22 import xml
.etree
.ElementTree
24 from ..compat
import functools
# isort: split
25 from ..compat
import (
26 compat_etree_fromstring
,
31 from ..cookies
import LenientSimpleCookie
32 from ..downloader
.f4m
import get_base_url
, remove_encrypted_media
33 from ..downloader
.hls
import HlsFD
34 from ..networking
import HEADRequest
, Request
35 from ..networking
.exceptions
import (
78 parse_m3u8_attributes
,
105 """Information Extractor class.
107 Information extractors are the classes that, given a URL, extract
108 information about the video (or videos) the URL refers to. This
109 information includes the real video URL, the video title, author and
110 others. The information is stored in a dictionary which is then
111 passed to the YoutubeDL. The YoutubeDL processes this
112 information possibly downloading the video to the file system, among
113 other possible outcomes.
115 The type field determines the type of the result.
116 By far the most common value (and the default if _type is missing) is
117 "video", which indicates a single video.
119 For a video, the dictionaries must include the following fields:
121 id: Video identifier.
122 title: Video title, unescaped. Set to an empty string if video has
123 no title as opposed to "None" which signifies that the
124 extractor failed to obtain a title
126 Additionally, it must contain either a formats entry or a url one:
128 formats: A list of dictionaries for each format available, ordered
129 from worst to best quality.
132 * url The mandatory URL representing the media:
133 for plain file media - HTTP URL of this file,
135 for HLS - URL of the M3U8 media playlist,
136 for HDS - URL of the F4M manifest,
138 - HTTP URL to plain file media (in case of
140 - URL of the MPD manifest or base URL
141 representing the media if MPD manifest
142 is parsed from a string (in case of
144 for MSS - URL of the ISM manifest.
145 * request_data Data to send in POST request to the URL
147 The URL of the manifest file in case of
149 for HLS - URL of the M3U8 master playlist,
150 for HDS - URL of the F4M manifest,
151 for DASH - URL of the MPD manifest,
152 for MSS - URL of the ISM manifest.
153 * manifest_stream_number (For internal use only)
154 The index of the stream in the manifest file
155 * ext Will be calculated from URL if missing
156 * format A human-readable description of the format
157 ("mp4 container with h264/opus").
158 Calculated from the format_id, width, height.
159 and format_note fields if missing.
160 * format_id A short description of the format
161 ("mp4_h264_opus" or "19").
162 Technically optional, but strongly recommended.
163 * format_note Additional info about the format
164 ("3D" or "DASH video")
165 * width Width of the video, if known
166 * height Height of the video, if known
167 * aspect_ratio Aspect ratio of the video, if known
168 Automatically calculated from width and height
169 * resolution Textual description of width and height
170 Automatically calculated from width and height
171 * dynamic_range The dynamic range of the video. One of:
172 "SDR" (None), "HDR10", "HDR10+, "HDR12", "HLG, "DV"
173 * tbr Average bitrate of audio and video in KBit/s
174 * abr Average audio bitrate in KBit/s
175 * acodec Name of the audio codec in use
176 * asr Audio sampling rate in Hertz
177 * audio_channels Number of audio channels
178 * vbr Average video bitrate in KBit/s
180 * vcodec Name of the video codec in use
181 * container Name of the container format
182 * filesize The number of bytes, if known in advance
183 * filesize_approx An estimate for the number of bytes
184 * player_url SWF Player URL (used for rtmpdump).
185 * protocol The protocol that will be used for the actual
186 download, lower-case. One of "http", "https" or
187 one of the protocols defined in downloader.PROTOCOL_MAP
189 Base URL for fragments. Each fragment's path
190 value (if present) will be relative to
192 * fragments A list of fragments of a fragmented media.
193 Each fragment entry must contain either an url
194 or a path. If an url is present it should be
195 considered by a client. Otherwise both path and
196 fragment_base_url must be present. Here is
197 the list of all potential fields:
198 * "url" - fragment's URL
199 * "path" - fragment's path relative to
201 * "duration" (optional, int or float)
202 * "filesize" (optional, int)
203 * is_from_start Is a live format that can be downloaded
204 from the start. Boolean
205 * preference Order number of this format. If this field is
206 present and not None, the formats get sorted
207 by this field, regardless of all other values.
208 -1 for default (order by other properties),
209 -2 or smaller for less than default.
210 < -1000 to hide the format (if there is
211 another one which is strictly better)
212 * language Language code, e.g. "de" or "en-US".
213 * language_preference Is this in the language mentioned in
215 10 if it's what the URL is about,
216 -1 for default (don't know),
217 -10 otherwise, other values reserved for now.
218 * quality Order number of the video quality of this
219 format, irrespective of the file format.
220 -1 for default (order by other properties),
221 -2 or smaller for less than default.
222 * source_preference Order number for this video source
223 (quality takes higher priority)
224 -1 for default (order by other properties),
225 -2 or smaller for less than default.
226 * http_headers A dictionary of additional HTTP headers
227 to add to the request.
228 * stretched_ratio If given and not 1, indicates that the
229 video's pixels are not square.
230 width : height ratio as float.
231 * no_resume The server does not support resuming the
232 (HTTP or RTMP) download. Boolean.
233 * has_drm True if the format has DRM and cannot be downloaded.
234 'maybe' if the format may have DRM and has to be tested before download.
235 * extra_param_to_segment_url A query string to append to each
236 fragment's URL, or to update each existing query string
237 with. Only applied by the native HLS/DASH downloaders.
238 * hls_aes A dictionary of HLS AES-128 decryption information
239 used by the native HLS downloader to override the
240 values in the media playlist when an '#EXT-X-KEY' tag
241 is present in the playlist:
242 * uri The URI from which the key will be downloaded
243 * key The key (as hex) used to decrypt fragments.
244 If `key` is given, any key URI will be ignored
245 * iv The IV (as hex) used to decrypt fragments
246 * downloader_options A dictionary of downloader options
247 (For internal use only)
248 * http_chunk_size Chunk size for HTTP downloads
249 * ffmpeg_args Extra arguments for ffmpeg downloader
250 RTMP formats can also have the additional fields: page_url,
251 app, play_path, tc_url, flash_version, rtmp_live, rtmp_conn,
252 rtmp_protocol, rtmp_real_time
254 url: Final video URL.
255 ext: Video filename extension.
256 format: The video format, defaults to ext (used for --get-format)
257 player_url: SWF Player URL (used for rtmpdump).
259 The following fields are optional:
261 direct: True if a direct video file was given (must only be set by GenericIE)
262 alt_title: A secondary title of the video.
263 display_id An alternative identifier for the video, not necessarily
264 unique, but available before title. Typically, id is
265 something like "4234987", title "Dancing naked mole rats",
266 and display_id "dancing-naked-mole-rats"
267 thumbnails: A list of dictionaries, with the following entries:
268 * "id" (optional, string) - Thumbnail format ID
270 * "preference" (optional, int) - quality of the image
271 * "width" (optional, int)
272 * "height" (optional, int)
273 * "resolution" (optional, string "{width}x{height}",
275 * "filesize" (optional, int)
276 * "http_headers" (dict) - HTTP headers for the request
277 thumbnail: Full URL to a video thumbnail image.
278 description: Full video description.
279 uploader: Full name of the video uploader.
280 license: License name the video is licensed under.
281 creator: The creator of the video.
282 timestamp: UNIX timestamp of the moment the video was uploaded
283 upload_date: Video upload date in UTC (YYYYMMDD).
284 If not explicitly set, calculated from timestamp
285 release_timestamp: UNIX timestamp of the moment the video was released.
286 If it is not clear whether to use timestamp or this, use the former
287 release_date: The date (YYYYMMDD) when the video was released in UTC.
288 If not explicitly set, calculated from release_timestamp
289 modified_timestamp: UNIX timestamp of the moment the video was last modified.
290 modified_date: The date (YYYYMMDD) when the video was last modified in UTC.
291 If not explicitly set, calculated from modified_timestamp
292 uploader_id: Nickname or id of the video uploader.
293 uploader_url: Full URL to a personal webpage of the video uploader.
294 channel: Full name of the channel the video is uploaded on.
295 Note that channel fields may or may not repeat uploader
296 fields. This depends on a particular extractor.
297 channel_id: Id of the channel.
298 channel_url: Full URL to a channel webpage.
299 channel_follower_count: Number of followers of the channel.
300 channel_is_verified: Whether the channel is verified on the platform.
301 location: Physical location where the video was filmed.
302 subtitles: The available subtitles as a dictionary in the format
303 {tag: subformats}. "tag" is usually a language code, and
304 "subformats" is a list sorted from lower to higher
305 preference, each element is a dictionary with the "ext"
307 * "data": The subtitles file contents
308 * "url": A URL pointing to the subtitles file
309 It can optionally also have:
310 * "name": Name or description of the subtitles
311 * "http_headers": A dictionary of additional HTTP headers
312 to add to the request.
313 "ext" will be calculated from URL if missing
314 automatic_captions: Like 'subtitles'; contains automatically generated
315 captions instead of normal subtitles
316 duration: Length of the video in seconds, as an integer or float.
317 view_count: How many users have watched the video on the platform.
318 concurrent_view_count: How many users are currently watching the video on the platform.
319 like_count: Number of positive ratings of the video
320 dislike_count: Number of negative ratings of the video
321 repost_count: Number of reposts of the video
322 average_rating: Average rating give by users, the scale used depends on the webpage
323 comment_count: Number of comments on the video
324 comments: A list of comments, each with one or more of the following
325 properties (all but one of text or html optional):
326 * "author" - human-readable name of the comment author
327 * "author_id" - user ID of the comment author
328 * "author_thumbnail" - The thumbnail of the comment author
329 * "author_url" - The url to the comment author's page
330 * "author_is_verified" - Whether the author is verified
332 * "author_is_uploader" - Whether the comment is made by
335 * "html" - Comment as HTML
336 * "text" - Plain text of the comment
337 * "timestamp" - UNIX timestamp of comment
338 * "parent" - ID of the comment this one is replying to.
339 Set to "root" to indicate that this is a
340 comment to the original video.
341 * "like_count" - Number of positive ratings of the comment
342 * "dislike_count" - Number of negative ratings of the comment
343 * "is_favorited" - Whether the comment is marked as
344 favorite by the video uploader
345 * "is_pinned" - Whether the comment is pinned to
346 the top of the comments
347 age_limit: Age restriction for the video, as an integer (years)
348 webpage_url: The URL to the video webpage, if given to yt-dlp it
349 should allow to get the same result again. (It will be set
350 by YoutubeDL if it's missing)
351 categories: A list of categories that the video falls in, for example
353 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
354 cast: A list of the video cast
355 is_live: True, False, or None (=unknown). Whether this video is a
356 live stream that goes on instead of a fixed-length video.
357 was_live: True, False, or None (=unknown). Whether this video was
358 originally a live stream.
359 live_status: None (=unknown), 'is_live', 'is_upcoming', 'was_live', 'not_live',
360 or 'post_live' (was live, but VOD is not yet processed)
361 If absent, automatically set from is_live, was_live
362 start_time: Time in seconds where the reproduction should start, as
363 specified in the URL.
364 end_time: Time in seconds where the reproduction should end, as
365 specified in the URL.
366 chapters: A list of dictionaries, with the following entries:
367 * "start_time" - The start time of the chapter in seconds
368 * "end_time" - The end time of the chapter in seconds
369 * "title" (optional, string)
370 heatmap: A list of dictionaries, with the following entries:
371 * "start_time" - The start time of the data point in seconds
372 * "end_time" - The end time of the data point in seconds
373 * "value" - The normalized value of the data point (float between 0 and 1)
374 playable_in_embed: Whether this video is allowed to play in embedded
375 players on other sites. Can be True (=always allowed),
376 False (=never allowed), None (=unknown), or a string
377 specifying the criteria for embedability; e.g. 'whitelist'
378 availability: Under what condition the video is available. One of
379 'private', 'premium_only', 'subscriber_only', 'needs_auth',
380 'unlisted' or 'public'. Use 'InfoExtractor._availability'
382 _old_archive_ids: A list of old archive ids needed for backward compatibility
383 _format_sort_fields: A list of fields to use for sorting formats
384 __post_extractor: A function to be called just before the metadata is
385 written to either disk, logger or console. The function
386 must return a dict which will be added to the info_dict.
387 This is usefull for additional information that is
388 time-consuming to extract. Note that the fields thus
389 extracted will not be available to output template and
390 match_filter. So, only "comments" and "comment_count" are
391 currently allowed to be extracted via this method.
393 The following fields should only be used when the video belongs to some logical
396 chapter: Name or title of the chapter the video belongs to.
397 chapter_number: Number of the chapter the video belongs to, as an integer.
398 chapter_id: Id of the chapter the video belongs to, as a unicode string.
400 The following fields should only be used when the video is an episode of some
401 series, programme or podcast:
403 series: Title of the series or programme the video episode belongs to.
404 series_id: Id of the series or programme the video episode belongs to, as a unicode string.
405 season: Title of the season the video episode belongs to.
406 season_number: Number of the season the video episode belongs to, as an integer.
407 season_id: Id of the season the video episode belongs to, as a unicode string.
408 episode: Title of the video episode. Unlike mandatory video title field,
409 this field should denote the exact title of the video episode
410 without any kind of decoration.
411 episode_number: Number of the video episode within a season, as an integer.
412 episode_id: Id of the video episode, as a unicode string.
414 The following fields should only be used when the media is a track or a part of
417 track: Title of the track.
418 track_number: Number of the track within an album or a disc, as an integer.
419 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
421 artist: Artist(s) of the track.
422 genre: Genre(s) of the track.
423 album: Title of the album the track belongs to.
424 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
425 album_artist: List of all artists appeared on the album (e.g.
426 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
428 disc_number: Number of the disc or other physical medium the track belongs to,
430 release_year: Year (YYYY) when the album was released.
431 composer: Composer of the piece
433 The following fields should only be set for clips that should be cut from the original video:
435 section_start: Start time of the section in seconds
436 section_end: End time of the section in seconds
438 The following fields should only be set for storyboards:
439 rows: Number of rows in each storyboard fragment, as an integer
440 columns: Number of columns in each storyboard fragment, as an integer
442 Unless mentioned otherwise, the fields should be Unicode strings.
444 Unless mentioned otherwise, None is equivalent to absence of information.
447 _type "playlist" indicates multiple videos.
448 There must be a key "entries", which is a list, an iterable, or a PagedList
449 object, each element of which is a valid dictionary by this specification.
451 Additionally, playlists can have "id", "title", and any other relevant
452 attributes with the same semantics as videos (see above).
454 It can also have the following optional fields:
456 playlist_count: The total number of videos in a playlist. If not given,
457 YoutubeDL tries to calculate it from "entries"
460 _type "multi_video" indicates that there are multiple videos that
461 form a single show, for examples multiple acts of an opera or TV episode.
462 It must have an entries key like a playlist and contain all the keys
463 required for a video at the same time.
466 _type "url" indicates that the video must be extracted from another
467 location, possibly by a different extractor. Its only required key is:
468 "url" - the next URL to extract.
469 The key "ie_key" can be set to the class name (minus the trailing "IE",
470 e.g. "Youtube") if the extractor class is known in advance.
471 Additionally, the dictionary may have any properties of the resolved entity
472 known in advance, for example "title" if the title of the referred video is
476 _type "url_transparent" entities have the same specification as "url", but
477 indicate that the given additional information is more precise than the one
478 associated with the resolved URL.
479 This is useful when a site employs a video service that hosts the video and
480 its technical metadata, but that video service does not embed a useful
481 title, description etc.
484 Subclasses of this should also be added to the list of extractors and
485 should define _VALID_URL as a regexp or a Sequence of regexps, and
486 re-define the _real_extract() and (optionally) _real_initialize() methods.
488 Subclasses may also override suitable() if necessary, but ensure the function
489 signature is preserved and that this function imports everything it needs
490 (except other extractors), so that lazy_extractors works correctly.
492 Subclasses can define a list of _EMBED_REGEX, which will be searched for in
493 the HTML of Generic webpages. It may also override _extract_embed_urls
494 or _extract_from_webpage as necessary. While these are normally classmethods,
495 _extract_from_webpage is allowed to be an instance method.
