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 (
40 from ..networking
.impersonate
import ImpersonateTarget
79 parse_m3u8_attributes
,
106 """Information Extractor class.
108 Information extractors are the classes that, given a URL, extract
109 information about the video (or videos) the URL refers to. This
110 information includes the real video URL, the video title, author and
111 others. The information is stored in a dictionary which is then
112 passed to the YoutubeDL. The YoutubeDL processes this
113 information possibly downloading the video to the file system, among
114 other possible outcomes.
116 The type field determines the type of the result.
117 By far the most common value (and the default if _type is missing) is
118 "video", which indicates a single video.
120 For a video, the dictionaries must include the following fields:
122 id: Video identifier.
123 title: Video title, unescaped. Set to an empty string if video has
124 no title as opposed to "None" which signifies that the
125 extractor failed to obtain a title
127 Additionally, it must contain either a formats entry or a url one:
129 formats: A list of dictionaries for each format available, ordered
130 from worst to best quality.
133 * url The mandatory URL representing the media:
134 for plain file media - HTTP URL of this file,
136 for HLS - URL of the M3U8 media playlist,
137 for HDS - URL of the F4M manifest,
139 - HTTP URL to plain file media (in case of
141 - URL of the MPD manifest or base URL
142 representing the media if MPD manifest
143 is parsed from a string (in case of
145 for MSS - URL of the ISM manifest.
146 * request_data Data to send in POST request to the URL
148 The URL of the manifest file in case of
150 for HLS - URL of the M3U8 master playlist,
151 for HDS - URL of the F4M manifest,
152 for DASH - URL of the MPD manifest,
153 for MSS - URL of the ISM manifest.
154 * manifest_stream_number (For internal use only)
155 The index of the stream in the manifest file
156 * ext Will be calculated from URL if missing
157 * format A human-readable description of the format
158 ("mp4 container with h264/opus").
159 Calculated from the format_id, width, height.
160 and format_note fields if missing.
161 * format_id A short description of the format
162 ("mp4_h264_opus" or "19").
163 Technically optional, but strongly recommended.
164 * format_note Additional info about the format
165 ("3D" or "DASH video")
166 * width Width of the video, if known
167 * height Height of the video, if known
168 * aspect_ratio Aspect ratio of the video, if known
169 Automatically calculated from width and height
170 * resolution Textual description of width and height
171 Automatically calculated from width and height
172 * dynamic_range The dynamic range of the video. One of:
173 "SDR" (None), "HDR10", "HDR10+, "HDR12", "HLG, "DV"
174 * tbr Average bitrate of audio and video in kbps (1000 bits/sec)
175 * abr Average audio bitrate in kbps (1000 bits/sec)
176 * acodec Name of the audio codec in use
177 * asr Audio sampling rate in Hertz
178 * audio_channels Number of audio channels
179 * vbr Average video bitrate in kbps (1000 bits/sec)
181 * vcodec Name of the video codec in use
182 * container Name of the container format
183 * filesize The number of bytes, if known in advance
184 * filesize_approx An estimate for the number of bytes
185 * player_url SWF Player URL (used for rtmpdump).
186 * protocol The protocol that will be used for the actual
187 download, lower-case. One of "http", "https" or
188 one of the protocols defined in downloader.PROTOCOL_MAP
190 Base URL for fragments. Each fragment's path
191 value (if present) will be relative to
193 * fragments A list of fragments of a fragmented media.
194 Each fragment entry must contain either an url
195 or a path. If an url is present it should be
196 considered by a client. Otherwise both path and
197 fragment_base_url must be present. Here is
198 the list of all potential fields:
199 * "url" - fragment's URL
200 * "path" - fragment's path relative to
202 * "duration" (optional, int or float)
203 * "filesize" (optional, int)
204 * is_from_start Is a live format that can be downloaded
205 from the start. Boolean
206 * preference Order number of this format. If this field is
207 present and not None, the formats get sorted
208 by this field, regardless of all other values.
209 -1 for default (order by other properties),
210 -2 or smaller for less than default.
211 < -1000 to hide the format (if there is
212 another one which is strictly better)
213 * language Language code, e.g. "de" or "en-US".
214 * language_preference Is this in the language mentioned in
216 10 if it's what the URL is about,
217 -1 for default (don't know),
218 -10 otherwise, other values reserved for now.
219 * quality Order number of the video quality of this
220 format, irrespective of the file format.
221 -1 for default (order by other properties),
222 -2 or smaller for less than default.
223 * source_preference Order number for this video source
224 (quality takes higher priority)
225 -1 for default (order by other properties),
226 -2 or smaller for less than default.
227 * http_headers A dictionary of additional HTTP headers
228 to add to the request.
229 * stretched_ratio If given and not 1, indicates that the
230 video's pixels are not square.
231 width : height ratio as float.
232 * no_resume The server does not support resuming the
233 (HTTP or RTMP) download. Boolean.
234 * has_drm True if the format has DRM and cannot be downloaded.
235 'maybe' if the format may have DRM and has to be tested before download.
236 * extra_param_to_segment_url A query string to append to each
237 fragment's URL, or to update each existing query string
238 with. Only applied by the native HLS/DASH downloaders.
239 * hls_aes A dictionary of HLS AES-128 decryption information
240 used by the native HLS downloader to override the
241 values in the media playlist when an '#EXT-X-KEY' tag
242 is present in the playlist:
243 * uri The URI from which the key will be downloaded
244 * key The key (as hex) used to decrypt fragments.
245 If `key` is given, any key URI will be ignored
246 * iv The IV (as hex) used to decrypt fragments
247 * downloader_options A dictionary of downloader options
248 (For internal use only)
249 * http_chunk_size Chunk size for HTTP downloads
250 * ffmpeg_args Extra arguments for ffmpeg downloader (input)
251 * ffmpeg_args_out Extra arguments for ffmpeg downloader (output)
252 * is_dash_periods Whether the format is a result of merging
253 multiple DASH periods.
254 RTMP formats can also have the additional fields: page_url,
255 app, play_path, tc_url, flash_version, rtmp_live, rtmp_conn,
256 rtmp_protocol, rtmp_real_time
258 url: Final video URL.
259 ext: Video filename extension.
260 format: The video format, defaults to ext (used for --get-format)
261 player_url: SWF Player URL (used for rtmpdump).
263 The following fields are optional:
265 direct: True if a direct video file was given (must only be set by GenericIE)
266 alt_title: A secondary title of the video.
267 display_id: An alternative identifier for the video, not necessarily
268 unique, but available before title. Typically, id is
269 something like "4234987", title "Dancing naked mole rats",
270 and display_id "dancing-naked-mole-rats"
271 thumbnails: A list of dictionaries, with the following entries:
272 * "id" (optional, string) - Thumbnail format ID
274 * "preference" (optional, int) - quality of the image
275 * "width" (optional, int)
276 * "height" (optional, int)
277 * "resolution" (optional, string "{width}x{height}",
279 * "filesize" (optional, int)
280 * "http_headers" (dict) - HTTP headers for the request
281 thumbnail: Full URL to a video thumbnail image.
282 description: Full video description.
283 uploader: Full name of the video uploader.
284 license: License name the video is licensed under.
285 creators: List of creators of the video.
286 timestamp: UNIX timestamp of the moment the video was uploaded
287 upload_date: Video upload date in UTC (YYYYMMDD).
288 If not explicitly set, calculated from timestamp
289 release_timestamp: UNIX timestamp of the moment the video was released.
290 If it is not clear whether to use timestamp or this, use the former
291 release_date: The date (YYYYMMDD) when the video was released in UTC.
292 If not explicitly set, calculated from release_timestamp
293 release_year: Year (YYYY) as integer when the video or album was released.
294 To be used if no exact release date is known.
295 If not explicitly set, calculated from release_date.
296 modified_timestamp: UNIX timestamp of the moment the video was last modified.
297 modified_date: The date (YYYYMMDD) when the video was last modified in UTC.
298 If not explicitly set, calculated from modified_timestamp
299 uploader_id: Nickname or id of the video uploader.
300 uploader_url: Full URL to a personal webpage of the video uploader.
301 channel: Full name of the channel the video is uploaded on.
302 Note that channel fields may or may not repeat uploader
303 fields. This depends on a particular extractor.
304 channel_id: Id of the channel.
305 channel_url: Full URL to a channel webpage.
306 channel_follower_count: Number of followers of the channel.
307 channel_is_verified: Whether the channel is verified on the platform.
308 location: Physical location where the video was filmed.
309 subtitles: The available subtitles as a dictionary in the format
310 {tag: subformats}. "tag" is usually a language code, and
311 "subformats" is a list sorted from lower to higher
312 preference, each element is a dictionary with the "ext"
314 * "data": The subtitles file contents
315 * "url": A URL pointing to the subtitles file
316 It can optionally also have:
317 * "name": Name or description of the subtitles
318 * "http_headers": A dictionary of additional HTTP headers
319 to add to the request.
320 "ext" will be calculated from URL if missing
321 automatic_captions: Like 'subtitles'; contains automatically generated
322 captions instead of normal subtitles
323 duration: Length of the video in seconds, as an integer or float.
324 view_count: How many users have watched the video on the platform.
325 concurrent_view_count: How many users are currently watching the video on the platform.
326 like_count: Number of positive ratings of the video
327 dislike_count: Number of negative ratings of the video
328 repost_count: Number of reposts of the video
329 average_rating: Average rating give by users, the scale used depends on the webpage
330 comment_count: Number of comments on the video
331 comments: A list of comments, each with one or more of the following
332 properties (all but one of text or html optional):
333 * "author" - human-readable name of the comment author
334 * "author_id" - user ID of the comment author
335 * "author_thumbnail" - The thumbnail of the comment author
336 * "author_url" - The url to the comment author's page
337 * "author_is_verified" - Whether the author is verified
339 * "author_is_uploader" - Whether the comment is made by
342 * "html" - Comment as HTML
343 * "text" - Plain text of the comment
344 * "timestamp" - UNIX timestamp of comment
345 * "parent" - ID of the comment this one is replying to.
346 Set to "root" to indicate that this is a
347 comment to the original video.
348 * "like_count" - Number of positive ratings of the comment
349 * "dislike_count" - Number of negative ratings of the comment
350 * "is_favorited" - Whether the comment is marked as
351 favorite by the video uploader
352 * "is_pinned" - Whether the comment is pinned to
353 the top of the comments
354 age_limit: Age restriction for the video, as an integer (years)
355 webpage_url: The URL to the video webpage, if given to yt-dlp it
356 should allow to get the same result again. (It will be set
357 by YoutubeDL if it's missing)
358 categories: A list of categories that the video falls in, for example
360 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
361 cast: A list of the video cast
362 is_live: True, False, or None (=unknown). Whether this video is a
363 live stream that goes on instead of a fixed-length video.
364 was_live: True, False, or None (=unknown). Whether this video was
365 originally a live stream.
366 live_status: None (=unknown), 'is_live', 'is_upcoming', 'was_live', 'not_live',
367 or 'post_live' (was live, but VOD is not yet processed)
368 If absent, automatically set from is_live, was_live
369 start_time: Time in seconds where the reproduction should start, as
370 specified in the URL.
371 end_time: Time in seconds where the reproduction should end, as
372 specified in the URL.
373 chapters: A list of dictionaries, with the following entries:
374 * "start_time" - The start time of the chapter in seconds
375 * "end_time" - The end time of the chapter in seconds
376 * "title" (optional, string)
377 heatmap: A list of dictionaries, with the following entries:
378 * "start_time" - The start time of the data point in seconds
379 * "end_time" - The end time of the data point in seconds
380 * "value" - The normalized value of the data point (float between 0 and 1)
381 playable_in_embed: Whether this video is allowed to play in embedded
382 players on other sites. Can be True (=always allowed),
383 False (=never allowed), None (=unknown), or a string
384 specifying the criteria for embedability; e.g. 'whitelist'
385 availability: Under what condition the video is available. One of
386 'private', 'premium_only', 'subscriber_only', 'needs_auth',
387 'unlisted' or 'public'. Use 'InfoExtractor._availability'
389 media_type: The type of media as classified by the site, e.g. "episode", "clip", "trailer"
390 _old_archive_ids: A list of old archive ids needed for backward compatibility
391 _format_sort_fields: A list of fields to use for sorting formats
392 __post_extractor: A function to be called just before the metadata is
393 written to either disk, logger or console. The function
394 must return a dict which will be added to the info_dict.
395 This is usefull for additional information that is
396 time-consuming to extract. Note that the fields thus
397 extracted will not be available to output template and
398 match_filter. So, only "comments" and "comment_count" are
399 currently allowed to be extracted via this method.
401 The following fields should only be used when the video belongs to some logical
404 chapter: Name or title of the chapter the video belongs to.
405 chapter_number: Number of the chapter the video belongs to, as an integer.
406 chapter_id: Id of the chapter the video belongs to, as a unicode string.
408 The following fields should only be used when the video is an episode of some
409 series, programme or podcast:
411 series: Title of the series or programme the video episode belongs to.
412 series_id: Id of the series or programme the video episode belongs to, as a unicode string.
413 season: Title of the season the video episode belongs to.
414 season_number: Number of the season the video episode belongs to, as an integer.
415 season_id: Id of the season the video episode belongs to, as a unicode string.
416 episode: Title of the video episode. Unlike mandatory video title field,
417 this field should denote the exact title of the video episode
418 without any kind of decoration.
419 episode_number: Number of the video episode within a season, as an integer.
420 episode_id: Id of the video episode, as a unicode string.
422 The following fields should only be used when the media is a track or a part of
425 track: Title of the track.
426 track_number: Number of the track within an album or a disc, as an integer.
427 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
429 artists: List of artists of the track.
430 composers: List of composers of the piece.
431 genres: List of genres of the track.
432 album: Title of the album the track belongs to.
433 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
434 album_artists: List of all artists appeared on the album.
435 E.g. ["Ash Borer", "Fell Voices"] or ["Various Artists"].
436 Useful for splits and compilations.
437 disc_number: Number of the disc or other physical medium the track belongs to,
440 The following fields should only be set for clips that should be cut from the original video:
442 section_start: Start time of the section in seconds
443 section_end: End time of the section in seconds
445 The following fields should only be set for storyboards:
446 rows: Number of rows in each storyboard fragment, as an integer
447 columns: Number of columns in each storyboard fragment, as an integer
449 The following fields are deprecated and should not be set by new code:
450 composer: Use "composers" instead.
451 Composer(s) of the piece, comma-separated.
452 artist: Use "artists" instead.
453 Artist(s) of the track, comma-separated.
454 genre: Use "genres" instead.
455 Genre(s) of the track, comma-separated.
456 album_artist: Use "album_artists" instead.
457 All artists appeared on the album, comma-separated.
458 creator: Use "creators" instead.
459 The creator of the video.
461 Unless mentioned otherwise, the fields should be Unicode strings.
463 Unless mentioned otherwise, None is equivalent to absence of information.
466 _type "playlist" indicates multiple videos.
467 There must be a key "entries", which is a list, an iterable, or a PagedList
468 object, each element of which is a valid dictionary by this specification.
470 Additionally, playlists can have "id", "title", and any other relevant
471 attributes with the same semantics as videos (see above).
473 It can also have the following optional fields:
475 playlist_count: The total number of videos in a playlist. If not given,
476 YoutubeDL tries to calculate it from "entries"
479 _type "multi_video" indicates that there are multiple videos that
480 form a single show, for examples multiple acts of an opera or TV episode.
481 It must have an entries key like a playlist and contain all the keys
482 required for a video at the same time.
485 _type "url" indicates that the video must be extracted from another
486 location, possibly by a different extractor. Its only required key is:
487 "url" - the next URL to extract.
488 The key "ie_key" can be set to the class name (minus the trailing "IE",
489 e.g. "Youtube") if the extractor class is known in advance.
490 Additionally, the dictionary may have any properties of the resolved entity
491 known in advance, for example "title" if the title of the referred video is
495 _type "url_transparent" entities have the same specification as "url", but
496 indicate that the given additional information is more precise than the one
497 associated with the resolved URL.
498 This is useful when a site employs a video service that hosts the video and
499 its technical metadata, but that video service does not embed a useful
500 title, description etc.
503 Subclasses of this should also be added to the list of extractors and
504 should define _VALID_URL as a regexp or a Sequence of regexps, and
505 re-define the _real_extract() and (optionally) _real_initialize() methods.
507 Subclasses may also override suitable() if necessary, but ensure the function
508 signature is preserved and that this function imports everything it needs
509 (except other extractors), so that lazy_extractors works correctly.
511 Subclasses can define a list of _EMBED_REGEX, which will be searched for in
512 the HTML of Generic webpages. It may also override _extract_embed_urls
513 or _extract_from_webpage as necessary. While these are normally classmethods,
514 _extract_from_webpage is allowed to be an instance method.
