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)
2253 return self._parse_smil_formats_and_subtitles(smil, urlh.url, video_id, f4m_params=f4m_params,
2254 namespace=self._parse_smil_namespace(smil))
2256 def _extract_smil_formats(self, *args, **kwargs):
2257 fmts, subs = self._extract_smil_formats_and_subtitles(*args, **kwargs)
2259 self._report_ignoring_subs('SMIL')
2262 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
2263 res = self._download_smil(smil_url, video_id, fatal=fatal)
2270 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
2272 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
2273 return self._download_xml_handle(
2274 smil_url, video_id, 'Downloading SMIL file',
2275 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
2277 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
2278 namespace = self._parse_smil_namespace(smil)
2280 formats, subtitles = self._parse_smil_formats_and_subtitles(
2281 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
2283 video_id = os.path.splitext(url_basename(smil_url))[0]
2287 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
2288 name = meta.attrib.get('name')
2289 content = meta.attrib.get('content')
2290 if not name or not content:
2292 if not title and name == 'title':
2294 elif not description and name in ('description', 'abstract'):
2295 description = content
2296 elif not upload_date and name == 'date':
2297 upload_date = unified_strdate(content)
2300 'id': image.get('type'),
2301 'url': image.get('src'),
2302 'width': int_or_none(image.get('width')),
2303 'height': int_or_none(image.get('height')),
2304 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
2308 'title': title or video_id,
2309 'description': description,
2310 'upload_date': upload_date,
2311 'thumbnails': thumbnails,
2313 'subtitles': subtitles,
2316 def _parse_smil_namespace(self, smil):
2317 return self._search_regex(
2318 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
2320 def _parse_smil_formats(self, *args, **kwargs):
2321 fmts, subs = self._parse_smil_formats_and_subtitles(*args, **kwargs)
2323 self._report_ignoring_subs('SMIL')
2326 def _parse_smil_formats_and_subtitles(
2327 self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
2329 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
2330 b = meta.get('base') or meta.get('httpBase')
2335 formats, subtitles = [], {}
2342 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
2343 for medium in media:
2344 src = medium.get('src')
2345 if not src or src in srcs:
2349 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
2350 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
2351 width = int_or_none(medium.get('width'))
2352 height = int_or_none(medium.get('height'))
2353 proto = medium.get('proto')
2354 ext = medium.get('ext')
2355 src_ext = determine_ext(src, default_ext=None) or ext or urlhandle_detect_ext(
2356 self._request_webpage(HEADRequest(src), video_id, note='Requesting extension info', fatal=False))
2357 streamer = medium.get('streamer') or base
2359 if proto == 'rtmp' or streamer.startswith('rtmp'):
2365 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
2367 'filesize': filesize,
2371 if transform_rtmp_url:
2372 streamer, src = transform_rtmp_url(streamer, src)
2373 formats[-1].update({
2379 src_url = src if src.startswith('http') else urllib.parse.urljoin(base, src)
2380 src_url = src_url.strip()
2382 if proto == 'm3u8' or src_ext == 'm3u8':
2383 m3u8_formats, m3u8_subs = self._extract_m3u8_formats_and_subtitles(
2384 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
2385 self._merge_subtitles(m3u8_subs, target=subtitles)
2386 if len(m3u8_formats) == 1:
2388 m3u8_formats[0].update({
2389 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
2394 formats.extend(m3u8_formats)
2395 elif src_ext == 'f4m':
2400 'plugin': 'flowplayer-3.2.0.1',
2402 f4m_url += '&' if '?' in f4m_url else '?'
2403 f4m_url += urllib.parse.urlencode(f4m_params)
2404 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
2405 elif src_ext == 'mpd':
2406 mpd_formats, mpd_subs = self._extract_mpd_formats_and_subtitles(
2407 src_url, video_id, mpd_id='dash', fatal=False)
2408 formats.extend(mpd_formats)
2409 self._merge_subtitles(mpd_subs, target=subtitles)
2410 elif re.search(r'\.ism/[Mm]anifest', src_url):
2411 ism_formats, ism_subs = self._extract_ism_formats_and_subtitles(
2412 src_url, video_id, ism_id='mss', fatal=False)
2413 formats.extend(ism_formats)
2414 self._merge_subtitles(ism_subs, target=subtitles)
2415 elif src_url.startswith('http') and self._is_valid_url(src, video_id):
2419 'ext': ext or src_ext or 'flv',
2420 'format_id': 'http-%d' % (bitrate or http_count),
2422 'filesize': filesize,
2427 for medium in smil.findall(self._xpath_ns('.//imagestream', namespace)):
2428 src = medium.get('src')
2429 if not src or src in srcs:
2435 'format_id': 'imagestream-%d' % (imgs_count),
2437 'ext': mimetype2ext(medium.get('type')),
2440 'width': int_or_none(medium.get('width')),
2441 'height': int_or_none(medium.get('height')),
2442 'format_note': 'SMIL storyboards',
2445 smil_subs = self._parse_smil_subtitles(smil, namespace=namespace)
2446 self._merge_subtitles(smil_subs, target=subtitles)
2448 return formats, subtitles
2450 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
2453 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
2454 src = textstream.get('src')
2455 if not src or src in urls:
2458 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
2459 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
2460 subtitles.setdefault(lang, []).append({
2466 def _extract_xspf_playlist(self, xspf_url, playlist_id, fatal=True):
2467 res = self._download_xml_handle(
2468 xspf_url, playlist_id, 'Downloading xpsf playlist',
2469 'Unable to download xspf manifest', fatal=fatal)
2476 return self._parse_xspf(
2477 xspf, playlist_id, xspf_url=xspf_url,
2478 xspf_base_url=base_url(xspf_url))
2480 def _parse_xspf(self, xspf_doc, playlist_id, xspf_url=None, xspf_base_url=None):
2482 'xspf': 'http://xspf.org/ns/0/',
2483 's1': 'http://static.streamone.nl/player/ns/0',
2487 for track in xspf_doc.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
2489 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
2490 description = xpath_text(
2491 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
2492 thumbnail = xpath_text(
2493 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
2494 duration = float_or_none(
2495 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
2498 for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP)):
2499 format_url = urljoin(xspf_base_url, location.text)
2504 'manifest_url': xspf_url,
2505 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
2506 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
2507 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
2513 'description': description,
2514 'thumbnail': thumbnail,
2515 'duration': duration,
2520 def _extract_mpd_formats(self, *args, **kwargs):
2521 fmts, subs = self._extract_mpd_formats_and_subtitles(*args, **kwargs)
2523 self._report_ignoring_subs('DASH')
2526 def _extract_mpd_formats_and_subtitles(
2527 self, mpd_url, video_id, mpd_id=None, note=None, errnote=None,
2528 fatal=True, data=None, headers={}, query={}):
2530 if self.get_param('ignore_no_formats_error'):
2533 res = self._download_xml_handle(
2535 note='Downloading MPD manifest' if note is None else note,
2536 errnote='Failed to download MPD manifest' if errnote is None else errnote,
2537 fatal=fatal, data=data, headers=headers, query=query)
2544 # We could have been redirected to a new url when we retrieved our mpd file.