497 _extract_from_webpage may raise self.StopExtraction() to stop further
498 processing of the webpage and obtain exclusive rights to it. This is useful
499 when the extractor cannot reliably be matched using just the URL,
500 e.g. invidious/peertube instances
502 Embed-only extractors can be defined by setting _VALID_URL = False.
504 To support username + password (or netrc) login, the extractor must define a
505 _NETRC_MACHINE and re-define _perform_login(username, password) and
506 (optionally) _initialize_pre_login() methods. The _perform_login method will
507 be called between _initialize_pre_login and _real_initialize if credentials
508 are passed by the user. In cases where it is necessary to have the login
509 process as part of the extraction rather than initialization, _perform_login
510 can be left undefined.
512 _GEO_BYPASS attribute may be set to False in order to disable
513 geo restriction bypass mechanisms for a particular extractor.
514 Though it won't disable explicit geo restriction bypass based on
515 country code provided with geo_bypass_country.
517 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
518 countries for this extractor. One of these countries will be used by
519 geo restriction bypass mechanism right away in order to bypass
520 geo restriction, of course, if the mechanism is not disabled.
522 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
523 IP blocks in CIDR notation for this extractor. One of these IP blocks
524 will be used by geo restriction bypass mechanism similarly
527 The _ENABLED attribute should be set to False for IEs that
528 are disabled by default and must be explicitly enabled.
530 The _WORKING attribute should be set to False for broken IEs
531 in order to warn the users and skip the tests.
536 _x_forwarded_for_ip
= None
538 _GEO_COUNTRIES
= None
539 _GEO_IP_BLOCKS
= None
542 _NETRC_MACHINE
= None
548 def _login_hint(self
, method
=NO_DEFAULT
, netrc
=None):
549 password_hint
= f
'--username and --password, --netrc-cmd, or --netrc ({netrc or self._NETRC_MACHINE}) to provide account credentials'
552 'any': f
'Use --cookies, --cookies-from-browser, {password_hint}',
553 'password': f
'Use {password_hint}',
555 'Use --cookies-from-browser or --cookies for the authentication. '
556 'See https://github.com/yt-dlp/yt-dlp/wiki/FAQ#how-do-i-pass-cookies-to-yt-dlp for how to manually pass cookies'),
557 }[method
if method
is not NO_DEFAULT
else 'any' if self
.supports_login() else 'cookies']
559 def __init__(self
, downloader
=None):
560 """Constructor. Receives an optional downloader (a YoutubeDL instance).
561 If a downloader is not passed during initialization,
562 it must be set using "set_downloader()" before "extract()" is called"""
564 self
._x
_forwarded
_for
_ip
= None
565 self
._printed
_messages
= set()
566 self
.set_downloader(downloader
)
569 def _match_valid_url(cls
, url
):
570 if cls
._VALID
_URL
is False:
572 # This does not use has/getattr intentionally - we want to know whether
573 # we have cached the regexp for *this* class, whereas getattr would also
574 # match the superclass
575 if '_VALID_URL_RE' not in cls
.__dict
__:
576 cls
._VALID
_URL
_RE
= tuple(map(re
.compile, variadic(cls
._VALID
_URL
)))
577 return next(filter(None, (regex
.match(url
) for regex
in cls
._VALID
_URL
_RE
)), None)
580 def suitable(cls
, url
):
581 """Receives a URL and returns True if suitable for this IE."""
582 # This function must import everything it needs (except other extractors),
583 # so that lazy_extractors works correctly
584 return cls
._match
_valid
_url
(url
) is not None
587 def _match_id(cls
, url
):
588 return cls
._match
_valid
_url
(url
).group('id')
591 def get_temp_id(cls
, url
):
593 return cls
._match
_id
(url
)
594 except (IndexError, AttributeError):
599 """Getter method for _WORKING."""
603 def supports_login(cls
):
604 return bool(cls
._NETRC
_MACHINE
)
606 def initialize(self
):
607 """Initializes an instance (authentication, etc)."""
608 self
._printed
_messages
= set()
609 self
._initialize
_geo
_bypass
({
610 'countries': self
._GEO
_COUNTRIES
,
611 'ip_blocks': self
._GEO
_IP
_BLOCKS
,
614 self
._initialize
_pre
_login
()
615 if self
.supports_login():
616 username
, password
= self
._get
_login
_info
()
618 self
._perform
_login
(username
, password
)
619 elif self
.get_param('username') and False not in (self
.IE_DESC
, self
._NETRC
_MACHINE
):
620 self
.report_warning(f
'Login with password is not supported for this website. {self._login_hint("cookies")}')
621 self
._real
_initialize
()
624 def _initialize_geo_bypass(self
, geo_bypass_context
):
626 Initialize geo restriction bypass mechanism.
628 This method is used to initialize geo bypass mechanism based on faking
629 X-Forwarded-For HTTP header. A random country from provided country list
630 is selected and a random IP belonging to this country is generated. This
631 IP will be passed as X-Forwarded-For HTTP header in all subsequent
634 This method will be used for initial geo bypass mechanism initialization
635 during the instance initialization with _GEO_COUNTRIES and
638 You may also manually call it from extractor's code if geo bypass
639 information is not available beforehand (e.g. obtained during
640 extraction) or due to some other reason. In this case you should pass
641 this information in geo bypass context passed as first argument. It may
642 contain following fields:
644 countries: List of geo unrestricted countries (similar
646 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
647 (similar to _GEO_IP_BLOCKS)
650 if not self
._x
_forwarded
_for
_ip
:
652 # Geo bypass mechanism is explicitly disabled by user
653 if not self
.get_param('geo_bypass', True):
656 if not geo_bypass_context
:
657 geo_bypass_context
= {}
659 # Backward compatibility: previously _initialize_geo_bypass
660 # expected a list of countries, some 3rd party code may still use
662 if isinstance(geo_bypass_context
, (list, tuple)):
663 geo_bypass_context
= {
664 'countries': geo_bypass_context
,
667 # The whole point of geo bypass mechanism is to fake IP
668 # as X-Forwarded-For HTTP header based on some IP block or
671 # Path 1: bypassing based on IP block in CIDR notation
673 # Explicit IP block specified by user, use it right away
674 # regardless of whether extractor is geo bypassable or not
675 ip_block
= self
.get_param('geo_bypass_ip_block', None)
677 # Otherwise use random IP block from geo bypass context but only
678 # if extractor is known as geo bypassable
680 ip_blocks
= geo_bypass_context
.get('ip_blocks')
681 if self
._GEO
_BYPASS
and ip_blocks
:
682 ip_block
= random
.choice(ip_blocks
)
685 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(ip_block
)
686 self
.write_debug(f
'Using fake IP {self._x_forwarded_for_ip} as X-Forwarded-For')
689 # Path 2: bypassing based on country code
691 # Explicit country code specified by user, use it right away
692 # regardless of whether extractor is geo bypassable or not
693 country
= self
.get_param('geo_bypass_country', None)
695 # Otherwise use random country code from geo bypass context but
696 # only if extractor is known as geo bypassable
698 countries
= geo_bypass_context
.get('countries')
699 if self
._GEO
_BYPASS
and countries
:
700 country
= random
.choice(countries
)
703 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country
)
704 self
._downloader
.write_debug(
705 f
'Using fake IP {self._x_forwarded_for_ip} ({country.upper()}) as X-Forwarded-For')
707 def extract(self
, url
):
708 """Extracts URL information and returns it in list of dicts."""
713 self
.to_screen('Extracting URL: %s' % (
714 url
if self
.get_param('verbose') else truncate_string(url
, 100, 20)))
715 ie_result
= self
._real
_extract
(url
)
716 if ie_result
is None:
718 if self
._x
_forwarded
_for
_ip
:
719 ie_result
['__x_forwarded_for_ip'] = self
._x
_forwarded
_for
_ip
720 subtitles
= ie_result
.get('subtitles') or {}
721 if 'no-live-chat' in self
.get_param('compat_opts'):
722 for lang
in ('live_chat', 'comments', 'danmaku'):
723 subtitles
.pop(lang
, None)
725 except GeoRestrictedError
as e
:
726 if self
.__maybe
_fake
_ip
_and
_retry
(e
.countries
):
729 except UnsupportedError
:
731 except ExtractorError
as e
:
732 e
.video_id
= e
.video_id
or self
.get_temp_id(url
),
733 e
.ie
= e
.ie
or self
.IE_NAME
,
734 e
.traceback
= e
.traceback
or sys
.exc_info()[2]
736 except IncompleteRead
as e
:
737 raise ExtractorError('A network error has occurred.', cause
=e
, expected
=True, video_id
=self
.get_temp_id(url
))
738 except (KeyError, StopIteration) as e
:
739 raise ExtractorError('An extractor error has occurred.', cause
=e
, video_id
=self
.get_temp_id(url
))
741 def __maybe_fake_ip_and_retry(self
, countries
):
742 if (not self
.get_param('geo_bypass_country', None)
744 and self
.get_param('geo_bypass', True)
745 and not self
._x
_forwarded
_for
_ip
747 country_code
= random
.choice(countries
)
748 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country_code
)
749 if self
._x
_forwarded
_for
_ip
:
751 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
752 % (self
._x
_forwarded
_for
_ip
, country_code
.upper()))
756 def set_downloader(self
, downloader
):
757 """Sets a YoutubeDL instance as the downloader for this IE."""
758 self
._downloader
= downloader
762 return self
._downloader
.cache
766 return self
._downloader
.cookiejar
768 def _initialize_pre_login(self
):
769 """ Initialization before login. Redefine in subclasses."""
772 def _perform_login(self
, username
, password
):
773 """ Login with username and password. Redefine in subclasses."""
776 def _real_initialize(self
):
777 """Real initialization process. Redefine in subclasses."""
780 def _real_extract(self
, url
):
781 """Real extraction process. Redefine in subclasses."""
782 raise NotImplementedError('This method must be implemented by subclasses')
786 """A string for getting the InfoExtractor with get_info_extractor"""
787 return cls
.__name
__[:-2]
791 return cls
.__name
__[:-2]
794 def __can_accept_status_code(err
, expected_status
):
795 assert isinstance(err
, HTTPError
)
796 if expected_status
is None:
798 elif callable(expected_status
):
799 return expected_status(err
.status
) is True
801 return err
.status
in variadic(expected_status
)
803 def _create_request(self
, url_or_request
, data
=None, headers
=None, query
=None):
804 if isinstance(url_or_request
, urllib
.request
.Request
):
805 self
._downloader
.deprecation_warning(
806 'Passing a urllib.request.Request to _create_request() is deprecated. '
807 'Use yt_dlp.networking.common.Request instead.')
808 url_or_request
= urllib_req_to_req(url_or_request
)
809 elif not isinstance(url_or_request
, Request
):
810 url_or_request
= Request(url_or_request
)
812 url_or_request
.update(data
=data
, headers
=headers
, query
=query
)
813 return url_or_request
815 def _request_webpage(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, data
=None, headers
=None, query
=None, expected_status
=None):
817 Return the response handle.
819 See _download_webpage docstring for arguments specification.
821 if not self
._downloader
._first
_webpage
_request
:
822 sleep_interval
= self
.get_param('sleep_interval_requests') or 0
823 if sleep_interval
> 0:
824 self
.to_screen('Sleeping %s seconds ...' % sleep_interval
)
825 time
.sleep(sleep_interval
)
827 self
._downloader
._first
_webpage
_request
= False
830 self
.report_download_webpage(video_id
)
831 elif note
is not False:
833 self
.to_screen(str(note
))
835 self
.to_screen(f
'{video_id}: {note}')
837 # Some sites check X-Forwarded-For HTTP header in order to figure out
838 # the origin of the client behind proxy. This allows bypassing geo
839 # restriction by faking this header's value to IP that belongs to some
840 # geo unrestricted country. We will do so once we encounter any
841 # geo restriction error.
842 if self
._x
_forwarded
_for
_ip
:
843 headers
= (headers
or {}).copy()
844 headers
.setdefault('X-Forwarded-For', self
._x
_forwarded
_for
_ip
)
847 return self
._downloader
.urlopen(self
._create
_request
(url_or_request
, data
, headers
, query
))
848 except network_exceptions
as err
:
849 if isinstance(err
, HTTPError
):
850 if self
.__can
_accept
_status
_code
(err
, expected_status
):
856 errnote
= 'Unable to download webpage'
858 errmsg
= f
'{errnote}: {error_to_compat_str(err)}'
860 raise ExtractorError(errmsg
, cause
=err
)
862 self
.report_warning(errmsg
)
865 def _download_webpage_handle(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True,
866 encoding
=None, data
=None, headers
={}, query={}
, expected_status
=None):
868 Return a tuple (page content as string, URL handle).
871 url_or_request -- plain text URL as a string or
872 a urllib.request.Request object
873 video_id -- Video/playlist/item identifier (string)
876 note -- note printed before downloading (string)
877 errnote -- note printed in case of an error (string)
878 fatal -- flag denoting whether error should be considered fatal,
879 i.e. whether it should cause ExtractionError to be raised,
880 otherwise a warning will be reported and extraction continued
881 encoding -- encoding for a page content decoding, guessed automatically
882 when not explicitly specified
883 data -- POST data (bytes)
884 headers -- HTTP headers (dict)
885 query -- URL query (dict)
886 expected_status -- allows to accept failed HTTP requests (non 2xx
887 status code) by explicitly specifying a set of accepted status
888 codes. Can be any of the following entities:
889 - an integer type specifying an exact failed status code to
891 - a list or a tuple of integer types specifying a list of
892 failed status codes to accept
893 - a callable accepting an actual failed status code and
894 returning True if it should be accepted
895 Note that this argument does not affect success status codes (2xx)
896 which are always accepted.
899 # Strip hashes from the URL (#1038)
900 if isinstance(url_or_request
, str):
901 url_or_request
= url_or_request
.partition('#')[0]
903 urlh
= self
._request
_webpage
(url_or_request
, video_id
, note
, errnote
, fatal
, data
=data
, headers
=headers
, query
=query
, expected_status
=expected_status
)
907 content
= self
._webpage
_read
_content
(urlh
, url_or_request
, video_id
, note
, errnote
, fatal
, encoding
=encoding
)
908 return (content
, urlh
)
911 def _guess_encoding_from_content(content_type
, webpage_bytes
):
912 m
= re
.match(r
'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type
)
914 encoding
= m
.group(1)
916 m
= re
.search(br
'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
917 webpage_bytes[:1024])
919 encoding = m.group(1).decode('ascii')
920 elif webpage_bytes.startswith(b'\xff\xfe'):
927 def __check_blocked(self, content):
928 first_block = content[:512]
929 if ('<title>Access to this site is blocked</title>' in content
930 and 'Websense' in first_block):
931 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
932 blocked_iframe = self._html_search_regex(
933 r'<iframe src="([^
"]+)"', content,
934 'Websense information URL
', default=None)
936 msg += ' Visit
%s for more details
' % blocked_iframe
937 raise ExtractorError(msg, expected=True)
938 if '<title
>The URL you requested has been blocked
</title
>' in first_block:
940 'Access to this webpage has been blocked by Indian censorship
. '
941 'Use a VPN
or proxy
server (with --proxy
) to route around it
.')
942 block_msg = self._html_search_regex(
943 r'</h1
><p
>(.*?