516 _extract_from_webpage may raise self.StopExtraction() to stop further
517 processing of the webpage and obtain exclusive rights to it. This is useful
518 when the extractor cannot reliably be matched using just the URL,
519 e.g. invidious/peertube instances
521 Embed-only extractors can be defined by setting _VALID_URL = False.
523 To support username + password (or netrc) login, the extractor must define a
524 _NETRC_MACHINE and re-define _perform_login(username, password) and
525 (optionally) _initialize_pre_login() methods. The _perform_login method will
526 be called between _initialize_pre_login and _real_initialize if credentials
527 are passed by the user. In cases where it is necessary to have the login
528 process as part of the extraction rather than initialization, _perform_login
529 can be left undefined.
531 _GEO_BYPASS attribute may be set to False in order to disable
532 geo restriction bypass mechanisms for a particular extractor.
533 Though it won't disable explicit geo restriction bypass based on
534 country code provided with geo_bypass_country.
536 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
537 countries for this extractor. One of these countries will be used by
538 geo restriction bypass mechanism right away in order to bypass
539 geo restriction, of course, if the mechanism is not disabled.
541 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
542 IP blocks in CIDR notation for this extractor. One of these IP blocks
543 will be used by geo restriction bypass mechanism similarly
546 The _ENABLED attribute should be set to False for IEs that
547 are disabled by default and must be explicitly enabled.
549 The _WORKING attribute should be set to False for broken IEs
550 in order to warn the users and skip the tests.
555 _x_forwarded_for_ip
= None
557 _GEO_COUNTRIES
= None
558 _GEO_IP_BLOCKS
= None
561 _NETRC_MACHINE
= None
567 def _login_hint(self
, method
=NO_DEFAULT
, netrc
=None):
568 password_hint
= f
'--username and --password, --netrc-cmd, or --netrc ({netrc or self._NETRC_MACHINE}) to provide account credentials'
571 'any': f
'Use --cookies, --cookies-from-browser, {password_hint}',
572 'password': f
'Use {password_hint}',
574 'Use --cookies-from-browser or --cookies for the authentication. '
575 'See https://github.com/yt-dlp/yt-dlp/wiki/FAQ#how-do-i-pass-cookies-to-yt-dlp for how to manually pass cookies'),
576 }[method
if method
is not NO_DEFAULT
else 'any' if self
.supports_login() else 'cookies']
578 def __init__(self
, downloader
=None):
579 """Constructor. Receives an optional downloader (a YoutubeDL instance).
580 If a downloader is not passed during initialization,
581 it must be set using "set_downloader()" before "extract()" is called"""
583 self
._x
_forwarded
_for
_ip
= None
584 self
._printed
_messages
= set()
585 self
.set_downloader(downloader
)
588 def _match_valid_url(cls
, url
):
589 if cls
._VALID
_URL
is False:
591 # This does not use has/getattr intentionally - we want to know whether
592 # we have cached the regexp for *this* class, whereas getattr would also
593 # match the superclass
594 if '_VALID_URL_RE' not in cls
.__dict
__:
595 cls
._VALID
_URL
_RE
= tuple(map(re
.compile, variadic(cls
._VALID
_URL
)))
596 return next(filter(None, (regex
.match(url
) for regex
in cls
._VALID
_URL
_RE
)), None)
599 def suitable(cls
, url
):
600 """Receives a URL and returns True if suitable for this IE."""
601 # This function must import everything it needs (except other extractors),
602 # so that lazy_extractors works correctly
603 return cls
._match
_valid
_url
(url
) is not None
606 def _match_id(cls
, url
):
607 return cls
._match
_valid
_url
(url
).group('id')
610 def get_temp_id(cls
, url
):
612 return cls
._match
_id
(url
)
613 except (IndexError, AttributeError):
618 """Getter method for _WORKING."""
622 def supports_login(cls
):
623 return bool(cls
._NETRC
_MACHINE
)
625 def initialize(self
):
626 """Initializes an instance (authentication, etc)."""
627 self
._printed
_messages
= set()
628 self
._initialize
_geo
_bypass
({
629 'countries': self
._GEO
_COUNTRIES
,
630 'ip_blocks': self
._GEO
_IP
_BLOCKS
,
633 self
._initialize
_pre
_login
()
634 if self
.supports_login():
635 username
, password
= self
._get
_login
_info
()
637 self
._perform
_login
(username
, password
)
638 elif self
.get_param('username') and False not in (self
.IE_DESC
, self
._NETRC
_MACHINE
):
639 self
.report_warning(f
'Login with password is not supported for this website. {self._login_hint("cookies")}')
640 self
._real
_initialize
()
643 def _initialize_geo_bypass(self
, geo_bypass_context
):
645 Initialize geo restriction bypass mechanism.
647 This method is used to initialize geo bypass mechanism based on faking
648 X-Forwarded-For HTTP header. A random country from provided country list
649 is selected and a random IP belonging to this country is generated. This
650 IP will be passed as X-Forwarded-For HTTP header in all subsequent
653 This method will be used for initial geo bypass mechanism initialization
654 during the instance initialization with _GEO_COUNTRIES and
657 You may also manually call it from extractor's code if geo bypass
658 information is not available beforehand (e.g. obtained during
659 extraction) or due to some other reason. In this case you should pass
660 this information in geo bypass context passed as first argument. It may
661 contain following fields:
663 countries: List of geo unrestricted countries (similar
665 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
666 (similar to _GEO_IP_BLOCKS)
669 if not self
._x
_forwarded
_for
_ip
:
671 # Geo bypass mechanism is explicitly disabled by user
672 if not self
.get_param('geo_bypass', True):
675 if not geo_bypass_context
:
676 geo_bypass_context
= {}
678 # Backward compatibility: previously _initialize_geo_bypass
679 # expected a list of countries, some 3rd party code may still use
681 if isinstance(geo_bypass_context
, (list, tuple)):
682 geo_bypass_context
= {
683 'countries': geo_bypass_context
,
686 # The whole point of geo bypass mechanism is to fake IP
687 # as X-Forwarded-For HTTP header based on some IP block or
690 # Path 1: bypassing based on IP block in CIDR notation
692 # Explicit IP block specified by user, use it right away
693 # regardless of whether extractor is geo bypassable or not
694 ip_block
= self
.get_param('geo_bypass_ip_block', None)
696 # Otherwise use random IP block from geo bypass context but only
697 # if extractor is known as geo bypassable
699 ip_blocks
= geo_bypass_context
.get('ip_blocks')
700 if self
._GEO
_BYPASS
and ip_blocks
:
701 ip_block
= random
.choice(ip_blocks
)
704 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(ip_block
)
705 self
.write_debug(f
'Using fake IP {self._x_forwarded_for_ip} as X-Forwarded-For')
708 # Path 2: bypassing based on country code
710 # Explicit country code specified by user, use it right away
711 # regardless of whether extractor is geo bypassable or not
712 country
= self
.get_param('geo_bypass_country', None)
714 # Otherwise use random country code from geo bypass context but
715 # only if extractor is known as geo bypassable
717 countries
= geo_bypass_context
.get('countries')
718 if self
._GEO
_BYPASS
and countries
:
719 country
= random
.choice(countries
)
722 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country
)
723 self
._downloader
.write_debug(
724 f
'Using fake IP {self._x_forwarded_for_ip} ({country.upper()}) as X-Forwarded-For')
726 def extract(self
, url
):
727 """Extracts URL information and returns it in list of dicts."""
732 self
.to_screen('Extracting URL: %s' % (
733 url
if self
.get_param('verbose') else truncate_string(url
, 100, 20)))
734 ie_result
= self
._real
_extract
(url
)
735 if ie_result
is None:
737 if self
._x
_forwarded
_for
_ip
:
738 ie_result
['__x_forwarded_for_ip'] = self
._x
_forwarded
_for
_ip
739 subtitles
= ie_result
.get('subtitles') or {}
740 if 'no-live-chat' in self
.get_param('compat_opts'):
741 for lang
in ('live_chat', 'comments', 'danmaku'):
742 subtitles
.pop(lang
, None)
744 except GeoRestrictedError
as e
:
745 if self
.__maybe
_fake
_ip
_and
_retry
(e
.countries
):
748 except UnsupportedError
:
750 except ExtractorError
as e
:
751 e
.video_id
= e
.video_id
or self
.get_temp_id(url
)
752 e
.ie
= e
.ie
or self
.IE_NAME
753 e
.traceback
= e
.traceback
or sys
.exc_info()[2]
755 except IncompleteRead
as e
:
756 raise ExtractorError('A network error has occurred.', cause
=e
, expected
=True, video_id
=self
.get_temp_id(url
))
757 except (KeyError, StopIteration) as e
:
758 raise ExtractorError('An extractor error has occurred.', cause
=e
, video_id
=self
.get_temp_id(url
))
760 def __maybe_fake_ip_and_retry(self
, countries
):
761 if (not self
.get_param('geo_bypass_country', None)
763 and self
.get_param('geo_bypass', True)
764 and not self
._x
_forwarded
_for
_ip
766 country_code
= random
.choice(countries
)
767 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country_code
)
768 if self
._x
_forwarded
_for
_ip
:
770 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
771 % (self
._x
_forwarded
_for
_ip
, country_code
.upper()))
775 def set_downloader(self
, downloader
):
776 """Sets a YoutubeDL instance as the downloader for this IE."""
777 self
._downloader
= downloader
781 return self
._downloader
.cache
785 return self
._downloader
.cookiejar
787 def _initialize_pre_login(self
):
788 """ Initialization before login. Redefine in subclasses."""
791 def _perform_login(self
, username
, password
):
792 """ Login with username and password. Redefine in subclasses."""
795 def _real_initialize(self
):
796 """Real initialization process. Redefine in subclasses."""
799 def _real_extract(self
, url
):
800 """Real extraction process. Redefine in subclasses."""
801 raise NotImplementedError('This method must be implemented by subclasses')
805 """A string for getting the InfoExtractor with get_info_extractor"""
806 return cls
.__name
__[:-2]
810 return cls
.__name
__[:-2]
813 def __can_accept_status_code(err
, expected_status
):
814 assert isinstance(err
, HTTPError
)
815 if expected_status
is None:
817 elif callable(expected_status
):
818 return expected_status(err
.status
) is True
820 return err
.status
in variadic(expected_status
)
822 def _create_request(self
, url_or_request
, data
=None, headers
=None, query
=None, extensions
=None):
823 if isinstance(url_or_request
, urllib
.request
.Request
):
824 self
._downloader
.deprecation_warning(
825 'Passing a urllib.request.Request to _create_request() is deprecated. '
826 'Use yt_dlp.networking.common.Request instead.')
827 url_or_request
= urllib_req_to_req(url_or_request
)
828 elif not isinstance(url_or_request
, Request
):
829 url_or_request
= Request(url_or_request
)
831 url_or_request
.update(data
=data
, headers
=headers
, query
=query
, extensions
=extensions
)
832 return url_or_request
834 def _request_webpage(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, data
=None,
835 headers
=None, query
=None, expected_status
=None, impersonate
=None, require_impersonation
=False):
837 Return the response handle.
839 See _download_webpage docstring for arguments specification.
841 if not self
._downloader
._first
_webpage
_request
:
842 sleep_interval
= self
.get_param('sleep_interval_requests') or 0
843 if sleep_interval
> 0:
844 self
.to_screen('Sleeping %s seconds ...' % sleep_interval
)
845 time
.sleep(sleep_interval
)
847 self
._downloader
._first
_webpage
_request
= False
850 self
.report_download_webpage(video_id
)
851 elif note
is not False:
853 self
.to_screen(str(note
))
855 self
.to_screen(f
'{video_id}: {note}')
857 # Some sites check X-Forwarded-For HTTP header in order to figure out
858 # the origin of the client behind proxy. This allows bypassing geo
859 # restriction by faking this header's value to IP that belongs to some
860 # geo unrestricted country. We will do so once we encounter any
861 # geo restriction error.
862 if self
._x
_forwarded
_for
_ip
:
863 headers
= (headers
or {}).copy()
864 headers
.setdefault('X-Forwarded-For', self
._x
_forwarded
_for
_ip
)
868 if impersonate
in (True, ''):
869 impersonate
= ImpersonateTarget()
870 requested_targets
= [
871 t
if isinstance(t
, ImpersonateTarget
) else ImpersonateTarget
.from_str(t
)
872 for t
in variadic(impersonate
)
873 ] if impersonate
else []
875 available_target
= next(filter(self
._downloader
._impersonate
_target
_available
, requested_targets
), None)
877 extensions
['impersonate'] = available_target
878 elif requested_targets
:
879 message
= 'The extractor is attempting impersonation, but '
881 'no impersonate target is available' if not str(impersonate
)
882 else f
'none of these impersonate targets are available: "{", ".join(map(str, requested_targets))}"')
883 info_msg
= ('see https://github.com/yt-dlp/yt-dlp#impersonation '
884 'for information on installing the required dependencies')
885 if require_impersonation
:
886 raise ExtractorError(f
'{message}; {info_msg}', expected
=True)
887 self
.report_warning(f
'{message}; if you encounter errors, then {info_msg}', only_once
=True)
890 return self
._downloader
.urlopen(self
._create
_request
(url_or_request
, data
, headers
, query
, extensions
))
891 except network_exceptions
as err
:
892 if isinstance(err
, HTTPError
):
893 if self
.__can
_accept
_status
_code
(err
, expected_status
):
899 errnote
= 'Unable to download webpage'
901 errmsg
= f
'{errnote}: {error_to_compat_str(err)}'
903 raise ExtractorError(errmsg
, cause
=err
)
905 self
.report_warning(errmsg
)
908 def _download_webpage_handle(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True,
909 encoding
=None, data
=None, headers
={}, query={}
, expected_status
=None,
910 impersonate
=None, require_impersonation
=False):
912 Return a tuple (page content as string, URL handle).
915 url_or_request -- plain text URL as a string or
916 a yt_dlp.networking.Request object
917 video_id -- Video/playlist/item identifier (string)
920 note -- note printed before downloading (string)
921 errnote -- note printed in case of an error (string)
922 fatal -- flag denoting whether error should be considered fatal,
923 i.e. whether it should cause ExtractionError to be raised,
924 otherwise a warning will be reported and extraction continued
925 encoding -- encoding for a page content decoding, guessed automatically
926 when not explicitly specified
927 data -- POST data (bytes)
928 headers -- HTTP headers (dict)
929 query -- URL query (dict)
930 expected_status -- allows to accept failed HTTP requests (non 2xx
931 status code) by explicitly specifying a set of accepted status
932 codes. Can be any of the following entities:
933 - an integer type specifying an exact failed status code to
935 - a list or a tuple of integer types specifying a list of
936 failed status codes to accept
937 - a callable accepting an actual failed status code and
938 returning True if it should be accepted
939 Note that this argument does not affect success status codes (2xx)
940 which are always accepted.
941 impersonate -- the impersonate target. Can be any of the following entities:
942 - an instance of yt_dlp.networking.impersonate.ImpersonateTarget
943 - a string in the format of CLIENT[:OS]
944 - a list or a tuple of CLIENT[:OS] strings or ImpersonateTarget instances
945 - a boolean value; True means any impersonate target is sufficient
946 require_impersonation -- flag to toggle whether the request should raise an error
947 if impersonation is not possible (bool, default: False)
950 # Strip hashes from the URL (#1038)
951 if isinstance(url_or_request
, str):
952 url_or_request
= url_or_request
.partition('#')[0]
954 urlh
= self
._request
_webpage
(url_or_request
, video_id
, note
, errnote
, fatal
, data
=data
,
955 headers
=headers
, query
=query
, expected_status
=expected_status
,
956 impersonate
=impersonate
, require_impersonation
=require_impersonation
)
960 content
= self
._webpage
_read
_content
(urlh
, url_or_request
, video_id
, note
, errnote
, fatal
, encoding
=encoding
)
961 return (content
, urlh
)
964 def _guess_encoding_from_content(content_type
, webpage_bytes
):
965 m
= re
.match(r
'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type
)
967 encoding
= m
.group(1)
969 m
= re
.search(br
'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
970 webpage_bytes[:1024])
972 encoding = m.group(1).decode('ascii')
973 elif webpage_bytes.startswith(b'\xff\xfe'):
980 def __check_blocked(self, content):
981 first_block = content[:512]
982 if ('<title>Access to this site is blocked</title>' in content
983 and 'Websense' in first_block):
984 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
985 blocked_iframe = self._html_search_regex(
986 r'<iframe src="([^
"]+)"', content,
987 'Websense information URL
', default=None)
989 msg += ' Visit
%s for more details
' % blocked_iframe
990 raise ExtractorError(msg, expected=True)
991 if '<title
>The URL you requested has been blocked
</title
>' in first_block:
993 'Access to this webpage has been blocked by Indian censorship
. '
994 'Use a VPN
or proxy
server (with --proxy
) to route around it
.')
995 block_msg = self._html_search_regex(
996 r'</h1
><p
>(.*?
