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
,
961 encoding
=encoding
, data
=data
)
962 return (content
, urlh
)
965 def _guess_encoding_from_content(content_type
, webpage_bytes
):
966 m
= re
.match(r
'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type
)
968 encoding
= m
.group(1)
970 m
= re
.search(br
'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
971 webpage_bytes[:1024])
973 encoding = m.group(1).decode('ascii')
974 elif webpage_bytes.startswith(b'\xff\xfe'):
981 def __check_blocked(self, content):
982 first_block = content[:512]
983 if ('<title>Access to this site is blocked</title>' in content
984 and 'Websense' in first_block):
985 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
986 blocked_iframe = self._html_search_regex(
987 r'<iframe src="([^
"]+)"', content,
988 'Websense information URL
', default=None)
990 msg += ' Visit
%s for more details
' % blocked_iframe
991 raise ExtractorError(msg, expected=True)
992 if '<title
>The URL you requested has been blocked
</title
>' in first_block:
994 'Access to this webpage has been blocked by Indian censorship
. '
995 'Use a VPN
or proxy
server (with --proxy
) to route around it
.')
996 block_msg = self._html_search_regex(
997 r'</h1
><p
>(.*?
)</p
>',
998 content, 'block message
', default=None)
1000 msg += ' (Message
: "%s")' % block_msg.replace('\n', ' ')
1001 raise ExtractorError(msg, expected=True)
1002 if ('<title
>TTK
:: Доступ к ресурсу ограничен
</title
>' in content
1003 and 'blocklist
.rkn
.gov
.ru
' in content):
1004 raise ExtractorError(
1005 'Access to this webpage has been blocked by decision of the Russian government
. '
1006 'Visit http
://blocklist
.rkn
.gov
.ru
/ for a block reason
.',
1009 def _request_dump_filename(self, url, video_id, data=None):
1010 if data is not None:
1011 data = hashlib.md5(data).hexdigest()
1012 basen = join_nonempty(video_id, data, url, delim='_
')
1013 trim_length = self.get_param('trim_file_name
') or 240
1014 if len(basen) > trim_length:
1015 h = '___
' + hashlib.md5(basen.encode('utf
-8')).hexdigest()
1016 basen = basen[:trim_length - len(h)] + h
1017 filename = sanitize_filename(f'{basen}
.dump
', restricted=True)
1018 # Working around MAX_PATH limitation on Windows (see
1019 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
1020 if compat_os_name == 'nt
':
1021 absfilepath = os.path.abspath(filename)
1022 if len(absfilepath) > 259:
1023 filename = fR'\\?\{absfilepath}
'
1026 def __decode_webpage(self, webpage_bytes, encoding, headers):
1028 encoding = self._guess_encoding_from_content(headers.get('Content
-Type
', ''), webpage_bytes)
1030 return webpage_bytes.decode(encoding, 'replace
')
1032 return webpage_bytes.decode('utf
-8', 'replace
')
1034 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True,
1035 prefix=None, encoding=None, data=None):
1036 webpage_bytes = urlh.read()
1037 if prefix is not None:
1038 webpage_bytes = prefix + webpage_bytes
1039 if self.get_param('dump_intermediate_pages
', False):
1040 self.to_screen('Dumping request to
' + urlh.url)
1041 dump = base64.b64encode(webpage_bytes).decode('ascii
')
1042 self._downloader.to_screen(dump)
1043 if self.get_param('write_pages
'):
1044 if isinstance(url_or_request, Request):
1045 data = self._create_request(url_or_request, data).data
1046 filename = self._request_dump_filename(urlh.url, video_id, data)
1047 self.to_screen(f'Saving request to {filename}
')
1048 with open(filename, 'wb
') as outf:
1049 outf.write(webpage_bytes)
1051 content = self.__decode_webpage(webpage_bytes, encoding, urlh.headers)
1052 self.__check_blocked(content)
1056 def __print_error(self, errnote, fatal, video_id, err):
1058 raise ExtractorError(f'{video_id}
: {errnote}
', cause=err)
1060 self.report_warning(f'{video_id}
: {errnote}
: {err}
')
1062 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True, errnote=None):
1063 if transform_source:
1064 xml_string = transform_source(xml_string)
1066 return compat_etree_fromstring(xml_string.encode('utf
-8'))
1067 except xml.etree.ElementTree.ParseError as ve:
1068 self.__print_error('Failed to parse XML
' if errnote is None else errnote, fatal, video_id, ve)
1070 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True, errnote=None, **parser_kwargs):
1073 json_string, cls=LenientJSONDecoder, strict=False, transform_source=transform_source, **parser_kwargs)
1074 except ValueError as ve:
1075 self.__print_error('Failed to parse JSON
' if errnote is None else errnote, fatal, video_id, ve)
1077 def _parse_socket_response_as_json(self, data, *args, **kwargs):
1078 return self._parse_json(data[data.find('{'):data.rfind('}
') + 1], *args, **kwargs)
1080 def __create_download_methods(name, parser, note, errnote, return_value):
1082 def parse(ie, content, *args, errnote=errnote, **kwargs):
1085 if errnote is False:
1086 kwargs['errnote
'] = errnote
1087 # parser is fetched by name so subclasses can override it
1088 return getattr(ie, parser)(content, *args, **kwargs)
1090 def download_handle(self, url_or_request, video_id, note=note, errnote=errnote, transform_source=None,
1091 fatal=True, encoding=None, data=None, headers={}, query={}, expected_status=None,
1092 impersonate=None, require_impersonation=False):
1093 res = self._download_webpage_handle(
1094 url_or_request, video_id, note=note, errnote=errnote, fatal=fatal, encoding=encoding,
1095 data=data, headers=headers, query=query, expected_status=expected_status,
1096 impersonate=impersonate, require_impersonation=require_impersonation)
1100 return parse(self, content, video_id, transform_source=transform_source, fatal=fatal, errnote=errnote), urlh
1102 def download_content(self, url_or_request, video_id, note=note, errnote=errnote, transform_source=None,
1103 fatal=True, encoding=None, data=None, headers={}, query={}, expected_status=None,
1104 impersonate=None, require_impersonation=False):
1105 if self.get_param('load_pages
'):
1106 url_or_request = self._create_request(url_or_request, data, headers, query)
1107 filename = self._request_dump_filename(url_or_request.url, video_id, url_or_request.data)
1108 self.to_screen(f'Loading request
from {filename}
')
1110 with open(filename, 'rb
') as dumpf:
1111 webpage_bytes = dumpf.read()
1112 except OSError as e:
1113 self.report_warning(f'Unable to load request
from disk
: {e}
')
1115 content = self.__decode_webpage(webpage_bytes, encoding, url_or_request.headers)
1116 return parse(self, content, video_id, transform_source=transform_source, fatal=fatal, errnote=errnote)
1120 'transform_source
': transform_source,
1122 'encoding
': encoding,
1126 'expected_status
': expected_status,
1127 'impersonate
': impersonate,
1128 'require_impersonation
': require_impersonation,
1131 kwargs.pop('transform_source
')
1132 # The method is fetched by name so subclasses can override _download_..._handle
1133 res = getattr(self, download_handle.__name__)(url_or_request, video_id, **kwargs)
1134 return res if res is False else res[0]
1136 def impersonate(func, name, return_value):
1137 func.__name__, func.__qualname__ = name, f'InfoExtractor
.{name}
'
1139 @param transform_source Apply this transformation before parsing
1140 @returns {return_value}
1142 See _download_webpage_handle docstring for other arguments specification
1145 impersonate(download_handle, f'_download_{name}_handle
', f'({return_value}
, URL handle
)')
1146 impersonate(download_content, f'_download_{name}
', f'{return_value}
')
1147 return download_handle, download_content
1149 _download_xml_handle, _download_xml = __create_download_methods(
1150 'xml
', '_parse_xml
', 'Downloading XML
', 'Unable to download XML
', 'xml
as an xml
.etree
.ElementTree
.Element
')
1151 _download_json_handle, _download_json = __create_download_methods(
1152 'json
', '_parse_json
', 'Downloading JSON metadata
', 'Unable to download JSON metadata
', 'JSON
object as a
dict')
1153 _download_socket_json_handle, _download_socket_json = __create_download_methods(
1154 'socket_json
', '_parse_socket_response_as_json
', 'Polling socket
', 'Unable to poll socket
', 'JSON
object as a
dict')
1155 __download_webpage = __create_download_methods('webpage
', None, None, None, 'data of the page
as a string
')[1]
1157 def _download_webpage(
1158 self, url_or_request, video_id, note=None, errnote=None,
1159 fatal=True, tries=1, timeout=NO_DEFAULT, *args, **kwargs):
1161 Return the data of the page as a string.
1164 tries -- number of tries
1165 timeout -- sleep interval between tries
1167 See _download_webpage_handle docstring for other arguments specification.
1170 R''' # NB: These are unused; should they be deprecated?
1172 self._downloader.deprecation_warning('tries argument
is deprecated
in InfoExtractor
._download
_webpage
')
1173 if timeout is NO_DEFAULT:
1176 self._downloader.deprecation_warning('timeout argument
is deprecated
in InfoExtractor
._download
_webpage
')
1182 return self.__download_webpage(url_or_request, video_id, note, errnote, None, fatal, *args, **kwargs)
1183 except IncompleteRead as e:
1185 if try_count >= tries:
1187 self._sleep(timeout, video_id)
1189 def report_warning(self, msg, video_id=None, *args, only_once=False, **kwargs):
1190 idstr = format_field(video_id, None, '%s: ')
1191 msg = f'[{self.IE_NAME}
] {idstr}{msg}
'
1193 if f'WARNING
: {msg}
' in self._printed_messages:
1195 self._printed_messages.add(f'WARNING
: {msg}
')
1196 self._downloader.report_warning(msg, *args, **kwargs)
1198 def to_screen(self, msg, *args, **kwargs):
1199 """Print msg to screen, prefixing it with '[ie_name
]'"""
1200 self._downloader.to_screen(f'[{self.IE_NAME}
] {msg}
', *args, **kwargs)
1202 def write_debug(self, msg, *args, **kwargs):
1203 self._downloader.write_debug(f'[{self.IE_NAME}
] {msg}
', *args, **kwargs)
1205 def get_param(self, name, default=None, *args, **kwargs):
1206 if self._downloader:
1207 return self._downloader.params.get(name, default, *args, **kwargs)
1210 def report_drm(self, video_id, partial=NO_DEFAULT):
1211 if partial is not NO_DEFAULT:
1212 self._downloader.deprecation_warning('InfoExtractor
.report_drm no longer accepts the argument partial
')
1213 self.raise_no_formats('This video
is DRM protected
', expected=True, video_id=video_id)
1215 def report_extraction(self, id_or_name):
1216 """Report information extraction."""
1217 self.to_screen('%s: Extracting information
' % id_or_name)
1219 def report_download_webpage(self, video_id):
1220 """Report webpage download."""
1221 self.to_screen('%s: Downloading webpage
' % video_id)
1223 def report_age_confirmation(self):
1224 """Report attempt to confirm age."""
1225 self.to_screen('Confirming age
')
1227 def report_login(self):
1228 """Report attempt to log in."""
1229 self.to_screen('Logging
in')
1231 def raise_login_required(
1232 self, msg='This video
is only available
for registered users
',
1233 metadata_available=False, method=NO_DEFAULT):
1234 if metadata_available and (
1235 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1236 self.report_warning(msg)
1238 msg += format_field(self._login_hint(method), None, '. %s')
1239 raise ExtractorError(msg, expected=True)
1241 def raise_geo_restricted(
1242 self, msg='This video
is not available
from your location due to geo restriction
',
1243 countries=None, metadata_available=False):
1244 if metadata_available and (
1245 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1246 self.report_warning(msg)
1248 raise GeoRestrictedError(msg, countries=countries)
1250 def raise_no_formats(self, msg, expected=False, video_id=None):
1252 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1253 self.report_warning(msg, video_id)
1254 elif isinstance(msg, ExtractorError):
1257 raise ExtractorError(msg, expected=expected, video_id=video_id)
1259 # Methods for following #608
1261 def url_result(url, ie=None, video_id=None, video_title=None, *, url_transparent=False, **kwargs):
1262 """Returns a URL that points to a page that should be processed"""
1264 kwargs['ie_key
'] = ie if isinstance(ie, str) else ie.ie_key()
1265 if video_id is not None:
1266 kwargs['id'] = video_id
1267 if video_title is not None:
1268 kwargs['title
'] = video_title
1271 '_type
': 'url_transparent
' if url_transparent else 'url
',
1276 def playlist_from_matches(cls, matches, playlist_id=None, playlist_title=None,
1277 getter=IDENTITY, ie=None, video_kwargs=None, **kwargs):
1278 return cls.playlist_result(
1279 (cls.url_result(m, ie, **(video_kwargs or {})) for m in orderedSet(map(getter, matches), lazy=True)),
1280 playlist_id, playlist_title, **kwargs)
1283 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None, *, multi_video=False, **kwargs):
1284 """Returns a playlist"""
1286 kwargs['id'] = playlist_id
1288 kwargs['title
'] = playlist_title
1289 if playlist_description is not None:
1290 kwargs['description
'] = playlist_description
1293 '_type
': 'multi_video
' if multi_video else 'playlist
',
1297 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1299 Perform a regex search on the given string, using a single or a list of
1300 patterns returning the first matching group.
