23 import xml
.etree
.ElementTree
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
78 parse_m3u8_attributes
,
105 """Information Extractor class.
107 Information extractors are the classes that, given a URL, extract
108 information about the video (or videos) the URL refers to. This
109 information includes the real video URL, the video title, author and
110 others. The information is stored in a dictionary which is then
111 passed to the YoutubeDL. The YoutubeDL processes this
112 information possibly downloading the video to the file system, among
113 other possible outcomes.
115 The type field determines the type of the result.
116 By far the most common value (and the default if _type is missing) is
117 "video", which indicates a single video.
119 For a video, the dictionaries must include the following fields:
121 id: Video identifier.
122 title: Video title, unescaped. Set to an empty string if video has
123 no title as opposed to "None" which signifies that the
124 extractor failed to obtain a title
126 Additionally, it must contain either a formats entry or a url one:
128 formats: A list of dictionaries for each format available, ordered
129 from worst to best quality.
132 * url The mandatory URL representing the media:
133 for plain file media - HTTP URL of this file,
135 for HLS - URL of the M3U8 media playlist,
136 for HDS - URL of the F4M manifest,
138 - HTTP URL to plain file media (in case of
140 - URL of the MPD manifest or base URL
141 representing the media if MPD manifest
142 is parsed from a string (in case of
144 for MSS - URL of the ISM manifest.
145 * request_data Data to send in POST request to the URL
147 The URL of the manifest file in case of
149 for HLS - URL of the M3U8 master playlist,
150 for HDS - URL of the F4M manifest,
151 for DASH - URL of the MPD manifest,
152 for MSS - URL of the ISM manifest.
153 * manifest_stream_number (For internal use only)
154 The index of the stream in the manifest file
155 * ext Will be calculated from URL if missing
156 * format A human-readable description of the format
157 ("mp4 container with h264/opus").
158 Calculated from the format_id, width, height.
159 and format_note fields if missing.
160 * format_id A short description of the format
161 ("mp4_h264_opus" or "19").
162 Technically optional, but strongly recommended.
163 * format_note Additional info about the format
164 ("3D" or "DASH video")
165 * width Width of the video, if known
166 * height Height of the video, if known
167 * aspect_ratio Aspect ratio of the video, if known
168 Automatically calculated from width and height
169 * resolution Textual description of width and height
170 Automatically calculated from width and height
171 * dynamic_range The dynamic range of the video. One of:
172 "SDR" (None), "HDR10", "HDR10+, "HDR12", "HLG, "DV"
173 * tbr Average bitrate of audio and video in kbps (1000 bits/sec)
174 * abr Average audio bitrate in kbps (1000 bits/sec)
175 * acodec Name of the audio codec in use
176 * asr Audio sampling rate in Hertz
177 * audio_channels Number of audio channels
178 * vbr Average video bitrate in kbps (1000 bits/sec)
180 * vcodec Name of the video codec in use
181 * container Name of the container format
182 * filesize The number of bytes, if known in advance
183 * filesize_approx An estimate for the number of bytes
184 * player_url SWF Player URL (used for rtmpdump).
185 * protocol The protocol that will be used for the actual
186 download, lower-case. One of "http", "https" or
187 one of the protocols defined in downloader.PROTOCOL_MAP
189 Base URL for fragments. Each fragment's path
190 value (if present) will be relative to
192 * fragments A list of fragments of a fragmented media.
193 Each fragment entry must contain either an url
194 or a path. If an url is present it should be
195 considered by a client. Otherwise both path and
196 fragment_base_url must be present. Here is
197 the list of all potential fields:
198 * "url" - fragment's URL
199 * "path" - fragment's path relative to
201 * "duration" (optional, int or float)
202 * "filesize" (optional, int)
203 * is_from_start Is a live format that can be downloaded
204 from the start. Boolean
205 * preference Order number of this format. If this field is
206 present and not None, the formats get sorted
207 by this field, regardless of all other values.
208 -1 for default (order by other properties),
209 -2 or smaller for less than default.
210 < -1000 to hide the format (if there is
211 another one which is strictly better)
212 * language Language code, e.g. "de" or "en-US".
213 * language_preference Is this in the language mentioned in
215 10 if it's what the URL is about,
216 -1 for default (don't know),
217 -10 otherwise, other values reserved for now.
218 * quality Order number of the video quality of this
219 format, irrespective of the file format.
220 -1 for default (order by other properties),
221 -2 or smaller for less than default.
222 * source_preference Order number for this video source
223 (quality takes higher priority)
224 -1 for default (order by other properties),
225 -2 or smaller for less than default.
226 * http_headers A dictionary of additional HTTP headers
227 to add to the request.
228 * stretched_ratio If given and not 1, indicates that the
229 video's pixels are not square.
230 width : height ratio as float.
231 * no_resume The server does not support resuming the
232 (HTTP or RTMP) download. Boolean.
233 * has_drm True if the format has DRM and cannot be downloaded.
234 'maybe' if the format may have DRM and has to be tested before download.
235 * extra_param_to_segment_url A query string to append to each
236 fragment's URL, or to update each existing query string
237 with. If it is an HLS stream with an AES-128 decryption key,
238 the query paramaters will be passed to the key URI as well,
239 unless there is an `extra_param_to_key_url` given,
240 or unless an external key URI is provided via `hls_aes`.
241 Only applied by the native HLS/DASH downloaders.
242 * extra_param_to_key_url A query string to append to the URL
243 of the format's HLS AES-128 decryption key.
244 Only applied by the native HLS downloader.
245 * hls_aes A dictionary of HLS AES-128 decryption information
246 used by the native HLS downloader to override the
247 values in the media playlist when an '#EXT-X-KEY' tag
248 is present in the playlist:
249 * uri The URI from which the key will be downloaded
250 * key The key (as hex) used to decrypt fragments.
251 If `key` is given, any key URI will be ignored
252 * iv The IV (as hex) used to decrypt fragments
253 * downloader_options A dictionary of downloader options
254 (For internal use only)
255 * http_chunk_size Chunk size for HTTP downloads
256 * ffmpeg_args Extra arguments for ffmpeg downloader (input)
257 * ffmpeg_args_out Extra arguments for ffmpeg downloader (output)
258 * is_dash_periods Whether the format is a result of merging
259 multiple DASH periods.
260 RTMP formats can also have the additional fields: page_url,
261 app, play_path, tc_url, flash_version, rtmp_live, rtmp_conn,
262 rtmp_protocol, rtmp_real_time
264 url: Final video URL.
265 ext: Video filename extension.
266 format: The video format, defaults to ext (used for --get-format)
267 player_url: SWF Player URL (used for rtmpdump).
269 The following fields are optional:
271 direct: True if a direct video file was given (must only be set by GenericIE)
272 alt_title: A secondary title of the video.
273 display_id: An alternative identifier for the video, not necessarily
274 unique, but available before title. Typically, id is
275 something like "4234987", title "Dancing naked mole rats",
276 and display_id "dancing-naked-mole-rats"
277 thumbnails: A list of dictionaries, with the following entries:
278 * "id" (optional, string) - Thumbnail format ID
280 * "preference" (optional, int) - quality of the image
281 * "width" (optional, int)
282 * "height" (optional, int)
283 * "resolution" (optional, string "{width}x{height}",
285 * "filesize" (optional, int)
286 * "http_headers" (dict) - HTTP headers for the request
287 thumbnail: Full URL to a video thumbnail image.
288 description: Full video description.
289 uploader: Full name of the video uploader.
290 license: License name the video is licensed under.
291 creators: List of creators of the video.
292 timestamp: UNIX timestamp of the moment the video was uploaded
293 upload_date: Video upload date in UTC (YYYYMMDD).
294 If not explicitly set, calculated from timestamp
295 release_timestamp: UNIX timestamp of the moment the video was released.
296 If it is not clear whether to use timestamp or this, use the former
297 release_date: The date (YYYYMMDD) when the video was released in UTC.
298 If not explicitly set, calculated from release_timestamp
299 release_year: Year (YYYY) as integer when the video or album was released.
300 To be used if no exact release date is known.
301 If not explicitly set, calculated from release_date.
302 modified_timestamp: UNIX timestamp of the moment the video was last modified.
303 modified_date: The date (YYYYMMDD) when the video was last modified in UTC.
304 If not explicitly set, calculated from modified_timestamp
305 uploader_id: Nickname or id of the video uploader.
306 uploader_url: Full URL to a personal webpage of the video uploader.
307 channel: Full name of the channel the video is uploaded on.
308 Note that channel fields may or may not repeat uploader
309 fields. This depends on a particular extractor.
310 channel_id: Id of the channel.
311 channel_url: Full URL to a channel webpage.
312 channel_follower_count: Number of followers of the channel.
313 channel_is_verified: Whether the channel is verified on the platform.
314 location: Physical location where the video was filmed.
315 subtitles: The available subtitles as a dictionary in the format
316 {tag: subformats}. "tag" is usually a language code, and
317 "subformats" is a list sorted from lower to higher
318 preference, each element is a dictionary with the "ext"
320 * "data": The subtitles file contents
321 * "url": A URL pointing to the subtitles file
322 It can optionally also have:
323 * "name": Name or description of the subtitles
324 * "http_headers": A dictionary of additional HTTP headers
325 to add to the request.
326 "ext" will be calculated from URL if missing
327 automatic_captions: Like 'subtitles'; contains automatically generated
328 captions instead of normal subtitles
329 duration: Length of the video in seconds, as an integer or float.
330 view_count: How many users have watched the video on the platform.
331 concurrent_view_count: How many users are currently watching the video on the platform.
332 like_count: Number of positive ratings of the video
333 dislike_count: Number of negative ratings of the video
334 repost_count: Number of reposts of the video
335 average_rating: Average rating give by users, the scale used depends on the webpage
336 comment_count: Number of comments on the video
337 comments: A list of comments, each with one or more of the following
338 properties (all but one of text or html optional):
339 * "author" - human-readable name of the comment author
340 * "author_id" - user ID of the comment author
341 * "author_thumbnail" - The thumbnail of the comment author
342 * "author_url" - The url to the comment author's page
343 * "author_is_verified" - Whether the author is verified
345 * "author_is_uploader" - Whether the comment is made by
348 * "html" - Comment as HTML
349 * "text" - Plain text of the comment
350 * "timestamp" - UNIX timestamp of comment
351 * "parent" - ID of the comment this one is replying to.
352 Set to "root" to indicate that this is a
353 comment to the original video.
354 * "like_count" - Number of positive ratings of the comment
355 * "dislike_count" - Number of negative ratings of the comment
356 * "is_favorited" - Whether the comment is marked as
357 favorite by the video uploader
358 * "is_pinned" - Whether the comment is pinned to
359 the top of the comments
360 age_limit: Age restriction for the video, as an integer (years)
361 webpage_url: The URL to the video webpage, if given to yt-dlp it
362 should allow to get the same result again. (It will be set
363 by YoutubeDL if it's missing)
364 categories: A list of categories that the video falls in, for example
366 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
367 cast: A list of the video cast
368 is_live: True, False, or None (=unknown). Whether this video is a
369 live stream that goes on instead of a fixed-length video.
370 was_live: True, False, or None (=unknown). Whether this video was
371 originally a live stream.
372 live_status: None (=unknown), 'is_live', 'is_upcoming', 'was_live', 'not_live',
373 or 'post_live' (was live, but VOD is not yet processed)
374 If absent, automatically set from is_live, was_live
375 start_time: Time in seconds where the reproduction should start, as
376 specified in the URL.
377 end_time: Time in seconds where the reproduction should end, as
378 specified in the URL.
379 chapters: A list of dictionaries, with the following entries:
380 * "start_time" - The start time of the chapter in seconds
381 * "end_time" - The end time of the chapter in seconds
382 * "title" (optional, string)
383 heatmap: A list of dictionaries, with the following entries:
384 * "start_time" - The start time of the data point in seconds
385 * "end_time" - The end time of the data point in seconds
386 * "value" - The normalized value of the data point (float between 0 and 1)
387 playable_in_embed: Whether this video is allowed to play in embedded
388 players on other sites. Can be True (=always allowed),
389 False (=never allowed), None (=unknown), or a string
390 specifying the criteria for embedability; e.g. 'whitelist'
391 availability: Under what condition the video is available. One of
392 'private', 'premium_only', 'subscriber_only', 'needs_auth',
393 'unlisted' or 'public'. Use 'InfoExtractor._availability'
395 media_type: The type of media as classified by the site, e.g. "episode", "clip", "trailer"
396 _old_archive_ids: A list of old archive ids needed for backward compatibility
397 _format_sort_fields: A list of fields to use for sorting formats
398 __post_extractor: A function to be called just before the metadata is
399 written to either disk, logger or console. The function
400 must return a dict which will be added to the info_dict.
401 This is usefull for additional information that is
402 time-consuming to extract. Note that the fields thus
403 extracted will not be available to output template and
404 match_filter. So, only "comments" and "comment_count" are
405 currently allowed to be extracted via this method.
407 The following fields should only be used when the video belongs to some logical
410 chapter: Name or title of the chapter the video belongs to.
411 chapter_number: Number of the chapter the video belongs to, as an integer.
412 chapter_id: Id of the chapter the video belongs to, as a unicode string.
414 The following fields should only be used when the video is an episode of some
415 series, programme or podcast:
417 series: Title of the series or programme the video episode belongs to.
418 series_id: Id of the series or programme the video episode belongs to, as a unicode string.
419 season: Title of the season the video episode belongs to.
420 season_number: Number of the season the video episode belongs to, as an integer.
421 season_id: Id of the season the video episode belongs to, as a unicode string.
422 episode: Title of the video episode. Unlike mandatory video title field,
423 this field should denote the exact title of the video episode
424 without any kind of decoration.
425 episode_number: Number of the video episode within a season, as an integer.
426 episode_id: Id of the video episode, as a unicode string.
428 The following fields should only be used when the media is a track or a part of
431 track: Title of the track.
432 track_number: Number of the track within an album or a disc, as an integer.
433 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
435 artists: List of artists of the track.
436 composers: List of composers of the piece.
437 genres: List of genres of the track.
438 album: Title of the album the track belongs to.
439 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
440 album_artists: List of all artists appeared on the album.
441 E.g. ["Ash Borer", "Fell Voices"] or ["Various Artists"].
442 Useful for splits and compilations.
443 disc_number: Number of the disc or other physical medium the track belongs to,
446 The following fields should only be set for clips that should be cut from the original video:
448 section_start: Start time of the section in seconds
449 section_end: End time of the section in seconds
451 The following fields should only be set for storyboards:
452 rows: Number of rows in each storyboard fragment, as an integer
453 columns: Number of columns in each storyboard fragment, as an integer
455 The following fields are deprecated and should not be set by new code:
456 composer: Use "composers" instead.
457 Composer(s) of the piece, comma-separated.
458 artist: Use "artists" instead.
459 Artist(s) of the track, comma-separated.
460 genre: Use "genres" instead.
461 Genre(s) of the track, comma-separated.
462 album_artist: Use "album_artists" instead.
463 All artists appeared on the album, comma-separated.
464 creator: Use "creators" instead.
465 The creator of the video.
467 Unless mentioned otherwise, the fields should be Unicode strings.
469 Unless mentioned otherwise, None is equivalent to absence of information.
472 _type "playlist" indicates multiple videos.
473 There must be a key "entries", which is a list, an iterable, or a PagedList
474 object, each element of which is a valid dictionary by this specification.
476 Additionally, playlists can have "id", "title", and any other relevant
477 attributes with the same semantics as videos (see above).
479 It can also have the following optional fields:
481 playlist_count: The total number of videos in a playlist. If not given,
482 YoutubeDL tries to calculate it from "entries"
485 _type "multi_video" indicates that there are multiple videos that
486 form a single show, for examples multiple acts of an opera or TV episode.
487 It must have an entries key like a playlist and contain all the keys
488 required for a video at the same time.
491 _type "url" indicates that the video must be extracted from another
492 location, possibly by a different extractor. Its only required key is:
493 "url" - the next URL to extract.
494 The key "ie_key" can be set to the class name (minus the trailing "IE",
495 e.g. "Youtube") if the extractor class is known in advance.
496 Additionally, the dictionary may have any properties of the resolved entity
497 known in advance, for example "title" if the title of the referred video is
501 _type "url_transparent" entities have the same specification as "url", but
502 indicate that the given additional information is more precise than the one
503 associated with the resolved URL.
504 This is useful when a site employs a video service that hosts the video and
505 its technical metadata, but that video service does not embed a useful
506 title, description etc.
509 Subclasses of this should also be added to the list of extractors and
510 should define _VALID_URL as a regexp or a Sequence of regexps, and
511 re-define the _real_extract() and (optionally) _real_initialize() methods.
513 Subclasses may also override suitable() if necessary, but ensure the function
514 signature is preserved and that this function imports everything it needs
515 (except other extractors), so that lazy_extractors works correctly.
517 Subclasses can define a list of _EMBED_REGEX, which will be searched for in
518 the HTML of Generic webpages. It may also override _extract_embed_urls
519 or _extract_from_webpage as necessary. While these are normally classmethods,
520 _extract_from_webpage is allowed to be an instance method.