)</p
>',
997 content, 'block message
', default=None)
999 msg += ' (Message
: "%s")' % block_msg.replace('\n', ' ')
1000 raise ExtractorError(msg, expected=True)
1001 if ('<title
>TTK
:: Доступ к ресурсу ограничен
</title
>' in content
1002 and 'blocklist
.rkn
.gov
.ru
' in content):
1003 raise ExtractorError(
1004 'Access to this webpage has been blocked by decision of the Russian government
. '
1005 'Visit http
://blocklist
.rkn
.gov
.ru
/ for a block reason
.',
1008 def _request_dump_filename(self, url, video_id):
1009 basen = f'{video_id}_{url}
'
1010 trim_length = self.get_param('trim_file_name
') or 240
1011 if len(basen) > trim_length:
1012 h = '___
' + hashlib.md5(basen.encode('utf
-8')).hexdigest()
1013 basen = basen[:trim_length - len(h)] + h
1014 filename = sanitize_filename(f'{basen}
.dump
', restricted=True)
1015 # Working around MAX_PATH limitation on Windows (see
1016 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
1017 if compat_os_name == 'nt
':
1018 absfilepath = os.path.abspath(filename)
1019 if len(absfilepath) > 259:
1020 filename = fR'\\?\{absfilepath}
'
1023 def __decode_webpage(self, webpage_bytes, encoding, headers):
1025 encoding = self._guess_encoding_from_content(headers.get('Content
-Type
', ''), webpage_bytes)
1027 return webpage_bytes.decode(encoding, 'replace
')
1029 return webpage_bytes.decode('utf
-8', 'replace
')
1031 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
1032 webpage_bytes = urlh.read()
1033 if prefix is not None:
1034 webpage_bytes = prefix + webpage_bytes
1035 if self.get_param('dump_intermediate_pages
', False):
1036 self.to_screen('Dumping request to
' + urlh.url)
1037 dump = base64.b64encode(webpage_bytes).decode('ascii
')
1038 self._downloader.to_screen(dump)
1039 if self.get_param('write_pages
'):
1040 filename = self._request_dump_filename(urlh.url, video_id)
1041 self.to_screen(f'Saving request to {filename}
')
1042 with open(filename, 'wb
') as outf:
1043 outf.write(webpage_bytes)
1045 content = self.__decode_webpage(webpage_bytes, encoding, urlh.headers)
1046 self.__check_blocked(content)
1050 def __print_error(self, errnote, fatal, video_id, err):
1052 raise ExtractorError(f'{video_id}
: {errnote}
', cause=err)
1054 self.report_warning(f'{video_id}
: {errnote}
: {err}
')
1056 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True, errnote=None):
1057 if transform_source:
1058 xml_string = transform_source(xml_string)
1060 return compat_etree_fromstring(xml_string.encode('utf
-8'))
1061 except xml.etree.ElementTree.ParseError as ve:
1062 self.__print_error('Failed to parse XML
' if errnote is None else errnote, fatal, video_id, ve)
1064 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True, errnote=None, **parser_kwargs):
1067 json_string, cls=LenientJSONDecoder, strict=False, transform_source=transform_source, **parser_kwargs)
1068 except ValueError as ve:
1069 self.__print_error('Failed to parse JSON
' if errnote is None else errnote, fatal, video_id, ve)
1071 def _parse_socket_response_as_json(self, data, *args, **kwargs):
1072 return self._parse_json(data[data.find('{'):data.rfind('}
') + 1], *args, **kwargs)
1074 def __create_download_methods(name, parser, note, errnote, return_value):
1076 def parse(ie, content, *args, errnote=errnote, **kwargs):
1079 if errnote is False:
1080 kwargs['errnote
'] = errnote
1081 # parser is fetched by name so subclasses can override it
1082 return getattr(ie, parser)(content, *args, **kwargs)
1084 def download_handle(self, url_or_request, video_id, note=note, errnote=errnote, transform_source=None,
1085 fatal=True, encoding=None, data=None, headers={}, query={}, expected_status=None,
1086 impersonate=None, require_impersonation=False):
1087 res = self._download_webpage_handle(
1088 url_or_request, video_id, note=note, errnote=errnote, fatal=fatal, encoding=encoding,
1089 data=data, headers=headers, query=query, expected_status=expected_status,
1090 impersonate=impersonate, require_impersonation=require_impersonation)
1094 return parse(self, content, video_id, transform_source=transform_source, fatal=fatal, errnote=errnote), urlh
1096 def download_content(self, url_or_request, video_id, note=note, errnote=errnote, transform_source=None,
1097 fatal=True, encoding=None, data=None, headers={}, query={}, expected_status=None,
1098 impersonate=None, require_impersonation=False):
1099 if self.get_param('load_pages
'):
1100 url_or_request = self._create_request(url_or_request, data, headers, query)
1101 filename = self._request_dump_filename(url_or_request.url, video_id)
1102 self.to_screen(f'Loading request
from {filename}
')
1104 with open(filename, 'rb
') as dumpf:
1105 webpage_bytes = dumpf.read()
1106 except OSError as e:
1107 self.report_warning(f'Unable to load request
from disk
: {e}
')
1109 content = self.__decode_webpage(webpage_bytes, encoding, url_or_request.headers)
1110 return parse(self, content, video_id, transform_source=transform_source, fatal=fatal, errnote=errnote)
1114 'transform_source
': transform_source,
1116 'encoding
': encoding,
1120 'expected_status
': expected_status,
1121 'impersonate
': impersonate,
1122 'require_impersonation
': require_impersonation,
1125 kwargs.pop('transform_source
')
1126 # The method is fetched by name so subclasses can override _download_..._handle
1127 res = getattr(self, download_handle.__name__)(url_or_request, video_id, **kwargs)
1128 return res if res is False else res[0]
1130 def impersonate(func, name, return_value):
1131 func.__name__, func.__qualname__ = name, f'InfoExtractor
.{name}
'
1133 @param transform_source Apply this transformation before parsing
1134 @returns {return_value}
1136 See _download_webpage_handle docstring for other arguments specification
1139 impersonate(download_handle, f'_download_{name}_handle
', f'({return_value}
, URL handle
)')
1140 impersonate(download_content, f'_download_{name}
', f'{return_value}
')
1141 return download_handle, download_content
1143 _download_xml_handle, _download_xml = __create_download_methods(
1144 'xml
', '_parse_xml
', 'Downloading XML
', 'Unable to download XML
', 'xml
as an xml
.etree
.ElementTree
.Element
')
1145 _download_json_handle, _download_json = __create_download_methods(
1146 'json
', '_parse_json
', 'Downloading JSON metadata
', 'Unable to download JSON metadata
', 'JSON
object as a
dict')
1147 _download_socket_json_handle, _download_socket_json = __create_download_methods(
1148 'socket_json
', '_parse_socket_response_as_json
', 'Polling socket
', 'Unable to poll socket
', 'JSON
object as a
dict')
1149 __download_webpage = __create_download_methods('webpage
', None, None, None, 'data of the page
as a string
')[1]
1151 def _download_webpage(
1152 self, url_or_request, video_id, note=None, errnote=None,
1153 fatal=True, tries=1, timeout=NO_DEFAULT, *args, **kwargs):
1155 Return the data of the page as a string.
1158 tries -- number of tries
1159 timeout -- sleep interval between tries
1161 See _download_webpage_handle docstring for other arguments specification.
1164 R''' # NB: These are unused; should they be deprecated?
1166 self._downloader.deprecation_warning('tries argument
is deprecated
in InfoExtractor
._download
_webpage
')
1167 if timeout is NO_DEFAULT:
1170 self._downloader.deprecation_warning('timeout argument
is deprecated
in InfoExtractor
._download
_webpage
')
1176 return self.__download_webpage(url_or_request, video_id, note, errnote, None, fatal, *args, **kwargs)
1177 except IncompleteRead as e:
1179 if try_count >= tries:
1181 self._sleep(timeout, video_id)
1183 def report_warning(self, msg, video_id=None, *args, only_once=False, **kwargs):
1184 idstr = format_field(video_id, None, '%s: ')
1185 msg = f'[{self.IE_NAME}
] {idstr}{msg}
'
1187 if f'WARNING
: {msg}
' in self._printed_messages:
1189 self._printed_messages.add(f'WARNING
: {msg}
')
1190 self._downloader.report_warning(msg, *args, **kwargs)
1192 def to_screen(self, msg, *args, **kwargs):
1193 """Print msg to screen, prefixing it with '[ie_name
]'"""
1194 self._downloader.to_screen(f'[{self.IE_NAME}
] {msg}
', *args, **kwargs)
1196 def write_debug(self, msg, *args, **kwargs):
1197 self._downloader.write_debug(f'[{self.IE_NAME}
] {msg}
', *args, **kwargs)
1199 def get_param(self, name, default=None, *args, **kwargs):
1200 if self._downloader:
1201 return self._downloader.params.get(name, default, *args, **kwargs)
1204 def report_drm(self, video_id, partial=NO_DEFAULT):
1205 if partial is not NO_DEFAULT:
1206 self._downloader.deprecation_warning('InfoExtractor
.report_drm no longer accepts the argument partial
')
1207 self.raise_no_formats('This video
is DRM protected
', expected=True, video_id=video_id)
1209 def report_extraction(self, id_or_name):
1210 """Report information extraction."""
1211 self.to_screen('%s: Extracting information
' % id_or_name)
1213 def report_download_webpage(self, video_id):
1214 """Report webpage download."""
1215 self.to_screen('%s: Downloading webpage
' % video_id)
1217 def report_age_confirmation(self):
1218 """Report attempt to confirm age."""
1219 self.to_screen('Confirming age
')
1221 def report_login(self):
1222 """Report attempt to log in."""
1223 self.to_screen('Logging
in')
1225 def raise_login_required(
1226 self, msg='This video
is only available
for registered users
',
1227 metadata_available=False, method=NO_DEFAULT):
1228 if metadata_available and (
1229 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1230 self.report_warning(msg)
1232 msg += format_field(self._login_hint(method), None, '. %s')
1233 raise ExtractorError(msg, expected=True)
1235 def raise_geo_restricted(
1236 self, msg='This video
is not available
from your location due to geo restriction
',
1237 countries=None, metadata_available=False):
1238 if metadata_available and (
1239 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1240 self.report_warning(msg)
1242 raise GeoRestrictedError(msg, countries=countries)
1244 def raise_no_formats(self, msg, expected=False, video_id=None):
1246 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1247 self.report_warning(msg, video_id)
1248 elif isinstance(msg, ExtractorError):
1251 raise ExtractorError(msg, expected=expected, video_id=video_id)
1253 # Methods for following #608
1255 def url_result(url, ie=None, video_id=None, video_title=None, *, url_transparent=False, **kwargs):
1256 """Returns a URL that points to a page that should be processed"""
1258 kwargs['ie_key
'] = ie if isinstance(ie, str) else ie.ie_key()
1259 if video_id is not None:
1260 kwargs['id'] = video_id
1261 if video_title is not None:
1262 kwargs['title
'] = video_title
1265 '_type
': 'url_transparent
' if url_transparent else 'url
',
1270 def playlist_from_matches(cls, matches, playlist_id=None, playlist_title=None,
1271 getter=IDENTITY, ie=None, video_kwargs=None, **kwargs):
1272 return cls.playlist_result(
1273 (cls.url_result(m, ie, **(video_kwargs or {})) for m in orderedSet(map(getter, matches), lazy=True)),
1274 playlist_id, playlist_title, **kwargs)
1277 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None, *, multi_video=False, **kwargs):
1278 """Returns a playlist"""
1280 kwargs['id'] = playlist_id
1282 kwargs['title
'] = playlist_title
1283 if playlist_description is not None:
1284 kwargs['description
'] = playlist_description
1287 '_type
': 'multi_video
' if multi_video else 'playlist
',
1291 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1293 Perform a regex search on the given string, using a single or a list of
1294 patterns returning the first matching group.
1295 In case of failure return a default value or raise a WARNING or a
1296 RegexNotFoundError, depending on fatal, specifying the field name.
1300 elif isinstance(pattern, (str, re.Pattern)):
1301 mobj = re.search(pattern, string, flags)
1304 mobj = re.search(p, string, flags)
1308 _name = self._downloader._format_err(name, self._downloader.Styles.EMPHASIS)
1312 # return the first matching group
1313 return next(g for g in mobj.groups() if g is not None)
1314 elif isinstance(group, (list, tuple)):
1315 return tuple(mobj.group(g) for g in group)
1317 return mobj.group(group)
1318 elif default is not NO_DEFAULT:
1321 raise RegexNotFoundError('Unable to extract
%s' % _name)
1323 self.report_warning('unable to extract
%s' % _name + bug_reports_message())
1326 def _search_json(self, start_pattern, string, name, video_id, *, end_pattern='',
1327 contains_pattern=r'{(?s:.+)}
', fatal=True, default=NO_DEFAULT, **kwargs):
1328 """Searches string for the JSON object specified by start_pattern"""
1329 # NB: end_pattern is only used to reduce the size of the initial match
1330 if default is NO_DEFAULT:
1331 default, has_default = {}, False
1333 fatal, has_default = False, True
1335 json_string = self._search_regex(
1336 rf'(?
:{start_pattern}
)\s
*(?P
<json
>{contains_pattern}
)\s
*(?
:{end_pattern}
)',
1337 string, name, group='json
', fatal=fatal, default=None if has_default else NO_DEFAULT)
1341 _name = self._downloader._format_err(name, self._downloader.Styles.EMPHASIS)
1343 return self._parse_json(json_string, video_id, ignore_extra=True, **kwargs)
1344 except ExtractorError as e:
1346 raise ExtractorError(
1347 f'Unable to extract {_name}
- Failed to parse JSON
', cause=e.cause, video_id=video_id)
1348 elif not has_default:
1349 self.report_warning(
1350 f'Unable to extract {_name}
- Failed to parse JSON
: {e}
', video_id=video_id)
1353 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1355 Like _search_regex, but strips HTML tags and unescapes entities.
1357 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1358 if isinstance(res, tuple):
1359 return tuple(map(clean_html, res))
1360 return clean_html(res)
1362 def _get_netrc_login_info(self, netrc_machine=None):
1363 netrc_machine = netrc_machine or self._NETRC_MACHINE
1365 cmd = self.get_param('netrc_cmd
')
1367 cmd = cmd.replace('{}', netrc_machine)
1368 self.to_screen(f'Executing command
: {cmd}
')
1369 stdout, _, ret = Popen.run(cmd, text=True, shell=True, stdout=subprocess.PIPE)
1371 raise OSError(f'Command returned error code {ret}
')
1372 info = netrc_from_content(stdout).authenticators(netrc_machine)
1374 elif self.get_param('usenetrc
', False):
1375 netrc_file = compat_expanduser(self.get_param('netrc_location
') or '~
')
1376 if os.path.isdir(netrc_file):
1377 netrc_file = os.path.join(netrc_file, '.netrc
')
1378 info = netrc.netrc(netrc_file).authenticators(netrc_machine)
1383 self.to_screen(f'No authenticators
for {netrc_machine}
')
1386 self.write_debug(f'Using netrc
for {netrc_machine} authentication
')
1387 return info[0], info[2]
1389 def _get_login_info(self, username_option='username
', password_option='password
', netrc_machine=None):
1391 Get the login info as (username, password)
1392 First look for the manually specified credentials using username_option
1393 and password_option as keys in params dictionary. If no such credentials
1394 are available try the netrc_cmd if it is defined or look in the
1395 netrc file using the netrc_machine or _NETRC_MACHINE value.
1396 If there's no info available
, return (None, None)
1399 username = self.get_param(username_option)
1400 if username is not None:
1401 password = self.get_param(password_option)
1404 username, password = self._get_netrc_login_info(netrc_machine)
1405 except (OSError, netrc.NetrcParseError) as err:
1406 self.report_warning(f'Failed to parse .netrc: {err}')
1408 return username, password
1410 def _get_tfa_info(self, note='two-factor verification code'):
1412 Get the two
-factor authentication info
1413 TODO
- asking the user will be required
for sms
/phone verify
1414 currently just uses the command line option
1415 If there
's no info available, return None
1418 tfa = self.get_param('twofactor
')
1422 return getpass.getpass('Type
%s and press
[Return
]: ' % note)
1424 # Helper functions for extracting OpenGraph info
1426 def _og_regexes(prop):
1427 content_re = r'content
=(?
:"([^"]+?
)"|\'([^\']+?)\'|\s*([^\s"\'=<>`
]+?
)(?
=\s|
/?
>))'
1428 property_re = (r'(?
:name|
property)=(?
:\'og
%(sep)s%(prop)s\'|
"og%(sep)s%(prop)s"|\s
*og
%(sep)s%(prop)s\b)'
1429 % {'prop': re.escape(prop), 'sep': '(?::|[:-])'})
1430 template = r'<meta
[^
>]+?
%s[^
>]+?