)</p
>',
944 content, 'block message
', default=None)
946 msg += ' (Message
: "%s")' % block_msg.replace('\n', ' ')
947 raise ExtractorError(msg, expected=True)
948 if ('<title
>TTK
:: Доступ к ресурсу ограничен
</title
>' in content
949 and 'blocklist
.rkn
.gov
.ru
' in content):
950 raise ExtractorError(
951 'Access to this webpage has been blocked by decision of the Russian government
. '
952 'Visit http
://blocklist
.rkn
.gov
.ru
/ for a block reason
.',
955 def _request_dump_filename(self, url, video_id):
956 basen = f'{video_id}_{url}
'
957 trim_length = self.get_param('trim_file_name
') or 240
958 if len(basen) > trim_length:
959 h = '___
' + hashlib.md5(basen.encode('utf
-8')).hexdigest()
960 basen = basen[:trim_length - len(h)] + h
961 filename = sanitize_filename(f'{basen}
.dump
', restricted=True)
962 # Working around MAX_PATH limitation on Windows (see
963 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
964 if compat_os_name == 'nt
':
965 absfilepath = os.path.abspath(filename)
966 if len(absfilepath) > 259:
967 filename = fR'\\?\{absfilepath}
'
970 def __decode_webpage(self, webpage_bytes, encoding, headers):
972 encoding = self._guess_encoding_from_content(headers.get('Content
-Type
', ''), webpage_bytes)
974 return webpage_bytes.decode(encoding, 'replace
')
976 return webpage_bytes.decode('utf
-8', 'replace
')
978 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
979 webpage_bytes = urlh.read()
980 if prefix is not None:
981 webpage_bytes = prefix + webpage_bytes
982 if self.get_param('dump_intermediate_pages
', False):
983 self.to_screen('Dumping request to
' + urlh.url)
984 dump = base64.b64encode(webpage_bytes).decode('ascii
')
985 self._downloader.to_screen(dump)
986 if self.get_param('write_pages
'):
987 filename = self._request_dump_filename(urlh.url, video_id)
988 self.to_screen(f'Saving request to {filename}
')
989 with open(filename, 'wb
') as outf:
990 outf.write(webpage_bytes)
992 content = self.__decode_webpage(webpage_bytes, encoding, urlh.headers)
993 self.__check_blocked(content)
997 def __print_error(self, errnote, fatal, video_id, err):
999 raise ExtractorError(f'{video_id}
: {errnote}
', cause=err)
1001 self.report_warning(f'{video_id}
: {errnote}
: {err}
')
1003 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True, errnote=None):
1004 if transform_source:
1005 xml_string = transform_source(xml_string)
1007 return compat_etree_fromstring(xml_string.encode('utf
-8'))
1008 except xml.etree.ElementTree.ParseError as ve:
1009 self.__print_error('Failed to parse XML
' if errnote is None else errnote, fatal, video_id, ve)
1011 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True, errnote=None, **parser_kwargs):
1014 json_string, cls=LenientJSONDecoder, strict=False, transform_source=transform_source, **parser_kwargs)
1015 except ValueError as ve:
1016 self.__print_error('Failed to parse JSON
' if errnote is None else errnote, fatal, video_id, ve)
1018 def _parse_socket_response_as_json(self, data, *args, **kwargs):
1019 return self._parse_json(data[data.find('{'):data.rfind('}
') + 1], *args, **kwargs)
1021 def __create_download_methods(name, parser, note, errnote, return_value):
1023 def parse(ie, content, *args, errnote=errnote, **kwargs):
1026 if errnote is False:
1027 kwargs['errnote
'] = errnote
1028 # parser is fetched by name so subclasses can override it
1029 return getattr(ie, parser)(content, *args, **kwargs)
1031 def download_handle(self, url_or_request, video_id, note=note, errnote=errnote, transform_source=None,
1032 fatal=True, encoding=None, data=None, headers={}, query={}, expected_status=None):
1033 res = self._download_webpage_handle(
1034 url_or_request, video_id, note=note, errnote=errnote, fatal=fatal, encoding=encoding,
1035 data=data, headers=headers, query=query, expected_status=expected_status)
1039 return parse(self, content, video_id, transform_source=transform_source, fatal=fatal, errnote=errnote), urlh
1041 def download_content(self, url_or_request, video_id, note=note, errnote=errnote, transform_source=None,
1042 fatal=True, encoding=None, data=None, headers={}, query={}, expected_status=None):
1043 if self.get_param('load_pages
'):
1044 url_or_request = self._create_request(url_or_request, data, headers, query)
1045 filename = self._request_dump_filename(url_or_request.url, video_id)
1046 self.to_screen(f'Loading request
from {filename}
')
1048 with open(filename, 'rb
') as dumpf:
1049 webpage_bytes = dumpf.read()
1050 except OSError as e:
1051 self.report_warning(f'Unable to load request
from disk
: {e}
')
1053 content = self.__decode_webpage(webpage_bytes, encoding, url_or_request.headers)
1054 return parse(self, content, video_id, transform_source=transform_source, fatal=fatal, errnote=errnote)
1058 'transform_source
': transform_source,
1060 'encoding
': encoding,
1064 'expected_status
': expected_status,
1067 kwargs.pop('transform_source
')
1068 # The method is fetched by name so subclasses can override _download_..._handle
1069 res = getattr(self, download_handle.__name__)(url_or_request, video_id, **kwargs)
1070 return res if res is False else res[0]
1072 def impersonate(func, name, return_value):
1073 func.__name__, func.__qualname__ = name, f'InfoExtractor
.{name}
'
1075 @param transform_source Apply this transformation before parsing
1076 @returns {return_value}
1078 See _download_webpage_handle docstring for other arguments specification
1081 impersonate(download_handle, f'_download_{name}_handle
', f'({return_value}
, URL handle
)')
1082 impersonate(download_content, f'_download_{name}
', f'{return_value}
')
1083 return download_handle, download_content
1085 _download_xml_handle, _download_xml = __create_download_methods(
1086 'xml
', '_parse_xml
', 'Downloading XML
', 'Unable to download XML
', 'xml
as an xml
.etree
.ElementTree
.Element
')
1087 _download_json_handle, _download_json = __create_download_methods(
1088 'json
', '_parse_json
', 'Downloading JSON metadata
', 'Unable to download JSON metadata
', 'JSON
object as a
dict')
1089 _download_socket_json_handle, _download_socket_json = __create_download_methods(
1090 'socket_json
', '_parse_socket_response_as_json
', 'Polling socket
', 'Unable to poll socket
', 'JSON
object as a
dict')
1091 __download_webpage = __create_download_methods('webpage
', None, None, None, 'data of the page
as a string
')[1]
1093 def _download_webpage(
1094 self, url_or_request, video_id, note=None, errnote=None,
1095 fatal=True, tries=1, timeout=NO_DEFAULT, *args, **kwargs):
1097 Return the data of the page as a string.
1100 tries -- number of tries
1101 timeout -- sleep interval between tries
1103 See _download_webpage_handle docstring for other arguments specification.
1106 R''' # NB: These are unused; should they be deprecated?
1108 self._downloader.deprecation_warning('tries argument
is deprecated
in InfoExtractor
._download
_webpage
')
1109 if timeout is NO_DEFAULT:
1112 self._downloader.deprecation_warning('timeout argument
is deprecated
in InfoExtractor
._download
_webpage
')
1118 return self.__download_webpage(url_or_request, video_id, note, errnote, None, fatal, *args, **kwargs)
1119 except IncompleteRead as e:
1121 if try_count >= tries:
1123 self._sleep(timeout, video_id)
1125 def report_warning(self, msg, video_id=None, *args, only_once=False, **kwargs):
1126 idstr = format_field(video_id, None, '%s: ')
1127 msg = f'[{self.IE_NAME}
] {idstr}{msg}
'
1129 if f'WARNING
: {msg}
' in self._printed_messages:
1131 self._printed_messages.add(f'WARNING
: {msg}
')
1132 self._downloader.report_warning(msg, *args, **kwargs)
1134 def to_screen(self, msg, *args, **kwargs):
1135 """Print msg to screen, prefixing it with '[ie_name
]'"""
1136 self._downloader.to_screen(f'[{self.IE_NAME}
] {msg}
', *args, **kwargs)
1138 def write_debug(self, msg, *args, **kwargs):
1139 self._downloader.write_debug(f'[{self.IE_NAME}
] {msg}
', *args, **kwargs)
1141 def get_param(self, name, default=None, *args, **kwargs):
1142 if self._downloader:
1143 return self._downloader.params.get(name, default, *args, **kwargs)
1146 def report_drm(self, video_id, partial=NO_DEFAULT):
1147 if partial is not NO_DEFAULT:
1148 self._downloader.deprecation_warning('InfoExtractor
.report_drm no longer accepts the argument partial
')
1149 self.raise_no_formats('This video
is DRM protected
', expected=True, video_id=video_id)
1151 def report_extraction(self, id_or_name):
1152 """Report information extraction."""
1153 self.to_screen('%s: Extracting information
' % id_or_name)
1155 def report_download_webpage(self, video_id):
1156 """Report webpage download."""
1157 self.to_screen('%s: Downloading webpage
' % video_id)
1159 def report_age_confirmation(self):
1160 """Report attempt to confirm age."""
1161 self.to_screen('Confirming age
')
1163 def report_login(self):
1164 """Report attempt to log in."""
1165 self.to_screen('Logging
in')
1167 def raise_login_required(
1168 self, msg='This video
is only available
for registered users
',
1169 metadata_available=False, method=NO_DEFAULT):
1170 if metadata_available and (
1171 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1172 self.report_warning(msg)
1174 msg += format_field(self._login_hint(method), None, '. %s')
1175 raise ExtractorError(msg, expected=True)
1177 def raise_geo_restricted(
1178 self, msg='This video
is not available
from your location due to geo restriction
',
1179 countries=None, metadata_available=False):
1180 if metadata_available and (
1181 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1182 self.report_warning(msg)
1184 raise GeoRestrictedError(msg, countries=countries)
1186 def raise_no_formats(self, msg, expected=False, video_id=None):
1188 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1189 self.report_warning(msg, video_id)
1190 elif isinstance(msg, ExtractorError):
1193 raise ExtractorError(msg, expected=expected, video_id=video_id)
1195 # Methods for following #608
1197 def url_result(url, ie=None, video_id=None, video_title=None, *, url_transparent=False, **kwargs):
1198 """Returns a URL that points to a page that should be processed"""
1200 kwargs['ie_key
'] = ie if isinstance(ie, str) else ie.ie_key()
1201 if video_id is not None:
1202 kwargs['id'] = video_id
1203 if video_title is not None:
1204 kwargs['title
'] = video_title
1207 '_type
': 'url_transparent
' if url_transparent else 'url
',
1212 def playlist_from_matches(cls, matches, playlist_id=None, playlist_title=None,
1213 getter=IDENTITY, ie=None, video_kwargs=None, **kwargs):
1214 return cls.playlist_result(
1215 (cls.url_result(m, ie, **(video_kwargs or {})) for m in orderedSet(map(getter, matches), lazy=True)),
1216 playlist_id, playlist_title, **kwargs)
1219 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None, *, multi_video=False, **kwargs):
1220 """Returns a playlist"""
1222 kwargs['id'] = playlist_id
1224 kwargs['title
'] = playlist_title
1225 if playlist_description is not None:
1226 kwargs['description
'] = playlist_description
1229 '_type
': 'multi_video
' if multi_video else 'playlist
',
1233 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1235 Perform a regex search on the given string, using a single or a list of
1236 patterns returning the first matching group.
1237 In case of failure return a default value or raise a WARNING or a
1238 RegexNotFoundError, depending on fatal, specifying the field name.
1242 elif isinstance(pattern, (str, re.Pattern)):
1243 mobj = re.search(pattern, string, flags)
1246 mobj = re.search(p, string, flags)
1250 _name = self._downloader._format_err(name, self._downloader.Styles.EMPHASIS)
1254 # return the first matching group
1255 return next(g for g in mobj.groups() if g is not None)
1256 elif isinstance(group, (list, tuple)):
1257 return tuple(mobj.group(g) for g in group)
1259 return mobj.group(group)
1260 elif default is not NO_DEFAULT:
1263 raise RegexNotFoundError('Unable to extract
%s' % _name)
1265 self.report_warning('unable to extract
%s' % _name + bug_reports_message())
1268 def _search_json(self, start_pattern, string, name, video_id, *, end_pattern='',
1269 contains_pattern=r'{(?s:.+)}
', fatal=True, default=NO_DEFAULT, **kwargs):
1270 """Searches string for the JSON object specified by start_pattern"""
1271 # NB: end_pattern is only used to reduce the size of the initial match
1272 if default is NO_DEFAULT:
1273 default, has_default = {}, False
1275 fatal, has_default = False, True
1277 json_string = self._search_regex(
1278 rf'(?
:{start_pattern}
)\s
*(?P
<json
>{contains_pattern}
)\s
*(?
:{end_pattern}
)',
1279 string, name, group='json
', fatal=fatal, default=None if has_default else NO_DEFAULT)
1283 _name = self._downloader._format_err(name, self._downloader.Styles.EMPHASIS)
1285 return self._parse_json(json_string, video_id, ignore_extra=True, **kwargs)
1286 except ExtractorError as e:
1288 raise ExtractorError(
1289 f'Unable to extract {_name}
- Failed to parse JSON
', cause=e.cause, video_id=video_id)
1290 elif not has_default:
1291 self.report_warning(
1292 f'Unable to extract {_name}
- Failed to parse JSON
: {e}
', video_id=video_id)
1295 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1297 Like _search_regex, but strips HTML tags and unescapes entities.
1299 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1300 if isinstance(res, tuple):
1301 return tuple(map(clean_html, res))
1302 return clean_html(res)
1304 def _get_netrc_login_info(self, netrc_machine=None):
1305 netrc_machine = netrc_machine or self._NETRC_MACHINE
1307 cmd = self.get_param('netrc_cmd
')
1309 cmd = cmd.replace('{}', netrc_machine)
1310 self.to_screen(f'Executing command
: {cmd}
')
1311 stdout, _, ret = Popen.run(cmd, text=True, shell=True, stdout=subprocess.PIPE)
1313 raise OSError(f'Command returned error code {ret}
')
1314 info = netrc_from_content(stdout).authenticators(netrc_machine)
1316 elif self.get_param('usenetrc
', False):
1317 netrc_file = compat_expanduser(self.get_param('netrc_location
') or '~
')
1318 if os.path.isdir(netrc_file):
1319 netrc_file = os.path.join(netrc_file, '.netrc
')
1320 info = netrc.netrc(netrc_file).authenticators(netrc_machine)
1325 raise netrc.NetrcParseError(f'No authenticators
for {netrc_machine}
')
1326 return info[0], info[2]
1328 def _get_login_info(self, username_option='username
', password_option='password
', netrc_machine=None):
1330 Get the login info as (username, password)
1331 First look for the manually specified credentials using username_option
1332 and password_option as keys in params dictionary. If no such credentials
1333 are available try the netrc_cmd if it is defined or look in the
1334 netrc file using the netrc_machine or _NETRC_MACHINE value.
1335 If there's no info available
, return (None, None)
1338 username = self.get_param(username_option)
1339 if username is not None:
1340 password = self.get_param(password_option)
1343 username, password = self._get_netrc_login_info(netrc_machine)
1344 except (OSError, netrc.NetrcParseError) as err:
1345 self.report_warning(f'Failed to parse .netrc: {err}')
1347 return username, password
1349 def _get_tfa_info(self, note='two-factor verification code'):
1351 Get the two
-factor authentication info
1352 TODO
- asking the user will be required
for sms
/phone verify
1353 currently just uses the command line option
1354 If there
's no info available, return None
1357 tfa = self.get_param('twofactor
')
1361 return getpass.getpass('Type
%s and press
[Return
]: ' % note)
1363 # Helper functions for extracting OpenGraph info
1365 def _og_regexes(prop):
1366 content_re = r'content
=(?
:"([^"]+?
)"|\'([^\']+?)\'|\s*([^\s"\'=<>`
]+?
)(?
=\s|
/?
>))'
1367 property_re = (r'(?
:name|
property)=(?
:\'og
%(sep)s%(prop)s\'|
"og%(sep)s%(prop)s"|\s
*og
%(sep)s%(prop)s\b)'
1368 % {'prop': re.escape(prop), 'sep': '(?::|[:-])'})
1369 template = r'<meta
[^
>]+?
%s[^
>]+?
%s'
1371 template % (property_re, content_re),
1372 template % (content_re, property_re),
1376 def _meta_regex(prop):
1377 return r'''(?isx)<meta
1378 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1379 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1381 def _og_search_property(self, prop, html, name=None, **kargs):
1382 prop = variadic(prop)
1384 name = 'OpenGraph
%s' % prop[0]
1387 og_regexes.extend(self._og_regexes(p))
1388 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1391 return unescapeHTML(escaped)
1393 def _og_search_thumbnail(self, html, **kargs):
1394 return self._og_search_property('image
', html, 'thumbnail URL
', fatal=False, **kargs)
1396 def _og_search_description(self, html, **kargs):
1397 return self._og_search_property('description
', html, fatal=False, **kargs)
1399 def _og_search_title(self, html, *, fatal=False, **kargs):
1400 return self._og_search_property('title
', html, fatal=fatal, **kargs)
1402 def _og_search_video_url(self, html, name='video url
', secure=True, **kargs):
1403 regexes = self._og_regexes('video
') + self._og_regexes('video
:url
')
1405 regexes = self._og_regexes('video
:secure_url
') + regexes
1406 return self._html_search_regex(regexes, html, name, **kargs)
1408 def _og_search_url(self, html, **kargs):
1409 return self._og_search_property('url
', html, **kargs)
1411 def _html_extract_title(self, html, name='title
', *, fatal=False, **kwargs):
1412 return self._html_search_regex(r'(?s
)<title
\b[^
>]*>([^
<]+)</title
>', html, name, fatal=fatal, **kwargs)
1414 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1415 name = variadic(name)
1416 if display_name is None:
1417 display_name = name[0]
1418 return self._html_search_regex(
1419 [self._meta_regex(n) for n in name],
1420 html, display_name, fatal=fatal, group='content
', **kwargs)
1422 def _dc_search_uploader(self, html):
1423 return self._html_search_meta('dc
.creator
', html, 'uploader
')
1426 def _rta_search(html):
1427 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1428 if re.search(r'(?ix
)<meta\s
+name
="rating"\s
+'
1429 r' content
="RTA-5042-1996-1400-1577-RTA"',
1433 # And then there are the jokers who advertise that they use RTA, but actually don't
.