522 _extract_from_webpage may raise self.StopExtraction() to stop further
523 processing of the webpage and obtain exclusive rights to it. This is useful
524 when the extractor cannot reliably be matched using just the URL,
525 e.g. invidious/peertube instances
527 Embed-only extractors can be defined by setting _VALID_URL = False.
529 To support username + password (or netrc) login, the extractor must define a
530 _NETRC_MACHINE and re-define _perform_login(username, password) and
531 (optionally) _initialize_pre_login() methods. The _perform_login method will
532 be called between _initialize_pre_login and _real_initialize if credentials
533 are passed by the user. In cases where it is necessary to have the login
534 process as part of the extraction rather than initialization, _perform_login
535 can be left undefined.
537 _GEO_BYPASS attribute may be set to False in order to disable
538 geo restriction bypass mechanisms for a particular extractor.
539 Though it won't disable explicit geo restriction bypass based on
540 country code provided with geo_bypass_country.
542 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
543 countries for this extractor. One of these countries will be used by
544 geo restriction bypass mechanism right away in order to bypass
545 geo restriction, of course, if the mechanism is not disabled.
547 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
548 IP blocks in CIDR notation for this extractor. One of these IP blocks
549 will be used by geo restriction bypass mechanism similarly
552 The _ENABLED attribute should be set to False for IEs that
553 are disabled by default and must be explicitly enabled.
555 The _WORKING attribute should be set to False for broken IEs
556 in order to warn the users and skip the tests.
561 _x_forwarded_for_ip
= None
563 _GEO_COUNTRIES
= None
564 _GEO_IP_BLOCKS
= None
567 _NETRC_MACHINE
= None
573 def _login_hint(self
, method
=NO_DEFAULT
, netrc
=None):
574 password_hint
= f
'--username and --password, --netrc-cmd, or --netrc ({netrc or self._NETRC_MACHINE}) to provide account credentials'
577 'any': f
'Use --cookies, --cookies-from-browser, {password_hint}',
578 'password': f
'Use {password_hint}',
580 'Use --cookies-from-browser or --cookies for the authentication. '
581 'See https://github.com/yt-dlp/yt-dlp/wiki/FAQ#how-do-i-pass-cookies-to-yt-dlp for how to manually pass cookies'),
582 }[method
if method
is not NO_DEFAULT
else 'any' if self
.supports_login() else 'cookies']
584 def __init__(self
, downloader
=None):
585 """Constructor. Receives an optional downloader (a YoutubeDL instance).
586 If a downloader is not passed during initialization,
587 it must be set using "set_downloader()" before "extract()" is called"""
589 self
._x
_forwarded
_for
_ip
= None
590 self
._printed
_messages
= set()
591 self
.set_downloader(downloader
)
594 def _match_valid_url(cls
, url
):
595 if cls
._VALID
_URL
is False:
597 # This does not use has/getattr intentionally - we want to know whether
598 # we have cached the regexp for *this* class, whereas getattr would also
599 # match the superclass
600 if '_VALID_URL_RE' not in cls
.__dict
__:
601 cls
._VALID
_URL
_RE
= tuple(map(re
.compile, variadic(cls
._VALID
_URL
)))
602 return next(filter(None, (regex
.match(url
) for regex
in cls
._VALID
_URL
_RE
)), None)
605 def suitable(cls
, url
):
606 """Receives a URL and returns True if suitable for this IE."""
607 # This function must import everything it needs (except other extractors),
608 # so that lazy_extractors works correctly
609 return cls
._match
_valid
_url
(url
) is not None
612 def _match_id(cls
, url
):
613 return cls
._match
_valid
_url
(url
).group('id')
616 def get_temp_id(cls
, url
):
618 return cls
._match
_id
(url
)
619 except (IndexError, AttributeError):
624 """Getter method for _WORKING."""
628 def supports_login(cls
):
629 return bool(cls
._NETRC
_MACHINE
)
631 def initialize(self
):
632 """Initializes an instance (authentication, etc)."""
633 self
._printed
_messages
= set()
634 self
._initialize
_geo
_bypass
({
635 'countries': self
._GEO
_COUNTRIES
,
636 'ip_blocks': self
._GEO
_IP
_BLOCKS
,
639 self
._initialize
_pre
_login
()
640 if self
.supports_login():
641 username
, password
= self
._get
_login
_info
()
643 self
._perform
_login
(username
, password
)
644 elif self
.get_param('username') and False not in (self
.IE_DESC
, self
._NETRC
_MACHINE
):
645 self
.report_warning(f
'Login with password is not supported for this website. {self._login_hint("cookies")}')
646 self
._real
_initialize
()
649 def _initialize_geo_bypass(self
, geo_bypass_context
):
651 Initialize geo restriction bypass mechanism.
653 This method is used to initialize geo bypass mechanism based on faking
654 X-Forwarded-For HTTP header. A random country from provided country list
655 is selected and a random IP belonging to this country is generated. This
656 IP will be passed as X-Forwarded-For HTTP header in all subsequent
659 This method will be used for initial geo bypass mechanism initialization
660 during the instance initialization with _GEO_COUNTRIES and
663 You may also manually call it from extractor's code if geo bypass
664 information is not available beforehand (e.g. obtained during
665 extraction) or due to some other reason. In this case you should pass
666 this information in geo bypass context passed as first argument. It may
667 contain following fields:
669 countries: List of geo unrestricted countries (similar
671 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
672 (similar to _GEO_IP_BLOCKS)
675 if not self
._x
_forwarded
_for
_ip
:
677 # Geo bypass mechanism is explicitly disabled by user
678 if not self
.get_param('geo_bypass', True):
681 if not geo_bypass_context
:
682 geo_bypass_context
= {}
684 # Backward compatibility: previously _initialize_geo_bypass
685 # expected a list of countries, some 3rd party code may still use
687 if isinstance(geo_bypass_context
, (list, tuple)):
688 geo_bypass_context
= {
689 'countries': geo_bypass_context
,
692 # The whole point of geo bypass mechanism is to fake IP
693 # as X-Forwarded-For HTTP header based on some IP block or
696 # Path 1: bypassing based on IP block in CIDR notation
698 # Explicit IP block specified by user, use it right away
699 # regardless of whether extractor is geo bypassable or not
700 ip_block
= self
.get_param('geo_bypass_ip_block', None)
702 # Otherwise use random IP block from geo bypass context but only
703 # if extractor is known as geo bypassable
705 ip_blocks
= geo_bypass_context
.get('ip_blocks')
706 if self
._GEO
_BYPASS
and ip_blocks
:
707 ip_block
= random
.choice(ip_blocks
)
710 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(ip_block
)
711 self
.write_debug(f
'Using fake IP {self._x_forwarded_for_ip} as X-Forwarded-For')
714 # Path 2: bypassing based on country code
716 # Explicit country code specified by user, use it right away
717 # regardless of whether extractor is geo bypassable or not
718 country
= self
.get_param('geo_bypass_country', None)
720 # Otherwise use random country code from geo bypass context but
721 # only if extractor is known as geo bypassable
723 countries
= geo_bypass_context
.get('countries')
724 if self
._GEO
_BYPASS
and countries
:
725 country
= random
.choice(countries
)
728 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country
)
729 self
._downloader
.write_debug(
730 f
'Using fake IP {self._x_forwarded_for_ip} ({country.upper()}) as X-Forwarded-For')
732 def extract(self
, url
):
733 """Extracts URL information and returns it in list of dicts."""
738 self
.to_screen('Extracting URL: %s' % (
739 url
if self
.get_param('verbose') else truncate_string(url
, 100, 20)))
740 ie_result
= self
._real
_extract
(url
)
741 if ie_result
is None:
743 if self
._x
_forwarded
_for
_ip
:
744 ie_result
['__x_forwarded_for_ip'] = self
._x
_forwarded
_for
_ip
745 subtitles
= ie_result
.get('subtitles') or {}
746 if 'no-live-chat' in self
.get_param('compat_opts'):
747 for lang
in ('live_chat', 'comments', 'danmaku'):
748 subtitles
.pop(lang
, None)
750 except GeoRestrictedError
as e
:
751 if self
.__maybe
_fake
_ip
_and
_retry
(e
.countries
):
754 except UnsupportedError
:
756 except ExtractorError
as e
:
757 e
.video_id
= e
.video_id
or self
.get_temp_id(url
)
758 e
.ie
= e
.ie
or self
.IE_NAME
759 e
.traceback
= e
.traceback
or sys
.exc_info()[2]
761 except IncompleteRead
as e
:
762 raise ExtractorError('A network error has occurred.', cause
=e
, expected
=True, video_id
=self
.get_temp_id(url
))
763 except (KeyError, StopIteration) as e
:
764 raise ExtractorError('An extractor error has occurred.', cause
=e
, video_id
=self
.get_temp_id(url
))
766 def __maybe_fake_ip_and_retry(self
, countries
):
767 if (not self
.get_param('geo_bypass_country', None)
769 and self
.get_param('geo_bypass', True)
770 and not self
._x
_forwarded
_for
_ip
772 country_code
= random
.choice(countries
)
773 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country_code
)
774 if self
._x
_forwarded
_for
_ip
:
776 'Video is geo restricted. Retrying extraction with fake IP '
777 f
'{self._x_forwarded_for_ip} ({country_code.upper()}) as X-Forwarded-For.')
781 def set_downloader(self
, downloader
):
782 """Sets a YoutubeDL instance as the downloader for this IE."""
783 self
._downloader
= downloader
787 return self
._downloader
.cache
791 return self
._downloader
.cookiejar
793 def _initialize_pre_login(self
):
794 """ Initialization before login. Redefine in subclasses."""
797 def _perform_login(self
, username
, password
):
798 """ Login with username and password. Redefine in subclasses."""
801 def _real_initialize(self
):
802 """Real initialization process. Redefine in subclasses."""
805 def _real_extract(self
, url
):
806 """Real extraction process. Redefine in subclasses."""
807 raise NotImplementedError('This method must be implemented by subclasses')
811 """A string for getting the InfoExtractor with get_info_extractor"""
812 return cls
.__name
__[:-2]
816 return cls
.__name
__[:-2]
819 def __can_accept_status_code(err
, expected_status
):
820 assert isinstance(err
, HTTPError
)
821 if expected_status
is None:
823 elif callable(expected_status
):
824 return expected_status(err
.status
) is True
826 return err
.status
in variadic(expected_status
)
828 def _create_request(self
, url_or_request
, data
=None, headers
=None, query
=None, extensions
=None):
829 if isinstance(url_or_request
, urllib
.request
.Request
):
830 self
._downloader
.deprecation_warning(
831 'Passing a urllib.request.Request to _create_request() is deprecated. '
832 'Use yt_dlp.networking.common.Request instead.')
833 url_or_request
= urllib_req_to_req(url_or_request
)
834 elif not isinstance(url_or_request
, Request
):
835 url_or_request
= Request(url_or_request
)
837 url_or_request
.update(data
=data
, headers
=headers
, query
=query
, extensions
=extensions
)
838 return url_or_request
840 def _request_webpage(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, data
=None,
841 headers
=None, query
=None, expected_status
=None, impersonate
=None, require_impersonation
=False):
843 Return the response handle.
845 See _download_webpage docstring for arguments specification.
847 if not self
._downloader
._first
_webpage
_request
:
848 sleep_interval
= self
.get_param('sleep_interval_requests') or 0
849 if sleep_interval
> 0:
850 self
.to_screen(f
'Sleeping {sleep_interval} seconds ...')
851 time
.sleep(sleep_interval
)
853 self
._downloader
._first
_webpage
_request
= False
856 self
.report_download_webpage(video_id
)
857 elif note
is not False:
859 self
.to_screen(str(note
))
861 self
.to_screen(f
'{video_id}: {note}')
863 # Some sites check X-Forwarded-For HTTP header in order to figure out
864 # the origin of the client behind proxy. This allows bypassing geo
865 # restriction by faking this header's value to IP that belongs to some
866 # geo unrestricted country. We will do so once we encounter any
867 # geo restriction error.
868 if self
._x
_forwarded
_for
_ip
:
869 headers
= (headers
or {}).copy()
870 headers
.setdefault('X-Forwarded-For', self
._x
_forwarded
_for
_ip
)
874 if impersonate
in (True, ''):
875 impersonate
= ImpersonateTarget()
876 requested_targets
= [
877 t
if isinstance(t
, ImpersonateTarget
) else ImpersonateTarget
.from_str(t
)
878 for t
in variadic(impersonate
)
879 ] if impersonate
else []
881 available_target
= next(filter(self
._downloader
._impersonate
_target
_available
, requested_targets
), None)
883 extensions
['impersonate'] = available_target
884 elif requested_targets
:
885 message
= 'The extractor is attempting impersonation, but '
887 'no impersonate target is available' if not str(impersonate
)
888 else f
'none of these impersonate targets are available: "{", ".join(map(str, requested_targets))}"')
889 info_msg
= ('see https://github.com/yt-dlp/yt-dlp#impersonation '
890 'for information on installing the required dependencies')
891 if require_impersonation
:
892 raise ExtractorError(f
'{message}; {info_msg}', expected
=True)
893 self
.report_warning(f
'{message}; if you encounter errors, then {info_msg}', only_once
=True)
896 return self
._downloader
.urlopen(self
._create
_request
(url_or_request
, data
, headers
, query
, extensions
))
897 except network_exceptions
as err
:
898 if isinstance(err
, HTTPError
):
899 if self
.__can
_accept
_status
_code
(err
, expected_status
):
905 errnote
= 'Unable to download webpage'
907 errmsg
= f
'{errnote}: {err}'
909 raise ExtractorError(errmsg
, cause
=err
)
911 self
.report_warning(errmsg
)
914 def _download_webpage_handle(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True,
915 encoding
=None, data
=None, headers
={}, query={}
, expected_status
=None,
916 impersonate
=None, require_impersonation
=False):
918 Return a tuple (page content as string, URL handle).
921 url_or_request -- plain text URL as a string or
922 a yt_dlp.networking.Request object
923 video_id -- Video/playlist/item identifier (string)
926 note -- note printed before downloading (string)
927 errnote -- note printed in case of an error (string)
928 fatal -- flag denoting whether error should be considered fatal,
929 i.e. whether it should cause ExtractionError to be raised,
930 otherwise a warning will be reported and extraction continued
931 encoding -- encoding for a page content decoding, guessed automatically
932 when not explicitly specified
933 data -- POST data (bytes)
934 headers -- HTTP headers (dict)
935 query -- URL query (dict)
936 expected_status -- allows to accept failed HTTP requests (non 2xx
937 status code) by explicitly specifying a set of accepted status
938 codes. Can be any of the following entities:
939 - an integer type specifying an exact failed status code to
941 - a list or a tuple of integer types specifying a list of
942 failed status codes to accept
943 - a callable accepting an actual failed status code and
944 returning True if it should be accepted
945 Note that this argument does not affect success status codes (2xx)
946 which are always accepted.
947 impersonate -- the impersonate target. Can be any of the following entities:
948 - an instance of yt_dlp.networking.impersonate.ImpersonateTarget
949 - a string in the format of CLIENT[:OS]
950 - a list or a tuple of CLIENT[:OS] strings or ImpersonateTarget instances
951 - a boolean value; True means any impersonate target is sufficient
952 require_impersonation -- flag to toggle whether the request should raise an error
953 if impersonation is not possible (bool, default: False)
956 # Strip hashes from the URL (#1038)
957 if isinstance(url_or_request
, str):
958 url_or_request
= url_or_request
.partition('#')[0]
960 urlh
= self
._request
_webpage
(url_or_request
, video_id
, note
, errnote
, fatal
, data
=data
,
961 headers
=headers
, query
=query
, expected_status
=expected_status
,
962 impersonate
=impersonate
, require_impersonation
=require_impersonation
)
966 content
= self
._webpage
_read
_content
(urlh
, url_or_request
, video_id
, note
, errnote
, fatal
,
967 encoding
=encoding
, data
=data
)
968 return (content
, urlh
)
971 def _guess_encoding_from_content(content_type
, webpage_bytes
):
972 m
= re
.match(r
'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type
)
974 encoding
= m
.group(1)
976 m
= re
.search(br
'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
977 webpage_bytes[:1024])
979 encoding = m.group(1).decode('ascii')
980 elif webpage_bytes.startswith(b'\xff\xfe'):
987 def __check_blocked(self, content):
988 first_block = content[:512]
989 if ('<title>Access to this site is blocked</title>' in content
990 and 'Websense' in first_block):
991 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
992 blocked_iframe = self._html_search_regex(
993 r'<iframe src="([^
"]+)"', content,
994 'Websense information URL
', default=None)
996 msg += f' Visit {blocked_iframe}
for more details
'
997 raise ExtractorError(msg, expected=True)
998 if '<title
>The URL you requested has been blocked
</title
>' in first_block:
1000 'Access to this webpage has been blocked by Indian censorship
. '
1001 'Use a VPN
or proxy
server (with --proxy
) to route around it
.')
1002 block_msg = self._html_search_regex(
1003 r'</h1
><p
>(.*?