%s'
1432 template % (property_re, content_re),
1433 template % (content_re, property_re),
1437 def _meta_regex(prop):
1438 return r'''(?isx)<meta
1439 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1440 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1442 def _og_search_property(self, prop, html, name=None, **kargs):
1443 prop = variadic(prop)
1445 name = 'OpenGraph
%s' % prop[0]
1448 og_regexes.extend(self._og_regexes(p))
1449 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1452 return unescapeHTML(escaped)
1454 def _og_search_thumbnail(self, html, **kargs):
1455 return self._og_search_property('image
', html, 'thumbnail URL
', fatal=False, **kargs)
1457 def _og_search_description(self, html, **kargs):
1458 return self._og_search_property('description
', html, fatal=False, **kargs)
1460 def _og_search_title(self, html, *, fatal=False, **kargs):
1461 return self._og_search_property('title
', html, fatal=fatal, **kargs)
1463 def _og_search_video_url(self, html, name='video url
', secure=True, **kargs):
1464 regexes = self._og_regexes('video
') + self._og_regexes('video
:url
')
1466 regexes = self._og_regexes('video
:secure_url
') + regexes
1467 return self._html_search_regex(regexes, html, name, **kargs)
1469 def _og_search_url(self, html, **kargs):
1470 return self._og_search_property('url
', html, **kargs)
1472 def _html_extract_title(self, html, name='title
', *, fatal=False, **kwargs):
1473 return self._html_search_regex(r'(?s
)<title
\b[^
>]*>([^
<]+)</title
>', html, name, fatal=fatal, **kwargs)
1475 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1476 name = variadic(name)
1477 if display_name is None:
1478 display_name = name[0]
1479 return self._html_search_regex(
1480 [self._meta_regex(n) for n in name],
1481 html, display_name, fatal=fatal, group='content
', **kwargs)
1483 def _dc_search_uploader(self, html):
1484 return self._html_search_meta('dc
.creator
', html, 'uploader
')
1487 def _rta_search(html):
1488 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1489 if re.search(r'(?ix
)<meta\s
+name
="rating"\s
+'
1490 r' content
="RTA-5042-1996-1400-1577-RTA"',
1494 # And then there are the jokers who advertise that they use RTA, but actually don't
.
1495 AGE_LIMIT_MARKERS
= [
1496 r
'Proudly Labeled <a href="http://www\.rtalabel\.org/" title="Restricted to Adults">RTA</a>',
1497 r
'>[^<]*you acknowledge you are at least (\d+) years old',
1498 r
'>\s*(?:18\s+U(?:\.S\.C\.|SC)\s+)?(?:§+\s*)?2257\b',
1502 for marker
in AGE_LIMIT_MARKERS
:
1503 mobj
= re
.search(marker
, html
)
1505 age_limit
= max(age_limit
, int(traverse_obj(mobj
, 1, default
=18)))
1508 def _media_rating_search(self
, html
):
1509 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1510 rating
= self
._html
_search
_meta
('rating', html
)
1522 return RATING_TABLE
.get(rating
.lower())
1524 def _family_friendly_search(self
, html
):
1525 # See http://schema.org/VideoObject
1526 family_friendly
= self
._html
_search
_meta
(
1527 'isFamilyFriendly', html
, default
=None)
1529 if not family_friendly
:
1538 return RATING_TABLE
.get(family_friendly
.lower())
1540 def _twitter_search_player(self
, html
):
1541 return self
._html
_search
_meta
('twitter:player', html
,
1542 'twitter card player')
1544 def _yield_json_ld(self
, html
, video_id
, *, fatal
=True, default
=NO_DEFAULT
):
1545 """Yield all json ld objects in the html"""
1546 if default
is not NO_DEFAULT
:
1548 for mobj
in re
.finditer(JSON_LD_RE
, html
):
1549 json_ld_item
= self
._parse
_json
(mobj
.group('json_ld'), video_id
, fatal
=fatal
)
1550 for json_ld
in variadic(json_ld_item
):
1551 if isinstance(json_ld
, dict):
1554 def _search_json_ld(self
, html
, video_id
, expected_type
=None, *, fatal
=True, default
=NO_DEFAULT
):
1555 """Search for a video in any json ld in the html"""
1556 if default
is not NO_DEFAULT
:
1558 info
= self
._json
_ld
(
1559 list(self
._yield
_json
_ld
(html
, video_id
, fatal
=fatal
, default
=default
)),
1560 video_id
, fatal
=fatal
, expected_type
=expected_type
)
1563 if default
is not NO_DEFAULT
:
1566 raise RegexNotFoundError('Unable to extract JSON-LD')
1568 self
.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1571 def _json_ld(self
, json_ld
, video_id
, fatal
=True, expected_type
=None):
1572 if isinstance(json_ld
, str):
1573 json_ld
= self
._parse
_json
(json_ld
, video_id
, fatal
=fatal
)
1578 INTERACTION_TYPE_MAP
= {
1579 'CommentAction': 'comment',
1580 'AgreeAction': 'like',
1581 'DisagreeAction': 'dislike',
1582 'LikeAction': 'like',
1583 'DislikeAction': 'dislike',
1584 'ListenAction': 'view',
1585 'WatchAction': 'view',
1586 'ViewAction': 'view',
1589 def is_type(e
, *expected_types
):
1590 type = variadic(traverse_obj(e
, '@type'))
1591 return any(x
in type for x
in expected_types
)
1593 def extract_interaction_type(e
):
1594 interaction_type
= e
.get('interactionType')
1595 if isinstance(interaction_type
, dict):
1596 interaction_type
= interaction_type
.get('@type')
1597 return str_or_none(interaction_type
)
1599 def extract_interaction_statistic(e
):
1600 interaction_statistic
= e
.get('interactionStatistic')
1601 if isinstance(interaction_statistic
, dict):
1602 interaction_statistic
= [interaction_statistic
]
1603 if not isinstance(interaction_statistic
, list):
1605 for is_e
in interaction_statistic
:
1606 if not is_type(is_e
, 'InteractionCounter'):
1608 interaction_type
= extract_interaction_type(is_e
)
1609 if not interaction_type
:
1611 # For interaction count some sites provide string instead of
1612 # an integer (as per spec) with non digit characters (e.g. ",")
1613 # so extracting count with more relaxed str_to_int
1614 interaction_count
= str_to_int(is_e
.get('userInteractionCount'))
1615 if interaction_count
is None:
1617 count_kind
= INTERACTION_TYPE_MAP
.get(interaction_type
.split('/')[-1])
1620 count_key
= '%s_count' % count_kind
1621 if info
.get(count_key
) is not None:
1623 info
[count_key
] = interaction_count
1625 def extract_chapter_information(e
):
1627 'title': part
.get('name'),
1628 'start_time': part
.get('startOffset'),
1629 'end_time': part
.get('endOffset'),
1630 } for part
in variadic(e
.get('hasPart') or []) if part
.get('@type') == 'Clip']
1631 for idx
, (last_c
, current_c
, next_c
) in enumerate(zip(
1632 [{'end_time': 0}
] + chapters
, chapters
, chapters
[1:])):
1633 current_c
['end_time'] = current_c
['end_time'] or next_c
['start_time']
1634 current_c
['start_time'] = current_c
['start_time'] or last_c
['end_time']
1635 if None in current_c
.values():
1636 self
.report_warning(f
'Chapter {idx} contains broken data. Not extracting chapters')
1639 chapters
[-1]['end_time'] = chapters
[-1]['end_time'] or info
['duration']
1640 info
['chapters'] = chapters
1642 def extract_video_object(e
):
1643 author
= e
.get('author')
1645 'url': url_or_none(e
.get('contentUrl')),
1646 'ext': mimetype2ext(e
.get('encodingFormat')),
1647 'title': unescapeHTML(e
.get('name')),
1648 'description': unescapeHTML(e
.get('description')),
1649 'thumbnails': [{'url': unescapeHTML(url)}
1650 for url
in variadic(traverse_obj(e
, 'thumbnailUrl', 'thumbnailURL'))
1651 if url_or_none(url
)],
1652 'duration': parse_duration(e
.get('duration')),
1653 'timestamp': unified_timestamp(e
.get('uploadDate')),
1654 # author can be an instance of 'Organization' or 'Person' types.
1655 # both types can have 'name' property(inherited from 'Thing' type). [1]
1656 # however some websites are using 'Text' type instead.
1657 # 1. https://schema.org/VideoObject
1658 'uploader': author
.get('name') if isinstance(author
, dict) else author
if isinstance(author
, str) else None,
1659 'artist': traverse_obj(e
, ('byArtist', 'name'), expected_type
=str),
1660 'filesize': int_or_none(float_or_none(e
.get('contentSize'))),
1661 'tbr': int_or_none(e
.get('bitrate')),
1662 'width': int_or_none(e
.get('width')),
1663 'height': int_or_none(e
.get('height')),
1664 'view_count': int_or_none(e
.get('interactionCount')),
1665 'tags': try_call(lambda: e
.get('keywords').split(',')),
1667 if is_type(e
, 'AudioObject'):
1670 'abr': int_or_none(e
.get('bitrate')),
1672 extract_interaction_statistic(e
)
1673 extract_chapter_information(e
)
1675 def traverse_json_ld(json_ld
, at_top_level
=True):
1676 for e
in variadic(json_ld
):
1677 if not isinstance(e
, dict):
1679 if at_top_level
and '@context' not in e
:
1681 if at_top_level
and set(e
.keys()) == {'@context', '@graph'}
:
1682 traverse_json_ld(e
['@graph'], at_top_level
=False)
1684 if expected_type
is not None and not is_type(e
, expected_type
):
1686 rating
= traverse_obj(e
, ('aggregateRating', 'ratingValue'), expected_type
=float_or_none
)
1687 if rating
is not None:
1688 info
['average_rating'] = rating
1689 if is_type(e
, 'TVEpisode', 'Episode'):
1690 episode_name
= unescapeHTML(e
.get('name'))
1692 'episode': episode_name
,
1693 'episode_number': int_or_none(e
.get('episodeNumber')),
1694 'description': unescapeHTML(e
.get('description')),
1696 if not info
.get('title') and episode_name
:
1697 info
['title'] = episode_name
1698 part_of_season
= e
.get('partOfSeason')
1699 if is_type(part_of_season
, 'TVSeason', 'Season', 'CreativeWorkSeason'):
1701 'season': unescapeHTML(part_of_season
.get('name')),
1702 'season_number': int_or_none(part_of_season
.get('seasonNumber')),
1704 part_of_series
= e
.get('partOfSeries') or e
.get('partOfTVSeries')
1705 if is_type(part_of_series
, 'TVSeries', 'Series', 'CreativeWorkSeries'):
1706 info
['series'] = unescapeHTML(part_of_series
.get('name'))
1707 elif is_type(e
, 'Movie'):
1709 'title': unescapeHTML(e
.get('name')),
1710 'description': unescapeHTML(e
.get('description')),
1711 'duration': parse_duration(e
.get('duration')),
1712 'timestamp': unified_timestamp(e
.get('dateCreated')),
1714 elif is_type(e
, 'Article', 'NewsArticle'):
1716 'timestamp': parse_iso8601(e
.get('datePublished')),
1717 'title': unescapeHTML(e
.get('headline')),
1718 'description': unescapeHTML(e
.get('articleBody') or e
.get('description')),
1720 if is_type(traverse_obj(e
, ('video', 0)), 'VideoObject'):
1721 extract_video_object(e
['video'][0])
1722 elif is_type(traverse_obj(e
, ('subjectOf', 0)), 'VideoObject'):
1723 extract_video_object(e
['subjectOf'][0])
1724 elif is_type(e
, 'VideoObject', 'AudioObject'):
1725 extract_video_object(e
)
1726 if expected_type
is None:
1730 video
= e
.get('video')
1731 if is_type(video
, 'VideoObject'):
1732 extract_video_object(video
)
1733 if expected_type
is None:
1738 traverse_json_ld(json_ld
)
1739 return filter_dict(info
)
1741 def _search_nextjs_data(self
, webpage
, video_id
, *, fatal
=True, default
=NO_DEFAULT
, **kw
):
1743 self
._downloader
.deprecation_warning('using `default=\'{}\'` is deprecated, use `default={}` instead')
1745 if default
is not NO_DEFAULT
:
1748 return self
._search
_json
(
1749 r
'<script[^>]+id=[\'"]__NEXT_DATA__[\'"][^
>]*>', webpage, 'next
.js data
',
1750 video_id, end_pattern='</script
>', fatal=fatal, default=default, **kw)
1752 def _search_nuxt_data(self, webpage, video_id, context_name='__NUXT__
', *, fatal=True, traverse=('data
', 0)):
1753 """Parses Nuxt.js metadata. This works as long as the function __NUXT__ invokes is a pure function"""
1754 rectx = re.escape(context_name)
1755 FUNCTION_RE = r'\
(function\
((?P
<arg_keys
>.*?
)\
){.*?\breturn\s+(?P<js>{.*?}
)\s
*;?\s
*}\
((?P
<arg_vals
>.*?
)\
)'
1756 js, arg_keys, arg_vals = self._search_regex(
1757 (rf'<script
>\s
*window\
.{rectx}
={FUNCTION_RE}\s
*\
)\s
*;?\s
*</script
>', rf'{rectx}\
(.*?{FUNCTION_RE}
'),
1758 webpage, context_name, group=('js
', 'arg_keys
', 'arg_vals
'),
1759 default=NO_DEFAULT if fatal else (None, None, None))
1763 args = dict(zip(arg_keys.split(','), map(json.dumps, self._parse_json(
1764 f'[{arg_vals}
]', video_id, transform_source=js_to_json, fatal=fatal) or ())))
1766 ret = self._parse_json(js, video_id, transform_source=functools.partial(js_to_json, vars=args), fatal=fatal)
1767 return traverse_obj(ret, traverse) or {}
1770 def _hidden_inputs(html):
1771 html = re.sub(r'<!--(?
:(?
!<!--).)*-->', '', html)
1773 for input in re.findall(r'(?i
)(<input[^
>]+>)', html):
1774 attrs = extract_attributes(input)
1777 if attrs.get('type') not in ('hidden
', 'submit
'):
1779 name = attrs.get('name
') or attrs.get('id')
1780 value = attrs.get('value
')
1781 if name and value is not None:
1782 hidden_inputs[name] = value
1783 return hidden_inputs
1785 def _form_hidden_inputs(self, form_id, html):
1786 form = self._search_regex(
1787 r'(?
is)<form
[^
>]+?
id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1788 html, '%s form' % form_id, group='form')
1789 return self._hidden_inputs(form)
1791 @classproperty(cache=True)
1792 def FormatSort(cls):
1793 class FormatSort(FormatSorter):
1794 def __init__(ie, *args, **kwargs):
1795 super().__init__(ie._downloader, *args, **kwargs)
1797 deprecation_warning(
1798 'yt_dlp.InfoExtractor.FormatSort is deprecated and may be removed in the future. '
1799 'Use yt_dlp.utils.FormatSorter instead')
1802 def _sort_formats(self, formats, field_preference=[]):
1803 if not field_preference:
1804 self._downloader.deprecation_warning(
1805 'yt_dlp.InfoExtractor._sort_formats is deprecated and is no longer required')
1807 self._downloader.deprecation_warning(
1808 'yt_dlp.InfoExtractor._sort_formats is deprecated and no longer works as expected. '
1809 'Return _format_sort_fields in the info_dict instead')
1811 formats[0]['__sort_fields'] = field_preference
1813 def _check_formats(self, formats, video_id):
1815 formats[:] = filter(
1816 lambda f: self._is_valid_url(
1818 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1822 def _remove_duplicate_formats(formats):
1826 if f['url'] not in format_urls:
1827 format_urls.add(f['url'])
1828 unique_formats.append(f)
1829 formats[:] = unique_formats
1831 def _is_valid_url(self, url, video_id, item='video', headers={}):
1832 url = self._proto_relative_url(url, scheme='http:')
1833 # For now assume non HTTP(S) URLs always valid
1834 if not (url.startswith('http://') or url.startswith('https://')):
1837 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1839 except ExtractorError as e:
1841 '%s: %s URL is invalid, skipping: %s'
1842 % (video_id, item, error_to_compat_str(e.cause)))
1845 def http_scheme(self):
1846 """ Either "http
:" or "https
:", depending on the user's preferences """
1849 if self.get_param('prefer_insecure', False)
1852 def _proto_relative_url(self, url, scheme=None):
1853 scheme = scheme or self.http_scheme()
1854 assert scheme.endswith(':')
1855 return sanitize_url(url, scheme=scheme[:-1])
1857 def _sleep(self, timeout, video_id, msg_template=None):
1858 if msg_template is None:
1859 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1860 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1864 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1865 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1866 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
1867 if self.get_param('ignore_no_formats_error'):
1870 res = self._download_xml_handle(
1871 manifest_url, video_id, 'Downloading f4m manifest',
1872 'Unable to download f4m manifest',
1873 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1874 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
1875 transform_source=transform_source,
1876 fatal=fatal, data=data, headers=headers, query=query)
1880 manifest, urlh = res
1881 manifest_url = urlh.url
1883 return self._parse_f4m_formats(
1884 manifest, manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1885 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1887 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1888 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1889 fatal=True, m3u8_id=None):
1890 if not isinstance(manifest, xml.etree.ElementTree.Element) and not fatal:
1893 # currently yt-dlp cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1894 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1895 if akamai_pv is not None and ';' in akamai_pv.text:
1896 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1897 if playerVerificationChallenge.strip() != '':
1901 manifest_version = '1.0'
1902 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1904 manifest_version = '2.0'
1905 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1906 # Remove unsupported DRM protected media from final formats
1907 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
1908 media_nodes = remove_encrypted_media(media_nodes)
1912 manifest_base_url = get_base_url(manifest)
1914 bootstrap_info = xpath_element(
1915 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1916 'bootstrap info', default=None)
1919 mime_type = xpath_text(
1920 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1921 'base URL', default=None)
1922 if mime_type and mime_type.startswith('audio/'):
1925 for i, media_el in enumerate(media_nodes):
1926 tbr = int_or_none(media_el.attrib.get('bitrate'))
1927 width = int_or_none(media_el.attrib.get('width'))
1928 height = int_or_none(media_el.attrib.get('height'))
1929 format_id = join_nonempty(f4m_id, tbr or i)
1930 # If <bootstrapInfo> is present, the specified f4m is a
1931 # stream-level manifest, and only set-level manifests may refer to
1932 # external resources. See section 11.4 and section 4 of F4M spec
1933 if bootstrap_info is None:
1935 # @href is introduced in 2.0, see section 11.6 of F4M spec
1936 if manifest_version == '2.0':
1937 media_url = media_el.attrib.get('href')
1938 if media_url is None:
1939 media_url = media_el.attrib.get('url')
1943 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1944 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1945 # If media_url is itself a f4m manifest do the recursive extraction
1946 # since bitrates in parent manifest (this one) and media_url manifest
1947 # may differ leading to inability to resolve the format by requested
1948 # bitrate in f4m downloader
1949 ext = determine_ext(manifest_url)
1951 f4m_formats = self._extract_f4m_formats(
1952 manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1953 transform_source=transform_source, fatal=fatal)
1954 # Sometimes stream-level manifest contains single media entry that
1955 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1956 # At the same time parent's media entry in set-level manifest may
1957 # contain it. We will copy it from parent in such cases.
