',
2119 def _report_ignoring_subs(self, name):
2120 self.report_warning(bug_reports_message(
2121 f'Ignoring subtitle tracks found in the {name} manifest; '
2122 'if any subtitle tracks are missing,'
2125 def _extract_m3u8_formats(self, *args, **kwargs):
2126 fmts, subs = self._extract_m3u8_formats_and_subtitles(*args, **kwargs)
2128 self._report_ignoring_subs('HLS')
2131 def _extract_m3u8_formats_and_subtitles(
2132 self, m3u8_url, video_id, ext=None, entry_protocol='m3u8_native',
2133 preference=None, quality=None, m3u8_id=None, note=None,
2134 errnote=None, fatal=True, live=False, data=None, headers={},
2137 res = self._download_webpage_handle(
2139 note='Downloading m3u8 information' if note is None else note,
2140 errnote='Failed to download m3u8 information' if errnote is None else errnote,
2141 fatal=fatal, data=data, headers=headers, query=query)
2146 m3u8_doc, urlh = res
2147 m3u8_url = urlh.geturl()
2149 return self._parse_m3u8_formats_and_subtitles(
2150 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
2151 preference=preference, quality=quality, m3u8_id=m3u8_id,
2152 note=note, errnote=errnote, fatal=fatal, live=live, data=data,
2153 headers=headers, query=query, video_id=video_id)
2155 def _parse_m3u8_formats_and_subtitles(
2156 self, m3u8_doc, m3u8_url=None, ext=None, entry_protocol='m3u8_native',
2157 preference=None, quality=None, m3u8_id=None, live=False, note=None,
2158 errnote=None, fatal=True, data=None, headers={}, query={},
2160 formats, subtitles = [], {}
2162 has_drm = re.search('|'.join([
2163 r'#EXT-X-FAXS-CM:', # Adobe Flash Access
2164 r'#EXT-X-(?:SESSION-)?KEY:.*?URI="skd
://', # Apple FairPlay
2167 def format_url(url):
2168 return url if re.match(r'^https?
://', url) else compat_urlparse.urljoin(m3u8_url, url)
2170 if self.get_param('hls_split_discontinuity
', False):
2171 def _extract_m3u8_playlist_indices(manifest_url=None, m3u8_doc=None):
2173 if not manifest_url:
2175 m3u8_doc = self._download_webpage(
2176 manifest_url, video_id, fatal=fatal, data=data, headers=headers,
2177 note=False, errnote='Failed to download m3u8 playlist information
')
2178 if m3u8_doc is False:
2180 return range(1 + sum(line.startswith('#EXT-X-DISCONTINUITY') for line in m3u8_doc.splitlines()))
2183 def _extract_m3u8_playlist_indices(*args
, **kwargs
):
2187 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
2188 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
2189 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
2191 # We should try extracting formats only from master playlists [1, 4.3.4],
2192 # i.e. playlists that describe available qualities. On the other hand
2193 # media playlists [1, 4.3.3] should be returned as is since they contain
2194 # just the media without qualities renditions.
2195 # Fortunately, master playlist can be easily distinguished from media
2196 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
2197 # master playlist tags MUST NOT appear in a media playlist and vice versa.
2198 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
2199 # media playlist and MUST NOT appear in master playlist thus we can
2200 # clearly detect media playlist with this criterion.
2202 if '#EXT-X-TARGETDURATION' in m3u8_doc
: # media playlist, return as is
2204 'format_id': join_nonempty(m3u8_id
, idx
),
2205 'format_index': idx
,
2206 'url': m3u8_url
or encode_data_uri(m3u8_doc
.encode('utf-8'), 'application/x-mpegurl'),
2208 'protocol': entry_protocol
,
2209 'preference': preference
,
2212 } for idx
in _extract_m3u8_playlist_indices(m3u8_doc
=m3u8_doc
)]
2214 return formats
, subtitles
2217 last_stream_inf
= {}
2219 def extract_media(x_media_line
):
2220 media
= parse_m3u8_attributes(x_media_line
)
2221 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
2222 media_type
, group_id
, name
= media
.get('TYPE'), media
.get('GROUP-ID'), media
.get('NAME')
2223 if not (media_type
and group_id
and name
):
2225 groups
.setdefault(group_id
, []).append(media
)
2226 # <https://tools.ietf.org/html/rfc8216#section-4.3.4.1>
2227 if media_type
== 'SUBTITLES':
2228 # According to RFC 8216 §4.3.4.2.1, URI is REQUIRED in the
2229 # EXT-X-MEDIA tag if the media type is SUBTITLES.
2230 # However, lack of URI has been spotted in the wild.
2231 # e.g. NebulaIE; see https://github.com/yt-dlp/yt-dlp/issues/339
2232 if not media
.get('URI'):
2234 url
= format_url(media
['URI'])
2237 'ext': determine_ext(url
),
2239 if sub_info
['ext'] == 'm3u8':
2240 # Per RFC 8216 §3.1, the only possible subtitle format m3u8
2241 # files may contain is WebVTT:
2242 # <https://tools.ietf.org/html/rfc8216#section-3.1>
2243 sub_info
['ext'] = 'vtt'
2244 sub_info
['protocol'] = 'm3u8_native'
2245 lang
= media
.get('LANGUAGE') or 'und'
2246 subtitles
.setdefault(lang
, []).append(sub_info
)
2247 if media_type
not in ('VIDEO', 'AUDIO'):
2249 media_url
= media
.get('URI')
2251 manifest_url
= format_url(media_url
)
2253 'format_id': join_nonempty(m3u8_id
, group_id
, name
, idx
),
2254 'format_note': name
,
2255 'format_index': idx
,
2256 'url': manifest_url
,
2257 'manifest_url': m3u8_url
,
2258 'language': media
.get('LANGUAGE'),
2260 'protocol': entry_protocol
,
2261 'preference': preference
,
2263 'vcodec': 'none' if media_type
== 'AUDIO' else None,
2264 } for idx
in _extract_m3u8_playlist_indices(manifest_url
))
2266 def build_stream_name():
2267 # Despite specification does not mention NAME attribute for
2268 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
2269 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
2270 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
2271 stream_name
= last_stream_inf
.get('NAME')
2274 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
2275 # from corresponding rendition group
2276 stream_group_id
= last_stream_inf
.get('VIDEO')
2277 if not stream_group_id
:
2279 stream_group
= groups
.get(stream_group_id
)
2280 if not stream_group
:
2281 return stream_group_id
2282 rendition
= stream_group
[0]
2283 return rendition
.get('NAME') or stream_group_id
2285 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
2286 # chance to detect video only formats when EXT-X-STREAM-INF tags
2287 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
2288 for line
in m3u8_doc
.splitlines():
2289 if line
.startswith('#EXT-X-MEDIA:'):
2292 for line
in m3u8_doc
.splitlines():
2293 if line
.startswith('#EXT-X-STREAM-INF:'):
2294 last_stream_inf
= parse_m3u8_attributes(line
)
2295 elif line
.startswith('#') or not line
.strip():
2298 tbr
= float_or_none(
2299 last_stream_inf
.get('AVERAGE-BANDWIDTH')
2300 or last_stream_inf
.get('BANDWIDTH'), scale
=1000)
2301 manifest_url
= format_url(line
.strip())
2303 for idx
in _extract_m3u8_playlist_indices(manifest_url
):
2304 format_id
= [m3u8_id
, None, idx
]
2305 # Bandwidth of live streams may differ over time thus making
2306 # format_id unpredictable. So it's better to keep provided
2309 stream_name
= build_stream_name()
2310 format_id
[1] = stream_name
or '%d' % (tbr
or len(formats
))
2312 'format_id': join_nonempty(*format_id
),
2313 'format_index': idx
,
2314 'url': manifest_url
,
2315 'manifest_url': m3u8_url
,
2318 'fps': float_or_none(last_stream_inf
.get('FRAME-RATE')),
2319 'protocol': entry_protocol
,
2320 'preference': preference
,
2323 resolution
= last_stream_inf
.get('RESOLUTION')
2325 mobj
= re
.search(r
'(?P<width>\d+)[xX](?P<height>\d+)', resolution
)
2327 f
['width'] = int(mobj
.group('width'))
2328 f
['height'] = int(mobj
.group('height'))
2329 # Unified Streaming Platform
2331 r
'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f
['url'])
2333 abr
, vbr
= mobj
.groups()
2334 abr
, vbr
= float_or_none(abr
, 1000), float_or_none(vbr
, 1000)
2339 codecs
= parse_codecs(last_stream_inf
.get('CODECS'))
2341 audio_group_id
= last_stream_inf
.get('AUDIO')
2342 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
2343 # references a rendition group MUST have a CODECS attribute.
