21 import xml
.etree
.ElementTree
23 from ..compat
import functools
# isort: split
24 from ..compat
import compat_etree_fromstring
, compat_expanduser
, compat_os_name
25 from ..cookies
import LenientSimpleCookie
26 from ..downloader
.f4m
import get_base_url
, remove_encrypted_media
65 parse_m3u8_attributes
,
95 """Information Extractor class.
97 Information extractors are the classes that, given a URL, extract
98 information about the video (or videos) the URL refers to. This
99 information includes the real video URL, the video title, author and
100 others. The information is stored in a dictionary which is then
101 passed to the YoutubeDL. The YoutubeDL processes this
102 information possibly downloading the video to the file system, among
103 other possible outcomes.
105 The type field determines the type of the result.
106 By far the most common value (and the default if _type is missing) is
107 "video", which indicates a single video.
109 For a video, the dictionaries must include the following fields:
111 id: Video identifier.
112 title: Video title, unescaped. Set to an empty string if video has
113 no title as opposed to "None" which signifies that the
114 extractor failed to obtain a title
116 Additionally, it must contain either a formats entry or a url one:
118 formats: A list of dictionaries for each format available, ordered
119 from worst to best quality.
122 * url The mandatory URL representing the media:
123 for plain file media - HTTP URL of this file,
125 for HLS - URL of the M3U8 media playlist,
126 for HDS - URL of the F4M manifest,
128 - HTTP URL to plain file media (in case of
130 - URL of the MPD manifest or base URL
131 representing the media if MPD manifest
132 is parsed from a string (in case of
134 for MSS - URL of the ISM manifest.
135 * request_data Data to send in POST request to the URL
137 The URL of the manifest file in case of
139 for HLS - URL of the M3U8 master playlist,
140 for HDS - URL of the F4M manifest,
141 for DASH - URL of the MPD manifest,
142 for MSS - URL of the ISM manifest.
143 * manifest_stream_number (For internal use only)
144 The index of the stream in the manifest file
145 * ext Will be calculated from URL if missing
146 * format A human-readable description of the format
147 ("mp4 container with h264/opus").
148 Calculated from the format_id, width, height.
149 and format_note fields if missing.
150 * format_id A short description of the format
151 ("mp4_h264_opus" or "19").
152 Technically optional, but strongly recommended.
153 * format_note Additional info about the format
154 ("3D" or "DASH video")
155 * width Width of the video, if known
156 * height Height of the video, if known
157 * aspect_ratio Aspect ratio of the video, if known
158 Automatically calculated from width and height
159 * resolution Textual description of width and height
160 Automatically calculated from width and height
161 * dynamic_range The dynamic range of the video. One of:
162 "SDR" (None), "HDR10", "HDR10+, "HDR12", "HLG, "DV"
163 * tbr Average bitrate of audio and video in KBit/s
164 * abr Average audio bitrate in KBit/s
165 * acodec Name of the audio codec in use
166 * asr Audio sampling rate in Hertz
167 * audio_channels Number of audio channels
168 * vbr Average video bitrate in KBit/s
170 * vcodec Name of the video codec in use
171 * container Name of the container format
172 * filesize The number of bytes, if known in advance
173 * filesize_approx An estimate for the number of bytes
174 * player_url SWF Player URL (used for rtmpdump).
175 * protocol The protocol that will be used for the actual
176 download, lower-case. One of "http", "https" or
177 one of the protocols defined in downloader.PROTOCOL_MAP
179 Base URL for fragments. Each fragment's path
180 value (if present) will be relative to
182 * fragments A list of fragments of a fragmented media.
183 Each fragment entry must contain either an url
184 or a path. If an url is present it should be
185 considered by a client. Otherwise both path and
186 fragment_base_url must be present. Here is
187 the list of all potential fields:
188 * "url" - fragment's URL
189 * "path" - fragment's path relative to
191 * "duration" (optional, int or float)
192 * "filesize" (optional, int)
193 * is_from_start Is a live format that can be downloaded
194 from the start. Boolean
195 * preference Order number of this format. If this field is
196 present and not None, the formats get sorted
197 by this field, regardless of all other values.
198 -1 for default (order by other properties),
199 -2 or smaller for less than default.
200 < -1000 to hide the format (if there is
201 another one which is strictly better)
202 * language Language code, e.g. "de" or "en-US".
203 * language_preference Is this in the language mentioned in
205 10 if it's what the URL is about,
206 -1 for default (don't know),
207 -10 otherwise, other values reserved for now.
208 * quality Order number of the video quality of this
209 format, irrespective of the file format.
210 -1 for default (order by other properties),
211 -2 or smaller for less than default.
212 * source_preference Order number for this video source
213 (quality takes higher priority)
214 -1 for default (order by other properties),
215 -2 or smaller for less than default.
216 * http_headers A dictionary of additional HTTP headers
217 to add to the request.
218 * stretched_ratio If given and not 1, indicates that the
219 video's pixels are not square.
220 width : height ratio as float.
221 * no_resume The server does not support resuming the
222 (HTTP or RTMP) download. Boolean.
223 * has_drm The format has DRM and cannot be downloaded. Boolean
224 * extra_param_to_segment_url A query string to append to each
225 fragment's URL, or to update each existing query string
226 with. Only applied by the native HLS/DASH downloaders.
227 * hls_aes A dictionary of HLS AES-128 decryption information
228 used by the native HLS downloader to override the
229 values in the media playlist when an '#EXT-X-KEY' tag
230 is present in the playlist:
231 * uri The URI from which the key will be downloaded
232 * key The key (as hex) used to decrypt fragments.
233 If `key` is given, any key URI will be ignored
234 * iv The IV (as hex) used to decrypt fragments
235 * downloader_options A dictionary of downloader options
236 (For internal use only)
237 * http_chunk_size Chunk size for HTTP downloads
238 * ffmpeg_args Extra arguments for ffmpeg downloader
239 RTMP formats can also have the additional fields: page_url,
240 app, play_path, tc_url, flash_version, rtmp_live, rtmp_conn,
241 rtmp_protocol, rtmp_real_time
243 url: Final video URL.
244 ext: Video filename extension.
245 format: The video format, defaults to ext (used for --get-format)
246 player_url: SWF Player URL (used for rtmpdump).
248 The following fields are optional:
250 direct: True if a direct video file was given (must only be set by GenericIE)
251 alt_title: A secondary title of the video.
252 display_id An alternative identifier for the video, not necessarily
253 unique, but available before title. Typically, id is
254 something like "4234987", title "Dancing naked mole rats",
255 and display_id "dancing-naked-mole-rats"
256 thumbnails: A list of dictionaries, with the following entries:
257 * "id" (optional, string) - Thumbnail format ID
259 * "preference" (optional, int) - quality of the image
260 * "width" (optional, int)
261 * "height" (optional, int)
262 * "resolution" (optional, string "{width}x{height}",
264 * "filesize" (optional, int)
265 * "http_headers" (dict) - HTTP headers for the request
266 thumbnail: Full URL to a video thumbnail image.
267 description: Full video description.
268 uploader: Full name of the video uploader.
269 license: License name the video is licensed under.
270 creator: The creator of the video.
271 timestamp: UNIX timestamp of the moment the video was uploaded
272 upload_date: Video upload date in UTC (YYYYMMDD).
273 If not explicitly set, calculated from timestamp
274 release_timestamp: UNIX timestamp of the moment the video was released.
275 If it is not clear whether to use timestamp or this, use the former
276 release_date: The date (YYYYMMDD) when the video was released in UTC.
277 If not explicitly set, calculated from release_timestamp
278 modified_timestamp: UNIX timestamp of the moment the video was last modified.
279 modified_date: The date (YYYYMMDD) when the video was last modified in UTC.
280 If not explicitly set, calculated from modified_timestamp
281 uploader_id: Nickname or id of the video uploader.
282 uploader_url: Full URL to a personal webpage of the video uploader.
283 channel: Full name of the channel the video is uploaded on.
284 Note that channel fields may or may not repeat uploader
285 fields. This depends on a particular extractor.
286 channel_id: Id of the channel.
287 channel_url: Full URL to a channel webpage.
288 channel_follower_count: Number of followers of the channel.
289 location: Physical location where the video was filmed.
290 subtitles: The available subtitles as a dictionary in the format
291 {tag: subformats}. "tag" is usually a language code, and
292 "subformats" is a list sorted from lower to higher
293 preference, each element is a dictionary with the "ext"
295 * "data": The subtitles file contents
296 * "url": A URL pointing to the subtitles file
297 It can optionally also have:
298 * "name": Name or description of the subtitles
299 * "http_headers": A dictionary of additional HTTP headers
300 to add to the request.
301 "ext" will be calculated from URL if missing
302 automatic_captions: Like 'subtitles'; contains automatically generated
303 captions instead of normal subtitles
304 duration: Length of the video in seconds, as an integer or float.
305 view_count: How many users have watched the video on the platform.
306 concurrent_view_count: How many users are currently watching the video on the platform.
307 like_count: Number of positive ratings of the video
308 dislike_count: Number of negative ratings of the video
309 repost_count: Number of reposts of the video
310 average_rating: Average rating give by users, the scale used depends on the webpage
311 comment_count: Number of comments on the video
312 comments: A list of comments, each with one or more of the following
313 properties (all but one of text or html optional):
314 * "author" - human-readable name of the comment author
315 * "author_id" - user ID of the comment author
316 * "author_thumbnail" - The thumbnail of the comment author
317 * "author_url" - The url to the comment author's page
318 * "author_is_verified" - Whether the author is verified
320 * "author_is_uploader" - Whether the comment is made by
323 * "html" - Comment as HTML
324 * "text" - Plain text of the comment
325 * "timestamp" - UNIX timestamp of comment
326 * "parent" - ID of the comment this one is replying to.
327 Set to "root" to indicate that this is a
328 comment to the original video.
329 * "like_count" - Number of positive ratings of the comment
330 * "dislike_count" - Number of negative ratings of the comment
331 * "is_favorited" - Whether the comment is marked as
332 favorite by the video uploader
333 * "is_pinned" - Whether the comment is pinned to
334 the top of the comments
335 age_limit: Age restriction for the video, as an integer (years)
336 webpage_url: The URL to the video webpage, if given to yt-dlp it
337 should allow to get the same result again. (It will be set
338 by YoutubeDL if it's missing)
339 categories: A list of categories that the video falls in, for example
341 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
342 cast: A list of the video cast
343 is_live: True, False, or None (=unknown). Whether this video is a
344 live stream that goes on instead of a fixed-length video.
345 was_live: True, False, or None (=unknown). Whether this video was
346 originally a live stream.
347 live_status: None (=unknown), 'is_live', 'is_upcoming', 'was_live', 'not_live',
348 or 'post_live' (was live, but VOD is not yet processed)
349 If absent, automatically set from is_live, was_live
350 start_time: Time in seconds where the reproduction should start, as
351 specified in the URL.
352 end_time: Time in seconds where the reproduction should end, as
353 specified in the URL.
354 chapters: A list of dictionaries, with the following entries:
355 * "start_time" - The start time of the chapter in seconds
356 * "end_time" - The end time of the chapter in seconds
357 * "title" (optional, string)
358 heatmap: A list of dictionaries, with the following entries:
359 * "start_time" - The start time of the data point in seconds
360 * "end_time" - The end time of the data point in seconds
361 * "value" - The normalized value of the data point (float between 0 and 1)
362 playable_in_embed: Whether this video is allowed to play in embedded
363 players on other sites. Can be True (=always allowed),
364 False (=never allowed), None (=unknown), or a string
365 specifying the criteria for embedability; e.g. 'whitelist'
366 availability: Under what condition the video is available. One of
367 'private', 'premium_only', 'subscriber_only', 'needs_auth',
368 'unlisted' or 'public'. Use 'InfoExtractor._availability'
370 _old_archive_ids: A list of old archive ids needed for backward compatibility
371 _format_sort_fields: A list of fields to use for sorting formats
372 __post_extractor: A function to be called just before the metadata is
373 written to either disk, logger or console. The function
374 must return a dict which will be added to the info_dict.
375 This is usefull for additional information that is
376 time-consuming to extract. Note that the fields thus
377 extracted will not be available to output template and
378 match_filter. So, only "comments" and "comment_count" are
379 currently allowed to be extracted via this method.
381 The following fields should only be used when the video belongs to some logical
384 chapter: Name or title of the chapter the video belongs to.
385 chapter_number: Number of the chapter the video belongs to, as an integer.
386 chapter_id: Id of the chapter the video belongs to, as a unicode string.
388 The following fields should only be used when the video is an episode of some
389 series, programme or podcast:
391 series: Title of the series or programme the video episode belongs to.
392 series_id: Id of the series or programme the video episode belongs to, as a unicode string.
393 season: Title of the season the video episode belongs to.
394 season_number: Number of the season the video episode belongs to, as an integer.
395 season_id: Id of the season the video episode belongs to, as a unicode string.
396 episode: Title of the video episode. Unlike mandatory video title field,
397 this field should denote the exact title of the video episode
398 without any kind of decoration.
399 episode_number: Number of the video episode within a season, as an integer.
400 episode_id: Id of the video episode, as a unicode string.
402 The following fields should only be used when the media is a track or a part of
405 track: Title of the track.
406 track_number: Number of the track within an album or a disc, as an integer.
407 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
409 artist: Artist(s) of the track.
410 genre: Genre(s) of the track.
411 album: Title of the album the track belongs to.
412 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
413 album_artist: List of all artists appeared on the album (e.g.
414 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
416 disc_number: Number of the disc or other physical medium the track belongs to,
418 release_year: Year (YYYY) when the album was released.
419 composer: Composer of the piece
421 The following fields should only be set for clips that should be cut from the original video:
423 section_start: Start time of the section in seconds
424 section_end: End time of the section in seconds
426 The following fields should only be set for storyboards:
427 rows: Number of rows in each storyboard fragment, as an integer
428 columns: Number of columns in each storyboard fragment, as an integer
430 Unless mentioned otherwise, the fields should be Unicode strings.
432 Unless mentioned otherwise, None is equivalent to absence of information.
435 _type "playlist" indicates multiple videos.
436 There must be a key "entries", which is a list, an iterable, or a PagedList
437 object, each element of which is a valid dictionary by this specification.
439 Additionally, playlists can have "id", "title", and any other relevant
440 attributes with the same semantics as videos (see above).
442 It can also have the following optional fields:
444 playlist_count: The total number of videos in a playlist. If not given,
445 YoutubeDL tries to calculate it from "entries"
448 _type "multi_video" indicates that there are multiple videos that
449 form a single show, for examples multiple acts of an opera or TV episode.
450 It must have an entries key like a playlist and contain all the keys
451 required for a video at the same time.
454 _type "url" indicates that the video must be extracted from another
455 location, possibly by a different extractor. Its only required key is:
456 "url" - the next URL to extract.
457 The key "ie_key" can be set to the class name (minus the trailing "IE",
458 e.g. "Youtube") if the extractor class is known in advance.
459 Additionally, the dictionary may have any properties of the resolved entity
460 known in advance, for example "title" if the title of the referred video is
464 _type "url_transparent" entities have the same specification as "url", but
465 indicate that the given additional information is more precise than the one
466 associated with the resolved URL.
467 This is useful when a site employs a video service that hosts the video and
468 its technical metadata, but that video service does not embed a useful
469 title, description etc.
472 Subclasses of this should also be added to the list of extractors and
473 should define a _VALID_URL regexp and, re-define the _real_extract() and
474 (optionally) _real_initialize() methods.
476 Subclasses may also override suitable() if necessary, but ensure the function
477 signature is preserved and that this function imports everything it needs
478 (except other extractors), so that lazy_extractors works correctly.
480 Subclasses can define a list of _EMBED_REGEX, which will be searched for in
481 the HTML of Generic webpages. It may also override _extract_embed_urls
482 or _extract_from_webpage as necessary. While these are normally classmethods,
483 _extract_from_webpage is allowed to be an instance method.
485 _extract_from_webpage may raise self.StopExtraction() to stop further
486 processing of the webpage and obtain exclusive rights to it. This is useful
487 when the extractor cannot reliably be matched using just the URL,
488 e.g. invidious/peertube instances
490 Embed-only extractors can be defined by setting _VALID_URL = False.
492 To support username + password (or netrc) login, the extractor must define a
493 _NETRC_MACHINE and re-define _perform_login(username, password) and
494 (optionally) _initialize_pre_login() methods. The _perform_login method will
495 be called between _initialize_pre_login and _real_initialize if credentials
496 are passed by the user. In cases where it is necessary to have the login
497 process as part of the extraction rather than initialization, _perform_login
498 can be left undefined.
500 _GEO_BYPASS attribute may be set to False in order to disable
501 geo restriction bypass mechanisms for a particular extractor.