1301 In case of failure return a default value or raise a WARNING or a
1302 RegexNotFoundError, depending on fatal, specifying the field name.
1306 elif isinstance(pattern, (str, re.Pattern)):
1307 mobj = re.search(pattern, string, flags)
1310 mobj = re.search(p, string, flags)
1314 _name = self._downloader._format_err(name, self._downloader.Styles.EMPHASIS)
1318 # return the first matching group
1319 return next(g for g in mobj.groups() if g is not None)
1320 elif isinstance(group, (list, tuple)):
1321 return tuple(mobj.group(g) for g in group)
1323 return mobj.group(group)
1324 elif default is not NO_DEFAULT:
1327 raise RegexNotFoundError('Unable to extract
%s' % _name)
1329 self.report_warning('unable to extract
%s' % _name + bug_reports_message())
1332 def _search_json(self, start_pattern, string, name, video_id, *, end_pattern='',
1333 contains_pattern=r'{(?s:.+)}
', fatal=True, default=NO_DEFAULT, **kwargs):
1334 """Searches string for the JSON object specified by start_pattern"""
1335 # NB: end_pattern is only used to reduce the size of the initial match
1336 if default is NO_DEFAULT:
1337 default, has_default = {}, False
1339 fatal, has_default = False, True
1341 json_string = self._search_regex(
1342 rf'(?
:{start_pattern}
)\s
*(?P
<json
>{contains_pattern}
)\s
*(?
:{end_pattern}
)',
1343 string, name, group='json
', fatal=fatal, default=None if has_default else NO_DEFAULT)
1347 _name = self._downloader._format_err(name, self._downloader.Styles.EMPHASIS)
1349 return self._parse_json(json_string, video_id, ignore_extra=True, **kwargs)
1350 except ExtractorError as e:
1352 raise ExtractorError(
1353 f'Unable to extract {_name}
- Failed to parse JSON
', cause=e.cause, video_id=video_id)
1354 elif not has_default:
1355 self.report_warning(
1356 f'Unable to extract {_name}
- Failed to parse JSON
: {e}
', video_id=video_id)
1359 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1361 Like _search_regex, but strips HTML tags and unescapes entities.
1363 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1364 if isinstance(res, tuple):
1365 return tuple(map(clean_html, res))
1366 return clean_html(res)
1368 def _get_netrc_login_info(self, netrc_machine=None):
1369 netrc_machine = netrc_machine or self._NETRC_MACHINE
1371 cmd = self.get_param('netrc_cmd
')
1373 cmd = cmd.replace('{}', netrc_machine)
1374 self.to_screen(f'Executing command
: {cmd}
')
1375 stdout, _, ret = Popen.run(cmd, text=True, shell=True, stdout=subprocess.PIPE)
1377 raise OSError(f'Command returned error code {ret}
')
1378 info = netrc_from_content(stdout).authenticators(netrc_machine)
1380 elif self.get_param('usenetrc
', False):
1381 netrc_file = compat_expanduser(self.get_param('netrc_location
') or '~
')
1382 if os.path.isdir(netrc_file):
1383 netrc_file = os.path.join(netrc_file, '.netrc
')
1384 info = netrc.netrc(netrc_file).authenticators(netrc_machine)
1389 self.to_screen(f'No authenticators
for {netrc_machine}
')
1392 self.write_debug(f'Using netrc
for {netrc_machine} authentication
')
1393 return info[0], info[2]
1395 def _get_login_info(self, username_option='username
', password_option='password
', netrc_machine=None):
1397 Get the login info as (username, password)
1398 First look for the manually specified credentials using username_option
1399 and password_option as keys in params dictionary. If no such credentials
1400 are available try the netrc_cmd if it is defined or look in the
1401 netrc file using the netrc_machine or _NETRC_MACHINE value.
1402 If there's no info available
, return (None, None)
1405 username = self.get_param(username_option)
1406 if username is not None:
1407 password = self.get_param(password_option)
1410 username, password = self._get_netrc_login_info(netrc_machine)
1411 except (OSError, netrc.NetrcParseError) as err:
1412 self.report_warning(f'Failed to parse .netrc: {err}')
1414 return username, password
1416 def _get_tfa_info(self, note='two-factor verification code'):
1418 Get the two
-factor authentication info
1419 TODO
- asking the user will be required
for sms
/phone verify
1420 currently just uses the command line option
1421 If there
's no info available, return None
1424 tfa = self.get_param('twofactor
')
1428 return getpass.getpass('Type
%s and press
[Return
]: ' % note)
1430 # Helper functions for extracting OpenGraph info
1432 def _og_regexes(prop):
1433 content_re = r'content
=(?
:"([^"]+?
)"|\'([^\']+?)\'|\s*([^\s"\'=<>`
]+?
)(?
=\s|
/?
>))'
1434 property_re = (r'(?
:name|
property)=(?
:\'og
%(sep)s%(prop)s\'|
"og%(sep)s%(prop)s"|\s
*og
%(sep)s%(prop)s\b)'
1435 % {'prop': re.escape(prop), 'sep': '(?::|[:-])'})
1436 template = r'<meta
[^
>]+?
%s[^
>]+?
%s'
1438 template % (property_re, content_re),
1439 template % (content_re, property_re),
1443 def _meta_regex(prop):
1444 return r'''(?isx)<meta
1445 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1446 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1448 def _og_search_property(self, prop, html, name=None, **kargs):
1449 prop = variadic(prop)
1451 name = 'OpenGraph
%s' % prop[0]
1454 og_regexes.extend(self._og_regexes(p))
1455 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1458 return unescapeHTML(escaped)
1460 def _og_search_thumbnail(self, html, **kargs):
1461 return self._og_search_property('image
', html, 'thumbnail URL
', fatal=False, **kargs)
1463 def _og_search_description(self, html, **kargs):
1464 return self._og_search_property('description
', html, fatal=False, **kargs)
1466 def _og_search_title(self, html, *, fatal=False, **kargs):
1467 return self._og_search_property('title
', html, fatal=fatal, **kargs)
1469 def _og_search_video_url(self, html, name='video url
', secure=True, **kargs):
1470 regexes = self._og_regexes('video
') + self._og_regexes('video
:url
')
1472 regexes = self._og_regexes('video
:secure_url
') + regexes
1473 return self._html_search_regex(regexes, html, name, **kargs)
1475 def _og_search_url(self, html, **kargs):
1476 return self._og_search_property('url
', html, **kargs)
1478 def _html_extract_title(self, html, name='title
', *, fatal=False, **kwargs):
1479 return self._html_search_regex(r'(?s
)<title
\b[^
>]*>([^
<]+)</title
>', html, name, fatal=fatal, **kwargs)
1481 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1482 name = variadic(name)
1483 if display_name is None:
1484 display_name = name[0]
1485 return self._html_search_regex(
1486 [self._meta_regex(n) for n in name],
1487 html, display_name, fatal=fatal, group='content
', **kwargs)
1489 def _dc_search_uploader(self, html):
1490 return self._html_search_meta('dc
.creator
', html, 'uploader
')
1493 def _rta_search(html):
1494 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1495 if re.search(r'(?ix
)<meta\s
+name
="rating"\s
+'
1496 r' content
="RTA-5042-1996-1400-1577-RTA"',
1500 # And then there are the jokers who advertise that they use RTA, but actually don't
.
1501 AGE_LIMIT_MARKERS
= [
1502 r
'Proudly Labeled <a href="http://www\.rtalabel\.org/" title="Restricted to Adults">RTA</a>',
1503 r
'>[^<]*you acknowledge you are at least (\d+) years old',
1504 r
'>\s*(?:18\s+U(?:\.S\.C\.|SC)\s+)?(?:§+\s*)?2257\b',
1508 for marker
in AGE_LIMIT_MARKERS
:
1509 mobj
= re
.search(marker
, html
)
1511 age_limit
= max(age_limit
, int(traverse_obj(mobj
, 1, default
=18)))
1514 def _media_rating_search(self
, html
):
1515 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1516 rating
= self
._html
_search
_meta
('rating', html
)
1528 return RATING_TABLE
.get(rating
.lower())
1530 def _family_friendly_search(self
, html
):
1531 # See http://schema.org/VideoObject
1532 family_friendly
= self
._html
_search
_meta
(
1533 'isFamilyFriendly', html
, default
=None)
1535 if not family_friendly
:
1544 return RATING_TABLE
.get(family_friendly
.lower())
1546 def _twitter_search_player(self
, html
):
1547 return self
._html
_search
_meta
('twitter:player', html
,
1548 'twitter card player')
1550 def _yield_json_ld(self
, html
, video_id
, *, fatal
=True, default
=NO_DEFAULT
):
1551 """Yield all json ld objects in the html"""
1552 if default
is not NO_DEFAULT
:
1554 for mobj
in re
.finditer(JSON_LD_RE
, html
):
1555 json_ld_item
= self
._parse
_json
(mobj
.group('json_ld'), video_id
, fatal
=fatal
)
1556 for json_ld
in variadic(json_ld_item
):
1557 if isinstance(json_ld
, dict):
1560 def _search_json_ld(self
, html
, video_id
, expected_type
=None, *, fatal
=True, default
=NO_DEFAULT
):
1561 """Search for a video in any json ld in the html"""
1562 if default
is not NO_DEFAULT
:
1564 info
= self
._json
_ld
(
1565 list(self
._yield
_json
_ld
(html
, video_id
, fatal
=fatal
, default
=default
)),
1566 video_id
, fatal
=fatal
, expected_type
=expected_type
)
1569 if default
is not NO_DEFAULT
:
1572 raise RegexNotFoundError('Unable to extract JSON-LD')
1574 self
.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1577 def _json_ld(self
, json_ld
, video_id
, fatal
=True, expected_type
=None):
1578 if isinstance(json_ld
, str):
1579 json_ld
= self
._parse
_json
(json_ld
, video_id
, fatal
=fatal
)
1584 INTERACTION_TYPE_MAP
= {
1585 'CommentAction': 'comment',
1586 'AgreeAction': 'like',
1587 'DisagreeAction': 'dislike',
1588 'LikeAction': 'like',
1589 'DislikeAction': 'dislike',
1590 'ListenAction': 'view',
1591 'WatchAction': 'view',
1592 'ViewAction': 'view',
1595 def is_type(e
, *expected_types
):
1596 type = variadic(traverse_obj(e
, '@type'))
1597 return any(x
in type for x
in expected_types
)
1599 def extract_interaction_type(e
):
1600 interaction_type
= e
.get('interactionType')
1601 if isinstance(interaction_type
, dict):
1602 interaction_type
= interaction_type
.get('@type')
1603 return str_or_none(interaction_type
)
1605 def extract_interaction_statistic(e
):
1606 interaction_statistic
= e
.get('interactionStatistic')
1607 if isinstance(interaction_statistic
, dict):
1608 interaction_statistic
= [interaction_statistic
]
1609 if not isinstance(interaction_statistic
, list):
1611 for is_e
in interaction_statistic
:
1612 if not is_type(is_e
, 'InteractionCounter'):
1614 interaction_type
= extract_interaction_type(is_e
)
1615 if not interaction_type
:
1617 # For interaction count some sites provide string instead of
1618 # an integer (as per spec) with non digit characters (e.g. ",")
1619 # so extracting count with more relaxed str_to_int
1620 interaction_count
= str_to_int(is_e
.get('userInteractionCount'))
1621 if interaction_count
is None:
1623 count_kind
= INTERACTION_TYPE_MAP
.get(interaction_type
.split('/')[-1])
1626 count_key
= '%s_count' % count_kind
1627 if info
.get(count_key
) is not None:
1629 info
[count_key
] = interaction_count
1631 def extract_chapter_information(e
):
1633 'title': part
.get('name'),
1634 'start_time': part
.get('startOffset'),
1635 'end_time': part
.get('endOffset'),
1636 } for part
in variadic(e
.get('hasPart') or []) if part
.get('@type') == 'Clip']
1637 for idx
, (last_c
, current_c
, next_c
) in enumerate(zip(
1638 [{'end_time': 0}
] + chapters
, chapters
, chapters
[1:])):
1639 current_c
['end_time'] = current_c
['end_time'] or next_c
['start_time']
1640 current_c
['start_time'] = current_c
['start_time'] or last_c
['end_time']
1641 if None in current_c
.values():
1642 self
.report_warning(f
'Chapter {idx} contains broken data. Not extracting chapters')
1645 chapters
[-1]['end_time'] = chapters
[-1]['end_time'] or info
['duration']
1646 info
['chapters'] = chapters
1648 def extract_video_object(e
):
1649 author
= e
.get('author')
1651 'url': url_or_none(e
.get('contentUrl')),
1652 'ext': mimetype2ext(e
.get('encodingFormat')),
1653 'title': unescapeHTML(e
.get('name')),
1654 'description': unescapeHTML(e
.get('description')),
1655 'thumbnails': [{'url': unescapeHTML(url)}
1656 for url
in variadic(traverse_obj(e
, 'thumbnailUrl', 'thumbnailURL'))
1657 if url_or_none(url
)],
1658 'duration': parse_duration(e
.get('duration')),
1659 'timestamp': unified_timestamp(e
.get('uploadDate')),
1660 # author can be an instance of 'Organization' or 'Person' types.