1958 if len(f4m_formats) == 1:
1961 'tbr': f.get('tbr') or tbr,
1962 'width': f.get('width') or width,
1963 'height': f.get('height') or height,
1964 'format_id': f.get('format_id') if not tbr else format_id,
1967 formats.extend(f4m_formats)
1970 formats.extend(self._extract_m3u8_formats(
1971 manifest_url, video_id, 'mp4', preference=preference,
1972 quality=quality, m3u8_id=m3u8_id, fatal=fatal))
1975 'format_id': format_id,
1976 'url': manifest_url,
1977 'manifest_url': manifest_url,
1978 'ext': 'flv' if bootstrap_info is not None else None,
1984 'preference': preference,
1989 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, quality=None, m3u8_id=None):
1991 'format_id': join_nonempty(m3u8_id, 'meta'),
1995 'preference': preference - 100 if preference else -100,
1997 'resolution': 'multiple',
1998 'format_note': 'Quality selection URL',
2001 def _report_ignoring_subs(self, name):
2002 self.report_warning(bug_reports_message(
2003 f'Ignoring subtitle tracks found in the {name} manifest; '
2004 'if any subtitle tracks are missing,'
2007 def _extract_m3u8_formats(self, *args, **kwargs):
2008 fmts, subs = self._extract_m3u8_formats_and_subtitles(*args, **kwargs)
2010 self._report_ignoring_subs('HLS')
2013 def _extract_m3u8_formats_and_subtitles(
2014 self, m3u8_url, video_id, ext=None, entry_protocol='m3u8_native',
2015 preference=None, quality=None, m3u8_id=None, note=None,
2016 errnote=None, fatal=True, live=False, data=None, headers={},
2019 if self.get_param('ignore_no_formats_error'):
2023 if errnote is not False:
2024 errnote = errnote or 'Failed to obtain m3u8 URL'
2026 raise ExtractorError(errnote, video_id=video_id)
2027 self.report_warning(f'{errnote}{bug_reports_message()}')
2030 res = self._download_webpage_handle(
2032 note='Downloading m3u8 information' if note is None else note,
2033 errnote='Failed to download m3u8 information' if errnote is None else errnote,
2034 fatal=fatal, data=data, headers=headers, query=query)
2039 m3u8_doc, urlh = res
2042 return self._parse_m3u8_formats_and_subtitles(
2043 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
2044 preference=preference, quality=quality, m3u8_id=m3u8_id,
2045 note=note, errnote=errnote, fatal=fatal, live=live, data=data,
2046 headers=headers, query=query, video_id=video_id)
2048 def _parse_m3u8_formats_and_subtitles(
2049 self, m3u8_doc, m3u8_url=None, ext=None, entry_protocol='m3u8_native',
2050 preference=None, quality=None, m3u8_id=None, live=False, note=None,
2051 errnote=None, fatal=True, data=None, headers={}, query={},
2053 formats, subtitles = [], {}
2054 has_drm = HlsFD._has_drm(m3u8_doc)
2056 def format_url(url):
2057 return url if re.match(r'^https?://', url) else urllib.parse.urljoin(m3u8_url, url)
2059 if self.get_param('hls_split_discontinuity', False):
2060 def _extract_m3u8_playlist_indices(manifest_url=None, m3u8_doc=None):
2062 if not manifest_url:
2064 m3u8_doc = self._download_webpage(
2065 manifest_url, video_id, fatal=fatal, data=data, headers=headers,
2066 note=False, errnote='Failed to download m3u8 playlist information')
2067 if m3u8_doc is False:
2069 return range(1 + sum(line.startswith('#EXT-X-DISCONTINUITY') for line in m3u8_doc.splitlines()))
2072 def _extract_m3u8_playlist_indices(*args, **kwargs):
2076 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
2077 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
2078 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
2080 # We should try extracting formats only from master playlists [1, 4.3.4],
2081 # i.e. playlists that describe available qualities. On the other hand
2082 # media playlists [1, 4.3.3] should be returned as is since they contain
2083 # just the media without qualities renditions.
2084 # Fortunately, master playlist can be easily distinguished from media
2085 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
2086 # master playlist tags MUST NOT appear in a media playlist and vice versa.
2087 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
2088 # media playlist and MUST NOT appear in master playlist thus we can
2089 # clearly detect media playlist with this criterion.
2091 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
2093 'format_id': join_nonempty(m3u8_id, idx),
2094 'format_index': idx,
2095 'url': m3u8_url or encode_data_uri(m3u8_doc.encode('utf-8'), 'application/x-mpegurl'),
2097 'protocol': entry_protocol,
2098 'preference': preference,
2101 } for idx in _extract_m3u8_playlist_indices(m3u8_doc=m3u8_doc)]
2103 return formats, subtitles
2106 last_stream_inf = {}
2108 def extract_media(x_media_line):
2109 media = parse_m3u8_attributes(x_media_line)
2110 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
2111 media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME')
2112 if not (media_type and group_id and name):
2114 groups.setdefault(group_id, []).append(media)
2115 # <https://tools.ietf.org/html/rfc8216#section-4.3.4.1>
2116 if media_type == 'SUBTITLES':
2117 # According to RFC 8216 §4.3.4.2.1, URI is REQUIRED in the
2118 # EXT-X-MEDIA tag if the media type is SUBTITLES.
2119 # However, lack of URI has been spotted in the wild.
2120 # e.g. NebulaIE; see https://github.com/yt-dlp/yt-dlp/issues/339
2121 if not media.get('URI'):
2123 url = format_url(media['URI'])
2126 'ext': determine_ext(url),
2128 if sub_info['ext'] == 'm3u8':
2129 # Per RFC 8216 §3.1, the only possible subtitle format m3u8
2130 # files may contain is WebVTT:
2131 # <https://tools.ietf.org/html/rfc8216#section-3.1>
2132 sub_info['ext'] = 'vtt'
2133 sub_info['protocol'] = 'm3u8_native'
2134 lang = media.get('LANGUAGE') or 'und'
2135 subtitles.setdefault(lang, []).append(sub_info)
2136 if media_type not in ('VIDEO', 'AUDIO'):
2138 media_url = media.get('URI')
2140 manifest_url = format_url(media_url)
2142 'format_id': join_nonempty(m3u8_id, group_id, name, idx),
2143 'format_note': name,
2144 'format_index': idx,
2145 'url': manifest_url,
2146 'manifest_url': m3u8_url,
2147 'language': media.get('LANGUAGE'),
2149 'protocol': entry_protocol,
2150 'preference': preference,
2153 'vcodec': 'none' if media_type == 'AUDIO' else None,
2154 } for idx in _extract_m3u8_playlist_indices(manifest_url))
2156 def build_stream_name():
2157 # Despite specification does not mention NAME attribute for
2158 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
2159 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
2160 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
2161 stream_name = last_stream_inf.get('NAME')
2164 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
2165 # from corresponding rendition group
2166 stream_group_id = last_stream_inf.get('VIDEO')
2167 if not stream_group_id:
2169 stream_group = groups.get(stream_group_id)
2170 if not stream_group:
2171 return stream_group_id
2172 rendition = stream_group[0]
2173 return rendition.get('NAME') or stream_group_id
2175 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
2176 # chance to detect video only formats when EXT-X-STREAM-INF tags
2177 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
2178 for line in m3u8_doc.splitlines():
2179 if line.startswith('#EXT-X-MEDIA:'):
2182 for line in m3u8_doc.splitlines():
2183 if line.startswith('#EXT-X-STREAM-INF:'):
2184 last_stream_inf = parse_m3u8_attributes(line)
2185 elif line.startswith('#') or not line.strip():
2188 tbr = float_or_none(
2189 last_stream_inf.get('AVERAGE-BANDWIDTH')
2190 or last_stream_inf.get('BANDWIDTH'), scale=1000)
2191 manifest_url = format_url(line.strip())
2193 for idx in _extract_m3u8_playlist_indices(manifest_url):
2194 format_id = [m3u8_id, None, idx]
2195 # Bandwidth of live streams may differ over time thus making
2196 # format_id unpredictable. So it's better to keep provided
2199 stream_name = build_stream_name()
2200 format_id[1] = stream_name or '%d' % (tbr or len(formats))
2202 'format_id': join_nonempty(*format_id),
2203 'format_index': idx,
2204 'url': manifest_url,
2205 'manifest_url': m3u8_url,
2208 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')),
2209 'protocol': entry_protocol,
2210 'preference': preference,
2214 resolution = last_stream_inf.get('RESOLUTION')
2216 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
2218 f['width'] = int(mobj.group('width'))
2219 f['height'] = int(mobj.group('height'))
2220 # Unified Streaming Platform
2222 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
2224 abr, vbr = mobj.groups()
2225 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
2230 codecs = parse_codecs(last_stream_inf.get('CODECS'))
2232 audio_group_id = last_stream_inf.get('AUDIO')
2233 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
2234 # references a rendition group MUST have a CODECS attribute.
2235 # However, this is not always respected. E.g. [2]
2236 # contains EXT-X-STREAM-INF tag which references AUDIO
2237 # rendition group but does not have CODECS and despite
2238 # referencing an audio group it represents a complete
2239 # (with audio and video) format. So, for such cases we will
2240 # ignore references to rendition groups and treat them
2241 # as complete formats.
2242 if audio_group_id and codecs and f.get('vcodec') != 'none':
2243 audio_group = groups.get(audio_group_id)
2244 if audio_group and audio_group[0].get('URI'):
2245 # TODO: update acodec for audio only formats with
2247 f['acodec'] = 'none'
2248 if not f.get('ext'):
2249 f['ext'] = 'm4a' if f.get('vcodec') == 'none' else 'mp4'
2253 progressive_uri = last_stream_inf.get('PROGRESSIVE-URI')
2256 del http_f['manifest_url']
2258 'format_id': f['format_id'].replace('hls-', 'http-'),
2260 'url': progressive_uri,
2262 formats.append(http_f)
2264 last_stream_inf = {}
2265 return formats, subtitles
2267 def _extract_m3u8_vod_duration(
2268 self, m3u8_vod_url, video_id, note=None, errnote=None, data=None, headers={}, query={}):
2270 m3u8_vod = self._download_webpage(
2271 m3u8_vod_url, video_id,
2272 note='Downloading m3u8 VOD manifest' if note is None else note,
2273 errnote='Failed to download VOD manifest' if errnote is None else errnote,
2274 fatal=False, data=data, headers=headers, query=query)
2276 return self._parse_m3u8_vod_duration(m3u8_vod or '', video_id)
2278 def _parse_m3u8_vod_duration(self, m3u8_vod, video_id):
2279 if '#EXT-X-ENDLIST' not in m3u8_vod:
2283 float(line[len('#EXTINF:'):].split(',')[0])
2284 for line in m3u8_vod.splitlines() if line.startswith('#EXTINF:'))) or None
2286 def _extract_mpd_vod_duration(
2287 self, mpd_url, video_id, note=None, errnote=None, data=None, headers={}, query={}):
2289 mpd_doc = self._download_xml(
2291 note='Downloading MPD VOD manifest' if note is None else note,
2292 errnote='Failed to download VOD manifest' if errnote is None else errnote,
2293 fatal=False, data=data, headers=headers, query=query)
2294 if not isinstance(mpd_doc, xml.etree.ElementTree.Element):
2296 return int_or_none(parse_duration(mpd_doc.get('mediaPresentationDuration')))
2299 def _xpath_ns(path, namespace=None):
2303 for c in path.split('/'):
2304 if not c or c == '.':
2307 out.append('{%s}%s' % (namespace, c))
2308 return '/'.join(out)
2310 def _extract_smil_formats_and_subtitles(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
2311 if self.get_param('ignore_no_formats_error'):
2314 res = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
2320 return self._parse_smil_formats_and_subtitles(smil, urlh.url, video_id, f4m_params=f4m_params,
2321 namespace=self._parse_smil_namespace(smil))
2323 def _extract_smil_formats(self, *args, **kwargs):
2324 fmts, subs = self._extract_smil_formats_and_subtitles(*args, **kwargs)
2326 self._report_ignoring_subs('SMIL')
2329 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
2330 res = self._download_smil(smil_url, video_id, fatal=fatal)
2337 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
2339 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
2340 return self._download_xml_handle(
2341 smil_url, video_id, 'Downloading SMIL file',
2342 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
2344 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
2345 namespace = self._parse_smil_namespace(smil)
2347 formats, subtitles = self._parse_smil_formats_and_subtitles(
2348 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
2350 video_id = os.path.splitext(url_basename(smil_url))[0]
2354 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
2355 name = meta.attrib.get('name')
2356 content = meta.attrib.get('content')
2357 if not name or not content:
2359 if not title and name == 'title':
2361 elif not description and name in ('description', 'abstract'):
2362 description = content
2363 elif not upload_date and name == 'date':
2364 upload_date = unified_strdate(content)
2367 'id': image.get('type'),
2368 'url': image.get('src'),
2369 'width': int_or_none(image.get('width')),
2370 'height': int_or_none(image.get('height')),
2371 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
2375 'title': title or video_id,
2376 'description': description,
2377 'upload_date': upload_date,
2378 'thumbnails': thumbnails,
2380 'subtitles': subtitles,
2383 def _parse_smil_namespace(self, smil):
2384 return self._search_regex(
2385 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
2387 def _parse_smil_formats(self, *args, **kwargs):
2388 fmts, subs = self._parse_smil_formats_and_subtitles(*args, **kwargs)
2390 self._report_ignoring_subs('SMIL')
2393 def _parse_smil_formats_and_subtitles(
2394 self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
2396 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
2397 b = meta.get('base') or meta.get('httpBase')
2402 formats, subtitles = [], {}
2409 media = itertools.chain.from_iterable(
2410 smil.findall(self._xpath_ns(arg, namespace))
2411 for arg in ['.//video', './/audio', './/media'])
2412 for medium in media:
2413 src = medium.get('src')
2414 if not src or src in srcs:
2418 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
2419 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
2420 width = int_or_none(medium.get('width'))
2421 height = int_or_none(medium.get('height'))
2422 proto = medium.get('proto')
2423 ext = medium.get('ext')
2424 src_ext = determine_ext(src, default_ext=None) or ext or urlhandle_detect_ext(
2425 self._request_webpage(HEADRequest(src), video_id, note='Requesting extension info', fatal=False))
2426 streamer = medium.get('streamer') or base
2428 if proto == 'rtmp' or streamer.startswith('rtmp'):
2434 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
2436 'filesize': filesize,
2440 if transform_rtmp_url:
2441 streamer, src = transform_rtmp_url(streamer, src)
2442 formats[-1].update({
2448 src_url = src if src.startswith('http') else urllib.parse.urljoin(base, src)
2449 src_url = src_url.strip()
2451 if proto == 'm3u8' or src_ext == 'm3u8':
2452 m3u8_formats, m3u8_subs = self._extract_m3u8_formats_and_subtitles(
2453 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
2454 self._merge_subtitles(m3u8_subs, target=subtitles)
2455 if len(m3u8_formats) == 1:
2457 m3u8_formats[0].update({
2458 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
2463 formats.extend(m3u8_formats)
2464 elif src_ext == 'f4m':
2469 'plugin': 'flowplayer-3.2.0.1',
2471 f4m_url += '&' if '?' in f4m_url else '?'