2344 # However, this is not always respected, for example, [2]
2345 # contains EXT-X-STREAM-INF tag which references AUDIO
2346 # rendition group but does not have CODECS and despite
2347 # referencing an audio group it represents a complete
2348 # (with audio and video) format. So, for such cases we will
2349 # ignore references to rendition groups and treat them
2350 # as complete formats.
2351 if audio_group_id
and codecs
and f
.get('vcodec') != 'none':
2352 audio_group
= groups
.get(audio_group_id
)
2353 if audio_group
and audio_group
[0].get('URI'):
2354 # TODO: update acodec for audio only formats with
2356 f
['acodec'] = 'none'
2357 if not f
.get('ext'):
2358 f
['ext'] = 'm4a' if f
.get('vcodec') == 'none' else 'mp4'
2362 progressive_uri
= last_stream_inf
.get('PROGRESSIVE-URI')
2365 del http_f
['manifest_url']
2367 'format_id': f
['format_id'].replace('hls-', 'http-'),
2369 'url': progressive_uri
,
2371 formats
.append(http_f
)
2373 last_stream_inf
= {}
2374 return formats
, subtitles
2376 def _extract_m3u8_vod_duration(
2377 self
, m3u8_vod_url
, video_id
, note
=None, errnote
=None, data
=None, headers
={}, query={}
):
2379 m3u8_vod
= self
._download
_webpage
(
2380 m3u8_vod_url
, video_id
,
2381 note
='Downloading m3u8 VOD manifest' if note
is None else note
,
2382 errnote
='Failed to download VOD manifest' if errnote
is None else errnote
,
2383 fatal
=False, data
=data
, headers
=headers
, query
=query
)
2385 return self
._parse
_m
3u8_vod
_duration
(m3u8_vod
or '', video_id
)
2387 def _parse_m3u8_vod_duration(self
, m3u8_vod
, video_id
):
2388 if '#EXT-X-PLAYLIST-TYPE:VOD' not in m3u8_vod
:
2392 float(line
[len('#EXTINF:'):].split(',')[0])
2393 for line
in m3u8_vod
.splitlines() if line
.startswith('#EXTINF:'))) or None
2396 def _xpath_ns(path
, namespace
=None):
2400 for c
in path
.split('/'):
2401 if not c
or c
== '.':
2404 out
.append('{%s}%s' % (namespace
, c
))
2405 return '/'.join(out
)
2407 def _extract_smil_formats_and_subtitles(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None, transform_source
=None):
2408 smil
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
, transform_source
=transform_source
)
2414 namespace
= self
._parse
_smil
_namespace
(smil
)
2416 fmts
= self
._parse
_smil
_formats
(
2417 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2418 subs
= self
._parse
_smil
_subtitles
(
2419 smil
, namespace
=namespace
)
2423 def _extract_smil_formats(self
, *args
, **kwargs
):
2424 fmts
, subs
= self
._extract
_smil
_formats
_and
_subtitles
(*args
, **kwargs
)
2426 self
._report
_ignoring
_subs
('SMIL')
2429 def _extract_smil_info(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None):
2430 smil
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
)
2433 return self
._parse
_smil
(smil
, smil_url
, video_id
, f4m_params
=f4m_params
)
2435 def _download_smil(self
, smil_url
, video_id
, fatal
=True, transform_source
=None):
2436 return self
._download
_xml
(
2437 smil_url
, video_id
, 'Downloading SMIL file',
2438 'Unable to download SMIL file', fatal
=fatal
, transform_source
=transform_source
)
2440 def _parse_smil(self
, smil
, smil_url
, video_id
, f4m_params
=None):
2441 namespace
= self
._parse
_smil
_namespace
(smil
)
2443 formats
= self
._parse
_smil
_formats
(
2444 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2445 subtitles
= self
._parse
_smil
_subtitles
(smil
, namespace
=namespace
)
2447 video_id
= os
.path
.splitext(url_basename(smil_url
))[0]
2451 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2452 name
= meta
.attrib
.get('name')
2453 content
= meta
.attrib
.get('content')
2454 if not name
or not content
:
2456 if not title
and name
== 'title':
2458 elif not description
and name
in ('description', 'abstract'):
2459 description
= content
2460 elif not upload_date
and name
== 'date':
2461 upload_date
= unified_strdate(content
)
2464 'id': image
.get('type'),
2465 'url': image
.get('src'),
2466 'width': int_or_none(image
.get('width')),
2467 'height': int_or_none(image
.get('height')),
2468 } for image
in smil
.findall(self
._xpath
_ns
('.//image', namespace
)) if image
.get('src')]
2472 'title': title
or video_id
,
2473 'description': description
,
2474 'upload_date': upload_date
,
2475 'thumbnails': thumbnails
,
2477 'subtitles': subtitles
,
2480 def _parse_smil_namespace(self
, smil
):
2481 return self
._search
_regex
(
2482 r
'(?i)^{([^}]+)?}smil$', smil
.tag
, 'namespace', default
=None)
2484 def _parse_smil_formats(self
, smil
, smil_url
, video_id
, namespace
=None, f4m_params
=None, transform_rtmp_url
=None):
2486 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2487 b
= meta
.get('base') or meta
.get('httpBase')
2499 media
= smil
.findall(self
._xpath
_ns
('.//video', namespace
)) + smil
.findall(self
._xpath
_ns
('.//audio', namespace
))
2500 for medium
in media
:
2501 src
= medium
.get('src')
2502 if not src
or src
in srcs
:
2506 bitrate
= float_or_none(medium
.get('system-bitrate') or medium
.get('systemBitrate'), 1000)
2507 filesize
= int_or_none(medium
.get('size') or medium
.get('fileSize'))
2508 width
= int_or_none(medium
.get('width'))
2509 height
= int_or_none(medium
.get('height'))
2510 proto
= medium
.get('proto')
2511 ext
= medium
.get('ext')
2512 src_ext
= determine_ext(src
)
2513 streamer
= medium
.get('streamer') or base
2515 if proto
== 'rtmp' or streamer
.startswith('rtmp'):
2521 'format_id': 'rtmp-%d' % (rtmp_count
if bitrate
is None else bitrate
),
2523 'filesize': filesize
,
2527 if transform_rtmp_url
:
2528 streamer
, src
= transform_rtmp_url(streamer
, src
)
2529 formats
[-1].update({
2535 src_url
= src
if src
.startswith('http') else compat_urlparse
.urljoin(base
, src
)
2536 src_url
= src_url
.strip()
2538 if proto
== 'm3u8' or src_ext
== 'm3u8':
2539 m3u8_formats
= self
._extract
_m
3u8_formats
(
2540 src_url
, video_id
, ext
or 'mp4', m3u8_id
='hls', fatal
=False)
2541 if len(m3u8_formats
) == 1:
2543 m3u8_formats
[0].update({
2544 'format_id': 'hls-%d' % (m3u8_count
if bitrate
is None else bitrate
),
2549 formats
.extend(m3u8_formats
)
2550 elif src_ext
== 'f4m':
2555 'plugin': 'flowplayer-3.2.0.1',
2557 f4m_url
+= '&' if '?' in f4m_url
else '?'