502 Though it won't disable explicit geo restriction bypass based on
503 country code provided with geo_bypass_country.
505 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
506 countries for this extractor. One of these countries will be used by
507 geo restriction bypass mechanism right away in order to bypass
508 geo restriction, of course, if the mechanism is not disabled.
510 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
511 IP blocks in CIDR notation for this extractor. One of these IP blocks
512 will be used by geo restriction bypass mechanism similarly
515 The _ENABLED attribute should be set to False for IEs that
516 are disabled by default and must be explicitly enabled.
518 The _WORKING attribute should be set to False for broken IEs
519 in order to warn the users and skip the tests.
524 _x_forwarded_for_ip
= None
526 _GEO_COUNTRIES
= None
527 _GEO_IP_BLOCKS
= None
530 _NETRC_MACHINE
= None
536 def _login_hint(self
, method
=NO_DEFAULT
, netrc
=None):
537 password_hint
= f
'--username and --password, or --netrc ({netrc or self._NETRC_MACHINE}) to provide account credentials'
540 'any': f
'Use --cookies, --cookies-from-browser, {password_hint}',
541 'password': f
'Use {password_hint}',
543 'Use --cookies-from-browser or --cookies for the authentication. '
544 'See https://github.com/yt-dlp/yt-dlp/wiki/FAQ#how-do-i-pass-cookies-to-yt-dlp for how to manually pass cookies'),
545 }[method
if method
is not NO_DEFAULT
else 'any' if self
.supports_login() else 'cookies']
547 def __init__(self
, downloader
=None):
548 """Constructor. Receives an optional downloader (a YoutubeDL instance).
549 If a downloader is not passed during initialization,
550 it must be set using "set_downloader()" before "extract()" is called"""
552 self
._x
_forwarded
_for
_ip
= None
553 self
._printed
_messages
= set()
554 self
.set_downloader(downloader
)
557 def _match_valid_url(cls
, url
):
558 if cls
._VALID
_URL
is False:
560 # This does not use has/getattr intentionally - we want to know whether
561 # we have cached the regexp for *this* class, whereas getattr would also
562 # match the superclass
563 if '_VALID_URL_RE' not in cls
.__dict
__:
564 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
565 return cls
._VALID
_URL
_RE
.match(url
)
568 def suitable(cls
, url
):
569 """Receives a URL and returns True if suitable for this IE."""
570 # This function must import everything it needs (except other extractors),
571 # so that lazy_extractors works correctly
572 return cls
._match
_valid
_url
(url
) is not None
575 def _match_id(cls
, url
):
576 return cls
._match
_valid
_url
(url
).group('id')
579 def get_temp_id(cls
, url
):
581 return cls
._match
_id
(url
)
582 except (IndexError, AttributeError):
587 """Getter method for _WORKING."""
591 def supports_login(cls
):
592 return bool(cls
._NETRC
_MACHINE
)
594 def initialize(self
):
595 """Initializes an instance (authentication, etc)."""
596 self
._printed
_messages
= set()
597 self
._initialize
_geo
_bypass
({
598 'countries': self
._GEO
_COUNTRIES
,
599 'ip_blocks': self
._GEO
_IP
_BLOCKS
,
602 self
._initialize
_pre
_login
()
603 if self
.supports_login():
604 username
, password
= self
._get
_login
_info
()
606 self
._perform
_login
(username
, password
)
607 elif self
.get_param('username') and False not in (self
.IE_DESC
, self
._NETRC
_MACHINE
):
608 self
.report_warning(f
'Login with password is not supported for this website. {self._login_hint("cookies")}')
609 self
._real
_initialize
()
612 def _initialize_geo_bypass(self
, geo_bypass_context
):
614 Initialize geo restriction bypass mechanism.
616 This method is used to initialize geo bypass mechanism based on faking
617 X-Forwarded-For HTTP header. A random country from provided country list
618 is selected and a random IP belonging to this country is generated. This
619 IP will be passed as X-Forwarded-For HTTP header in all subsequent
622 This method will be used for initial geo bypass mechanism initialization
623 during the instance initialization with _GEO_COUNTRIES and
626 You may also manually call it from extractor's code if geo bypass
627 information is not available beforehand (e.g. obtained during
628 extraction) or due to some other reason. In this case you should pass
629 this information in geo bypass context passed as first argument. It may
630 contain following fields:
632 countries: List of geo unrestricted countries (similar
634 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
635 (similar to _GEO_IP_BLOCKS)
638 if not self
._x
_forwarded
_for
_ip
:
640 # Geo bypass mechanism is explicitly disabled by user
641 if not self
.get_param('geo_bypass', True):
644 if not geo_bypass_context
:
645 geo_bypass_context
= {}
647 # Backward compatibility: previously _initialize_geo_bypass
648 # expected a list of countries, some 3rd party code may still use
650 if isinstance(geo_bypass_context
, (list, tuple)):
651 geo_bypass_context
= {
652 'countries': geo_bypass_context
,
655 # The whole point of geo bypass mechanism is to fake IP
656 # as X-Forwarded-For HTTP header based on some IP block or
659 # Path 1: bypassing based on IP block in CIDR notation
661 # Explicit IP block specified by user, use it right away
662 # regardless of whether extractor is geo bypassable or not
663 ip_block
= self
.get_param('geo_bypass_ip_block', None)
665 # Otherwise use random IP block from geo bypass context but only
666 # if extractor is known as geo bypassable
668 ip_blocks
= geo_bypass_context
.get('ip_blocks')
669 if self
._GEO
_BYPASS
and ip_blocks
:
670 ip_block
= random
.choice(ip_blocks
)
673 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(ip_block
)
674 self
.write_debug(f
'Using fake IP {self._x_forwarded_for_ip} as X-Forwarded-For')
677 # Path 2: bypassing based on country code
679 # Explicit country code specified by user, use it right away
680 # regardless of whether extractor is geo bypassable or not
681 country
= self
.get_param('geo_bypass_country', None)
683 # Otherwise use random country code from geo bypass context but
684 # only if extractor is known as geo bypassable
686 countries
= geo_bypass_context
.get('countries')
687 if self
._GEO
_BYPASS
and countries
:
688 country
= random
.choice(countries
)
691 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country
)
692 self
._downloader
.write_debug(
693 f
'Using fake IP {self._x_forwarded_for_ip} ({country.upper()}) as X-Forwarded-For')
695 def extract(self
, url
):
696 """Extracts URL information and returns it in list of dicts."""
701 self
.to_screen('Extracting URL: %s' % (
702 url
if self
.get_param('verbose') else truncate_string(url
, 100, 20)))
703 ie_result
= self
._real
_extract
(url
)
704 if ie_result
is None:
706 if self
._x
_forwarded
_for
_ip
:
707 ie_result
['__x_forwarded_for_ip'] = self
._x
_forwarded
_for
_ip
708 subtitles
= ie_result
.get('subtitles') or {}
709 if 'no-live-chat' in self
.get_param('compat_opts'):
710 for lang
in ('live_chat', 'comments', 'danmaku'):
711 subtitles
.pop(lang
, None)
713 except GeoRestrictedError
as e
:
714 if self
.__maybe
_fake
_ip
_and
_retry
(e
.countries
):
717 except UnsupportedError
:
719 except ExtractorError
as e
:
720 e
.video_id
= e
.video_id
or self
.get_temp_id(url
),
721 e
.ie
= e
.ie
or self
.IE_NAME
,
722 e
.traceback
= e
.traceback
or sys
.exc_info()[2]
724 except http
.client
.IncompleteRead
as e
:
725 raise ExtractorError('A network error has occurred.', cause
=e
, expected
=True, video_id
=self
.get_temp_id(url
))
726 except (KeyError, StopIteration) as e
:
727 raise ExtractorError('An extractor error has occurred.', cause
=e
, video_id
=self
.get_temp_id(url
))
729 def __maybe_fake_ip_and_retry(self
, countries
):
730 if (not self
.get_param('geo_bypass_country', None)
732 and self
.get_param('geo_bypass', True)
733 and not self
._x
_forwarded
_for
_ip
735 country_code
= random
.choice(countries
)
736 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country_code
)
737 if self
._x
_forwarded
_for
_ip
:
739 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
740 % (self
._x
_forwarded
_for
_ip
, country_code
.upper()))
744 def set_downloader(self
, downloader
):
745 """Sets a YoutubeDL instance as the downloader for this IE."""
746 self
._downloader
= downloader
750 return self
._downloader
.cache
754 return self
._downloader
.cookiejar
756 def _initialize_pre_login(self
):
757 """ Initialization before login. Redefine in subclasses."""
760 def _perform_login(self
, username
, password
):
761 """ Login with username and password. Redefine in subclasses."""
764 def _real_initialize(self
):
765 """Real initialization process. Redefine in subclasses."""
768 def _real_extract(self
, url
):
769 """Real extraction process. Redefine in subclasses."""
770 raise NotImplementedError('This method must be implemented by subclasses')
774 """A string for getting the InfoExtractor with get_info_extractor"""
775 return cls
.__name
__[:-2]
779 return cls
.__name
__[:-2]
782 def __can_accept_status_code(err
, expected_status
):
783 assert isinstance(err
, urllib
.error
.HTTPError
)
784 if expected_status
is None:
786 elif callable(expected_status
):
787 return expected_status(err
.code
) is True
789 return err
.code
in variadic(expected_status
)
791 def _create_request(self
, url_or_request
, data
=None, headers
=None, query
=None):
792 if isinstance(url_or_request
, urllib
.request
.Request
):
793 return update_Request(url_or_request
, data
=data
, headers
=headers
, query
=query
)
795 url_or_request
= update_url_query(url_or_request
, query
)
796 return sanitized_Request(url_or_request
, data
, headers
or {})
798 def _request_webpage(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, data
=None, headers
=None, query
=None, expected_status
=None):
800 Return the response handle.
802 See _download_webpage docstring for arguments specification.
804 if not self
._downloader
._first
_webpage
_request
:
805 sleep_interval
= self
.get_param('sleep_interval_requests') or 0
806 if sleep_interval
> 0:
807 self
.to_screen('Sleeping %s seconds ...' % sleep_interval
)
808 time
.sleep(sleep_interval
)
810 self
._downloader
._first
_webpage
_request
= False
813 self
.report_download_webpage(video_id
)
814 elif note
is not False:
816 self
.to_screen(str(note
))
818 self
.to_screen(f
'{video_id}: {note}')
820 # Some sites check X-Forwarded-For HTTP header in order to figure out
821 # the origin of the client behind proxy. This allows bypassing geo
822 # restriction by faking this header's value to IP that belongs to some
823 # geo unrestricted country. We will do so once we encounter any
824 # geo restriction error.
825 if self
._x
_forwarded
_for
_ip
:
826 headers
= (headers
or {}).copy()
827 headers
.setdefault('X-Forwarded-For', self
._x
_forwarded
_for
_ip
)
830 return self
._downloader
.urlopen(self
._create
_request
(url_or_request
, data
, headers
, query
))
831 except network_exceptions
as err
:
832 if isinstance(err
, urllib
.error
.HTTPError
):
833 if self
.__can
_accept
_status
_code
(err
, expected_status
):
834 # Retain reference to error to prevent file object from
835 # being closed before it can be read. Works around the
836 # effects of <https://bugs.python.org/issue15002>
837 # introduced in Python 3.4.1.
844 errnote
= 'Unable to download webpage'
846 errmsg
= f
'{errnote}: {error_to_compat_str(err)}'
848 raise ExtractorError(errmsg
, cause
=err
)
850 self
.report_warning(errmsg
)
853 def _download_webpage_handle(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True,
854 encoding
=None, data
=None, headers
={}, query={}
, expected_status
=None):
856 Return a tuple (page content as string, URL handle).
859 url_or_request -- plain text URL as a string or
860 a urllib.request.Request object
861 video_id -- Video/playlist/item identifier (string)
864 note -- note printed before downloading (string)
865 errnote -- note printed in case of an error (string)
866 fatal -- flag denoting whether error should be considered fatal,
867 i.e. whether it should cause ExtractionError to be raised,
868 otherwise a warning will be reported and extraction continued
869 encoding -- encoding for a page content decoding, guessed automatically
870 when not explicitly specified
871 data -- POST data (bytes)
872 headers -- HTTP headers (dict)
873 query -- URL query (dict)
874 expected_status -- allows to accept failed HTTP requests (non 2xx
875 status code) by explicitly specifying a set of accepted status
876 codes. Can be any of the following entities:
877 - an integer type specifying an exact failed status code to
879 - a list or a tuple of integer types specifying a list of
880 failed status codes to accept
881 - a callable accepting an actual failed status code and
882 returning True if it should be accepted
883 Note that this argument does not affect success status codes (2xx)
884 which are always accepted.
887 # Strip hashes from the URL (#1038)
888 if isinstance(url_or_request
, str):
889 url_or_request
= url_or_request
.partition('#')[0]
891 urlh
= self
._request
_webpage
(url_or_request
, video_id
, note
, errnote
, fatal
, data
=data
, headers
=headers
, query
=query
, expected_status
=expected_status
)
895 content
= self
._webpage
_read
_content
(urlh
, url_or_request
, video_id
, note
, errnote
, fatal
, encoding
=encoding
)
896 return (content
, urlh
)
899 def _guess_encoding_from_content(content_type
, webpage_bytes
):
900 m
= re
.match(r
'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type
)
902 encoding
= m
.group(1)
904 m
= re
.search(br
'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
905 webpage_bytes[:1024])
907 encoding = m.group(1).decode('ascii')
908 elif webpage_bytes.startswith(b'\xff\xfe'):
915 def __check_blocked(self, content):
916 first_block = content[:512]
917 if ('<title>Access to this site is blocked</title>' in content
918 and 'Websense' in first_block):
919 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
920 blocked_iframe = self._html_search_regex(
921 r'<iframe src="([^
"]+)"', content,
922 'Websense information URL
', default=None)
924 msg += ' Visit
%s for more details
' % blocked_iframe
925 raise ExtractorError(msg, expected=True)
926 if '<title
>The URL you requested has been blocked
</title
>' in first_block:
928 'Access to this webpage has been blocked by Indian censorship
. '
929 'Use a VPN
or proxy
server (with --proxy
) to route around it
.')
930 block_msg = self._html_search_regex(
931 r'</h1
><p
>(.*?