)</p
>',
1004 content, 'block message
', default=None)
1006 msg += ' (Message
: "{}")'.format(block_msg.replace('\n', ' '))
1007 raise ExtractorError(msg, expected=True)
1008 if ('<title
>TTK
:: Доступ к ресурсу ограничен
</title
>' in content
1009 and 'blocklist
.rkn
.gov
.ru
' in content):
1010 raise ExtractorError(
1011 'Access to this webpage has been blocked by decision of the Russian government
. '
1012 'Visit http
://blocklist
.rkn
.gov
.ru
/ for a block reason
.',
1015 def _request_dump_filename(self, url, video_id, data=None):
1016 if data is not None:
1017 data = hashlib.md5(data).hexdigest()
1018 basen = join_nonempty(video_id, data, url, delim='_
')
1019 trim_length = self.get_param('trim_file_name
') or 240
1020 if len(basen) > trim_length:
1021 h = '___
' + hashlib.md5(basen.encode()).hexdigest()
1022 basen = basen[:trim_length - len(h)] + h
1023 filename = sanitize_filename(f'{basen}
.dump
', restricted=True)
1024 # Working around MAX_PATH limitation on Windows (see
1025 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
1026 if compat_os_name == 'nt
':
1027 absfilepath = os.path.abspath(filename)
1028 if len(absfilepath) > 259:
1029 filename = fR'\\?\{absfilepath}
'
1032 def __decode_webpage(self, webpage_bytes, encoding, headers):
1034 encoding = self._guess_encoding_from_content(headers.get('Content
-Type
', ''), webpage_bytes)
1036 return webpage_bytes.decode(encoding, 'replace
')
1038 return webpage_bytes.decode('utf
-8', 'replace
')
1040 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True,
1041 prefix=None, encoding=None, data=None):
1042 webpage_bytes = urlh.read()
1043 if prefix is not None:
1044 webpage_bytes = prefix + webpage_bytes
1045 if self.get_param('dump_intermediate_pages
', False):
1046 self.to_screen('Dumping request to
' + urlh.url)
1047 dump = base64.b64encode(webpage_bytes).decode('ascii
')
1048 self._downloader.to_screen(dump)
1049 if self.get_param('write_pages
'):
1050 if isinstance(url_or_request, Request):
1051 data = self._create_request(url_or_request, data).data
1052 filename = self._request_dump_filename(urlh.url, video_id, data)
1053 self.to_screen(f'Saving request to {filename}
')
1054 with open(filename, 'wb
') as outf:
1055 outf.write(webpage_bytes)
1057 content = self.__decode_webpage(webpage_bytes, encoding, urlh.headers)
1058 self.__check_blocked(content)
1062 def __print_error(self, errnote, fatal, video_id, err):
1064 raise ExtractorError(f'{video_id}
: {errnote}
', cause=err)
1066 self.report_warning(f'{video_id}
: {errnote}
: {err}
')
1068 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True, errnote=None):
1069 if transform_source:
1070 xml_string = transform_source(xml_string)
1072 return compat_etree_fromstring(xml_string.encode())
1073 except xml.etree.ElementTree.ParseError as ve:
1074 self.__print_error('Failed to parse XML
' if errnote is None else errnote, fatal, video_id, ve)
1076 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True, errnote=None, **parser_kwargs):
1079 json_string, cls=LenientJSONDecoder, strict=False, transform_source=transform_source, **parser_kwargs)
1080 except ValueError as ve:
1081 self.__print_error('Failed to parse JSON
' if errnote is None else errnote, fatal, video_id, ve)
1083 def _parse_socket_response_as_json(self, data, *args, **kwargs):
1084 return self._parse_json(data[data.find('{'):data.rfind('}
') + 1], *args, **kwargs)
1086 def __create_download_methods(name, parser, note, errnote, return_value):
1088 def parse(ie, content, *args, errnote=errnote, **kwargs):
1091 if errnote is False:
1092 kwargs['errnote
'] = errnote
1093 # parser is fetched by name so subclasses can override it
1094 return getattr(ie, parser)(content, *args, **kwargs)
1096 def download_handle(self, url_or_request, video_id, note=note, errnote=errnote, transform_source=None,
1097 fatal=True, encoding=None, data=None, headers={}, query={}, expected_status=None,
1098 impersonate=None, require_impersonation=False):
1099 res = self._download_webpage_handle(
1100 url_or_request, video_id, note=note, errnote=errnote, fatal=fatal, encoding=encoding,
1101 data=data, headers=headers, query=query, expected_status=expected_status,
1102 impersonate=impersonate, require_impersonation=require_impersonation)
1106 return parse(self, content, video_id, transform_source=transform_source, fatal=fatal, errnote=errnote), urlh
1108 def download_content(self, url_or_request, video_id, note=note, errnote=errnote, transform_source=None,
1109 fatal=True, encoding=None, data=None, headers={}, query={}, expected_status=None,
1110 impersonate=None, require_impersonation=False):
1111 if self.get_param('load_pages
'):
1112 url_or_request = self._create_request(url_or_request, data, headers, query)
1113 filename = self._request_dump_filename(url_or_request.url, video_id, url_or_request.data)
1114 self.to_screen(f'Loading request
from {filename}
')
1116 with open(filename, 'rb
') as dumpf:
1117 webpage_bytes = dumpf.read()
1118 except OSError as e:
1119 self.report_warning(f'Unable to load request
from disk
: {e}
')
1121 content = self.__decode_webpage(webpage_bytes, encoding, url_or_request.headers)
1122 return parse(self, content, video_id, transform_source=transform_source, fatal=fatal, errnote=errnote)
1126 'transform_source
': transform_source,
1128 'encoding
': encoding,
1132 'expected_status
': expected_status,
1133 'impersonate
': impersonate,
1134 'require_impersonation
': require_impersonation,
1137 kwargs.pop('transform_source
')
1138 # The method is fetched by name so subclasses can override _download_..._handle
1139 res = getattr(self, download_handle.__name__)(url_or_request, video_id, **kwargs)
1140 return res if res is False else res[0]
1142 def impersonate(func, name, return_value):
1143 func.__name__, func.__qualname__ = name, f'InfoExtractor
.{name}
'
1145 @param transform_source Apply this transformation before parsing
1146 @returns {return_value}
1148 See _download_webpage_handle docstring for other arguments specification
1151 impersonate(download_handle, f'_download_{name}_handle
', f'({return_value}
, URL handle
)')
1152 impersonate(download_content, f'_download_{name}
', f'{return_value}
')
1153 return download_handle, download_content
1155 _download_xml_handle, _download_xml = __create_download_methods(
1156 'xml
', '_parse_xml
', 'Downloading XML
', 'Unable to download XML
', 'xml
as an xml
.etree
.ElementTree
.Element
')
1157 _download_json_handle, _download_json = __create_download_methods(
1158 'json
', '_parse_json
', 'Downloading JSON metadata
', 'Unable to download JSON metadata
', 'JSON
object as a
dict')
1159 _download_socket_json_handle, _download_socket_json = __create_download_methods(
1160 'socket_json
', '_parse_socket_response_as_json
', 'Polling socket
', 'Unable to poll socket
', 'JSON
object as a
dict')
1161 __download_webpage = __create_download_methods('webpage
', None, None, None, 'data of the page
as a string
')[1]
1163 def _download_webpage(
1164 self, url_or_request, video_id, note=None, errnote=None,
1165 fatal=True, tries=1, timeout=NO_DEFAULT, *args, **kwargs):
1167 Return the data of the page as a string.
1170 tries -- number of tries
1171 timeout -- sleep interval between tries
1173 See _download_webpage_handle docstring for other arguments specification.
1176 R''' # NB: These are unused; should they be deprecated?
1178 self._downloader.deprecation_warning('tries argument
is deprecated
in InfoExtractor
._download
_webpage
')
1179 if timeout is NO_DEFAULT:
1182 self._downloader.deprecation_warning('timeout argument
is deprecated
in InfoExtractor
._download
_webpage
')
1188 return self.__download_webpage(url_or_request, video_id, note, errnote, None, fatal, *args, **kwargs)
1189 except IncompleteRead as e:
1191 if try_count >= tries:
1193 self._sleep(timeout, video_id)
1195 def report_warning(self, msg, video_id=None, *args, only_once=False, **kwargs):
1196 idstr = format_field(video_id, None, '%s: ')
1197 msg = f'[{self.IE_NAME}
] {idstr}{msg}
'
1199 if f'WARNING
: {msg}
' in self._printed_messages:
1201 self._printed_messages.add(f'WARNING
: {msg}
')
1202 self._downloader.report_warning(msg, *args, **kwargs)
1204 def to_screen(self, msg, *args, **kwargs):
1205 """Print msg to screen, prefixing it with '[ie_name
]'"""
1206 self._downloader.to_screen(f'[{self.IE_NAME}
] {msg}
', *args, **kwargs)
1208 def write_debug(self, msg, *args, **kwargs):
1209 self._downloader.write_debug(f'[{self.IE_NAME}
] {msg}
', *args, **kwargs)
1211 def get_param(self, name, default=None, *args, **kwargs):
1212 if self._downloader:
1213 return self._downloader.params.get(name, default, *args, **kwargs)
1216 def report_drm(self, video_id, partial=NO_DEFAULT):
1217 if partial is not NO_DEFAULT:
1218 self._downloader.deprecation_warning('InfoExtractor
.report_drm no longer accepts the argument partial
')
1219 self.raise_no_formats('This video
is DRM protected
', expected=True, video_id=video_id)
1221 def report_extraction(self, id_or_name):
1222 """Report information extraction."""
1223 self.to_screen(f'{id_or_name}
: Extracting information
')
1225 def report_download_webpage(self, video_id):
1226 """Report webpage download."""
1227 self.to_screen(f'{video_id}
: Downloading webpage
')
1229 def report_age_confirmation(self):
1230 """Report attempt to confirm age."""
1231 self.to_screen('Confirming age
')
1233 def report_login(self):
1234 """Report attempt to log in."""
1235 self.to_screen('Logging
in')
1237 def raise_login_required(
1238 self, msg='This video
is only available
for registered users
',
1239 metadata_available=False, method=NO_DEFAULT):
1240 if metadata_available and (
1241 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1242 self.report_warning(msg)
1244 msg += format_field(self._login_hint(method), None, '. %s')
1245 raise ExtractorError(msg, expected=True)
1247 def raise_geo_restricted(
1248 self, msg='This video
is not available
from your location due to geo restriction
',
1249 countries=None, metadata_available=False):
1250 if metadata_available and (
1251 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1252 self.report_warning(msg)
1254 raise GeoRestrictedError(msg, countries=countries)
1256 def raise_no_formats(self, msg, expected=False, video_id=None):
1258 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1259 self.report_warning(msg, video_id)
1260 elif isinstance(msg, ExtractorError):
1263 raise ExtractorError(msg, expected=expected, video_id=video_id)
1265 # Methods for following #608
1267 def url_result(url, ie=None, video_id=None, video_title=None, *, url_transparent=False, **kwargs):
1268 """Returns a URL that points to a page that should be processed"""
1270 kwargs['ie_key
'] = ie if isinstance(ie, str) else ie.ie_key()
1271 if video_id is not None:
1272 kwargs['id'] = video_id
1273 if video_title is not None:
1274 kwargs['title
'] = video_title
1277 '_type
': 'url_transparent
' if url_transparent else 'url
',
1282 def playlist_from_matches(cls, matches, playlist_id=None, playlist_title=None,
1283 getter=IDENTITY, ie=None, video_kwargs=None, **kwargs):
1284 return cls.playlist_result(
1285 (cls.url_result(m, ie, **(video_kwargs or {})) for m in orderedSet(map(getter, matches), lazy=True)),
1286 playlist_id, playlist_title, **kwargs)
1289 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None, *, multi_video=False, **kwargs):
1290 """Returns a playlist"""
1292 kwargs['id'] = playlist_id
1294 kwargs['title
'] = playlist_title
1295 if playlist_description is not None:
1296 kwargs['description
'] = playlist_description
1299 '_type
': 'multi_video
' if multi_video else 'playlist
',
1303 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1305 Perform a regex search on the given string, using a single or a list of
1306 patterns returning the first matching group.
1307 In case of failure return a default value or raise a WARNING or a
1308 RegexNotFoundError, depending on fatal, specifying the field name.
1312 elif isinstance(pattern, (str, re.Pattern)):
1313 mobj = re.search(pattern, string, flags)
1316 mobj = re.search(p, string, flags)
1320 _name = self._downloader._format_err(name, self._downloader.Styles.EMPHASIS)
1324 # return the first matching group
1325 return next(g for g in mobj.groups() if g is not None)
1326 elif isinstance(group, (list, tuple)):
1327 return tuple(mobj.group(g) for g in group)
1329 return mobj.group(group)
1330 elif default is not NO_DEFAULT:
1333 raise RegexNotFoundError(f'Unable to extract {_name}
')
1335 self.report_warning(f'unable to extract {_name}
' + bug_reports_message())
1338 def _search_json(self, start_pattern, string, name, video_id, *, end_pattern='',
1339 contains_pattern=r'{(?s:.+)}
', fatal=True, default=NO_DEFAULT, **kwargs):
1340 """Searches string for the JSON object specified by start_pattern"""
1341 # NB: end_pattern is only used to reduce the size of the initial match
1342 if default is NO_DEFAULT:
1343 default, has_default = {}, False
1345 fatal, has_default = False, True
1347 json_string = self._search_regex(
1348 rf'(?
:{start_pattern}
)\s
*(?P
<json
>{contains_pattern}
)\s
*(?
:{end_pattern}
)',
1349 string, name, group='json
', fatal=fatal, default=None if has_default else NO_DEFAULT)
1353 _name = self._downloader._format_err(name, self._downloader.Styles.EMPHASIS)
1355 return self._parse_json(json_string, video_id, ignore_extra=True, **kwargs)
1356 except ExtractorError as e:
1358 raise ExtractorError(
1359 f'Unable to extract {_name}
- Failed to parse JSON
', cause=e.cause, video_id=video_id)
1360 elif not has_default:
1361 self.report_warning(
1362 f'Unable to extract {_name}
- Failed to parse JSON
: {e}
', video_id=video_id)
1365 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1367 Like _search_regex, but strips HTML tags and unescapes entities.
1369 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1370 if isinstance(res, tuple):
1371 return tuple(map(clean_html, res))
1372 return clean_html(res)
1374 def _get_netrc_login_info(self, netrc_machine=None):
1375 netrc_machine = netrc_machine or self._NETRC_MACHINE
1377 cmd = self.get_param('netrc_cmd
')
1379 cmd = cmd.replace('{}', netrc_machine)
1380 self.to_screen(f'Executing command
: {cmd}
')
1381 stdout, _, ret = Popen.run(cmd, text=True, shell=True, stdout=subprocess.PIPE)
1383 raise OSError(f'Command returned error code {ret}
')
1384 info = netrc_from_content(stdout).authenticators(netrc_machine)
1386 elif self.get_param('usenetrc
', False):
1387 netrc_file = compat_expanduser(self.get_param('netrc_location
') or '~
')
1388 if os.path.isdir(netrc_file):
1389 netrc_file = os.path.join(netrc_file, '.netrc
')
1390 info = netrc.netrc(netrc_file).authenticators(netrc_machine)
1395 self.to_screen(f'No authenticators
for {netrc_machine}
')
1398 self.write_debug(f'Using netrc
for {netrc_machine} authentication
')
1399 return info[0], info[2]
1401 def _get_login_info(self, username_option='username
', password_option='password
', netrc_machine=None):
1403 Get the login info as (username, password)
1404 First look for the manually specified credentials using username_option
1405 and password_option as keys in params dictionary. If no such credentials
1406 are available try the netrc_cmd if it is defined or look in the
1407 netrc file using the netrc_machine or _NETRC_MACHINE value.
1408 If there's no info available
, return (None, None)
1411 username = self.get_param(username_option)
1412 if username is not None:
1413 password = self.get_param(password_option)
1416 username, password = self._get_netrc_login_info(netrc_machine)
1417 except (OSError, netrc.NetrcParseError) as err:
1418 self.report_warning(f'Failed to parse .netrc: {err}')
1420 return username, password
1422 def _get_tfa_info(self, note='two-factor verification code'):
1424 Get the two
-factor authentication info
1425 TODO
- asking the user will be required
for sms
/phone verify
1426 currently just uses the command line option
1427 If there
's no info available, return None
1430 tfa = self.get_param('twofactor
')
1434 return getpass.getpass(f'Type {note}
and press
[Return
]: ')
1436 # Helper functions for extracting OpenGraph info
1438 def _og_regexes(prop):
1439 content_re = r'content
=(?
:"([^"]+?
)"|\'([^\']+?)\'|\s*([^\s"\'=<>`
]+?
)(?
=\s|
/?
>))'
1440 property_re = r'(?
:name|
property)=(?
:\'og{sep}{prop}
\'|
"og{sep}{prop}"|\s
*og{sep}{prop}
\b)'.format(
1441 prop=re.escape(prop), sep='(?