2546 mpd_base_url = base_url(mpd_url)
2548 return self._parse_mpd_formats_and_subtitles(
2549 mpd_doc, mpd_id, mpd_base_url, mpd_url)
2551 def _parse_mpd_formats(self, *args, **kwargs):
2552 fmts, subs = self._parse_mpd_formats_and_subtitles(*args, **kwargs)
2554 self._report_ignoring_subs('DASH')
2557 def _parse_mpd_formats_and_subtitles(
2558 self, mpd_doc, mpd_id=None, mpd_base_url='', mpd_url=None):
2560 Parse formats from MPD manifest.
2562 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
2563 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
2564 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
2566 if not self.get_param('dynamic_mpd', True):
2567 if mpd_doc.get('type') == 'dynamic':
2570 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
2573 return self._xpath_ns(path, namespace)
2575 def is_drm_protected(element):
2576 return element.find(_add_ns('ContentProtection')) is not None
2578 def extract_multisegment_info(element, ms_parent_info):
2579 ms_info = ms_parent_info.copy()
2581 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2582 # common attributes and elements. We will only extract relevant
2584 def extract_common(source):
2585 segment_timeline = source.find(_add_ns('SegmentTimeline'))
2586 if segment_timeline is not None:
2587 s_e = segment_timeline.findall(_add_ns('S'))
2589 ms_info['total_number'] = 0
2592 r = int(s.get('r', 0))
2593 ms_info['total_number'] += 1 + r
2594 ms_info['s'].append({
2595 't': int(s.get('t', 0)),
2596 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2597 'd': int(s.attrib['d']),
2600 start_number = source.get('startNumber')
2602 ms_info['start_number'] = int(start_number)
2603 timescale = source.get('timescale')
2605 ms_info['timescale'] = int(timescale)
2606 segment_duration = source.get('duration')
2607 if segment_duration:
2608 ms_info['segment_duration'] = float(segment_duration)
2610 def extract_Initialization(source):
2611 initialization = source.find(_add_ns('Initialization'))
2612 if initialization is not None:
2613 ms_info['initialization_url'] = initialization.attrib['sourceURL']
2615 segment_list = element.find(_add_ns('SegmentList'))
2616 if segment_list is not None:
2617 extract_common(segment_list)
2618 extract_Initialization(segment_list)
2619 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
2621 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
2623 segment_template = element.find(_add_ns('SegmentTemplate'))
2624 if segment_template is not None:
2625 extract_common(segment_template)
2626 media = segment_template.get('media')
2628 ms_info['media'] = media
2629 initialization = segment_template.get('initialization')
2631 ms_info['initialization'] = initialization
2633 extract_Initialization(segment_template)
2636 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
2637 formats, subtitles = [], {}
2638 stream_numbers = collections.defaultdict(int)
2639 for period in mpd_doc.findall(_add_ns('Period')):
2640 period_duration = parse_duration(period.get('duration')) or mpd_duration
2641 period_ms_info = extract_multisegment_info(period, {
2645 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
2646 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
2647 for representation in adaptation_set.findall(_add_ns('Representation')):
2648 representation_attrib = adaptation_set.attrib.copy()
2649 representation_attrib.update(representation.attrib)
2650 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2651 mime_type = representation_attrib['mimeType']
2652 content_type = representation_attrib.get('contentType', mime_type.split('/')[0])
2654 codec_str = representation_attrib.get('codecs', '')
2655 # Some kind of binary subtitle found in some youtube livestreams
2656 if mime_type == 'application/x-rawcc':
2657 codecs = {'scodec': codec_str}
2659 codecs = parse_codecs(codec_str)
2660 if content_type not in ('video', 'audio', 'text'):
2661 if mime_type == 'image/jpeg':
2662 content_type = mime_type
2663 elif codecs.get('vcodec', 'none') != 'none':
2664 content_type = 'video'
2665 elif codecs.get('acodec', 'none') != 'none':
2666 content_type = 'audio'
2667 elif codecs.get('scodec', 'none') != 'none':
2668 content_type = 'text'
2669 elif mimetype2ext(mime_type) in ('tt', 'dfxp', 'ttml', 'xml', 'json'):
2670 content_type = 'text'
2672 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
2676 for element in (representation, adaptation_set, period, mpd_doc):
2677 base_url_e = element.find(_add_ns('BaseURL'))
2678 if try_call(lambda: base_url_e.text) is not None:
2679 base_url = base_url_e.text + base_url
2680 if re.match(r'^https?://', base_url):
2682 if mpd_base_url and base_url.startswith('/'):
2683 base_url = urllib.parse.urljoin(mpd_base_url, base_url)
2684 elif mpd_base_url and not re.match(r'^https?://', base_url):
2685 if not mpd_base_url.endswith('/'):
2687 base_url = mpd_base_url + base_url
2688 representation_id = representation_attrib.get('id')
2689 lang = representation_attrib.get('lang')
2690 url_el = representation.find(_add_ns('BaseURL'))
2691 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
2692 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
2693 if representation_id is not None:
2694 format_id = representation_id
2696 format_id = content_type
2698 format_id = mpd_id + '-' + format_id
2699 if content_type in ('video', 'audio'):
2701 'format_id': format_id,
2702 'manifest_url': mpd_url,
2703 'ext': mimetype2ext(mime_type),
2704 'width': int_or_none(representation_attrib.get('width')),
2705 'height': int_or_none(representation_attrib.get('height')),
2706 'tbr': float_or_none(bandwidth, 1000),
2707 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
2708 'fps': int_or_none(representation_attrib.get('frameRate')),
2709 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
2710 'format_note': 'DASH %s' % content_type,
2711 'filesize': filesize,
2712 'container': mimetype2ext(mime_type) + '_dash',
2715 elif content_type == 'text':
2717 'ext': mimetype2ext(mime_type),
2718 'manifest_url': mpd_url,
2719 'filesize': filesize,
2721 elif content_type == 'image/jpeg':
2722 # See test case in VikiIE
2723 # https://www.viki.com/videos/1175236v-choosing-spouse-by-lottery-episode-1
2725 'format_id': format_id,
2727 'manifest_url': mpd_url,
2728 'format_note': 'DASH storyboards (jpeg)',
2732 if is_drm_protected(adaptation_set) or is_drm_protected(representation):
2734 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
2736 def prepare_template(template_name, identifiers):