)</p
>',
932 content, 'block message
', default=None)
934 msg += ' (Message
: "%s")' % block_msg.replace('\n', ' ')
935 raise ExtractorError(msg, expected=True)
936 if ('<title
>TTK
:: Доступ к ресурсу ограничен
</title
>' in content
937 and 'blocklist
.rkn
.gov
.ru
' in content):
938 raise ExtractorError(
939 'Access to this webpage has been blocked by decision of the Russian government
. '
940 'Visit http
://blocklist
.rkn
.gov
.ru
/ for a block reason
.',
943 def _request_dump_filename(self, url, video_id):
944 basen = f'{video_id}_{url}
'
945 trim_length = self.get_param('trim_file_name
') or 240
946 if len(basen) > trim_length:
947 h = '___
' + hashlib.md5(basen.encode('utf
-8')).hexdigest()
948 basen = basen[:trim_length - len(h)] + h
949 filename = sanitize_filename(f'{basen}
.dump
', restricted=True)
950 # Working around MAX_PATH limitation on Windows (see
951 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
952 if compat_os_name == 'nt
':
953 absfilepath = os.path.abspath(filename)
954 if len(absfilepath) > 259:
955 filename = fR'\\?\{absfilepath}
'
958 def __decode_webpage(self, webpage_bytes, encoding, headers):
960 encoding = self._guess_encoding_from_content(headers.get('Content
-Type
', ''), webpage_bytes)
962 return webpage_bytes.decode(encoding, 'replace
')
964 return webpage_bytes.decode('utf
-8', 'replace
')
966 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
967 webpage_bytes = urlh.read()
968 if prefix is not None:
969 webpage_bytes = prefix + webpage_bytes
970 if self.get_param('dump_intermediate_pages
', False):
971 self.to_screen('Dumping request to
' + urlh.geturl())
972 dump = base64.b64encode(webpage_bytes).decode('ascii
')
973 self._downloader.to_screen(dump)
974 if self.get_param('write_pages
'):
975 filename = self._request_dump_filename(urlh.geturl(), video_id)
976 self.to_screen(f'Saving request to {filename}
')
977 with open(filename, 'wb
') as outf:
978 outf.write(webpage_bytes)
980 content = self.__decode_webpage(webpage_bytes, encoding, urlh.headers)
981 self.__check_blocked(content)
985 def __print_error(self, errnote, fatal, video_id, err):
987 raise ExtractorError(f'{video_id}
: {errnote}
', cause=err)
989 self.report_warning(f'{video_id}
: {errnote}
: {err}
')
991 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True, errnote=None):
993 xml_string = transform_source(xml_string)
995 return compat_etree_fromstring(xml_string.encode('utf
-8'))
996 except xml.etree.ElementTree.ParseError as ve:
997 self.__print_error('Failed to parse XML
' if errnote is None else errnote, fatal, video_id, ve)
999 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True, errnote=None, **parser_kwargs):
1002 json_string, cls=LenientJSONDecoder, strict=False, transform_source=transform_source, **parser_kwargs)
1003 except ValueError as ve:
1004 self.__print_error('Failed to parse JSON
' if errnote is None else errnote, fatal, video_id, ve)
1006 def _parse_socket_response_as_json(self, data, *args, **kwargs):
1007 return self._parse_json(data[data.find('{'):data.rfind('}
') + 1], *args, **kwargs)
1009 def __create_download_methods(name, parser, note, errnote, return_value):
1011 def parse(ie, content, *args, errnote=errnote, **kwargs):
1014 if errnote is False:
1015 kwargs['errnote
'] = errnote
1016 # parser is fetched by name so subclasses can override it
1017 return getattr(ie, parser)(content, *args, **kwargs)
1019 def download_handle(self, url_or_request, video_id, note=note, errnote=errnote, transform_source=None,
1020 fatal=True, encoding=None, data=None, headers={}, query={}, expected_status=None):
1021 res = self._download_webpage_handle(
1022 url_or_request, video_id, note=note, errnote=errnote, fatal=fatal, encoding=encoding,
1023 data=data, headers=headers, query=query, expected_status=expected_status)
1027 return parse(self, content, video_id, transform_source=transform_source, fatal=fatal, errnote=errnote), urlh
1029 def download_content(self, url_or_request, video_id, note=note, errnote=errnote, transform_source=None,
1030 fatal=True, encoding=None, data=None, headers={}, query={}, expected_status=None):
1031 if self.get_param('load_pages
'):
1032 url_or_request = self._create_request(url_or_request, data, headers, query)
1033 filename = self._request_dump_filename(url_or_request.full_url, video_id)
1034 self.to_screen(f'Loading request
from {filename}
')
1036 with open(filename, 'rb
') as dumpf:
1037 webpage_bytes = dumpf.read()
1038 except OSError as e:
1039 self.report_warning(f'Unable to load request
from disk
: {e}
')
1041 content = self.__decode_webpage(webpage_bytes, encoding, url_or_request.headers)
1042 return parse(self, content, video_id, transform_source=transform_source, fatal=fatal, errnote=errnote)
1046 'transform_source
': transform_source,
1048 'encoding
': encoding,
1052 'expected_status
': expected_status,
1055 kwargs.pop('transform_source
')
1056 # The method is fetched by name so subclasses can override _download_..._handle
1057 res = getattr(self, download_handle.__name__)(url_or_request, video_id, **kwargs)
1058 return res if res is False else res[0]
1060 def impersonate(func, name, return_value):
1061 func.__name__, func.__qualname__ = name, f'InfoExtractor
.{name}
'
1063 @param transform_source Apply this transformation before parsing
1064 @returns {return_value}
1066 See _download_webpage_handle docstring for other arguments specification
1069 impersonate(download_handle, f'_download_{name}_handle
', f'({return_value}
, URL handle
)')
1070 impersonate(download_content, f'_download_{name}
', f'{return_value}
')
1071 return download_handle, download_content
1073 _download_xml_handle, _download_xml = __create_download_methods(
1074 'xml
', '_parse_xml
', 'Downloading XML
', 'Unable to download XML
', 'xml
as an xml
.etree
.ElementTree
.Element
')
1075 _download_json_handle, _download_json = __create_download_methods(
1076 'json
', '_parse_json
', 'Downloading JSON metadata
', 'Unable to download JSON metadata
', 'JSON
object as a
dict')
1077 _download_socket_json_handle, _download_socket_json = __create_download_methods(
1078 'socket_json
', '_parse_socket_response_as_json
', 'Polling socket
', 'Unable to poll socket
', 'JSON
object as a
dict')
1079 __download_webpage = __create_download_methods('webpage
', None, None, None, 'data of the page
as a string
')[1]
1081 def _download_webpage(
1082 self, url_or_request, video_id, note=None, errnote=None,
1083 fatal=True, tries=1, timeout=NO_DEFAULT, *args, **kwargs):
1085 Return the data of the page as a string.
1088 tries -- number of tries
1089 timeout -- sleep interval between tries
1091 See _download_webpage_handle docstring for other arguments specification.
1094 R''' # NB: These are unused; should they be deprecated?
1096 self._downloader.deprecation_warning('tries argument
is deprecated
in InfoExtractor
._download
_webpage
')
1097 if timeout is NO_DEFAULT:
1100 self._downloader.deprecation_warning('timeout argument
is deprecated
in InfoExtractor
._download
_webpage
')
1106 return self.__download_webpage(url_or_request, video_id, note, errnote, None, fatal, *args, **kwargs)
1107 except http.client.IncompleteRead as e:
1109 if try_count >= tries:
1111 self._sleep(timeout, video_id)
1113 def report_warning(self, msg, video_id=None, *args, only_once=False, **kwargs):
1114 idstr = format_field(video_id, None, '%s: ')
1115 msg = f'[{self.IE_NAME}
] {idstr}{msg}
'
1117 if f'WARNING
: {msg}
' in self._printed_messages:
1119 self._printed_messages.add(f'WARNING
: {msg}
')
1120 self._downloader.report_warning(msg, *args, **kwargs)
1122 def to_screen(self, msg, *args, **kwargs):
1123 """Print msg to screen, prefixing it with '[ie_name
]'"""
1124 self._downloader.to_screen(f'[{self.IE_NAME}
] {msg}
', *args, **kwargs)
1126 def write_debug(self, msg, *args, **kwargs):
1127 self._downloader.write_debug(f'[{self.IE_NAME}
] {msg}
', *args, **kwargs)
1129 def get_param(self, name, default=None, *args, **kwargs):
1130 if self._downloader:
1131 return self._downloader.params.get(name, default, *args, **kwargs)
1134 def report_drm(self, video_id, partial=NO_DEFAULT):
1135 if partial is not NO_DEFAULT:
1136 self._downloader.deprecation_warning('InfoExtractor
.report_drm no longer accepts the argument partial
')
1137 self.raise_no_formats('This video
is DRM protected
', expected=True, video_id=video_id)
1139 def report_extraction(self, id_or_name):
1140 """Report information extraction."""
1141 self.to_screen('%s: Extracting information
' % id_or_name)
1143 def report_download_webpage(self, video_id):
1144 """Report webpage download."""
1145 self.to_screen('%s: Downloading webpage
' % video_id)
1147 def report_age_confirmation(self):
1148 """Report attempt to confirm age."""
1149 self.to_screen('Confirming age
')
1151 def report_login(self):
1152 """Report attempt to log in."""
1153 self.to_screen('Logging
in')
1155 def raise_login_required(
1156 self, msg='This video
is only available
for registered users
',
1157 metadata_available=False, method=NO_DEFAULT):
1158 if metadata_available and (
1159 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1160 self.report_warning(msg)
1162 msg += format_field(self._login_hint(method), None, '. %s')
1163 raise ExtractorError(msg, expected=True)
1165 def raise_geo_restricted(
1166 self, msg='This video
is not available
from your location due to geo restriction
',
1167 countries=None, metadata_available=False):
1168 if metadata_available and (
1169 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1170 self.report_warning(msg)
1172 raise GeoRestrictedError(msg, countries=countries)
1174 def raise_no_formats(self, msg, expected=False, video_id=None):
1176 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1177 self.report_warning(msg, video_id)
1178 elif isinstance(msg, ExtractorError):
1181 raise ExtractorError(msg, expected=expected, video_id=video_id)
1183 # Methods for following #608
1185 def url_result(url, ie=None, video_id=None, video_title=None, *, url_transparent=False, **kwargs):
1186 """Returns a URL that points to a page that should be processed"""
1188 kwargs['ie_key
'] = ie if isinstance(ie, str) else ie.ie_key()
1189 if video_id is not None:
1190 kwargs['id'] = video_id
1191 if video_title is not None:
1192 kwargs['title
'] = video_title
1195 '_type
': 'url_transparent
' if url_transparent else 'url
',
1200 def playlist_from_matches(cls, matches, playlist_id=None, playlist_title=None,
1201 getter=IDENTITY, ie=None, video_kwargs=None, **kwargs):
1202 return cls.playlist_result(
1203 (cls.url_result(m, ie, **(video_kwargs or {})) for m in orderedSet(map(getter, matches), lazy=True)),
1204 playlist_id, playlist_title, **kwargs)
1207 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None, *, multi_video=False, **kwargs):
1208 """Returns a playlist"""
1210 kwargs['id'] = playlist_id
1212 kwargs['title
'] = playlist_title
1213 if playlist_description is not None:
1214 kwargs['description
'] = playlist_description
1217 '_type
': 'multi_video
' if multi_video else 'playlist
',
1221 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1223 Perform a regex search on the given string, using a single or a list of
1224 patterns returning the first matching group.
1225 In case of failure return a default value or raise a WARNING or a
1226 RegexNotFoundError, depending on fatal, specifying the field name.
1230 elif isinstance(pattern, (str, re.Pattern)):
1231 mobj = re.search(pattern, string, flags)
1234 mobj = re.search(p, string, flags)
1238 _name = self._downloader._format_err(name, self._downloader.Styles.EMPHASIS)
1242 # return the first matching group
1243 return next(g for g in mobj.groups() if g is not None)
1244 elif isinstance(group, (list, tuple)):
1245 return tuple(mobj.group(g) for g in group)
1247 return mobj.group(group)
1248 elif default is not NO_DEFAULT:
1251 raise RegexNotFoundError('Unable to extract
%s' % _name)
1253 self.report_warning('unable to extract
%s' % _name + bug_reports_message())
1256 def _search_json(self, start_pattern, string, name, video_id, *, end_pattern='',
1257 contains_pattern=r'{(?s:.+)}
', fatal=True, default=NO_DEFAULT, **kwargs):
1258 """Searches string for the JSON object specified by start_pattern"""
1259 # NB: end_pattern is only used to reduce the size of the initial match
1260 if default is NO_DEFAULT:
1261 default, has_default = {}, False
1263 fatal, has_default = False, True
1265 json_string = self._search_regex(
1266 rf'(?
:{start_pattern}
)\s
*(?P
<json
>{contains_pattern}
)\s
*(?
:{end_pattern}
)',
1267 string, name, group='json
', fatal=fatal, default=None if has_default else NO_DEFAULT)
1271 _name = self._downloader._format_err(name, self._downloader.Styles.EMPHASIS)
1273 return self._parse_json(json_string, video_id, ignore_extra=True, **kwargs)
1274 except ExtractorError as e:
1276 raise ExtractorError(
1277 f'Unable to extract {_name}
- Failed to parse JSON
', cause=e.cause, video_id=video_id)
1278 elif not has_default:
1279 self.report_warning(
1280 f'Unable to extract {_name}
- Failed to parse JSON
: {e}
', video_id=video_id)
1283 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1285 Like _search_regex, but strips HTML tags and unescapes entities.
1287 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1288 if isinstance(res, tuple):
1289 return tuple(map(clean_html, res))
1290 return clean_html(res)
1292 def _get_netrc_login_info(self, netrc_machine=None):
1295 netrc_machine = netrc_machine or self._NETRC_MACHINE
1297 if self.get_param('usenetrc
', False):
1299 netrc_file = compat_expanduser(self.get_param('netrc_location
') or '~
')
1300 if os.path.isdir(netrc_file):
1301 netrc_file = os.path.join(netrc_file, '.netrc
')
1302 info = netrc.netrc(file=netrc_file).authenticators(netrc_machine)
1303 if info is not None:
1307 raise netrc.NetrcParseError(
1308 'No authenticators
for %s' % netrc_machine)
1309 except (OSError, netrc.NetrcParseError) as err:
1310 self.report_warning(
1311 'parsing
.netrc
: %s' % error_to_compat_str(err))
1313 return username, password
1315 def _get_login_info(self, username_option='username
', password_option='password
', netrc_machine=None):
1317 Get the login info as (username, password)
1318 First look for the manually specified credentials using username_option
1319 and password_option as keys in params dictionary. If no such credentials
1320 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1322 If there's no info available
, return (None, None)
1325 # Attempt to use provided username and password or .netrc data
1326 username = self.get_param(username_option)
1327 if username is not None:
1328 password = self.get_param(password_option)
1330 username, password = self._get_netrc_login_info(netrc_machine)
1332 return username, password
1334 def _get_tfa_info(self, note='two-factor verification code'):
1336 Get the two
-factor authentication info
1337 TODO
- asking the user will be required
for sms
/phone verify
1338 currently just uses the command line option
1339 If there
's no info available, return None
1342 tfa = self.get_param('twofactor
')
1346 return getpass.getpass('Type
%s and press
[Return
]: ' % note)
1348 # Helper functions for extracting OpenGraph info
1350 def _og_regexes(prop):
1351 content_re = r'content
=(?
:"([^"]+?
)"|\'([^\']+?)\'|\s*([^\s"\'=<>`
]+?
)(?
=\s|
/?
>))'
1352 property_re = (r'(?
:name|
property)=(?
:\'og
%(sep)s%(prop)s\'|
"og%(sep)s%(prop)s"|\s
*og
%(sep)s%(prop)s\b)'
1353 % {'prop': re.escape(prop), 'sep': '(?::|[:-])'})
1354 template = r'<meta
[^
>]+?
%s[^
>]+?
%s'
1356 template % (property_re, content_re),
1357 template % (content_re, property_re),
1361 def _meta_regex(prop):
1362 return r'''(?isx)<meta
1363 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1364 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1366 def _og_search_property(self, prop, html, name=None, **kargs):
1367 prop = variadic(prop)
1369 name = 'OpenGraph
%s' % prop[0]
1372 og_regexes.extend(self._og_regexes(p))
1373 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1376 return unescapeHTML(escaped)
1378 def _og_search_thumbnail(self, html, **kargs):
1379 return self._og_search_property('image
', html, 'thumbnail URL
', fatal=False, **kargs)
1381 def _og_search_description(self, html, **kargs):
1382 return self._og_search_property('description
', html, fatal=False, **kargs)
1384 def _og_search_title(self, html, *, fatal=False, **kargs):
1385 return self._og_search_property('title
', html, fatal=fatal, **kargs)
1387 def _og_search_video_url(self, html, name='video url
', secure=True, **kargs):
1388 regexes = self._og_regexes('video
') + self._og_regexes('video
:url
')
1390 regexes = self._og_regexes('video
:secure_url
') + regexes
1391 return self._html_search_regex(regexes, html, name, **kargs)
1393 def _og_search_url(self, html, **kargs):
1394 return self._og_search_property('url
', html, **kargs)
1396 def _html_extract_title(self, html, name='title
', *, fatal=False, **kwargs):
1397 return self._html_search_regex(r'(?s
)<title
\b[^
>]*>([^
<]+)</title
>', html, name, fatal=fatal, **kwargs)
1399 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1400 name = variadic(name)
1401 if display_name is None:
1402 display_name = name[0]
1403 return self._html_search_regex(
1404 [self._meta_regex(n) for n in name],
1405 html, display_name, fatal=fatal, group='content
', **kwargs)
1407 def _dc_search_uploader(self, html):
1408 return self._html_search_meta('dc
.creator
', html, 'uploader
')
1411 def _rta_search(html):
1412 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1413 if re.search(r'(?ix
)<meta\s
+name
="rating"\s
+'
1414 r' content
="RTA-5042-1996-1400-1577-RTA"',
1418 # And then there are the jokers who advertise that they use RTA, but actually don't
.