::|[:-])')
1442 template
= r
'<meta[^>]+?%s[^>]+?%s'
1444 template
% (property_re
, content_re
),
1445 template
% (content_re
, property_re
),
1449 def _meta_regex(prop
):
1450 return rf
'''(?isx)<meta
1451 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?){re.escape(prop)}\1)
1452 [^>]+?content=(["\'])(?P<content>.*?)\2'''
1454 def _og_search_property(self
, prop
, html
, name
=None, **kargs
):
1455 prop
= variadic(prop
)
1457 name
= f
'OpenGraph {prop[0]}'
1460 og_regexes
.extend(self
._og
_regexes
(p
))
1461 escaped
= self
._search
_regex
(og_regexes
, html
, name
, flags
=re
.DOTALL
, **kargs
)
1464 return unescapeHTML(escaped
)
1466 def _og_search_thumbnail(self
, html
, **kargs
):
1467 return self
._og
_search
_property
('image', html
, 'thumbnail URL', fatal
=False, **kargs
)
1469 def _og_search_description(self
, html
, **kargs
):
1470 return self
._og
_search
_property
('description', html
, fatal
=False, **kargs
)
1472 def _og_search_title(self
, html
, *, fatal
=False, **kargs
):
1473 return self
._og
_search
_property
('title', html
, fatal
=fatal
, **kargs
)
1475 def _og_search_video_url(self
, html
, name
='video url', secure
=True, **kargs
):
1476 regexes
= self
._og
_regexes
('video') + self
._og
_regexes
('video:url')
1478 regexes
= self
._og
_regexes
('video:secure_url') + regexes
1479 return self
._html
_search
_regex
(regexes
, html
, name
, **kargs
)
1481 def _og_search_url(self
, html
, **kargs
):
1482 return self
._og
_search
_property
('url', html
, **kargs
)
1484 def _html_extract_title(self
, html
, name
='title', *, fatal
=False, **kwargs
):
1485 return self
._html
_search
_regex
(r
'(?s)<title\b[^>]*>([^<]+)</title>', html
, name
, fatal
=fatal
, **kwargs
)
1487 def _html_search_meta(self
, name
, html
, display_name
=None, fatal
=False, **kwargs
):
1488 name
= variadic(name
)
1489 if display_name
is None:
1490 display_name
= name
[0]
1491 return self
._html
_search
_regex
(
1492 [self
._meta
_regex
(n
) for n
in name
],
1493 html
, display_name
, fatal
=fatal
, group
='content', **kwargs
)
1495 def _dc_search_uploader(self
, html
):
1496 return self
._html
_search
_meta
('dc.creator', html
, 'uploader')
1499 def _rta_search(html
):
1500 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1501 if re
.search(r
'(?ix)<meta\s+name="rating"\s+'
1502 r
' content="RTA-5042-1996-1400-1577-RTA"',
1506 # And then there are the jokers who advertise that they use RTA, but actually don't.
1507 AGE_LIMIT_MARKERS
= [
1508 r
'Proudly Labeled <a href="http://www\.rtalabel\.org/" title="Restricted to Adults">RTA</a>',
1509 r
'>[^<]*you acknowledge you are at least (\d+) years old',
1510 r
'>\s*(?:18\s+U(?:\.S\.C\.|SC)\s+)?(?:§+\s*)?2257\b',
1514 for marker
in AGE_LIMIT_MARKERS
:
1515 mobj
= re
.search(marker
, html
)
1517 age_limit
= max(age_limit
, int(traverse_obj(mobj
, 1, default
=18)))
1520 def _media_rating_search(self
, html
):
1521 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1522 rating
= self
._html
_search
_meta
('rating', html
)
1534 return RATING_TABLE
.get(rating
.lower())
1536 def _family_friendly_search(self
, html
):
1537 # See http://schema.org/VideoObject
1538 family_friendly
= self
._html
_search
_meta
(
1539 'isFamilyFriendly', html
, default
=None)
1541 if not family_friendly
:
1550 return RATING_TABLE
.get(family_friendly
.lower())
1552 def _twitter_search_player(self
, html
):
1553 return self
._html
_search
_meta
('twitter:player', html
,
1554 'twitter card player')
1556 def _yield_json_ld(self
, html
, video_id
, *, fatal
=True, default
=NO_DEFAULT
):
1557 """Yield all json ld objects in the html"""
1558 if default
is not NO_DEFAULT
:
1560 for mobj
in re
.finditer(JSON_LD_RE
, html
):
1561 json_ld_item
= self
._parse
_json
(mobj
.group('json_ld'), video_id
, fatal
=fatal
)
1562 for json_ld
in variadic(json_ld_item
):
1563 if isinstance(json_ld
, dict):
1566 def _search_json_ld(self
, html
, video_id
, expected_type
=None, *, fatal
=True, default
=NO_DEFAULT
):
1567 """Search for a video in any json ld in the html"""
1568 if default
is not NO_DEFAULT
:
1570 info
= self
._json
_ld
(
1571 list(self
._yield
_json
_ld
(html
, video_id
, fatal
=fatal
, default
=default
)),
1572 video_id
, fatal
=fatal
, expected_type
=expected_type
)
1575 if default
is not NO_DEFAULT
:
1578 raise RegexNotFoundError('Unable to extract JSON-LD')
1580 self
.report_warning(f
'unable to extract JSON-LD {bug_reports_message()}')
1583 def _json_ld(self
, json_ld
, video_id
, fatal
=True, expected_type
=None):
1584 if isinstance(json_ld
, str):
1585 json_ld
= self
._parse
_json
(json_ld
, video_id
, fatal
=fatal
)
1590 INTERACTION_TYPE_MAP
= {
1591 'CommentAction': 'comment',
1592 'AgreeAction': 'like',
1593 'DisagreeAction': 'dislike',
1594 'LikeAction': 'like',
1595 'DislikeAction': 'dislike',
1596 'ListenAction': 'view',
1597 'WatchAction': 'view',
1598 'ViewAction': 'view',
1601 def is_type(e
, *expected_types
):
1602 type_
= variadic(traverse_obj(e
, '@type'))
1603 return any(x
in type_
for x
in expected_types
)
1605 def extract_interaction_type(e
):
1606 interaction_type
= e
.get('interactionType')
1607 if isinstance(interaction_type
, dict):
1608 interaction_type
= interaction_type
.get('@type')
1609 return str_or_none(interaction_type
)
1611 def extract_interaction_statistic(e
):
1612 interaction_statistic
= e
.get('interactionStatistic')
1613 if isinstance(interaction_statistic
, dict):
1614 interaction_statistic
= [interaction_statistic
]
1615 if not isinstance(interaction_statistic
, list):
1617 for is_e
in interaction_statistic
:
1618 if not is_type(is_e
, 'InteractionCounter'):
1620 interaction_type
= extract_interaction_type(is_e
)
1621 if not interaction_type
:
1623 # For interaction count some sites provide string instead of
1624 # an integer (as per spec) with non digit characters (e.g. ",")
1625 # so extracting count with more relaxed str_to_int
1626 interaction_count
= str_to_int(is_e
.get('userInteractionCount'))
1627 if interaction_count
is None:
1629 count_kind
= INTERACTION_TYPE_MAP
.get(interaction_type
.split('/')[-1])
1632 count_key
= f
'{count_kind}_count'
1633 if info
.get(count_key
) is not None:
1635 info
[count_key
] = interaction_count
1637 def extract_chapter_information(e
):
1639 'title': part
.get('name'),
1640 'start_time': part
.get('startOffset'),
1641 'end_time': part
.get('endOffset'),
1642 } for part
in variadic(e
.get('hasPart') or []) if part
.get('@type') == 'Clip']
1643 for idx
, (last_c
, current_c
, next_c
) in enumerate(zip(
1644 [{'end_time': 0}
, *chapters
], chapters
, chapters
[1:])):
1645 current_c
['end_time'] = current_c
['end_time'] or next_c
['start_time']
1646 current_c
['start_time'] = current_c
['start_time'] or last_c
['end_time']
1647 if None in current_c
.values():
1648 self
.report_warning(f
'Chapter {idx} contains broken data. Not extracting chapters')
1651 chapters
[-1]['end_time'] = chapters
[-1]['end_time'] or info
['duration']
1652 info
['chapters'] = chapters
1654 def extract_video_object(e
):
1655 author
= e
.get('author')
1657 'url': url_or_none(e
.get('contentUrl')),
1658 'ext': mimetype2ext(e
.get('encodingFormat')),
1659 'title': unescapeHTML(e
.get('name')),
1660 'description': unescapeHTML(e
.get('description')),
1661 'thumbnails': [{'url': unescapeHTML(url)}
1662 for url
in variadic(traverse_obj(e
, 'thumbnailUrl', 'thumbnailURL'))
1663 if url_or_none(url
)],
1664 'duration': parse_duration(e
.get('duration')),
1665 'timestamp': unified_timestamp(e
.get('uploadDate')),
1666 # author can be an instance of 'Organization' or 'Person' types.
1667 # both types can have 'name' property(inherited from 'Thing' type). [1]
1668 # however some websites are using 'Text' type instead.
1669 # 1. https://schema.org/VideoObject
1670 'uploader': author
.get('name') if isinstance(author
, dict) else author
if isinstance(author
, str) else None,
1671 'artist': traverse_obj(e
, ('byArtist', 'name'), expected_type
=str),
1672 'filesize': int_or_none(float_or_none(e
.get('contentSize'))),
1673 'tbr': int_or_none(e
.get('bitrate')),
1674 'width': int_or_none(e
.get('width')),
1675 'height': int_or_none(e
.get('height')),
1676 'view_count': int_or_none(e
.get('interactionCount')),
1677 'tags': try_call(lambda: e
.get('keywords').split(',')),
1679 if is_type(e
, 'AudioObject'):
1682 'abr': int_or_none(e
.get('bitrate')),
1684 extract_interaction_statistic(e
)
1685 extract_chapter_information(e
)
1687 def traverse_json_ld(json_ld
, at_top_level
=True):
1688 for e
in variadic(json_ld
):
1689 if not isinstance(e
, dict):
1691 if at_top_level
and '@context' not in e
:
1693 if at_top_level
and set(e
.keys()) == {'@context', '@graph'}
:
1694 traverse_json_ld(e
['@graph'], at_top_level
=False)
1696 if expected_type
is not None and not is_type(e
, expected_type
):
1698 rating
= traverse_obj(e
, ('aggregateRating', 'ratingValue'), expected_type
=float_or_none
)
1699 if rating
is not None:
1700 info
['average_rating'] = rating
1701 if is_type(e
, 'TVEpisode', 'Episode'):
1702 episode_name
= unescapeHTML(e
.get('name'))
1704 'episode': episode_name
,
1705 'episode_number': int_or_none(e
.get('episodeNumber')),
1706 'description': unescapeHTML(e
.get('description')),
1708 if not info
.get('title') and episode_name
:
1709 info
['title'] = episode_name
1710 part_of_season
= e
.get('partOfSeason')
1711 if is_type(part_of_season
, 'TVSeason', 'Season', 'CreativeWorkSeason'):
1713 'season': unescapeHTML(part_of_season
.get('name')),
1714 'season_number': int_or_none(part_of_season
.get('seasonNumber')),
1716 part_of_series
= e
.get('partOfSeries') or e
.get('partOfTVSeries')
1717 if is_type(part_of_series
, 'TVSeries', 'Series', 'CreativeWorkSeries'):
1718 info
['series'] = unescapeHTML(part_of_series
.get('name'))
1719 elif is_type(e
, 'Movie'):
1721 'title': unescapeHTML(e
.get('name')),
1722 'description': unescapeHTML(e
.get('description')),
1723 'duration': parse_duration(e
.get('duration')),
1724 'timestamp': unified_timestamp(e
.get('dateCreated')),
1726 elif is_type(e
, 'Article', 'NewsArticle'):
1728 'timestamp': parse_iso8601(e
.get('datePublished')),
1729 'title': unescapeHTML(e
.get('headline')),
1730 'description': unescapeHTML(e
.get('articleBody') or e
.get('description')),
1732 if is_type(traverse_obj(e
, ('video', 0)), 'VideoObject'):
1733 extract_video_object(e
['video'][0])
1734 elif is_type(traverse_obj(e
, ('subjectOf', 0)), 'VideoObject'):
1735 extract_video_object(e
['subjectOf'][0])
1736 elif is_type(e
, 'VideoObject', 'AudioObject'):
1737 extract_video_object(e
)
1738 if expected_type
is None:
1742 video
= e
.get('video')
1743 if is_type(video
, 'VideoObject'):
1744 extract_video_object(video
)
1745 if expected_type
is None:
1750 traverse_json_ld(json_ld
)
1751 return filter_dict(info
)
1753 def _search_nextjs_data(self
, webpage
, video_id
, *, fatal
=True, default
=NO_DEFAULT
, **kw
):
1755 self
._downloader
.deprecation_warning('using `default=\'{}\'` is deprecated, use `default={}` instead')
1757 if default
is not NO_DEFAULT
:
1760 return self
._search
_json
(
1761 r
'<script[^>]+id=[\'"]__NEXT_DATA__[\'"][^
>]*>', webpage, 'next
.js data
',
1762 video_id, end_pattern='</script
>', fatal=fatal, default=default, **kw)
1764 def _search_nuxt_data(self, webpage, video_id, context_name='__NUXT__
', *, fatal=True, traverse=('data
', 0)):
1765 """Parses Nuxt.js metadata. This works as long as the function __NUXT__ invokes is a pure function"""
1766 rectx = re.escape(context_name)
1767 FUNCTION_RE = r'\
(function\
((?P
<arg_keys
>.*?
)\
){.*?\breturn\s+(?P<js>{.*?}
)\s
*;?\s
*}\
((?P
<arg_vals
>.*?
)\
)'
1768 js, arg_keys, arg_vals = self._search_regex(
1769 (rf'<script
>\s
*window\
.{rectx}
={FUNCTION_RE}\s
*\
)\s
*;?\s
*</script
>', rf'{rectx}\
(.*?{FUNCTION_RE}
'),
1770 webpage, context_name, group=('js
', 'arg_keys
', 'arg_vals
'),
1771 default=NO_DEFAULT if fatal else (None, None, None))
1775 args = dict(zip(arg_keys.split(','), map(json.dumps, self._parse_json(
1776 f'[{arg_vals}
]', video_id, transform_source=js_to_json, fatal=fatal) or ())))
1778 ret = self._parse_json(js, video_id, transform_source=functools.partial(js_to_json, vars=args), fatal=fatal)
1779 return traverse_obj(ret, traverse) or {}
1782 def _hidden_inputs(html):
1783 html = re.sub(r'<!--(?
:(?
!<!--).)*-->', '', html)
1785 for input_el in re.findall(r'(?i
)(<input[^
>]+>)', html):
1786 attrs = extract_attributes(input_el)
1789 if attrs.get('type') not in ('hidden
', 'submit
'):
1791 name = attrs.get('name
') or attrs.get('id')
1792 value = attrs.get('value
')
1793 if name and value is not None:
1794 hidden_inputs[name] = value
1795 return hidden_inputs
1797 def _form_hidden_inputs(self, form_id, html):
1798 form = self._search_regex(
1799 rf'(?
is)<form
[^
>]+?
id=(["\']){form_id}\1[^>]*>(?P<form>.+?)</form>',
1800 html, f'{form_id} form', group='form')
1801 return self._hidden_inputs(form)
1803 @classproperty(cache=True)
1804 def FormatSort(cls):
1805 class FormatSort(FormatSorter):
1806 def __init__(ie, *args, **kwargs):
1807 super().__init__(ie._downloader, *args, **kwargs)
1809 deprecation_warning(
1810 'yt_dlp.InfoExtractor.FormatSort is deprecated and may be removed in the future. '
1811 'Use yt_dlp.utils.FormatSorter instead')
1814 def _sort_formats(self, formats, field_preference=[]):
1815 if not field_preference:
1816 self._downloader.deprecation_warning(
1817 'yt_dlp.InfoExtractor._sort_formats is deprecated and is no longer required')
1819 self._downloader.deprecation_warning(
1820 'yt_dlp.InfoExtractor._sort_formats is deprecated and no longer works as expected. '
1821 'Return _format_sort_fields in the info_dict instead')
1823 formats[0]['__sort_fields'] = field_preference
1825 def _check_formats(self, formats, video_id):
1827 formats[:] = filter(
1828 lambda f: self._is_valid_url(
1830 item='{} video format'.format(f.get('format_id')) if f.get('format_id') else 'video'),
1834 def _remove_duplicate_formats(formats):
1838 if f['url'] not in format_urls:
1839 format_urls.add(f['url'])
1840 unique_formats.append(f)
1841 formats[:] = unique_formats
1843 def _is_valid_url(self, url, video_id, item='video', headers={}):
1844 url = self._proto_relative_url(url, scheme='http:')
1845 # For now assume non HTTP(S) URLs always valid
1846 if not url.startswith(('http://', 'https://')):
1849 self._request_webpage(url, video_id, f'Checking {item} URL', headers=headers)
1851 except ExtractorError as e:
1853 f'{video_id}: {item} URL is invalid, skipping: {e.cause!s}')
1856 def http_scheme(self):
1857 """ Either "http
:" or "https
:", depending on the user's preferences """
1860 if self.get_param('prefer_insecure', False)
1863 def _proto_relative_url(self, url, scheme=None):
1864 scheme = scheme or self.http_scheme()
1865 assert scheme.endswith(':')
1866 return sanitize_url(url, scheme=scheme[:-1])
1868 def _sleep(self, timeout, video_id, msg_template=None):
1869 if msg_template is None:
1870 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1871 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1875 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1876 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1877 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
1878 if self.get_param('ignore_no_formats_error'):
1881 res = self._download_xml_handle(
1882 manifest_url, video_id, 'Downloading f4m manifest',
1883 'Unable to download f4m manifest',
1884 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1885 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
1886 transform_source=transform_source,
1887 fatal=fatal, data=data, headers=headers, query=query)
1891 manifest, urlh = res
1892 manifest_url = urlh.url
1894 return self._parse_f4m_formats(
1895 manifest, manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1896 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1898 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1899 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1900 fatal=True, m3u8_id=None):
1901 if not isinstance(manifest, xml.etree.ElementTree.Element) and not fatal:
1904 # currently yt-dlp cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1905 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1906 if akamai_pv is not None and ';' in akamai_pv.text:
1907 player_verification_challenge = akamai_pv.text.split(';')[0]
1908 if player_verification_challenge.strip() != '':
1912 manifest_version = '1.0'
1913 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1915 manifest_version = '2.0'
1916 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1917 # Remove unsupported DRM protected media from final formats
1918 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
1919 media_nodes = remove_encrypted_media(media_nodes)
1923 manifest_base_url = get_base_url(manifest)
1925 bootstrap_info = xpath_element(
1926 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1927 'bootstrap info', default=None)
1930 mime_type = xpath_text(
1931 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1932 'base URL', default=None)
1933 if mime_type and mime_type.startswith('audio/'):
1936 for i, media_el in enumerate(media_nodes):
1937 tbr = int_or_none(media_el.attrib.get('bitrate'))
1938 width = int_or_none(media_el.attrib.get('width'))
1939 height = int_or_none(media_el.attrib.get('height'))
1940 format_id = join_nonempty(f4m_id, tbr or i)
1941 # If <bootstrapInfo> is present, the specified f4m is a
1942 # stream-level manifest, and only set-level manifests may refer to
1943 # external resources. See section 11.4 and section 4 of F4M spec
1944 if bootstrap_info is None:
1946 # @href is introduced in 2.0, see section 11.6 of F4M spec
1947 if manifest_version == '2.0':
1948 media_url = media_el.attrib.get('href')
1949 if media_url is None:
1950 media_url = media_el.attrib.get('url')
1954 media_url if media_url.startswith(('http://', 'https://'))
1955 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1956 # If media_url is itself a f4m manifest do the recursive extraction
1957 # since bitrates in parent manifest (this one) and media_url manifest
1958 # may differ leading to inability to resolve the format by requested
1959 # bitrate in f4m downloader
1960 ext = determine_ext(manifest_url)
1962 f4m_formats = self._extract_f4m_formats(
1963 manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1964 transform_source=transform_source, fatal=fatal)
1965 # Sometimes stream-level manifest contains single media entry that
1966 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1967 # At the same time parent's media entry in set-level manifest may
1968 # contain it. We will copy it from parent in such cases.