2737 tmpl = representation_ms_info[template_name]
2738 if representation_id is not None:
2739 tmpl = tmpl.replace('$RepresentationID$', representation_id)
2740 # First of, % characters outside $...$ templates
2741 # must be escaped by doubling for proper processing
2742 # by % operator string formatting used further (see
2743 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2749 in_template = not in_template
2750 elif c == '%' and not in_template:
2752 # Next, $...$ templates are translated to their
2753 # %(...) counterparts to be used with % operator
2754 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
2755 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
2756 t.replace('$$', '$')
2759 # @initialization is a regular template like @media one
2760 # so it should be handled just the same way (see
2761 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2762 if 'initialization' in representation_ms_info:
2763 initialization_template = prepare_template(
2765 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2766 # $Time$ shall not be included for @initialization thus
2767 # only $Bandwidth$ remains
2769 representation_ms_info['initialization_url'] = initialization_template % {
2770 'Bandwidth': bandwidth,
2773 def location_key(location):
2774 return 'url' if re.match(r'^https?://', location) else 'path'
2776 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
2778 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2779 media_location_key = location_key(media_template)
2781 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2782 # can't be used at the same time
2783 if '%(Number' in media_template and 's' not in representation_ms_info:
2784 segment_duration = None
2785 if 'total_number' not in representation_ms_info and 'segment_duration' in representation_ms_info:
2786 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
2787 representation_ms_info['total_number'] = int(math.ceil(
2788 float_or_none(period_duration, segment_duration, default=0)))
2789 representation_ms_info['fragments'] = [{
2790 media_location_key: media_template % {
2791 'Number': segment_number,
2792 'Bandwidth': bandwidth,
2794 'duration': segment_duration,
2795 } for segment_number in range(
2796 representation_ms_info['start_number'],
2797 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
2799 # $Number*$ or $Time$ in media template with S list available
2800 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2801 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2802 representation_ms_info['fragments'] = []
2805 segment_number = representation_ms_info['start_number']
2807 def add_segment_url():
2808 segment_url = media_template % {
2809 'Time': segment_time,
2810 'Bandwidth': bandwidth,
2811 'Number': segment_number,
2813 representation_ms_info['fragments'].append({
2814 media_location_key: segment_url,
2815 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
2818 for num, s in enumerate(representation_ms_info['s']):
2819 segment_time = s.get('t') or segment_time
2823 for r in range(s.get('r', 0)):
2824 segment_time += segment_d
2827 segment_time += segment_d
2828 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
2829 # No media template,
2830 # e.g. https://www.youtube.com/watch?v=iXZV5uAYMJI
2831 # or any YouTube dashsegments video
2834 timescale = representation_ms_info['timescale']
2835 for s in representation_ms_info['s']:
2836 duration = float_or_none(s['d'], timescale)
2837 for r in range(s.get('r', 0) + 1):
2838 segment_uri = representation_ms_info['segment_urls'][segment_index]
2840 location_key(segment_uri): segment_uri,
2841 'duration': duration,
2844 representation_ms_info['fragments'] = fragments
2845 elif 'segment_urls' in representation_ms_info:
2846 # Segment URLs with no SegmentTimeline
2847 # E.g. https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2848 # https://github.com/ytdl-org/youtube-dl/pull/14844
2850 segment_duration = float_or_none(
2851 representation_ms_info['segment_duration'],
2852 representation_ms_info['timescale']) if 'segment_duration' in representation_ms_info else None
2853 for segment_url in representation_ms_info['segment_urls']:
2855 location_key(segment_url): segment_url,
2857 if segment_duration:
2858 fragment['duration'] = segment_duration
2859 fragments.append(fragment)
2860 representation_ms_info['fragments'] = fragments
2861 # If there is a fragments key available then we correctly recognized fragmented media.
2862 # Otherwise we will assume unfragmented media with direct access. Technically, such
2863 # assumption is not necessarily correct since we may simply have no support for
2864 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2865 if 'fragments' in representation_ms_info:
2867 # NB: mpd_url may be empty when MPD manifest is parsed from a string
2868 'url': mpd_url or base_url,
2869 'fragment_base_url': base_url,
2871 'protocol': 'http_dash_segments' if mime_type != 'image/jpeg' else 'mhtml',
2873 if 'initialization_url' in representation_ms_info:
2874 initialization_url = representation_ms_info['initialization_url']
2875 if not f.get('url'):
2876 f['url'] = initialization_url
2877 f['fragments'].append({location_key(initialization_url): initialization_url})
2878 f['fragments'].extend(representation_ms_info['fragments'])
2879 if not period_duration:
2880 period_duration = try_get(
2881 representation_ms_info,
2882 lambda r: sum(frag['duration'] for frag in r['fragments']), float)
2884 # Assuming direct URL to unfragmented media.
2886 if content_type in ('video', 'audio', 'image/jpeg'):
2887 f['manifest_stream_number'] = stream_numbers[f['url']]
2888 stream_numbers[f['url']] += 1
2890 elif content_type == 'text':
2891 subtitles.setdefault(lang or 'und', []).append(f)
2893 return formats, subtitles
2895 def _extract_ism_formats(self, *args, **kwargs):
2896 fmts, subs = self._extract_ism_formats_and_subtitles(*args, **kwargs)
2898 self._report_ignoring_subs('ISM')
2901 def _extract_ism_formats_and_subtitles(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
2902 if self.get_param('ignore_no_formats_error'):
2905 res = self._download_xml_handle(
2907 note='Downloading ISM manifest' if note is None else note,
2908 errnote='Failed to download ISM manifest' if errnote is None else errnote,
2909 fatal=fatal, data=data, headers=headers, query=query)
2916 return self._parse_ism_formats_and_subtitles(ism_doc, urlh.url, ism_id)
2918 def _parse_ism_formats_and_subtitles(self, ism_doc, ism_url, ism_id=None):