1419 AGE_LIMIT_MARKERS
= [
1420 r
'Proudly Labeled <a href="http://www\.rtalabel\.org/" title="Restricted to Adults">RTA</a>',
1421 r
'>[^<]*you acknowledge you are at least (\d+) years old',
1422 r
'>\s*(?:18\s+U(?:\.S\.C\.|SC)\s+)?(?:§+\s*)?2257\b',
1426 for marker
in AGE_LIMIT_MARKERS
:
1427 mobj
= re
.search(marker
, html
)
1429 age_limit
= max(age_limit
, int(traverse_obj(mobj
, 1, default
=18)))
1432 def _media_rating_search(self
, html
):
1433 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1434 rating
= self
._html
_search
_meta
('rating', html
)
1446 return RATING_TABLE
.get(rating
.lower())
1448 def _family_friendly_search(self
, html
):
1449 # See http://schema.org/VideoObject
1450 family_friendly
= self
._html
_search
_meta
(
1451 'isFamilyFriendly', html
, default
=None)
1453 if not family_friendly
:
1462 return RATING_TABLE
.get(family_friendly
.lower())
1464 def _twitter_search_player(self
, html
):
1465 return self
._html
_search
_meta
('twitter:player', html
,
1466 'twitter card player')
1468 def _yield_json_ld(self
, html
, video_id
, *, fatal
=True, default
=NO_DEFAULT
):
1469 """Yield all json ld objects in the html"""
1470 if default
is not NO_DEFAULT
:
1472 for mobj
in re
.finditer(JSON_LD_RE
, html
):
1473 json_ld_item
= self
._parse
_json
(mobj
.group('json_ld'), video_id
, fatal
=fatal
)
1474 for json_ld
in variadic(json_ld_item
):
1475 if isinstance(json_ld
, dict):
1478 def _search_json_ld(self
, html
, video_id
, expected_type
=None, *, fatal
=True, default
=NO_DEFAULT
):
1479 """Search for a video in any json ld in the html"""
1480 if default
is not NO_DEFAULT
:
1482 info
= self
._json
_ld
(
1483 list(self
._yield
_json
_ld
(html
, video_id
, fatal
=fatal
, default
=default
)),
1484 video_id
, fatal
=fatal
, expected_type
=expected_type
)
1487 if default
is not NO_DEFAULT
:
1490 raise RegexNotFoundError('Unable to extract JSON-LD')
1492 self
.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1495 def _json_ld(self
, json_ld
, video_id
, fatal
=True, expected_type
=None):
1496 if isinstance(json_ld
, str):
1497 json_ld
= self
._parse
_json
(json_ld
, video_id
, fatal
=fatal
)
1502 INTERACTION_TYPE_MAP
= {
1503 'CommentAction': 'comment',
1504 'AgreeAction': 'like',
1505 'DisagreeAction': 'dislike',
1506 'LikeAction': 'like',
1507 'DislikeAction': 'dislike',
1508 'ListenAction': 'view',
1509 'WatchAction': 'view',
1510 'ViewAction': 'view',
1513 def is_type(e
, *expected_types
):
1514 type = variadic(traverse_obj(e
, '@type'))
1515 return any(x
in type for x
in expected_types
)
1517 def extract_interaction_type(e
):
1518 interaction_type
= e
.get('interactionType')
1519 if isinstance(interaction_type
, dict):
1520 interaction_type
= interaction_type
.get('@type')
1521 return str_or_none(interaction_type
)
1523 def extract_interaction_statistic(e
):
1524 interaction_statistic
= e
.get('interactionStatistic')
1525 if isinstance(interaction_statistic
, dict):
1526 interaction_statistic
= [interaction_statistic
]
1527 if not isinstance(interaction_statistic
, list):
1529 for is_e
in interaction_statistic
:
1530 if not is_type(is_e
, 'InteractionCounter'):
1532 interaction_type
= extract_interaction_type(is_e
)
1533 if not interaction_type
:
1535 # For interaction count some sites provide string instead of
1536 # an integer (as per spec) with non digit characters (e.g. ",")
1537 # so extracting count with more relaxed str_to_int
1538 interaction_count
= str_to_int(is_e
.get('userInteractionCount'))
1539 if interaction_count
is None:
1541 count_kind
= INTERACTION_TYPE_MAP
.get(interaction_type
.split('/')[-1])
1544 count_key
= '%s_count' % count_kind
1545 if info
.get(count_key
) is not None:
1547 info
[count_key
] = interaction_count
1549 def extract_chapter_information(e
):
1551 'title': part
.get('name'),
1552 'start_time': part
.get('startOffset'),
1553 'end_time': part
.get('endOffset'),
1554 } for part
in variadic(e
.get('hasPart') or []) if part
.get('@type') == 'Clip']
1555 for idx
, (last_c
, current_c
, next_c
) in enumerate(zip(
1556 [{'end_time': 0}
] + chapters
, chapters
, chapters
[1:])):
1557 current_c
['end_time'] = current_c
['end_time'] or next_c
['start_time']
1558 current_c
['start_time'] = current_c
['start_time'] or last_c
['end_time']
1559 if None in current_c
.values():
1560 self
.report_warning(f
'Chapter {idx} contains broken data. Not extracting chapters')
1563 chapters
[-1]['end_time'] = chapters
[-1]['end_time'] or info
['duration']
1564 info
['chapters'] = chapters
1566 def extract_video_object(e
):
1567 author
= e
.get('author')
1569 'url': url_or_none(e
.get('contentUrl')),
1570 'ext': mimetype2ext(e
.get('encodingFormat')),
1571 'title': unescapeHTML(e
.get('name')),
1572 'description': unescapeHTML(e
.get('description')),
1573 'thumbnails': [{'url': unescapeHTML(url)}
1574 for url
in variadic(traverse_obj(e
, 'thumbnailUrl', 'thumbnailURL'))
1575 if url_or_none(url
)],
1576 'duration': parse_duration(e
.get('duration')),
1577 'timestamp': unified_timestamp(e
.get('uploadDate')),
1578 # author can be an instance of 'Organization' or 'Person' types.
1579 # both types can have 'name' property(inherited from 'Thing' type). [1]
1580 # however some websites are using 'Text' type instead.
1581 # 1. https://schema.org/VideoObject
1582 'uploader': author
.get('name') if isinstance(author
, dict) else author
if isinstance(author
, str) else None,
1583 'artist': traverse_obj(e
, ('byArtist', 'name'), expected_type
=str),
1584 'filesize': int_or_none(float_or_none(e
.get('contentSize'))),
1585 'tbr': int_or_none(e
.get('bitrate')),
1586 'width': int_or_none(e
.get('width')),
1587 'height': int_or_none(e
.get('height')),
1588 'view_count': int_or_none(e
.get('interactionCount')),
1589 'tags': try_call(lambda: e
.get('keywords').split(',')),
1591 if is_type(e
, 'AudioObject'):
1594 'abr': int_or_none(e
.get('bitrate')),
1596 extract_interaction_statistic(e
)
1597 extract_chapter_information(e
)
1599 def traverse_json_ld(json_ld
, at_top_level
=True):
1600 for e
in variadic(json_ld
):
1601 if not isinstance(e
, dict):
1603 if at_top_level
and '@context' not in e
:
1605 if at_top_level
and set(e
.keys()) == {'@context', '@graph'}
:
1606 traverse_json_ld(e
['@graph'], at_top_level
=False)
1608 if expected_type
is not None and not is_type(e
, expected_type
):
1610 rating
= traverse_obj(e
, ('aggregateRating', 'ratingValue'), expected_type
=float_or_none
)
1611 if rating
is not None:
1612 info
['average_rating'] = rating
1613 if is_type(e
, 'TVEpisode', 'Episode'):
1614 episode_name
= unescapeHTML(e
.get('name'))
1616 'episode': episode_name
,
1617 'episode_number': int_or_none(e
.get('episodeNumber')),
1618 'description': unescapeHTML(e
.get('description')),
1620 if not info
.get('title') and episode_name
:
1621 info
['title'] = episode_name
1622 part_of_season
= e
.get('partOfSeason')
1623 if is_type(part_of_season
, 'TVSeason', 'Season', 'CreativeWorkSeason'):
1625 'season': unescapeHTML(part_of_season
.get('name')),
1626 'season_number': int_or_none(part_of_season
.get('seasonNumber')),
1628 part_of_series
= e
.get('partOfSeries') or e
.get('partOfTVSeries')
1629 if is_type(part_of_series
, 'TVSeries', 'Series', 'CreativeWorkSeries'):
1630 info
['series'] = unescapeHTML(part_of_series
.get('name'))
1631 elif is_type(e
, 'Movie'):
1633 'title': unescapeHTML(e
.get('name')),
1634 'description': unescapeHTML(e
.get('description')),
1635 'duration': parse_duration(e
.get('duration')),
1636 'timestamp': unified_timestamp(e
.get('dateCreated')),
1638 elif is_type(e
, 'Article', 'NewsArticle'):
1640 'timestamp': parse_iso8601(e
.get('datePublished')),
1641 'title': unescapeHTML(e
.get('headline')),
1642 'description': unescapeHTML(e
.get('articleBody') or e
.get('description')),
1644 if is_type(traverse_obj(e
, ('video', 0)), 'VideoObject'):
1645 extract_video_object(e
['video'][0])
1646 elif is_type(traverse_obj(e
, ('subjectOf', 0)), 'VideoObject'):
1647 extract_video_object(e
['subjectOf'][0])
1648 elif is_type(e
, 'VideoObject', 'AudioObject'):
1649 extract_video_object(e
)
1650 if expected_type
is None:
1654 video
= e
.get('video')
1655 if is_type(video
, 'VideoObject'):
1656 extract_video_object(video
)
1657 if expected_type
is None:
1662 traverse_json_ld(json_ld
)
1663 return filter_dict(info
)
1665 def _search_nextjs_data(self
, webpage
, video_id
, *, transform_source
=None, fatal
=True, **kw
):
1666 return self
._parse
_json
(
1668 r
'(?s)<script[^>]+id=[\'"]__NEXT_DATA__[\'"][^
>]*>([^
<]+)</script
>',
1669 webpage, 'next
.js data
', fatal=fatal, **kw),
1670 video_id, transform_source=transform_source, fatal=fatal)
1672 def _search_nuxt_data(self, webpage, video_id, context_name='__NUXT__
', *, fatal=True, traverse=('data
', 0)):
1673 """Parses Nuxt.js metadata. This works as long as the function __NUXT__ invokes is a pure function"""
1674 rectx = re.escape(context_name)
1675 FUNCTION_RE = r'\
(function\
((?P
<arg_keys
>.*?
)\
){return\s+(?P<js>{.*?}
)\s
*;?\s
*}\
((?P
<arg_vals
>.*?
)\
)'
1676 js, arg_keys, arg_vals = self._search_regex(
1677 (rf'<script
>\s
*window\
.{rectx}
={FUNCTION_RE}\s
*\
)\s
*;?\s
*</script
>', rf'{rectx}\
(.*?{FUNCTION_RE}
'),
1678 webpage, context_name, group=('js
', 'arg_keys
', 'arg_vals
'),
1679 default=NO_DEFAULT if fatal else (None, None, None))
1683 args = dict(zip(arg_keys.split(','), map(json.dumps, self._parse_json(
1684 f'[{arg_vals}
]', video_id, transform_source=js_to_json, fatal=fatal) or ())))
1686 ret = self._parse_json(js, video_id, transform_source=functools.partial(js_to_json, vars=args), fatal=fatal)
1687 return traverse_obj(ret, traverse) or {}
1690 def _hidden_inputs(html):
1691 html = re.sub(r'<!--(?
:(?
!<!--).)*-->', '', html)
1693 for input in re.findall(r'(?i
)(<input[^
>]+>)', html):
1694 attrs = extract_attributes(input)
1697 if attrs.get('type') not in ('hidden
', 'submit
'):
1699 name = attrs.get('name
') or attrs.get('id')
1700 value = attrs.get('value
')
1701 if name and value is not None:
1702 hidden_inputs[name] = value
1703 return hidden_inputs
1705 def _form_hidden_inputs(self, form_id, html):
1706 form = self._search_regex(
1707 r'(?
is)<form
[^
>]+?
id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1708 html, '%s form' % form_id, group='form')
1709 return self._hidden_inputs(form)
1711 @classproperty(cache=True)
1712 def FormatSort(cls):
1713 class FormatSort(FormatSorter):
1714 def __init__(ie, *args, **kwargs):
1715 super().__init__(ie._downloader, *args, **kwargs)
1717 deprecation_warning(
1718 'yt_dlp.InfoExtractor.FormatSort is deprecated and may be removed in the future. '
1719 'Use yt_dlp.utils.FormatSorter instead')
1722 def _sort_formats(self, formats, field_preference=[]):
1723 if not field_preference:
1724 self._downloader.deprecation_warning(
1725 'yt_dlp.InfoExtractor._sort_formats is deprecated and is no longer required')
1727 self._downloader.deprecation_warning(
1728 'yt_dlp.InfoExtractor._sort_formats is deprecated and no longer works as expected. '
1729 'Return _format_sort_fields in the info_dict instead')
1731 formats[0]['__sort_fields'] = field_preference
1733 def _check_formats(self, formats, video_id):
1735 formats[:] = filter(
1736 lambda f: self._is_valid_url(
1738 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1742 def _remove_duplicate_formats(formats):
1746 if f['url'] not in format_urls:
1747 format_urls.add(f['url'])
1748 unique_formats.append(f)
1749 formats[:] = unique_formats
1751 def _is_valid_url(self, url, video_id, item='video', headers={}):
1752 url = self._proto_relative_url(url, scheme='http:')
1753 # For now assume non HTTP(S) URLs always valid
1754 if not (url.startswith('http://') or url.startswith('https://')):
1757 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1759 except ExtractorError as e:
1761 '%s: %s URL is invalid, skipping: %s'
1762 % (video_id, item, error_to_compat_str(e.cause)))
1765 def http_scheme(self):
1766 """ Either "http
:" or "https
:", depending on the user's preferences """
1769 if self.get_param('prefer_insecure', False)
1772 def _proto_relative_url(self, url, scheme=None):
1773 scheme = scheme or self.http_scheme()
1774 assert scheme.endswith(':')
1775 return sanitize_url(url, scheme=scheme[:-1])
1777 def _sleep(self, timeout, video_id, msg_template=None):
1778 if msg_template is None:
1779 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1780 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1784 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1785 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1786 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
1787 if self.get_param('ignore_no_formats_error'):
1790 res = self._download_xml_handle(
1791 manifest_url, video_id, 'Downloading f4m manifest',
1792 'Unable to download f4m manifest',
1793 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1794 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
1795 transform_source=transform_source,
1796 fatal=fatal, data=data, headers=headers, query=query)
1800 manifest, urlh = res
1801 manifest_url = urlh.geturl()
1803 return self._parse_f4m_formats(
1804 manifest, manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1805 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1807 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1808 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1809 fatal=True, m3u8_id=None):
1810 if not isinstance(manifest, xml.etree.ElementTree.Element) and not fatal:
1813 # currently yt-dlp cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1814 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1815 if akamai_pv is not None and ';' in akamai_pv.text:
1816 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1817 if playerVerificationChallenge.strip() != '':
1821 manifest_version = '1.0'
1822 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1824 manifest_version = '2.0'
1825 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1826 # Remove unsupported DRM protected media from final formats
1827 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
1828 media_nodes = remove_encrypted_media(media_nodes)
1832 manifest_base_url = get_base_url(manifest)
1834 bootstrap_info = xpath_element(
1835 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1836 'bootstrap info', default=None)
1839 mime_type = xpath_text(
1840 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1841 'base URL', default=None)
1842 if mime_type and mime_type.startswith('audio/'):
1845 for i, media_el in enumerate(media_nodes):
1846 tbr = int_or_none(media_el.attrib.get('bitrate'))
1847 width = int_or_none(media_el.attrib.get('width'))
1848 height = int_or_none(media_el.attrib.get('height'))
1849 format_id = join_nonempty(f4m_id, tbr or i)
1850 # If <bootstrapInfo> is present, the specified f4m is a
1851 # stream-level manifest, and only set-level manifests may refer to
1852 # external resources. See section 11.4 and section 4 of F4M spec
1853 if bootstrap_info is None:
1855 # @href is introduced in 2.0, see section 11.6 of F4M spec
1856 if manifest_version == '2.0':
1857 media_url = media_el.attrib.get('href')
1858 if media_url is None:
1859 media_url = media_el.attrib.get('url')
1863 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1864 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1865 # If media_url is itself a f4m manifest do the recursive extraction
1866 # since bitrates in parent manifest (this one) and media_url manifest
1867 # may differ leading to inability to resolve the format by requested
1868 # bitrate in f4m downloader
1869 ext = determine_ext(manifest_url)
1871 f4m_formats = self._extract_f4m_formats(
1872 manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1873 transform_source=transform_source, fatal=fatal)
1874 # Sometimes stream-level manifest contains single media entry that
1875 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1876 # At the same time parent's media entry in set-level manifest may
1877 # contain it. We will copy it from parent in such cases.