1969 if len(f4m_formats) == 1:
1972 'tbr': f.get('tbr') or tbr,
1973 'width': f.get('width') or width,
1974 'height': f.get('height') or height,
1975 'format_id': f.get('format_id') if not tbr else format_id,
1978 formats.extend(f4m_formats)
1981 formats.extend(self._extract_m3u8_formats(
1982 manifest_url, video_id, 'mp4', preference=preference,
1983 quality=quality, m3u8_id=m3u8_id, fatal=fatal))
1986 'format_id': format_id,
1987 'url': manifest_url,
1988 'manifest_url': manifest_url,
1989 'ext': 'flv' if bootstrap_info is not None else None,
1995 'preference': preference,
2000 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, quality=None, m3u8_id=None):
2002 'format_id': join_nonempty(m3u8_id, 'meta'),
2006 'preference': preference - 100 if preference else -100,
2008 'resolution': 'multiple',
2009 'format_note': 'Quality selection URL',
2012 def _report_ignoring_subs(self, name):
2013 self.report_warning(bug_reports_message(
2014 f'Ignoring subtitle tracks found in the {name} manifest; '
2015 'if any subtitle tracks are missing,',
2018 def _extract_m3u8_formats(self, *args, **kwargs):
2019 fmts, subs = self._extract_m3u8_formats_and_subtitles(*args, **kwargs)
2021 self._report_ignoring_subs('HLS')
2024 def _extract_m3u8_formats_and_subtitles(
2025 self, m3u8_url, video_id, ext=None, entry_protocol='m3u8_native',
2026 preference=None, quality=None, m3u8_id=None, note=None,
2027 errnote=None, fatal=True, live=False, data=None, headers={},
2030 if self.get_param('ignore_no_formats_error'):
2034 if errnote is not False:
2035 errnote = errnote or 'Failed to obtain m3u8 URL'
2037 raise ExtractorError(errnote, video_id=video_id)
2038 self.report_warning(f'{errnote}{bug_reports_message()}')
2041 res = self._download_webpage_handle(
2043 note='Downloading m3u8 information' if note is None else note,
2044 errnote='Failed to download m3u8 information' if errnote is None else errnote,
2045 fatal=fatal, data=data, headers=headers, query=query)
2050 m3u8_doc, urlh = res
2053 return self._parse_m3u8_formats_and_subtitles(
2054 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
2055 preference=preference, quality=quality, m3u8_id=m3u8_id,
2056 note=note, errnote=errnote, fatal=fatal, live=live, data=data,
2057 headers=headers, query=query, video_id=video_id)
2059 def _parse_m3u8_formats_and_subtitles(
2060 self, m3u8_doc, m3u8_url=None, ext=None, entry_protocol='m3u8_native',
2061 preference=None, quality=None, m3u8_id=None, live=False, note=None,
2062 errnote=None, fatal=True, data=None, headers={}, query={},
2064 formats, subtitles = [], {}
2065 has_drm = HlsFD._has_drm(m3u8_doc)
2067 def format_url(url):
2068 return url if re.match(r'^https?://', url) else urllib.parse.urljoin(m3u8_url, url)
2070 if self.get_param('hls_split_discontinuity', False):
2071 def _extract_m3u8_playlist_indices(manifest_url=None, m3u8_doc=None):
2073 if not manifest_url:
2075 m3u8_doc = self._download_webpage(
2076 manifest_url, video_id, fatal=fatal, data=data, headers=headers,
2077 note=False, errnote='Failed to download m3u8 playlist information')
2078 if m3u8_doc is False:
2080 return range(1 + sum(line.startswith('#EXT-X-DISCONTINUITY') for line in m3u8_doc.splitlines()))
2083 def _extract_m3u8_playlist_indices(*args, **kwargs):
2087 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
2088 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
2089 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
2091 # We should try extracting formats only from master playlists [1, 4.3.4],
2092 # i.e. playlists that describe available qualities. On the other hand
2093 # media playlists [1, 4.3.3] should be returned as is since they contain
2094 # just the media without qualities renditions.
2095 # Fortunately, master playlist can be easily distinguished from media
2096 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
2097 # master playlist tags MUST NOT appear in a media playlist and vice versa.
2098 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
2099 # media playlist and MUST NOT appear in master playlist thus we can
2100 # clearly detect media playlist with this criterion.
2102 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
2104 'format_id': join_nonempty(m3u8_id, idx),
2105 'format_index': idx,
2106 'url': m3u8_url or encode_data_uri(m3u8_doc.encode(), 'application/x-mpegurl'),
2108 'protocol': entry_protocol,
2109 'preference': preference,
2112 } for idx in _extract_m3u8_playlist_indices(m3u8_doc=m3u8_doc)]
2114 return formats, subtitles
2117 last_stream_inf = {}
2119 def extract_media(x_media_line):
2120 media = parse_m3u8_attributes(x_media_line)
2121 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
2122 media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME')
2123 if not (media_type and group_id and name):
2125 groups.setdefault(group_id, []).append(media)
2126 # <https://tools.ietf.org/html/rfc8216#section-4.3.4.1>
2127 if media_type == 'SUBTITLES':
2128 # According to RFC 8216 §4.3.4.2.1, URI is REQUIRED in the
2129 # EXT-X-MEDIA tag if the media type is SUBTITLES.
2130 # However, lack of URI has been spotted in the wild.
2131 # e.g. NebulaIE; see https://github.com/yt-dlp/yt-dlp/issues/339
2132 if not media.get('URI'):
2134 url = format_url(media['URI'])
2137 'ext': determine_ext(url),
2139 if sub_info['ext'] == 'm3u8':
2140 # Per RFC 8216 §3.1, the only possible subtitle format m3u8
2141 # files may contain is WebVTT:
2142 # <https://tools.ietf.org/html/rfc8216#section-3.1>
2143 sub_info['ext'] = 'vtt'
2144 sub_info['protocol'] = 'm3u8_native'
2145 lang = media.get('LANGUAGE') or 'und'
2146 subtitles.setdefault(lang, []).append(sub_info)
2147 if media_type not in ('VIDEO', 'AUDIO'):
2149 media_url = media.get('URI')
2151 manifest_url = format_url(media_url)
2153 'format_id': join_nonempty(m3u8_id, group_id, name, idx),
2154 'format_note': name,
2155 'format_index': idx,
2156 'url': manifest_url,
2157 'manifest_url': m3u8_url,
2158 'language': media.get('LANGUAGE'),
2160 'protocol': entry_protocol,
2161 'preference': preference,
2164 'vcodec': 'none' if media_type == 'AUDIO' else None,
2165 } for idx in _extract_m3u8_playlist_indices(manifest_url))
2167 def build_stream_name():
2168 # Despite specification does not mention NAME attribute for
2169 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
2170 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
2171 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
2172 stream_name = last_stream_inf.get('NAME')
2175 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
2176 # from corresponding rendition group
2177 stream_group_id = last_stream_inf.get('VIDEO')
2178 if not stream_group_id:
2180 stream_group = groups.get(stream_group_id)
2181 if not stream_group:
2182 return stream_group_id
2183 rendition = stream_group[0]
2184 return rendition.get('NAME') or stream_group_id
2186 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
2187 # chance to detect video only formats when EXT-X-STREAM-INF tags
2188 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
2189 for line in m3u8_doc.splitlines():
2190 if line.startswith('#EXT-X-MEDIA:'):
2193 for line in m3u8_doc.splitlines():
2194 if line.startswith('#EXT-X-STREAM-INF:'):
2195 last_stream_inf = parse_m3u8_attributes(line)
2196 elif line.startswith('#') or not line.strip():
2199 tbr = float_or_none(
2200 last_stream_inf.get('AVERAGE-BANDWIDTH')
2201 or last_stream_inf.get('BANDWIDTH'), scale=1000)
2202 manifest_url = format_url(line.strip())
2204 for idx in _extract_m3u8_playlist_indices(manifest_url):
2205 format_id = [m3u8_id, None, idx]
2206 # Bandwidth of live streams may differ over time thus making
2207 # format_id unpredictable. So it's better to keep provided
2210 stream_name = build_stream_name()
2211 format_id[1] = stream_name or '%d' % (tbr or len(formats))
2213 'format_id': join_nonempty(*format_id),
2214 'format_index': idx,
2215 'url': manifest_url,
2216 'manifest_url': m3u8_url,
2219 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')),
2220 'protocol': entry_protocol,
2221 'preference': preference,
2227 if yt_audio_content_id := last_stream_inf.get('YT-EXT-AUDIO-CONTENT-ID'):
2228 f['language'] = yt_audio_content_id.split('.')[0]
2230 resolution = last_stream_inf.get('RESOLUTION')
2232 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
2234 f['width'] = int(mobj.group('width'))
2235 f['height'] = int(mobj.group('height'))
2236 # Unified Streaming Platform
2238 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
2240 abr, vbr = mobj.groups()
2241 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
2246 codecs = parse_codecs(last_stream_inf.get('CODECS'))
2248 audio_group_id = last_stream_inf.get('AUDIO')
2249 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
2250 # references a rendition group MUST have a CODECS attribute.
2251 # However, this is not always respected. E.g. [2]
2252 # contains EXT-X-STREAM-INF tag which references AUDIO
2253 # rendition group but does not have CODECS and despite
2254 # referencing an audio group it represents a complete
2255 # (with audio and video) format. So, for such cases we will
2256 # ignore references to rendition groups and treat them
2257 # as complete formats.
2258 if audio_group_id and codecs and f.get('vcodec') != 'none':
2259 audio_group = groups.get(audio_group_id)
2260 if audio_group and audio_group[0].get('URI'):
2261 # TODO: update acodec for audio only formats with
2263 f['acodec'] = 'none'
2264 if not f.get('ext'):
2265 f['ext'] = 'm4a' if f.get('vcodec') == 'none' else 'mp4'
2269 progressive_uri = last_stream_inf.get('PROGRESSIVE-URI')
2272 del http_f['manifest_url']
2274 'format_id': f['format_id'].replace('hls-', 'http-'),
2276 'url': progressive_uri,
2278 formats.append(http_f)
2280 last_stream_inf = {}
2281 return formats, subtitles
2283 def _extract_m3u8_vod_duration(
2284 self, m3u8_vod_url, video_id, note=None, errnote=None, data=None, headers={}, query={}):
2286 m3u8_vod = self._download_webpage(
2287 m3u8_vod_url, video_id,
2288 note='Downloading m3u8 VOD manifest' if note is None else note,
2289 errnote='Failed to download VOD manifest' if errnote is None else errnote,
2290 fatal=False, data=data, headers=headers, query=query)
2292 return self._parse_m3u8_vod_duration(m3u8_vod or '', video_id)
2294 def _parse_m3u8_vod_duration(self, m3u8_vod, video_id):
2295 if '#EXT-X-ENDLIST' not in m3u8_vod:
2299 float(line[len('#EXTINF:'):].split(',')[0])
2300 for line in m3u8_vod.splitlines() if line.startswith('#EXTINF:'))) or None
2302 def _extract_mpd_vod_duration(
2303 self, mpd_url, video_id, note=None, errnote=None, data=None, headers={}, query={}):
2305 mpd_doc = self._download_xml(
2307 note='Downloading MPD VOD manifest' if note is None else note,
2308 errnote='Failed to download VOD manifest' if errnote is None else errnote,
2309 fatal=False, data=data, headers=headers, query=query)
2310 if not isinstance(mpd_doc, xml.etree.ElementTree.Element):
2312 return int_or_none(parse_duration(mpd_doc.get('mediaPresentationDuration')))
2315 def _xpath_ns(path, namespace=None):
2319 for c in path.split('/'):
2320 if not c or c == '.':
2323 out.append(f'{{{namespace}}}{c}')
2324 return '/'.join(out)
2326 def _extract_smil_formats_and_subtitles(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
2327 if self.get_param('ignore_no_formats_error'):
2330 res = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
2336 return self._parse_smil_formats_and_subtitles(smil, urlh.url, video_id, f4m_params=f4m_params,
2337 namespace=self._parse_smil_namespace(smil))
2339 def _extract_smil_formats(self, *args, **kwargs):
2340 fmts, subs = self._extract_smil_formats_and_subtitles(*args, **kwargs)
2342 self._report_ignoring_subs('SMIL')
2345 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
2346 res = self._download_smil(smil_url, video_id, fatal=fatal)
2353 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
2355 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
2356 return self._download_xml_handle(
2357 smil_url, video_id, 'Downloading SMIL file',
2358 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
2360 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
2361 namespace = self._parse_smil_namespace(smil)
2363 formats, subtitles = self._parse_smil_formats_and_subtitles(
2364 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
2366 video_id = os.path.splitext(url_basename(smil_url))[0]
2370 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
2371 name = meta.attrib.get('name')
2372 content = meta.attrib.get('content')
2373 if not name or not content:
2375 if not title and name == 'title':
2377 elif not description and name in ('description', 'abstract'):
2378 description = content
2379 elif not upload_date and name == 'date':
2380 upload_date = unified_strdate(content)
2383 'id': image.get('type'),
2384 'url': image.get('src'),
2385 'width': int_or_none(image.get('width')),
2386 'height': int_or_none(image.get('height')),
2387 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
2391 'title': title or video_id,
2392 'description': description,
2393 'upload_date': upload_date,
2394 'thumbnails': thumbnails,
2396 'subtitles': subtitles,
2399 def _parse_smil_namespace(self, smil):
2400 return self._search_regex(
2401 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
2403 def _parse_smil_formats(self, *args, **kwargs):
2404 fmts, subs = self._parse_smil_formats_and_subtitles(*args, **kwargs)
2406 self._report_ignoring_subs('SMIL')
2409 def _parse_smil_formats_and_subtitles(
2410 self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
2412 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
2413 b = meta.get('base') or meta.get('httpBase')
2418 formats, subtitles = [], {}
2425 media = itertools.chain.from_iterable(
2426 smil.findall(self._xpath_ns(arg, namespace))
2427 for arg in ['.//video', './/audio', './/media'])
2428 for medium in media:
2429 src = medium.get('src')
2430 if not src or src in srcs:
2434 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
2435 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
2436 width = int_or_none(medium.get('width'))
2437 height = int_or_none(medium.get('height'))
2438 proto = medium.get('proto')
2439 ext = medium.get('ext')
2440 src_ext = determine_ext(src, default_ext=None) or ext or urlhandle_detect_ext(
2441 self._request_webpage(HEADRequest(src), video_id, note='Requesting extension info', fatal=False))
2442 streamer = medium.get('streamer') or base
2444 if proto == 'rtmp' or streamer.startswith('rtmp'):
2450 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
2452 'filesize': filesize,
2456 if transform_rtmp_url:
2457 streamer, src = transform_rtmp_url(streamer, src)
2458 formats[-1].update({
2464 src_url = src if src.startswith('http') else urllib.parse.urljoin(f'{base}/', src)
2465 src_url = src_url.strip()
2467 if proto == 'm3u8' or src_ext == 'm3u8':
2468 m3u8_formats, m3u8_subs = self._extract_m3u8_formats_and_subtitles(
2469 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
2470 self._merge_subtitles(m3u8_subs, target=subtitles)
2471 if len(m3u8_formats) == 1:
2473 m3u8_formats[0].update({
2474 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
2479 formats.extend(m3u8_formats)
2480 elif src_ext == 'f4m':
2485 'plugin': 'flowplayer-3.2.0.1',
2487 f4m_url += '&' if '?' in f4m_url else '?'