1434 AGE_LIMIT_MARKERS
= [
1435 r
'Proudly Labeled <a href="http://www\.rtalabel\.org/" title="Restricted to Adults">RTA</a>',
1436 r
'>[^<]*you acknowledge you are at least (\d+) years old',
1437 r
'>\s*(?:18\s+U(?:\.S\.C\.|SC)\s+)?(?:§+\s*)?2257\b',
1441 for marker
in AGE_LIMIT_MARKERS
:
1442 mobj
= re
.search(marker
, html
)
1444 age_limit
= max(age_limit
, int(traverse_obj(mobj
, 1, default
=18)))
1447 def _media_rating_search(self
, html
):
1448 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1449 rating
= self
._html
_search
_meta
('rating', html
)
1461 return RATING_TABLE
.get(rating
.lower())
1463 def _family_friendly_search(self
, html
):
1464 # See http://schema.org/VideoObject
1465 family_friendly
= self
._html
_search
_meta
(
1466 'isFamilyFriendly', html
, default
=None)
1468 if not family_friendly
:
1477 return RATING_TABLE
.get(family_friendly
.lower())
1479 def _twitter_search_player(self
, html
):
1480 return self
._html
_search
_meta
('twitter:player', html
,
1481 'twitter card player')
1483 def _yield_json_ld(self
, html
, video_id
, *, fatal
=True, default
=NO_DEFAULT
):
1484 """Yield all json ld objects in the html"""
1485 if default
is not NO_DEFAULT
:
1487 for mobj
in re
.finditer(JSON_LD_RE
, html
):
1488 json_ld_item
= self
._parse
_json
(mobj
.group('json_ld'), video_id
, fatal
=fatal
)
1489 for json_ld
in variadic(json_ld_item
):
1490 if isinstance(json_ld
, dict):
1493 def _search_json_ld(self
, html
, video_id
, expected_type
=None, *, fatal
=True, default
=NO_DEFAULT
):
1494 """Search for a video in any json ld in the html"""
1495 if default
is not NO_DEFAULT
:
1497 info
= self
._json
_ld
(
1498 list(self
._yield
_json
_ld
(html
, video_id
, fatal
=fatal
, default
=default
)),
1499 video_id
, fatal
=fatal
, expected_type
=expected_type
)
1502 if default
is not NO_DEFAULT
:
1505 raise RegexNotFoundError('Unable to extract JSON-LD')
1507 self
.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1510 def _json_ld(self
, json_ld
, video_id
, fatal
=True, expected_type
=None):
1511 if isinstance(json_ld
, str):
1512 json_ld
= self
._parse
_json
(json_ld
, video_id
, fatal
=fatal
)
1517 INTERACTION_TYPE_MAP
= {
1518 'CommentAction': 'comment',
1519 'AgreeAction': 'like',
1520 'DisagreeAction': 'dislike',
1521 'LikeAction': 'like',
1522 'DislikeAction': 'dislike',
1523 'ListenAction': 'view',
1524 'WatchAction': 'view',
1525 'ViewAction': 'view',
1528 def is_type(e
, *expected_types
):
1529 type = variadic(traverse_obj(e
, '@type'))
1530 return any(x
in type for x
in expected_types
)
1532 def extract_interaction_type(e
):
1533 interaction_type
= e
.get('interactionType')
1534 if isinstance(interaction_type
, dict):
1535 interaction_type
= interaction_type
.get('@type')
1536 return str_or_none(interaction_type
)
1538 def extract_interaction_statistic(e
):
1539 interaction_statistic
= e
.get('interactionStatistic')
1540 if isinstance(interaction_statistic
, dict):
1541 interaction_statistic
= [interaction_statistic
]
1542 if not isinstance(interaction_statistic
, list):
1544 for is_e
in interaction_statistic
:
1545 if not is_type(is_e
, 'InteractionCounter'):
1547 interaction_type
= extract_interaction_type(is_e
)
1548 if not interaction_type
:
1550 # For interaction count some sites provide string instead of
1551 # an integer (as per spec) with non digit characters (e.g. ",")
1552 # so extracting count with more relaxed str_to_int
1553 interaction_count
= str_to_int(is_e
.get('userInteractionCount'))
1554 if interaction_count
is None:
1556 count_kind
= INTERACTION_TYPE_MAP
.get(interaction_type
.split('/')[-1])
1559 count_key
= '%s_count' % count_kind
1560 if info
.get(count_key
) is not None:
1562 info
[count_key
] = interaction_count
1564 def extract_chapter_information(e
):
1566 'title': part
.get('name'),
1567 'start_time': part
.get('startOffset'),
1568 'end_time': part
.get('endOffset'),
1569 } for part
in variadic(e
.get('hasPart') or []) if part
.get('@type') == 'Clip']
1570 for idx
, (last_c
, current_c
, next_c
) in enumerate(zip(
1571 [{'end_time': 0}
] + chapters
, chapters
, chapters
[1:])):
1572 current_c
['end_time'] = current_c
['end_time'] or next_c
['start_time']
1573 current_c
['start_time'] = current_c
['start_time'] or last_c
['end_time']
1574 if None in current_c
.values():
1575 self
.report_warning(f
'Chapter {idx} contains broken data. Not extracting chapters')
1578 chapters
[-1]['end_time'] = chapters
[-1]['end_time'] or info
['duration']
1579 info
['chapters'] = chapters
1581 def extract_video_object(e
):
1582 author
= e
.get('author')
1584 'url': url_or_none(e
.get('contentUrl')),
1585 'ext': mimetype2ext(e
.get('encodingFormat')),
1586 'title': unescapeHTML(e
.get('name')),
1587 'description': unescapeHTML(e
.get('description')),
1588 'thumbnails': [{'url': unescapeHTML(url)}
1589 for url
in variadic(traverse_obj(e
, 'thumbnailUrl', 'thumbnailURL'))
1590 if url_or_none(url
)],
1591 'duration': parse_duration(e
.get('duration')),
1592 'timestamp': unified_timestamp(e
.get('uploadDate')),
1593 # author can be an instance of 'Organization' or 'Person' types.
1594 # both types can have 'name' property(inherited from 'Thing' type). [1]
1595 # however some websites are using 'Text' type instead.
1596 # 1. https://schema.org/VideoObject
1597 'uploader': author
.get('name') if isinstance(author
, dict) else author
if isinstance(author
, str) else None,
1598 'artist': traverse_obj(e
, ('byArtist', 'name'), expected_type
=str),
1599 'filesize': int_or_none(float_or_none(e
.get('contentSize'))),
1600 'tbr': int_or_none(e
.get('bitrate')),
1601 'width': int_or_none(e
.get('width')),
1602 'height': int_or_none(e
.get('height')),
1603 'view_count': int_or_none(e
.get('interactionCount')),
1604 'tags': try_call(lambda: e
.get('keywords').split(',')),
1606 if is_type(e
, 'AudioObject'):
1609 'abr': int_or_none(e
.get('bitrate')),
1611 extract_interaction_statistic(e
)
1612 extract_chapter_information(e
)
1614 def traverse_json_ld(json_ld
, at_top_level
=True):
1615 for e
in variadic(json_ld
):
1616 if not isinstance(e
, dict):
1618 if at_top_level
and '@context' not in e
:
1620 if at_top_level
and set(e
.keys()) == {'@context', '@graph'}
:
1621 traverse_json_ld(e
['@graph'], at_top_level
=False)
1623 if expected_type
is not None and not is_type(e
, expected_type
):
1625 rating
= traverse_obj(e
, ('aggregateRating', 'ratingValue'), expected_type
=float_or_none
)
1626 if rating
is not None:
1627 info
['average_rating'] = rating
1628 if is_type(e
, 'TVEpisode', 'Episode'):
1629 episode_name
= unescapeHTML(e
.get('name'))
1631 'episode': episode_name
,
1632 'episode_number': int_or_none(e
.get('episodeNumber')),
1633 'description': unescapeHTML(e
.get('description')),
1635 if not info
.get('title') and episode_name
:
1636 info
['title'] = episode_name
1637 part_of_season
= e
.get('partOfSeason')
1638 if is_type(part_of_season
, 'TVSeason', 'Season', 'CreativeWorkSeason'):
1640 'season': unescapeHTML(part_of_season
.get('name')),
1641 'season_number': int_or_none(part_of_season
.get('seasonNumber')),
1643 part_of_series
= e
.get('partOfSeries') or e
.get('partOfTVSeries')
1644 if is_type(part_of_series
, 'TVSeries', 'Series', 'CreativeWorkSeries'):
1645 info
['series'] = unescapeHTML(part_of_series
.get('name'))
1646 elif is_type(e
, 'Movie'):
1648 'title': unescapeHTML(e
.get('name')),
1649 'description': unescapeHTML(e
.get('description')),
1650 'duration': parse_duration(e
.get('duration')),
1651 'timestamp': unified_timestamp(e
.get('dateCreated')),
1653 elif is_type(e
, 'Article', 'NewsArticle'):
1655 'timestamp': parse_iso8601(e
.get('datePublished')),
1656 'title': unescapeHTML(e
.get('headline')),
1657 'description': unescapeHTML(e
.get('articleBody') or e
.get('description')),
1659 if is_type(traverse_obj(e
, ('video', 0)), 'VideoObject'):
1660 extract_video_object(e
['video'][0])
1661 elif is_type(traverse_obj(e
, ('subjectOf', 0)), 'VideoObject'):
1662 extract_video_object(e
['subjectOf'][0])
1663 elif is_type(e
, 'VideoObject', 'AudioObject'):
1664 extract_video_object(e
)
1665 if expected_type
is None:
1669 video
= e
.get('video')
1670 if is_type(video
, 'VideoObject'):
1671 extract_video_object(video
)
1672 if expected_type
is None:
1677 traverse_json_ld(json_ld
)
1678 return filter_dict(info
)
1680 def _search_nextjs_data(self
, webpage
, video_id
, *, transform_source
=None, fatal
=True, **kw
):
1681 return self
._parse
_json
(
1683 r
'(?s)<script[^>]+id=[\'"]__NEXT_DATA__[\'"][^
>]*>([^
<]+)</script
>',
1684 webpage, 'next
.js data
', fatal=fatal, **kw),
1685 video_id, transform_source=transform_source, fatal=fatal)
1687 def _search_nuxt_data(self, webpage, video_id, context_name='__NUXT__
', *, fatal=True, traverse=('data
', 0)):
1688 """Parses Nuxt.js metadata. This works as long as the function __NUXT__ invokes is a pure function"""
1689 rectx = re.escape(context_name)
1690 FUNCTION_RE = r'\
(function\
((?P
<arg_keys
>.*?
)\
){return\s+(?P<js>{.*?}
)\s
*;?\s
*}\
((?P
<arg_vals
>.*?
)\
)'
1691 js, arg_keys, arg_vals = self._search_regex(
1692 (rf'<script
>\s
*window\
.{rectx}
={FUNCTION_RE}\s
*\
)\s
*;?\s
*</script
>', rf'{rectx}\
(.*?{FUNCTION_RE}
'),
1693 webpage, context_name, group=('js
', 'arg_keys
', 'arg_vals
'),
1694 default=NO_DEFAULT if fatal else (None, None, None))
1698 args = dict(zip(arg_keys.split(','), map(json.dumps, self._parse_json(
1699 f'[{arg_vals}
]', video_id, transform_source=js_to_json, fatal=fatal) or ())))
1701 ret = self._parse_json(js, video_id, transform_source=functools.partial(js_to_json, vars=args), fatal=fatal)
1702 return traverse_obj(ret, traverse) or {}
1705 def _hidden_inputs(html):
1706 html = re.sub(r'<!--(?
:(?
!<!--).)*-->', '', html)
1708 for input in re.findall(r'(?i
)(<input[^
>]+>)', html):
1709 attrs = extract_attributes(input)
1712 if attrs.get('type') not in ('hidden
', 'submit
'):
1714 name = attrs.get('name
') or attrs.get('id')
1715 value = attrs.get('value
')
1716 if name and value is not None:
1717 hidden_inputs[name] = value
1718 return hidden_inputs
1720 def _form_hidden_inputs(self, form_id, html):
1721 form = self._search_regex(
1722 r'(?
is)<form
[^
>]+?
id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1723 html, '%s form' % form_id, group='form')
1724 return self._hidden_inputs(form)
1726 @classproperty(cache=True)
1727 def FormatSort(cls):
1728 class FormatSort(FormatSorter):
1729 def __init__(ie, *args, **kwargs):
1730 super().__init__(ie._downloader, *args, **kwargs)
1732 deprecation_warning(
1733 'yt_dlp.InfoExtractor.FormatSort is deprecated and may be removed in the future. '
1734 'Use yt_dlp.utils.FormatSorter instead')
1737 def _sort_formats(self, formats, field_preference=[]):
1738 if not field_preference:
1739 self._downloader.deprecation_warning(
1740 'yt_dlp.InfoExtractor._sort_formats is deprecated and is no longer required')
1742 self._downloader.deprecation_warning(
1743 'yt_dlp.InfoExtractor._sort_formats is deprecated and no longer works as expected. '
1744 'Return _format_sort_fields in the info_dict instead')
1746 formats[0]['__sort_fields'] = field_preference
1748 def _check_formats(self, formats, video_id):
1750 formats[:] = filter(
1751 lambda f: self._is_valid_url(
1753 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1757 def _remove_duplicate_formats(formats):
1761 if f['url'] not in format_urls:
1762 format_urls.add(f['url'])
1763 unique_formats.append(f)
1764 formats[:] = unique_formats
1766 def _is_valid_url(self, url, video_id, item='video', headers={}):
1767 url = self._proto_relative_url(url, scheme='http:')
1768 # For now assume non HTTP(S) URLs always valid
1769 if not (url.startswith('http://') or url.startswith('https://')):
1772 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1774 except ExtractorError as e:
1776 '%s: %s URL is invalid, skipping: %s'
1777 % (video_id, item, error_to_compat_str(e.cause)))
1780 def http_scheme(self):
1781 """ Either "http
:" or "https
:", depending on the user's preferences """
1784 if self.get_param('prefer_insecure', False)
1787 def _proto_relative_url(self, url, scheme=None):
1788 scheme = scheme or self.http_scheme()
1789 assert scheme.endswith(':')
1790 return sanitize_url(url, scheme=scheme[:-1])
1792 def _sleep(self, timeout, video_id, msg_template=None):
1793 if msg_template is None:
1794 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1795 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1799 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1800 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1801 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
1802 if self.get_param('ignore_no_formats_error'):
1805 res = self._download_xml_handle(
1806 manifest_url, video_id, 'Downloading f4m manifest',
1807 'Unable to download f4m manifest',
1808 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1809 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
1810 transform_source=transform_source,
1811 fatal=fatal, data=data, headers=headers, query=query)
1815 manifest, urlh = res
1816 manifest_url = urlh.url
1818 return self._parse_f4m_formats(
1819 manifest, manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1820 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1822 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1823 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1824 fatal=True, m3u8_id=None):
1825 if not isinstance(manifest, xml.etree.ElementTree.Element) and not fatal:
1828 # currently yt-dlp cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1829 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1830 if akamai_pv is not None and ';' in akamai_pv.text:
1831 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1832 if playerVerificationChallenge.strip() != '':
1836 manifest_version = '1.0'
1837 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1839 manifest_version = '2.0'
1840 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1841 # Remove unsupported DRM protected media from final formats
1842 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
1843 media_nodes = remove_encrypted_media(media_nodes)
1847 manifest_base_url = get_base_url(manifest)
1849 bootstrap_info = xpath_element(
1850 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1851 'bootstrap info', default=None)
1854 mime_type = xpath_text(
1855 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1856 'base URL', default=None)
1857 if mime_type and mime_type.startswith('audio/'):
1860 for i, media_el in enumerate(media_nodes):
1861 tbr = int_or_none(media_el.attrib.get('bitrate'))
1862 width = int_or_none(media_el.attrib.get('width'))
1863 height = int_or_none(media_el.attrib.get('height'))
1864 format_id = join_nonempty(f4m_id, tbr or i)
1865 # If <bootstrapInfo> is present, the specified f4m is a
1866 # stream-level manifest, and only set-level manifests may refer to
1867 # external resources. See section 11.4 and section 4 of F4M spec
1868 if bootstrap_info is None:
1870 # @href is introduced in 2.0, see section 11.6 of F4M spec
1871 if manifest_version == '2.0':
1872 media_url = media_el.attrib.get('href')
1873 if media_url is None:
1874 media_url = media_el.attrib.get('url')
1878 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1879 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1880 # If media_url is itself a f4m manifest do the recursive extraction
1881 # since bitrates in parent manifest (this one) and media_url manifest
1882 # may differ leading to inability to resolve the format by requested
1883 # bitrate in f4m downloader
1884 ext = determine_ext(manifest_url)
1886 f4m_formats = self._extract_f4m_formats(
1887 manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1888 transform_source=transform_source, fatal=fatal)
1889 # Sometimes stream-level manifest contains single media entry that
1890 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1891 # At the same time parent's media entry in set-level manifest may
1892 # contain it. We will copy it from parent in such cases.