1878 if len(f4m_formats) == 1:
1881 'tbr': f.get('tbr') or tbr,
1882 'width': f.get('width') or width,
1883 'height': f.get('height') or height,
1884 'format_id': f.get('format_id') if not tbr else format_id,
1887 formats.extend(f4m_formats)
1890 formats.extend(self._extract_m3u8_formats(
1891 manifest_url, video_id, 'mp4', preference=preference,
1892 quality=quality, m3u8_id=m3u8_id, fatal=fatal))
1895 'format_id': format_id,
1896 'url': manifest_url,
1897 'manifest_url': manifest_url,
1898 'ext': 'flv' if bootstrap_info is not None else None,
1904 'preference': preference,
1909 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, quality=None, m3u8_id=None):
1911 'format_id': join_nonempty(m3u8_id, 'meta'),
1915 'preference': preference - 100 if preference else -100,
1917 'resolution': 'multiple',
1918 'format_note': 'Quality selection URL',
1921 def _report_ignoring_subs(self, name):
1922 self.report_warning(bug_reports_message(
1923 f'Ignoring subtitle tracks found in the {name} manifest; '
1924 'if any subtitle tracks are missing,'
1927 def _extract_m3u8_formats(self, *args, **kwargs):
1928 fmts, subs = self._extract_m3u8_formats_and_subtitles(*args, **kwargs)
1930 self._report_ignoring_subs('HLS')
1933 def _extract_m3u8_formats_and_subtitles(
1934 self, m3u8_url, video_id, ext=None, entry_protocol='m3u8_native',
1935 preference=None, quality=None, m3u8_id=None, note=None,
1936 errnote=None, fatal=True, live=False, data=None, headers={},
1939 if self.get_param('ignore_no_formats_error'):
1943 if errnote is not False:
1944 errnote = errnote or 'Failed to obtain m3u8 URL'
1946 raise ExtractorError(errnote, video_id=video_id)
1947 self.report_warning(f'{errnote}{bug_reports_message()}')
1950 res = self._download_webpage_handle(
1952 note='Downloading m3u8 information' if note is None else note,
1953 errnote='Failed to download m3u8 information' if errnote is None else errnote,
1954 fatal=fatal, data=data, headers=headers, query=query)
1959 m3u8_doc, urlh = res
1960 m3u8_url = urlh.geturl()
1962 return self._parse_m3u8_formats_and_subtitles(
1963 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
1964 preference=preference, quality=quality, m3u8_id=m3u8_id,
1965 note=note, errnote=errnote, fatal=fatal, live=live, data=data,
1966 headers=headers, query=query, video_id=video_id)
1968 def _parse_m3u8_formats_and_subtitles(
1969 self, m3u8_doc, m3u8_url=None, ext=None, entry_protocol='m3u8_native',
1970 preference=None, quality=None, m3u8_id=None, live=False, note=None,
1971 errnote=None, fatal=True, data=None, headers={}, query={},
1973 formats, subtitles = [], {}
1975 has_drm = re.search('|'.join([
1976 r'#EXT-X-FAXS-CM:', # Adobe Flash Access
1977 r'#EXT-X-(?:SESSION-)?KEY:.*?URI="skd
://', # Apple FairPlay
1980 def format_url(url):
1981 return url if re.match(r'^https?
://', url) else urllib.parse.urljoin(m3u8_url, url)
1983 if self.get_param('hls_split_discontinuity
', False):
1984 def _extract_m3u8_playlist_indices(manifest_url=None, m3u8_doc=None):
1986 if not manifest_url:
1988 m3u8_doc = self._download_webpage(
1989 manifest_url, video_id, fatal=fatal, data=data, headers=headers,
1990 note=False, errnote='Failed to download m3u8 playlist information
')
1991 if m3u8_doc is False:
1993 return range(1 + sum(line.startswith('#EXT-X-DISCONTINUITY') for line in m3u8_doc.splitlines()))
1996 def _extract_m3u8_playlist_indices(*args
, **kwargs
):
2000 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
2001 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
2002 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
2004 # We should try extracting formats only from master playlists [1, 4.3.4],
2005 # i.e. playlists that describe available qualities. On the other hand
2006 # media playlists [1, 4.3.3] should be returned as is since they contain
2007 # just the media without qualities renditions.
2008 # Fortunately, master playlist can be easily distinguished from media
2009 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
2010 # master playlist tags MUST NOT appear in a media playlist and vice versa.
2011 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
2012 # media playlist and MUST NOT appear in master playlist thus we can
2013 # clearly detect media playlist with this criterion.
2015 if '#EXT-X-TARGETDURATION' in m3u8_doc
: # media playlist, return as is
2017 'format_id': join_nonempty(m3u8_id
, idx
),
2018 'format_index': idx
,
2019 'url': m3u8_url
or encode_data_uri(m3u8_doc
.encode('utf-8'), 'application/x-mpegurl'),
2021 'protocol': entry_protocol
,
2022 'preference': preference
,
2025 } for idx
in _extract_m3u8_playlist_indices(m3u8_doc
=m3u8_doc
)]
2027 return formats
, subtitles
2030 last_stream_inf
= {}
2032 def extract_media(x_media_line
):
2033 media
= parse_m3u8_attributes(x_media_line
)
2034 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
2035 media_type
, group_id
, name
= media
.get('TYPE'), media
.get('GROUP-ID'), media
.get('NAME')
2036 if not (media_type
and group_id
and name
):
2038 groups
.setdefault(group_id
, []).append(media
)
2039 # <https://tools.ietf.org/html/rfc8216#section-4.3.4.1>
2040 if media_type
== 'SUBTITLES':
2041 # According to RFC 8216 §4.3.4.2.1, URI is REQUIRED in the
2042 # EXT-X-MEDIA tag if the media type is SUBTITLES.
2043 # However, lack of URI has been spotted in the wild.
2044 # e.g. NebulaIE; see https://github.com/yt-dlp/yt-dlp/issues/339
2045 if not media
.get('URI'):
2047 url
= format_url(media
['URI'])
2050 'ext': determine_ext(url
),
2052 if sub_info
['ext'] == 'm3u8':
2053 # Per RFC 8216 §3.1, the only possible subtitle format m3u8
2054 # files may contain is WebVTT:
2055 # <https://tools.ietf.org/html/rfc8216#section-3.1>
2056 sub_info
['ext'] = 'vtt'
2057 sub_info
['protocol'] = 'm3u8_native'
2058 lang
= media
.get('LANGUAGE') or 'und'
2059 subtitles
.setdefault(lang
, []).append(sub_info
)
2060 if media_type
not in ('VIDEO', 'AUDIO'):
2062 media_url
= media
.get('URI')
2064 manifest_url
= format_url(media_url
)
2066 'format_id': join_nonempty(m3u8_id
, group_id
, name
, idx
),
2067 'format_note': name
,
2068 'format_index': idx
,
2069 'url': manifest_url
,
2070 'manifest_url': m3u8_url
,
2071 'language': media
.get('LANGUAGE'),
2073 'protocol': entry_protocol
,
2074 'preference': preference
,
2077 'vcodec': 'none' if media_type
== 'AUDIO' else None,
2078 } for idx
in _extract_m3u8_playlist_indices(manifest_url
))
2080 def build_stream_name():
2081 # Despite specification does not mention NAME attribute for
2082 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
2083 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
2084 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
2085 stream_name
= last_stream_inf
.get('NAME')
2088 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
2089 # from corresponding rendition group
2090 stream_group_id
= last_stream_inf
.get('VIDEO')
2091 if not stream_group_id
:
2093 stream_group
= groups
.get(stream_group_id
)
2094 if not stream_group
:
2095 return stream_group_id
2096 rendition
= stream_group
[0]
2097 return rendition
.get('NAME') or stream_group_id
2099 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
2100 # chance to detect video only formats when EXT-X-STREAM-INF tags
2101 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
2102 for line
in m3u8_doc
.splitlines():
2103 if line
.startswith('#EXT-X-MEDIA:'):
2106 for line
in m3u8_doc
.splitlines():
2107 if line
.startswith('#EXT-X-STREAM-INF:'):
2108 last_stream_inf
= parse_m3u8_attributes(line
)
2109 elif line
.startswith('#') or not line
.strip():
2112 tbr
= float_or_none(
2113 last_stream_inf
.get('AVERAGE-BANDWIDTH')
2114 or last_stream_inf
.get('BANDWIDTH'), scale
=1000)
2115 manifest_url
= format_url(line
.strip())
2117 for idx
in _extract_m3u8_playlist_indices(manifest_url
):
2118 format_id
= [m3u8_id
, None, idx
]
2119 # Bandwidth of live streams may differ over time thus making
2120 # format_id unpredictable. So it's better to keep provided
2123 stream_name
= build_stream_name()
2124 format_id
[1] = stream_name
or '%d' % (tbr
or len(formats
))
2126 'format_id': join_nonempty(*format_id
),
2127 'format_index': idx
,
2128 'url': manifest_url
,
2129 'manifest_url': m3u8_url
,
2132 'fps': float_or_none(last_stream_inf
.get('FRAME-RATE')),
2133 'protocol': entry_protocol
,
2134 'preference': preference
,
2138 resolution
= last_stream_inf
.get('RESOLUTION')
2140 mobj
= re
.search(r
'(?P<width>\d+)[xX](?P<height>\d+)', resolution
)
2142 f
['width'] = int(mobj
.group('width'))
2143 f
['height'] = int(mobj
.group('height'))
2144 # Unified Streaming Platform
2146 r
'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f
['url'])
2148 abr
, vbr
= mobj
.groups()
2149 abr
, vbr
= float_or_none(abr
, 1000), float_or_none(vbr
, 1000)
2154 codecs
= parse_codecs(last_stream_inf
.get('CODECS'))
2156 audio_group_id
= last_stream_inf
.get('AUDIO')
2157 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
2158 # references a rendition group MUST have a CODECS attribute.
2159 # However, this is not always respected. E.g. [2]
2160 # contains EXT-X-STREAM-INF tag which references AUDIO
2161 # rendition group but does not have CODECS and despite
2162 # referencing an audio group it represents a complete
2163 # (with audio and video) format. So, for such cases we will
2164 # ignore references to rendition groups and treat them
2165 # as complete formats.
2166 if audio_group_id
and codecs
and f
.get('vcodec') != 'none':
2167 audio_group
= groups
.get(audio_group_id
)
2168 if audio_group
and audio_group
[0].get('URI'):
2169 # TODO: update acodec for audio only formats with
2171 f
['acodec'] = 'none'
2172 if not f
.get('ext'):
2173 f
['ext'] = 'm4a' if f
.get('vcodec') == 'none' else 'mp4'
2177 progressive_uri
= last_stream_inf
.get('PROGRESSIVE-URI')
2180 del http_f
['manifest_url']
2182 'format_id': f
['format_id'].replace('hls-', 'http-'),
2184 'url': progressive_uri
,
2186 formats
.append(http_f
)
2188 last_stream_inf
= {}
2189 return formats
, subtitles
2191 def _extract_m3u8_vod_duration(
2192 self
, m3u8_vod_url
, video_id
, note
=None, errnote
=None, data
=None, headers
={}, query={}
):
2194 m3u8_vod
= self
._download
_webpage
(
2195 m3u8_vod_url
, video_id
,
2196 note
='Downloading m3u8 VOD manifest' if note
is None else note
,
2197 errnote
='Failed to download VOD manifest' if errnote
is None else errnote
,
2198 fatal
=False, data
=data
, headers
=headers
, query
=query
)
2200 return self
._parse
_m
3u8_vod
_duration
(m3u8_vod
or '', video_id
)
2202 def _parse_m3u8_vod_duration(self
, m3u8_vod
, video_id
):
2203 if '#EXT-X-ENDLIST' not in m3u8_vod
:
2207 float(line
[len('#EXTINF:'):].split(',')[0])
2208 for line
in m3u8_vod
.splitlines() if line
.startswith('#EXTINF:'))) or None
2210 def _extract_mpd_vod_duration(
2211 self
, mpd_url
, video_id
, note
=None, errnote
=None, data
=None, headers
={}, query={}
):
2213 mpd_doc
= self
._download
_xml
(
2215 note
='Downloading MPD VOD manifest' if note
is None else note
,
2216 errnote
='Failed to download VOD manifest' if errnote
is None else errnote
,
2217 fatal
=False, data
=data
, headers
=headers
, query
=query
) or {}
2218 return int_or_none(parse_duration(mpd_doc
.get('mediaPresentationDuration')))
2221 def _xpath_ns(path
, namespace
=None):
2225 for c
in path
.split('/'):
2226 if not c
or c
== '.':
2229 out
.append('{%s}%s' % (namespace
, c
))
2230 return '/'.join(out
)
2232 def _extract_smil_formats_and_subtitles(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None, transform_source
=None):
2233 if self
.get_param('ignore_no_formats_error'):
2236 res
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
, transform_source
=transform_source
)
2242 smil_url
= urlh
.geturl()
2244 namespace
= self
._parse
_smil
_namespace
(smil
)
2246 fmts
= self
._parse
_smil
_formats
(
2247 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2248 subs
= self
._parse
_smil
_subtitles
(
2249 smil
, namespace
=namespace
)
2253 def _extract_smil_formats(self
, *args
, **kwargs
):
2254 fmts
, subs
= self
._extract
_smil
_formats
_and
_subtitles
(*args
, **kwargs
)
2256 self
._report
_ignoring
_subs
('SMIL')
2259 def _extract_smil_info(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None):
2260 res
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
)
2265 smil_url
= urlh
.geturl()
2267 return self
._parse
_smil
(smil
, smil_url
, video_id
, f4m_params
=f4m_params
)
2269 def _download_smil(self
, smil_url
, video_id
, fatal
=True, transform_source
=None):
2270 return self
._download
_xml
_handle
(
2271 smil_url
, video_id
, 'Downloading SMIL file',
2272 'Unable to download SMIL file', fatal
=fatal
, transform_source
=transform_source
)
2274 def _parse_smil(self
, smil
, smil_url
, video_id
, f4m_params
=None):
2275 namespace
= self
._parse
_smil
_namespace
(smil
)
2277 formats
= self
._parse
_smil
_formats
(
2278 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2279 subtitles
= self
._parse
_smil
_subtitles
(smil
, namespace
=namespace
)
2281 video_id
= os
.path
.splitext(url_basename(smil_url
))[0]
2285 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2286 name
= meta
.attrib
.get('name')
2287 content
= meta
.attrib
.get('content')
2288 if not name
or not content
:
2290 if not title
and name
== 'title':
2292 elif not description
and name
in ('description', 'abstract'):
2293 description
= content
2294 elif not upload_date
and name
== 'date':
2295 upload_date
= unified_strdate(content
)
2298 'id': image
.get('type'),
2299 'url': image
.get('src'),
2300 'width': int_or_none(image
.get('width')),
2301 'height': int_or_none(image
.get('height')),
2302 } for image
in smil
.findall(self
._xpath
_ns
('.//image', namespace
)) if image
.get('src')]
2306 'title': title
or video_id
,
2307 'description': description
,
2308 'upload_date': upload_date
,
2309 'thumbnails': thumbnails
,
2311 'subtitles': subtitles
,
2314 def _parse_smil_namespace(self
, smil
):
2315 return self
._search
_regex
(
2316 r
'(?i)^{([^}]+)?}smil$', smil
.tag
, 'namespace', default
=None)
2318 def _parse_smil_formats(self
, smil
, smil_url
, video_id
, namespace
=None, f4m_params
=None, transform_rtmp_url
=None):
2320 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2321 b
= meta
.get('base') or meta
.get('httpBase')
2333 media
= smil
.findall(self
._xpath
_ns
('.//video', namespace
)) + smil
.findall(self
._xpath
_ns
('.//audio', namespace
))
2334 for medium
in media
:
2335 src
= medium
.get('src')
2336 if not src
or src
in srcs
:
2340 bitrate
= float_or_none(medium
.get('system-bitrate') or medium
.get('systemBitrate'), 1000)
2341 filesize
= int_or_none(medium
.get('size') or medium
.get('fileSize'))
2342 width
= int_or_none(medium
.get('width'))
2343 height
= int_or_none(medium
.get('height'))
2344 proto
= medium
.get('proto')
2345 ext
= medium
.get('ext')
2346 src_ext
= determine_ext(src
, default_ext
=None) or ext
or urlhandle_detect_ext(
2347 self
._request
_webpage
(HEADRequest(src
), video_id
, note
='Requesting extension info', fatal
=False))
2348 streamer
= medium
.get('streamer') or base
2350 if proto
== 'rtmp' or streamer
.startswith('rtmp'):
2356 'format_id': 'rtmp-%d' % (rtmp_count
if bitrate
is None else bitrate
),
2358 'filesize': filesize
,
2362 if transform_rtmp_url
:
2363 streamer
, src
= transform_rtmp_url(streamer
, src
)
2364 formats
[-1].update({
2370 src_url
= src
if src
.startswith('http') else urllib
.parse
.urljoin(base
, src
)
2371 src_url
= src_url
.strip()
2373 if proto
== 'm3u8' or src_ext
== 'm3u8':
2374 m3u8_formats
= self
._extract
_m
3u8_formats
(
2375 src_url
, video_id
, ext
or 'mp4', m3u8_id
='hls', fatal
=False)
2376 if len(m3u8_formats
) == 1:
2378 m3u8_formats
[0].update({
2379 'format_id': 'hls-%d' % (m3u8_count
if bitrate
is None else bitrate
),
2384 formats
.extend(m3u8_formats
)
2385 elif src_ext
== 'f4m':
2390 'plugin': 'flowplayer-3.2.0.1',
2392 f4m_url
+= '&' if '?' in f4m_url
else '?'