2488 f4m_url += urllib.parse.urlencode(f4m_params)
2489 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
2490 elif src_ext == 'mpd':
2491 mpd_formats, mpd_subs = self._extract_mpd_formats_and_subtitles(
2492 src_url, video_id, mpd_id='dash', fatal=False)
2493 formats.extend(mpd_formats)
2494 self._merge_subtitles(mpd_subs, target=subtitles)
2495 elif re.search(r'\.ism/[Mm]anifest', src_url):
2496 ism_formats, ism_subs = self._extract_ism_formats_and_subtitles(
2497 src_url, video_id, ism_id='mss', fatal=False)
2498 formats.extend(ism_formats)
2499 self._merge_subtitles(ism_subs, target=subtitles)
2500 elif src_url.startswith('http') and self._is_valid_url(src, video_id):
2504 'ext': ext or src_ext or 'flv',
2505 'format_id': 'http-%d' % (bitrate or http_count),
2507 'filesize': filesize,
2512 for medium in smil.findall(self._xpath_ns('.//imagestream', namespace)):
2513 src = medium.get('src')
2514 if not src or src in srcs:
2520 'format_id': f'imagestream-{imgs_count}',
2522 'ext': mimetype2ext(medium.get('type')),
2525 'width': int_or_none(medium.get('width')),
2526 'height': int_or_none(medium.get('height')),
2527 'format_note': 'SMIL storyboards',
2530 smil_subs = self._parse_smil_subtitles(smil, namespace=namespace)
2531 self._merge_subtitles(smil_subs, target=subtitles)
2533 return formats, subtitles
2535 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
2538 for textstream in smil.findall(self._xpath_ns('.//textstream', namespace)):
2539 src = textstream.get('src')
2540 if not src or src in urls:
2543 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
2544 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
2545 subtitles.setdefault(lang, []).append({
2551 def _extract_xspf_playlist(self, xspf_url, playlist_id, fatal=True):
2552 res = self._download_xml_handle(
2553 xspf_url, playlist_id, 'Downloading xpsf playlist',
2554 'Unable to download xspf manifest', fatal=fatal)
2561 return self._parse_xspf(
2562 xspf, playlist_id, xspf_url=xspf_url,
2563 xspf_base_url=base_url(xspf_url))
2565 def _parse_xspf(self, xspf_doc, playlist_id, xspf_url=None, xspf_base_url=None):
2567 'xspf': 'http://xspf.org/ns/0/',
2568 's1': 'http://static.streamone.nl/player/ns/0',
2572 for track in xspf_doc.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
2574 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
2575 description = xpath_text(
2576 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
2577 thumbnail = xpath_text(
2578 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
2579 duration = float_or_none(
2580 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
2583 for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP)):
2584 format_url = urljoin(xspf_base_url, location.text)
2589 'manifest_url': xspf_url,
2590 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
2591 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
2592 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
2598 'description': description,
2599 'thumbnail': thumbnail,
2600 'duration': duration,
2605 def _extract_mpd_formats(self, *args, **kwargs):
2606 fmts, subs = self._extract_mpd_formats_and_subtitles(*args, **kwargs)
2608 self._report_ignoring_subs('DASH')
2611 def _extract_mpd_formats_and_subtitles(self, *args, **kwargs):
2612 periods = self._extract_mpd_periods(*args, **kwargs)
2613 return self._merge_mpd_periods(periods)
2615 def _extract_mpd_periods(
2616 self, mpd_url, video_id, mpd_id=None, note=None, errnote=None,
2617 fatal=True, data=None, headers={}, query={}):
2619 if self.get_param('ignore_no_formats_error'):
2622 res = self._download_xml_handle(
2624 note='Downloading MPD manifest' if note is None else note,
2625 errnote='Failed to download MPD manifest' if errnote is None else errnote,
2626 fatal=fatal, data=data, headers=headers, query=query)
2633 # We could have been redirected to a new url when we retrieved our mpd file.
2635 mpd_base_url = base_url(mpd_url)
2637 return self._parse_mpd_periods(mpd_doc, mpd_id, mpd_base_url, mpd_url)
2639 def _parse_mpd_formats(self, *args, **kwargs):
2640 fmts, subs = self._parse_mpd_formats_and_subtitles(*args, **kwargs)
2642 self._report_ignoring_subs('DASH')
2645 def _parse_mpd_formats_and_subtitles(self, *args, **kwargs):
2646 periods = self._parse_mpd_periods(*args, **kwargs)
2647 return self._merge_mpd_periods(periods)
2649 def _merge_mpd_periods(self, periods):
2651 Combine all formats and subtitles from an MPD manifest into a single list,
2652 by concatenate streams with similar formats.
2654 formats, subtitles = {}, {}
2655 for period in periods:
2656 for f in period['formats']:
2657 assert 'is_dash_periods' not in f, 'format already processed'
2658 f['is_dash_periods'] = True
2659 format_key = tuple(v for k, v in f.items() if k not in (
2660 ('format_id', 'fragments', 'manifest_stream_number')))
2661 if format_key not in formats:
2662 formats[format_key] = f
2663 elif 'fragments' in f:
2664 formats[format_key].setdefault('fragments', []).extend(f['fragments'])
2666 if subtitles and period['subtitles']:
2667 self.report_warning(bug_reports_message(
2668 'Found subtitles in multiple periods in the DASH manifest; '
2669 'if part of the subtitles are missing,',
2672 for sub_lang, sub_info in period['subtitles'].items():
2673 subtitles.setdefault(sub_lang, []).extend(sub_info)
2675 return list(formats.values()), subtitles
2677 def _parse_mpd_periods(self, mpd_doc, mpd_id=None, mpd_base_url='', mpd_url=None):
2679 Parse formats from MPD manifest.
2681 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
2682 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
2683 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
2685 if not self.get_param('dynamic_mpd', True):
2686 if mpd_doc.get('type') == 'dynamic':
2689 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
2692 return self._xpath_ns(path, namespace)
2694 def is_drm_protected(element):
2695 return element.find(_add_ns('ContentProtection')) is not None
2697 def extract_multisegment_info(element, ms_parent_info):
2698 ms_info = ms_parent_info.copy()
2700 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2701 # common attributes and elements. We will only extract relevant
2703 def extract_common(source):
2704 segment_timeline = source.find(_add_ns('SegmentTimeline'))
2705 if segment_timeline is not None:
2706 s_e = segment_timeline.findall(_add_ns('S'))
2708 ms_info['total_number'] = 0
2711 r = int(s.get('r', 0))
2712 ms_info['total_number'] += 1 + r
2713 ms_info['s'].append({
2714 't': int(s.get('t', 0)),
2715 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2716 'd': int(s.attrib['d']),
2719 start_number = source.get('startNumber')
2721 ms_info['start_number'] = int(start_number)
2722 timescale = source.get('timescale')
2724 ms_info['timescale'] = int(timescale)
2725 segment_duration = source.get('duration')
2726 if segment_duration:
2727 ms_info['segment_duration'] = float(segment_duration)
2729 def extract_Initialization(source):
2730 initialization = source.find(_add_ns('Initialization'))
2731 if initialization is not None:
2732 ms_info['initialization_url'] = initialization.attrib['sourceURL']
2734 segment_list = element.find(_add_ns('SegmentList'))
2735 if segment_list is not None:
2736 extract_common(segment_list)
2737 extract_Initialization(segment_list)
2738 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
2740 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
2742 segment_template = element.find(_add_ns('SegmentTemplate'))
2743 if segment_template is not None:
2744 extract_common(segment_template)
2745 media = segment_template.get('media')
2747 ms_info['media'] = media
2748 initialization = segment_template.get('initialization')
2750 ms_info['initialization'] = initialization
2752 extract_Initialization(segment_template)
2755 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
2756 stream_numbers = collections.defaultdict(int)
2757 for period_idx, period in enumerate(mpd_doc.findall(_add_ns('Period'))):
2759 'id': period.get('id', f'period-{period_idx}'),
2761 'subtitles': collections.defaultdict(list),
2763 period_duration = parse_duration(period.get('duration')) or mpd_duration
2764 period_ms_info = extract_multisegment_info(period, {
2768 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
2769 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
2770 for representation in adaptation_set.findall(_add_ns('Representation')):
2771 representation_attrib = adaptation_set.attrib.copy()
2772 representation_attrib.update(representation.attrib)
2773 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2774 mime_type = representation_attrib['mimeType']
2775 content_type = representation_attrib.get('contentType', mime_type.split('/')[0])
2777 codec_str = representation_attrib.get('codecs', '')
2778 # Some kind of binary subtitle found in some youtube livestreams
2779 if mime_type == 'application/x-rawcc':
2780 codecs = {'scodec': codec_str}
2782 codecs = parse_codecs(codec_str)
2783 if content_type not in ('video', 'audio', 'text'):
2784 if mime_type == 'image/jpeg':
2785 content_type = mime_type
2786 elif codecs.get('vcodec', 'none') != 'none':
2787 content_type = 'video'
2788 elif codecs.get('acodec', 'none') != 'none':
2789 content_type = 'audio'
2790 elif codecs.get('scodec', 'none') != 'none':
2791 content_type = 'text'
2792 elif mimetype2ext(mime_type) in ('tt', 'dfxp', 'ttml', 'xml', 'json'):
2793 content_type = 'text'
2795 self.report_warning(f'Unknown MIME type {mime_type} in DASH manifest')
2799 for element in (representation, adaptation_set, period, mpd_doc):
2800 base_url_e = element.find(_add_ns('BaseURL'))
2801 if try_call(lambda: base_url_e.text) is not None:
2802 base_url = base_url_e.text + base_url
2803 if re.match(r'^https?://', base_url):
2805 if mpd_base_url and base_url.startswith('/'):
2806 base_url = urllib.parse.urljoin(mpd_base_url, base_url)
2807 elif mpd_base_url and not re.match(r'^https?://', base_url):
2808 if not mpd_base_url.endswith('/'):
2810 base_url = mpd_base_url + base_url
2811 representation_id = representation_attrib.get('id')
2812 lang = representation_attrib.get('lang')
2813 url_el = representation.find(_add_ns('BaseURL'))
2814 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
2815 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
2816 if representation_id is not None:
2817 format_id = representation_id
2819 format_id = content_type
2821 format_id = mpd_id + '-' + format_id
2822 if content_type in ('video', 'audio'):
2824 'format_id': format_id,
2825 'manifest_url': mpd_url,
2826 'ext': mimetype2ext(mime_type),
2827 'width': int_or_none(representation_attrib.get('width')),
2828 'height': int_or_none(representation_attrib.get('height')),
2829 'tbr': float_or_none(bandwidth, 1000),
2830 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
2831 'fps': int_or_none(representation_attrib.get('frameRate')),
2832 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
2833 'format_note': f'DASH {content_type}',
2834 'filesize': filesize,
2835 'container': mimetype2ext(mime_type) + '_dash',
2838 elif content_type == 'text':
2840 'ext': mimetype2ext(mime_type),
2841 'manifest_url': mpd_url,
2842 'filesize': filesize,
2844 elif content_type == 'image/jpeg':
2845 # See test case in VikiIE
2846 # https://www.viki.com/videos/1175236v-choosing-spouse-by-lottery-episode-1
2848 'format_id': format_id,
2850 'manifest_url': mpd_url,
2851 'format_note': 'DASH storyboards (jpeg)',
2855 if is_drm_protected(adaptation_set) or is_drm_protected(representation):
2857 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
2859 def prepare_template(template_name, identifiers):
2860 tmpl = representation_ms_info[template_name]
2861 if representation_id is not None:
2862 tmpl = tmpl.replace('$RepresentationID$', representation_id)
2863 # First of, % characters outside $...$ templates
2864 # must be escaped by doubling for proper processing
2865 # by % operator string formatting used further (see
2866 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2872 in_template = not in_template
2873 elif c == '%' and not in_template:
2875 # Next, $...$ templates are translated to their
2876 # %(...) counterparts to be used with % operator
2877 t = re.sub(r'\$({})\$'.format('|'.join(identifiers)), r'%(\1)d', t)
2878 t = re.sub(r'\$({})%([^$]+)\$'.format('|'.join(identifiers)), r'%(\1)\2', t)
2879 t.replace('$$', '$')
2882 # @initialization is a regular template like @media one
2883 # so it should be handled just the same way (see
2884 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2885 if 'initialization' in representation_ms_info:
2886 initialization_template = prepare_template(
2888 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2889 # $Time$ shall not be included for @initialization thus
2890 # only $Bandwidth$ remains
2892 representation_ms_info['initialization_url'] = initialization_template % {
2893 'Bandwidth': bandwidth,
2896 def location_key(location):
2897 return 'url' if re.match(r'^https?://', location) else 'path'
2899 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
2901 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2902 media_location_key = location_key(media_template)
2904 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2905 # can't be used at the same time
2906 if '%(Number' in media_template and 's' not in representation_ms_info:
2907 segment_duration = None
2908 if 'total_number' not in representation_ms_info and 'segment_duration' in representation_ms_info:
2909 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
2910 representation_ms_info['total_number'] = int(math.ceil(
2911 float_or_none(period_duration, segment_duration, default=0)))
2912 representation_ms_info['fragments'] = [{
2913 media_location_key: media_template % {
2914 'Number': segment_number,
2915 'Bandwidth': bandwidth,
2917 'duration': segment_duration,
2918 } for segment_number in range(
2919 representation_ms_info['start_number'],
2920 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
2922 # $Number*$ or $Time$ in media template with S list available
2923 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2924 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2925 representation_ms_info['fragments'] = []
2928 segment_number = representation_ms_info['start_number']
2930 def add_segment_url():
2931 segment_url = media_template % {
2932 'Time': segment_time,
2933 'Bandwidth': bandwidth,
2934 'Number': segment_number,
2936 representation_ms_info['fragments'].append({
2937 media_location_key: segment_url,
2938 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
2941 for s in representation_ms_info['s']:
2942 segment_time = s.get('t') or segment_time
2946 for _ in range(s.get('r', 0)):
2947 segment_time += segment_d
2950 segment_time += segment_d
2951 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
2952 # No media template,
2953 # e.g. https://www.youtube.com/watch?v=iXZV5uAYMJI
2954 # or any YouTube dashsegments video
2957 timescale = representation_ms_info['timescale']
2958 for s in representation_ms_info['s']:
2959 duration = float_or_none(s['d'], timescale)
2960 for _ in range(s.get('r', 0) + 1):
2961 segment_uri = representation_ms_info['segment_urls'][segment_index]
2963 location_key(segment_uri): segment_uri,
2964 'duration': duration,
2967 representation_ms_info['fragments'] = fragments
2968 elif 'segment_urls' in representation_ms_info:
2969 # Segment URLs with no SegmentTimeline
2970 # E.g. https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2971 # https://github.com/ytdl-org/youtube-dl/pull/14844
2973 segment_duration = float_or_none(
2974 representation_ms_info['segment_duration'],
2975 representation_ms_info['timescale']) if 'segment_duration' in representation_ms_info else None
2976 for segment_url in representation_ms_info['segment_urls']:
2978 location_key(segment_url): segment_url,
2980 if segment_duration:
2981 fragment['duration'] = segment_duration
2982 fragments.append(fragment)
2983 representation_ms_info['fragments'] = fragments
2984 # If there is a fragments key available then we correctly recognized fragmented media.
2985 # Otherwise we will assume unfragmented media with direct access. Technically, such
2986 # assumption is not necessarily correct since we may simply have no support for
2987 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2988 if 'fragments' in representation_ms_info:
2990 # NB: mpd_url may be empty when MPD manifest is parsed from a string
2991 'url': mpd_url or base_url,
2992 'fragment_base_url': base_url,
2994 'protocol': 'http_dash_segments' if mime_type != 'image/jpeg' else 'mhtml',
2996 if 'initialization_url' in representation_ms_info:
2997 initialization_url = representation_ms_info['initialization_url']
2998 if not f.get('url'):
2999 f['url'] = initialization_url
3000 f['fragments'].append({location_key(initialization_url): initialization_url})
3001 f['fragments'].extend(representation_ms_info['fragments'])
3002 if not period_duration:
3003 period_duration = try_get(
3004 representation_ms_info,
3005 lambda r: sum(frag['duration'] for frag in r['fragments']), float)
3007 # Assuming direct URL to unfragmented media.