2472 f4m_url += urllib.parse.urlencode(f4m_params)
2473 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
2474 elif src_ext == 'mpd':
2475 mpd_formats, mpd_subs = self._extract_mpd_formats_and_subtitles(
2476 src_url, video_id, mpd_id='dash', fatal=False)
2477 formats.extend(mpd_formats)
2478 self._merge_subtitles(mpd_subs, target=subtitles)
2479 elif re.search(r'\.ism/[Mm]anifest', src_url):
2480 ism_formats, ism_subs = self._extract_ism_formats_and_subtitles(
2481 src_url, video_id, ism_id='mss', fatal=False)
2482 formats.extend(ism_formats)
2483 self._merge_subtitles(ism_subs, target=subtitles)
2484 elif src_url.startswith('http') and self._is_valid_url(src, video_id):
2488 'ext': ext or src_ext or 'flv',
2489 'format_id': 'http-%d' % (bitrate or http_count),
2491 'filesize': filesize,
2496 for medium in smil.findall(self._xpath_ns('.//imagestream', namespace)):
2497 src = medium.get('src')
2498 if not src or src in srcs:
2504 'format_id': 'imagestream-%d' % (imgs_count),
2506 'ext': mimetype2ext(medium.get('type')),
2509 'width': int_or_none(medium.get('width')),
2510 'height': int_or_none(medium.get('height')),
2511 'format_note': 'SMIL storyboards',
2514 smil_subs = self._parse_smil_subtitles(smil, namespace=namespace)
2515 self._merge_subtitles(smil_subs, target=subtitles)
2517 return formats, subtitles
2519 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
2522 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
2523 src = textstream.get('src')
2524 if not src or src in urls:
2527 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
2528 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
2529 subtitles.setdefault(lang, []).append({
2535 def _extract_xspf_playlist(self, xspf_url, playlist_id, fatal=True):
2536 res = self._download_xml_handle(
2537 xspf_url, playlist_id, 'Downloading xpsf playlist',
2538 'Unable to download xspf manifest', fatal=fatal)
2545 return self._parse_xspf(
2546 xspf, playlist_id, xspf_url=xspf_url,
2547 xspf_base_url=base_url(xspf_url))
2549 def _parse_xspf(self, xspf_doc, playlist_id, xspf_url=None, xspf_base_url=None):
2551 'xspf': 'http://xspf.org/ns/0/',
2552 's1': 'http://static.streamone.nl/player/ns/0',
2556 for track in xspf_doc.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
2558 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
2559 description = xpath_text(
2560 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
2561 thumbnail = xpath_text(
2562 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
2563 duration = float_or_none(
2564 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
2567 for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP)):
2568 format_url = urljoin(xspf_base_url, location.text)
2573 'manifest_url': xspf_url,
2574 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
2575 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
2576 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
2582 'description': description,
2583 'thumbnail': thumbnail,
2584 'duration': duration,
2589 def _extract_mpd_formats(self, *args, **kwargs):
2590 fmts, subs = self._extract_mpd_formats_and_subtitles(*args, **kwargs)
2592 self._report_ignoring_subs('DASH')
2595 def _extract_mpd_formats_and_subtitles(self, *args, **kwargs):
2596 periods = self._extract_mpd_periods(*args, **kwargs)
2597 return self._merge_mpd_periods(periods)
2599 def _extract_mpd_periods(
2600 self, mpd_url, video_id, mpd_id=None, note=None, errnote=None,
2601 fatal=True, data=None, headers={}, query={}):
2603 if self.get_param('ignore_no_formats_error'):
2606 res = self._download_xml_handle(
2608 note='Downloading MPD manifest' if note is None else note,
2609 errnote='Failed to download MPD manifest' if errnote is None else errnote,
2610 fatal=fatal, data=data, headers=headers, query=query)
2617 # We could have been redirected to a new url when we retrieved our mpd file.
2619 mpd_base_url = base_url(mpd_url)
2621 return self._parse_mpd_periods(mpd_doc, mpd_id, mpd_base_url, mpd_url)
2623 def _parse_mpd_formats(self, *args, **kwargs):
2624 fmts, subs = self._parse_mpd_formats_and_subtitles(*args, **kwargs)
2626 self._report_ignoring_subs('DASH')
2629 def _parse_mpd_formats_and_subtitles(self, *args, **kwargs):
2630 periods = self._parse_mpd_periods(*args, **kwargs)
2631 return self._merge_mpd_periods(periods)
2633 def _merge_mpd_periods(self, periods):
2635 Combine all formats and subtitles from an MPD manifest into a single list,
2636 by concatenate streams with similar formats.
2638 formats, subtitles = {}, {}
2639 for period in periods:
2640 for f in period['formats']:
2641 assert 'is_dash_periods' not in f, 'format already processed'
2642 f['is_dash_periods'] = True
2643 format_key = tuple(v for k, v in f.items() if k not in (
2644 ('format_id', 'fragments', 'manifest_stream_number')))
2645 if format_key not in formats:
2646 formats[format_key] = f
2647 elif 'fragments' in f:
2648 formats[format_key].setdefault('fragments', []).extend(f['fragments'])
2650 if subtitles and period['subtitles']:
2651 self.report_warning(bug_reports_message(
2652 'Found subtitles in multiple periods in the DASH manifest; '
2653 'if part of the subtitles are missing,'
2656 for sub_lang, sub_info in period['subtitles'].items():
2657 subtitles.setdefault(sub_lang, []).extend(sub_info)
2659 return list(formats.values()), subtitles
2661 def _parse_mpd_periods(self, mpd_doc, mpd_id=None, mpd_base_url='', mpd_url=None):
2663 Parse formats from MPD manifest.
2665 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
2666 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
2667 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
2669 if not self.get_param('dynamic_mpd', True):
2670 if mpd_doc.get('type') == 'dynamic':
2673 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
2676 return self._xpath_ns(path, namespace)
2678 def is_drm_protected(element):
2679 return element.find(_add_ns('ContentProtection')) is not None
2681 def extract_multisegment_info(element, ms_parent_info):
2682 ms_info = ms_parent_info.copy()
2684 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2685 # common attributes and elements. We will only extract relevant
2687 def extract_common(source):
2688 segment_timeline = source.find(_add_ns('SegmentTimeline'))
2689 if segment_timeline is not None:
2690 s_e = segment_timeline.findall(_add_ns('S'))
2692 ms_info['total_number'] = 0
2695 r = int(s.get('r', 0))
2696 ms_info['total_number'] += 1 + r
2697 ms_info['s'].append({
2698 't': int(s.get('t', 0)),
2699 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2700 'd': int(s.attrib['d']),
2703 start_number = source.get('startNumber')
2705 ms_info['start_number'] = int(start_number)
2706 timescale = source.get('timescale')
2708 ms_info['timescale'] = int(timescale)
2709 segment_duration = source.get('duration')
2710 if segment_duration:
2711 ms_info['segment_duration'] = float(segment_duration)
2713 def extract_Initialization(source):
2714 initialization = source.find(_add_ns('Initialization'))
2715 if initialization is not None:
2716 ms_info['initialization_url'] = initialization.attrib['sourceURL']
2718 segment_list = element.find(_add_ns('SegmentList'))
2719 if segment_list is not None:
2720 extract_common(segment_list)
2721 extract_Initialization(segment_list)
2722 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
2724 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
2726 segment_template = element.find(_add_ns('SegmentTemplate'))
2727 if segment_template is not None:
2728 extract_common(segment_template)
2729 media = segment_template.get('media')
2731 ms_info['media'] = media
2732 initialization = segment_template.get('initialization')
2734 ms_info['initialization'] = initialization
2736 extract_Initialization(segment_template)
2739 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
2740 stream_numbers = collections.defaultdict(int)
2741 for period_idx, period in enumerate(mpd_doc.findall(_add_ns('Period'))):
2743 'id': period.get('id', f'period-{period_idx}'),
2745 'subtitles': collections.defaultdict(list),
2747 period_duration = parse_duration(period.get('duration')) or mpd_duration
2748 period_ms_info = extract_multisegment_info(period, {
2752 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
2753 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
2754 for representation in adaptation_set.findall(_add_ns('Representation')):
2755 representation_attrib = adaptation_set.attrib.copy()
2756 representation_attrib.update(representation.attrib)
2757 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2758 mime_type = representation_attrib['mimeType']
2759 content_type = representation_attrib.get('contentType', mime_type.split('/')[0])
2761 codec_str = representation_attrib.get('codecs', '')
2762 # Some kind of binary subtitle found in some youtube livestreams
2763 if mime_type == 'application/x-rawcc':
2764 codecs = {'scodec': codec_str}
2766 codecs = parse_codecs(codec_str)
2767 if content_type not in ('video', 'audio', 'text'):
2768 if mime_type == 'image/jpeg':
2769 content_type = mime_type
2770 elif codecs.get('vcodec', 'none') != 'none':
2771 content_type = 'video'
2772 elif codecs.get('acodec', 'none') != 'none':
2773 content_type = 'audio'
2774 elif codecs.get('scodec', 'none') != 'none':
2775 content_type = 'text'
2776 elif mimetype2ext(mime_type) in ('tt', 'dfxp', 'ttml', 'xml', 'json'):
2777 content_type = 'text'
2779 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
2783 for element in (representation, adaptation_set, period, mpd_doc):
2784 base_url_e = element.find(_add_ns('BaseURL'))
2785 if try_call(lambda: base_url_e.text) is not None:
2786 base_url = base_url_e.text + base_url
2787 if re.match(r'^https?://', base_url):
2789 if mpd_base_url and base_url.startswith('/'):
2790 base_url = urllib.parse.urljoin(mpd_base_url, base_url)
2791 elif mpd_base_url and not re.match(r'^https?://', base_url):
2792 if not mpd_base_url.endswith('/'):
2794 base_url = mpd_base_url + base_url
2795 representation_id = representation_attrib.get('id')
2796 lang = representation_attrib.get('lang')
2797 url_el = representation.find(_add_ns('BaseURL'))
2798 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
2799 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
2800 if representation_id is not None:
2801 format_id = representation_id
2803 format_id = content_type
2805 format_id = mpd_id + '-' + format_id
2806 if content_type in ('video', 'audio'):
2808 'format_id': format_id,
2809 'manifest_url': mpd_url,
2810 'ext': mimetype2ext(mime_type),
2811 'width': int_or_none(representation_attrib.get('width')),
2812 'height': int_or_none(representation_attrib.get('height')),
2813 'tbr': float_or_none(bandwidth, 1000),
2814 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
2815 'fps': int_or_none(representation_attrib.get('frameRate')),
2816 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
2817 'format_note': 'DASH %s' % content_type,
2818 'filesize': filesize,
2819 'container': mimetype2ext(mime_type) + '_dash',
2822 elif content_type == 'text':
2824 'ext': mimetype2ext(mime_type),
2825 'manifest_url': mpd_url,
2826 'filesize': filesize,
2828 elif content_type == 'image/jpeg':
2829 # See test case in VikiIE
2830 # https://www.viki.com/videos/1175236v-choosing-spouse-by-lottery-episode-1
2832 'format_id': format_id,
2834 'manifest_url': mpd_url,
2835 'format_note': 'DASH storyboards (jpeg)',
2839 if is_drm_protected(adaptation_set) or is_drm_protected(representation):
2841 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
2843 def prepare_template(template_name, identifiers):
2844 tmpl = representation_ms_info[template_name]
2845 if representation_id is not None:
2846 tmpl = tmpl.replace('$RepresentationID$', representation_id)
2847 # First of, % characters outside $...$ templates
2848 # must be escaped by doubling for proper processing
2849 # by % operator string formatting used further (see
2850 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2856 in_template = not in_template
2857 elif c == '%' and not in_template:
2859 # Next, $...$ templates are translated to their
2860 # %(...) counterparts to be used with % operator
2861 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
2862 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
2863 t.replace('$$', '$')
2866 # @initialization is a regular template like @media one
2867 # so it should be handled just the same way (see
2868 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2869 if 'initialization' in representation_ms_info:
2870 initialization_template = prepare_template(
2872 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2873 # $Time$ shall not be included for @initialization thus
2874 # only $Bandwidth$ remains
2876 representation_ms_info['initialization_url'] = initialization_template % {
2877 'Bandwidth': bandwidth,
2880 def location_key(location):
2881 return 'url' if re.match(r'^https?://', location) else 'path'
2883 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
2885 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2886 media_location_key = location_key(media_template)
2888 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2889 # can't be used at the same time
2890 if '%(Number' in media_template and 's' not in representation_ms_info:
2891 segment_duration = None
2892 if 'total_number' not in representation_ms_info and 'segment_duration' in representation_ms_info:
2893 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
2894 representation_ms_info['total_number'] = int(math.ceil(
2895 float_or_none(period_duration, segment_duration, default=0)))
2896 representation_ms_info['fragments'] = [{
2897 media_location_key: media_template % {
2898 'Number': segment_number,
2899 'Bandwidth': bandwidth,
2901 'duration': segment_duration,
2902 } for segment_number in range(
2903 representation_ms_info['start_number'],
2904 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
2906 # $Number*$ or $Time$ in media template with S list available
2907 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2908 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2909 representation_ms_info['fragments'] = []
2912 segment_number = representation_ms_info['start_number']
2914 def add_segment_url():
2915 segment_url = media_template % {
2916 'Time': segment_time,
2917 'Bandwidth': bandwidth,
2918 'Number': segment_number,
2920 representation_ms_info['fragments'].append({
2921 media_location_key: segment_url,
2922 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
2925 for num, s in enumerate(representation_ms_info['s']):
2926 segment_time = s.get('t') or segment_time
2930 for r in range(s.get('r', 0)):
2931 segment_time += segment_d
2934 segment_time += segment_d
2935 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
2936 # No media template,
2937 # e.g. https://www.youtube.com/watch?v=iXZV5uAYMJI
2938 # or any YouTube dashsegments video
2941 timescale = representation_ms_info['timescale']
2942 for s in representation_ms_info['s']:
2943 duration = float_or_none(s['d'], timescale)
2944 for r in range(s.get('r', 0) + 1):
2945 segment_uri = representation_ms_info['segment_urls'][segment_index]
2947 location_key(segment_uri): segment_uri,
2948 'duration': duration,
2951 representation_ms_info['fragments'] = fragments
2952 elif 'segment_urls' in representation_ms_info:
2953 # Segment URLs with no SegmentTimeline
2954 # E.g. https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2955 # https://github.com/ytdl-org/youtube-dl/pull/14844
2957 segment_duration = float_or_none(
2958 representation_ms_info['segment_duration'],
2959 representation_ms_info['timescale']) if 'segment_duration' in representation_ms_info else None
2960 for segment_url in representation_ms_info['segment_urls']:
2962 location_key(segment_url): segment_url,
2964 if segment_duration:
2965 fragment['duration'] = segment_duration
2966 fragments.append(fragment)
2967 representation_ms_info['fragments'] = fragments
2968 # If there is a fragments key available then we correctly recognized fragmented media.
2969 # Otherwise we will assume unfragmented media with direct access. Technically, such
2970 # assumption is not necessarily correct since we may simply have no support for
2971 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2972 if 'fragments' in representation_ms_info:
2974 # NB: mpd_url may be empty when MPD manifest is parsed from a string
2975 'url': mpd_url or base_url,
2976 'fragment_base_url': base_url,
2978 'protocol': 'http_dash_segments' if mime_type != 'image/jpeg' else 'mhtml',
2980 if 'initialization_url' in representation_ms_info:
2981 initialization_url = representation_ms_info['initialization_url']
2982 if not f.get('url'):
2983 f['url'] = initialization_url
2984 f['fragments'].append({location_key(initialization_url): initialization_url})
2985 f['fragments'].extend(representation_ms_info['fragments'])
2986 if not period_duration:
2987 period_duration = try_get(
2988 representation_ms_info,
2989 lambda r: sum(frag['duration'] for frag in r['fragments']), float)
2991 # Assuming direct URL to unfragmented media.