2393 f4m_url
+= urllib
.parse
.urlencode(f4m_params
)
2394 formats
.extend(self
._extract
_f
4m
_formats
(f4m_url
, video_id
, f4m_id
='hds', fatal
=False))
2395 elif src_ext
== 'mpd':
2396 formats
.extend(self
._extract
_mpd
_formats
(
2397 src_url
, video_id
, mpd_id
='dash', fatal
=False))
2398 elif re
.search(r
'\.ism/[Mm]anifest', src_url
):
2399 formats
.extend(self
._extract
_ism
_formats
(
2400 src_url
, video_id
, ism_id
='mss', fatal
=False))
2401 elif src_url
.startswith('http') and self
._is
_valid
_url
(src
, video_id
):
2405 'ext': ext
or src_ext
or 'flv',
2406 'format_id': 'http-%d' % (bitrate
or http_count
),
2408 'filesize': filesize
,
2413 for medium
in smil
.findall(self
._xpath
_ns
('.//imagestream', namespace
)):
2414 src
= medium
.get('src')
2415 if not src
or src
in srcs
:
2421 'format_id': 'imagestream-%d' % (imgs_count
),
2423 'ext': mimetype2ext(medium
.get('type')),
2426 'width': int_or_none(medium
.get('width')),
2427 'height': int_or_none(medium
.get('height')),
2428 'format_note': 'SMIL storyboards',
2433 def _parse_smil_subtitles(self
, smil
, namespace
=None, subtitles_lang
='en'):
2436 for num
, textstream
in enumerate(smil
.findall(self
._xpath
_ns
('.//textstream', namespace
))):
2437 src
= textstream
.get('src')
2438 if not src
or src
in urls
:
2441 ext
= textstream
.get('ext') or mimetype2ext(textstream
.get('type')) or determine_ext(src
)
2442 lang
= textstream
.get('systemLanguage') or textstream
.get('systemLanguageName') or textstream
.get('lang') or subtitles_lang
2443 subtitles
.setdefault(lang
, []).append({
2449 def _extract_xspf_playlist(self
, xspf_url
, playlist_id
, fatal
=True):
2450 res
= self
._download
_xml
_handle
(
2451 xspf_url
, playlist_id
, 'Downloading xpsf playlist',
2452 'Unable to download xspf manifest', fatal
=fatal
)
2457 xspf_url
= urlh
.geturl()
2459 return self
._parse
_xspf
(
2460 xspf
, playlist_id
, xspf_url
=xspf_url
,
2461 xspf_base_url
=base_url(xspf_url
))
2463 def _parse_xspf(self
, xspf_doc
, playlist_id
, xspf_url
=None, xspf_base_url
=None):
2465 'xspf': 'http://xspf.org/ns/0/',
2466 's1': 'http://static.streamone.nl/player/ns/0',
2470 for track
in xspf_doc
.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP
)):
2472 track
, xpath_with_ns('./xspf:title', NS_MAP
), 'title', default
=playlist_id
)
2473 description
= xpath_text(
2474 track
, xpath_with_ns('./xspf:annotation', NS_MAP
), 'description')
2475 thumbnail
= xpath_text(
2476 track
, xpath_with_ns('./xspf:image', NS_MAP
), 'thumbnail')
2477 duration
= float_or_none(
2478 xpath_text(track
, xpath_with_ns('./xspf:duration', NS_MAP
), 'duration'), 1000)
2481 for location
in track
.findall(xpath_with_ns('./xspf:location', NS_MAP
)):
2482 format_url
= urljoin(xspf_base_url
, location
.text
)
2487 'manifest_url': xspf_url
,
2488 'format_id': location
.get(xpath_with_ns('s1:label', NS_MAP
)),
2489 'width': int_or_none(location
.get(xpath_with_ns('s1:width', NS_MAP
))),
2490 'height': int_or_none(location
.get(xpath_with_ns('s1:height', NS_MAP
))),
2496 'description': description
,
2497 'thumbnail': thumbnail
,
2498 'duration': duration
,
2503 def _extract_mpd_formats(self
, *args
, **kwargs
):
2504 fmts
, subs
= self
._extract
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2506 self
._report
_ignoring
_subs
('DASH')
2509 def _extract_mpd_formats_and_subtitles(
2510 self
, mpd_url
, video_id
, mpd_id
=None, note
=None, errnote
=None,
2511 fatal
=True, data
=None, headers
={}, query={}
):
2513 if self
.get_param('ignore_no_formats_error'):
2516 res
= self
._download
_xml
_handle
(
2518 note
='Downloading MPD manifest' if note
is None else note
,
2519 errnote
='Failed to download MPD manifest' if errnote
is None else errnote
,
2520 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
2527 # We could have been redirected to a new url when we retrieved our mpd file.
2528 mpd_url
= urlh
.geturl()
2529 mpd_base_url
= base_url(mpd_url
)
2531 return self
._parse
_mpd
_formats
_and
_subtitles
(
2532 mpd_doc
, mpd_id
, mpd_base_url
, mpd_url
)
2534 def _parse_mpd_formats(self
, *args
, **kwargs
):
2535 fmts
, subs
= self
._parse
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2537 self
._report
_ignoring
_subs
('DASH')
2540 def _parse_mpd_formats_and_subtitles(
2541 self
, mpd_doc
, mpd_id
=None, mpd_base_url
='', mpd_url
=None):
2543 Parse formats from MPD manifest.
2545 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
2546 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
2547 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
2549 if not self
.get_param('dynamic_mpd', True):
2550 if mpd_doc
.get('type') == 'dynamic':
2553 namespace
= self
._search
_regex
(r
'(?i)^{([^}]+)?}MPD$', mpd_doc
.tag
, 'namespace', default
=None)
2556 return self
._xpath
_ns
(path
, namespace
)
2558 def is_drm_protected(element
):
2559 return element
.find(_add_ns('ContentProtection')) is not None
2561 def extract_multisegment_info(element
, ms_parent_info
):
2562 ms_info
= ms_parent_info
.copy()
2564 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2565 # common attributes and elements. We will only extract relevant
2567 def extract_common(source
):
2568 segment_timeline
= source
.find(_add_ns('SegmentTimeline'))
2569 if segment_timeline
is not None:
2570 s_e
= segment_timeline
.findall(_add_ns('S'))
2572 ms_info
['total_number'] = 0
2575 r
= int(s
.get('r', 0))
2576 ms_info
['total_number'] += 1 + r
2577 ms_info
['s'].append({
2578 't': int(s
.get('t', 0)),
2579 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2580 'd': int(s
.attrib
['d']),
2583 start_number
= source
.get('startNumber')
2585 ms_info
['start_number'] = int(start_number
)
2586 timescale
= source
.get('timescale')
2588 ms_info
['timescale'] = int(timescale
)
2589 segment_duration
= source
.get('duration')
2590 if segment_duration
:
2591 ms_info
['segment_duration'] = float(segment_duration
)
2593 def extract_Initialization(source
):
2594 initialization
= source
.find(_add_ns('Initialization'))
2595 if initialization
is not None:
2596 ms_info
['initialization_url'] = initialization
.attrib
['sourceURL']
2598 segment_list
= element
.find(_add_ns('SegmentList'))
2599 if segment_list
is not None:
2600 extract_common(segment_list
)
2601 extract_Initialization(segment_list
)
2602 segment_urls_e
= segment_list
.findall(_add_ns('SegmentURL'))
2604 ms_info
['segment_urls'] = [segment
.attrib
['media'] for segment
in segment_urls_e
]
2606 segment_template
= element
.find(_add_ns('SegmentTemplate'))
2607 if segment_template
is not None:
2608 extract_common(segment_template
)
2609 media
= segment_template
.get('media')
2611 ms_info
['media'] = media
2612 initialization
= segment_template
.get('initialization')
2614 ms_info
['initialization'] = initialization
2616 extract_Initialization(segment_template
)
2619 mpd_duration
= parse_duration(mpd_doc
.get('mediaPresentationDuration'))
2620 formats
, subtitles
= [], {}
2621 stream_numbers
= collections
.defaultdict(int)
2622 for period
in mpd_doc
.findall(_add_ns('Period')):
2623 period_duration
= parse_duration(period
.get('duration')) or mpd_duration
2624 period_ms_info
= extract_multisegment_info(period
, {
2628 for adaptation_set
in period
.findall(_add_ns('AdaptationSet')):
2629 adaption_set_ms_info
= extract_multisegment_info(adaptation_set
, period_ms_info
)
2630 for representation
in adaptation_set
.findall(_add_ns('Representation')):
2631 representation_attrib
= adaptation_set
.attrib
.copy()
2632 representation_attrib
.update(representation
.attrib
)
2633 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2634 mime_type
= representation_attrib
['mimeType']
2635 content_type
= representation_attrib
.get('contentType', mime_type
.split('/')[0])
2637 codec_str
= representation_attrib
.get('codecs', '')
2638 # Some kind of binary subtitle found in some youtube livestreams
2639 if mime_type
== 'application/x-rawcc':
2640 codecs
= {'scodec': codec_str}
2642 codecs
= parse_codecs(codec_str
)
2643 if content_type
not in ('video', 'audio', 'text'):
2644 if mime_type
== 'image/jpeg':
2645 content_type
= mime_type
2646 elif codecs
.get('vcodec', 'none') != 'none':
2647 content_type
= 'video'
2648 elif codecs
.get('acodec', 'none') != 'none':
2649 content_type
= 'audio'
2650 elif codecs
.get('scodec', 'none') != 'none':
2651 content_type
= 'text'
2652 elif mimetype2ext(mime_type
) in ('tt', 'dfxp', 'ttml', 'xml', 'json'):
2653 content_type
= 'text'
2655 self
.report_warning('Unknown MIME type %s in DASH manifest' % mime_type
)
2659 for element
in (representation
, adaptation_set
, period
, mpd_doc
):
2660 base_url_e
= element
.find(_add_ns('BaseURL'))
2661 if try_call(lambda: base_url_e
.text
) is not None:
2662 base_url
= base_url_e
.text
+ base_url
2663 if re
.match(r
'^https?://', base_url
):
2665 if mpd_base_url
and base_url
.startswith('/'):
2666 base_url
= urllib
.parse
.urljoin(mpd_base_url
, base_url
)
2667 elif mpd_base_url
and not re
.match(r
'^https?://', base_url
):
2668 if not mpd_base_url
.endswith('/'):
2670 base_url
= mpd_base_url
+ base_url
2671 representation_id
= representation_attrib
.get('id')
2672 lang
= representation_attrib
.get('lang')
2673 url_el
= representation
.find(_add_ns('BaseURL'))
2674 filesize
= int_or_none(url_el
.attrib
.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el
is not None else None)
2675 bandwidth
= int_or_none(representation_attrib
.get('bandwidth'))
2676 if representation_id
is not None:
2677 format_id
= representation_id
2679 format_id
= content_type
2681 format_id
= mpd_id
+ '-' + format_id
2682 if content_type
in ('video', 'audio'):
2684 'format_id': format_id
,
2685 'manifest_url': mpd_url
,
2686 'ext': mimetype2ext(mime_type
),
2687 'width': int_or_none(representation_attrib
.get('width')),
2688 'height': int_or_none(representation_attrib
.get('height')),
2689 'tbr': float_or_none(bandwidth
, 1000),
2690 'asr': int_or_none(representation_attrib
.get('audioSamplingRate')),
2691 'fps': int_or_none(representation_attrib
.get('frameRate')),
2692 'language': lang
if lang
not in ('mul', 'und', 'zxx', 'mis') else None,
2693 'format_note': 'DASH %s' % content_type
,
2694 'filesize': filesize
,
2695 'container': mimetype2ext(mime_type
) + '_dash',
2698 elif content_type
== 'text':
2700 'ext': mimetype2ext(mime_type
),
2701 'manifest_url': mpd_url
,
2702 'filesize': filesize
,
2704 elif content_type
== 'image/jpeg':
2705 # See test case in VikiIE
2706 # https://www.viki.com/videos/1175236v-choosing-spouse-by-lottery-episode-1
2708 'format_id': format_id
,
2710 'manifest_url': mpd_url
,
2711 'format_note': 'DASH storyboards (jpeg)',
2715 if is_drm_protected(adaptation_set
) or is_drm_protected(representation
):
2717 representation_ms_info
= extract_multisegment_info(representation
, adaption_set_ms_info
)
2719 def prepare_template(template_name
, identifiers
):
2720 tmpl
= representation_ms_info
[template_name
]
2721 if representation_id
is not None:
2722 tmpl
= tmpl
.replace('$RepresentationID$', representation_id
)
2723 # First of, % characters outside $...$ templates
2724 # must be escaped by doubling for proper processing
2725 # by % operator string formatting used further (see
2726 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2732 in_template
= not in_template
2733 elif c
== '%' and not in_template
:
2735 # Next, $...$ templates are translated to their
2736 # %(...) counterparts to be used with % operator
2737 t
= re
.sub(r
'\$(%s)\$' % '|'.join(identifiers
), r
'%(\1)d', t
)
2738 t
= re
.sub(r
'\$(%s)%%([^$]+)\$' % '|'.join(identifiers
), r
'%(\1)\2', t
)
2739 t
.replace('$$', '$')
2742 # @initialization is a regular template like @media one
2743 # so it should be handled just the same way (see
2744 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2745 if 'initialization' in representation_ms_info
:
2746 initialization_template
= prepare_template(
2748 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2749 # $Time$ shall not be included for @initialization thus
2750 # only $Bandwidth$ remains
2752 representation_ms_info
['initialization_url'] = initialization_template
% {
2753 'Bandwidth': bandwidth
,
2756 def location_key(location
):
2757 return 'url' if re
.match(r
'^https?://', location
) else 'path'
2759 if 'segment_urls' not in representation_ms_info
and 'media' in representation_ms_info
:
2761 media_template
= prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2762 media_location_key
= location_key(media_template
)
2764 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2765 # can't be used at the same time
2766 if '%(Number' in media_template
and 's' not in representation_ms_info
:
2767 segment_duration
= None
2768 if 'total_number' not in representation_ms_info
and 'segment_duration' in representation_ms_info
:
2769 segment_duration
= float_or_none(representation_ms_info
['segment_duration'], representation_ms_info
['timescale'])
2770 representation_ms_info
['total_number'] = int(math
.ceil(
2771 float_or_none(period_duration
, segment_duration
, default
=0)))
2772 representation_ms_info
['fragments'] = [{
2773 media_location_key
: media_template
% {
2774 'Number': segment_number
,
2775 'Bandwidth': bandwidth
,
2777 'duration': segment_duration
,
2778 } for segment_number
in range(
2779 representation_ms_info
['start_number'],
2780 representation_ms_info
['total_number'] + representation_ms_info
['start_number'])]
2782 # $Number*$ or $Time$ in media template with S list available
2783 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2784 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2785 representation_ms_info
['fragments'] = []
2788 segment_number
= representation_ms_info
['start_number']
2790 def add_segment_url():
2791 segment_url
= media_template
% {
2792 'Time': segment_time
,
2793 'Bandwidth': bandwidth
,
2794 'Number': segment_number
,
2796 representation_ms_info
['fragments'].append({
2797 media_location_key
: segment_url
,
2798 'duration': float_or_none(segment_d
, representation_ms_info
['timescale']),
2801 for num
, s
in enumerate(representation_ms_info
['s']):
2802 segment_time
= s
.get('t') or segment_time
2806 for r
in range(s
.get('r', 0)):
2807 segment_time
+= segment_d
2810 segment_time
+= segment_d
2811 elif 'segment_urls' in representation_ms_info
and 's' in representation_ms_info
:
2812 # No media template,
2813 # e.g. https://www.youtube.com/watch?v=iXZV5uAYMJI
2814 # or any YouTube dashsegments video
2817 timescale
= representation_ms_info
['timescale']
2818 for s
in representation_ms_info
['s']:
2819 duration
= float_or_none(s
['d'], timescale
)
2820 for r
in range(s
.get('r', 0) + 1):
2821 segment_uri
= representation_ms_info
['segment_urls'][segment_index
]
2823 location_key(segment_uri
): segment_uri
,
2824 'duration': duration
,
2827 representation_ms_info
['fragments'] = fragments
2828 elif 'segment_urls' in representation_ms_info
:
2829 # Segment URLs with no SegmentTimeline
2830 # E.g. https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2831 # https://github.com/ytdl-org/youtube-dl/pull/14844
2833 segment_duration
= float_or_none(
2834 representation_ms_info
['segment_duration'],
2835 representation_ms_info
['timescale']) if 'segment_duration' in representation_ms_info
else None
2836 for segment_url
in representation_ms_info
['segment_urls']:
2838 location_key(segment_url
): segment_url
,
2840 if segment_duration
:
2841 fragment
['duration'] = segment_duration
2842 fragments
.append(fragment
)
2843 representation_ms_info
['fragments'] = fragments
2844 # If there is a fragments key available then we correctly recognized fragmented media.