1661 # both types can have 'name' property(inherited from 'Thing' type). [1]
1662 # however some websites are using 'Text' type instead.
1663 # 1. https://schema.org/VideoObject
1664 'uploader': author
.get('name') if isinstance(author
, dict) else author
if isinstance(author
, str) else None,
1665 'artist': traverse_obj(e
, ('byArtist', 'name'), expected_type
=str),
1666 'filesize': int_or_none(float_or_none(e
.get('contentSize'))),
1667 'tbr': int_or_none(e
.get('bitrate')),
1668 'width': int_or_none(e
.get('width')),
1669 'height': int_or_none(e
.get('height')),
1670 'view_count': int_or_none(e
.get('interactionCount')),
1671 'tags': try_call(lambda: e
.get('keywords').split(',')),
1673 if is_type(e
, 'AudioObject'):
1676 'abr': int_or_none(e
.get('bitrate')),
1678 extract_interaction_statistic(e
)
1679 extract_chapter_information(e
)
1681 def traverse_json_ld(json_ld
, at_top_level
=True):
1682 for e
in variadic(json_ld
):
1683 if not isinstance(e
, dict):
1685 if at_top_level
and '@context' not in e
:
1687 if at_top_level
and set(e
.keys()) == {'@context', '@graph'}
:
1688 traverse_json_ld(e
['@graph'], at_top_level
=False)
1690 if expected_type
is not None and not is_type(e
, expected_type
):
1692 rating
= traverse_obj(e
, ('aggregateRating', 'ratingValue'), expected_type
=float_or_none
)
1693 if rating
is not None:
1694 info
['average_rating'] = rating
1695 if is_type(e
, 'TVEpisode', 'Episode'):
1696 episode_name
= unescapeHTML(e
.get('name'))
1698 'episode': episode_name
,
1699 'episode_number': int_or_none(e
.get('episodeNumber')),
1700 'description': unescapeHTML(e
.get('description')),
1702 if not info
.get('title') and episode_name
:
1703 info
['title'] = episode_name
1704 part_of_season
= e
.get('partOfSeason')
1705 if is_type(part_of_season
, 'TVSeason', 'Season', 'CreativeWorkSeason'):
1707 'season': unescapeHTML(part_of_season
.get('name')),
1708 'season_number': int_or_none(part_of_season
.get('seasonNumber')),
1710 part_of_series
= e
.get('partOfSeries') or e
.get('partOfTVSeries')
1711 if is_type(part_of_series
, 'TVSeries', 'Series', 'CreativeWorkSeries'):
1712 info
['series'] = unescapeHTML(part_of_series
.get('name'))
1713 elif is_type(e
, 'Movie'):
1715 'title': unescapeHTML(e
.get('name')),
1716 'description': unescapeHTML(e
.get('description')),
1717 'duration': parse_duration(e
.get('duration')),
1718 'timestamp': unified_timestamp(e
.get('dateCreated')),
1720 elif is_type(e
, 'Article', 'NewsArticle'):
1722 'timestamp': parse_iso8601(e
.get('datePublished')),
1723 'title': unescapeHTML(e
.get('headline')),
1724 'description': unescapeHTML(e
.get('articleBody') or e
.get('description')),
1726 if is_type(traverse_obj(e
, ('video', 0)), 'VideoObject'):
1727 extract_video_object(e
['video'][0])
1728 elif is_type(traverse_obj(e
, ('subjectOf', 0)), 'VideoObject'):
1729 extract_video_object(e
['subjectOf'][0])
1730 elif is_type(e
, 'VideoObject', 'AudioObject'):
1731 extract_video_object(e
)
1732 if expected_type
is None:
1736 video
= e
.get('video')
1737 if is_type(video
, 'VideoObject'):
1738 extract_video_object(video
)
1739 if expected_type
is None:
1744 traverse_json_ld(json_ld
)
1745 return filter_dict(info
)
1747 def _search_nextjs_data(self
, webpage
, video_id
, *, fatal
=True, default
=NO_DEFAULT
, **kw
):
1749 self
._downloader
.deprecation_warning('using `default=\'{}\'` is deprecated, use `default={}` instead')
1751 if default
is not NO_DEFAULT
:
1754 return self
._search
_json
(
1755 r
'<script[^>]+id=[\'"]__NEXT_DATA__[\'"][^
>]*>', webpage, 'next
.js data
',
1756 video_id, end_pattern='</script
>', fatal=fatal, default=default, **kw)
1758 def _search_nuxt_data(self, webpage, video_id, context_name='__NUXT__
', *, fatal=True, traverse=('data
', 0)):
1759 """Parses Nuxt.js metadata. This works as long as the function __NUXT__ invokes is a pure function"""
1760 rectx = re.escape(context_name)
1761 FUNCTION_RE = r'\
(function\
((?P
<arg_keys
>.*?
)\
){.*?\breturn\s+(?P<js>{.*?}
)\s
*;?\s
*}\
((?P
<arg_vals
>.*?
)\
)'
1762 js, arg_keys, arg_vals = self._search_regex(
1763 (rf'<script
>\s
*window\
.{rectx}
={FUNCTION_RE}\s
*\
)\s
*;?\s
*</script
>', rf'{rectx}\
(.*?{FUNCTION_RE}
'),
1764 webpage, context_name, group=('js
', 'arg_keys
', 'arg_vals
'),
1765 default=NO_DEFAULT if fatal else (None, None, None))
1769 args = dict(zip(arg_keys.split(','), map(json.dumps, self._parse_json(
1770 f'[{arg_vals}
]', video_id, transform_source=js_to_json, fatal=fatal) or ())))
1772 ret = self._parse_json(js, video_id, transform_source=functools.partial(js_to_json, vars=args), fatal=fatal)
1773 return traverse_obj(ret, traverse) or {}
1776 def _hidden_inputs(html):
1777 html = re.sub(r'<!--(?
:(?
!<!--).)*-->', '', html)
1779 for input in re.findall(r'(?i
)(<input[^
>]+>)', html):
1780 attrs = extract_attributes(input)
1783 if attrs.get('type') not in ('hidden
', 'submit
'):
1785 name = attrs.get('name
') or attrs.get('id')
1786 value = attrs.get('value
')
1787 if name and value is not None:
1788 hidden_inputs[name] = value
1789 return hidden_inputs
1791 def _form_hidden_inputs(self, form_id, html):
1792 form = self._search_regex(
1793 r'(?
is)<form
[^
>]+?
id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1794 html, '%s form' % form_id, group='form')
1795 return self._hidden_inputs(form)
1797 @classproperty(cache=True)
1798 def FormatSort(cls):
1799 class FormatSort(FormatSorter):
1800 def __init__(ie, *args, **kwargs):
1801 super().__init__(ie._downloader, *args, **kwargs)
1803 deprecation_warning(
1804 'yt_dlp.InfoExtractor.FormatSort is deprecated and may be removed in the future. '
1805 'Use yt_dlp.utils.FormatSorter instead')
1808 def _sort_formats(self, formats, field_preference=[]):
1809 if not field_preference:
1810 self._downloader.deprecation_warning(
1811 'yt_dlp.InfoExtractor._sort_formats is deprecated and is no longer required')
1813 self._downloader.deprecation_warning(
1814 'yt_dlp.InfoExtractor._sort_formats is deprecated and no longer works as expected. '
1815 'Return _format_sort_fields in the info_dict instead')
1817 formats[0]['__sort_fields'] = field_preference
1819 def _check_formats(self, formats, video_id):
1821 formats[:] = filter(
1822 lambda f: self._is_valid_url(
1824 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1828 def _remove_duplicate_formats(formats):
1832 if f['url'] not in format_urls:
1833 format_urls.add(f['url'])
1834 unique_formats.append(f)
1835 formats[:] = unique_formats
1837 def _is_valid_url(self, url, video_id, item='video', headers={}):
1838 url = self._proto_relative_url(url, scheme='http:')
1839 # For now assume non HTTP(S) URLs always valid
1840 if not (url.startswith('http://') or url.startswith('https://')):
1843 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1845 except ExtractorError as e:
1847 '%s: %s URL is invalid, skipping: %s'
1848 % (video_id, item, error_to_compat_str(e.cause)))
1851 def http_scheme(self):
1852 """ Either "http
:" or "https
:", depending on the user's preferences """
1855 if self.get_param('prefer_insecure', False)
1858 def _proto_relative_url(self, url, scheme=None):
1859 scheme = scheme or self.http_scheme()
1860 assert scheme.endswith(':')
1861 return sanitize_url(url, scheme=scheme[:-1])
1863 def _sleep(self, timeout, video_id, msg_template=None):
1864 if msg_template is None:
1865 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1866 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1870 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1871 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1872 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
1873 if self.get_param('ignore_no_formats_error'):
1876 res = self._download_xml_handle(
1877 manifest_url, video_id, 'Downloading f4m manifest',
1878 'Unable to download f4m manifest',
1879 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1880 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
1881 transform_source=transform_source,
1882 fatal=fatal, data=data, headers=headers, query=query)
1886 manifest, urlh = res
1887 manifest_url = urlh.url
1889 return self._parse_f4m_formats(
1890 manifest, manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1891 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1893 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1894 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1895 fatal=True, m3u8_id=None):
1896 if not isinstance(manifest, xml.etree.ElementTree.Element) and not fatal:
1899 # currently yt-dlp cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1900 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1901 if akamai_pv is not None and ';' in akamai_pv.text:
1902 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1903 if playerVerificationChallenge.strip() != '':
1907 manifest_version = '1.0'
1908 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1910 manifest_version = '2.0'
1911 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1912 # Remove unsupported DRM protected media from final formats
1913 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
1914 media_nodes = remove_encrypted_media(media_nodes)
1918 manifest_base_url = get_base_url(manifest)
1920 bootstrap_info = xpath_element(
1921 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1922 'bootstrap info', default=None)
1925 mime_type = xpath_text(
1926 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1927 'base URL', default=None)
1928 if mime_type and mime_type.startswith('audio/'):
1931 for i, media_el in enumerate(media_nodes):
1932 tbr = int_or_none(media_el.attrib.get('bitrate'))
1933 width = int_or_none(media_el.attrib.get('width'))
1934 height = int_or_none(media_el.attrib.get('height'))
1935 format_id = join_nonempty(f4m_id, tbr or i)
1936 # If <bootstrapInfo> is present, the specified f4m is a
1937 # stream-level manifest, and only set-level manifests may refer to
1938 # external resources. See section 11.4 and section 4 of F4M spec
1939 if bootstrap_info is None:
1941 # @href is introduced in 2.0, see section 11.6 of F4M spec
1942 if manifest_version == '2.0':
1943 media_url = media_el.attrib.get('href')
1944 if media_url is None:
1945 media_url = media_el.attrib.get('url')
1949 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1950 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1951 # If media_url is itself a f4m manifest do the recursive extraction
1952 # since bitrates in parent manifest (this one) and media_url manifest
1953 # may differ leading to inability to resolve the format by requested
1954 # bitrate in f4m downloader
1955 ext = determine_ext(manifest_url)
1957 f4m_formats = self._extract_f4m_formats(
1958 manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1959 transform_source=transform_source, fatal=fatal)
1960 # Sometimes stream-level manifest contains single media entry that
1961 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1962 # At the same time parent's media entry in set-level manifest may
1963 # contain it. We will copy it from parent in such cases.