2558 f4m_url
+= compat_urllib_parse_urlencode(f4m_params
)
2559 formats
.extend(self
._extract
_f
4m
_formats
(f4m_url
, video_id
, f4m_id
='hds', fatal
=False))
2560 elif src_ext
== 'mpd':
2561 formats
.extend(self
._extract
_mpd
_formats
(
2562 src_url
, video_id
, mpd_id
='dash', fatal
=False))
2563 elif re
.search(r
'\.ism/[Mm]anifest', src_url
):
2564 formats
.extend(self
._extract
_ism
_formats
(
2565 src_url
, video_id
, ism_id
='mss', fatal
=False))
2566 elif src_url
.startswith('http') and self
._is
_valid
_url
(src
, video_id
):
2570 'ext': ext
or src_ext
or 'flv',
2571 'format_id': 'http-%d' % (bitrate
or http_count
),
2573 'filesize': filesize
,
2578 for medium
in smil
.findall(self
._xpath
_ns
('.//imagestream', namespace
)):
2579 src
= medium
.get('src')
2580 if not src
or src
in srcs
:
2586 'format_id': 'imagestream-%d' % (imgs_count
),
2588 'ext': mimetype2ext(medium
.get('type')),
2591 'width': int_or_none(medium
.get('width')),
2592 'height': int_or_none(medium
.get('height')),
2593 'format_note': 'SMIL storyboards',
2598 def _parse_smil_subtitles(self
, smil
, namespace
=None, subtitles_lang
='en'):
2601 for num
, textstream
in enumerate(smil
.findall(self
._xpath
_ns
('.//textstream', namespace
))):
2602 src
= textstream
.get('src')
2603 if not src
or src
in urls
:
2606 ext
= textstream
.get('ext') or mimetype2ext(textstream
.get('type')) or determine_ext(src
)
2607 lang
= textstream
.get('systemLanguage') or textstream
.get('systemLanguageName') or textstream
.get('lang') or subtitles_lang
2608 subtitles
.setdefault(lang
, []).append({
2614 def _extract_xspf_playlist(self
, xspf_url
, playlist_id
, fatal
=True):
2615 xspf
= self
._download
_xml
(
2616 xspf_url
, playlist_id
, 'Downloading xpsf playlist',
2617 'Unable to download xspf manifest', fatal
=fatal
)
2620 return self
._parse
_xspf
(
2621 xspf
, playlist_id
, xspf_url
=xspf_url
,
2622 xspf_base_url
=base_url(xspf_url
))
2624 def _parse_xspf(self
, xspf_doc
, playlist_id
, xspf_url
=None, xspf_base_url
=None):
2626 'xspf': 'http://xspf.org/ns/0/',
2627 's1': 'http://static.streamone.nl/player/ns/0',
2631 for track
in xspf_doc
.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP
)):
2633 track
, xpath_with_ns('./xspf:title', NS_MAP
), 'title', default
=playlist_id
)
2634 description
= xpath_text(
2635 track
, xpath_with_ns('./xspf:annotation', NS_MAP
), 'description')
2636 thumbnail
= xpath_text(
2637 track
, xpath_with_ns('./xspf:image', NS_MAP
), 'thumbnail')
2638 duration
= float_or_none(
2639 xpath_text(track
, xpath_with_ns('./xspf:duration', NS_MAP
), 'duration'), 1000)
2642 for location
in track
.findall(xpath_with_ns('./xspf:location', NS_MAP
)):
2643 format_url
= urljoin(xspf_base_url
, location
.text
)
2648 'manifest_url': xspf_url
,
2649 'format_id': location
.get(xpath_with_ns('s1:label', NS_MAP
)),
2650 'width': int_or_none(location
.get(xpath_with_ns('s1:width', NS_MAP
))),
2651 'height': int_or_none(location
.get(xpath_with_ns('s1:height', NS_MAP
))),
2653 self
._sort
_formats
(formats
)
2658 'description': description
,
2659 'thumbnail': thumbnail
,
2660 'duration': duration
,
2665 def _extract_mpd_formats(self
, *args
, **kwargs
):
2666 fmts
, subs
= self
._extract
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2668 self
._report
_ignoring
_subs
('DASH')
2671 def _extract_mpd_formats_and_subtitles(
2672 self
, mpd_url
, video_id
, mpd_id
=None, note
=None, errnote
=None,
2673 fatal
=True, data
=None, headers
={}, query={}
):
2674 res
= self
._download
_xml
_handle
(
2676 note
='Downloading MPD manifest' if note
is None else note
,
2677 errnote
='Failed to download MPD manifest' if errnote
is None else errnote
,
2678 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
2684 mpd_base_url
= base_url(urlh
.geturl())
2686 return self
._parse
_mpd
_formats
_and
_subtitles
(
2687 mpd_doc
, mpd_id
, mpd_base_url
, mpd_url
)
2689 def _parse_mpd_formats(self
, *args
, **kwargs
):
2690 fmts
, subs
= self
._parse
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2692 self
._report
_ignoring
_subs
('DASH')
2695 def _parse_mpd_formats_and_subtitles(
2696 self
, mpd_doc
, mpd_id
=None, mpd_base_url
='', mpd_url
=None):
2698 Parse formats from MPD manifest.
2700 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
2701 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
2702 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
2704 if not self
.get_param('dynamic_mpd', True):
2705 if mpd_doc
.get('type') == 'dynamic':
2708 namespace
= self
._search
_regex
(r
'(?i)^{([^}]+)?}MPD$', mpd_doc
.tag
, 'namespace', default
=None)
2711 return self
._xpath
_ns
(path
, namespace
)
2713 def is_drm_protected(element
):
2714 return element
.find(_add_ns('ContentProtection')) is not None
2716 def extract_multisegment_info(element
, ms_parent_info
):
2717 ms_info
= ms_parent_info
.copy()
2719 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2720 # common attributes and elements. We will only extract relevant
2722 def extract_common(source
):
2723 segment_timeline
= source
.find(_add_ns('SegmentTimeline'))
2724 if segment_timeline
is not None:
2725 s_e
= segment_timeline
.findall(_add_ns('S'))
2727 ms_info
['total_number'] = 0
2730 r
= int(s
.get('r', 0))
2731 ms_info
['total_number'] += 1 + r
2732 ms_info
['s'].append({
2733 't': int(s
.get('t', 0)),
2734 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2735 'd': int(s
.attrib
['d']),
2738 start_number
= source
.get('startNumber')
2740 ms_info
['start_number'] = int(start_number
)
2741 timescale
= source
.get('timescale')
2743 ms_info
['timescale'] = int(timescale
)
2744 segment_duration
= source
.get('duration')
2745 if segment_duration
:
2746 ms_info
['segment_duration'] = float(segment_duration
)
2748 def extract_Initialization(source
):
2749 initialization
= source
.find(_add_ns('Initialization'))
2750 if initialization
is not None:
2751 ms_info
['initialization_url'] = initialization
.attrib
['sourceURL']
2753 segment_list
= element
.find(_add_ns('SegmentList'))
2754 if segment_list
is not None:
2755 extract_common(segment_list
)
2756 extract_Initialization(segment_list
)
2757 segment_urls_e
= segment_list
.findall(_add_ns('SegmentURL'))
2759 ms_info
['segment_urls'] = [segment
.attrib
['media'] for segment
in segment_urls_e
]
2761 segment_template
= element
.find(_add_ns('SegmentTemplate'))
2762 if segment_template
is not None:
2763 extract_common(segment_template
)
2764 media
= segment_template
.get('media')
2766 ms_info
['media'] = media
2767 initialization
= segment_template
.get('initialization')
2769 ms_info
['initialization'] = initialization
2771 extract_Initialization(segment_template
)
2774 mpd_duration
= parse_duration(mpd_doc
.get('mediaPresentationDuration'))
2775 formats
, subtitles
= [], {}
2776 stream_numbers
= collections
.defaultdict(int)
2777 for period
in mpd_doc
.findall(_add_ns('Period')):
2778 period_duration
= parse_duration(period
.get('duration')) or mpd_duration
2779 period_ms_info
= extract_multisegment_info(period
, {
2783 for adaptation_set
in period
.findall(_add_ns('AdaptationSet')):
2784 adaption_set_ms_info
= extract_multisegment_info(adaptation_set
, period_ms_info
)
2785 for representation
in adaptation_set
.findall(_add_ns('Representation')):
2786 representation_attrib
= adaptation_set
.attrib
.copy()
2787 representation_attrib
.update(representation
.attrib
)
2788 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2789 mime_type
= representation_attrib
['mimeType']
2790 content_type
= representation_attrib
.get('contentType', mime_type
.split('/')[0])
2792 codecs