2845 # Otherwise we will assume unfragmented media with direct access. Technically, such
2846 # assumption is not necessarily correct since we may simply have no support for
2847 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2848 if 'fragments' in representation_ms_info
:
2850 # NB: mpd_url may be empty when MPD manifest is parsed from a string
2851 'url': mpd_url
or base_url
,
2852 'fragment_base_url': base_url
,
2854 'protocol': 'http_dash_segments' if mime_type
!= 'image/jpeg' else 'mhtml',
2856 if 'initialization_url' in representation_ms_info
:
2857 initialization_url
= representation_ms_info
['initialization_url']
2858 if not f
.get('url'):
2859 f
['url'] = initialization_url
2860 f
['fragments'].append({location_key(initialization_url): initialization_url}
)
2861 f
['fragments'].extend(representation_ms_info
['fragments'])
2862 if not period_duration
:
2863 period_duration
= try_get(
2864 representation_ms_info
,
2865 lambda r
: sum(frag
['duration'] for frag
in r
['fragments']), float)
2867 # Assuming direct URL to unfragmented media.
2869 if content_type
in ('video', 'audio', 'image/jpeg'):
2870 f
['manifest_stream_number'] = stream_numbers
[f
['url']]
2871 stream_numbers
[f
['url']] += 1
2873 elif content_type
== 'text':
2874 subtitles
.setdefault(lang
or 'und', []).append(f
)
2876 return formats
, subtitles
2878 def _extract_ism_formats(self
, *args
, **kwargs
):
2879 fmts
, subs
= self
._extract
_ism
_formats
_and
_subtitles
(*args
, **kwargs
)
2881 self
._report
_ignoring
_subs
('ISM')
2884 def _extract_ism_formats_and_subtitles(self
, ism_url
, video_id
, ism_id
=None, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
):
2885 if self
.get_param('ignore_no_formats_error'):
2888 res
= self
._download
_xml
_handle
(
2890 note
='Downloading ISM manifest' if note
is None else note
,
2891 errnote
='Failed to download ISM manifest' if errnote
is None else errnote
,
2892 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
2899 return self
._parse
_ism
_formats
_and
_subtitles
(ism_doc
, urlh
.geturl(), ism_id
)
2901 def _parse_ism_formats_and_subtitles(self
, ism_doc
, ism_url
, ism_id
=None):
2903 Parse formats from ISM manifest.
2905 1. [MS-SSTR]: Smooth Streaming Protocol,
2906 https://msdn.microsoft.com/en-us/library/ff469518.aspx
2908 if ism_doc
.get('IsLive') == 'TRUE':
2911 duration
= int(ism_doc
.attrib
['Duration'])
2912 timescale
= int_or_none(ism_doc
.get('TimeScale')) or 10000000
2916 for stream
in ism_doc
.findall('StreamIndex'):
2917 stream_type
= stream
.get('Type')
2918 if stream_type
not in ('video', 'audio', 'text'):
2920 url_pattern
= stream
.attrib
['Url']
2921 stream_timescale
= int_or_none(stream
.get('TimeScale')) or timescale
2922 stream_name
= stream
.get('Name')
2923 stream_language
= stream
.get('Language', 'und')
2924 for track
in stream
.findall('QualityLevel'):
2925 KNOWN_TAGS
= {'255': 'AACL', '65534': 'EC-3'}
2926 fourcc
= track
.get('FourCC') or KNOWN_TAGS
.get(track
.get('AudioTag'))
2927 # TODO: add support for WVC1 and WMAP
2928 if fourcc
not in ('H264', 'AVC1', 'AACL', 'TTML', 'EC-3'):
2929 self
.report_warning('%s is not a supported codec' % fourcc
)
2931 tbr
= int(track
.attrib
['Bitrate']) // 1000
2932 # [1] does not mention Width and Height attributes. However,
2933 # they're often present while MaxWidth and MaxHeight are
2934 # missing, so should be used as fallbacks
2935 width
= int_or_none(track
.get('MaxWidth') or track
.get('Width'))
2936 height
= int_or_none(track
.get('MaxHeight') or track
.get('Height'))
2937 sampling_rate
= int_or_none(track
.get('SamplingRate'))
2939 track_url_pattern
= re
.sub(r
'{[Bb]itrate}', track
.attrib
['Bitrate'], url_pattern
)
2940 track_url_pattern
= urllib
.parse
.urljoin(ism_url
, track_url_pattern
)
2946 stream_fragments
= stream
.findall('c')
2947 for stream_fragment_index
, stream_fragment
in enumerate(stream_fragments
):
2948 fragment_ctx
['time'] = int_or_none(stream_fragment
.get('t')) or fragment_ctx
['time']
2949 fragment_repeat
= int_or_none(stream_fragment
.get('r')) or 1
2950 fragment_ctx
['duration'] = int_or_none(stream_fragment
.get('d'))
2951 if not fragment_ctx
['duration']:
2953 next_fragment_time
= int(stream_fragment
[stream_fragment_index
+ 1].attrib
['t'])
2955 next_fragment_time
= duration
2956 fragment_ctx
['duration'] = (next_fragment_time
- fragment_ctx
['time']) / fragment_repeat
2957 for _
in range(fragment_repeat
):
2959 'url': re
.sub(r
'{start[ _]time}', str(fragment_ctx
['time']), track_url_pattern
),
2960 'duration': fragment_ctx
['duration'] / stream_timescale
,
2962 fragment_ctx
['time'] += fragment_ctx
['duration']
2964 if stream_type
== 'text':
2965 subtitles
.setdefault(stream_language
, []).append({
2969 'manifest_url': ism_url
,
2970 'fragments': fragments
,
2971 '_download_params': {
2972 'stream_type': stream_type
,
2973 'duration': duration
,
2974 'timescale': stream_timescale
,
2976 'language': stream_language
,
2977 'codec_private_data': track
.get('CodecPrivateData'),
2980 elif stream_type
in ('video', 'audio'):
2982 'format_id': join_nonempty(ism_id
, stream_name
, tbr
),
2984 'manifest_url': ism_url
,
2985 'ext': 'ismv' if stream_type
== 'video' else 'isma',
2989 'asr': sampling_rate
,
2990 'vcodec': 'none' if stream_type
== 'audio' else fourcc
,
2991 'acodec': 'none' if stream_type
== 'video' else fourcc
,
2993 'fragments': fragments
,
2994 'has_drm': ism_doc
.find('Protection') is not None,
2995 'language': stream_language
,
2996 'audio_channels': int_or_none(track
.get('Channels')),
2997 '_download_params': {
2998 'stream_type': stream_type
,
2999 'duration': duration
,
3000 'timescale': stream_timescale
,
3001 'width': width
or 0,
3002 'height': height
or 0,
3004 'language': stream_language
,
3005 'codec_private_data': track
.get('CodecPrivateData'),
3006 'sampling_rate': sampling_rate
,
3007 'channels': int_or_none(track
.get('Channels', 2)),
3008 'bits_per_sample': int_or_none(track
.get('BitsPerSample', 16)),
3009 'nal_unit_length_field': int_or_none(track
.get('NALUnitLengthField', 4)),
3012 return formats
, subtitles
3014 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):
3015 def absolute_url(item_url
):
3016 return urljoin(base_url
, item_url
)
3018 def parse_content_type(content_type
):
3019 if not content_type
:
3021 ctr
= re
.search(r
'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type
)
3023 mimetype
, codecs
= ctr
.groups()
3024 f
= parse_codecs(codecs
)
3025 f
['ext'] = mimetype2ext(mimetype
)
3029 def _media_formats(src
, cur_media_type
, type_info
=None):
3030 type_info
= type_info
or {}
3031 full_url
= absolute_url(src
)
3032 ext
= type_info
.get('ext') or determine_ext(full_url
)
3034 is_plain_url
= False
3035 formats
= self
._extract
_m
3u8_formats
(
3036 full_url
, video_id
, ext
='mp4',
3037 entry_protocol
=m3u8_entry_protocol
, m3u8_id
=m3u8_id
,
3038 preference
=preference
, quality
=quality
, fatal
=False)
3040 is_plain_url
= False
3041 formats
= self
._extract
_mpd
_formats
(
3042 full_url
, video_id
, mpd_id
=mpd_id
, fatal
=False)
3047 'vcodec': 'none' if cur_media_type
== 'audio' else None,
3050 return is_plain_url
, formats
3053 # amp-video and amp-audio are very similar to their HTML5 counterparts
3054 # so we will include them right here (see
3055 # https://www.ampproject.org/docs/reference/components/amp-video)
3056 # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/
3057 _MEDIA_TAG_NAME_RE
= r
'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)'
3058 media_tags
= [(media_tag
, media_tag_name
, media_type
, '')
3059 for media_tag
, media_tag_name
, media_type
3060 in re
.findall(r
'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE
, webpage
)]
3061 media_tags
.extend(re
.findall(
3062 # We only allow video|audio followed by a whitespace or '>'.
3063 # Allowing more characters may end up in significant slow down (see
3064 # https://github.com/ytdl-org/youtube-dl/issues/11979,
3065 # e.g. http://www.porntrex.com/maps/videositemap.xml).
3066 r
'(?s)(<(?P<tag>%s)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>' % _MEDIA_TAG_NAME_RE
, webpage
))
3067 for media_tag
, _
, media_type
, media_content
in media_tags
:
3072 media_attributes
= extract_attributes(media_tag
)
3073 src
= strip_or_none(dict_get(media_attributes
, ('src', 'data-video-src', 'data-src', 'data-source')))
3075 f
= parse_content_type(media_attributes
.get('type'))
3076 _
, formats
= _media_formats(src
, media_type
, f
)
3077 media_info
['formats'].extend(formats
)
3078 media_info
['thumbnail'] = absolute_url(media_attributes
.get('poster'))
3080 for source_tag
in re
.findall(r
'<source[^>]+>', media_content
):
3081 s_attr
= extract_attributes(source_tag
)
3082 # data-video-src and data-src are non standard but seen
3083 # several times in the wild
3084 src
= strip_or_none(dict_get(s_attr
, ('src', 'data-video-src', 'data-src', 'data-source')))
3087 f
= parse_content_type(s_attr
.get('type'))
3088 is_plain_url
, formats
= _media_formats(src
, media_type
, f
)
3090 # width, height, res, label and title attributes are
3091 # all not standard but seen several times in the wild
3094 for lbl
in ('label', 'title')
3095 if str_or_none(s_attr
.get(lbl
))
3097 width
= int_or_none(s_attr
.get('width'))
3098 height
= (int_or_none(s_attr
.get('height'))
3099 or int_or_none(s_attr
.get('res')))
3100 if not width
or not height
:
3102 resolution
= parse_resolution(lbl
)
3105 width
= width
or resolution
.get('width')
3106 height
= height
or resolution
.get('height')
3108 tbr
= parse_bitrate(lbl
)
3117 'format_id': s_attr
.get('label') or s_attr
.get('title'),
3119 f
.update(formats
[0])
3120 media_info
['formats'].append(f
)
3122 media_info
['formats'].extend(formats
)
3123 for track_tag
in re
.findall(r
'<track[^>]+>', media_content
):
3124 track_attributes
= extract_attributes(track_tag
)
3125 kind
= track_attributes
.get('kind')
3126 if not kind
or kind
in ('subtitles', 'captions'):
3127 src
= strip_or_none(track_attributes
.get('src'))
3130 lang
= track_attributes
.get('srclang') or track_attributes
.get('lang') or track_attributes
.get('label')
3131 media_info
['subtitles'].setdefault(lang
, []).append({
3132 'url': absolute_url(src
),
3134 for f
in media_info
['formats']:
3135 f
.setdefault('http_headers', {})['Referer'] = base_url
3136 if media_info
['formats'] or media_info
['subtitles']:
3137 entries
.append(media_info
)
3140 def _extract_akamai_formats(self
, *args
, **kwargs
):
3141 fmts
, subs
= self
._extract
_akamai
_formats
_and
_subtitles
(*args
, **kwargs
)
3143 self
._report
_ignoring
_subs
('akamai')
3146 def _extract_akamai_formats_and_subtitles(self
, manifest_url
, video_id
, hosts
={}):
3147 signed
= 'hdnea=' in manifest_url
3149 # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
3150 manifest_url
= re
.sub(
3151 r
'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
3152 '', manifest_url
).strip('?')