2920 Parse formats from ISM manifest.
2922 1. [MS-SSTR]: Smooth Streaming Protocol,
2923 https://msdn.microsoft.com/en-us/library/ff469518.aspx
2925 if ism_doc.get('IsLive') == 'TRUE':
2928 duration = int(ism_doc.attrib['Duration'])
2929 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
2933 for stream in ism_doc.findall('StreamIndex'):
2934 stream_type = stream.get('Type')
2935 if stream_type not in ('video', 'audio', 'text'):
2937 url_pattern = stream.attrib['Url']
2938 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
2939 stream_name = stream.get('Name')
2940 stream_language = stream.get('Language', 'und')
2941 for track in stream.findall('QualityLevel'):
2942 KNOWN_TAGS = {'255': 'AACL', '65534': 'EC-3'}
2943 fourcc = track.get('FourCC') or KNOWN_TAGS.get(track.get('AudioTag'))
2944 # TODO: add support for WVC1 and WMAP
2945 if fourcc not in ('H264', 'AVC1', 'AACL', 'TTML', 'EC-3'):
2946 self.report_warning('%s is not a supported codec' % fourcc)
2948 tbr = int(track.attrib['Bitrate']) // 1000
2949 # [1] does not mention Width and Height attributes. However,
2950 # they're often present while MaxWidth and MaxHeight are
2951 # missing, so should be used as fallbacks
2952 width = int_or_none(track.get('MaxWidth') or track.get('Width'))
2953 height = int_or_none(track.get('MaxHeight') or track.get('Height'))
2954 sampling_rate = int_or_none(track.get('SamplingRate'))
2956 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
2957 track_url_pattern = urllib.parse.urljoin(ism_url, track_url_pattern)
2963 stream_fragments = stream.findall('c')
2964 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
2965 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
2966 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
2967 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
2968 if not fragment_ctx['duration']:
2970 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
2972 next_fragment_time = duration
2973 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
2974 for _ in range(fragment_repeat):
2976 'url': re.sub(r'{start[ _]time}', str(fragment_ctx['time']), track_url_pattern),
2977 'duration': fragment_ctx['duration'] / stream_timescale,
2979 fragment_ctx['time'] += fragment_ctx['duration']
2981 if stream_type == 'text':
2982 subtitles.setdefault(stream_language, []).append({
2986 'manifest_url': ism_url,
2987 'fragments': fragments,
2988 '_download_params': {
2989 'stream_type': stream_type,
2990 'duration': duration,
2991 'timescale': stream_timescale,
2993 'language': stream_language,
2994 'codec_private_data': track.get('CodecPrivateData'),
2997 elif stream_type in ('video', 'audio'):
2999 'format_id': join_nonempty(ism_id, stream_name, tbr),
3001 'manifest_url': ism_url,
3002 'ext': 'ismv' if stream_type == 'video' else 'isma',
3006 'asr': sampling_rate,
3007 'vcodec': 'none' if stream_type == 'audio' else fourcc,
3008 'acodec': 'none' if stream_type == 'video' else fourcc,
3010 'fragments': fragments,
3011 'has_drm': ism_doc.find('Protection') is not None,
3012 'language': stream_language,
3013 'audio_channels': int_or_none(track.get('Channels')),
3014 '_download_params': {
3015 'stream_type': stream_type,
3016 'duration': duration,
3017 'timescale': stream_timescale,
3018 'width': width or 0,
3019 'height': height or 0,
3021 'language': stream_language,
3022 'codec_private_data': track.get('CodecPrivateData'),
3023 'sampling_rate': sampling_rate,
3024 'channels': int_or_none(track.get('Channels', 2)),
3025 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
3026 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
3029 return formats, subtitles
3031 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):
3032 def absolute_url(item_url):
3033 return urljoin(base_url, item_url)
3035 def parse_content_type(content_type):
3036 if not content_type:
3038 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?
(?P
<codecs
>[^
"]+))?', content_type)
3040 mimetype, codecs = ctr.groups()
3041 f = parse_codecs(codecs)
3042 f['ext'] = mimetype2ext(mimetype)
3046 def _media_formats(src, cur_media_type, type_info=None):
3047 type_info = type_info or {}
3048 full_url = absolute_url(src)
3049 ext = type_info.get('ext') or determine_ext(full_url)
3051 is_plain_url = False
3052 formats = self._extract_m3u8_formats(
3053 full_url, video_id, ext='mp4',
3054 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
3055 preference=preference, quality=quality, fatal=False)
3057 is_plain_url = False
3058 formats = self._extract_mpd_formats(
3059 full_url, video_id, mpd_id=mpd_id, fatal=False)
3064 'vcodec': 'none' if cur_media_type == 'audio' else None,
3067 return is_plain_url, formats
3070 # amp-video and amp-audio are very similar to their HTML5 counterparts
3071 # so we will include them right here (see
3072 # https://www.ampproject.org/docs/reference/components/amp-video)
3073 # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/
3074 _MEDIA_TAG_NAME_RE = r'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)'
3075 media_tags = [(media_tag, media_tag_name, media_type, '')
3076 for media_tag, media_tag_name, media_type
3077 in re.findall(r'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE, webpage)]
3078 media_tags.extend(re.findall(
3079 # We only allow video|audio followed by a whitespace or '>'.
3080 # Allowing more characters may end up in significant slow down (see
3081 # https://github.com/ytdl-org/youtube-dl/issues/11979,
3082 # e.g. http://www.porntrex.com/maps/videositemap.xml).
3083 r'(?s)(<(?P<tag>%s)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>' % _MEDIA_TAG_NAME_RE, webpage))
3084 for media_tag, _, media_type, media_content in media_tags:
3089 media_attributes = extract_attributes(media_tag)
3090 src = strip_or_none(dict_get(media_attributes, ('src', 'data-video-src', 'data-src', 'data-source')))
3092 f = parse_content_type(media_attributes.get('type'))
3093 _, formats = _media_formats(src, media_type, f)
3094 media_info['formats'].extend(formats)
3095 media_info['thumbnail'] = absolute_url(media_attributes.get('poster'))
3097 for source_tag in re.findall(r'<source[^>]+>', media_content):
3098 s_attr = extract_attributes(source_tag)
3099 # data-video-src and data-src are non standard but seen
3100 # several times in the wild
3101 src = strip_or_none(dict_get(s_attr, ('src', 'data-video-src', 'data-src', 'data-source')))
3104 f = parse_content_type(s_attr.get('type'))
3105 is_plain_url, formats = _media_formats(src, media_type, f)
3107 # width, height, res, label and title attributes are
3108 # all not standard but seen several times in the wild
3111 for lbl in ('label', 'title')
3112 if str_or_none(s_attr.get(lbl))
3114 width = int_or_none(s_attr.get('width'))
3115 height = (int_or_none(s_attr.get('height'))
3116 or int_or_none(s_attr.get('res')))
3117 if not width or not height:
3119 resolution = parse_resolution(lbl)
3122 width = width or resolution.get('width')
3123 height = height or resolution.get('height')
3125 tbr = parse_bitrate(lbl)
3134 'format_id': s_attr.get('label') or s_attr.get('title'),
3136 f.update(formats[0])
3137 media_info['formats'].append(f)
3139 media_info['formats'].extend(formats)
3140 for track_tag in re.findall(r'<track[^>]+>', media_content):
3141 track_attributes = extract_attributes(track_tag)
3142 kind = track_attributes.get('kind')
3143 if not kind or kind in ('subtitles', 'captions'):
3144 src = strip_or_none(track_attributes.get('src'))
3147 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
3148 media_info['subtitles'].setdefault(lang, []).append({
3149 'url': absolute_url(src),
3151 for f in media_info['formats']:
3152 f.setdefault('http_headers', {})['Referer'] = base_url
3153 if media_info['formats'] or media_info['subtitles']:
3154 entries.append(media_info)
3157 def _extract_akamai_formats(self, *args, **kwargs):
3158 fmts, subs = self._extract_akamai_formats_and_subtitles(*args, **kwargs)
3160 self._report_ignoring_subs('akamai')
3163 def _extract_akamai_formats_and_subtitles(self, manifest_url, video_id, hosts={}):
3164 signed = 'hdnea=' in manifest_url
3166 # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
3167 manifest_url = re.sub(
3168 r'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
3169 '', manifest_url).strip('?')