= parse_codecs(representation_attrib
.get('codecs', ''))
2793 if content_type
not in ('video', 'audio', 'text'):
2794 if mime_type
== 'image/jpeg':
2795 content_type
= mime_type
2796 elif codecs
['vcodec'] != 'none':
2797 content_type
= 'video'
2798 elif codecs
['acodec'] != 'none':
2799 content_type
= 'audio'
2800 elif codecs
.get('tcodec', 'none') != 'none':
2801 content_type
= 'text'
2802 elif mimetype2ext(mime_type
) in ('tt', 'dfxp', 'ttml', 'xml', 'json'):
2803 content_type
= 'text'
2805 self
.report_warning('Unknown MIME type %s in DASH manifest' % mime_type
)
2809 for element
in (representation
, adaptation_set
, period
, mpd_doc
):
2810 base_url_e
= element
.find(_add_ns('BaseURL'))
2811 if base_url_e
is not None:
2812 base_url
= base_url_e
.text
+ base_url
2813 if re
.match(r
'^https?://', base_url
):
2815 if mpd_base_url
and base_url
.startswith('/'):
2816 base_url
= compat_urlparse
.urljoin(mpd_base_url
, base_url
)
2817 elif mpd_base_url
and not re
.match(r
'^https?://', base_url
):
2818 if not mpd_base_url
.endswith('/'):
2820 base_url
= mpd_base_url
+ base_url
2821 representation_id
= representation_attrib
.get('id')
2822 lang
= representation_attrib
.get('lang')
2823 url_el
= representation
.find(_add_ns('BaseURL'))
2824 filesize
= int_or_none(url_el
.attrib
.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el
is not None else None)
2825 bandwidth
= int_or_none(representation_attrib
.get('bandwidth'))
2826 if representation_id
is not None:
2827 format_id
= representation_id
2829 format_id
= content_type
2831 format_id
= mpd_id
+ '-' + format_id
2832 if content_type
in ('video', 'audio'):
2834 'format_id': format_id
,
2835 'manifest_url': mpd_url
,
2836 'ext': mimetype2ext(mime_type
),
2837 'width': int_or_none(representation_attrib
.get('width')),
2838 'height': int_or_none(representation_attrib
.get('height')),
2839 'tbr': float_or_none(bandwidth
, 1000),
2840 'asr': int_or_none(representation_attrib
.get('audioSamplingRate')),
2841 'fps': int_or_none(representation_attrib
.get('frameRate')),
2842 'language': lang
if lang
not in ('mul', 'und', 'zxx', 'mis') else None,
2843 'format_note': 'DASH %s' % content_type
,
2844 'filesize': filesize
,
2845 'container': mimetype2ext(mime_type
) + '_dash',
2848 elif content_type
== 'text':
2850 'ext': mimetype2ext(mime_type
),
2851 'manifest_url': mpd_url
,
2852 'filesize': filesize
,
2854 elif content_type
== 'image/jpeg':
2855 # See test case in VikiIE
2856 # https://www.viki.com/videos/1175236v-choosing-spouse-by-lottery-episode-1
2858 'format_id': format_id
,
2860 'manifest_url': mpd_url
,
2861 'format_note': 'DASH storyboards (jpeg)',
2865 if is_drm_protected(adaptation_set
) or is_drm_protected(representation
):
2867 representation_ms_info
= extract_multisegment_info(representation
, adaption_set_ms_info
)
2869 def prepare_template(template_name
, identifiers
):
2870 tmpl
= representation_ms_info
[template_name
]
2871 # First of, % characters outside $...$ templates
2872 # must be escaped by doubling for proper processing
2873 # by % operator string formatting used further (see
2874 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2880 in_template
= not in_template
2881 elif c
== '%' and not in_template
:
2883 # Next, $...$ templates are translated to their
2884 # %(...) counterparts to be used with % operator
2885 if representation_id
is not None:
2886 t
= t
.replace('$RepresentationID$', representation_id
)
2887 t
= re
.sub(r
'\$(%s)\$' % '|'.join(identifiers
), r
'%(\1)d', t
)
2888 t
= re
.sub(r
'\$(%s)%%([^$]+)\$' % '|'.join(identifiers
), r
'%(\1)\2', t
)
2889 t
.replace('$$', '$')
2892 # @initialization is a regular template like @media one
2893 # so it should be handled just the same way (see
2894 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2895 if 'initialization' in representation_ms_info
:
2896 initialization_template
= prepare_template(
2898 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2899 # $Time$ shall not be included for @initialization thus
2900 # only $Bandwidth$ remains
2902 representation_ms_info
['initialization_url'] = initialization_template
% {
2903 'Bandwidth': bandwidth
,
2906 def location_key(location
):
2907 return 'url' if re
.match(r
'^https?://', location
) else 'path'
2909 if 'segment_urls' not in representation_ms_info
and 'media' in representation_ms_info
:
2911 media_template
= prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2912 media_location_key
= location_key(media_template
)
2914 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2915 # can't be used at the same time
2916 if '%(Number' in media_template
and 's' not in representation_ms_info
:
2917 segment_duration
= None
2918 if 'total_number' not in representation_ms_info
and 'segment_duration' in representation_ms_info
:
2919 segment_duration
= float_or_none(representation_ms_info
['segment_duration'], representation_ms_info
['timescale'])
2920 representation_ms_info
['total_number'] = int(math
.ceil(
2921 float_or_none(period_duration
, segment_duration
, default
=0)))
2922 representation_ms_info
['fragments'] = [{
2923 media_location_key
: media_template
% {
2924 'Number': segment_number
,
2925 'Bandwidth': bandwidth
,
2927 'duration': segment_duration
,
2928 } for segment_number
in range(
2929 representation_ms_info
['start_number'],
2930 representation_ms_info
['total_number'] + representation_ms_info
['start_number'])]
2932 # $Number*$ or $Time$ in media template with S list available
2933 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2934 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2935 representation_ms_info
['fragments'] = []
2938 segment_number
= representation_ms_info
['start_number']
2940 def add_segment_url():
2941 segment_url
= media_template
% {
2942 'Time': segment_time
,
2943 'Bandwidth': bandwidth
,
2944 'Number': segment_number
,
2946 representation_ms_info
['fragments'].append({
2947 media_location_key
: segment_url
,
2948 'duration': float_or_none(segment_d
, representation_ms_info
['timescale']),
2951 for num
, s
in enumerate(representation_ms_info
['s']):
2952 segment_time
= s
.get('t') or segment_time
2956 for r
in range(s
.get('r', 0)):
2957 segment_time
+= segment_d
2960 segment_time
+= segment_d
2961 elif 'segment_urls' in representation_ms_info
and 's' in representation_ms_info
:
2963 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2964 # or any YouTube dashsegments video
2967 timescale
= representation_ms_info
['timescale']
2968 for s
in representation_ms_info
['s']:
2969 duration
= float_or_none(s
['d'], timescale
)
2970 for r
in range(s
.get('r', 0) + 1):
2971 segment_uri
= representation_ms_info
['segment_urls'][segment_index
]
2973 location_key(segment_uri
): segment_uri
,
2974 'duration': duration
,
2977 representation_ms_info
['fragments'] = fragments
2978 elif 'segment_urls' in representation_ms_info
:
2979 # Segment URLs with no SegmentTimeline
2980 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2981 # https://github.com/ytdl-org/youtube-dl/pull/14844
2983 segment_duration
= float_or_none(
2984 representation_ms_info
['segment_duration'],
2985 representation_ms_info
['timescale']) if 'segment_duration' in representation_ms_info
else None
2986 for segment_url
in representation_ms_info
['segment_urls']:
2988 location_key(segment_url
): segment_url
,
2990 if segment_duration
:
2991 fragment
['duration'] = segment_duration
2992 fragments
.append(fragment
)
2993 representation_ms_info
['fragments'] = fragments
2994 # If there is a fragments key available then we correctly recognized fragmented media.