1893 if len(f4m_formats) == 1:
1896 'tbr': f.get('tbr') or tbr,
1897 'width': f.get('width') or width,
1898 'height': f.get('height') or height,
1899 'format_id': f.get('format_id') if not tbr else format_id,
1902 formats.extend(f4m_formats)
1905 formats.extend(self._extract_m3u8_formats(
1906 manifest_url, video_id, 'mp4', preference=preference,
1907 quality=quality, m3u8_id=m3u8_id, fatal=fatal))
1910 'format_id': format_id,
1911 'url': manifest_url,
1912 'manifest_url': manifest_url,
1913 'ext': 'flv' if bootstrap_info is not None else None,
1919 'preference': preference,
1924 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, quality=None, m3u8_id=None):
1926 'format_id': join_nonempty(m3u8_id, 'meta'),
1930 'preference': preference - 100 if preference else -100,
1932 'resolution': 'multiple',
1933 'format_note': 'Quality selection URL',
1936 def _report_ignoring_subs(self, name):
1937 self.report_warning(bug_reports_message(
1938 f'Ignoring subtitle tracks found in the {name} manifest; '
1939 'if any subtitle tracks are missing,'
1942 def _extract_m3u8_formats(self, *args, **kwargs):
1943 fmts, subs = self._extract_m3u8_formats_and_subtitles(*args, **kwargs)
1945 self._report_ignoring_subs('HLS')
1948 def _extract_m3u8_formats_and_subtitles(
1949 self, m3u8_url, video_id, ext=None, entry_protocol='m3u8_native',
1950 preference=None, quality=None, m3u8_id=None, note=None,
1951 errnote=None, fatal=True, live=False, data=None, headers={},
1954 if self.get_param('ignore_no_formats_error'):
1958 if errnote is not False:
1959 errnote = errnote or 'Failed to obtain m3u8 URL'
1961 raise ExtractorError(errnote, video_id=video_id)
1962 self.report_warning(f'{errnote}{bug_reports_message()}')
1965 res = self._download_webpage_handle(
1967 note='Downloading m3u8 information' if note is None else note,
1968 errnote='Failed to download m3u8 information' if errnote is None else errnote,
1969 fatal=fatal, data=data, headers=headers, query=query)
1974 m3u8_doc, urlh = res
1977 return self._parse_m3u8_formats_and_subtitles(
1978 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
1979 preference=preference, quality=quality, m3u8_id=m3u8_id,
1980 note=note, errnote=errnote, fatal=fatal, live=live, data=data,
1981 headers=headers, query=query, video_id=video_id)
1983 def _parse_m3u8_formats_and_subtitles(
1984 self, m3u8_doc, m3u8_url=None, ext=None, entry_protocol='m3u8_native',
1985 preference=None, quality=None, m3u8_id=None, live=False, note=None,
1986 errnote=None, fatal=True, data=None, headers={}, query={},
1988 formats, subtitles = [], {}
1989 has_drm = HlsFD._has_drm(m3u8_doc)
1991 def format_url(url):
1992 return url if re.match(r'^https?://', url) else urllib.parse.urljoin(m3u8_url, url)
1994 if self.get_param('hls_split_discontinuity', False):
1995 def _extract_m3u8_playlist_indices(manifest_url=None, m3u8_doc=None):
1997 if not manifest_url:
1999 m3u8_doc = self._download_webpage(
2000 manifest_url, video_id, fatal=fatal, data=data, headers=headers,
2001 note=False, errnote='Failed to download m3u8 playlist information')
2002 if m3u8_doc is False:
2004 return range(1 + sum(line.startswith('#EXT-X-DISCONTINUITY') for line in m3u8_doc.splitlines()))
2007 def _extract_m3u8_playlist_indices(*args, **kwargs):
2011 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
2012 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
2013 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
2015 # We should try extracting formats only from master playlists [1, 4.3.4],
2016 # i.e. playlists that describe available qualities. On the other hand
2017 # media playlists [1, 4.3.3] should be returned as is since they contain
2018 # just the media without qualities renditions.
2019 # Fortunately, master playlist can be easily distinguished from media
2020 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
2021 # master playlist tags MUST NOT appear in a media playlist and vice versa.
2022 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
2023 # media playlist and MUST NOT appear in master playlist thus we can
2024 # clearly detect media playlist with this criterion.
2026 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
2028 'format_id': join_nonempty(m3u8_id, idx),
2029 'format_index': idx,
2030 'url': m3u8_url or encode_data_uri(m3u8_doc.encode('utf-8'), 'application/x-mpegurl'),
2032 'protocol': entry_protocol,
2033 'preference': preference,
2036 } for idx in _extract_m3u8_playlist_indices(m3u8_doc=m3u8_doc)]
2038 return formats, subtitles
2041 last_stream_inf = {}
2043 def extract_media(x_media_line):
2044 media = parse_m3u8_attributes(x_media_line)
2045 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
2046 media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME')
2047 if not (media_type and group_id and name):
2049 groups.setdefault(group_id, []).append(media)
2050 # <https://tools.ietf.org/html/rfc8216#section-4.3.4.1>
2051 if media_type == 'SUBTITLES':
2052 # According to RFC 8216 §4.3.4.2.1, URI is REQUIRED in the
2053 # EXT-X-MEDIA tag if the media type is SUBTITLES.
2054 # However, lack of URI has been spotted in the wild.
2055 # e.g. NebulaIE; see https://github.com/yt-dlp/yt-dlp/issues/339
2056 if not media.get('URI'):
2058 url = format_url(media['URI'])
2061 'ext': determine_ext(url),
2063 if sub_info['ext'] == 'm3u8':
2064 # Per RFC 8216 §3.1, the only possible subtitle format m3u8
2065 # files may contain is WebVTT:
2066 # <https://tools.ietf.org/html/rfc8216#section-3.1>
2067 sub_info['ext'] = 'vtt'
2068 sub_info['protocol'] = 'm3u8_native'
2069 lang = media.get('LANGUAGE') or 'und'
2070 subtitles.setdefault(lang, []).append(sub_info)
2071 if media_type not in ('VIDEO', 'AUDIO'):
2073 media_url = media.get('URI')
2075 manifest_url = format_url(media_url)
2077 'format_id': join_nonempty(m3u8_id, group_id, name, idx),
2078 'format_note': name,
2079 'format_index': idx,
2080 'url': manifest_url,
2081 'manifest_url': m3u8_url,
2082 'language': media.get('LANGUAGE'),
2084 'protocol': entry_protocol,
2085 'preference': preference,
2088 'vcodec': 'none' if media_type == 'AUDIO' else None,
2089 } for idx in _extract_m3u8_playlist_indices(manifest_url))
2091 def build_stream_name():
2092 # Despite specification does not mention NAME attribute for
2093 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
2094 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
2095 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
2096 stream_name = last_stream_inf.get('NAME')
2099 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
2100 # from corresponding rendition group
2101 stream_group_id = last_stream_inf.get('VIDEO')
2102 if not stream_group_id:
2104 stream_group = groups.get(stream_group_id)
2105 if not stream_group:
2106 return stream_group_id
2107 rendition = stream_group[0]
2108 return rendition.get('NAME') or stream_group_id
2110 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
2111 # chance to detect video only formats when EXT-X-STREAM-INF tags
2112 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
2113 for line in m3u8_doc.splitlines():
2114 if line.startswith('#EXT-X-MEDIA:'):
2117 for line in m3u8_doc.splitlines():
2118 if line.startswith('#EXT-X-STREAM-INF:'):
2119 last_stream_inf = parse_m3u8_attributes(line)
2120 elif line.startswith('#') or not line.strip():
2123 tbr = float_or_none(
2124 last_stream_inf.get('AVERAGE-BANDWIDTH')
2125 or last_stream_inf.get('BANDWIDTH'), scale=1000)
2126 manifest_url = format_url(line.strip())
2128 for idx in _extract_m3u8_playlist_indices(manifest_url):
2129 format_id = [m3u8_id, None, idx]
2130 # Bandwidth of live streams may differ over time thus making
2131 # format_id unpredictable. So it's better to keep provided
2134 stream_name = build_stream_name()
2135 format_id[1] = stream_name or '%d' % (tbr or len(formats))
2137 'format_id': join_nonempty(*format_id),
2138 'format_index': idx,
2139 'url': manifest_url,
2140 'manifest_url': m3u8_url,
2143 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')),
2144 'protocol': entry_protocol,
2145 'preference': preference,
2149 resolution = last_stream_inf.get('RESOLUTION')
2151 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
2153 f['width'] = int(mobj.group('width'))
2154 f['height'] = int(mobj.group('height'))
2155 # Unified Streaming Platform
2157 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
2159 abr, vbr = mobj.groups()
2160 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
2165 codecs = parse_codecs(last_stream_inf.get('CODECS'))
2167 audio_group_id = last_stream_inf.get('AUDIO')
2168 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
2169 # references a rendition group MUST have a CODECS attribute.
2170 # However, this is not always respected. E.g. [2]
2171 # contains EXT-X-STREAM-INF tag which references AUDIO
2172 # rendition group but does not have CODECS and despite
2173 # referencing an audio group it represents a complete
2174 # (with audio and video) format. So, for such cases we will
2175 # ignore references to rendition groups and treat them
2176 # as complete formats.
2177 if audio_group_id and codecs and f.get('vcodec') != 'none':
2178 audio_group = groups.get(audio_group_id)
2179 if audio_group and audio_group[0].get('URI'):
2180 # TODO: update acodec for audio only formats with
2182 f['acodec'] = 'none'
2183 if not f.get('ext'):
2184 f['ext'] = 'm4a' if f.get('vcodec') == 'none' else 'mp4'
2188 progressive_uri = last_stream_inf.get('PROGRESSIVE-URI')
2191 del http_f['manifest_url']
2193 'format_id': f['format_id'].replace('hls-', 'http-'),
2195 'url': progressive_uri,
2197 formats.append(http_f)
2199 last_stream_inf = {}
2200 return formats, subtitles
2202 def _extract_m3u8_vod_duration(
2203 self, m3u8_vod_url, video_id, note=None, errnote=None, data=None, headers={}, query={}):
2205 m3u8_vod = self._download_webpage(
2206 m3u8_vod_url, video_id,
2207 note='Downloading m3u8 VOD manifest' if note is None else note,
2208 errnote='Failed to download VOD manifest' if errnote is None else errnote,
2209 fatal=False, data=data, headers=headers, query=query)
2211 return self._parse_m3u8_vod_duration(m3u8_vod or '', video_id)
2213 def _parse_m3u8_vod_duration(self, m3u8_vod, video_id):
2214 if '#EXT-X-ENDLIST' not in m3u8_vod:
2218 float(line[len('#EXTINF:'):].split(',')[0])
2219 for line in m3u8_vod.splitlines() if line.startswith('#EXTINF:'))) or None
2221 def _extract_mpd_vod_duration(
2222 self, mpd_url, video_id, note=None, errnote=None, data=None, headers={}, query={}):
2224 mpd_doc = self._download_xml(
2226 note='Downloading MPD VOD manifest' if note is None else note,
2227 errnote='Failed to download VOD manifest' if errnote is None else errnote,
2228 fatal=False, data=data, headers=headers, query=query) or {}
2229 return int_or_none(parse_duration(mpd_doc.get('mediaPresentationDuration')))
2232 def _xpath_ns(path, namespace=None):
2236 for c in path.split('/'):
2237 if not c or c == '.':
2240 out.append('{%s}%s' % (namespace, c))
2241 return '/'.join(out)
2243 def _extract_smil_formats_and_subtitles(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
2244 if self.get_param('ignore_no_formats_error'):
2247 res = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
2255 namespace = self._parse_smil_namespace(smil)
2257 fmts = self._parse_smil_formats(
2258 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
2259 subs = self._parse_smil_subtitles(
2260 smil, namespace=namespace)
2264 def _extract_smil_formats(self, *args, **kwargs):
2265 fmts, subs = self._extract_smil_formats_and_subtitles(*args, **kwargs)
2267 self._report_ignoring_subs('SMIL')
2270 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
2271 res = self._download_smil(smil_url, video_id, fatal=fatal)
2278 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
2280 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
2281 return self._download_xml_handle(
2282 smil_url, video_id, 'Downloading SMIL file',
2283 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
2285 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
2286 namespace = self._parse_smil_namespace(smil)
2288 formats = self._parse_smil_formats(
2289 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
2290 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
2292 video_id = os.path.splitext(url_basename(smil_url))[0]
2296 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
2297 name = meta.attrib.get('name')
2298 content = meta.attrib.get('content')
2299 if not name or not content:
2301 if not title and name == 'title':
2303 elif not description and name in ('description', 'abstract'):
2304 description = content
2305 elif not upload_date and name == 'date':
2306 upload_date = unified_strdate(content)
2309 'id': image.get('type'),
2310 'url': image.get('src'),
2311 'width': int_or_none(image.get('width')),
2312 'height': int_or_none(image.get('height')),
2313 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
2317 'title': title or video_id,
2318 'description': description,
2319 'upload_date': upload_date,
2320 'thumbnails': thumbnails,
2322 'subtitles': subtitles,
2325 def _parse_smil_namespace(self, smil):
2326 return self._search_regex(
2327 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
2329 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
2331 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
2332 b = meta.get('base') or meta.get('httpBase')
2344 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
2345 for medium in media:
2346 src = medium.get('src')
2347 if not src or src in srcs:
2351 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
2352 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
2353 width = int_or_none(medium.get('width'))
2354 height = int_or_none(medium.get('height'))
2355 proto = medium.get('proto')
2356 ext = medium.get('ext')
2357 src_ext = determine_ext(src, default_ext=None) or ext or urlhandle_detect_ext(
2358 self._request_webpage(HEADRequest(src), video_id, note='Requesting extension info', fatal=False))
2359 streamer = medium.get('streamer') or base
2361 if proto == 'rtmp' or streamer.startswith('rtmp'):
2367 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
2369 'filesize': filesize,
2373 if transform_rtmp_url:
2374 streamer, src = transform_rtmp_url(streamer, src)
2375 formats[-1].update({
2381 src_url = src if src.startswith('http') else urllib.parse.urljoin(base, src)
2382 src_url = src_url.strip()
2384 if proto == 'm3u8' or src_ext == 'm3u8':
2385 m3u8_formats = self._extract_m3u8_formats(
2386 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
2387 if len(m3u8_formats) == 1:
2389 m3u8_formats[0].update({
2390 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
2395 formats.extend(m3u8_formats)
2396 elif src_ext == 'f4m':
2401 'plugin': 'flowplayer-3.2.0.1',
2403 f4m_url += '&' if '?' in f4m_url else '?'