1964 if len(f4m_formats) == 1:
1967 'tbr': f.get('tbr') or tbr,
1968 'width': f.get('width') or width,
1969 'height': f.get('height') or height,
1970 'format_id': f.get('format_id') if not tbr else format_id,
1973 formats.extend(f4m_formats)
1976 formats.extend(self._extract_m3u8_formats(
1977 manifest_url, video_id, 'mp4', preference=preference,
1978 quality=quality, m3u8_id=m3u8_id, fatal=fatal))
1981 'format_id': format_id,
1982 'url': manifest_url,
1983 'manifest_url': manifest_url,
1984 'ext': 'flv' if bootstrap_info is not None else None,
1990 'preference': preference,
1995 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, quality=None, m3u8_id=None):
1997 'format_id': join_nonempty(m3u8_id, 'meta'),
2001 'preference': preference - 100 if preference else -100,
2003 'resolution': 'multiple',
2004 'format_note': 'Quality selection URL',
2007 def _report_ignoring_subs(self, name):
2008 self.report_warning(bug_reports_message(
2009 f'Ignoring subtitle tracks found in the {name} manifest; '
2010 'if any subtitle tracks are missing,'
2013 def _extract_m3u8_formats(self, *args, **kwargs):
2014 fmts, subs = self._extract_m3u8_formats_and_subtitles(*args, **kwargs)
2016 self._report_ignoring_subs('HLS')
2019 def _extract_m3u8_formats_and_subtitles(
2020 self, m3u8_url, video_id, ext=None, entry_protocol='m3u8_native',
2021 preference=None, quality=None, m3u8_id=None, note=None,
2022 errnote=None, fatal=True, live=False, data=None, headers={},
2025 if self.get_param('ignore_no_formats_error'):
2029 if errnote is not False:
2030 errnote = errnote or 'Failed to obtain m3u8 URL'
2032 raise ExtractorError(errnote, video_id=video_id)
2033 self.report_warning(f'{errnote}{bug_reports_message()}')
2036 res = self._download_webpage_handle(
2038 note='Downloading m3u8 information' if note is None else note,
2039 errnote='Failed to download m3u8 information' if errnote is None else errnote,
2040 fatal=fatal, data=data, headers=headers, query=query)
2045 m3u8_doc, urlh = res
2048 return self._parse_m3u8_formats_and_subtitles(
2049 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
2050 preference=preference, quality=quality, m3u8_id=m3u8_id,
2051 note=note, errnote=errnote, fatal=fatal, live=live, data=data,
2052 headers=headers, query=query, video_id=video_id)
2054 def _parse_m3u8_formats_and_subtitles(
2055 self, m3u8_doc, m3u8_url=None, ext=None, entry_protocol='m3u8_native',
2056 preference=None, quality=None, m3u8_id=None, live=False, note=None,
2057 errnote=None, fatal=True, data=None, headers={}, query={},
2059 formats, subtitles = [], {}
2060 has_drm = HlsFD._has_drm(m3u8_doc)
2062 def format_url(url):
2063 return url if re.match(r'^https?://', url) else urllib.parse.urljoin(m3u8_url, url)
2065 if self.get_param('hls_split_discontinuity', False):
2066 def _extract_m3u8_playlist_indices(manifest_url=None, m3u8_doc=None):
2068 if not manifest_url:
2070 m3u8_doc = self._download_webpage(
2071 manifest_url, video_id, fatal=fatal, data=data, headers=headers,
2072 note=False, errnote='Failed to download m3u8 playlist information')
2073 if m3u8_doc is False:
2075 return range(1 + sum(line.startswith('#EXT-X-DISCONTINUITY') for line in m3u8_doc.splitlines()))
2078 def _extract_m3u8_playlist_indices(*args, **kwargs):
2082 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
2083 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
2084 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
2086 # We should try extracting formats only from master playlists [1, 4.3.4],
2087 # i.e. playlists that describe available qualities. On the other hand
2088 # media playlists [1, 4.3.3] should be returned as is since they contain
2089 # just the media without qualities renditions.
2090 # Fortunately, master playlist can be easily distinguished from media
2091 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
2092 # master playlist tags MUST NOT appear in a media playlist and vice versa.
2093 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
2094 # media playlist and MUST NOT appear in master playlist thus we can
2095 # clearly detect media playlist with this criterion.
2097 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
2099 'format_id': join_nonempty(m3u8_id, idx),
2100 'format_index': idx,
2101 'url': m3u8_url or encode_data_uri(m3u8_doc.encode('utf-8'), 'application/x-mpegurl'),
2103 'protocol': entry_protocol,
2104 'preference': preference,
2107 } for idx in _extract_m3u8_playlist_indices(m3u8_doc=m3u8_doc)]
2109 return formats, subtitles
2112 last_stream_inf = {}
2114 def extract_media(x_media_line):
2115 media = parse_m3u8_attributes(x_media_line)
2116 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
2117 media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME')
2118 if not (media_type and group_id and name):
2120 groups.setdefault(group_id, []).append(media)
2121 # <https://tools.ietf.org/html/rfc8216#section-4.3.4.1>
2122 if media_type == 'SUBTITLES':
2123 # According to RFC 8216 §4.3.4.2.1, URI is REQUIRED in the
2124 # EXT-X-MEDIA tag if the media type is SUBTITLES.
2125 # However, lack of URI has been spotted in the wild.
2126 # e.g. NebulaIE; see https://github.com/yt-dlp/yt-dlp/issues/339
2127 if not media.get('URI'):
2129 url = format_url(media['URI'])
2132 'ext': determine_ext(url),
2134 if sub_info['ext'] == 'm3u8':
2135 # Per RFC 8216 §3.1, the only possible subtitle format m3u8
2136 # files may contain is WebVTT:
2137 # <https://tools.ietf.org/html/rfc8216#section-3.1>
2138 sub_info['ext'] = 'vtt'
2139 sub_info['protocol'] = 'm3u8_native'
2140 lang = media.get('LANGUAGE') or 'und'
2141 subtitles.setdefault(lang, []).append(sub_info)
2142 if media_type not in ('VIDEO', 'AUDIO'):
2144 media_url = media.get('URI')
2146 manifest_url = format_url(media_url)
2148 'format_id': join_nonempty(m3u8_id, group_id, name, idx),
2149 'format_note': name,
2150 'format_index': idx,
2151 'url': manifest_url,
2152 'manifest_url': m3u8_url,
2153 'language': media.get('LANGUAGE'),
2155 'protocol': entry_protocol,
2156 'preference': preference,
2159 'vcodec': 'none' if media_type == 'AUDIO' else None,
2160 } for idx in _extract_m3u8_playlist_indices(manifest_url))
2162 def build_stream_name():
2163 # Despite specification does not mention NAME attribute for
2164 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
2165 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
2166 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
2167 stream_name = last_stream_inf.get('NAME')
2170 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
2171 # from corresponding rendition group
2172 stream_group_id = last_stream_inf.get('VIDEO')
2173 if not stream_group_id:
2175 stream_group = groups.get(stream_group_id)
2176 if not stream_group:
2177 return stream_group_id
2178 rendition = stream_group[0]
2179 return rendition.get('NAME') or stream_group_id
2181 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
2182 # chance to detect video only formats when EXT-X-STREAM-INF tags
2183 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
2184 for line in m3u8_doc.splitlines():
2185 if line.startswith('#EXT-X-MEDIA:'):
2188 for line in m3u8_doc.splitlines():
2189 if line.startswith('#EXT-X-STREAM-INF:'):
2190 last_stream_inf = parse_m3u8_attributes(line)
2191 elif line.startswith('#') or not line.strip():
2194 tbr = float_or_none(
2195 last_stream_inf.get('AVERAGE-BANDWIDTH')
2196 or last_stream_inf.get('BANDWIDTH'), scale=1000)
2197 manifest_url = format_url(line.strip())
2199 for idx in _extract_m3u8_playlist_indices(manifest_url):
2200 format_id = [m3u8_id, None, idx]
2201 # Bandwidth of live streams may differ over time thus making
2202 # format_id unpredictable. So it's better to keep provided
2205 stream_name = build_stream_name()
2206 format_id[1] = stream_name or '%d' % (tbr or len(formats))
2208 'format_id': join_nonempty(*format_id),
2209 'format_index': idx,
2210 'url': manifest_url,
2211 'manifest_url': m3u8_url,
2214 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')),
2215 'protocol': entry_protocol,
2216 'preference': preference,
2220 resolution = last_stream_inf.get('RESOLUTION')
2222 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
2224 f['width'] = int(mobj.group('width'))
2225 f['height'] = int(mobj.group('height'))
2226 # Unified Streaming Platform
2228 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
2230 abr, vbr = mobj.groups()
2231 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
2236 codecs = parse_codecs(last_stream_inf.get('CODECS'))
2238 audio_group_id = last_stream_inf.get('AUDIO')
2239 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
2240 # references a rendition group MUST have a CODECS attribute.
2241 # However, this is not always respected. E.g. [2]
2242 # contains EXT-X-STREAM-INF tag which references AUDIO
2243 # rendition group but does not have CODECS and despite
2244 # referencing an audio group it represents a complete
2245 # (with audio and video) format. So, for such cases we will
2246 # ignore references to rendition groups and treat them
2247 # as complete formats.
2248 if audio_group_id and codecs and f.get('vcodec') != 'none':
2249 audio_group = groups.get(audio_group_id)
2250 if audio_group and audio_group[0].get('URI'):
2251 # TODO: update acodec for audio only formats with
2253 f['acodec'] = 'none'
2254 if not f.get('ext'):
2255 f['ext'] = 'm4a' if f.get('vcodec') == 'none' else 'mp4'
2259 progressive_uri = last_stream_inf.get('PROGRESSIVE-URI')
2262 del http_f['manifest_url']
2264 'format_id': f['format_id'].replace('hls-', 'http-'),
2266 'url': progressive_uri,
2268 formats.append(http_f)
2270 last_stream_inf = {}
2271 return formats, subtitles
2273 def _extract_m3u8_vod_duration(
2274 self, m3u8_vod_url, video_id, note=None, errnote=None, data=None, headers={}, query={}):
2276 m3u8_vod = self._download_webpage(
2277 m3u8_vod_url, video_id,
2278 note='Downloading m3u8 VOD manifest' if note is None else note,
2279 errnote='Failed to download VOD manifest' if errnote is None else errnote,
2280 fatal=False, data=data, headers=headers, query=query)
2282 return self._parse_m3u8_vod_duration(m3u8_vod or '', video_id)
2284 def _parse_m3u8_vod_duration(self, m3u8_vod, video_id):
2285 if '#EXT-X-ENDLIST' not in m3u8_vod:
2289 float(line[len('#EXTINF:'):].split(',')[0])
2290 for line in m3u8_vod.splitlines() if line.startswith('#EXTINF:'))) or None
2292 def _extract_mpd_vod_duration(
2293 self, mpd_url, video_id, note=None, errnote=None, data=None, headers={}, query={}):
2295 mpd_doc = self._download_xml(
2297 note='Downloading MPD VOD manifest' if note is None else note,
2298 errnote='Failed to download VOD manifest' if errnote is None else errnote,
2299 fatal=False, data=data, headers=headers, query=query)
2300 if not isinstance(mpd_doc, xml.etree.ElementTree.Element):
2302 return int_or_none(parse_duration(mpd_doc.get('mediaPresentationDuration')))
2305 def _xpath_ns(path, namespace=None):
2309 for c in path.split('/'):
2310 if not c or c == '.':
2313 out.append('{%s}%s' % (namespace, c))
2314 return '/'.join(out)
2316 def _extract_smil_formats_and_subtitles(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
2317 if self.get_param('ignore_no_formats_error'):
2320 res = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
2326 return self._parse_smil_formats_and_subtitles(smil, urlh.url, video_id, f4m_params=f4m_params,
2327 namespace=self._parse_smil_namespace(smil))
2329 def _extract_smil_formats(self, *args, **kwargs):
2330 fmts, subs = self._extract_smil_formats_and_subtitles(*args, **kwargs)
2332 self._report_ignoring_subs('SMIL')
2335 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
2336 res = self._download_smil(smil_url, video_id, fatal=fatal)
2343 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
2345 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
2346 return self._download_xml_handle(
2347 smil_url, video_id, 'Downloading SMIL file',
2348 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
2350 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
2351 namespace = self._parse_smil_namespace(smil)
2353 formats, subtitles = self._parse_smil_formats_and_subtitles(
2354 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
2356 video_id = os.path.splitext(url_basename(smil_url))[0]
2360 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
2361 name = meta.attrib.get('name')
2362 content = meta.attrib.get('content')
2363 if not name or not content:
2365 if not title and name == 'title':
2367 elif not description and name in ('description', 'abstract'):
2368 description = content
2369 elif not upload_date and name == 'date':
2370 upload_date = unified_strdate(content)
2373 'id': image.get('type'),
2374 'url': image.get('src'),
2375 'width': int_or_none(image.get('width')),
2376 'height': int_or_none(image.get('height')),
2377 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
2381 'title': title or video_id,
2382 'description': description,
2383 'upload_date': upload_date,
2384 'thumbnails': thumbnails,
2386 'subtitles': subtitles,
2389 def _parse_smil_namespace(self, smil):
2390 return self._search_regex(
2391 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
2393 def _parse_smil_formats(self, *args, **kwargs):
2394 fmts, subs = self._parse_smil_formats_and_subtitles(*args, **kwargs)
2396 self._report_ignoring_subs('SMIL')
2399 def _parse_smil_formats_and_subtitles(
2400 self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
2402 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
2403 b = meta.get('base') or meta.get('httpBase')
2408 formats, subtitles = [], {}
2415 media = itertools.chain.from_iterable(
2416 smil.findall(self._xpath_ns(arg, namespace))
2417 for arg in ['.//video', './/audio', './/media'])
2418 for medium in media:
2419 src = medium.get('src')
2420 if not src or src in srcs:
2424 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
2425 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
2426 width = int_or_none(medium.get('width'))
2427 height = int_or_none(medium.get('height'))
2428 proto = medium.get('proto')
2429 ext = medium.get('ext')
2430 src_ext = determine_ext(src, default_ext=None) or ext or urlhandle_detect_ext(
2431 self._request_webpage(HEADRequest(src), video_id, note='Requesting extension info', fatal=False))
2432 streamer = medium.get('streamer') or base
2434 if proto == 'rtmp' or streamer.startswith('rtmp'):
2440 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
2442 'filesize': filesize,
2446 if transform_rtmp_url:
2447 streamer, src = transform_rtmp_url(streamer, src)
2448 formats[-1].update({
2454 src_url = src if src.startswith('http') else urllib.parse.urljoin(base, src)
2455 src_url = src_url.strip()
2457 if proto == 'm3u8' or src_ext == 'm3u8':
2458 m3u8_formats, m3u8_subs = self._extract_m3u8_formats_and_subtitles(
2459 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
2460 self._merge_subtitles(m3u8_subs, target=subtitles)
2461 if len(m3u8_formats) == 1:
2463 m3u8_formats[0].update({
2464 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
2469 formats.extend(m3u8_formats)
2470 elif src_ext == 'f4m':
2475 'plugin': 'flowplayer-3.2.0.1',
2477 f4m_url += '&' if '?' in f4m_url else '?'