3174 hdcore_sign = 'hdcore=3.7.0'
3175 f4m_url = re.sub(r'(https?://[^/]+)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
3176 hds_host = hosts.get('hds')
3178 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
3179 if 'hdcore=' not in f4m_url:
3180 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
3181 f4m_formats = self._extract_f4m_formats(
3182 f4m_url, video_id, f4m_id='hds', fatal=False)
3183 for entry in f4m_formats:
3184 entry.update({'extra_param_to_segment_url': hdcore_sign})
3185 formats.extend(f4m_formats)
3187 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
3188 hls_host = hosts.get('hls')
3190 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
3191 m3u8_formats, m3u8_subtitles = self._extract_m3u8_formats_and_subtitles(
3192 m3u8_url, video_id, 'mp4', 'm3u8_native',
3193 m3u8_id='hls', fatal=False)
3194 formats.extend(m3u8_formats)
3195 subtitles = self._merge_subtitles(subtitles, m3u8_subtitles)
3197 http_host = hosts.get('http')
3198 if http_host and m3u8_formats and not signed:
3199 REPL_REGEX = r'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
3200 qualities = re.match(REPL_REGEX, m3u8_url).group(2).split(',')
3201 qualities_length = len(qualities)
3202 if len(m3u8_formats) in (qualities_length, qualities_length + 1):
3204 for f in m3u8_formats:
3205 if f['vcodec'] != 'none':
3206 for protocol in ('http', 'https'):
3208 del http_f['manifest_url']
3210 REPL_REGEX, protocol + fr'://{http_host}/\g<1>{qualities[i]}\3', f['url'])
3212 'format_id': http_f['format_id'].replace('hls-', protocol + '-'),
3214 'protocol': protocol,
3216 formats.append(http_f)
3219 return formats, subtitles
3221 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
3222 query = urllib.parse.urlparse(url).query
3223 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
3225 r'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url)
3226 url_base = mobj.group('url')
3227 http_base_url = '%s%s:%s' % ('http', mobj.group('s') or '', url_base)
3230 def manifest_url(manifest):
3231 m_url = f'{http_base_url}/{manifest}'
3233 m_url += '?%s' % query
3236 if 'm3u8' not in skip_protocols:
3237 formats.extend(self._extract_m3u8_formats(
3238 manifest_url('playlist.m3u8'), video_id, 'mp4',
3239 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
3240 if 'f4m' not in skip_protocols:
3241 formats.extend(self._extract_f4m_formats(
3242 manifest_url('manifest.f4m'),
3243 video_id, f4m_id='hds', fatal=False))
3244 if 'dash' not in skip_protocols:
3245 formats.extend(self._extract_mpd_formats(
3246 manifest_url('manifest.mpd'),
3247 video_id, mpd_id='dash', fatal=False))
3248 if re.search(r'(?:/smil:|\.smil)', url_base):
3249 if 'smil' not in skip_protocols:
3250 rtmp_formats = self._extract_smil_formats(
3251 manifest_url('jwplayer.smil'),
3252 video_id, fatal=False)
3253 for rtmp_format in rtmp_formats:
3254 rtsp_format = rtmp_format.copy()
3255 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
3256 del rtsp_format['play_path']
3257 del rtsp_format['ext']
3258 rtsp_format.update({
3259 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
3260 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
3263 formats.extend([rtmp_format, rtsp_format])
3265 for protocol in ('rtmp', 'rtsp'):
3266 if protocol not in skip_protocols:
3268 'url': f'{protocol}:{url_base}',
3269 'format_id': protocol,
3270 'protocol': protocol,
3274 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
3276 r'''(?s)jwplayer\s*\(\s*(?P<q>'|")(?
!(?P
=q
)).+(?P
=q
)\s
*\
)(?
!</script
>).*?\
.\s
*setup\s
*\
(\s
*(?P
<options
>(?
:\
([^
)]*\
)|
[^
)])+)\s
*\
)''',
3280 jwplayer_data = self._parse_json(mobj.group('options'),
3282 transform_source=transform_source)
3283 except ExtractorError:
3286 if isinstance(jwplayer_data, dict):
3287 return jwplayer_data
3289 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
3290 jwplayer_data = self._find_jwplayer_data(
3291 webpage, video_id, transform_source=js_to_json)
3292 return self._parse_jwplayer_data(
3293 jwplayer_data, video_id, *args, **kwargs)
3295 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
3296 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3298 if not isinstance(jwplayer_data, dict):
3301 playlist_items = jwplayer_data.get('playlist')
3302 # JWPlayer backward compatibility: single playlist item/flattened playlists
3303 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
3304 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
3305 if not isinstance(playlist_items, list):
3306 playlist_items = (playlist_items or jwplayer_data, )
3308 for video_data in playlist_items:
3309 if not isinstance(video_data, dict):
3311 # JWPlayer backward compatibility: flattened sources
3312 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
3313 if 'sources' not in video_data:
3314 video_data['sources'] = [video_data]
3316 this_video_id = video_id or video_data['mediaid']
3318 formats = self._parse_jwplayer_formats(
3319 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
3320 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
3323 tracks = video_data.get('tracks')
3324 if tracks and isinstance(tracks, list):
3325 for track in tracks:
3326 if not isinstance(track, dict):
3328 track_kind = track.get('kind')
3329 if not track_kind or not isinstance(track_kind, str):
3331 if track_kind.lower() not in ('captions', 'subtitles'):
3333 track_url = urljoin(base_url, track.get('file'))
3336 subtitles.setdefault(track.get('label') or 'en', []).append({
3337 'url': self._proto_relative_url(track_url)
3341 'id': this_video_id,
3342 'title': unescapeHTML(video_data['title'] if require_title else video_data.get('title')),
3343 'description': clean_html(video_data.get('description')),
3344 'thumbnail': urljoin(base_url, self._proto_relative_url(video_data.get('image'))),
3345 'timestamp': int_or_none(video_data.get('pubdate')),
3346 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
3347 'subtitles': subtitles,
3348 'alt_title': clean_html(video_data.get('subtitle')), # attributes used e.g. by Tele5 ...