2404 f4m_url += urllib.parse.urlencode(f4m_params)
2405 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
2406 elif src_ext == 'mpd':
2407 formats.extend(self._extract_mpd_formats(
2408 src_url, video_id, mpd_id='dash', fatal=False))
2409 elif re.search(r'\.ism/[Mm]anifest', src_url):
2410 formats.extend(self._extract_ism_formats(
2411 src_url, video_id, ism_id='mss', fatal=False))
2412 elif src_url.startswith('http') and self._is_valid_url(src, video_id):
2416 'ext': ext or src_ext or 'flv',
2417 'format_id': 'http-%d' % (bitrate or http_count),
2419 'filesize': filesize,
2424 for medium in smil.findall(self._xpath_ns('.//imagestream', namespace)):
2425 src = medium.get('src')
2426 if not src or src in srcs:
2432 'format_id': 'imagestream-%d' % (imgs_count),
2434 'ext': mimetype2ext(medium.get('type')),
2437 'width': int_or_none(medium.get('width')),
2438 'height': int_or_none(medium.get('height')),
2439 'format_note': 'SMIL storyboards',
2444 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
2447 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
2448 src = textstream.get('src')
2449 if not src or src in urls:
2452 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
2453 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
2454 subtitles.setdefault(lang, []).append({
2460 def _extract_xspf_playlist(self, xspf_url, playlist_id, fatal=True):
2461 res = self._download_xml_handle(
2462 xspf_url, playlist_id, 'Downloading xpsf playlist',
2463 'Unable to download xspf manifest', fatal=fatal)
2470 return self._parse_xspf(
2471 xspf, playlist_id, xspf_url=xspf_url,
2472 xspf_base_url=base_url(xspf_url))
2474 def _parse_xspf(self, xspf_doc, playlist_id, xspf_url=None, xspf_base_url=None):
2476 'xspf': 'http://xspf.org/ns/0/',
2477 's1': 'http://static.streamone.nl/player/ns/0',
2481 for track in xspf_doc.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
2483 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
2484 description = xpath_text(
2485 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
2486 thumbnail = xpath_text(
2487 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
2488 duration = float_or_none(
2489 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
2492 for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP)):
2493 format_url = urljoin(xspf_base_url, location.text)
2498 'manifest_url': xspf_url,
2499 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
2500 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
2501 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
2507 'description': description,
2508 'thumbnail': thumbnail,
2509 'duration': duration,
2514 def _extract_mpd_formats(self, *args, **kwargs):
2515 fmts, subs = self._extract_mpd_formats_and_subtitles(*args, **kwargs)
2517 self._report_ignoring_subs('DASH')
2520 def _extract_mpd_formats_and_subtitles(
2521 self, mpd_url, video_id, mpd_id=None, note=None, errnote=None,
2522 fatal=True, data=None, headers={}, query={}):
2524 if self.get_param('ignore_no_formats_error'):
2527 res = self._download_xml_handle(
2529 note='Downloading MPD manifest' if note is None else note,
2530 errnote='Failed to download MPD manifest' if errnote is None else errnote,
2531 fatal=fatal, data=data, headers=headers, query=query)
2538 # We could have been redirected to a new url when we retrieved our mpd file.
2540 mpd_base_url = base_url(mpd_url)
2542 return self._parse_mpd_formats_and_subtitles(
2543 mpd_doc, mpd_id, mpd_base_url, mpd_url)
2545 def _parse_mpd_formats(self, *args, **kwargs):
2546 fmts, subs = self._parse_mpd_formats_and_subtitles(*args, **kwargs)
2548 self._report_ignoring_subs('DASH')
2551 def _parse_mpd_formats_and_subtitles(
2552 self, mpd_doc, mpd_id=None, mpd_base_url='', mpd_url=None):
2554 Parse formats from MPD manifest.
2556 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
2557 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
2558 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
2560 if not self.get_param('dynamic_mpd', True):
2561 if mpd_doc.get('type') == 'dynamic':
2564 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
2567 return self._xpath_ns(path, namespace)
2569 def is_drm_protected(element):
2570 return element.find(_add_ns('ContentProtection')) is not None
2572 def extract_multisegment_info(element, ms_parent_info):
2573 ms_info = ms_parent_info.copy()
2575 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2576 # common attributes and elements. We will only extract relevant
2578 def extract_common(source):
2579 segment_timeline = source.find(_add_ns('SegmentTimeline'))
2580 if segment_timeline is not None:
2581 s_e = segment_timeline.findall(_add_ns('S'))
2583 ms_info['total_number'] = 0
2586 r = int(s.get('r', 0))
2587 ms_info['total_number'] += 1 + r
2588 ms_info['s'].append({
2589 't': int(s.get('t', 0)),
2590 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2591 'd': int(s.attrib['d']),
2594 start_number = source.get('startNumber')
2596 ms_info['start_number'] = int(start_number)
2597 timescale = source.get('timescale')
2599 ms_info['timescale'] = int(timescale)
2600 segment_duration = source.get('duration')
2601 if segment_duration:
2602 ms_info['segment_duration'] = float(segment_duration)
2604 def extract_Initialization(source):
2605 initialization = source.find(_add_ns('Initialization'))
2606 if initialization is not None:
2607 ms_info['initialization_url'] = initialization.attrib['sourceURL']
2609 segment_list = element.find(_add_ns('SegmentList'))
2610 if segment_list is not None:
2611 extract_common(segment_list)
2612 extract_Initialization(segment_list)
2613 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
2615 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
2617 segment_template = element.find(_add_ns('SegmentTemplate'))
2618 if segment_template is not None:
2619 extract_common(segment_template)
2620 media = segment_template.get('media')
2622 ms_info['media'] = media
2623 initialization = segment_template.get('initialization')
2625 ms_info['initialization'] = initialization
2627 extract_Initialization(segment_template)
2630 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
2631 formats, subtitles = [], {}
2632 stream_numbers = collections.defaultdict(int)
2633 for period in mpd_doc.findall(_add_ns('Period')):
2634 period_duration = parse_duration(period.get('duration')) or mpd_duration
2635 period_ms_info = extract_multisegment_info(period, {
2639 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
2640 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
2641 for representation in adaptation_set.findall(_add_ns('Representation')):
2642 representation_attrib = adaptation_set.attrib.copy()
2643 representation_attrib.update(representation.attrib)
2644 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2645 mime_type = representation_attrib['mimeType']
2646 content_type = representation_attrib.get('contentType', mime_type.split('/')[0])
2648 codec_str = representation_attrib.get('codecs', '')
2649 # Some kind of binary subtitle found in some youtube livestreams
2650 if mime_type == 'application/x-rawcc':
2651 codecs = {'scodec': codec_str}
2653 codecs = parse_codecs(codec_str)
2654 if content_type not in ('video', 'audio', 'text'):
2655 if mime_type == 'image/jpeg':
2656 content_type = mime_type
2657 elif codecs.get('vcodec', 'none') != 'none':
2658 content_type = 'video'
2659 elif codecs.get('acodec', 'none') != 'none':
2660 content_type = 'audio'
2661 elif codecs.get('scodec', 'none') != 'none':
2662 content_type = 'text'
2663 elif mimetype2ext(mime_type) in ('tt', 'dfxp', 'ttml', 'xml', 'json'):
2664 content_type = 'text'
2666 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
2670 for element in (representation, adaptation_set, period, mpd_doc):
2671 base_url_e = element.find(_add_ns('BaseURL'))
2672 if try_call(lambda: base_url_e.text) is not None:
2673 base_url = base_url_e.text + base_url
2674 if re.match(r'^https?://', base_url):
2676 if mpd_base_url and base_url.startswith('/'):
2677 base_url = urllib.parse.urljoin(mpd_base_url, base_url)
2678 elif mpd_base_url and not re.match(r'^https?://', base_url):
2679 if not mpd_base_url.endswith('/'):
2681 base_url = mpd_base_url + base_url
2682 representation_id = representation_attrib.get('id')
2683 lang = representation_attrib.get('lang')
2684 url_el = representation.find(_add_ns('BaseURL'))
2685 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
2686 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
2687 if representation_id is not None:
2688 format_id = representation_id
2690 format_id = content_type
2692 format_id = mpd_id + '-' + format_id
2693 if content_type in ('video', 'audio'):
2695 'format_id': format_id,
2696 'manifest_url': mpd_url,
2697 'ext': mimetype2ext(mime_type),
2698 'width': int_or_none(representation_attrib.get('width')),
2699 'height': int_or_none(representation_attrib.get('height')),
2700 'tbr': float_or_none(bandwidth, 1000),
2701 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
2702 'fps': int_or_none(representation_attrib.get('frameRate')),
2703 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
2704 'format_note': 'DASH %s' % content_type,
2705 'filesize': filesize,
2706 'container': mimetype2ext(mime_type) + '_dash',
2709 elif content_type == 'text':
2711 'ext': mimetype2ext(mime_type),
2712 'manifest_url': mpd_url,
2713 'filesize': filesize,
2715 elif content_type == 'image/jpeg':
2716 # See test case in VikiIE
2717 # https://www.viki.com/videos/1175236v-choosing-spouse-by-lottery-episode-1
2719 'format_id': format_id,
2721 'manifest_url': mpd_url,
2722 'format_note': 'DASH storyboards (jpeg)',
2726 if is_drm_protected(adaptation_set) or is_drm_protected(representation):
2728 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
2730 def prepare_template(template_name, identifiers):
2731 tmpl = representation_ms_info[template_name]
2732 if representation_id is not None:
2733 tmpl = tmpl.replace('$RepresentationID$', representation_id)
2734 # First of, % characters outside $...$ templates
2735 # must be escaped by doubling for proper processing
2736 # by % operator string formatting used further (see
2737 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2743 in_template = not in_template
2744 elif c == '%' and not in_template:
2746 # Next, $...$ templates are translated to their
2747 # %(...) counterparts to be used with % operator
2748 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
2749 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
2750 t.replace('$$', '$')
2753 # @initialization is a regular template like @media one
2754 # so it should be handled just the same way (see
2755 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2756 if 'initialization' in representation_ms_info:
2757 initialization_template = prepare_template(
2759 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2760 # $Time$ shall not be included for @initialization thus
2761 # only $Bandwidth$ remains
2763 representation_ms_info['initialization_url'] = initialization_template % {
2764 'Bandwidth': bandwidth,
2767 def location_key(location):
2768 return 'url' if re.match(r'^https?://', location) else 'path'
2770 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
2772 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2773 media_location_key = location_key(media_template)
2775 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2776 # can't be used at the same time
2777 if '%(Number' in media_template and 's' not in representation_ms_info:
2778 segment_duration = None
2779 if 'total_number' not in representation_ms_info and 'segment_duration' in representation_ms_info:
2780 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
2781 representation_ms_info['total_number'] = int(math.ceil(
2782 float_or_none(period_duration, segment_duration, default=0)))
2783 representation_ms_info['fragments'] = [{
2784 media_location_key: media_template % {
2785 'Number': segment_number,
2786 'Bandwidth': bandwidth,
2788 'duration': segment_duration,
2789 } for segment_number in range(
2790 representation_ms_info['start_number'],
2791 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
2793 # $Number*$ or $Time$ in media template with S list available
2794 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2795 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2796 representation_ms_info['fragments'] = []
2799 segment_number = representation_ms_info['start_number']
2801 def add_segment_url():
2802 segment_url = media_template % {
2803 'Time': segment_time,
2804 'Bandwidth': bandwidth,
2805 'Number': segment_number,
2807 representation_ms_info['fragments'].append({
2808 media_location_key: segment_url,
2809 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
2812 for num, s in enumerate(representation_ms_info['s']):
2813 segment_time = s.get('t') or segment_time
2817 for r in range(s.get('r', 0)):
2818 segment_time += segment_d
2821 segment_time += segment_d
2822 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
2823 # No media template,
2824 # e.g. https://www.youtube.com/watch?v=iXZV5uAYMJI
2825 # or any YouTube dashsegments video
2828 timescale = representation_ms_info['timescale']
2829 for s in representation_ms_info['s']:
2830 duration = float_or_none(s['d'], timescale)
2831 for r in range(s.get('r', 0) + 1):
2832 segment_uri = representation_ms_info['segment_urls'][segment_index]
2834 location_key(segment_uri): segment_uri,
2835 'duration': duration,
2838 representation_ms_info['fragments'] = fragments
2839 elif 'segment_urls' in representation_ms_info:
2840 # Segment URLs with no SegmentTimeline
2841 # E.g. https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2842 # https://github.com/ytdl-org/youtube-dl/pull/14844
2844 segment_duration = float_or_none(
2845 representation_ms_info['segment_duration'],
2846 representation_ms_info['timescale']) if 'segment_duration' in representation_ms_info else None
2847 for segment_url in representation_ms_info['segment_urls']:
2849 location_key(segment_url): segment_url,
2851 if segment_duration:
2852 fragment['duration'] = segment_duration
2853 fragments.append(fragment)
2854 representation_ms_info['fragments'] = fragments
2855 # If there is a fragments key available then we correctly recognized fragmented media.
2856 # Otherwise we will assume unfragmented media with direct access. Technically, such
2857 # assumption is not necessarily correct since we may simply have no support for
2858 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2859 if 'fragments' in representation_ms_info:
2861 # NB: mpd_url may be empty when MPD manifest is parsed from a string
2862 'url': mpd_url or base_url,
2863 'fragment_base_url': base_url,
2865 'protocol': 'http_dash_segments' if mime_type != 'image/jpeg' else 'mhtml',
2867 if 'initialization_url' in representation_ms_info:
2868 initialization_url = representation_ms_info['initialization_url']
2869 if not f.get('url'):
2870 f['url'] = initialization_url
2871 f['fragments'].append({location_key(initialization_url): initialization_url})
2872 f['fragments'].extend(representation_ms_info['fragments'])
2873 if not period_duration:
2874 period_duration = try_get(
2875 representation_ms_info,
2876 lambda r: sum(frag['duration'] for frag in r['fragments']), float)
2878 # Assuming direct URL to unfragmented media.
2880 if content_type in ('video', 'audio', 'image/jpeg'):
2881 f['manifest_stream_number'] = stream_numbers[f['url']]
2882 stream_numbers[f['url']] += 1
2884 elif content_type == 'text':
2885 subtitles.setdefault(lang or 'und', []).append(f)
2887 return formats, subtitles
2889 def _extract_ism_formats(self, *args, **kwargs):
2890 fmts, subs = self._extract_ism_formats_and_subtitles(*args, **kwargs)
2892 self._report_ignoring_subs('ISM')
2895 def _extract_ism_formats_and_subtitles(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
2896 if self.get_param('ignore_no_formats_error'):
2899 res = self._download_xml_handle(
2901 note='Downloading ISM manifest' if note is None else note,
2902 errnote='Failed to download ISM manifest' if errnote is None else errnote,
2903 fatal=fatal, data=data, headers=headers, query=query)
2910 return self._parse_ism_formats_and_subtitles(ism_doc, urlh.url, ism_id)
2912 def _parse_ism_formats_and_subtitles(self, ism_doc, ism_url, ism_id=None):
2914 Parse formats from ISM manifest.
2916 1. [MS-SSTR]: Smooth Streaming Protocol,
2917 https://msdn.microsoft.com/en-us/library/ff469518.aspx
2919 if ism_doc.get('IsLive') == 'TRUE':
2922 duration = int(ism_doc.attrib['Duration'])
2923 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
2927 for stream in ism_doc.findall('StreamIndex'):
2928 stream_type = stream.get('Type')
2929 if stream_type not in ('video', 'audio', 'text'):
2931 url_pattern = stream.attrib['Url']
2932 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
2933 stream_name = stream.get('Name')
2934 stream_language = stream.get('Language', 'und')
2935 for track in stream.findall('QualityLevel'):
2936 KNOWN_TAGS = {'255': 'AACL', '65534': 'EC-3'}
2937 fourcc = track.get('FourCC') or KNOWN_TAGS.get(track.get('AudioTag'))
2938 # TODO: add support for WVC1 and WMAP
2939 if fourcc not in ('H264', 'AVC1', 'AACL', 'TTML', 'EC-3'):
2940 self.report_warning('%s is not a supported codec' % fourcc)
2942 tbr = int(track.attrib['Bitrate']) // 1000
2943 # [1] does not mention Width and Height attributes. However,
2944 # they're often present while MaxWidth and MaxHeight are
2945 # missing, so should be used as fallbacks
2946 width = int_or_none(track.get('MaxWidth') or track.get('Width'))
2947 height = int_or_none(track.get('MaxHeight') or track.get('Height'))
2948 sampling_rate = int_or_none(track.get('SamplingRate'))
2950 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
2951 track_url_pattern = urllib.parse.urljoin(ism_url, track_url_pattern)
2957 stream_fragments = stream.findall('c')
2958 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
2959 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
2960 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
2961 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
2962 if not fragment_ctx['duration']:
2964 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
2966 next_fragment_time = duration
2967 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
2968 for _ in range(fragment_repeat):
2970 'url': re.sub(r'{start[ _]time}', str(fragment_ctx['time']), track_url_pattern),
2971 'duration': fragment_ctx['duration'] / stream_timescale,
2973 fragment_ctx['time'] += fragment_ctx['duration']
2975 if stream_type == 'text':
2976 subtitles.setdefault(stream_language, []).append({
2980 'manifest_url': ism_url,
2981 'fragments': fragments,
2982 '_download_params': {
2983 'stream_type': stream_type,
2984 'duration': duration,
2985 'timescale': stream_timescale,
2987 'language': stream_language,
2988 'codec_private_data': track.get('CodecPrivateData'),
2991 elif stream_type in ('video', 'audio'):
2993 'format_id': join_nonempty(ism_id, stream_name, tbr),
2995 'manifest_url': ism_url,
2996 'ext': 'ismv' if stream_type == 'video' else 'isma',
3000 'asr': sampling_rate,
3001 'vcodec': 'none' if stream_type == 'audio' else fourcc,
3002 'acodec': 'none' if stream_type == 'video' else fourcc,
3004 'fragments': fragments,
3005 'has_drm': ism_doc.find('Protection') is not None,
3006 'language': stream_language,
3007 'audio_channels': int_or_none(track.get('Channels')),
3008 '_download_params': {
3009 'stream_type': stream_type,
3010 'duration': duration,
3011 'timescale': stream_timescale,
3012 'width': width or 0,
3013 'height': height or 0,
3015 'language': stream_language,
3016 'codec_private_data': track.get('CodecPrivateData'),
3017 'sampling_rate': sampling_rate,
3018 'channels': int_or_none(track.get('Channels', 2)),
3019 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
3020 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
3023 return formats, subtitles
3025 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):
3026 def absolute_url(item_url):
3027 return urljoin(base_url, item_url)
3029 def parse_content_type(content_type):
3030 if not content_type:
3032 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?