2478 f4m_url += urllib.parse.urlencode(f4m_params)
2479 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
2480 elif src_ext == 'mpd':
2481 mpd_formats, mpd_subs = self._extract_mpd_formats_and_subtitles(
2482 src_url, video_id, mpd_id='dash', fatal=False)
2483 formats.extend(mpd_formats)
2484 self._merge_subtitles(mpd_subs, target=subtitles)
2485 elif re.search(r'\.ism/[Mm]anifest', src_url):
2486 ism_formats, ism_subs = self._extract_ism_formats_and_subtitles(
2487 src_url, video_id, ism_id='mss', fatal=False)
2488 formats.extend(ism_formats)
2489 self._merge_subtitles(ism_subs, target=subtitles)
2490 elif src_url.startswith('http') and self._is_valid_url(src, video_id):
2494 'ext': ext or src_ext or 'flv',
2495 'format_id': 'http-%d' % (bitrate or http_count),
2497 'filesize': filesize,
2502 for medium in smil.findall(self._xpath_ns('.//imagestream', namespace)):
2503 src = medium.get('src')
2504 if not src or src in srcs:
2510 'format_id': 'imagestream-%d' % (imgs_count),
2512 'ext': mimetype2ext(medium.get('type')),
2515 'width': int_or_none(medium.get('width')),
2516 'height': int_or_none(medium.get('height')),
2517 'format_note': 'SMIL storyboards',
2520 smil_subs = self._parse_smil_subtitles(smil, namespace=namespace)
2521 self._merge_subtitles(smil_subs, target=subtitles)
2523 return formats, subtitles
2525 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
2528 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
2529 src = textstream.get('src')
2530 if not src or src in urls:
2533 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
2534 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
2535 subtitles.setdefault(lang, []).append({
2541 def _extract_xspf_playlist(self, xspf_url, playlist_id, fatal=True):
2542 res = self._download_xml_handle(
2543 xspf_url, playlist_id, 'Downloading xpsf playlist',
2544 'Unable to download xspf manifest', fatal=fatal)
2551 return self._parse_xspf(
2552 xspf, playlist_id, xspf_url=xspf_url,
2553 xspf_base_url=base_url(xspf_url))
2555 def _parse_xspf(self, xspf_doc, playlist_id, xspf_url=None, xspf_base_url=None):
2557 'xspf': 'http://xspf.org/ns/0/',
2558 's1': 'http://static.streamone.nl/player/ns/0',
2562 for track in xspf_doc.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
2564 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
2565 description = xpath_text(
2566 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
2567 thumbnail = xpath_text(
2568 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
2569 duration = float_or_none(
2570 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
2573 for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP)):
2574 format_url = urljoin(xspf_base_url, location.text)
2579 'manifest_url': xspf_url,
2580 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
2581 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
2582 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
2588 'description': description,
2589 'thumbnail': thumbnail,
2590 'duration': duration,
2595 def _extract_mpd_formats(self, *args, **kwargs):
2596 fmts, subs = self._extract_mpd_formats_and_subtitles(*args, **kwargs)
2598 self._report_ignoring_subs('DASH')
2601 def _extract_mpd_formats_and_subtitles(self, *args, **kwargs):
2602 periods = self._extract_mpd_periods(*args, **kwargs)
2603 return self._merge_mpd_periods(periods)
2605 def _extract_mpd_periods(
2606 self, mpd_url, video_id, mpd_id=None, note=None, errnote=None,
2607 fatal=True, data=None, headers={}, query={}):
2609 if self.get_param('ignore_no_formats_error'):
2612 res = self._download_xml_handle(
2614 note='Downloading MPD manifest' if note is None else note,
2615 errnote='Failed to download MPD manifest' if errnote is None else errnote,
2616 fatal=fatal, data=data, headers=headers, query=query)
2623 # We could have been redirected to a new url when we retrieved our mpd file.
2625 mpd_base_url = base_url(mpd_url)
2627 return self._parse_mpd_periods(mpd_doc, mpd_id, mpd_base_url, mpd_url)
2629 def _parse_mpd_formats(self, *args, **kwargs):
2630 fmts, subs = self._parse_mpd_formats_and_subtitles(*args, **kwargs)
2632 self._report_ignoring_subs('DASH')
2635 def _parse_mpd_formats_and_subtitles(self, *args, **kwargs):
2636 periods = self._parse_mpd_periods(*args, **kwargs)
2637 return self._merge_mpd_periods(periods)
2639 def _merge_mpd_periods(self, periods):
2641 Combine all formats and subtitles from an MPD manifest into a single list,
2642 by concatenate streams with similar formats.
2644 formats, subtitles = {}, {}
2645 for period in periods:
2646 for f in period['formats']:
2647 assert 'is_dash_periods' not in f, 'format already processed'
2648 f['is_dash_periods'] = True
2649 format_key = tuple(v for k, v in f.items() if k not in (
2650 ('format_id', 'fragments', 'manifest_stream_number')))
2651 if format_key not in formats:
2652 formats[format_key] = f
2653 elif 'fragments' in f:
2654 formats[format_key].setdefault('fragments', []).extend(f['fragments'])
2656 if subtitles and period['subtitles']:
2657 self.report_warning(bug_reports_message(
2658 'Found subtitles in multiple periods in the DASH manifest; '
2659 'if part of the subtitles are missing,'
2662 for sub_lang, sub_info in period['subtitles'].items():
2663 subtitles.setdefault(sub_lang, []).extend(sub_info)
2665 return list(formats.values()), subtitles
2667 def _parse_mpd_periods(self, mpd_doc, mpd_id=None, mpd_base_url='', mpd_url=None):
2669 Parse formats from MPD manifest.
2671 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
2672 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
2673 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
2675 if not self.get_param('dynamic_mpd', True):
2676 if mpd_doc.get('type') == 'dynamic':
2679 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
2682 return self._xpath_ns(path, namespace)
2684 def is_drm_protected(element):
2685 return element.find(_add_ns('ContentProtection')) is not None
2687 def extract_multisegment_info(element, ms_parent_info):
2688 ms_info = ms_parent_info.copy()
2690 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2691 # common attributes and elements. We will only extract relevant
2693 def extract_common(source):
2694 segment_timeline = source.find(_add_ns('SegmentTimeline'))
2695 if segment_timeline is not None:
2696 s_e = segment_timeline.findall(_add_ns('S'))
2698 ms_info['total_number'] = 0
2701 r = int(s.get('r', 0))
2702 ms_info['total_number'] += 1 + r
2703 ms_info['s'].append({
2704 't': int(s.get('t', 0)),
2705 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2706 'd': int(s.attrib['d']),
2709 start_number = source.get('startNumber')
2711 ms_info['start_number'] = int(start_number)
2712 timescale = source.get('timescale')
2714 ms_info['timescale'] = int(timescale)
2715 segment_duration = source.get('duration')
2716 if segment_duration:
2717 ms_info['segment_duration'] = float(segment_duration)
2719 def extract_Initialization(source):
2720 initialization = source.find(_add_ns('Initialization'))
2721 if initialization is not None:
2722 ms_info['initialization_url'] = initialization.attrib['sourceURL']
2724 segment_list = element.find(_add_ns('SegmentList'))
2725 if segment_list is not None:
2726 extract_common(segment_list)
2727 extract_Initialization(segment_list)
2728 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
2730 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
2732 segment_template = element.find(_add_ns('SegmentTemplate'))
2733 if segment_template is not None:
2734 extract_common(segment_template)
2735 media = segment_template.get('media')
2737 ms_info['media'] = media
2738 initialization = segment_template.get('initialization')
2740 ms_info['initialization'] = initialization
2742 extract_Initialization(segment_template)
2745 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
2746 stream_numbers = collections.defaultdict(int)
2747 for period_idx, period in enumerate(mpd_doc.findall(_add_ns('Period'))):
2749 'id': period.get('id', f'period-{period_idx}'),
2751 'subtitles': collections.defaultdict(list),
2753 period_duration = parse_duration(period.get('duration')) or mpd_duration
2754 period_ms_info = extract_multisegment_info(period, {
2758 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
2759 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
2760 for representation in adaptation_set.findall(_add_ns('Representation')):
2761 representation_attrib = adaptation_set.attrib.copy()
2762 representation_attrib.update(representation.attrib)
2763 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2764 mime_type = representation_attrib['mimeType']
2765 content_type = representation_attrib.get('contentType', mime_type.split('/')[0])
2767 codec_str = representation_attrib.get('codecs', '')
2768 # Some kind of binary subtitle found in some youtube livestreams
2769 if mime_type == 'application/x-rawcc':
2770 codecs = {'scodec': codec_str}
2772 codecs = parse_codecs(codec_str)
2773 if content_type not in ('video', 'audio', 'text'):
2774 if mime_type == 'image/jpeg':
2775 content_type = mime_type
2776 elif codecs.get('vcodec', 'none') != 'none':
2777 content_type = 'video'
2778 elif codecs.get('acodec', 'none') != 'none':
2779 content_type = 'audio'
2780 elif codecs.get('scodec', 'none') != 'none':
2781 content_type = 'text'
2782 elif mimetype2ext(mime_type) in ('tt', 'dfxp', 'ttml', 'xml', 'json'):
2783 content_type = 'text'
2785 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
2789 for element in (representation, adaptation_set, period, mpd_doc):
2790 base_url_e = element.find(_add_ns('BaseURL'))
2791 if try_call(lambda: base_url_e.text) is not None:
2792 base_url = base_url_e.text + base_url
2793 if re.match(r'^https?://', base_url):
2795 if mpd_base_url and base_url.startswith('/'):
2796 base_url = urllib.parse.urljoin(mpd_base_url, base_url)
2797 elif mpd_base_url and not re.match(r'^https?://', base_url):
2798 if not mpd_base_url.endswith('/'):
2800 base_url = mpd_base_url + base_url
2801 representation_id = representation_attrib.get('id')
2802 lang = representation_attrib.get('lang')
2803 url_el = representation.find(_add_ns('BaseURL'))
2804 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
2805 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
2806 if representation_id is not None:
2807 format_id = representation_id
2809 format_id = content_type
2811 format_id = mpd_id + '-' + format_id
2812 if content_type in ('video', 'audio'):
2814 'format_id': format_id,
2815 'manifest_url': mpd_url,
2816 'ext': mimetype2ext(mime_type),
2817 'width': int_or_none(representation_attrib.get('width')),
2818 'height': int_or_none(representation_attrib.get('height')),
2819 'tbr': float_or_none(bandwidth, 1000),
2820 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
2821 'fps': int_or_none(representation_attrib.get('frameRate')),
2822 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
2823 'format_note': 'DASH %s' % content_type,
2824 'filesize': filesize,
2825 'container': mimetype2ext(mime_type) + '_dash',
2828 elif content_type == 'text':
2830 'ext': mimetype2ext(mime_type),
2831 'manifest_url': mpd_url,
2832 'filesize': filesize,
2834 elif content_type == 'image/jpeg':
2835 # See test case in VikiIE
2836 # https://www.viki.com/videos/1175236v-choosing-spouse-by-lottery-episode-1
2838 'format_id': format_id,
2840 'manifest_url': mpd_url,
2841 'format_note': 'DASH storyboards (jpeg)',
2845 if is_drm_protected(adaptation_set) or is_drm_protected(representation):
2847 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
2849 def prepare_template(template_name, identifiers):
2850 tmpl = representation_ms_info[template_name]
2851 if representation_id is not None:
2852 tmpl = tmpl.replace('$RepresentationID$', representation_id)
2853 # First of, % characters outside $...$ templates
2854 # must be escaped by doubling for proper processing
2855 # by % operator string formatting used further (see
2856 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2862 in_template = not in_template
2863 elif c == '%' and not in_template:
2865 # Next, $...$ templates are translated to their
2866 # %(...) counterparts to be used with % operator
2867 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
2868 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
2869 t.replace('$$', '$')
2872 # @initialization is a regular template like @media one
2873 # so it should be handled just the same way (see
2874 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2875 if 'initialization' in representation_ms_info:
2876 initialization_template = prepare_template(
2878 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2879 # $Time$ shall not be included for @initialization thus
2880 # only $Bandwidth$ remains
2882 representation_ms_info['initialization_url'] = initialization_template % {
2883 'Bandwidth': bandwidth,
2886 def location_key(location):
2887 return 'url' if re.match(r'^https?://', location) else 'path'
2889 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
2891 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2892 media_location_key = location_key(media_template)
2894 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2895 # can't be used at the same time
2896 if '%(Number' in media_template and 's' not in representation_ms_info:
2897 segment_duration = None
2898 if 'total_number' not in representation_ms_info and 'segment_duration' in representation_ms_info:
2899 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
2900 representation_ms_info['total_number'] = int(math.ceil(
2901 float_or_none(period_duration, segment_duration, default=0)))
2902 representation_ms_info['fragments'] = [{
2903 media_location_key: media_template % {
2904 'Number': segment_number,
2905 'Bandwidth': bandwidth,
2907 'duration': segment_duration,
2908 } for segment_number in range(
2909 representation_ms_info['start_number'],
2910 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
2912 # $Number*$ or $Time$ in media template with S list available
2913 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2914 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2915 representation_ms_info['fragments'] = []
2918 segment_number = representation_ms_info['start_number']
2920 def add_segment_url():
2921 segment_url = media_template % {
2922 'Time': segment_time,
2923 'Bandwidth': bandwidth,
2924 'Number': segment_number,
2926 representation_ms_info['fragments'].append({
2927 media_location_key: segment_url,
2928 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
2931 for num, s in enumerate(representation_ms_info['s']):
2932 segment_time = s.get('t') or segment_time
2936 for r in range(s.get('r', 0)):
2937 segment_time += segment_d
2940 segment_time += segment_d
2941 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
2942 # No media template,
2943 # e.g. https://www.youtube.com/watch?v=iXZV5uAYMJI
2944 # or any YouTube dashsegments video
2947 timescale = representation_ms_info['timescale']
2948 for s in representation_ms_info['s']:
2949 duration = float_or_none(s['d'], timescale)
2950 for r in range(s.get('r', 0) + 1):
2951 segment_uri = representation_ms_info['segment_urls'][segment_index]
2953 location_key(segment_uri): segment_uri,
2954 'duration': duration,
2957 representation_ms_info['fragments'] = fragments
2958 elif 'segment_urls' in representation_ms_info:
2959 # Segment URLs with no SegmentTimeline
2960 # E.g. https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2961 # https://github.com/ytdl-org/youtube-dl/pull/14844
2963 segment_duration = float_or_none(
2964 representation_ms_info['segment_duration'],
2965 representation_ms_info['timescale']) if 'segment_duration' in representation_ms_info else None
2966 for segment_url in representation_ms_info['segment_urls']:
2968 location_key(segment_url): segment_url,
2970 if segment_duration:
2971 fragment['duration'] = segment_duration
2972 fragments.append(fragment)
2973 representation_ms_info['fragments'] = fragments
2974 # If there is a fragments key available then we correctly recognized fragmented media.