2995 # Otherwise we will assume unfragmented media with direct access. Technically, such
2996 # assumption is not necessarily correct since we may simply have no support for
2997 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2998 if 'fragments' in representation_ms_info
:
3000 # NB: mpd_url may be empty when MPD manifest is parsed from a string
3001 'url': mpd_url
or base_url
,
3002 'fragment_base_url': base_url
,
3004 'protocol': 'http_dash_segments' if mime_type
!= 'image/jpeg' else 'mhtml',
3006 if 'initialization_url' in representation_ms_info
:
3007 initialization_url
= representation_ms_info
['initialization_url']
3008 if not f
.get('url'):
3009 f
['url'] = initialization_url
3010 f
['fragments'].append({location_key(initialization_url): initialization_url}
)
3011 f
['fragments'].extend(representation_ms_info
['fragments'])
3012 if not period_duration
:
3013 period_duration
= try_get(
3014 representation_ms_info
,
3015 lambda r
: sum(frag
['duration'] for frag
in r
['fragments']), float)
3017 # Assuming direct URL to unfragmented media.
3019 if content_type
in ('video', 'audio', 'image/jpeg'):
3020 f
['manifest_stream_number'] = stream_numbers
[f
['url']]
3021 stream_numbers
[f
['url']] += 1
3023 elif content_type
== 'text':
3024 subtitles
.setdefault(lang
or 'und', []).append(f
)
3026 return formats
, subtitles
3028 def _extract_ism_formats(self
, *args
, **kwargs
):
3029 fmts
, subs
= self
._extract
_ism
_formats
_and
_subtitles
(*args
, **kwargs
)
3031 self
._report
_ignoring
_subs
('ISM')
3034 def _extract_ism_formats_and_subtitles(self
, ism_url
, video_id
, ism_id
=None, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
):
3035 res
= self
._download
_xml
_handle
(
3037 note
='Downloading ISM manifest' if note
is None else note
,
3038 errnote
='Failed to download ISM manifest' if errnote
is None else errnote
,
3039 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
3046 return self
._parse
_ism
_formats
_and
_subtitles
(ism_doc
, urlh
.geturl(), ism_id
)
3048 def _parse_ism_formats_and_subtitles(self
, ism_doc
, ism_url
, ism_id
=None):
3050 Parse formats from ISM manifest.
3052 1. [MS-SSTR]: Smooth Streaming Protocol,
3053 https://msdn.microsoft.com/en-us/library/ff469518.aspx
3055 if ism_doc
.get('IsLive') == 'TRUE':
3058 duration
= int(ism_doc
.attrib
['Duration'])
3059 timescale
= int_or_none(ism_doc
.get('TimeScale')) or 10000000
3063 for stream
in ism_doc
.findall('StreamIndex'):
3064 stream_type
= stream
.get('Type')
3065 if stream_type
not in ('video', 'audio', 'text'):
3067 url_pattern
= stream
.attrib
['Url']
3068 stream_timescale
= int_or_none(stream
.get('TimeScale')) or timescale
3069 stream_name
= stream
.get('Name')
3070 stream_language
= stream
.get('Language', 'und')
3071 for track
in stream
.findall('QualityLevel'):
3072 fourcc
= track
.get('FourCC') or ('AACL' if track
.get('AudioTag') == '255' else None)
3073 # TODO: add support for WVC1 and WMAP
3074 if fourcc
not in ('H264', 'AVC1', 'AACL', 'TTML'):
3075 self
.report_warning('%s is not a supported codec' % fourcc
)
3077 tbr
= int(track
.attrib
['Bitrate']) // 1000
3078 # [1] does not mention Width and Height attributes. However,
3079 # they're often present while MaxWidth and MaxHeight are
3080 # missing, so should be used as fallbacks
3081 width
= int_or_none(track
.get('MaxWidth') or track
.get('Width'))
3082 height
= int_or_none(track
.get('MaxHeight') or track
.get('Height'))
3083 sampling_rate
= int_or_none(track
.get('SamplingRate'))
3085 track_url_pattern
= re
.sub(r
'{[Bb]itrate}', track
.attrib
['Bitrate'], url_pattern
)
3086 track_url_pattern
= compat_urlparse
.urljoin(ism_url
, track_url_pattern
)
3092 stream_fragments
= stream
.findall('c')
3093 for stream_fragment_index
, stream_fragment
in enumerate(stream_fragments
):
3094 fragment_ctx
['time'] = int_or_none(stream_fragment
.get('t')) or fragment_ctx
['time']
3095 fragment_repeat
= int_or_none(stream_fragment
.get('r')) or 1
3096 fragment_ctx
['duration'] = int_or_none(stream_fragment
.get('d'))
3097 if not fragment_ctx
['duration']:
3099 next_fragment_time
= int(stream_fragment
[stream_fragment_index
+ 1].attrib
['t'])
3101 next_fragment_time
= duration
3102 fragment_ctx
['duration'] = (next_fragment_time
- fragment_ctx
['time']) / fragment_repeat
3103 for _
in range(fragment_repeat
):
3105 'url': re
.sub(r
'{start[ _]time}', compat_str(fragment_ctx
['time']), track_url_pattern
),
3106 'duration': fragment_ctx
['duration'] / stream_timescale
,
3108 fragment_ctx
['time'] += fragment_ctx
['duration']
3110 if stream_type
== 'text':
3111 subtitles
.setdefault(stream_language
, []).append({
3115 'manifest_url': ism_url
,
3116 'fragments': fragments
,
3117 '_download_params': {
3118 'stream_type': stream_type
,
3119 'duration': duration
,
3120 'timescale': stream_timescale
,
3122 'language': stream_language
,
3123 'codec_private_data': track
.get('CodecPrivateData'),
3126 elif stream_type
in ('video', 'audio'):
3128 'format_id': join_nonempty(ism_id
, stream_name
, tbr
),
3130 'manifest_url': ism_url
,
3131 'ext': 'ismv' if stream_type
== 'video' else 'isma',
3135 'asr': sampling_rate
,
3136 'vcodec': 'none' if stream_type
== 'audio' else fourcc
,
3137 'acodec': 'none' if stream_type
== 'video' else fourcc
,
3139 'fragments': fragments
,
3140 'has_drm': ism_doc
.find('Protection') is not None,
3141 '_download_params': {
3142 'stream_type': stream_type
,
3143 'duration': duration
,
3144 'timescale': stream_timescale
,
3145 'width': width
or 0,
3146 'height': height
or 0,
3148 'language': stream_language
,
3149 'codec_private_data': track
.get('CodecPrivateData'),
3150 'sampling_rate': sampling_rate
,
3151 'channels': int_or_none(track
.get('Channels', 2)),
3152 'bits_per_sample': int_or_none(track
.get('BitsPerSample', 16)),
3153 'nal_unit_length_field': int_or_none(track
.get('NALUnitLengthField', 4)),
3156 return formats
, subtitles
3158 def _parse_html5_media_entries(self
, base_url
, webpage
, video_id
, m3u8_id
=None, m3u8_entry_protocol
='m3u8_native', mpd_id
=None, preference
=None, quality
=None):
3159 def absolute_url(item_url
):
3160 return urljoin(base_url
, item_url
)
3162 def parse_content_type(content_type
):
3163 if not content_type
:
3165 ctr
= re
.search(r
'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type
)
3167 mimetype
, codecs
= ctr
.groups()
3168 f
= parse_codecs(codecs
)
3169 f
['ext'] = mimetype2ext(mimetype
)
3173 def _media_formats(src
, cur_media_type
, type_info
={}):
3174 full_url
= absolute_url(src
)
3175 ext
= type_info
.get('ext') or determine_ext(full_url
)
3177 is_plain_url
= False
3178 formats
= self
._extract
_m
3u8_formats
(
3179 full_url
, video_id
, ext
='mp4',
3180 entry_protocol
=m3u8_entry_protocol
, m3u8_id
=m3u8_id
,
3181 preference
=preference
, quality
=quality
, fatal
=False)
3183 is_plain_url
= False
3184 formats
= self
._extract
_mpd
_formats
(
3185 full_url
, video_id
, mpd_id
=mpd_id
, fatal
=False)
3190 'vcodec': 'none' if cur_media_type
== 'audio' else None,
3192 return is_plain_url
, formats
3195 # amp-video and amp-audio are very similar to their HTML5 counterparts
3196 # so we wll include them right here (see
3197 # https://www.ampproject.org/docs/reference/components/amp-video)
3198 # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/
3199 _MEDIA_TAG_NAME_RE
= r
'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)'
3200 media_tags
= [(media_tag
, media_tag_name
, media_type
, '')
3201 for media_tag
, media_tag_name
, media_type
3202 in re
.findall(r
'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE
, webpage
)]
3203 media_tags
.extend(re
.findall(
3204 # We only allow video|audio followed by a whitespace or '>'.