3009 if content_type in ('video', 'audio', 'image/jpeg'):
3010 f['manifest_stream_number'] = stream_numbers[f['url']]
3011 stream_numbers[f['url']] += 1
3012 period_entry['formats'].append(f)
3013 elif content_type == 'text':
3014 period_entry['subtitles'][lang or 'und'].append(f)
3017 def _extract_ism_formats(self, *args, **kwargs):
3018 fmts, subs = self._extract_ism_formats_and_subtitles(*args, **kwargs)
3020 self._report_ignoring_subs('ISM')
3023 def _extract_ism_formats_and_subtitles(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
3024 if self.get_param('ignore_no_formats_error'):
3027 res = self._download_xml_handle(
3029 note='Downloading ISM manifest' if note is None else note,
3030 errnote='Failed to download ISM manifest' if errnote is None else errnote,
3031 fatal=fatal, data=data, headers=headers, query=query)
3038 return self._parse_ism_formats_and_subtitles(ism_doc, urlh.url, ism_id)
3040 def _parse_ism_formats_and_subtitles(self, ism_doc, ism_url, ism_id=None):
3042 Parse formats from ISM manifest.
3044 1. [MS-SSTR]: Smooth Streaming Protocol,
3045 https://msdn.microsoft.com/en-us/library/ff469518.aspx
3047 if ism_doc.get('IsLive') == 'TRUE':
3050 duration = int(ism_doc.attrib['Duration'])
3051 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
3055 for stream in ism_doc.findall('StreamIndex'):
3056 stream_type = stream.get('Type')
3057 if stream_type not in ('video', 'audio', 'text'):
3059 url_pattern = stream.attrib['Url']
3060 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
3061 stream_name = stream.get('Name')
3062 stream_language = stream.get('Language', 'und')
3063 for track in stream.findall('QualityLevel'):
3064 KNOWN_TAGS = {'255': 'AACL', '65534': 'EC-3'}
3065 fourcc = track.get('FourCC') or KNOWN_TAGS.get(track.get('AudioTag'))
3066 # TODO: add support for WVC1 and WMAP
3067 if fourcc not in ('H264', 'AVC1', 'AACL', 'TTML', 'EC-3'):
3068 self.report_warning(f'{fourcc} is not a supported codec')
3070 tbr = int(track.attrib['Bitrate']) // 1000
3071 # [1] does not mention Width and Height attributes. However,
3072 # they're often present while MaxWidth and MaxHeight are
3073 # missing, so should be used as fallbacks
3074 width = int_or_none(track.get('MaxWidth') or track.get('Width'))
3075 height = int_or_none(track.get('MaxHeight') or track.get('Height'))
3076 sampling_rate = int_or_none(track.get('SamplingRate'))
3078 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
3079 track_url_pattern = urllib.parse.urljoin(ism_url, track_url_pattern)
3085 stream_fragments = stream.findall('c')
3086 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
3087 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
3088 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
3089 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
3090 if not fragment_ctx['duration']:
3092 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
3094 next_fragment_time = duration
3095 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
3096 for _ in range(fragment_repeat):
3098 'url': re.sub(r'{start[ _]time}', str(fragment_ctx['time']), track_url_pattern),
3099 'duration': fragment_ctx['duration'] / stream_timescale,
3101 fragment_ctx['time'] += fragment_ctx['duration']
3103 if stream_type == 'text':
3104 subtitles.setdefault(stream_language, []).append({
3108 'manifest_url': ism_url,
3109 'fragments': fragments,
3110 '_download_params': {
3111 'stream_type': stream_type,
3112 'duration': duration,
3113 'timescale': stream_timescale,
3115 'language': stream_language,
3116 'codec_private_data': track.get('CodecPrivateData'),
3119 elif stream_type in ('video', 'audio'):
3121 'format_id': join_nonempty(ism_id, stream_name, tbr),
3123 'manifest_url': ism_url,
3124 'ext': 'ismv' if stream_type == 'video' else 'isma',
3128 'asr': sampling_rate,
3129 'vcodec': 'none' if stream_type == 'audio' else fourcc,
3130 'acodec': 'none' if stream_type == 'video' else fourcc,
3132 'fragments': fragments,
3133 'has_drm': ism_doc.find('Protection') is not None,
3134 'language': stream_language,
3135 'audio_channels': int_or_none(track.get('Channels')),
3136 '_download_params': {
3137 'stream_type': stream_type,
3138 'duration': duration,
3139 'timescale': stream_timescale,
3140 'width': width or 0,
3141 'height': height or 0,
3143 'language': stream_language,
3144 'codec_private_data': track.get('CodecPrivateData'),
3145 'sampling_rate': sampling_rate,
3146 'channels': int_or_none(track.get('Channels', 2)),
3147 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
3148 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
3151 return formats, subtitles
3153 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):
3154 def absolute_url(item_url):
3155 return urljoin(base_url, item_url)
3157 def parse_content_type(content_type):
3158 if not content_type:
3160 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?
(?P
<codecs
>[^
"]+))?', content_type)
3162 mimetype, codecs = ctr.groups()
3163 f = parse_codecs(codecs)
3164 f['ext'] = mimetype2ext(mimetype)
3168 def _media_formats(src, cur_media_type, type_info=None):
3169 type_info = type_info or {}
3170 full_url = absolute_url(src)
3171 ext = type_info.get('ext') or determine_ext(full_url)
3173 is_plain_url = False
3174 formats = self._extract_m3u8_formats(
3175 full_url, video_id, ext='mp4',
3176 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
3177 preference=preference, quality=quality, fatal=False)
3179 is_plain_url = False
3180 formats = self._extract_mpd_formats(
3181 full_url, video_id, mpd_id=mpd_id, fatal=False)
3186 'vcodec': 'none' if cur_media_type == 'audio' else None,
3189 return is_plain_url, formats
3192 # amp-video and amp-audio are very similar to their HTML5 counterparts
3193 # so we will include them right here (see
3194 # https://www.ampproject.org/docs/reference/components/amp-video)
3195 # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/
3196 _MEDIA_TAG_NAME_RE = r'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)'
3197 media_tags = [(media_tag, media_tag_name, media_type, '')
3198 for media_tag, media_tag_name, media_type
3199 in re.findall(rf'(?s)(<({_MEDIA_TAG_NAME_RE})[^>]*/>)', webpage)]
3200 media_tags.extend(re.findall(
3201 # We only allow video|audio followed by a whitespace or '>'.
3202 # Allowing more characters may end up in significant slow down (see
3203 # https://github.com/ytdl-org/youtube-dl/issues/11979,
3204 # e.g. http://www.porntrex.com/maps/videositemap.xml).
3205 rf'(?s)(<(?P<tag>{_MEDIA_TAG_NAME_RE})(?:\s+[^>]*)?>)(.*?)</(?P=tag)>', webpage))
3206 for media_tag, _, media_type, media_content in media_tags:
3211 media_attributes = extract_attributes(media_tag)
3212 src = strip_or_none(dict_get(media_attributes, ('src', 'data-video-src', 'data-src', 'data-source')))
3214 f = parse_content_type(media_attributes.get('type'))
3215 _, formats = _media_formats(src, media_type, f)
3216 media_info['formats'].extend(formats)
3217 media_info['thumbnail'] = absolute_url(media_attributes.get('poster'))
3219 for source_tag in re.findall(r'<source[^>]+>', media_content):
3220 s_attr = extract_attributes(source_tag)
3221 # data-video-src and data-src are non standard but seen
3222 # several times in the wild
3223 src = strip_or_none(dict_get(s_attr, ('src', 'data-video-src', 'data-src', 'data-source')))
3226 f = parse_content_type(s_attr.get('type'))
3227 is_plain_url, formats = _media_formats(src, media_type, f)
3229 # width, height, res, label and title attributes are
3230 # all not standard but seen several times in the wild
3233 for lbl in ('label', 'title')
3234 if str_or_none(s_attr.get(lbl))
3236 width = int_or_none(s_attr.get('width'))
3237 height = (int_or_none(s_attr.get('height'))
3238 or int_or_none(s_attr.get('res')))
3239 if not width or not height:
3241 resolution = parse_resolution(lbl)
3244 width = width or resolution.get('width')
3245 height = height or resolution.get('height')
3247 tbr = parse_bitrate(lbl)
3256 'format_id': s_attr.get('label') or s_attr.get('title'),
3258 f.update(formats[0])
3259 media_info['formats'].append(f)
3261 media_info['formats'].extend(formats)
3262 for track_tag in re.findall(r'<track[^>]+>', media_content):
3263 track_attributes = extract_attributes(track_tag)
3264 kind = track_attributes.get('kind')
3265 if not kind or kind in ('subtitles', 'captions'):
3266 src = strip_or_none(track_attributes.get('src'))
3269 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
3270 media_info['subtitles'].setdefault(lang, []).append({
3271 'url': absolute_url(src),
3273 for f in media_info['formats']:
3274 f.setdefault('http_headers', {})['Referer'] = base_url
3275 if media_info['formats'] or media_info['subtitles']:
3276 entries.append(media_info)
3279 def _extract_akamai_formats(self, *args, **kwargs):
3280 fmts, subs = self._extract_akamai_formats_and_subtitles(*args, **kwargs)
3282 self._report_ignoring_subs('akamai')
3285 def _extract_akamai_formats_and_subtitles(self, manifest_url, video_id, hosts={}):
3286 signed = 'hdnea=' in manifest_url
3288 # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
3289 manifest_url = re.sub(
3290 r'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
3291 '', manifest_url).strip('?')
3296 hdcore_sign = 'hdcore=3.7.0'
3297 f4m_url = re.sub(r'(https?://[^/]+)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
3298 hds_host = hosts.get('hds')
3300 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
3301 if 'hdcore=' not in f4m_url:
3302 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
3303 f4m_formats = self._extract_f4m_formats(
3304 f4m_url, video_id, f4m_id='hds', fatal=False)
3305 for entry in f4m_formats:
3306 entry.update({'extra_param_to_segment_url': hdcore_sign})
3307 formats.extend(f4m_formats)
3309 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
3310 hls_host = hosts.get('hls')
3312 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
3313 m3u8_formats, m3u8_subtitles = self._extract_m3u8_formats_and_subtitles(
3314 m3u8_url, video_id, 'mp4', 'm3u8_native',
3315 m3u8_id='hls', fatal=False)
3316 formats.extend(m3u8_formats)
3317 subtitles = self._merge_subtitles(subtitles, m3u8_subtitles)
3319 http_host = hosts.get('http')
3320 if http_host and m3u8_formats and not signed:
3321 REPL_REGEX = r'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
3322 qualities = re.match(REPL_REGEX, m3u8_url).group(2).split(',')
3323 qualities_length = len(qualities)
3324 if len(m3u8_formats) in (qualities_length, qualities_length + 1):
3326 for f in m3u8_formats:
3327 if f['vcodec'] != 'none':
3328 for protocol in ('http', 'https'):
3330 del http_f['manifest_url']
3332 REPL_REGEX, protocol + fr'://{http_host}/\g<1>{qualities[i]}\3', f['url'])
3334 'format_id': http_f['format_id'].replace('hls-', protocol + '-'),
3336 'protocol': protocol,
3338 formats.append(http_f)
3341 return formats, subtitles
3343 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
3344 query = urllib.parse.urlparse(url).query
3345 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
3347 r'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url)
3348 url_base = mobj.group('url')
3349 http_base_url = '{}{}:{}'.format('http', mobj.group('s') or '', url_base)
3352 def manifest_url(manifest):
3353 m_url = f'{http_base_url}/{manifest}'
3355 m_url += f'?{query}'
3358 if 'm3u8' not in skip_protocols:
3359 formats.extend(self._extract_m3u8_formats(
3360 manifest_url('playlist.m3u8'), video_id, 'mp4',
3361 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
3362 if 'f4m' not in skip_protocols:
3363 formats.extend(self._extract_f4m_formats(
3364 manifest_url('manifest.f4m'),
3365 video_id, f4m_id='hds', fatal=False))
3366 if 'dash' not in skip_protocols:
3367 formats.extend(self._extract_mpd_formats(
3368 manifest_url('manifest.mpd'),
3369 video_id, mpd_id='dash', fatal=False))
3370 if re.search(r'(?:/smil:|\.smil)', url_base):
3371 if 'smil' not in skip_protocols:
3372 rtmp_formats = self._extract_smil_formats(
3373 manifest_url('jwplayer.smil'),
3374 video_id, fatal=False)
3375 for rtmp_format in rtmp_formats:
3376 rtsp_format = rtmp_format.copy()
3377 rtsp_format['url'] = '{}/{}'.format(rtmp_format['url'], rtmp_format['play_path'])
3378 del rtsp_format['play_path']
3379 del rtsp_format['ext']
3380 rtsp_format.update({
3381 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
3382 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
3385 formats.extend([rtmp_format, rtsp_format])
3387 for protocol in ('rtmp', 'rtsp'):
3388 if protocol not in skip_protocols:
3390 'url': f'{protocol}:{url_base}',
3391 'format_id': protocol,
3392 'protocol': protocol,
3396 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
3397 return self._search_json(
3398 r'''(?<!-)\bjwplayer\s*\(\s*(?P<q>'|")(?
!(?P
=q
)).+(?P
=q
)\s
*\
)(?
:(?
!</script
>).)*?\
.\s
*(?
:setup\s
*\
(|
(?P
<load
>load
)\s
*\
(\s
*\
[)''',
3399 webpage, 'JWPlayer data', video_id,
3400 # must be a {...} or sequence, ending
3401 contains_pattern=r'\{(?s:.*)}(?(load)(?:\s*,\s*\{(?s:.*)})*)', end_pattern=r'(?(load)\]|\))',
3402 transform_source=transform_source, default=None)
3404 def _extract_jwplayer_data(self, webpage, video_id, *args, transform_source=js_to_json, **kwargs):
3405 jwplayer_data = self._find_jwplayer_data(
3406 webpage, video_id, transform_source=transform_source)
3407 return self._parse_jwplayer_data(
3408 jwplayer_data, video_id, *args, **kwargs)
3410 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
3411 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3413 if not isinstance(jwplayer_data, dict):
3416 playlist_items = jwplayer_data.get('playlist')
3417 # JWPlayer backward compatibility: single playlist item/flattened playlists
3418 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
3419 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
3420 if not isinstance(playlist_items, list):
3421 playlist_items = (playlist_items or jwplayer_data, )
3423 for video_data in playlist_items:
3424 if not isinstance(video_data, dict):
3426 # JWPlayer backward compatibility: flattened sources
3427 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
3428 if 'sources' not in video_data:
3429 video_data['sources'] = [video_data]
3431 this_video_id = video_id or video_data['mediaid']
3433 formats = self._parse_jwplayer_formats(
3434 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
3435 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
3438 for track in traverse_obj(video_data, (
3439 'tracks', lambda _, v: v['kind'].lower() in ('captions', 'subtitles'))):
3440 track_url = urljoin(base_url, track.get('file'))
3443 subtitles.setdefault(track.get('label') or 'en', []).append({
3444 'url': self._proto_relative_url(track_url),
3448 'id': this_video_id,
3449 'title': unescapeHTML(video_data['title'] if require_title else video_data.get('title')),
3450 'description': clean_html(video_data.get('description')),
3451 'thumbnail': urljoin(base_url, self._proto_relative_url(video_data.get('image'))),
3452 'timestamp': int_or_none(video_data.get('pubdate')),
3453 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
3454 'subtitles': subtitles,
3455 'alt_title': clean_html(video_data.get('subtitle')), # attributes used e.g. by Tele5 ...
3456 'genre': clean_html(video_data.get('genre')),
3457 'channel': clean_html(dict_get(video_data, ('category', 'channel'))),
3458 'season_number': int_or_none(video_data.get('season')),
3459 'episode_number': int_or_none(video_data.get('episode')),
3460 'release_year': int_or_none(video_data.get('releasedate')),
3461 'age_limit': int_or_none(video_data.get('age_restriction')),
3463 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
3464 if len(formats) == 1 and re.search(r'^(?:http|//).*(?:youtube\.com|youtu\.be)/.+', formats[0]['url']):
3466 '_type': 'url_transparent',
3467 'url': formats[0]['url'],
3470 entry['formats'] = formats
3471 entries.append(entry)
3472 if len(entries) == 1:
3475 return self.playlist_result(entries)
3477 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
3478 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3481 for source in jwplayer_sources_data:
3482 if not isinstance(source, dict):
3484 source_url = urljoin(
3485 base_url, self._proto_relative_url(source.get('file')))
3486 if not source_url or source_url in urls:
3488 urls.add(source_url)
3489 source_type = source.get('type') or ''
3490 ext = mimetype2ext(source_type) or determine_ext(source_url)
3491 if source_type == 'hls' or ext == 'm3u8' or 'format=m3u8-aapl' in source_url:
3492 formats.extend(self._extract_m3u8_formats(
3493 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
3494 m3u8_id=m3u8_id, fatal=False))
3495 elif source_type == 'dash' or ext == 'mpd' or 'format=mpd-time-csf' in source_url:
3496 formats.extend(self._extract_mpd_formats(
3497 source_url, video_id, mpd_id=mpd_id, fatal=False))
3499 formats.extend(self._extract_smil_formats(
3500 source_url, video_id, fatal=False))
3501 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3502 elif source_type.startswith('audio') or ext in (
3503 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
3510 format_id = str_or_none(source.get('label'))
3511 height = int_or_none(source.get('height'))
3512 if height is None and format_id:
3513 # Often no height is provided but there is a label in
3514 # format like "1080p", "720p SD", or 1080.