2993 if content_type in ('video', 'audio', 'image/jpeg'):
2994 f['manifest_stream_number'] = stream_numbers[f['url']]
2995 stream_numbers[f['url']] += 1
2996 period_entry['formats'].append(f)
2997 elif content_type == 'text':
2998 period_entry['subtitles'][lang or 'und'].append(f)
3001 def _extract_ism_formats(self, *args, **kwargs):
3002 fmts, subs = self._extract_ism_formats_and_subtitles(*args, **kwargs)
3004 self._report_ignoring_subs('ISM')
3007 def _extract_ism_formats_and_subtitles(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
3008 if self.get_param('ignore_no_formats_error'):
3011 res = self._download_xml_handle(
3013 note='Downloading ISM manifest' if note is None else note,
3014 errnote='Failed to download ISM manifest' if errnote is None else errnote,
3015 fatal=fatal, data=data, headers=headers, query=query)
3022 return self._parse_ism_formats_and_subtitles(ism_doc, urlh.url, ism_id)
3024 def _parse_ism_formats_and_subtitles(self, ism_doc, ism_url, ism_id=None):
3026 Parse formats from ISM manifest.
3028 1. [MS-SSTR]: Smooth Streaming Protocol,
3029 https://msdn.microsoft.com/en-us/library/ff469518.aspx
3031 if ism_doc.get('IsLive') == 'TRUE':
3034 duration = int(ism_doc.attrib['Duration'])
3035 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
3039 for stream in ism_doc.findall('StreamIndex'):
3040 stream_type = stream.get('Type')
3041 if stream_type not in ('video', 'audio', 'text'):
3043 url_pattern = stream.attrib['Url']
3044 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
3045 stream_name = stream.get('Name')
3046 stream_language = stream.get('Language', 'und')
3047 for track in stream.findall('QualityLevel'):
3048 KNOWN_TAGS = {'255': 'AACL', '65534': 'EC-3'}
3049 fourcc = track.get('FourCC') or KNOWN_TAGS.get(track.get('AudioTag'))
3050 # TODO: add support for WVC1 and WMAP
3051 if fourcc not in ('H264', 'AVC1', 'AACL', 'TTML', 'EC-3'):
3052 self.report_warning('%s is not a supported codec' % fourcc)
3054 tbr = int(track.attrib['Bitrate']) // 1000
3055 # [1] does not mention Width and Height attributes. However,
3056 # they're often present while MaxWidth and MaxHeight are
3057 # missing, so should be used as fallbacks
3058 width = int_or_none(track.get('MaxWidth') or track.get('Width'))
3059 height = int_or_none(track.get('MaxHeight') or track.get('Height'))
3060 sampling_rate = int_or_none(track.get('SamplingRate'))
3062 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
3063 track_url_pattern = urllib.parse.urljoin(ism_url, track_url_pattern)
3069 stream_fragments = stream.findall('c')
3070 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
3071 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
3072 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
3073 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
3074 if not fragment_ctx['duration']:
3076 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
3078 next_fragment_time = duration
3079 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
3080 for _ in range(fragment_repeat):
3082 'url': re.sub(r'{start[ _]time}', str(fragment_ctx['time']), track_url_pattern),
3083 'duration': fragment_ctx['duration'] / stream_timescale,
3085 fragment_ctx['time'] += fragment_ctx['duration']
3087 if stream_type == 'text':
3088 subtitles.setdefault(stream_language, []).append({
3092 'manifest_url': ism_url,
3093 'fragments': fragments,
3094 '_download_params': {
3095 'stream_type': stream_type,
3096 'duration': duration,
3097 'timescale': stream_timescale,
3099 'language': stream_language,
3100 'codec_private_data': track.get('CodecPrivateData'),
3103 elif stream_type in ('video', 'audio'):
3105 'format_id': join_nonempty(ism_id, stream_name, tbr),
3107 'manifest_url': ism_url,
3108 'ext': 'ismv' if stream_type == 'video' else 'isma',
3112 'asr': sampling_rate,
3113 'vcodec': 'none' if stream_type == 'audio' else fourcc,
3114 'acodec': 'none' if stream_type == 'video' else fourcc,
3116 'fragments': fragments,
3117 'has_drm': ism_doc.find('Protection') is not None,
3118 'language': stream_language,
3119 'audio_channels': int_or_none(track.get('Channels')),
3120 '_download_params': {
3121 'stream_type': stream_type,
3122 'duration': duration,
3123 'timescale': stream_timescale,
3124 'width': width or 0,
3125 'height': height or 0,
3127 'language': stream_language,
3128 'codec_private_data': track.get('CodecPrivateData'),
3129 'sampling_rate': sampling_rate,
3130 'channels': int_or_none(track.get('Channels', 2)),
3131 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
3132 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
3135 return formats, subtitles
3137 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):
3138 def absolute_url(item_url):
3139 return urljoin(base_url, item_url)
3141 def parse_content_type(content_type):
3142 if not content_type:
3144 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?
(?P
<codecs
>[^
"]+))?', content_type)
3146 mimetype, codecs = ctr.groups()
3147 f = parse_codecs(codecs)
3148 f['ext'] = mimetype2ext(mimetype)
3152 def _media_formats(src, cur_media_type, type_info=None):
3153 type_info = type_info or {}
3154 full_url = absolute_url(src)
3155 ext = type_info.get('ext') or determine_ext(full_url)
3157 is_plain_url = False
3158 formats = self._extract_m3u8_formats(
3159 full_url, video_id, ext='mp4',
3160 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
3161 preference=preference, quality=quality, fatal=False)
3163 is_plain_url = False
3164 formats = self._extract_mpd_formats(
3165 full_url, video_id, mpd_id=mpd_id, fatal=False)
3170 'vcodec': 'none' if cur_media_type == 'audio' else None,
3173 return is_plain_url, formats
3176 # amp-video and amp-audio are very similar to their HTML5 counterparts
3177 # so we will include them right here (see
3178 # https://www.ampproject.org/docs/reference/components/amp-video)
3179 # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/
3180 _MEDIA_TAG_NAME_RE = r'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)'
3181 media_tags = [(media_tag, media_tag_name, media_type, '')
3182 for media_tag, media_tag_name, media_type
3183 in re.findall(r'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE, webpage)]
3184 media_tags.extend(re.findall(
3185 # We only allow video|audio followed by a whitespace or '>'.
3186 # Allowing more characters may end up in significant slow down (see
3187 # https://github.com/ytdl-org/youtube-dl/issues/11979,
3188 # e.g. http://www.porntrex.com/maps/videositemap.xml).
3189 r'(?s)(<(?P<tag>%s)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>' % _MEDIA_TAG_NAME_RE, webpage))
3190 for media_tag, _, media_type, media_content in media_tags:
3195 media_attributes = extract_attributes(media_tag)
3196 src = strip_or_none(dict_get(media_attributes, ('src', 'data-video-src', 'data-src', 'data-source')))
3198 f = parse_content_type(media_attributes.get('type'))
3199 _, formats = _media_formats(src, media_type, f)
3200 media_info['formats'].extend(formats)
3201 media_info['thumbnail'] = absolute_url(media_attributes.get('poster'))
3203 for source_tag in re.findall(r'<source[^>]+>', media_content):
3204 s_attr = extract_attributes(source_tag)
3205 # data-video-src and data-src are non standard but seen
3206 # several times in the wild
3207 src = strip_or_none(dict_get(s_attr, ('src', 'data-video-src', 'data-src', 'data-source')))
3210 f = parse_content_type(s_attr.get('type'))
3211 is_plain_url, formats = _media_formats(src, media_type, f)
3213 # width, height, res, label and title attributes are
3214 # all not standard but seen several times in the wild
3217 for lbl in ('label', 'title')
3218 if str_or_none(s_attr.get(lbl))
3220 width = int_or_none(s_attr.get('width'))
3221 height = (int_or_none(s_attr.get('height'))
3222 or int_or_none(s_attr.get('res')))
3223 if not width or not height:
3225 resolution = parse_resolution(lbl)
3228 width = width or resolution.get('width')
3229 height = height or resolution.get('height')
3231 tbr = parse_bitrate(lbl)
3240 'format_id': s_attr.get('label') or s_attr.get('title'),
3242 f.update(formats[0])
3243 media_info['formats'].append(f)
3245 media_info['formats'].extend(formats)
3246 for track_tag in re.findall(r'<track[^>]+>', media_content):
3247 track_attributes = extract_attributes(track_tag)
3248 kind = track_attributes.get('kind')
3249 if not kind or kind in ('subtitles', 'captions'):
3250 src = strip_or_none(track_attributes.get('src'))
3253 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
3254 media_info['subtitles'].setdefault(lang, []).append({
3255 'url': absolute_url(src),
3257 for f in media_info['formats']:
3258 f.setdefault('http_headers', {})['Referer'] = base_url
3259 if media_info['formats'] or media_info['subtitles']:
3260 entries.append(media_info)
3263 def _extract_akamai_formats(self, *args, **kwargs):
3264 fmts, subs = self._extract_akamai_formats_and_subtitles(*args, **kwargs)
3266 self._report_ignoring_subs('akamai')
3269 def _extract_akamai_formats_and_subtitles(self, manifest_url, video_id, hosts={}):
3270 signed = 'hdnea=' in manifest_url
3272 # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
3273 manifest_url = re.sub(
3274 r'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
3275 '', manifest_url).strip('?')
3280 hdcore_sign = 'hdcore=3.7.0'
3281 f4m_url = re.sub(r'(https?://[^/]+)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
3282 hds_host = hosts.get('hds')
3284 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
3285 if 'hdcore=' not in f4m_url:
3286 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
3287 f4m_formats = self._extract_f4m_formats(
3288 f4m_url, video_id, f4m_id='hds', fatal=False)
3289 for entry in f4m_formats:
3290 entry.update({'extra_param_to_segment_url': hdcore_sign})
3291 formats.extend(f4m_formats)
3293 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
3294 hls_host = hosts.get('hls')
3296 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
3297 m3u8_formats, m3u8_subtitles = self._extract_m3u8_formats_and_subtitles(
3298 m3u8_url, video_id, 'mp4', 'm3u8_native',
3299 m3u8_id='hls', fatal=False)
3300 formats.extend(m3u8_formats)
3301 subtitles = self._merge_subtitles(subtitles, m3u8_subtitles)
3303 http_host = hosts.get('http')
3304 if http_host and m3u8_formats and not signed:
3305 REPL_REGEX = r'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
3306 qualities = re.match(REPL_REGEX, m3u8_url).group(2).split(',')
3307 qualities_length = len(qualities)
3308 if len(m3u8_formats) in (qualities_length, qualities_length + 1):
3310 for f in m3u8_formats:
3311 if f['vcodec'] != 'none':
3312 for protocol in ('http', 'https'):
3314 del http_f['manifest_url']
3316 REPL_REGEX, protocol + fr'://{http_host}/\g<1>{qualities[i]}\3', f['url'])
3318 'format_id': http_f['format_id'].replace('hls-', protocol + '-'),
3320 'protocol': protocol,
3322 formats.append(http_f)
3325 return formats, subtitles
3327 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
3328 query = urllib.parse.urlparse(url).query
3329 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
3331 r'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url)
3332 url_base = mobj.group('url')
3333 http_base_url = '%s%s:%s' % ('http', mobj.group('s') or '', url_base)
3336 def manifest_url(manifest):
3337 m_url = f'{http_base_url}/{manifest}'
3339 m_url += '?%s' % query
3342 if 'm3u8' not in skip_protocols:
3343 formats.extend(self._extract_m3u8_formats(
3344 manifest_url('playlist.m3u8'), video_id, 'mp4',
3345 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
3346 if 'f4m' not in skip_protocols:
3347 formats.extend(self._extract_f4m_formats(
3348 manifest_url('manifest.f4m'),
3349 video_id, f4m_id='hds', fatal=False))
3350 if 'dash' not in skip_protocols:
3351 formats.extend(self._extract_mpd_formats(
3352 manifest_url('manifest.mpd'),
3353 video_id, mpd_id='dash', fatal=False))
3354 if re.search(r'(?:/smil:|\.smil)', url_base):
3355 if 'smil' not in skip_protocols:
3356 rtmp_formats = self._extract_smil_formats(
3357 manifest_url('jwplayer.smil'),
3358 video_id, fatal=False)
3359 for rtmp_format in rtmp_formats:
3360 rtsp_format = rtmp_format.copy()
3361 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
3362 del rtsp_format['play_path']
3363 del rtsp_format['ext']
3364 rtsp_format.update({
3365 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
3366 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
3369 formats.extend([rtmp_format, rtsp_format])
3371 for protocol in ('rtmp', 'rtsp'):
3372 if protocol not in skip_protocols:
3374 'url': f'{protocol}:{url_base}',
3375 'format_id': protocol,
3376 'protocol': protocol,
3380 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
3382 r'''(?s)jwplayer\s*\(\s*(?P<q>'|")(?
!(?P
=q
)).+(?P
=q
)\s
*\
)(?
!</script
>).*?\
.\s
*setup\s
*\
(\s
*(?P
<options
>(?
:\
([^
)]*\
)|
[^
)])+)\s
*\
)''',
3386 jwplayer_data = self._parse_json(mobj.group('options'),
3388 transform_source=transform_source)
3389 except ExtractorError:
3392 if isinstance(jwplayer_data, dict):
3393 return jwplayer_data
3395 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
3396 jwplayer_data = self._find_jwplayer_data(
3397 webpage, video_id, transform_source=js_to_json)
3398 return self._parse_jwplayer_data(
3399 jwplayer_data, video_id, *args, **kwargs)
3401 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
3402 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3404 if not isinstance(jwplayer_data, dict):
3407 playlist_items = jwplayer_data.get('playlist')
3408 # JWPlayer backward compatibility: single playlist item/flattened playlists
3409 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
3410 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
3411 if not isinstance(playlist_items, list):
3412 playlist_items = (playlist_items or jwplayer_data, )
3414 for video_data in playlist_items:
3415 if not isinstance(video_data, dict):
3417 # JWPlayer backward compatibility: flattened sources
3418 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
3419 if 'sources' not in video_data:
3420 video_data['sources'] = [video_data]
3422 this_video_id = video_id or video_data['mediaid']
3424 formats = self._parse_jwplayer_formats(
3425 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
3426 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
3429 tracks = video_data.get('tracks')
3430 if tracks and isinstance(tracks, list):
3431 for track in tracks:
3432 if not isinstance(track, dict):
3434 track_kind = track.get('kind')
3435 if not track_kind or not isinstance(track_kind, str):
3437 if track_kind.lower() not in ('captions', 'subtitles'):
3439 track_url = urljoin(base_url, track.get('file'))
3442 subtitles.setdefault(track.get('label') or 'en', []).append({
3443 'url': self._proto_relative_url(track_url)
3447 'id': this_video_id,
3448 'title': unescapeHTML(video_data['title'] if require_title else video_data.get('title')),
3449 'description': clean_html(video_data.get('description')),
3450 'thumbnail': urljoin(base_url, self._proto_relative_url(video_data.get('image'))),
3451 'timestamp': int_or_none(video_data.get('pubdate')),
3452 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
3453 'subtitles': subtitles,
3454 'alt_title': clean_html(video_data.get('subtitle')), # attributes used e.g. by Tele5 ...
3455 'genre': clean_html(video_data.get('genre')),
3456 'channel': clean_html(dict_get(video_data, ('category', 'channel'))),
3457 'season_number': int_or_none(video_data.get('season')),
3458 'episode_number': int_or_none(video_data.get('episode')),
3459 'release_year': int_or_none(video_data.get('releasedate')),
3460 'age_limit': int_or_none(video_data.get('age_restriction')),
3462 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
3463 if len(formats) == 1 and re.search(r'^(?:http|//).*(?:youtube\.com|youtu\.be)/.+', formats[0]['url']):
3465 '_type': 'url_transparent',
3466 'url': formats[0]['url'],
3469 entry['formats'] = formats
3470 entries.append(entry)
3471 if len(entries) == 1:
3474 return self.playlist_result(entries)
3476 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
3477 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3480 for source in jwplayer_sources_data:
3481 if not isinstance(source, dict):
3483 source_url = urljoin(
3484 base_url, self._proto_relative_url(source.get('file')))
3485 if not source_url or source_url in urls:
3487 urls.add(source_url)
3488 source_type = source.get('type') or ''
3489 ext = mimetype2ext(source_type) or determine_ext(source_url)
3490 if source_type == 'hls' or ext == 'm3u8' or 'format=m3u8-aapl' in source_url:
3491 formats.extend(self._extract_m3u8_formats(
3492 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
3493 m3u8_id=m3u8_id, fatal=False))
3494 elif source_type == 'dash' or ext == 'mpd' or 'format=mpd-time-csf' in source_url:
3495 formats.extend(self._extract_mpd_formats(
3496 source_url, video_id, mpd_id=mpd_id, fatal=False))
3498 formats.extend(self._extract_smil_formats(
3499 source_url, video_id, fatal=False))
3500 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3501 elif source_type.startswith('audio') or ext in (
3502 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
3509 format_id = str_or_none(source.get('label'))
3510 height = int_or_none(source.get('height'))
3511 if height is None and format_id:
3512 # Often no height is provided but there is a label in
3513 # format like "1080p", "720p SD", or 1080.