3349 'genre': clean_html(video_data.get('genre')),
3350 'channel': clean_html(dict_get(video_data, ('category', 'channel'))),
3351 'season_number': int_or_none(video_data.get('season')),
3352 'episode_number': int_or_none(video_data.get('episode')),
3353 'release_year': int_or_none(video_data.get('releasedate')),
3354 'age_limit': int_or_none(video_data.get('age_restriction')),
3356 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
3357 if len(formats) == 1 and re.search(r'^(?:http|//).*(?:youtube\.com|youtu\.be)/.+', formats[0]['url']):
3359 '_type': 'url_transparent',
3360 'url': formats[0]['url'],
3363 entry['formats'] = formats
3364 entries.append(entry)
3365 if len(entries) == 1:
3368 return self.playlist_result(entries)
3370 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
3371 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3374 for source in jwplayer_sources_data:
3375 if not isinstance(source, dict):
3377 source_url = urljoin(
3378 base_url, self._proto_relative_url(source.get('file')))
3379 if not source_url or source_url in urls:
3381 urls.add(source_url)
3382 source_type = source.get('type') or ''
3383 ext = mimetype2ext(source_type) or determine_ext(source_url)
3384 if source_type == 'hls' or ext == 'm3u8' or 'format=m3u8-aapl' in source_url:
3385 formats.extend(self._extract_m3u8_formats(
3386 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
3387 m3u8_id=m3u8_id, fatal=False))
3388 elif source_type == 'dash' or ext == 'mpd' or 'format=mpd-time-csf' in source_url:
3389 formats.extend(self._extract_mpd_formats(
3390 source_url, video_id, mpd_id=mpd_id, fatal=False))
3392 formats.extend(self._extract_smil_formats(
3393 source_url, video_id, fatal=False))
3394 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3395 elif source_type.startswith('audio') or ext in (
3396 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
3403 format_id = str_or_none(source.get('label'))
3404 height = int_or_none(source.get('height'))
3405 if height is None and format_id:
3406 # Often no height is provided but there is a label in
3407 # format like "1080p", "720p SD", or 1080.
3408 height = parse_resolution(format_id).get('height')
3411 'width': int_or_none(source.get('width')),
3413 'tbr': int_or_none(source.get('bitrate'), scale=1000),
3414 'filesize': int_or_none(source.get('filesize')),
3416 'format_id': format_id
3418 if source_url.startswith('rtmp'):
3419 a_format['ext'] = 'flv'
3420 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3421 # of jwplayer.flash.swf
3422 rtmp_url_parts = re.split(
3423 r'((?:mp4|mp3|flv):)', source_url, 1)
3424 if len(rtmp_url_parts) == 3:
3425 rtmp_url, prefix, play_path = rtmp_url_parts
3428 'play_path': prefix + play_path,
3431 a_format.update(rtmp_params)
3432 formats.append(a_format)
3435 def _live_title(self, name):
3436 self._downloader.deprecation_warning('yt_dlp.InfoExtractor._live_title is deprecated and does not work as expected')
3439 def _int(self, v, name, fatal=False, **kwargs):
3440 res = int_or_none(v, **kwargs)
3442 msg = f'Failed to extract {name}: Could not parse value {v!r}'
3444 raise ExtractorError(msg)
3446 self.report_warning(msg)
3449 def _float(self, v, name, fatal=False, **kwargs):
3450 res = float_or_none(v, **kwargs)
3452 msg = f'Failed to extract {name}: Could not parse value {v!r}'
3454 raise ExtractorError(msg)
3456 self.report_warning(msg)
3459 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
3460 path='/', secure=False, discard=False, rest={}, **kwargs):
3461 cookie = http.cookiejar.Cookie(
3462 0, name, value, port, port is not None, domain, True,
3463 domain.startswith('.'), path, True, secure, expire_time,
3464 discard, None, None, rest)
3465 self.cookiejar.set_cookie(cookie)
3467 def _get_cookies(self, url):
3468 """ Return a http.cookies.SimpleCookie with the cookies for the url """
3469 return LenientSimpleCookie(self._downloader.cookiejar.get_cookie_header(url))
3471 def _apply_first_set_cookie_header(self, url_handle, cookie):
3473 Apply first Set-Cookie header instead of the last. Experimental.
3475 Some sites (e.g. [1-3]) may serve two cookies under the same name
3476 in Set-Cookie header and expect the first (old) one to be set rather
3477 than second (new). However, as of RFC6265 the newer one cookie
3478 should be set into cookie store what actually happens.
3479 We will workaround this issue by resetting the cookie to
3480 the first one manually.
3481 1. https://new.vk.com/
3482 2. https://github.com/ytdl-org/youtube-dl/issues/9841#issuecomment-227871201
3483 3. https://learning.oreilly.com/
3485 for header, cookies in url_handle.headers.items():
3486 if header.lower() != 'set-cookie':
3488 cookies = cookies.encode('iso-8859-1').decode('utf-8')
3489 cookie_value = re.search(
3490 r'%s=(.+?);.*?\b[Dd]omain=(.+?)(?:[,;]|$)' % cookie, cookies)
3492 value, domain = cookie_value.groups()
3493 self._set_cookie(domain, cookie, value)
3497 def get_testcases(cls, include_onlymatching=False):
3498 # Do not look in super classes
3499 t = vars(cls).get('_TEST')
3501 assert not hasattr(cls, '_TESTS'), f'{cls.ie_key()}IE has _TEST and _TESTS'
3504 tests = vars(cls).get('_TESTS', [])
3506 if not include_onlymatching and t.get('only_matching', False):
3508 t['name'] = cls.ie_key()
3510 if getattr(cls, '__wrapped__', None):
3511 yield from cls.__wrapped__.get_testcases(include_onlymatching)
3514 def get_webpage_testcases(cls):
3515 tests = vars(cls).get('_WEBPAGE_TESTS', [])
3517 t['name'] = cls.ie_key()
3519 if getattr(cls, '__wrapped__', None):
3520 yield from cls.__wrapped__.get_webpage_testcases()
3522 @classproperty(cache=True)
3524 """Get age limit from the testcases"""
3525 return max(traverse_obj(
3526 (*cls.get_testcases(include_onlymatching=False), *cls.get_webpage_testcases()),
3527 (..., (('playlist', 0), None), 'info_dict', 'age_limit')) or [0])
3529 @classproperty(cache=True)
3530 def _RETURN_TYPE(cls):
3531 """What the extractor returns: "video", "playlist", "any", or None (Unknown)"""
3532 tests = tuple(cls.get_testcases(include_onlymatching=False))
3535 elif not any(k.startswith('playlist') for test in tests for k in test):
3537 elif all(any(k.startswith('playlist') for k in test) for test in tests):