3515 height = parse_resolution(format_id).get('height')
3518 'width': int_or_none(source.get('width')),
3520 'tbr': int_or_none(source.get('bitrate'), scale=1000),
3521 'filesize': int_or_none(source.get('filesize')),
3523 'format_id': format_id,
3525 if source_url.startswith('rtmp'):
3526 a_format['ext'] = 'flv'
3527 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3528 # of jwplayer.flash.swf
3529 rtmp_url_parts = re.split(
3530 r'((?:mp4|mp3|flv):)', source_url, maxsplit=1)
3531 if len(rtmp_url_parts) == 3:
3532 rtmp_url, prefix, play_path = rtmp_url_parts
3535 'play_path': prefix + play_path,
3538 a_format.update(rtmp_params)
3539 formats.append(a_format)
3542 def _live_title(self, name):
3543 self._downloader.deprecation_warning('yt_dlp.InfoExtractor._live_title is deprecated and does not work as expected')
3546 def _int(self, v, name, fatal=False, **kwargs):
3547 res = int_or_none(v, **kwargs)
3549 msg = f'Failed to extract {name}: Could not parse value {v!r}'
3551 raise ExtractorError(msg)
3553 self.report_warning(msg)
3556 def _float(self, v, name, fatal=False, **kwargs):
3557 res = float_or_none(v, **kwargs)
3559 msg = f'Failed to extract {name}: Could not parse value {v!r}'
3561 raise ExtractorError(msg)
3563 self.report_warning(msg)
3566 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
3567 path='/', secure=False, discard=False, rest={}, **kwargs):
3568 cookie = http.cookiejar.Cookie(
3569 0, name, value, port, port is not None, domain, True,
3570 domain.startswith('.'), path, True, secure, expire_time,
3571 discard, None, None, rest)
3572 self.cookiejar.set_cookie(cookie)
3574 def _get_cookies(self, url):
3575 """ Return a http.cookies.SimpleCookie with the cookies for the url """
3576 return LenientSimpleCookie(self._downloader.cookiejar.get_cookie_header(url))
3578 def _apply_first_set_cookie_header(self, url_handle, cookie):
3580 Apply first Set-Cookie header instead of the last. Experimental.
3582 Some sites (e.g. [1-3]) may serve two cookies under the same name
3583 in Set-Cookie header and expect the first (old) one to be set rather
3584 than second (new). However, as of RFC6265 the newer one cookie
3585 should be set into cookie store what actually happens.
3586 We will workaround this issue by resetting the cookie to
3587 the first one manually.
3588 1. https://new.vk.com/
3589 2. https://github.com/ytdl-org/youtube-dl/issues/9841#issuecomment-227871201
3590 3. https://learning.oreilly.com/
3592 for header, cookies in url_handle.headers.items():
3593 if header.lower() != 'set-cookie':
3595 cookies = cookies.encode('iso-8859-1').decode('utf-8')
3596 cookie_value = re.search(
3597 rf'{cookie}=(.+?);.*?\b[Dd]omain=(.+?)(?:[,;]|$)', cookies)
3599 value, domain = cookie_value.groups()
3600 self._set_cookie(domain, cookie, value)
3604 def get_testcases(cls, include_onlymatching=False):
3605 # Do not look in super classes
3606 t = vars(cls).get('_TEST')
3608 assert not hasattr(cls, '_TESTS'), f'{cls.ie_key()}IE has _TEST and _TESTS'
3611 tests = vars(cls).get('_TESTS', [])
3613 if not include_onlymatching and t.get('only_matching', False):
3615 t['name'] = cls.ie_key()
3617 if getattr(cls, '__wrapped__', None):
3618 yield from cls.__wrapped__.get_testcases(include_onlymatching)
3621 def get_webpage_testcases(cls):
3622 tests = vars(cls).get('_WEBPAGE_TESTS', [])
3624 t['name'] = cls.ie_key()
3626 if getattr(cls, '__wrapped__', None):
3627 yield from cls.__wrapped__.get_webpage_testcases()
3629 @classproperty(cache=True)
3631 """Get age limit from the testcases"""
3632 return max(traverse_obj(
3633 (*cls.get_testcases(include_onlymatching=False), *cls.get_webpage_testcases()),
3634 (..., (('playlist', 0), None), 'info_dict', 'age_limit')) or [0])
3636 @classproperty(cache=True)
3637 def _RETURN_TYPE(cls):
3638 """What the extractor returns: "video", "playlist", "any", or None (Unknown)"""
3639 tests = tuple(cls.get_testcases(include_onlymatching=False))
3642 elif not any(k.startswith('playlist') for test in tests for k in test):
3644 elif all(any(k.startswith('playlist') for k in test) for test in tests):
3649 def is_single_video(cls, url):
3650 """Returns whether the URL is of a single video, None if unknown"""
3651 if cls.suitable(url):
3652 return {'video': True, 'playlist': False}.get(cls._RETURN_TYPE)
3655 def is_suitable(cls, age_limit):
3656 """Test whether the extractor is generally suitable for the given age limit"""
3657 return not age_restricted(cls.age_limit, age_limit)
3660 def description(cls, *, markdown=True, search_examples=None):
3661 """Description of the extractor"""
3663 if cls._NETRC_MACHINE:
3665 desc += f' [*{cls._NETRC_MACHINE}*](## "netrc machine")'
3667 desc += f' [{cls._NETRC_MACHINE}]'
3668 if cls.IE_DESC is False:
3671 desc += f' {cls.IE_DESC}'
3673 desc += f'{";" if cls.IE_DESC else ""} "{cls.SEARCH_KEY}:" prefix'
3675 _COUNTS = ('', '5', '10', 'all')
3676 desc += f' (e.g. "{cls.SEARCH_KEY}{random.choice(_COUNTS)}:{random.choice(search_examples)}")'
3677 if not cls.working():
3678 desc += ' (**Currently broken**)' if markdown else ' (Currently broken)'
3680 # Escape emojis. Ref: https://github.com/github/markup/issues/1153
3681 name = (' - **{}**'.format(re.sub(r':(\w+:)', ':\u200B\\g<1>', cls.IE_NAME))) if markdown else cls.IE_NAME
3682 return f'{name}:{desc}' if desc else name
3684 def extract_subtitles(self, *args, **kwargs):
3685 if (self.get_param('writesubtitles', False)
3686 or self.get_param('listsubtitles')):
3687 return self._get_subtitles(*args, **kwargs)
3690 def _get_subtitles(self, *args, **kwargs):
3691 raise NotImplementedError('This method must be implemented by subclasses')
3693 class CommentsDisabled(Exception):
3694 """Raise in _get_comments if comments are disabled for the video"""
3696 def extract_comments(self, *args, **kwargs):
3697 if not self.get_param('getcomments'):
3699 generator = self._get_comments(*args, **kwargs)
3706 comments.append(next(generator))
3707 except StopIteration:
3709 except KeyboardInterrupt:
3710 self.to_screen('Interrupted by user')
3711 except self.CommentsDisabled:
3712 return {'comments': None, 'comment_count': None}
3713 except Exception as e:
3714 if self.get_param('ignoreerrors') is not True:
3716 self._downloader.report_error(e)
3717 comment_count = len(comments)
3718 self.to_screen(f'Extracted {comment_count} comments')
3720 'comments': comments,
3721 'comment_count': None if interrupted else comment_count,
3725 def _get_comments(self, *args, **kwargs):
3726 raise NotImplementedError('This method must be implemented by subclasses')
3729 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
3730 """ Merge subtitle items for one language. Items with duplicated URLs/data
3731 will be dropped. """
3732 list1_data = {(item.get('url'), item.get('data')) for item in subtitle_list1}
3733 ret = list(subtitle_list1)
3734 ret.extend(item for item in subtitle_list2 if (item.get('url'), item.get('data')) not in list1_data)
3738 def _merge_subtitles(cls, *dicts, target=None):
3739 """ Merge subtitle dictionaries, language by language. """
3743 for lang, subs in d.items():
3744 target[lang] = cls._merge_subtitle_items(target.get(lang, []), subs)
3747 def extract_automatic_captions(self, *args, **kwargs):
3748 if (self.get_param('writeautomaticsub', False)
3749 or self.get_param('listsubtitles')):
3750 return self._get_automatic_captions(*args, **kwargs)
3753 def _get_automatic_captions(self, *args, **kwargs):
3754 raise NotImplementedError('This method must be implemented by subclasses')
3756 @functools.cached_property
3757 def _cookies_passed(self):
3758 """Whether cookies have been passed to YoutubeDL"""
3759 return self.get_param('cookiefile') is not None or self.get_param('cookiesfrombrowser') is not None
3761 def mark_watched(self, *args, **kwargs):
3762 if not self.get_param('mark_watched', False):
3764 if self.supports_login() and self._get_login_info()[0] is not None or self._cookies_passed:
3765 self._mark_watched(*args, **kwargs)
3767 def _mark_watched(self, *args, **kwargs):
3768 raise NotImplementedError('This method must be implemented by subclasses')
3770 def geo_verification_headers(self):
3772 geo_verification_proxy = self.get_param('geo_verification_proxy')
3773 if geo_verification_proxy:
3774 headers['Ytdl-request-proxy'] = geo_verification_proxy
3778 def _generic_id(url):
3779 return urllib.parse.unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
3781 def _generic_title(self, url='', webpage='', *, default=None):
3782 return (self._og_search_title(webpage, default=None)
3783 or self._html_extract_title(webpage, default=None)
3784 or urllib.parse.unquote(os.path.splitext(url_basename(url))[0])
3787 def _extract_chapters_helper(self, chapter_list, start_function, title_function, duration, strict=True):
3791 'start_time': start_function(chapter),
3792 'title': title_function(chapter),
3793 } for chapter in chapter_list or []]
3795 warn = self.report_warning
3797 warn = self.write_debug
3798 chapter_list.sort(key=lambda c: c['start_time'] or 0)
3800 chapters = [{'start_time': 0}]
3801 for idx, chapter in enumerate(chapter_list):
3802 if chapter['start_time'] is None:
3803 warn(f'Incomplete chapter {idx}')
3804 elif chapters[-1]['start_time'] <= chapter['start_time'] <= duration:
3805 chapters.append(chapter)
3806 elif chapter not in chapters:
3807 issue = (f'{chapter["start_time"]} > {duration}' if chapter['start_time'] > duration
3808 else f'{chapter["start_time"]} < {chapters[-1]["start_time"]}')
3809 warn(f'Invalid start time ({issue}) for chapter "{chapter["title"]}"')
3812 def _extract_chapters_from_description(self, description, duration):
3813 duration_re = r'(?:\d+:)?\d{1,2}:\d{2}'
3814 sep_re = r'(?m)^\s*(%s)\b\W*\s(%s)\s*$'
3815 return self._extract_chapters_helper(
3816 re.findall(sep_re % (duration_re, r'.+?'), description or ''),
3817 start_function=lambda x: parse_duration(x[0]), title_function=lambda x: x[1],
3818 duration=duration, strict=False) or self._extract_chapters_helper(
3819 re.findall(sep_re % (r'.+?', duration_re), description or ''),
3820 start_function=lambda x: parse_duration(x[1]), title_function=lambda x: x[0],
3821 duration=duration, strict=False)
3824 def _availability(is_private=None, needs_premium=None, needs_subscription=None, needs_auth=None, is_unlisted=None):
3826 x is not None for x in
3827 (is_private, needs_premium, needs_subscription, needs_auth, is_unlisted))
3829 'private' if is_private
3830 else 'premium_only' if needs_premium
3831 else 'subscriber_only' if needs_subscription
3832 else 'needs_auth' if needs_auth
3833 else 'unlisted' if is_unlisted
3834 else 'public' if all_known
3837 def _configuration_arg(self, key, default=NO_DEFAULT, *, ie_key=None, casesense=False):
3839 @returns A
list of values
for the extractor argument given by
"key"
3840 or "default" if no such key
is present
3841 @param default The default value to
return when the key
is not present (default
: [])
3842 @param casesense When false
, the values are converted to lower case
3844 ie_key = ie_key if isinstance(ie_key, str) else (ie_key or self).ie_key()
3845 val = traverse_obj(self._downloader.params, ('extractor_args', ie_key.lower(), key))
3847 return [] if default is NO_DEFAULT else default
3848 return list(val) if casesense else [x.lower() for x in val]
3850 def _yes_playlist(self, playlist_id, video_id, smuggled_data=None, *, playlist_label='playlist', video_label='video'):
3851 if not playlist_id or not video_id:
3854 no_playlist = (smuggled_data or {}).get('force_noplaylist')
3855 if no_playlist is not None:
3856 return not no_playlist
3858 video_id = '' if video_id is True else f' {video_id}'
3859 playlist_id = '' if playlist_id is True else f' {playlist_id}'
3860 if self.get_param('noplaylist'):
3861 self.to_screen(f'Downloading just the {video_label}{video_id} because of --no-playlist')
3863 self.to_screen(f'Downloading {playlist_label}{playlist_id} - add --no-playlist to download just the {video_label}{video_id}')
3866 def _error_or_warning(self, err, _count=None, _retries=0, *, fatal=True):
3867 RetryManager.report_retry(
3868 err, _count or int(fatal), _retries,
3869 info=self.to_screen, warn=self.report_warning, error=None if fatal else self.report_warning,
3870 sleep_func=self.get_param('retry_sleep_functions', {}).get('extractor'))
3872 def RetryManager(self, **kwargs):
3873 return RetryManager(self.get_param('extractor_retries', 3), self._error_or_warning, **kwargs)
3875 def _extract_generic_embeds(self, url, *args, info_dict={}, note='Extracting generic embeds', **kwargs):
3876 display_id = traverse_obj(info_dict, 'display_id', 'id')
3877 self.to_screen(f'{format_field(display_id, None, "%s: ")}{note}')
3878 return self._downloader.get_info_extractor('Generic')._extract_embeds(
3879 smuggle_url(url, {'block_ies': [self.ie_key()]}), *args, **kwargs)
3882 def extract_from_webpage(cls, ydl, url, webpage):
3883 ie = (cls if isinstance(cls._extract_from_webpage, types.MethodType)
3884 else ydl.get_info_extractor(cls.ie_key()))
3885 for info in ie._extract_from_webpage(url, webpage) or []:
3886 # url = None since we do not want to set (webpage/original)_url
3887 ydl.add_default_extra_info(info, ie, None)
3891 def _extract_from_webpage(cls, url, webpage):
3892 for embed_url in orderedSet(
3893 cls._extract_embed_urls(url, webpage) or [], lazy=True):
3894 yield cls.url_result(embed_url, None if cls._VALID_URL is False else cls)
3897 def _extract_embed_urls(cls, url, webpage):
3898 """@returns all the embed urls on the webpage"""
3899 if '_EMBED_URL_RE' not in cls.__dict__:
3900 assert isinstance(cls._EMBED_REGEX, (list, tuple))
3901 for idx, regex in enumerate(cls._EMBED_REGEX):
3902 assert regex.count('(?P<url>') == 1, \
3903 f'{cls.__name__}._EMBED_REGEX[{idx}] must have exactly 1 url group\n\t{regex}'
3904 cls._EMBED_URL_RE = tuple(map(re.compile, cls._EMBED_REGEX))
3906 for regex in cls._EMBED_URL_RE:
3907 for mobj in regex.finditer(webpage):
3908 embed_url = urllib.parse.urljoin(url, unescapeHTML(mobj.group('url')))
3909 if cls._VALID_URL is False or cls.suitable(embed_url):
3912 class StopExtraction(Exception):
3916 def _extract_url(cls, webpage): # TODO: Remove
3917 """Only for compatibility with some older extractors"""
3918 return next(iter(cls._extract_embed_urls(None, webpage) or []), None)
3921 def __init_subclass__(cls, *, plugin_name=None, **kwargs):
3923 mro = inspect.getmro(cls)
3924 super_class = cls.__wrapped__ = mro[mro.index(cls) + 1]
3925 cls.PLUGIN_NAME, cls.ie_key = plugin_name, super_class.ie_key
3926 cls.IE_NAME = f'{super_class.IE_NAME}+{plugin_name}'
3927 while getattr(super_class, '__wrapped__', None):
3928 super_class = super_class.__wrapped__
3929 setattr(sys.modules[super_class.__module__], super_class.__name__, cls)
3930 _PLUGIN_OVERRIDES[super_class].append(cls)
3932 return super().__init_subclass__(**kwargs)
3935 class SearchInfoExtractor(InfoExtractor):
3937 Base class for paged search queries extractors.
3938 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
3939 Instances should define _SEARCH_KEY and optionally _MAX_RESULTS
3942 _MAX_RESULTS = float('inf')
3943 _RETURN_TYPE = 'playlist'
3946 def _VALID_URL(cls):
3947 return rf'{cls._SEARCH_KEY}(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)'
3949 def _real_extract(self, query):
3950 prefix, query = self._match_valid_url(query).group('prefix', 'query')
3952 return self._get_n_results(query, 1)
3953 elif prefix == 'all':
3954 return self._get_n_results(query, self._MAX_RESULTS)
3958 raise ExtractorError(f'invalid download number {n} for query "{query}"')
3959 elif n > self._MAX_RESULTS:
3960 self.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3961 n = self._MAX_RESULTS
3962 return self._get_n_results(query, n)
3964 def _get_n_results(self, query, n):
3965 """Get a specified number of results for a query.
3966 Either this function or _search_results must be overridden by subclasses """
3967 return self.playlist_result(
3968 itertools.islice(self._search_results(query), 0, None if n == float('inf') else n),
3971 def _search_results(self, query):
3972 """Returns an iterator of search results"""
3973 raise NotImplementedError('This method must be implemented by subclasses')
3976 def SEARCH_KEY(cls):
3977 return cls._SEARCH_KEY
3980 class UnsupportedURLIE(InfoExtractor):
3985 def _real_extract(self, url):
3986 raise UnsupportedError(url)
3989 _PLUGIN_OVERRIDES = collections.defaultdict(list)