3205 # Allowing more characters may end up in significant slow down (see
3206 # https://github.com/ytdl-org/youtube-dl/issues/11979, example URL:
3207 # http://www.porntrex.com/maps/videositemap.xml).
3208 r
'(?s)(<(?P<tag>%s)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>' % _MEDIA_TAG_NAME_RE
, webpage
))
3209 for media_tag
, _
, media_type
, media_content
in media_tags
:
3214 media_attributes
= extract_attributes(media_tag
)
3215 src
= strip_or_none(media_attributes
.get('src'))
3217 _
, formats
= _media_formats(src
, media_type
)
3218 media_info
['formats'].extend(formats
)
3219 media_info
['thumbnail'] = absolute_url(media_attributes
.get('poster'))
3221 for source_tag
in re
.findall(r
'<source[^>]+>', media_content
):
3222 s_attr
= extract_attributes(source_tag
)
3223 # data-video-src and data-src are non standard but seen
3224 # several times in the wild
3225 src
= strip_or_none(dict_get(s_attr
, ('src', 'data-video-src', 'data-src')))
3228 f
= parse_content_type(s_attr
.get('type'))
3229 is_plain_url
, formats
= _media_formats(src
, media_type
, f
)
3231 # width, height, res, label and title attributes are
3232 # all not standard but seen several times in the wild
3235 for lbl
in ('label', 'title')
3236 if str_or_none(s_attr
.get(lbl
))
3238 width
= int_or_none(s_attr
.get('width'))
3239 height
= (int_or_none(s_attr
.get('height'))
3240 or int_or_none(s_attr
.get('res')))
3241 if not width
or not height
:
3243 resolution
= parse_resolution(lbl
)
3246 width
= width
or resolution
.get('width')
3247 height
= height
or resolution
.get('height')
3249 tbr
= parse_bitrate(lbl
)
3258 'format_id': s_attr
.get('label') or s_attr
.get('title'),
3260 f
.update(formats
[0])
3261 media_info
['formats'].append(f
)
3263 media_info
['formats'].extend(formats
)
3264 for track_tag
in re
.findall(r
'<track[^>]+>', media_content
):
3265 track_attributes
= extract_attributes(track_tag
)
3266 kind
= track_attributes
.get('kind')
3267 if not kind
or kind
in ('subtitles', 'captions'):
3268 src
= strip_or_none(track_attributes
.get('src'))
3271 lang
= track_attributes
.get('srclang') or track_attributes
.get('lang') or track_attributes
.get('label')
3272 media_info
['subtitles'].setdefault(lang
, []).append({
3273 'url': absolute_url(src
),
3275 for f
in media_info
['formats']:
3276 f
.setdefault('http_headers', {})['Referer'] = base_url
3277 if media_info
['formats'] or media_info
['subtitles']:
3278 entries
.append(media_info
)
3281 def _extract_akamai_formats(self
, *args
, **kwargs
):
3282 fmts
, subs
= self
._extract
_akamai
_formats
_and
_subtitles
(*args
, **kwargs
)
3284 self
._report
_ignoring
_subs
('akamai')
3287 def _extract_akamai_formats_and_subtitles(self
, manifest_url
, video_id
, hosts
={}):
3288 signed
= 'hdnea=' in manifest_url
3290 # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
3291 manifest_url
= re
.sub(
3292 r
'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
3293 '', manifest_url
).strip('?')
3298 hdcore_sign
= 'hdcore=3.7.0'
3299 f4m_url
= re
.sub(r
'(https?://[^/]+)/i/', r
'\1/z/', manifest_url
).replace('/master.m3u8', '/manifest.f4m')
3300 hds_host
= hosts
.get('hds')
3302 f4m_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hds_host
, f4m_url
)
3303 if 'hdcore=' not in f4m_url
:
3304 f4m_url
+= ('&' if '?' in f4m_url
else '?') + hdcore_sign
3305 f4m_formats
= self
._extract
_f
4m
_formats
(
3306 f4m_url
, video_id
, f4m_id
='hds', fatal
=False)
3307 for entry
in f4m_formats
:
3308 entry
.update({'extra_param_to_segment_url': hdcore_sign}
)
3309 formats
.extend(f4m_formats
)
3311 m3u8_url
= re
.sub(r
'(https?://[^/]+)/z/', r
'\1/i/', manifest_url
).replace('/manifest.f4m', '/master.m3u8')
3312 hls_host
= hosts
.get('hls')
3314 m3u8_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hls_host
, m3u8_url
)
3315 m3u8_formats
, m3u8_subtitles
= self
._extract
_m
3u8_formats
_and
_subtitles
(
3316 m3u8_url
, video_id
, 'mp4', 'm3u8_native',
3317 m3u8_id
='hls', fatal
=False)
3318 formats
.extend(m3u8_formats
)
3319 subtitles
= self
._merge
_subtitles
(subtitles
, m3u8_subtitles
)
3321 http_host
= hosts
.get('http')
3322 if http_host
and m3u8_formats
and not signed
:
3323 REPL_REGEX
= r
'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
3324 qualities
= re
.match(REPL_REGEX
, m3u8_url
).group(2).split(',')
3325 qualities_length
= len(qualities
)
3326 if len(m3u8_formats
) in (qualities_length
, qualities_length
+ 1):
3328 for f
in m3u8_formats
:
3329 if f
['vcodec'] != 'none':
3330 for protocol
in ('http', 'https'):
3332 del http_f
['manifest_url']
3334 REPL_REGEX
, protocol
+ r
'://%s/\g<1>%s\3' % (http_host
, qualities
[i
]), f
['url'])
3336 'format_id': http_f
['format_id'].replace('hls-', protocol
+ '-'),
3338 'protocol': protocol
,
3340 formats
.append(http_f
)
3343 return formats
, subtitles
3345 def _extract_wowza_formats(self
, url
, video_id
, m3u8_entry_protocol
='m3u8_native', skip_protocols
=[]):
3346 query
= compat_urlparse
.urlparse(url
).query
3347 url
= re
.sub(r
'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url
)
3349 r
'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url
)
3350 url_base
= mobj
.group('url')
3351 http_base_url
= '%s%s:%s' % ('http', mobj
.group('s') or '', url_base
)
3354 def manifest_url(manifest
):
3355 m_url
= '%s/%s' % (http_base_url
, manifest
)
3357 m_url
+= '?%s' % query
3360 if 'm3u8' not in skip_protocols
:
3361 formats
.extend(self
._extract
_m
3u8_formats
(
3362 manifest_url('playlist.m3u8'), video_id
, 'mp4',
3363 m3u8_entry_protocol
, m3u8_id
='hls', fatal
=False))
3364 if 'f4m' not in skip_protocols
:
3365 formats
.extend(self
._extract
_f
4m
_formats
(
3366 manifest_url('manifest.f4m'),
3367 video_id
, f4m_id
='hds', fatal
=False))
3368 if 'dash' not in skip_protocols
:
3369 formats
.extend(self
._extract
_mpd
_formats
(
3370 manifest_url('manifest.mpd'),
3371 video_id
, mpd_id
='dash', fatal
=False))
3372 if re
.search(r
'(?:/smil:|\.smil)', url_base
):
3373 if 'smil' not in skip_protocols
:
3374 rtmp_formats
= self
._extract
_smil
_formats
(
3375 manifest_url('jwplayer.smil'),
3376 video_id
, fatal
=False)
3377 for rtmp_format
in rtmp_formats
:
3378 rtsp_format
= rtmp_format
.copy()
3379 rtsp_format
['url'] = '%s/%s' % (rtmp_format
['url'], rtmp_format
['play_path'])
3380 del rtsp_format
['play_path']
3381 del rtsp_format
['ext']
3382 rtsp_format
.update({
3383 'url': rtsp_format
['url'].replace('rtmp://', 'rtsp://'),
3384 'format_id': rtmp_format
['format_id'].replace('rtmp', 'rtsp'),
3387 formats
.extend([rtmp_format
, rtsp_format
])
3389 for protocol
in ('rtmp', 'rtsp'):
3390 if protocol
not in skip_protocols
:
3392 'url': '%s:%s' % (protocol
, url_base
),
3393 'format_id': protocol
,
3394 'protocol': protocol
,
3398 def _find_jwplayer_data(self
, webpage
, video_id
=None, transform_source
=js_to_json
):
3400 r
'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P
=quote
)\
)(?