2975 # Otherwise we will assume unfragmented media with direct access. Technically, such
2976 # assumption is not necessarily correct since we may simply have no support for
2977 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2978 if 'fragments' in representation_ms_info:
2980 # NB: mpd_url may be empty when MPD manifest is parsed from a string
2981 'url': mpd_url or base_url,
2982 'fragment_base_url': base_url,
2984 'protocol': 'http_dash_segments' if mime_type != 'image/jpeg' else 'mhtml',
2986 if 'initialization_url' in representation_ms_info:
2987 initialization_url = representation_ms_info['initialization_url']
2988 if not f.get('url'):
2989 f['url'] = initialization_url
2990 f['fragments'].append({location_key(initialization_url): initialization_url})
2991 f['fragments'].extend(representation_ms_info['fragments'])
2992 if not period_duration:
2993 period_duration = try_get(
2994 representation_ms_info,
2995 lambda r: sum(frag['duration'] for frag in r['fragments']), float)
2997 # Assuming direct URL to unfragmented media.
2999 if content_type in ('video', 'audio', 'image/jpeg'):
3000 f['manifest_stream_number'] = stream_numbers[f['url']]
3001 stream_numbers[f['url']] += 1
3002 period_entry['formats'].append(f)
3003 elif content_type == 'text':
3004 period_entry['subtitles'][lang or 'und'].append(f)
3007 def _extract_ism_formats(self, *args, **kwargs):
3008 fmts, subs = self._extract_ism_formats_and_subtitles(*args, **kwargs)
3010 self._report_ignoring_subs('ISM')
3013 def _extract_ism_formats_and_subtitles(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
3014 if self.get_param('ignore_no_formats_error'):
3017 res = self._download_xml_handle(
3019 note='Downloading ISM manifest' if note is None else note,
3020 errnote='Failed to download ISM manifest' if errnote is None else errnote,
3021 fatal=fatal, data=data, headers=headers, query=query)
3028 return self._parse_ism_formats_and_subtitles(ism_doc, urlh.url, ism_id)
3030 def _parse_ism_formats_and_subtitles(self, ism_doc, ism_url, ism_id=None):
3032 Parse formats from ISM manifest.
3034 1. [MS-SSTR]: Smooth Streaming Protocol,
3035 https://msdn.microsoft.com/en-us/library/ff469518.aspx
3037 if ism_doc.get('IsLive') == 'TRUE':
3040 duration = int(ism_doc.attrib['Duration'])
3041 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
3045 for stream in ism_doc.findall('StreamIndex'):
3046 stream_type = stream.get('Type')
3047 if stream_type not in ('video', 'audio', 'text'):
3049 url_pattern = stream.attrib['Url']
3050 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
3051 stream_name = stream.get('Name')
3052 stream_language = stream.get('Language', 'und')
3053 for track in stream.findall('QualityLevel'):
3054 KNOWN_TAGS = {'255': 'AACL', '65534': 'EC-3'}
3055 fourcc = track.get('FourCC') or KNOWN_TAGS.get(track.get('AudioTag'))
3056 # TODO: add support for WVC1 and WMAP
3057 if fourcc not in ('H264', 'AVC1', 'AACL', 'TTML', 'EC-3'):
3058 self.report_warning('%s is not a supported codec' % fourcc)
3060 tbr = int(track.attrib['Bitrate']) // 1000
3061 # [1] does not mention Width and Height attributes. However,
3062 # they're often present while MaxWidth and MaxHeight are
3063 # missing, so should be used as fallbacks
3064 width = int_or_none(track.get('MaxWidth') or track.get('Width'))
3065 height = int_or_none(track.get('MaxHeight') or track.get('Height'))
3066 sampling_rate = int_or_none(track.get('SamplingRate'))
3068 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
3069 track_url_pattern = urllib.parse.urljoin(ism_url, track_url_pattern)
3075 stream_fragments = stream.findall('c')
3076 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
3077 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
3078 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
3079 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
3080 if not fragment_ctx['duration']:
3082 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
3084 next_fragment_time = duration
3085 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
3086 for _ in range(fragment_repeat):
3088 'url': re.sub(r'{start[ _]time}', str(fragment_ctx['time']), track_url_pattern),
3089 'duration': fragment_ctx['duration'] / stream_timescale,
3091 fragment_ctx['time'] += fragment_ctx['duration']
3093 if stream_type == 'text':
3094 subtitles.setdefault(stream_language, []).append({
3098 'manifest_url': ism_url,
3099 'fragments': fragments,
3100 '_download_params': {
3101 'stream_type': stream_type,
3102 'duration': duration,
3103 'timescale': stream_timescale,
3105 'language': stream_language,
3106 'codec_private_data': track.get('CodecPrivateData'),
3109 elif stream_type in ('video', 'audio'):
3111 'format_id': join_nonempty(ism_id, stream_name, tbr),
3113 'manifest_url': ism_url,
3114 'ext': 'ismv' if stream_type == 'video' else 'isma',
3118 'asr': sampling_rate,
3119 'vcodec': 'none' if stream_type == 'audio' else fourcc,
3120 'acodec': 'none' if stream_type == 'video' else fourcc,
3122 'fragments': fragments,
3123 'has_drm': ism_doc.find('Protection') is not None,
3124 'language': stream_language,
3125 'audio_channels': int_or_none(track.get('Channels')),
3126 '_download_params': {
3127 'stream_type': stream_type,
3128 'duration': duration,
3129 'timescale': stream_timescale,
3130 'width': width or 0,
3131 'height': height or 0,
3133 'language': stream_language,
3134 'codec_private_data': track.get('CodecPrivateData'),
3135 'sampling_rate': sampling_rate,
3136 'channels': int_or_none(track.get('Channels', 2)),
3137 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
3138 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
3141 return formats, subtitles
3143 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):
3144 def absolute_url(item_url):
3145 return urljoin(base_url, item_url)
3147 def parse_content_type(content_type):
3148 if not content_type:
3150 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?
(?P
<codecs
>[^
"]+))?', content_type)
3152 mimetype, codecs = ctr.groups()
3153 f = parse_codecs(codecs)
3154 f['ext'] = mimetype2ext(mimetype)
3158 def _media_formats(src, cur_media_type, type_info=None):
3159 type_info = type_info or {}
3160 full_url = absolute_url(src)
3161 ext = type_info.get('ext') or determine_ext(full_url)
3163 is_plain_url = False
3164 formats = self._extract_m3u8_formats(
3165 full_url, video_id, ext='mp4',
3166 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
3167 preference=preference, quality=quality, fatal=False)
3169 is_plain_url = False
3170 formats = self._extract_mpd_formats(
3171 full_url, video_id, mpd_id=mpd_id, fatal=False)
3176 'vcodec': 'none' if cur_media_type == 'audio' else None,
3179 return is_plain_url, formats
3182 # amp-video and amp-audio are very similar to their HTML5 counterparts
3183 # so we will include them right here (see
3184 # https://www.ampproject.org/docs/reference/components/amp-video)
3185 # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/
3186 _MEDIA_TAG_NAME_RE = r'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)'
3187 media_tags = [(media_tag, media_tag_name, media_type, '')
3188 for media_tag, media_tag_name, media_type
3189 in re.findall(r'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE, webpage)]
3190 media_tags.extend(re.findall(
3191 # We only allow video|audio followed by a whitespace or '>'.
3192 # Allowing more characters may end up in significant slow down (see
3193 # https://github.com/ytdl-org/youtube-dl/issues/11979,
3194 # e.g. http://www.porntrex.com/maps/videositemap.xml).
3195 r'(?s)(<(?P<tag>%s)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>' % _MEDIA_TAG_NAME_RE, webpage))
3196 for media_tag, _, media_type, media_content in media_tags:
3201 media_attributes = extract_attributes(media_tag)
3202 src = strip_or_none(dict_get(media_attributes, ('src', 'data-video-src', 'data-src', 'data-source')))
3204 f = parse_content_type(media_attributes.get('type'))
3205 _, formats = _media_formats(src, media_type, f)
3206 media_info['formats'].extend(formats)
3207 media_info['thumbnail'] = absolute_url(media_attributes.get('poster'))
3209 for source_tag in re.findall(r'<source[^>]+>', media_content):
3210 s_attr = extract_attributes(source_tag)
3211 # data-video-src and data-src are non standard but seen
3212 # several times in the wild
3213 src = strip_or_none(dict_get(s_attr, ('src', 'data-video-src', 'data-src', 'data-source')))
3216 f = parse_content_type(s_attr.get('type'))
3217 is_plain_url, formats = _media_formats(src, media_type, f)
3219 # width, height, res, label and title attributes are
3220 # all not standard but seen several times in the wild
3223 for lbl in ('label', 'title')
3224 if str_or_none(s_attr.get(lbl))
3226 width = int_or_none(s_attr.get('width'))
3227 height = (int_or_none(s_attr.get('height'))
3228 or int_or_none(s_attr.get('res')))
3229 if not width or not height:
3231 resolution = parse_resolution(lbl)
3234 width = width or resolution.get('width')
3235 height = height or resolution.get('height')
3237 tbr = parse_bitrate(lbl)
3246 'format_id': s_attr.get('label') or s_attr.get('title'),
3248 f.update(formats[0])
3249 media_info['formats'].append(f)
3251 media_info['formats'].extend(formats)
3252 for track_tag in re.findall(r'<track[^>]+>', media_content):
3253 track_attributes = extract_attributes(track_tag)
3254 kind = track_attributes.get('kind')
3255 if not kind or kind in ('subtitles', 'captions'):
3256 src = strip_or_none(track_attributes.get('src'))
3259 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
3260 media_info['subtitles'].setdefault(lang, []).append({
3261 'url': absolute_url(src),
3263 for f in media_info['formats']:
3264 f.setdefault('http_headers', {})['Referer'] = base_url
3265 if media_info['formats'] or media_info['subtitles']:
3266 entries.append(media_info)
3269 def _extract_akamai_formats(self, *args, **kwargs):
3270 fmts, subs = self._extract_akamai_formats_and_subtitles(*args, **kwargs)
3272 self._report_ignoring_subs('akamai')
3275 def _extract_akamai_formats_and_subtitles(self, manifest_url, video_id, hosts={}):
3276 signed = 'hdnea=' in manifest_url
3278 # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
3279 manifest_url = re.sub(
3280 r'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
3281 '', manifest_url).strip('?')