(?P
<codecs
>[^
"]+))?', content_type)
3034 mimetype, codecs = ctr.groups()
3035 f = parse_codecs(codecs)
3036 f['ext'] = mimetype2ext(mimetype)
3040 def _media_formats(src, cur_media_type, type_info=None):
3041 type_info = type_info or {}
3042 full_url = absolute_url(src)
3043 ext = type_info.get('ext') or determine_ext(full_url)
3045 is_plain_url = False
3046 formats = self._extract_m3u8_formats(
3047 full_url, video_id, ext='mp4',
3048 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
3049 preference=preference, quality=quality, fatal=False)
3051 is_plain_url = False
3052 formats = self._extract_mpd_formats(
3053 full_url, video_id, mpd_id=mpd_id, fatal=False)
3058 'vcodec': 'none' if cur_media_type == 'audio' else None,
3061 return is_plain_url, formats
3064 # amp-video and amp-audio are very similar to their HTML5 counterparts
3065 # so we will include them right here (see
3066 # https://www.ampproject.org/docs/reference/components/amp-video)
3067 # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/
3068 _MEDIA_TAG_NAME_RE = r'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)'
3069 media_tags = [(media_tag, media_tag_name, media_type, '')
3070 for media_tag, media_tag_name, media_type
3071 in re.findall(r'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE, webpage)]
3072 media_tags.extend(re.findall(
3073 # We only allow video|audio followed by a whitespace or '>'.
3074 # Allowing more characters may end up in significant slow down (see
3075 # https://github.com/ytdl-org/youtube-dl/issues/11979,
3076 # e.g. http://www.porntrex.com/maps/videositemap.xml).
3077 r'(?s)(<(?P<tag>%s)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>' % _MEDIA_TAG_NAME_RE, webpage))
3078 for media_tag, _, media_type, media_content in media_tags:
3083 media_attributes = extract_attributes(media_tag)
3084 src = strip_or_none(dict_get(media_attributes, ('src', 'data-video-src', 'data-src', 'data-source')))
3086 f = parse_content_type(media_attributes.get('type'))
3087 _, formats = _media_formats(src, media_type, f)
3088 media_info['formats'].extend(formats)
3089 media_info['thumbnail'] = absolute_url(media_attributes.get('poster'))
3091 for source_tag in re.findall(r'<source[^>]+>', media_content):
3092 s_attr = extract_attributes(source_tag)
3093 # data-video-src and data-src are non standard but seen
3094 # several times in the wild
3095 src = strip_or_none(dict_get(s_attr, ('src', 'data-video-src', 'data-src', 'data-source')))
3098 f = parse_content_type(s_attr.get('type'))
3099 is_plain_url, formats = _media_formats(src, media_type, f)
3101 # width, height, res, label and title attributes are
3102 # all not standard but seen several times in the wild
3105 for lbl in ('label', 'title')
3106 if str_or_none(s_attr.get(lbl))
3108 width = int_or_none(s_attr.get('width'))
3109 height = (int_or_none(s_attr.get('height'))
3110 or int_or_none(s_attr.get('res')))
3111 if not width or not height:
3113 resolution = parse_resolution(lbl)
3116 width = width or resolution.get('width')
3117 height = height or resolution.get('height')
3119 tbr = parse_bitrate(lbl)
3128 'format_id': s_attr.get('label') or s_attr.get('title'),
3130 f.update(formats[0])
3131 media_info['formats'].append(f)
3133 media_info['formats'].extend(formats)
3134 for track_tag in re.findall(r'<track[^>]+>', media_content):
3135 track_attributes = extract_attributes(track_tag)
3136 kind = track_attributes.get('kind')
3137 if not kind or kind in ('subtitles', 'captions'):
3138 src = strip_or_none(track_attributes.get('src'))
3141 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
3142 media_info['subtitles'].setdefault(lang, []).append({
3143 'url': absolute_url(src),
3145 for f in media_info['formats']:
3146 f.setdefault('http_headers', {})['Referer'] = base_url
3147 if media_info['formats'] or media_info['subtitles']:
3148 entries.append(media_info)
3151 def _extract_akamai_formats(self, *args, **kwargs):
3152 fmts, subs = self._extract_akamai_formats_and_subtitles(*args, **kwargs)
3154 self._report_ignoring_subs('akamai')
3157 def _extract_akamai_formats_and_subtitles(self, manifest_url, video_id, hosts={}):
3158 signed = 'hdnea=' in manifest_url
3160 # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
3161 manifest_url = re.sub(
3162 r'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
3163 '', manifest_url).strip('?')
3168 hdcore_sign = 'hdcore=3.7.0'
3169 f4m_url = re.sub(r'(https?://[^/]+)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
3170 hds_host = hosts.get('hds')
3172 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
3173 if 'hdcore=' not in f4m_url:
3174 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
3175 f4m_formats = self._extract_f4m_formats(
3176 f4m_url, video_id, f4m_id='hds', fatal=False)
3177 for entry in f4m_formats:
3178 entry.update({'extra_param_to_segment_url': hdcore_sign})
3179 formats.extend(f4m_formats)
3181 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
3182 hls_host = hosts.get('hls')
3184 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
3185 m3u8_formats, m3u8_subtitles = self._extract_m3u8_formats_and_subtitles(
3186 m3u8_url, video_id, 'mp4', 'm3u8_native',
3187 m3u8_id='hls', fatal=False)
3188 formats.extend(m3u8_formats)
3189 subtitles = self._merge_subtitles(subtitles, m3u8_subtitles)
3191 http_host = hosts.get('http')
3192 if http_host and m3u8_formats and not signed:
3193 REPL_REGEX = r'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
3194 qualities = re.match(REPL_REGEX, m3u8_url).group(2).split(',')
3195 qualities_length = len(qualities)
3196 if len(m3u8_formats) in (qualities_length, qualities_length + 1):
3198 for f in m3u8_formats:
3199 if f['vcodec'] != 'none':
3200 for protocol in ('http', 'https'):
3202 del http_f['manifest_url']
3204 REPL_REGEX, protocol + fr'://{http_host}/\g<1>{qualities[i]}\3', f['url'])
3206 'format_id': http_f['format_id'].replace('hls-', protocol + '-'),
3208 'protocol': protocol,
3210 formats.append(http_f)
3213 return formats, subtitles
3215 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
3216 query = urllib.parse.urlparse(url).query
3217 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
3219 r'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url)
3220 url_base = mobj.group('url')
3221 http_base_url = '%s%s:%s' % ('http', mobj.group('s') or '', url_base)
3224 def manifest_url(manifest):
3225 m_url = f'{http_base_url}/{manifest}'
3227 m_url += '?%s' % query
3230 if 'm3u8' not in skip_protocols:
3231 formats.extend(self._extract_m3u8_formats(
3232 manifest_url('playlist.m3u8'), video_id, 'mp4',
3233 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
3234 if 'f4m' not in skip_protocols:
3235 formats.extend(self._extract_f4m_formats(
3236 manifest_url('manifest.f4m'),
3237 video_id, f4m_id='hds', fatal=False))
3238 if 'dash' not in skip_protocols:
3239 formats.extend(self._extract_mpd_formats(
3240 manifest_url('manifest.mpd'),
3241 video_id, mpd_id='dash', fatal=False))
3242 if re.search(r'(?:/smil:|\.smil)', url_base):
3243 if 'smil' not in skip_protocols:
3244 rtmp_formats = self._extract_smil_formats(
3245 manifest_url('jwplayer.smil'),
3246 video_id, fatal=False)
3247 for rtmp_format in rtmp_formats:
3248 rtsp_format = rtmp_format.copy()
3249 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
3250 del rtsp_format['play_path']
3251 del rtsp_format['ext']
3252 rtsp_format.update({
3253 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
3254 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
3257 formats.extend([rtmp_format, rtsp_format])
3259 for protocol in ('rtmp', 'rtsp'):
3260 if protocol not in skip_protocols:
3262 'url': f'{protocol}:{url_base}',
3263 'format_id': protocol,
3264 'protocol': protocol,
3268 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
3270 r'''(?s)jwplayer\s*\(\s*(?P<q>'|")(?
!(?P
=q
)).+(?P
=q
)\s
*\
)(?
!</script
>).*?\
.\s
*setup\s
*\
(\s
*(?P
<options
>(?
:\
([^
)]*\
)|
[^
)])+)\s
*\
)''',
3274 jwplayer_data = self._parse_json(mobj.group('options'),
3276 transform_source=transform_source)
3277 except ExtractorError:
3280 if isinstance(jwplayer_data, dict):
3281 return jwplayer_data
3283 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
3284 jwplayer_data = self._find_jwplayer_data(
3285 webpage, video_id, transform_source=js_to_json)
3286 return self._parse_jwplayer_data(
3287 jwplayer_data, video_id, *args, **kwargs)
3289 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
3290 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3292 if not isinstance(jwplayer_data, dict):
3295 playlist_items = jwplayer_data.get('playlist')
3296 # JWPlayer backward compatibility: single playlist item/flattened playlists
3297 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
3298 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
3299 if not isinstance(playlist_items, list):
3300 playlist_items = (playlist_items or jwplayer_data, )
3302 for video_data in playlist_items:
3303 if not isinstance(video_data, dict):
3305 # JWPlayer backward compatibility: flattened sources
3306 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
3307 if 'sources' not in video_data:
3308 video_data['sources'] = [video_data]
3310 this_video_id = video_id or video_data['mediaid']
3312 formats = self._parse_jwplayer_formats(
3313 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
3314 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
3317 tracks = video_data.get('tracks')
3318 if tracks and isinstance(tracks, list):
3319 for track in tracks:
3320 if not isinstance(track, dict):
3322 track_kind = track.get('kind')
3323 if not track_kind or not isinstance(track_kind, str):
3325 if track_kind.lower() not in ('captions', 'subtitles'):
3327 track_url = urljoin(base_url, track.get('file'))
3330 subtitles.setdefault(track.get('label') or 'en', []).append({
3331 'url': self._proto_relative_url(track_url)
3335 'id': this_video_id,
3336 'title': unescapeHTML(video_data['title'] if require_title else video_data.get('title')),
3337 'description': clean_html(video_data.get('description')),
3338 'thumbnail': urljoin(base_url, self._proto_relative_url(video_data.get('image'))),
3339 'timestamp': int_or_none(video_data.get('pubdate')),
3340 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
3341 'subtitles': subtitles,
3342 'alt_title': clean_html(video_data.get('subtitle')), # attributes used e.g. by Tele5 ...
3343 'genre': clean_html(video_data.get('genre')),
3344 'channel': clean_html(dict_get(video_data, ('category', 'channel'))),
3345 'season_number': int_or_none(video_data.get('season')),
3346 'episode_number': int_or_none(video_data.get('episode')),
3347 'release_year': int_or_none(video_data.get('releasedate')),
3348 'age_limit': int_or_none(video_data.get('age_restriction')),
3350 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
3351 if len(formats) == 1 and re.search(r'^(?:http|//).*(?:youtube\.com|youtu\.be)/.+', formats[0]['url']):
3353 '_type': 'url_transparent',
3354 'url': formats[0]['url'],
3357 entry['formats'] = formats
3358 entries.append(entry)
3359 if len(entries) == 1:
3362 return self.playlist_result(entries)
3364 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
3365 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3368 for source in jwplayer_sources_data:
3369 if not isinstance(source, dict):
3371 source_url = urljoin(
3372 base_url, self._proto_relative_url(source.get('file')))
3373 if not source_url or source_url in urls:
3375 urls.add(source_url)
3376 source_type = source.get('type') or ''
3377 ext = mimetype2ext(source_type) or determine_ext(source_url)
3378 if source_type == 'hls' or ext == 'm3u8' or 'format=m3u8-aapl' in source_url:
3379 formats.extend(self._extract_m3u8_formats(
3380 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
3381 m3u8_id=m3u8_id, fatal=False))
3382 elif source_type == 'dash' or ext == 'mpd' or 'format=mpd-time-csf' in source_url:
3383 formats.extend(self._extract_mpd_formats(
3384 source_url, video_id, mpd_id=mpd_id, fatal=False))
3386 formats.extend(self._extract_smil_formats(
3387 source_url, video_id, fatal=False))
3388 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3389 elif source_type.startswith('audio') or ext in (
3390 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
3397 format_id = str_or_none(source.get('label'))
3398 height = int_or_none(source.get('height'))
3399 if height is None and format_id:
3400 # Often no height is provided but there is a label in
3401 # format like "1080p", "720p SD", or 1080.
3402 height = parse_resolution(format_id).get('height')
3405 'width': int_or_none(source.get('width')),
3407 'tbr': int_or_none(source.get('bitrate'), scale=1000),
3408 'filesize': int_or_none(source.get('filesize')),
3410 'format_id': format_id
3412 if source_url.startswith('rtmp'):
3413 a_format['ext'] = 'flv'
3414 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3415 # of jwplayer.flash.swf
3416 rtmp_url_parts = re.split(
3417 r'((?:mp4|mp3|flv):)', source_url, 1)
3418 if len(rtmp_url_parts) == 3:
3419 rtmp_url, prefix, play_path = rtmp_url_parts
3422 'play_path': prefix + play_path,
3425 a_format.update(rtmp_params)
3426 formats.append(a_format)
3429 def _live_title(self, name):
3430 self._downloader.deprecation_warning('yt_dlp.InfoExtractor._live_title is deprecated and does not work as expected')
3433 def _int(self, v, name, fatal=False, **kwargs):
3434 res = int_or_none(v, **kwargs)
3436 msg = f'Failed to extract {name}: Could not parse value {v!r}'
3438 raise ExtractorError(msg)
3440 self.report_warning(msg)
3443 def _float(self, v, name, fatal=False, **kwargs):
3444 res = float_or_none(v, **kwargs)
3446 msg = f'Failed to extract {name}: Could not parse value {v!r}'
3448 raise ExtractorError(msg)
3450 self.report_warning(msg)
3453 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
3454 path='/', secure=False, discard=False, rest={}, **kwargs):
3455 cookie = http.cookiejar.Cookie(
3456 0, name, value, port, port is not None, domain, True,
3457 domain.startswith('.'), path, True, secure, expire_time,
3458 discard, None, None, rest)
3459 self.cookiejar.set_cookie(cookie)
3461 def _get_cookies(self, url):
3462 """ Return a http.cookies.SimpleCookie with the cookies for the url """
3463 return LenientSimpleCookie(self._downloader.cookiejar.get_cookie_header(url))
3465 def _apply_first_set_cookie_header(self, url_handle, cookie):
3467 Apply first Set-Cookie header instead of the last. Experimental.
3469 Some sites (e.g. [1-3]) may serve two cookies under the same name
3470 in Set-Cookie header and expect the first (old) one to be set rather
3471 than second (new). However, as of RFC6265 the newer one cookie
3472 should be set into cookie store what actually happens.