!</script
>).*?\
.setup\s
*\
((?P
<options
>[^
)]+)\
)',
3404 jwplayer_data = self._parse_json(mobj.group('options
'),
3406 transform_source=transform_source)
3407 except ExtractorError:
3410 if isinstance(jwplayer_data, dict):
3411 return jwplayer_data
3413 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
3414 jwplayer_data = self._find_jwplayer_data(
3415 webpage, video_id, transform_source=js_to_json)
3416 return self._parse_jwplayer_data(
3417 jwplayer_data, video_id, *args, **kwargs)
3419 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
3420 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3421 # JWPlayer backward compatibility: flattened playlists
3422 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
3423 if 'playlist
' not in jwplayer_data:
3424 jwplayer_data = {'playlist': [jwplayer_data]}
3428 # JWPlayer backward compatibility: single playlist item
3429 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
3430 if not isinstance(jwplayer_data['playlist
'], list):
3431 jwplayer_data['playlist
'] = [jwplayer_data['playlist
']]
3433 for video_data in jwplayer_data['playlist
']:
3434 # JWPlayer backward compatibility: flattened sources
3435 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
3436 if 'sources
' not in video_data:
3437 video_data['sources
'] = [video_data]
3439 this_video_id = video_id or video_data['mediaid
']
3441 formats = self._parse_jwplayer_formats(
3442 video_data['sources
'], video_id=this_video_id, m3u8_id=m3u8_id,
3443 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
3446 tracks = video_data.get('tracks
')
3447 if tracks and isinstance(tracks, list):
3448 for track in tracks:
3449 if not isinstance(track, dict):
3451 track_kind = track.get('kind
')
3452 if not track_kind or not isinstance(track_kind, compat_str):
3454 if track_kind.lower() not in ('captions
', 'subtitles
'):
3456 track_url = urljoin(base_url, track.get('file'))
3459 subtitles.setdefault(track.get('label
') or 'en
', []).append({
3460 'url
': self._proto_relative_url(track_url)
3464 'id': this_video_id,
3465 'title
': unescapeHTML(video_data['title
'] if require_title else video_data.get('title
')),
3466 'description
': clean_html(video_data.get('description
')),
3467 'thumbnail
': urljoin(base_url, self._proto_relative_url(video_data.get('image
'))),
3468 'timestamp
': int_or_none(video_data.get('pubdate
')),
3469 'duration
': float_or_none(jwplayer_data.get('duration
') or video_data.get('duration
')),
3470 'subtitles
': subtitles,
3472 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
3473 if len(formats) == 1 and re.search(r'^
(?
:http|
//).*(?
:youtube\
.com|youtu\
.be
)/.+', formats[0]['url
']):
3475 '_type
': 'url_transparent
',
3476 'url
': formats[0]['url
'],
3479 self._sort_formats(formats)
3480 entry['formats
'] = formats
3481 entries.append(entry)
3482 if len(entries) == 1:
3485 return self.playlist_result(entries)
3487 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
3488 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3491 for source in jwplayer_sources_data:
3492 if not isinstance(source, dict):
3494 source_url = urljoin(
3495 base_url, self._proto_relative_url(source.get('file')))
3496 if not source_url or source_url in urls:
3498 urls.append(source_url)
3499 source_type = source.get('type') or ''
3500 ext = mimetype2ext(source_type) or determine_ext(source_url)
3501 if source_type == 'hls
' or ext == 'm3u8
':
3502 formats.extend(self._extract_m3u8_formats(
3503 source_url, video_id, 'mp4
', entry_protocol='m3u8_native
',
3504 m3u8_id=m3u8_id, fatal=False))
3505 elif source_type == 'dash
' or ext == 'mpd
':
3506 formats.extend(self._extract_mpd_formats(
3507 source_url, video_id, mpd_id=mpd_id, fatal=False))
3509 formats.extend(self._extract_smil_formats(
3510 source_url, video_id, fatal=False))
3511 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3512 elif source_type.startswith('audio
') or ext in (
3513 'oga
', 'aac
', 'mp3
', 'mpeg
', 'vorbis
'):
3520 height = int_or_none(source.get('height
'))
3522 # Often no height is provided but there is a label in
3523 # format like "1080p", "720p SD", or 1080.
3524 height = int_or_none(self._search_regex(
3525 r'^
(\d{3,4}
)[pP
]?
(?
:\b|$
)', compat_str(source.get('label
') or ''),
3526 'height
', default=None))
3529 'width
': int_or_none(source.get('width
')),
3531 'tbr
': int_or_none(source.get('bitrate
')),
3534 if source_url.startswith('rtmp
'):
3535 a_format['ext
'] = 'flv
'
3536 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3537 # of jwplayer.flash.swf
3538 rtmp_url_parts = re.split(
3539 r'((?
:mp4|mp3|flv
):)', source_url, 1)
3540 if len(rtmp_url_parts) == 3:
3541 rtmp_url, prefix, play_path = rtmp_url_parts
3544 'play_path
': prefix + play_path,
3547 a_format.update(rtmp_params)
3548 formats.append(a_format)
3551 def _live_title(self, name):
3552 self._downloader.deprecation_warning('yt_dlp
.InfoExtractor
._live
_title
is deprecated
and does
not work
as expected
')
3555 def _int(self, v, name, fatal=False, **kwargs):
3556 res = int_or_none(v, **kwargs)
3558 msg = 'Failed to extract
%s: Could
not parse value
%r' % (name, v)
3560 raise ExtractorError(msg)
3562 self.report_warning(msg)
3565 def _float(self, v, name, fatal=False, **kwargs):
3566 res = float_or_none(v, **kwargs)
3568 msg = 'Failed to extract
%s: Could
not parse value
%r' % (name, v)
3570 raise ExtractorError(msg)
3572 self.report_warning(msg)
3575 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
3576 path='/', secure=False, discard=False, rest={}, **kwargs):
3577 cookie = compat_cookiejar_Cookie(
3578 0, name, value, port, port is not None, domain, True,
3579 domain.startswith('.'), path, True, secure, expire_time,
3580 discard, None, None, rest)
3581 self._downloader.cookiejar.set_cookie(cookie)
3583 def _get_cookies(self, url):
3584 """ Return a compat_cookies_SimpleCookie with the cookies for the url """
3585 req = sanitized_Request(url)
3586 self._downloader.cookiejar.add_cookie_header(req)
3587 return compat_cookies_SimpleCookie(req.get_header('Cookie
'))
3589 def _apply_first_set_cookie_header(self, url_handle, cookie):
3591 Apply first Set-Cookie header instead of the last. Experimental.
3593 Some sites (e.g. [1-3]) may serve two cookies under the same name
3594 in Set-Cookie header and expect the first (old) one to be set rather
3595 than second (new). However, as of RFC6265 the newer one cookie
3596 should be set into cookie store what actually happens.