3286 hdcore_sign = 'hdcore=3.7.0'
3287 f4m_url = re.sub(r'(https?://[^/]+)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
3288 hds_host = hosts.get('hds')
3290 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
3291 if 'hdcore=' not in f4m_url:
3292 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
3293 f4m_formats = self._extract_f4m_formats(
3294 f4m_url, video_id, f4m_id='hds', fatal=False)
3295 for entry in f4m_formats:
3296 entry.update({'extra_param_to_segment_url': hdcore_sign})
3297 formats.extend(f4m_formats)
3299 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
3300 hls_host = hosts.get('hls')
3302 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
3303 m3u8_formats, m3u8_subtitles = self._extract_m3u8_formats_and_subtitles(
3304 m3u8_url, video_id, 'mp4', 'm3u8_native',
3305 m3u8_id='hls', fatal=False)
3306 formats.extend(m3u8_formats)
3307 subtitles = self._merge_subtitles(subtitles, m3u8_subtitles)
3309 http_host = hosts.get('http')
3310 if http_host and m3u8_formats and not signed:
3311 REPL_REGEX = r'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
3312 qualities = re.match(REPL_REGEX, m3u8_url).group(2).split(',')
3313 qualities_length = len(qualities)
3314 if len(m3u8_formats) in (qualities_length, qualities_length + 1):
3316 for f in m3u8_formats:
3317 if f['vcodec'] != 'none':
3318 for protocol in ('http', 'https'):
3320 del http_f['manifest_url']
3322 REPL_REGEX, protocol + fr'://{http_host}/\g<1>{qualities[i]}\3', f['url'])
3324 'format_id': http_f['format_id'].replace('hls-', protocol + '-'),
3326 'protocol': protocol,
3328 formats.append(http_f)
3331 return formats, subtitles
3333 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
3334 query = urllib.parse.urlparse(url).query
3335 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
3337 r'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url)
3338 url_base = mobj.group('url')
3339 http_base_url = '%s%s:%s' % ('http', mobj.group('s') or '', url_base)
3342 def manifest_url(manifest):
3343 m_url = f'{http_base_url}/{manifest}'
3345 m_url += '?%s' % query
3348 if 'm3u8' not in skip_protocols:
3349 formats.extend(self._extract_m3u8_formats(
3350 manifest_url('playlist.m3u8'), video_id, 'mp4',
3351 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
3352 if 'f4m' not in skip_protocols:
3353 formats.extend(self._extract_f4m_formats(
3354 manifest_url('manifest.f4m'),
3355 video_id, f4m_id='hds', fatal=False))
3356 if 'dash' not in skip_protocols:
3357 formats.extend(self._extract_mpd_formats(
3358 manifest_url('manifest.mpd'),
3359 video_id, mpd_id='dash', fatal=False))
3360 if re.search(r'(?:/smil:|\.smil)', url_base):
3361 if 'smil' not in skip_protocols:
3362 rtmp_formats = self._extract_smil_formats(
3363 manifest_url('jwplayer.smil'),
3364 video_id, fatal=False)
3365 for rtmp_format in rtmp_formats:
3366 rtsp_format = rtmp_format.copy()
3367 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
3368 del rtsp_format['play_path']
3369 del rtsp_format['ext']
3370 rtsp_format.update({
3371 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
3372 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
3375 formats.extend([rtmp_format, rtsp_format])
3377 for protocol in ('rtmp', 'rtsp'):
3378 if protocol not in skip_protocols:
3380 'url': f'{protocol}:{url_base}',
3381 'format_id': protocol,
3382 'protocol': protocol,
3386 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
3388 r'''(?s)jwplayer\s*\(\s*(?P<q>'|")(?
!(?P
=q
)).+(?P
=q
)\s
*\
)(?
!</script
>).*?\
.\s
*setup\s
*\
(\s
*(?P
<options
>(?
:\
([^
)]*\
)|
[^
)])+)\s
*\
)''',
3392 jwplayer_data = self._parse_json(mobj.group('options'),
3394 transform_source=transform_source)
3395 except ExtractorError:
3398 if isinstance(jwplayer_data, dict):
3399 return jwplayer_data
3401 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
3402 jwplayer_data = self._find_jwplayer_data(
3403 webpage, video_id, transform_source=js_to_json)
3404 return self._parse_jwplayer_data(
3405 jwplayer_data, video_id, *args, **kwargs)
3407 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
3408 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3410 if not isinstance(jwplayer_data, dict):
3413 playlist_items = jwplayer_data.get('playlist')
3414 # JWPlayer backward compatibility: single playlist item/flattened playlists
3415 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
3416 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
3417 if not isinstance(playlist_items, list):
3418 playlist_items = (playlist_items or jwplayer_data, )
3420 for video_data in playlist_items:
3421 if not isinstance(video_data, dict):
3423 # JWPlayer backward compatibility: flattened sources
3424 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
3425 if 'sources' not in video_data:
3426 video_data['sources'] = [video_data]
3428 this_video_id = video_id or video_data['mediaid']
3430 formats = self._parse_jwplayer_formats(
3431 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
3432 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
3435 tracks = video_data.get('tracks')
3436 if tracks and isinstance(tracks, list):
3437 for track in tracks:
3438 if not isinstance(track, dict):
3440 track_kind = track.get('kind')
3441 if not track_kind or not isinstance(track_kind, str):
3443 if track_kind.lower() not in ('captions', 'subtitles'):
3445 track_url = urljoin(base_url, track.get('file'))
3448 subtitles.setdefault(track.get('label') or 'en', []).append({
3449 'url': self._proto_relative_url(track_url)
3453 'id': this_video_id,
3454 'title': unescapeHTML(video_data['title'] if require_title else video_data.get('title')),
3455 'description': clean_html(video_data.get('description')),
3456 'thumbnail': urljoin(base_url, self._proto_relative_url(video_data.get('image'))),
3457 'timestamp': int_or_none(video_data.get('pubdate')),
3458 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
3459 'subtitles': subtitles,
3460 'alt_title': clean_html(video_data.get('subtitle')), # attributes used e.g. by Tele5 ...
3461 'genre': clean_html(video_data.get('genre')),
3462 'channel': clean_html(dict_get(video_data, ('category', 'channel'))),
3463 'season_number': int_or_none(video_data.get('season')),
3464 'episode_number': int_or_none(video_data.get('episode')),
3465 'release_year': int_or_none(video_data.get('releasedate')),
3466 'age_limit': int_or_none(video_data.get('age_restriction')),
3468 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
3469 if len(formats) == 1 and re.search(r'^(?:http|//).*(?:youtube\.com|youtu\.be)/.+', formats[0]['url']):
3471 '_type': 'url_transparent',
3472 'url': formats[0]['url'],
3475 entry['formats'] = formats
3476 entries.append(entry)
3477 if len(entries) == 1:
3480 return self.playlist_result(entries)
3482 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
3483 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3486 for source in jwplayer_sources_data:
3487 if not isinstance(source, dict):
3489 source_url = urljoin(
3490 base_url, self._proto_relative_url(source.get('file')))
3491 if not source_url or source_url in urls:
3493 urls.add(source_url)
3494 source_type = source.get('type') or ''
3495 ext = mimetype2ext(source_type) or determine_ext(source_url)
3496 if source_type == 'hls' or ext == 'm3u8' or 'format=m3u8-aapl' in source_url:
3497 formats.extend(self._extract_m3u8_formats(
3498 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
3499 m3u8_id=m3u8_id, fatal=False))
3500 elif source_type == 'dash' or ext == 'mpd' or 'format=mpd-time-csf' in source_url:
3501 formats.extend(self._extract_mpd_formats(
3502 source_url, video_id, mpd_id=mpd_id, fatal=False))
3504 formats.extend(self._extract_smil_formats(
3505 source_url, video_id, fatal=False))
3506 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3507 elif source_type.startswith('audio') or ext in (
3508 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
3515 format_id = str_or_none(source.get('label'))
3516 height = int_or_none(source.get('height'))
3517 if height is None and format_id:
3518 # Often no height is provided but there is a label in
3519 # format like "1080p", "720p SD", or 1080.
3520 height = parse_resolution(format_id).get('height')
3523 'width': int_or_none(source.get('width')),
3525 'tbr': int_or_none(source.get('bitrate'), scale=1000),
3526 'filesize': int_or_none(source.get('filesize')),
3528 'format_id': format_id
3530 if source_url.startswith('rtmp'):
3531 a_format['ext'] = 'flv'
3532 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3533 # of jwplayer.flash.swf
3534 rtmp_url_parts = re.split(
3535 r'((?:mp4|mp3|flv):)', source_url, 1)
3536 if len(rtmp_url_parts) == 3:
3537 rtmp_url, prefix, play_path = rtmp_url_parts
3540 'play_path': prefix + play_path,
3543 a_format.update(rtmp_params)
3544 formats.append(a_format)
3547 def _live_title(self, name):
3548 self._downloader.deprecation_warning('yt_dlp.InfoExtractor._live_title is deprecated and does not work as expected')
3551 def _int(self, v, name, fatal=False, **kwargs):
3552 res = int_or_none(v, **kwargs)
3554 msg = f'Failed to extract {name}: Could not parse value {v!r}'
3556 raise ExtractorError(msg)
3558 self.report_warning(msg)
3561 def _float(self, v, name, fatal=False, **kwargs):
3562 res = float_or_none(v, **kwargs)
3564 msg = f'Failed to extract {name}: Could not parse value {v!r}'
3566 raise ExtractorError(msg)
3568 self.report_warning(msg)
3571 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
3572 path='/', secure=False, discard=False, rest={}, **kwargs):
3573 cookie = http.cookiejar.Cookie(
3574 0, name, value, port, port is not None, domain, True,
3575 domain.startswith('.'), path, True, secure, expire_time,
3576 discard, None, None, rest)
3577 self.cookiejar.set_cookie(cookie)
3579 def _get_cookies(self, url):
3580 """ Return a http.cookies.SimpleCookie with the cookies for the url """
3581 return LenientSimpleCookie(self._downloader.cookiejar.get_cookie_header(url))
3583 def _apply_first_set_cookie_header(self, url_handle, cookie):
3585 Apply first Set-Cookie header instead of the last. Experimental.
3587 Some sites (e.g. [1-3]) may serve two cookies under the same name
3588 in Set-Cookie header and expect the first (old) one to be set rather
3589 than second (new). However, as of RFC6265 the newer one cookie
3590 should be set into cookie store what actually happens.
3591 We will workaround this issue by resetting the cookie to
3592 the first one manually.