3473 We will workaround this issue by resetting the cookie to
3474 the first one manually.
3475 1. https://new.vk.com/
3476 2. https://github.com/ytdl-org/youtube-dl/issues/9841#issuecomment-227871201
3477 3. https://learning.oreilly.com/
3479 for header, cookies in url_handle.headers.items():
3480 if header.lower() != 'set-cookie':
3482 cookies = cookies.encode('iso-8859-1').decode('utf-8')
3483 cookie_value = re.search(
3484 r'%s=(.+?);.*?\b[Dd]omain=(.+?)(?:[,;]|$)' % cookie, cookies)
3486 value, domain = cookie_value.groups()
3487 self._set_cookie(domain, cookie, value)
3491 def get_testcases(cls, include_onlymatching=False):
3492 # Do not look in super classes
3493 t = vars(cls).get('_TEST')
3495 assert not hasattr(cls, '_TESTS'), f'{cls.ie_key()}IE has _TEST and _TESTS'
3498 tests = vars(cls).get('_TESTS', [])
3500 if not include_onlymatching and t.get('only_matching', False):
3502 t['name'] = cls.ie_key()
3504 if getattr(cls, '__wrapped__', None):
3505 yield from cls.__wrapped__.get_testcases(include_onlymatching)
3508 def get_webpage_testcases(cls):
3509 tests = vars(cls).get('_WEBPAGE_TESTS', [])
3511 t['name'] = cls.ie_key()
3513 if getattr(cls, '__wrapped__', None):
3514 yield from cls.__wrapped__.get_webpage_testcases()
3516 @classproperty(cache=True)
3518 """Get age limit from the testcases"""
3519 return max(traverse_obj(
3520 (*cls.get_testcases(include_onlymatching=False), *cls.get_webpage_testcases()),
3521 (..., (('playlist', 0), None), 'info_dict', 'age_limit')) or [0])
3523 @classproperty(cache=True)
3524 def _RETURN_TYPE(cls):
3525 """What the extractor returns: "video", "playlist", "any", or None (Unknown)"""
3526 tests = tuple(cls.get_testcases(include_onlymatching=False))
3529 elif not any(k.startswith('playlist') for test in tests for k in test):
3531 elif all(any(k.startswith('playlist') for k in test) for test in tests):
3536 def is_single_video(cls, url):
3537 """Returns whether the URL is of a single video, None if unknown"""
3538 if cls.suitable(url):
3539 return {'video': True, 'playlist': False}.get(cls._RETURN_TYPE)
3542 def is_suitable(cls, age_limit):
3543 """Test whether the extractor is generally suitable for the given age limit"""
3544 return not age_restricted(cls.age_limit, age_limit)
3547 def description(cls, *, markdown=True, search_examples=None):
3548 """Description of the extractor"""
3550 if cls._NETRC_MACHINE:
3552 desc += f' [*{cls._NETRC_MACHINE}*](## "netrc machine")'
3554 desc += f' [{cls._NETRC_MACHINE}]'
3555 if cls.IE_DESC is False:
3558 desc += f' {cls.IE_DESC}'
3560 desc += f'{";" if cls.IE_DESC else ""} "{cls.SEARCH_KEY}:" prefix'
3562 _COUNTS = ('', '5', '10', 'all')
3563 desc += f' (e.g. "{cls.SEARCH_KEY}{random.choice(_COUNTS)}:{random.choice(search_examples)}")'
3564 if not cls.working():
3565 desc += ' (**Currently broken**)' if markdown else ' (Currently broken)'
3567 # Escape emojis. Ref: https://github.com/github/markup/issues/1153
3568 name = (' - **%s**' % re.sub(r':(\w+:)', ':\u200B\\g<1>', cls.IE_NAME)) if markdown else cls.IE_NAME
3569 return f'{name}:{desc}' if desc else name
3571 def extract_subtitles(self, *args, **kwargs):
3572 if (self.get_param('writesubtitles', False)
3573 or self.get_param('listsubtitles')):
3574 return self._get_subtitles(*args, **kwargs)
3577 def _get_subtitles(self, *args, **kwargs):
3578 raise NotImplementedError('This method must be implemented by subclasses')
3580 class CommentsDisabled(Exception):
3581 """Raise in _get_comments if comments are disabled for the video"""
3583 def extract_comments(self, *args, **kwargs):
3584 if not self.get_param('getcomments'):
3586 generator = self._get_comments(*args, **kwargs)
3593 comments.append(next(generator))
3594 except StopIteration:
3596 except KeyboardInterrupt:
3597 self.to_screen('Interrupted by user')
3598 except self.CommentsDisabled:
3599 return {'comments': None, 'comment_count': None}
3600 except Exception as e:
3601 if self.get_param('ignoreerrors') is not True:
3603 self._downloader.report_error(e)
3604 comment_count = len(comments)
3605 self.to_screen(f'Extracted {comment_count} comments')
3607 'comments': comments,
3608 'comment_count': None if interrupted else comment_count
3612 def _get_comments(self, *args, **kwargs):
3613 raise NotImplementedError('This method must be implemented by subclasses')
3616 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
3617 """ Merge subtitle items for one language. Items with duplicated URLs/data
3618 will be dropped. """
3619 list1_data = {(item.get('url'), item.get('data')) for item in subtitle_list1}
3620 ret = list(subtitle_list1)
3621 ret.extend(item for item in subtitle_list2 if (item.get('url'), item.get('data')) not in list1_data)
3625 def _merge_subtitles(cls, *dicts, target=None):
3626 """ Merge subtitle dictionaries, language by language. """
3630 for lang, subs in d.items():
3631 target[lang] = cls._merge_subtitle_items(target.get(lang, []), subs)
3634 def extract_automatic_captions(self, *args, **kwargs):
3635 if (self.get_param('writeautomaticsub', False)
3636 or self.get_param('listsubtitles')):
3637 return self._get_automatic_captions(*args, **kwargs)
3640 def _get_automatic_captions(self, *args, **kwargs):
3641 raise NotImplementedError('This method must be implemented by subclasses')
3643 @functools.cached_property
3644 def _cookies_passed(self):
3645 """Whether cookies have been passed to YoutubeDL"""
3646 return self.get_param('cookiefile') is not None or self.get_param('cookiesfrombrowser') is not None
3648 def mark_watched(self, *args, **kwargs):
3649 if not self.get_param('mark_watched', False):
3651 if self.supports_login() and self._get_login_info()[0] is not None or self._cookies_passed:
3652 self._mark_watched(*args, **kwargs)
3654 def _mark_watched(self, *args, **kwargs):
3655 raise NotImplementedError('This method must be implemented by subclasses')
3657 def geo_verification_headers(self):
3659 geo_verification_proxy = self.get_param('geo_verification_proxy')
3660 if geo_verification_proxy:
3661 headers['Ytdl-request-proxy'] = geo_verification_proxy
3665 def _generic_id(url):
3666 return urllib.parse.unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
3668 def _generic_title(self, url='', webpage='', *, default=None):
3669 return (self._og_search_title(webpage, default=None)
3670 or self._html_extract_title(webpage, default=None)
3671 or urllib.parse.unquote(os.path.splitext(url_basename(url))[0])
3674 def _extract_chapters_helper(self, chapter_list, start_function, title_function, duration, strict=True):
3678 'start_time': start_function(chapter),
3679 'title': title_function(chapter),
3680 } for chapter in chapter_list or []]
3682 warn = self.report_warning
3684 warn = self.write_debug
3685 chapter_list.sort(key=lambda c: c['start_time'] or 0)
3687 chapters = [{'start_time': 0}]
3688 for idx, chapter in enumerate(chapter_list):
3689 if chapter['start_time'] is None:
3690 warn(f'Incomplete chapter {idx}')
3691 elif chapters[-1]['start_time'] <= chapter['start_time'] <= duration:
3692 chapters.append(chapter)
3693 elif chapter not in chapters:
3694 issue = (f'{chapter["start_time"]} > {duration}' if chapter['start_time'] > duration
3695 else f'{chapter["start_time"]} < {chapters[-1]["start_time"]}')
3696 warn(f'Invalid start time ({issue}) for chapter "{chapter["title"]}"')
3699 def _extract_chapters_from_description(self, description, duration):
3700 duration_re = r'(?:\d+:)?\d{1,2}:\d{2}'
3701 sep_re = r'(?m)^\s*(%s)\b\W*\s(%s)\s*$'
3702 return self._extract_chapters_helper(
3703 re.findall(sep_re % (duration_re, r'.+?'), description or ''),
3704 start_function=lambda x: parse_duration(x[0]), title_function=lambda x: x[1],
3705 duration=duration, strict=False) or self._extract_chapters_helper(
3706 re.findall(sep_re % (r'.+?', duration_re), description or ''),
3707 start_function=lambda x: parse_duration(x[1]), title_function=lambda x: x[0],
3708 duration=duration, strict=False)
3711 def _availability(is_private=None, needs_premium=None, needs_subscription=None, needs_auth=None, is_unlisted=None):
3712 all_known = all(map(
3713 lambda x: x is not None,
3714 (is_private, needs_premium, needs_subscription, needs_auth, is_unlisted)))
3716 'private' if is_private
3717 else 'premium_only' if needs_premium
3718 else 'subscriber_only' if needs_subscription
3719 else 'needs_auth' if needs_auth
3720 else 'unlisted' if is_unlisted
3721 else 'public' if all_known
3724 def _configuration_arg(self, key, default=NO_DEFAULT, *, ie_key=None, casesense=False):
3726 @returns A
list of values
for the extractor argument given by
"key"
3727 or "default" if no such key
is present
3728 @param default The default value to
return when the key
is not present (default
: [])
3729 @param casesense When false
, the values are converted to lower case
3731 ie_key = ie_key if isinstance(ie_key, str) else (ie_key or self).ie_key()
3732 val = traverse_obj(self._downloader.params, ('extractor_args', ie_key.lower(), key))
3734 return [] if default is NO_DEFAULT else default
3735 return list(val) if casesense else [x.lower() for x in val]
3737 def _yes_playlist(self, playlist_id, video_id, smuggled_data=None, *, playlist_label='playlist', video_label='video'):
3738 if not playlist_id or not video_id:
3741 no_playlist = (smuggled_data or {}).get('force_noplaylist')
3742 if no_playlist is not None:
3743 return not no_playlist
3745 video_id = '' if video_id is True else f' {video_id}'
3746 playlist_id = '' if playlist_id is True else f' {playlist_id}'
3747 if self.get_param('noplaylist'):
3748 self.to_screen(f'Downloading just the {video_label}{video_id} because of --no-playlist')
3750 self.to_screen(f'Downloading {playlist_label}{playlist_id} - add --no-playlist to download just the {video_label}{video_id}')
3753 def _error_or_warning(self, err, _count=None, _retries=0, *, fatal=True):
3754 RetryManager.report_retry(
3755 err, _count or int(fatal), _retries,
3756 info=self.to_screen, warn=self.report_warning, error=None if fatal else self.report_warning,
3757 sleep_func=self.get_param('retry_sleep_functions', {}).get('extractor'))
3759 def RetryManager(self, **kwargs):
3760 return RetryManager(self.get_param('extractor_retries', 3), self._error_or_warning, **kwargs)
3762 def _extract_generic_embeds(self, url, *args, info_dict={}, note='Extracting generic embeds', **kwargs):
3763 display_id = traverse_obj(info_dict, 'display_id', 'id')
3764 self.to_screen(f'{format_field(display_id, None, "%s: ")}{note}')
3765 return self._downloader.get_info_extractor('Generic')._extract_embeds(
3766 smuggle_url(url, {'block_ies': [self.ie_key()]}), *args, **kwargs)
3769 def extract_from_webpage(cls, ydl, url, webpage):
3770 ie = (cls if isinstance(cls._extract_from_webpage, types.MethodType)
3771 else ydl.get_info_extractor(cls.ie_key()))
3772 for info in ie._extract_from_webpage(url, webpage) or []:
3773 # url = None since we do not want to set (webpage/original)_url
3774 ydl.add_default_extra_info(info, ie, None)
3778 def _extract_from_webpage(cls, url, webpage):
3779 for embed_url in orderedSet(
3780 cls._extract_embed_urls(url, webpage) or [], lazy=True):
3781 yield cls.url_result(embed_url, None if cls._VALID_URL is False else cls)
3784 def _extract_embed_urls(cls, url, webpage):
3785 """@returns all the embed urls on the webpage"""
3786 if '_EMBED_URL_RE' not in cls.__dict__:
3787 assert isinstance(cls._EMBED_REGEX, (list, tuple))
3788 for idx, regex in enumerate(cls._EMBED_REGEX):
3789 assert regex.count('(?P<url>') == 1, \
3790 f'{cls.__name__}._EMBED_REGEX[{idx}] must have exactly 1 url group\n\t{regex}'
3791 cls._EMBED_URL_RE = tuple(map(re.compile, cls._EMBED_REGEX))
3793 for regex in cls._EMBED_URL_RE:
3794 for mobj in regex.finditer(webpage):
3795 embed_url = urllib.parse.urljoin(url, unescapeHTML(mobj.group('url')))
3796 if cls._VALID_URL is False or cls.suitable(embed_url):
3799 class StopExtraction(Exception):
3803 def _extract_url(cls, webpage): # TODO: Remove
3804 """Only for compatibility with some older extractors"""
3805 return next(iter(cls._extract_embed_urls(None, webpage) or []), None)
3808 def __init_subclass__(cls, *, plugin_name=None, **kwargs):
3810 mro = inspect.getmro(cls)
3811 super_class = cls.__wrapped__ = mro[mro.index(cls) + 1]
3812 cls.PLUGIN_NAME, cls.ie_key = plugin_name, super_class.ie_key
3813 cls.IE_NAME = f'{super_class.IE_NAME}+{plugin_name}'
3814 while getattr(super_class, '__wrapped__', None):
3815 super_class = super_class.__wrapped__
3816 setattr(sys.modules[super_class.__module__], super_class.__name__, cls)
3817 _PLUGIN_OVERRIDES[super_class].append(cls)
3819 return super().__init_subclass__(**kwargs)
3822 class SearchInfoExtractor(InfoExtractor):
3824 Base class for paged search queries extractors.
3825 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
3826 Instances should define _SEARCH_KEY and optionally _MAX_RESULTS
3829 _MAX_RESULTS = float('inf')
3830 _RETURN_TYPE = 'playlist'
3833 def _VALID_URL(cls):
3834 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
3836 def _real_extract(self, query):
3837 prefix, query = self._match_valid_url(query).group('prefix', 'query')
3839 return self._get_n_results(query, 1)
3840 elif prefix == 'all':
3841 return self._get_n_results(query, self._MAX_RESULTS)
3845 raise ExtractorError(f'invalid download number {n} for query "{query}"')
3846 elif n > self._MAX_RESULTS:
3847 self.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3848 n = self._MAX_RESULTS
3849 return self._get_n_results(query, n)
3851 def _get_n_results(self, query, n):
3852 """Get a specified number of results for a query.
3853 Either this function or _search_results must be overridden by subclasses """
3854 return self.playlist_result(
3855 itertools.islice(self._search_results(query), 0, None if n == float('inf') else n),
3858 def _search_results(self, query):
3859 """Returns an iterator of search results"""
3860 raise NotImplementedError('This method must be implemented by subclasses')
3863 def SEARCH_KEY(cls):
3864 return cls._SEARCH_KEY
3867 class UnsupportedURLIE(InfoExtractor):
3872 def _real_extract(self, url):
3873 raise UnsupportedError(url)
3876 _PLUGIN_OVERRIDES = collections.defaultdict(list)