3514 height = parse_resolution(format_id).get('height')
3517 'width': int_or_none(source.get('width')),
3519 'tbr': int_or_none(source.get('bitrate'), scale=1000),
3520 'filesize': int_or_none(source.get('filesize')),
3522 'format_id': format_id
3524 if source_url.startswith('rtmp'):
3525 a_format['ext'] = 'flv'
3526 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3527 # of jwplayer.flash.swf
3528 rtmp_url_parts = re.split(
3529 r'((?:mp4|mp3|flv):)', source_url, 1)
3530 if len(rtmp_url_parts) == 3:
3531 rtmp_url, prefix, play_path = rtmp_url_parts
3534 'play_path': prefix + play_path,
3537 a_format.update(rtmp_params)
3538 formats.append(a_format)
3541 def _live_title(self, name):
3542 self._downloader.deprecation_warning('yt_dlp.InfoExtractor._live_title is deprecated and does not work as expected')
3545 def _int(self, v, name, fatal=False, **kwargs):
3546 res = int_or_none(v, **kwargs)
3548 msg = f'Failed to extract {name}: Could not parse value {v!r}'
3550 raise ExtractorError(msg)
3552 self.report_warning(msg)
3555 def _float(self, v, name, fatal=False, **kwargs):
3556 res = float_or_none(v, **kwargs)
3558 msg = f'Failed to extract {name}: Could not parse value {v!r}'
3560 raise ExtractorError(msg)
3562 self.report_warning(msg)
3565 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
3566 path='/', secure=False, discard=False, rest={}, **kwargs):
3567 cookie = http.cookiejar.Cookie(
3568 0, name, value, port, port is not None, domain, True,
3569 domain.startswith('.'), path, True, secure, expire_time,
3570 discard, None, None, rest)
3571 self.cookiejar.set_cookie(cookie)
3573 def _get_cookies(self, url):
3574 """ Return a http.cookies.SimpleCookie with the cookies for the url """
3575 return LenientSimpleCookie(self._downloader.cookiejar.get_cookie_header(url))
3577 def _apply_first_set_cookie_header(self, url_handle, cookie):
3579 Apply first Set-Cookie header instead of the last. Experimental.
3581 Some sites (e.g. [1-3]) may serve two cookies under the same name
3582 in Set-Cookie header and expect the first (old) one to be set rather
3583 than second (new). However, as of RFC6265 the newer one cookie
3584 should be set into cookie store what actually happens.
3585 We will workaround this issue by resetting the cookie to
3586 the first one manually.
3587 1. https://new.vk.com/
3588 2. https://github.com/ytdl-org/youtube-dl/issues/9841#issuecomment-227871201
3589 3. https://learning.oreilly.com/
3591 for header, cookies in url_handle.headers.items():
3592 if header.lower() != 'set-cookie':
3594 cookies = cookies.encode('iso-8859-1').decode('utf-8')
3595 cookie_value = re.search(
3596 r'%s=(.+?);.*?\b[Dd]omain=(.+?)(?:[,;]|$)' % cookie, cookies)
3598 value, domain = cookie_value.groups()
3599 self._set_cookie(domain, cookie, value)
3603 def get_testcases(cls, include_onlymatching=False):
3604 # Do not look in super classes
3605 t = vars(cls).get('_TEST')
3607 assert not hasattr(cls, '_TESTS'), f'{cls.ie_key()}IE has _TEST and _TESTS'
3610 tests = vars(cls).get('_TESTS', [])
3612 if not include_onlymatching and t.get('only_matching', False):
3614 t['name'] = cls.ie_key()
3616 if getattr(cls, '__wrapped__', None):
3617 yield from cls.__wrapped__.get_testcases(include_onlymatching)
3620 def get_webpage_testcases(cls):
3621 tests = vars(cls).get('_WEBPAGE_TESTS', [])
3623 t['name'] = cls.ie_key()
3625 if getattr(cls, '__wrapped__', None):
3626 yield from cls.__wrapped__.get_webpage_testcases()
3628 @classproperty(cache=True)
3630 """Get age limit from the testcases"""
3631 return max(traverse_obj(
3632 (*cls.get_testcases(include_onlymatching=False), *cls.get_webpage_testcases()),
3633 (..., (('playlist', 0), None), 'info_dict', 'age_limit')) or [0])
3635 @classproperty(cache=True)
3636 def _RETURN_TYPE(cls):
3637 """What the extractor returns: "video", "playlist", "any", or None (Unknown)"""
3638 tests = tuple(cls.get_testcases(include_onlymatching=False))
3641 elif not any(k.startswith('playlist') for test in tests for k in test):
3643 elif all(any(k.startswith('playlist') for k in test) for test in tests):
3648 def is_single_video(cls, url):
3649 """Returns whether the URL is of a single video, None if unknown"""
3650 if cls.suitable(url):
3651 return {'video': True, 'playlist': False}.get(cls._RETURN_TYPE)
3654 def is_suitable(cls, age_limit):
3655 """Test whether the extractor is generally suitable for the given age limit"""
3656 return not age_restricted(cls.age_limit, age_limit)
3659 def description(cls, *, markdown=True, search_examples=None):
3660 """Description of the extractor"""
3662 if cls._NETRC_MACHINE:
3664 desc += f' [*{cls._NETRC_MACHINE}*](## "netrc machine")'
3666 desc += f' [{cls._NETRC_MACHINE}]'
3667 if cls.IE_DESC is False:
3670 desc += f' {cls.IE_DESC}'
3672 desc += f'{";" if cls.IE_DESC else ""} "{cls.SEARCH_KEY}:" prefix'
3674 _COUNTS = ('', '5', '10', 'all')
3675 desc += f' (e.g. "{cls.SEARCH_KEY}{random.choice(_COUNTS)}:{random.choice(search_examples)}")'
3676 if not cls.working():
3677 desc += ' (**Currently broken**)' if markdown else ' (Currently broken)'
3679 # Escape emojis. Ref: https://github.com/github/markup/issues/1153
3680 name = (' - **%s**' % re.sub(r':(\w+:)', ':\u200B\\g<1>', cls.IE_NAME)) if markdown else cls.IE_NAME
3681 return f'{name}:{desc}' if desc else name
3683 def extract_subtitles(self, *args, **kwargs):
3684 if (self.get_param('writesubtitles', False)
3685 or self.get_param('listsubtitles')):
3686 return self._get_subtitles(*args, **kwargs)
3689 def _get_subtitles(self, *args, **kwargs):
3690 raise NotImplementedError('This method must be implemented by subclasses')
3692 class CommentsDisabled(Exception):
3693 """Raise in _get_comments if comments are disabled for the video"""
3695 def extract_comments(self, *args, **kwargs):
3696 if not self.get_param('getcomments'):
3698 generator = self._get_comments(*args, **kwargs)
3705 comments.append(next(generator))
3706 except StopIteration:
3708 except KeyboardInterrupt:
3709 self.to_screen('Interrupted by user')
3710 except self.CommentsDisabled:
3711 return {'comments': None, 'comment_count': None}
3712 except Exception as e:
3713 if self.get_param('ignoreerrors') is not True:
3715 self._downloader.report_error(e)
3716 comment_count = len(comments)
3717 self.to_screen(f'Extracted {comment_count} comments')
3719 'comments': comments,
3720 'comment_count': None if interrupted else comment_count
3724 def _get_comments(self, *args, **kwargs):
3725 raise NotImplementedError('This method must be implemented by subclasses')
3728 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
3729 """ Merge subtitle items for one language. Items with duplicated URLs/data
3730 will be dropped. """
3731 list1_data = {(item.get('url'), item.get('data')) for item in subtitle_list1}
3732 ret = list(subtitle_list1)
3733 ret.extend(item for item in subtitle_list2 if (item.get('url'), item.get('data')) not in list1_data)
3737 def _merge_subtitles(cls, *dicts, target=None):
3738 """ Merge subtitle dictionaries, language by language. """
3742 for lang, subs in d.items():
3743 target[lang] = cls._merge_subtitle_items(target.get(lang, []), subs)
3746 def extract_automatic_captions(self, *args, **kwargs):
3747 if (self.get_param('writeautomaticsub', False)
3748 or self.get_param('listsubtitles')):
3749 return self._get_automatic_captions(*args, **kwargs)
3752 def _get_automatic_captions(self, *args, **kwargs):
3753 raise NotImplementedError('This method must be implemented by subclasses')
3755 @functools.cached_property
3756 def _cookies_passed(self):
3757 """Whether cookies have been passed to YoutubeDL"""
3758 return self.get_param('cookiefile') is not None or self.get_param('cookiesfrombrowser') is not None
3760 def mark_watched(self, *args, **kwargs):
3761 if not self.get_param('mark_watched', False):
3763 if self.supports_login() and self._get_login_info()[0] is not None or self._cookies_passed:
3764 self._mark_watched(*args, **kwargs)
3766 def _mark_watched(self, *args, **kwargs):
3767 raise NotImplementedError('This method must be implemented by subclasses')
3769 def geo_verification_headers(self):
3771 geo_verification_proxy = self.get_param('geo_verification_proxy')
3772 if geo_verification_proxy:
3773 headers['Ytdl-request-proxy'] = geo_verification_proxy
3777 def _generic_id(url):
3778 return urllib.parse.unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
3780 def _generic_title(self, url='', webpage='', *, default=None):
3781 return (self._og_search_title(webpage, default=None)
3782 or self._html_extract_title(webpage, default=None)
3783 or urllib.parse.unquote(os.path.splitext(url_basename(url))[0])
3786 def _extract_chapters_helper(self, chapter_list, start_function, title_function, duration, strict=True):
3790 'start_time': start_function(chapter),
3791 'title': title_function(chapter),
3792 } for chapter in chapter_list or []]
3794 warn = self.report_warning
3796 warn = self.write_debug
3797 chapter_list.sort(key=lambda c: c['start_time'] or 0)
3799 chapters = [{'start_time': 0}]
3800 for idx, chapter in enumerate(chapter_list):
3801 if chapter['start_time'] is None:
3802 warn(f'Incomplete chapter {idx}')
3803 elif chapters[-1]['start_time'] <= chapter['start_time'] <= duration:
3804 chapters.append(chapter)
3805 elif chapter not in chapters:
3806 issue = (f'{chapter["start_time"]} > {duration}' if chapter['start_time'] > duration
3807 else f'{chapter["start_time"]} < {chapters[-1]["start_time"]}')
3808 warn(f'Invalid start time ({issue}) for chapter "{chapter["title"]}"')
3811 def _extract_chapters_from_description(self, description, duration):
3812 duration_re = r'(?:\d+:)?\d{1,2}:\d{2}'
3813 sep_re = r'(?m)^\s*(%s)\b\W*\s(%s)\s*$'
3814 return self._extract_chapters_helper(
3815 re.findall(sep_re % (duration_re, r'.+?'), description or ''),
3816 start_function=lambda x: parse_duration(x[0]), title_function=lambda x: x[1],
3817 duration=duration, strict=False) or self._extract_chapters_helper(
3818 re.findall(sep_re % (r'.+?', duration_re), description or ''),
3819 start_function=lambda x: parse_duration(x[1]), title_function=lambda x: x[0],
3820 duration=duration, strict=False)
3823 def _availability(is_private=None, needs_premium=None, needs_subscription=None, needs_auth=None, is_unlisted=None):
3824 all_known = all(map(
3825 lambda x: x is not None,
3826 (is_private, needs_premium, needs_subscription, needs_auth, is_unlisted)))
3828 'private' if is_private
3829 else 'premium_only' if needs_premium
3830 else 'subscriber_only' if needs_subscription
3831 else 'needs_auth' if needs_auth
3832 else 'unlisted' if is_unlisted
3833 else 'public' if all_known
3836 def _configuration_arg(self, key, default=NO_DEFAULT, *, ie_key=None, casesense=False):
3838 @returns A
list of values
for the extractor argument given by
"key"
3839 or "default" if no such key
is present
3840 @param default The default value to
return when the key
is not present (default
: [])
3841 @param casesense When false
, the values are converted to lower case
3843 ie_key = ie_key if isinstance(ie_key, str) else (ie_key or self).ie_key()
3844 val = traverse_obj(self._downloader.params, ('extractor_args', ie_key.lower(), key))
3846 return [] if default is NO_DEFAULT else default
3847 return list(val) if casesense else [x.lower() for x in val]
3849 def _yes_playlist(self, playlist_id, video_id, smuggled_data=None, *, playlist_label='playlist', video_label='video'):
3850 if not playlist_id or not video_id:
3853 no_playlist = (smuggled_data or {}).get('force_noplaylist')
3854 if no_playlist is not None:
3855 return not no_playlist
3857 video_id = '' if video_id is True else f' {video_id}'
3858 playlist_id = '' if playlist_id is True else f' {playlist_id}'
3859 if self.get_param('noplaylist'):
3860 self.to_screen(f'Downloading just the {video_label}{video_id} because of --no-playlist')
3862 self.to_screen(f'Downloading {playlist_label}{playlist_id} - add --no-playlist to download just the {video_label}{video_id}')
3865 def _error_or_warning(self, err, _count=None, _retries=0, *, fatal=True):
3866 RetryManager.report_retry(
3867 err, _count or int(fatal), _retries,
3868 info=self.to_screen, warn=self.report_warning, error=None if fatal else self.report_warning,
3869 sleep_func=self.get_param('retry_sleep_functions', {}).get('extractor'))
3871 def RetryManager(self, **kwargs):
3872 return RetryManager(self.get_param('extractor_retries', 3), self._error_or_warning, **kwargs)
3874 def _extract_generic_embeds(self, url, *args, info_dict={}, note='Extracting generic embeds', **kwargs):
3875 display_id = traverse_obj(info_dict, 'display_id', 'id')
3876 self.to_screen(f'{format_field(display_id, None, "%s: ")}{note}')
3877 return self._downloader.get_info_extractor('Generic')._extract_embeds(
3878 smuggle_url(url, {'block_ies': [self.ie_key()]}), *args, **kwargs)
3881 def extract_from_webpage(cls, ydl, url, webpage):
3882 ie = (cls if isinstance(cls._extract_from_webpage, types.MethodType)
3883 else ydl.get_info_extractor(cls.ie_key()))
3884 for info in ie._extract_from_webpage(url, webpage) or []:
3885 # url = None since we do not want to set (webpage/original)_url
3886 ydl.add_default_extra_info(info, ie, None)
3890 def _extract_from_webpage(cls, url, webpage):
3891 for embed_url in orderedSet(
3892 cls._extract_embed_urls(url, webpage) or [], lazy=True):
3893 yield cls.url_result(embed_url, None if cls._VALID_URL is False else cls)
3896 def _extract_embed_urls(cls, url, webpage):
3897 """@returns all the embed urls on the webpage"""
3898 if '_EMBED_URL_RE' not in cls.__dict__:
3899 assert isinstance(cls._EMBED_REGEX, (list, tuple))
3900 for idx, regex in enumerate(cls._EMBED_REGEX):
3901 assert regex.count('(?P<url>') == 1, \
3902 f'{cls.__name__}._EMBED_REGEX[{idx}] must have exactly 1 url group\n\t{regex}'
3903 cls._EMBED_URL_RE = tuple(map(re.compile, cls._EMBED_REGEX))
3905 for regex in cls._EMBED_URL_RE:
3906 for mobj in regex.finditer(webpage):
3907 embed_url = urllib.parse.urljoin(url, unescapeHTML(mobj.group('url')))
3908 if cls._VALID_URL is False or cls.suitable(embed_url):
3911 class StopExtraction(Exception):
3915 def _extract_url(cls, webpage): # TODO: Remove
3916 """Only for compatibility with some older extractors"""
3917 return next(iter(cls._extract_embed_urls(None, webpage) or []), None)
3920 def __init_subclass__(cls, *, plugin_name=None, **kwargs):
3922 mro = inspect.getmro(cls)
3923 super_class = cls.__wrapped__ = mro[mro.index(cls) + 1]
3924 cls.PLUGIN_NAME, cls.ie_key = plugin_name, super_class.ie_key
3925 cls.IE_NAME = f'{super_class.IE_NAME}+{plugin_name}'
3926 while getattr(super_class, '__wrapped__', None):
3927 super_class = super_class.__wrapped__
3928 setattr(sys.modules[super_class.__module__], super_class.__name__, cls)
3929 _PLUGIN_OVERRIDES[super_class].append(cls)
3931 return super().__init_subclass__(**kwargs)
3934 class SearchInfoExtractor(InfoExtractor):
3936 Base class for paged search queries extractors.
3937 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
3938 Instances should define _SEARCH_KEY and optionally _MAX_RESULTS
3941 _MAX_RESULTS = float('inf')
3942 _RETURN_TYPE = 'playlist'
3945 def _VALID_URL(cls):
3946 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
3948 def _real_extract(self, query):
3949 prefix, query = self._match_valid_url(query).group('prefix', 'query')
3951 return self._get_n_results(query, 1)
3952 elif prefix == 'all':
3953 return self._get_n_results(query, self._MAX_RESULTS)
3957 raise ExtractorError(f'invalid download number {n} for query "{query}"')
3958 elif n > self._MAX_RESULTS:
3959 self.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3960 n = self._MAX_RESULTS
3961 return self._get_n_results(query, n)
3963 def _get_n_results(self, query, n):
3964 """Get a specified number of results for a query.
3965 Either this function or _search_results must be overridden by subclasses """
3966 return self.playlist_result(
3967 itertools.islice(self._search_results(query), 0, None if n == float('inf') else n),
3970 def _search_results(self, query):
3971 """Returns an iterator of search results"""
3972 raise NotImplementedError('This method must be implemented by subclasses')
3975 def SEARCH_KEY(cls):
3976 return cls._SEARCH_KEY
3979 class UnsupportedURLIE(InfoExtractor):
3984 def _real_extract(self, url):
3985 raise UnsupportedError(url)
3988 _PLUGIN_OVERRIDES = collections.defaultdict(list)