3593 1. https://new.vk.com/
3594 2. https://github.com/ytdl-org/youtube-dl/issues/9841#issuecomment-227871201
3595 3. https://learning.oreilly.com/
3597 for header, cookies in url_handle.headers.items():
3598 if header.lower() != 'set-cookie':
3600 cookies = cookies.encode('iso-8859-1').decode('utf-8')
3601 cookie_value = re.search(
3602 r'%s=(.+?);.*?\b[Dd]omain=(.+?)(?:[,;]|$)' % cookie, cookies)
3604 value, domain = cookie_value.groups()
3605 self._set_cookie(domain, cookie, value)
3609 def get_testcases(cls, include_onlymatching=False):
3610 # Do not look in super classes
3611 t = vars(cls).get('_TEST')
3613 assert not hasattr(cls, '_TESTS'), f'{cls.ie_key()}IE has _TEST and _TESTS'
3616 tests = vars(cls).get('_TESTS', [])
3618 if not include_onlymatching and t.get('only_matching', False):
3620 t['name'] = cls.ie_key()
3622 if getattr(cls, '__wrapped__', None):
3623 yield from cls.__wrapped__.get_testcases(include_onlymatching)
3626 def get_webpage_testcases(cls):
3627 tests = vars(cls).get('_WEBPAGE_TESTS', [])
3629 t['name'] = cls.ie_key()
3631 if getattr(cls, '__wrapped__', None):
3632 yield from cls.__wrapped__.get_webpage_testcases()
3634 @classproperty(cache=True)
3636 """Get age limit from the testcases"""
3637 return max(traverse_obj(
3638 (*cls.get_testcases(include_onlymatching=False), *cls.get_webpage_testcases()),
3639 (..., (('playlist', 0), None), 'info_dict', 'age_limit')) or [0])
3641 @classproperty(cache=True)
3642 def _RETURN_TYPE(cls):
3643 """What the extractor returns: "video", "playlist", "any", or None (Unknown)"""
3644 tests = tuple(cls.get_testcases(include_onlymatching=False))
3647 elif not any(k.startswith('playlist') for test in tests for k in test):
3649 elif all(any(k.startswith('playlist') for k in test) for test in tests):
3654 def is_single_video(cls, url):
3655 """Returns whether the URL is of a single video, None if unknown"""
3656 if cls.suitable(url):
3657 return {'video': True, 'playlist': False}.get(cls._RETURN_TYPE)
3660 def is_suitable(cls, age_limit):
3661 """Test whether the extractor is generally suitable for the given age limit"""
3662 return not age_restricted(cls.age_limit, age_limit)
3665 def description(cls, *, markdown=True, search_examples=None):
3666 """Description of the extractor"""
3668 if cls._NETRC_MACHINE:
3670 desc += f' [*{cls._NETRC_MACHINE}*](## "netrc machine")'
3672 desc += f' [{cls._NETRC_MACHINE}]'
3673 if cls.IE_DESC is False:
3676 desc += f' {cls.IE_DESC}'
3678 desc += f'{";" if cls.IE_DESC else ""} "{cls.SEARCH_KEY}:" prefix'
3680 _COUNTS = ('', '5', '10', 'all')
3681 desc += f' (e.g. "{cls.SEARCH_KEY}{random.choice(_COUNTS)}:{random.choice(search_examples)}")'
3682 if not cls.working():
3683 desc += ' (**Currently broken**)' if markdown else ' (Currently broken)'
3685 # Escape emojis. Ref: https://github.com/github/markup/issues/1153
3686 name = (' - **%s**' % re.sub(r':(\w+:)', ':\u200B\\g<1>', cls.IE_NAME)) if markdown else cls.IE_NAME
3687 return f'{name}:{desc}' if desc else name
3689 def extract_subtitles(self, *args, **kwargs):
3690 if (self.get_param('writesubtitles', False)
3691 or self.get_param('listsubtitles')):
3692 return self._get_subtitles(*args, **kwargs)
3695 def _get_subtitles(self, *args, **kwargs):
3696 raise NotImplementedError('This method must be implemented by subclasses')
3698 class CommentsDisabled(Exception):
3699 """Raise in _get_comments if comments are disabled for the video"""
3701 def extract_comments(self, *args, **kwargs):
3702 if not self.get_param('getcomments'):
3704 generator = self._get_comments(*args, **kwargs)
3711 comments.append(next(generator))
3712 except StopIteration:
3714 except KeyboardInterrupt:
3715 self.to_screen('Interrupted by user')
3716 except self.CommentsDisabled:
3717 return {'comments': None, 'comment_count': None}
3718 except Exception as e:
3719 if self.get_param('ignoreerrors') is not True:
3721 self._downloader.report_error(e)
3722 comment_count = len(comments)
3723 self.to_screen(f'Extracted {comment_count} comments')
3725 'comments': comments,
3726 'comment_count': None if interrupted else comment_count
3730 def _get_comments(self, *args, **kwargs):
3731 raise NotImplementedError('This method must be implemented by subclasses')
3734 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
3735 """ Merge subtitle items for one language. Items with duplicated URLs/data
3736 will be dropped. """
3737 list1_data = {(item.get('url'), item.get('data')) for item in subtitle_list1}
3738 ret = list(subtitle_list1)
3739 ret.extend(item for item in subtitle_list2 if (item.get('url'), item.get('data')) not in list1_data)
3743 def _merge_subtitles(cls, *dicts, target=None):
3744 """ Merge subtitle dictionaries, language by language. """
3748 for lang, subs in d.items():
3749 target[lang] = cls._merge_subtitle_items(target.get(lang, []), subs)
3752 def extract_automatic_captions(self, *args, **kwargs):
3753 if (self.get_param('writeautomaticsub', False)
3754 or self.get_param('listsubtitles')):
3755 return self._get_automatic_captions(*args, **kwargs)
3758 def _get_automatic_captions(self, *args, **kwargs):
3759 raise NotImplementedError('This method must be implemented by subclasses')
3761 @functools.cached_property
3762 def _cookies_passed(self):
3763 """Whether cookies have been passed to YoutubeDL"""
3764 return self.get_param('cookiefile') is not None or self.get_param('cookiesfrombrowser') is not None
3766 def mark_watched(self, *args, **kwargs):
3767 if not self.get_param('mark_watched', False):
3769 if self.supports_login() and self._get_login_info()[0] is not None or self._cookies_passed:
3770 self._mark_watched(*args, **kwargs)
3772 def _mark_watched(self, *args, **kwargs):
3773 raise NotImplementedError('This method must be implemented by subclasses')
3775 def geo_verification_headers(self):
3777 geo_verification_proxy = self.get_param('geo_verification_proxy')
3778 if geo_verification_proxy:
3779 headers['Ytdl-request-proxy'] = geo_verification_proxy
3783 def _generic_id(url):
3784 return urllib.parse.unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
3786 def _generic_title(self, url='', webpage='', *, default=None):
3787 return (self._og_search_title(webpage, default=None)
3788 or self._html_extract_title(webpage, default=None)
3789 or urllib.parse.unquote(os.path.splitext(url_basename(url))[0])
3792 def _extract_chapters_helper(self, chapter_list, start_function, title_function, duration, strict=True):
3796 'start_time': start_function(chapter),
3797 'title': title_function(chapter),
3798 } for chapter in chapter_list or []]
3800 warn = self.report_warning
3802 warn = self.write_debug
3803 chapter_list.sort(key=lambda c: c['start_time'] or 0)
3805 chapters = [{'start_time': 0}]
3806 for idx, chapter in enumerate(chapter_list):
3807 if chapter['start_time'] is None:
3808 warn(f'Incomplete chapter {idx}')
3809 elif chapters[-1]['start_time'] <= chapter['start_time'] <= duration:
3810 chapters.append(chapter)
3811 elif chapter not in chapters:
3812 issue = (f'{chapter["start_time"]} > {duration}' if chapter['start_time'] > duration
3813 else f'{chapter["start_time"]} < {chapters[-1]["start_time"]}')
3814 warn(f'Invalid start time ({issue}) for chapter "{chapter["title"]}"')
3817 def _extract_chapters_from_description(self, description, duration):
3818 duration_re = r'(?:\d+:)?\d{1,2}:\d{2}'
3819 sep_re = r'(?m)^\s*(%s)\b\W*\s(%s)\s*$'
3820 return self._extract_chapters_helper(
3821 re.findall(sep_re % (duration_re, r'.+?'), description or ''),
3822 start_function=lambda x: parse_duration(x[0]), title_function=lambda x: x[1],
3823 duration=duration, strict=False) or self._extract_chapters_helper(
3824 re.findall(sep_re % (r'.+?', duration_re), description or ''),
3825 start_function=lambda x: parse_duration(x[1]), title_function=lambda x: x[0],
3826 duration=duration, strict=False)
3829 def _availability(is_private=None, needs_premium=None, needs_subscription=None, needs_auth=None, is_unlisted=None):
3830 all_known = all(map(
3831 lambda x: x is not None,
3832 (is_private, needs_premium, needs_subscription, needs_auth, is_unlisted)))
3834 'private' if is_private
3835 else 'premium_only' if needs_premium
3836 else 'subscriber_only' if needs_subscription
3837 else 'needs_auth' if needs_auth
3838 else 'unlisted' if is_unlisted
3839 else 'public' if all_known
3842 def _configuration_arg(self, key, default=NO_DEFAULT, *, ie_key=None, casesense=False):
3844 @returns A
list of values
for the extractor argument given by
"key"
3845 or "default" if no such key
is present
3846 @param default The default value to
return when the key
is not present (default
: [])
3847 @param casesense When false
, the values are converted to lower case
3849 ie_key = ie_key if isinstance(ie_key, str) else (ie_key or self).ie_key()
3850 val = traverse_obj(self._downloader.params, ('extractor_args', ie_key.lower(), key))
3852 return [] if default is NO_DEFAULT else default
3853 return list(val) if casesense else [x.lower() for x in val]
3855 def _yes_playlist(self, playlist_id, video_id, smuggled_data=None, *, playlist_label='playlist', video_label='video'):
3856 if not playlist_id or not video_id:
3859 no_playlist = (smuggled_data or {}).get('force_noplaylist')
3860 if no_playlist is not None:
3861 return not no_playlist
3863 video_id = '' if video_id is True else f' {video_id}'
3864 playlist_id = '' if playlist_id is True else f' {playlist_id}'
3865 if self.get_param('noplaylist'):
3866 self.to_screen(f'Downloading just the {video_label}{video_id} because of --no-playlist')
3868 self.to_screen(f'Downloading {playlist_label}{playlist_id} - add --no-playlist to download just the {video_label}{video_id}')
3871 def _error_or_warning(self, err, _count=None, _retries=0, *, fatal=True):
3872 RetryManager.report_retry(
3873 err, _count or int(fatal), _retries,
3874 info=self.to_screen, warn=self.report_warning, error=None if fatal else self.report_warning,
3875 sleep_func=self.get_param('retry_sleep_functions', {}).get('extractor'))
3877 def RetryManager(self, **kwargs):
3878 return RetryManager(self.get_param('extractor_retries', 3), self._error_or_warning, **kwargs)
3880 def _extract_generic_embeds(self, url, *args, info_dict={}, note='Extracting generic embeds', **kwargs):
3881 display_id = traverse_obj(info_dict, 'display_id', 'id')
3882 self.to_screen(f'{format_field(display_id, None, "%s: ")}{note}')
3883 return self._downloader.get_info_extractor('Generic')._extract_embeds(
3884 smuggle_url(url, {'block_ies': [self.ie_key()]}), *args, **kwargs)
3887 def extract_from_webpage(cls, ydl, url, webpage):
3888 ie = (cls if isinstance(cls._extract_from_webpage, types.MethodType)
3889 else ydl.get_info_extractor(cls.ie_key()))
3890 for info in ie._extract_from_webpage(url, webpage) or []:
3891 # url = None since we do not want to set (webpage/original)_url
3892 ydl.add_default_extra_info(info, ie, None)
3896 def _extract_from_webpage(cls, url, webpage):
3897 for embed_url in orderedSet(
3898 cls._extract_embed_urls(url, webpage) or [], lazy=True):
3899 yield cls.url_result(embed_url, None if cls._VALID_URL is False else cls)
3902 def _extract_embed_urls(cls, url, webpage):
3903 """@returns all the embed urls on the webpage"""
3904 if '_EMBED_URL_RE' not in cls.__dict__:
3905 assert isinstance(cls._EMBED_REGEX, (list, tuple))
3906 for idx, regex in enumerate(cls._EMBED_REGEX):
3907 assert regex.count('(?P<url>') == 1, \
3908 f'{cls.__name__}._EMBED_REGEX[{idx}] must have exactly 1 url group\n\t{regex}'
3909 cls._EMBED_URL_RE = tuple(map(re.compile, cls._EMBED_REGEX))
3911 for regex in cls._EMBED_URL_RE:
3912 for mobj in regex.finditer(webpage):
3913 embed_url = urllib.parse.urljoin(url, unescapeHTML(mobj.group('url')))
3914 if cls._VALID_URL is False or cls.suitable(embed_url):
3917 class StopExtraction(Exception):
3921 def _extract_url(cls, webpage): # TODO: Remove
3922 """Only for compatibility with some older extractors"""
3923 return next(iter(cls._extract_embed_urls(None, webpage) or []), None)
3926 def __init_subclass__(cls, *, plugin_name=None, **kwargs):
3928 mro = inspect.getmro(cls)
3929 super_class = cls.__wrapped__ = mro[mro.index(cls) + 1]
3930 cls.PLUGIN_NAME, cls.ie_key = plugin_name, super_class.ie_key
3931 cls.IE_NAME = f'{super_class.IE_NAME}+{plugin_name}'
3932 while getattr(super_class, '__wrapped__', None):
3933 super_class = super_class.__wrapped__
3934 setattr(sys.modules[super_class.__module__], super_class.__name__, cls)
3935 _PLUGIN_OVERRIDES[super_class].append(cls)
3937 return super().__init_subclass__(**kwargs)
3940 class SearchInfoExtractor(InfoExtractor):
3942 Base class for paged search queries extractors.
3943 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
3944 Instances should define _SEARCH_KEY and optionally _MAX_RESULTS
3947 _MAX_RESULTS = float('inf')
3948 _RETURN_TYPE = 'playlist'
3951 def _VALID_URL(cls):
3952 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
3954 def _real_extract(self, query):
3955 prefix, query = self._match_valid_url(query).group('prefix', 'query')
3957 return self._get_n_results(query, 1)
3958 elif prefix == 'all':
3959 return self._get_n_results(query, self._MAX_RESULTS)
3963 raise ExtractorError(f'invalid download number {n} for query "{query}"')
3964 elif n > self._MAX_RESULTS:
3965 self.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3966 n = self._MAX_RESULTS
3967 return self._get_n_results(query, n)
3969 def _get_n_results(self, query, n):
3970 """Get a specified number of results for a query.
3971 Either this function or _search_results must be overridden by subclasses """
3972 return self.playlist_result(
3973 itertools.islice(self._search_results(query), 0, None if n == float('inf') else n),
3976 def _search_results(self, query):
3977 """Returns an iterator of search results"""
3978 raise NotImplementedError('This method must be implemented by subclasses')
3981 def SEARCH_KEY(cls):
3982 return cls._SEARCH_KEY
3985 class UnsupportedURLIE(InfoExtractor):
3990 def _real_extract(self, url):
3991 raise UnsupportedError(url)
3994 _PLUGIN_OVERRIDES = collections.defaultdict(list)