3542 def is_single_video(cls, url):
3543 """Returns whether the URL is of a single video, None if unknown"""
3544 if cls.suitable(url):
3545 return {'video': True, 'playlist': False}.get(cls._RETURN_TYPE)
3548 def is_suitable(cls, age_limit):
3549 """Test whether the extractor is generally suitable for the given age limit"""
3550 return not age_restricted(cls.age_limit, age_limit)
3553 def description(cls, *, markdown=True, search_examples=None):
3554 """Description of the extractor"""
3556 if cls._NETRC_MACHINE:
3558 desc += f' [*{cls._NETRC_MACHINE}*](## "netrc machine")'
3560 desc += f' [{cls._NETRC_MACHINE}]'
3561 if cls.IE_DESC is False:
3564 desc += f' {cls.IE_DESC}'
3566 desc += f'{";" if cls.IE_DESC else ""} "{cls.SEARCH_KEY}:" prefix'
3568 _COUNTS = ('', '5', '10', 'all')
3569 desc += f' (e.g. "{cls.SEARCH_KEY}{random.choice(_COUNTS)}:{random.choice(search_examples)}")'
3570 if not cls.working():
3571 desc += ' (**Currently broken**)' if markdown else ' (Currently broken)'
3573 # Escape emojis. Ref: https://github.com/github/markup/issues/1153
3574 name = (' - **%s**' % re.sub(r':(\w+:)', ':\u200B\\g<1>', cls.IE_NAME)) if markdown else cls.IE_NAME
3575 return f'{name}:{desc}' if desc else name
3577 def extract_subtitles(self, *args, **kwargs):
3578 if (self.get_param('writesubtitles', False)
3579 or self.get_param('listsubtitles')):
3580 return self._get_subtitles(*args, **kwargs)
3583 def _get_subtitles(self, *args, **kwargs):
3584 raise NotImplementedError('This method must be implemented by subclasses')
3586 class CommentsDisabled(Exception):
3587 """Raise in _get_comments if comments are disabled for the video"""
3589 def extract_comments(self, *args, **kwargs):
3590 if not self.get_param('getcomments'):
3592 generator = self._get_comments(*args, **kwargs)
3599 comments.append(next(generator))
3600 except StopIteration:
3602 except KeyboardInterrupt:
3603 self.to_screen('Interrupted by user')
3604 except self.CommentsDisabled:
3605 return {'comments': None, 'comment_count': None}
3606 except Exception as e:
3607 if self.get_param('ignoreerrors') is not True:
3609 self._downloader.report_error(e)
3610 comment_count = len(comments)
3611 self.to_screen(f'Extracted {comment_count} comments')
3613 'comments': comments,
3614 'comment_count': None if interrupted else comment_count
3618 def _get_comments(self, *args, **kwargs):
3619 raise NotImplementedError('This method must be implemented by subclasses')
3622 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
3623 """ Merge subtitle items for one language. Items with duplicated URLs/data
3624 will be dropped. """
3625 list1_data = {(item.get('url'), item.get('data')) for item in subtitle_list1}
3626 ret = list(subtitle_list1)
3627 ret.extend(item for item in subtitle_list2 if (item.get('url'), item.get('data')) not in list1_data)
3631 def _merge_subtitles(cls, *dicts, target=None):
3632 """ Merge subtitle dictionaries, language by language. """
3636 for lang, subs in d.items():
3637 target[lang] = cls._merge_subtitle_items(target.get(lang, []), subs)
3640 def extract_automatic_captions(self, *args, **kwargs):
3641 if (self.get_param('writeautomaticsub', False)
3642 or self.get_param('listsubtitles')):
3643 return self._get_automatic_captions(*args, **kwargs)
3646 def _get_automatic_captions(self, *args, **kwargs):
3647 raise NotImplementedError('This method must be implemented by subclasses')
3649 @functools.cached_property
3650 def _cookies_passed(self):
3651 """Whether cookies have been passed to YoutubeDL"""
3652 return self.get_param('cookiefile') is not None or self.get_param('cookiesfrombrowser') is not None
3654 def mark_watched(self, *args, **kwargs):
3655 if not self.get_param('mark_watched', False):
3657 if self.supports_login() and self._get_login_info()[0] is not None or self._cookies_passed:
3658 self._mark_watched(*args, **kwargs)
3660 def _mark_watched(self, *args, **kwargs):
3661 raise NotImplementedError('This method must be implemented by subclasses')
3663 def geo_verification_headers(self):
3665 geo_verification_proxy = self.get_param('geo_verification_proxy')
3666 if geo_verification_proxy:
3667 headers['Ytdl-request-proxy'] = geo_verification_proxy
3671 def _generic_id(url):
3672 return urllib.parse.unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
3674 def _generic_title(self, url='', webpage='', *, default=None):
3675 return (self._og_search_title(webpage, default=None)
3676 or self._html_extract_title(webpage, default=None)
3677 or urllib.parse.unquote(os.path.splitext(url_basename(url))[0])
3680 def _extract_chapters_helper(self, chapter_list, start_function, title_function, duration, strict=True):
3684 'start_time': start_function(chapter),
3685 'title': title_function(chapter),
3686 } for chapter in chapter_list or []]
3688 warn = self.report_warning
3690 warn = self.write_debug
3691 chapter_list.sort(key=lambda c: c['start_time'] or 0)
3693 chapters = [{'start_time': 0}]
3694 for idx, chapter in enumerate(chapter_list):
3695 if chapter['start_time'] is None:
3696 warn(f'Incomplete chapter {idx}')
3697 elif chapters[-1]['start_time'] <= chapter['start_time'] <= duration:
3698 chapters.append(chapter)
3699 elif chapter not in chapters:
3700 issue = (f'{chapter["start_time"]} > {duration}' if chapter['start_time'] > duration
3701 else f'{chapter["start_time"]} < {chapters[-1]["start_time"]}')
3702 warn(f'Invalid start time ({issue}) for chapter "{chapter["title"]}"')
3705 def _extract_chapters_from_description(self, description, duration):
3706 duration_re = r'(?:\d+:)?\d{1,2}:\d{2}'
3707 sep_re = r'(?m)^\s*(%s)\b\W*\s(%s)\s*$'
3708 return self._extract_chapters_helper(
3709 re.findall(sep_re % (duration_re, r'.+?'), description or ''),
3710 start_function=lambda x: parse_duration(x[0]), title_function=lambda x: x[1],
3711 duration=duration, strict=False) or self._extract_chapters_helper(
3712 re.findall(sep_re % (r'.+?', duration_re), description or ''),
3713 start_function=lambda x: parse_duration(x[1]), title_function=lambda x: x[0],
3714 duration=duration, strict=False)
3717 def _availability(is_private=None, needs_premium=None, needs_subscription=None, needs_auth=None, is_unlisted=None):