3597 We will workaround this issue by resetting the cookie to
3598 the first one manually.
3599 1. https://new.vk.com/
3600 2. https://github.com/ytdl-org/youtube-dl/issues/9841#issuecomment-227871201
3601 3. https://learning.oreilly.com/
3603 for header, cookies in url_handle.headers.items():
3604 if header.lower() != 'set-cookie
':
3606 if sys.version_info[0] >= 3:
3607 cookies = cookies.encode('iso
-8859-1')
3608 cookies = cookies.decode('utf
-8')
3609 cookie_value = re.search(
3610 r'%s=(.+?
);.*?
\b[Dd
]omain
=(.+?
)(?
:[,;]|$
)' % cookie, cookies)
3612 value, domain = cookie_value.groups()
3613 self._set_cookie(domain, cookie, value)
3616 def get_testcases(self, include_onlymatching=False):
3617 t = getattr(self, '_TEST
', None)
3619 assert not hasattr(self, '_TESTS
'), \
3620 '%s has _TEST
and _TESTS
' % type(self).__name__
3623 tests = getattr(self, '_TESTS
', [])
3625 if not include_onlymatching and t.get('only_matching
', False):
3627 t['name
'] = type(self).__name__[:-len('IE
')]
3630 def is_suitable(self, age_limit):
3631 """ Test whether the extractor is generally suitable for the given
3632 age limit (i.e. pornographic sites are not, all others usually are) """
3634 any_restricted = False
3635 for tc in self.get_testcases(include_onlymatching=False):
3636 if tc.get('playlist
', []):
3637 tc = tc['playlist
'][0]
3638 is_restricted = age_restricted(
3639 tc.get('info_dict
', {}).get('age_limit
'), age_limit)
3640 if not is_restricted:
3642 any_restricted = any_restricted or is_restricted
3643 return not any_restricted
3645 def extract_subtitles(self, *args, **kwargs):
3646 if (self.get_param('writesubtitles
', False)
3647 or self.get_param('listsubtitles
')):
3648 return self._get_subtitles(*args, **kwargs)
3651 def _get_subtitles(self, *args, **kwargs):
3652 raise NotImplementedError('This method must be implemented by subclasses
')
3654 def extract_comments(self, *args, **kwargs):
3655 if not self.get_param('getcomments
'):
3657 generator = self._get_comments(*args, **kwargs)
3664 comments.append(next(generator))
3665 except StopIteration:
3667 except KeyboardInterrupt:
3668 self.to_screen('Interrupted by user
')
3669 except Exception as e:
3670 if self.get_param('ignoreerrors
') is not True:
3672 self._downloader.report_error(e)
3673 comment_count = len(comments)
3674 self.to_screen(f'Extracted {comment_count} comments
')
3676 'comments
': comments,
3677 'comment_count
': None if interrupted else comment_count
3681 def _get_comments(self, *args, **kwargs):
3682 raise NotImplementedError('This method must be implemented by subclasses
')
3685 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
3686 """ Merge subtitle items for one language. Items with duplicated URLs/data
3687 will be dropped. """
3688 list1_data = set([item.get('url
') or item['data
'] for item in subtitle_list1])
3689 ret = list(subtitle_list1)
3690 ret.extend([item for item in subtitle_list2 if (item.get('url
') or item['data
']) not in list1_data])
3694 def _merge_subtitles(cls, *dicts, target=None):
3695 """ Merge subtitle dictionaries, language by language. """
3699 for lang, subs in d.items():
3700 target[lang] = cls._merge_subtitle_items(target.get(lang, []), subs)
3703 def extract_automatic_captions(self, *args, **kwargs):
3704 if (self.get_param('writeautomaticsub
', False)
3705 or self.get_param('listsubtitles
')):
3706 return self._get_automatic_captions(*args, **kwargs)
3709 def _get_automatic_captions(self, *args, **kwargs):
3710 raise NotImplementedError('This method must be implemented by subclasses
')
3712 def mark_watched(self, *args, **kwargs):
3713 if not self.get_param('mark_watched
', False):
3715 if (self.supports_login() and self._get_login_info()[0] is not None
3716 or self.get_param('cookiefile
') or self.get_param('cookiesfrombrowser
')):
3717 self._mark_watched(*args, **kwargs)
3719 def _mark_watched(self, *args, **kwargs):
3720 raise NotImplementedError('This method must be implemented by subclasses
')
3722 def geo_verification_headers(self):
3724 geo_verification_proxy = self.get_param('geo_verification_proxy
')
3725 if geo_verification_proxy:
3726 headers['Ytdl
-request
-proxy
'] = geo_verification_proxy
3729 def _generic_id(self, url):
3730 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
3732 def _generic_title(self, url):
3733 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
3736 def _availability(is_private=None, needs_premium=None, needs_subscription=None, needs_auth=None, is_unlisted=None):
3737 all_known = all(map(
3738 lambda x: x is not None,
3739 (is_private, needs_premium, needs_subscription, needs_auth, is_unlisted)))
3741 'private
' if is_private
3742 else 'premium_only
' if needs_premium
3743 else 'subscriber_only
' if needs_subscription
3744 else 'needs_auth
' if needs_auth
3745 else 'unlisted
' if is_unlisted
3746 else 'public
' if all_known
3749 def _configuration_arg(self, key, default=NO_DEFAULT, *, ie_key=None, casesense=False):
3751 @returns A list of values for the extractor argument given by "key"
3752 or "default" if no such key is present
3753 @param default The default value to return when the key is not present (default: [])
3754 @param casesense When false, the values are converted to lower case
3757 self._downloader.params, ('extractor_args
', (ie_key or self.ie_key()).lower(), key))
3759 return [] if default is NO_DEFAULT else default
3760 return list(val) if casesense else [x.lower() for x in val]
3762 def _yes_playlist(self, playlist_id, video_id, smuggled_data=None, *, playlist_label='playlist
', video_label='video
'):
3763 if not playlist_id or not video_id:
3766 no_playlist = (smuggled_data or {}).get('force_noplaylist
')
3767 if no_playlist is not None:
3768 return not no_playlist
3770 video_id = '' if video_id is True else f' {video_id}
'
3771 playlist_id = '' if playlist_id is True else f' {playlist_id}
'
3772 if self.get_param('noplaylist
'):
3773 self.to_screen(f'Downloading just the {video_label}{video_id} because of
--no
-playlist
')
3775 self.to_screen(f'Downloading {playlist_label}{playlist_id}
- add
--no
-playlist to download just the {video_label}{video_id}
')
3779 class SearchInfoExtractor(InfoExtractor):
3781 Base class for paged search queries extractors.
3782 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
3783 Instances should define _SEARCH_KEY and optionally _MAX_RESULTS
3786 _MAX_RESULTS = float('inf
')
3789 def _make_valid_url(cls):
3790 return r'%s(?P
<prefix
>|
[1-9][0-9]*|all
):(?P
<query
>[\s\S
]+)' % cls._SEARCH_KEY
3792 def _real_extract(self, query):
3793 prefix, query = self._match_valid_url(query).group('prefix
', 'query
')
3795 return self._get_n_results(query, 1)
3796 elif prefix == 'all
':
3797 return self._get_n_results(query, self._MAX_RESULTS)
3801 raise ExtractorError('invalid download number
%s for query
"%s"' % (n, query))
3802 elif n > self._MAX_RESULTS:
3803 self.report_warning('%s returns
max %i results (you requested
%i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3804 n = self._MAX_RESULTS
3805 return self._get_n_results(query, n)
3807 def _get_n_results(self, query, n):
3808 """Get a specified number of results for a query.
3809 Either this function or _search_results must be overridden by subclasses """
3810 return self.playlist_result(
3811 itertools.islice(self._search_results(query), 0, None if n == float('inf
') else n),
3814 def _search_results(self, query):
3815 """Returns an iterator of search results"""
3816 raise NotImplementedError('This method must be implemented by subclasses
')
3819 def SEARCH_KEY(self):
3820 return self._SEARCH_KEY