3718 all_known = all(map(
3719 lambda x: x is not None,
3720 (is_private, needs_premium, needs_subscription, needs_auth, is_unlisted)))
3722 'private' if is_private
3723 else 'premium_only' if needs_premium
3724 else 'subscriber_only' if needs_subscription
3725 else 'needs_auth' if needs_auth
3726 else 'unlisted' if is_unlisted
3727 else 'public' if all_known
3730 def _configuration_arg(self, key, default=NO_DEFAULT, *, ie_key=None, casesense=False):
3732 @returns A
list of values
for the extractor argument given by
"key"
3733 or "default" if no such key
is present
3734 @param default The default value to
return when the key
is not present (default
: [])
3735 @param casesense When false
, the values are converted to lower case
3737 ie_key = ie_key if isinstance(ie_key, str) else (ie_key or self).ie_key()
3738 val = traverse_obj(self._downloader.params, ('extractor_args', ie_key.lower(), key))
3740 return [] if default is NO_DEFAULT else default
3741 return list(val) if casesense else [x.lower() for x in val]
3743 def _yes_playlist(self, playlist_id, video_id, smuggled_data=None, *, playlist_label='playlist', video_label='video'):
3744 if not playlist_id or not video_id:
3747 no_playlist = (smuggled_data or {}).get('force_noplaylist')
3748 if no_playlist is not None:
3749 return not no_playlist
3751 video_id = '' if video_id is True else f' {video_id}'
3752 playlist_id = '' if playlist_id is True else f' {playlist_id}'
3753 if self.get_param('noplaylist'):
3754 self.to_screen(f'Downloading just the {video_label}{video_id} because of --no-playlist')
3756 self.to_screen(f'Downloading {playlist_label}{playlist_id} - add --no-playlist to download just the {video_label}{video_id}')
3759 def _error_or_warning(self, err, _count=None, _retries=0, *, fatal=True):
3760 RetryManager.report_retry(
3761 err, _count or int(fatal), _retries,
3762 info=self.to_screen, warn=self.report_warning, error=None if fatal else self.report_warning,
3763 sleep_func=self.get_param('retry_sleep_functions', {}).get('extractor'))
3765 def RetryManager(self, **kwargs):
3766 return RetryManager(self.get_param('extractor_retries', 3), self._error_or_warning, **kwargs)
3768 def _extract_generic_embeds(self, url, *args, info_dict={}, note='Extracting generic embeds', **kwargs):
3769 display_id = traverse_obj(info_dict, 'display_id', 'id')
3770 self.to_screen(f'{format_field(display_id, None, "%s: ")}{note}')
3771 return self._downloader.get_info_extractor('Generic')._extract_embeds(
3772 smuggle_url(url, {'block_ies': [self.ie_key()]}), *args, **kwargs)
3775 def extract_from_webpage(cls, ydl, url, webpage):
3776 ie = (cls if isinstance(cls._extract_from_webpage, types.MethodType)
3777 else ydl.get_info_extractor(cls.ie_key()))
3778 for info in ie._extract_from_webpage(url, webpage) or []:
3779 # url = None since we do not want to set (webpage/original)_url
3780 ydl.add_default_extra_info(info, ie, None)
3784 def _extract_from_webpage(cls, url, webpage):
3785 for embed_url in orderedSet(
3786 cls._extract_embed_urls(url, webpage) or [], lazy=True):
3787 yield cls.url_result(embed_url, None if cls._VALID_URL is False else cls)
3790 def _extract_embed_urls(cls, url, webpage):
3791 """@returns all the embed urls on the webpage"""
3792 if '_EMBED_URL_RE' not in cls.__dict__:
3793 assert isinstance(cls._EMBED_REGEX, (list, tuple))
3794 for idx, regex in enumerate(cls._EMBED_REGEX):
3795 assert regex.count('(?P<url>') == 1, \
3796 f'{cls.__name__}._EMBED_REGEX[{idx}] must have exactly 1 url group\n\t{regex}'
3797 cls._EMBED_URL_RE = tuple(map(re.compile, cls._EMBED_REGEX))
3799 for regex in cls._EMBED_URL_RE:
3800 for mobj in regex.finditer(webpage):
3801 embed_url = urllib.parse.urljoin(url, unescapeHTML(mobj.group('url')))
3802 if cls._VALID_URL is False or cls.suitable(embed_url):
3805 class StopExtraction(Exception):
3809 def _extract_url(cls, webpage): # TODO: Remove
3810 """Only for compatibility with some older extractors"""
3811 return next(iter(cls._extract_embed_urls(None, webpage) or []), None)
3814 def __init_subclass__(cls, *, plugin_name=None, **kwargs):
3816 mro = inspect.getmro(cls)
3817 super_class = cls.__wrapped__ = mro[mro.index(cls) + 1]
3818 cls.PLUGIN_NAME, cls.ie_key = plugin_name, super_class.ie_key
3819 cls.IE_NAME = f'{super_class.IE_NAME}+{plugin_name}'
3820 while getattr(super_class, '__wrapped__', None):
3821 super_class = super_class.__wrapped__
3822 setattr(sys.modules[super_class.__module__], super_class.__name__, cls)
3823 _PLUGIN_OVERRIDES[super_class].append(cls)
3825 return super().__init_subclass__(**kwargs)
3828 class SearchInfoExtractor(InfoExtractor):
3830 Base class for paged search queries extractors.
3831 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
3832 Instances should define _SEARCH_KEY and optionally _MAX_RESULTS
3835 _MAX_RESULTS = float('inf')
3836 _RETURN_TYPE = 'playlist'
3839 def _VALID_URL(cls):
3840 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
3842 def _real_extract(self, query):
3843 prefix, query = self._match_valid_url(query).group('prefix', 'query')
3845 return self._get_n_results(query, 1)
3846 elif prefix == 'all':
3847 return self._get_n_results(query, self._MAX_RESULTS)
3851 raise ExtractorError(f'invalid download number {n} for query "{query}"')
3852 elif n > self._MAX_RESULTS:
3853 self.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3854 n = self._MAX_RESULTS
3855 return self._get_n_results(query, n)
3857 def _get_n_results(self, query, n):
3858 """Get a specified number of results for a query.
3859 Either this function or _search_results must be overridden by subclasses """
3860 return self.playlist_result(
3861 itertools.islice(self._search_results(query), 0, None if n == float('inf') else n),
3864 def _search_results(self, query):
3865 """Returns an iterator of search results"""
3866 raise NotImplementedError('This method must be implemented by subclasses')
3869 def SEARCH_KEY(cls):
3870 return cls._SEARCH_KEY
3873 class UnsupportedURLIE(InfoExtractor):
3878 def _real_extract(self, url):
3879 raise UnsupportedError(url)
3882 _PLUGIN_OVERRIDES = collections.defaultdict(list)