3157 hdcore_sign
= 'hdcore=3.7.0'
3158 f4m_url
= re
.sub(r
'(https?://[^/]+)/i/', r
'\1/z/', manifest_url
).replace('/master.m3u8', '/manifest.f4m')
3159 hds_host
= hosts
.get('hds')
3161 f4m_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hds_host
, f4m_url
)
3162 if 'hdcore=' not in f4m_url
:
3163 f4m_url
+= ('&' if '?' in f4m_url
else '?') + hdcore_sign
3164 f4m_formats
= self
._extract
_f
4m
_formats
(
3165 f4m_url
, video_id
, f4m_id
='hds', fatal
=False)
3166 for entry
in f4m_formats
:
3167 entry
.update({'extra_param_to_segment_url': hdcore_sign}
)
3168 formats
.extend(f4m_formats
)
3170 m3u8_url
= re
.sub(r
'(https?://[^/]+)/z/', r
'\1/i/', manifest_url
).replace('/manifest.f4m', '/master.m3u8')
3171 hls_host
= hosts
.get('hls')
3173 m3u8_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hls_host
, m3u8_url
)
3174 m3u8_formats
, m3u8_subtitles
= self
._extract
_m
3u8_formats
_and
_subtitles
(
3175 m3u8_url
, video_id
, 'mp4', 'm3u8_native',
3176 m3u8_id
='hls', fatal
=False)
3177 formats
.extend(m3u8_formats
)
3178 subtitles
= self
._merge
_subtitles
(subtitles
, m3u8_subtitles
)
3180 http_host
= hosts
.get('http')
3181 if http_host
and m3u8_formats
and not signed
:
3182 REPL_REGEX
= r
'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
3183 qualities
= re
.match(REPL_REGEX
, m3u8_url
).group(2).split(',')
3184 qualities_length
= len(qualities
)
3185 if len(m3u8_formats
) in (qualities_length
, qualities_length
+ 1):
3187 for f
in m3u8_formats
:
3188 if f
['vcodec'] != 'none':
3189 for protocol
in ('http', 'https'):
3191 del http_f
['manifest_url']
3193 REPL_REGEX
, protocol
+ fr
'://{http_host}/\g<1>{qualities[i]}\3', f
['url'])
3195 'format_id': http_f
['format_id'].replace('hls-', protocol
+ '-'),
3197 'protocol': protocol
,
3199 formats
.append(http_f
)
3202 return formats
, subtitles
3204 def _extract_wowza_formats(self
, url
, video_id
, m3u8_entry_protocol
='m3u8_native', skip_protocols
=[]):
3205 query
= urllib
.parse
.urlparse(url
).query
3206 url
= re
.sub(r
'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url
)
3208 r
'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url
)
3209 url_base
= mobj
.group('url')
3210 http_base_url
= '%s%s:%s' % ('http', mobj
.group('s') or '', url_base
)
3213 def manifest_url(manifest
):
3214 m_url
= f
'{http_base_url}/{manifest}'
3216 m_url
+= '?%s' % query
3219 if 'm3u8' not in skip_protocols
:
3220 formats
.extend(self
._extract
_m
3u8_formats
(
3221 manifest_url('playlist.m3u8'), video_id
, 'mp4',
3222 m3u8_entry_protocol
, m3u8_id
='hls', fatal
=False))
3223 if 'f4m' not in skip_protocols
:
3224 formats
.extend(self
._extract
_f
4m
_formats
(
3225 manifest_url('manifest.f4m'),
3226 video_id
, f4m_id
='hds', fatal
=False))
3227 if 'dash' not in skip_protocols
:
3228 formats
.extend(self
._extract
_mpd
_formats
(
3229 manifest_url('manifest.mpd'),
3230 video_id
, mpd_id
='dash', fatal
=False))
3231 if re
.search(r
'(?:/smil:|\.smil)', url_base
):
3232 if 'smil' not in skip_protocols
:
3233 rtmp_formats
= self
._extract
_smil
_formats
(
3234 manifest_url('jwplayer.smil'),
3235 video_id
, fatal
=False)
3236 for rtmp_format
in rtmp_formats
:
3237 rtsp_format
= rtmp_format
.copy()
3238 rtsp_format
['url'] = '%s/%s' % (rtmp_format
['url'], rtmp_format
['play_path'])
3239 del rtsp_format
['play_path']
3240 del rtsp_format
['ext']
3241 rtsp_format
.update({
3242 'url': rtsp_format
['url'].replace('rtmp://', 'rtsp://'),
3243 'format_id': rtmp_format
['format_id'].replace('rtmp', 'rtsp'),
3246 formats
.extend([rtmp_format
, rtsp_format
])
3248 for protocol
in ('rtmp', 'rtsp'):
3249 if protocol
not in skip_protocols
:
3251 'url': f
'{protocol}:{url_base}',
3252 'format_id': protocol
,
3253 'protocol': protocol
,
3257 def _find_jwplayer_data(self
, webpage
, video_id
=None, transform_source
=js_to_json
):
3259 r
'''(?s)jwplayer\s*\(\s*(?P<q>'|")(?!(?P=q)).+(?P=q)\s*\)(?!</script>).*?\.\s*setup\s*\(\s*(?P<options>(?:\([^)]*\)|[^)])+)\s*\)''',
3263 jwplayer_data
= self
._parse
_json
(mobj
.group('options'),
3265 transform_source
=transform_source
)
3266 except ExtractorError
:
3269 if isinstance(jwplayer_data
, dict):
3270 return jwplayer_data
3272 def _extract_jwplayer_data(self
, webpage
, video_id
, *args
, **kwargs
):
3273 jwplayer_data
= self
._find
_jwplayer
_data
(
3274 webpage
, video_id
, transform_source
=js_to_json
)
3275 return self
._parse
_jwplayer
_data
(
3276 jwplayer_data
, video_id
, *args
, **kwargs
)
3278 def _parse_jwplayer_data(self
, jwplayer_data
, video_id
=None, require_title
=True,
3279 m3u8_id
=None, mpd_id
=None, rtmp_params
=None, base_url
=None):
3281 if not isinstance(jwplayer_data
, dict):
3284 playlist_items
= jwplayer_data
.get('playlist')
3285 # JWPlayer backward compatibility: single playlist item/flattened playlists
3286 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
3287 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
3288 if not isinstance(playlist_items
, list):
3289 playlist_items
= (playlist_items
or jwplayer_data
, )
3291 for video_data
in playlist_items
:
3292 if not isinstance(video_data
, dict):
3294 # JWPlayer backward compatibility: flattened sources
3295 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
3296 if 'sources' not in video_data
:
3297 video_data
['sources'] = [video_data
]
3299 this_video_id
= video_id
or video_data
['mediaid']
3301 formats
= self
._parse
_jwplayer
_formats
(
3302 video_data
['sources'], video_id
=this_video_id
, m3u8_id
=m3u8_id
,
3303 mpd_id
=mpd_id
, rtmp_params
=rtmp_params
, base_url
=base_url
)
3306 tracks
= video_data
.get('tracks')
3307 if tracks
and isinstance(tracks
, list):
3308 for track
in tracks
:
3309 if not isinstance(track
, dict):
3311 track_kind
= track
.get('kind')
3312 if not track_kind
or not isinstance(track_kind
, str):
3314 if track_kind
.lower() not in ('captions', 'subtitles'):
3316 track_url
= urljoin(base_url
, track
.get('file'))
3319 subtitles
.setdefault(track
.get('label') or 'en', []).append({
3320 'url': self
._proto
_relative
_url
(track_url
)
3324 'id': this_video_id
,
3325 'title': unescapeHTML(video_data
['title'] if require_title
else video_data
.get('title')),
3326 'description': clean_html(video_data
.get('description')),
3327 'thumbnail': urljoin(base_url
, self
._proto
_relative
_url
(video_data
.get('image'))),
3328 'timestamp': int_or_none(video_data
.get('pubdate')),
3329 'duration': float_or_none(jwplayer_data
.get('duration') or video_data
.get('duration')),
3330 'subtitles': subtitles
,
3331 'alt_title': clean_html(video_data
.get('subtitle')), # attributes used e.g. by Tele5 ...
3332 'genre': clean_html(video_data
.get('genre')),
3333 'channel': clean_html(dict_get(video_data
, ('category', 'channel'))),
3334 'season_number': int_or_none(video_data
.get('season')),
3335 'episode_number': int_or_none(video_data
.get('episode')),
3336 'release_year': int_or_none(video_data
.get('releasedate')),
3337 'age_limit': int_or_none(video_data
.get('age_restriction')),
3339 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
3340 if len(formats
) == 1 and re
.search(r
'^(?:http|//).*(?:youtube\.com|youtu\.be)/.+', formats
[0]['url']):
3342 '_type': 'url_transparent',
3343 'url': formats
[0]['url'],
3346 entry
['formats'] = formats
3347 entries
.append(entry
)
3348 if len(entries
) == 1:
3351 return self
.playlist_result(entries
)
3353 def _parse_jwplayer_formats(self
, jwplayer_sources_data
, video_id
=None,
3354 m3u8_id
=None, mpd_id
=None, rtmp_params
=None, base_url
=None):
3357 for source
in jwplayer_sources_data
:
3358 if not isinstance(source
, dict):
3360 source_url
= urljoin(
3361 base_url
, self
._proto
_relative
_url
(source
.get('file')))
3362 if not source_url
or source_url
in urls
:
3364 urls
.add(source_url
)
3365 source_type
= source
.get('type') or ''
3366 ext
= mimetype2ext(source_type
) or determine_ext(source_url
)
3367 if source_type
== 'hls' or ext
== 'm3u8' or 'format=m3u8-aapl' in source_url
:
3368 formats
.extend(self
._extract
_m
3u8_formats
(
3369 source_url
, video_id
, 'mp4', entry_protocol
='m3u8_native',
3370 m3u8_id
=m3u8_id
, fatal
=False))
3371 elif source_type
== 'dash' or ext
== 'mpd' or 'format=mpd-time-csf' in source_url
:
3372 formats
.extend(self
._extract
_mpd
_formats
(
3373 source_url
, video_id
, mpd_id
=mpd_id
, fatal
=False))
3375 formats
.extend(self
._extract
_smil
_formats
(
3376 source_url
, video_id
, fatal
=False))
3377 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3378 elif source_type
.startswith('audio') or ext
in (
3379 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
3386 format_id
= str_or_none(source
.get('label'))
3387 height
= int_or_none(source
.get('height'))
3388 if height
is None and format_id
:
3389 # Often no height is provided but there is a label in
3390 # format like "1080p", "720p SD", or 1080.
3391 height
= parse_resolution(format_id
).get('height')
3394 'width': int_or_none(source
.get('width')),
3396 'tbr': int_or_none(source
.get('bitrate'), scale
=1000),
3397 'filesize': int_or_none(source
.get('filesize')),
3399 'format_id': format_id
3401 if source_url
.startswith('rtmp'):
3402 a_format
['ext'] = 'flv'
3403 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3404 # of jwplayer.flash.swf
3405 rtmp_url_parts
= re
.split(
3406 r
'((?:mp4|mp3|flv):)', source_url
, 1)
3407 if len(rtmp_url_parts
) == 3:
3408 rtmp_url
, prefix
, play_path
= rtmp_url_parts
3411 'play_path': prefix
+ play_path
,
3414 a_format
.update(rtmp_params
)
3415 formats
.append(a_format
)
3418 def _live_title(self
, name
):
3419 self
._downloader
.deprecation_warning('yt_dlp.InfoExtractor._live_title is deprecated and does not work as expected')
3422 def _int(self
, v
, name
, fatal
=False, **kwargs
):
3423 res
= int_or_none(v
, **kwargs
)
3425 msg
= f
'Failed to extract {name}: Could not parse value {v!r}'
3427 raise ExtractorError(msg
)
3429 self
.report_warning(msg
)
3432 def _float(self
, v
, name
, fatal
=False, **kwargs
):
3433 res
= float_or_none(v
, **kwargs
)
3435 msg
= f
'Failed to extract {name}: Could not parse value {v!r}'
3437 raise ExtractorError(msg
)
3439 self
.report_warning(msg
)
3442 def _set_cookie(self
, domain
, name
, value
, expire_time
=None, port
=None,
3443 path
='/', secure
=False, discard
=False, rest
={}, **kwargs
):
3444 cookie
= http
.cookiejar
.Cookie(
3445 0, name
, value
, port
, port
is not None, domain
, True,
3446 domain
.startswith('.'), path
, True, secure
, expire_time
,
3447 discard
, None, None, rest
)
3448 self
.cookiejar
.set_cookie(cookie
)
3450 def _get_cookies(self
, url
):
3451 """ Return a http.cookies.SimpleCookie with the cookies for the url """
3452 return LenientSimpleCookie(self
._downloader
.cookiejar
.get_cookie_header(url
))
3454 def _apply_first_set_cookie_header(self
, url_handle
, cookie
):
3456 Apply first Set-Cookie header instead of the last. Experimental.
3458 Some sites (e.g. [1-3]) may serve two cookies under the same name
3459 in Set-Cookie header and expect the first (old) one to be set rather
3460 than second (new). However, as of RFC6265 the newer one cookie
3461 should be set into cookie store what actually happens.
3462 We will workaround this issue by resetting the cookie to
3463 the first one manually.
3464 1. https://new.vk.com/
3465 2. https://github.com/ytdl-org/youtube-dl/issues/9841#issuecomment-227871201
3466 3. https://learning.oreilly.com/
3468 for header
, cookies
in url_handle
.headers
.items():
3469 if header
.lower() != 'set-cookie':
3471 cookies
= cookies
.encode('iso-8859-1').decode('utf-8')
3472 cookie_value
= re
.search(
3473 r
'%s=(.+?);.*?\b[Dd]omain=(.+?)(?:[,;]|$)' % cookie
, cookies
)
3475 value
, domain
= cookie_value
.groups()
3476 self
._set
_cookie
(domain
, cookie
, value
)
3480 def get_testcases(cls
, include_onlymatching
=False):
3481 # Do not look in super classes
3482 t
= vars(cls
).get('_TEST')
3484 assert not hasattr(cls
, '_TESTS'), f
'{cls.ie_key()}IE has _TEST and _TESTS'
3487 tests
= vars(cls
).get('_TESTS', [])
3489 if not include_onlymatching
and t
.get('only_matching', False):
3491 t
['name'] = cls
.ie_key()
3493 if getattr(cls
, '__wrapped__', None):
3494 yield from cls
.__wrapped
__.get_testcases(include_onlymatching
)
3497 def get_webpage_testcases(cls
):
3498 tests
= vars(cls
).get('_WEBPAGE_TESTS', [])
3500 t
['name'] = cls
.ie_key()
3502 if getattr(cls
, '__wrapped__', None):
3503 yield from cls
.__wrapped
__.get_webpage_testcases()
3505 @classproperty(cache
=True)
3507 """Get age limit from the testcases"""
3508 return max(traverse_obj(
3509 (*cls
.get_testcases(include_onlymatching
=False), *cls
.get_webpage_testcases()),
3510 (..., (('playlist', 0), None), 'info_dict', 'age_limit')) or [0])
3512 @classproperty(cache
=True)
3513 def _RETURN_TYPE(cls
):
3514 """What the extractor returns: "video", "playlist", "any", or None (Unknown)"""
3515 tests
= tuple(cls
.get_testcases(include_onlymatching
=False))
3518 elif not any(k
.startswith('playlist') for test
in tests
for k
in test
):
3520 elif all(any(k
.startswith('playlist') for k
in test
) for test
in tests
):
3525 def is_single_video(cls
, url
):
3526 """Returns whether the URL is of a single video, None if unknown"""
3527 if cls
.suitable(url
):
3528 return {'video': True, 'playlist': False}
.get(cls
._RETURN
_TYPE
)
3531 def is_suitable(cls
, age_limit
):
3532 """Test whether the extractor is generally suitable for the given age limit"""
3533 return not age_restricted(cls
.age_limit
, age_limit
)
3536 def description(cls
, *, markdown
=True, search_examples
=None):
3537 """Description of the extractor"""
3539 if cls
._NETRC
_MACHINE
:
3541 desc
+= f
' [*{cls._NETRC_MACHINE}*](## "netrc machine")'
3543 desc
+= f
' [{cls._NETRC_MACHINE}]'
3544 if cls
.IE_DESC
is False:
3547 desc
+= f
' {cls.IE_DESC}'
3549 desc
+= f
'{";" if cls.IE_DESC else ""} "{cls.SEARCH_KEY}:" prefix'
3551 _COUNTS
= ('', '5', '10', 'all')
3552 desc
+= f
' (e.g. "{cls.SEARCH_KEY}{random.choice(_COUNTS)}:{random.choice(search_examples)}")'
3553 if not cls
.working():
3554 desc
+= ' (**Currently broken**)' if markdown
else ' (Currently broken)'
3556 # Escape emojis. Ref: https://github.com/github/markup/issues/1153
3557 name
= (' - **%s**' % re
.sub(r
':(\w+:)', ':\u200B\\g<1>', cls
.IE_NAME
)) if markdown
else cls
.IE_NAME
3558 return f
'{name}:{desc}' if desc
else name
3560 def extract_subtitles(self
, *args
, **kwargs
):
3561 if (self
.get_param('writesubtitles', False)
3562 or self
.get_param('listsubtitles')):
3563 return self
._get
_subtitles
(*args
, **kwargs
)
3566 def _get_subtitles(self
, *args
, **kwargs
):
3567 raise NotImplementedError('This method must be implemented by subclasses')
3569 class CommentsDisabled(Exception):
3570 """Raise in _get_comments if comments are disabled for the video"""
3572 def extract_comments(self
, *args
, **kwargs
):
3573 if not self
.get_param('getcomments'):
3575 generator
= self
._get
_comments
(*args
, **kwargs
)
3582 comments
.append(next(generator
))
3583 except StopIteration:
3585 except KeyboardInterrupt:
3586 self
.to_screen('Interrupted by user')
3587 except self
.CommentsDisabled
:
3588 return {'comments': None, 'comment_count': None}
3589 except Exception as e
:
3590 if self
.get_param('ignoreerrors') is not True:
3592 self
._downloader
.report_error(e
)
3593 comment_count
= len(comments
)
3594 self
.to_screen(f
'Extracted {comment_count} comments')
3596 'comments': comments
,
3597 'comment_count': None if interrupted
else comment_count
3601 def _get_comments(self
, *args
, **kwargs
):
3602 raise NotImplementedError('This method must be implemented by subclasses')
3605 def _merge_subtitle_items(subtitle_list1
, subtitle_list2
):
3606 """ Merge subtitle items for one language. Items with duplicated URLs/data
3607 will be dropped. """
3608 list1_data
= {(item.get('url'), item.get('data')) for item in subtitle_list1}
3609 ret
= list(subtitle_list1
)
3610 ret
.extend(item
for item
in subtitle_list2
if (item
.get('url'), item
.get('data')) not in list1_data
)
3614 def _merge_subtitles(cls
, *dicts
, target
=None):
3615 """ Merge subtitle dictionaries, language by language. """
3619 for lang
, subs
in d
.items():
3620 target
[lang
] = cls
._merge
_subtitle
_items
(target
.get(lang
, []), subs
)
3623 def extract_automatic_captions(self
, *args
, **kwargs
):
3624 if (self
.get_param('writeautomaticsub', False)
3625 or self
.get_param('listsubtitles')):
3626 return self
._get
_automatic
_captions
(*args
, **kwargs
)
3629 def _get_automatic_captions(self
, *args
, **kwargs
):
3630 raise NotImplementedError('This method must be implemented by subclasses')
3632 @functools.cached_property
3633 def _cookies_passed(self
):
3634 """Whether cookies have been passed to YoutubeDL"""
3635 return self
.get_param('cookiefile') is not None or self
.get_param('cookiesfrombrowser') is not None
3637 def mark_watched(self
, *args
, **kwargs
):
3638 if not self
.get_param('mark_watched', False):
3640 if self
.supports_login() and self
._get
_login
_info
()[0] is not None or self
._cookies
_passed
:
3641 self
._mark
_watched
(*args
, **kwargs
)
3643 def _mark_watched(self
, *args
, **kwargs
):
3644 raise NotImplementedError('This method must be implemented by subclasses')
3646 def geo_verification_headers(self
):
3648 geo_verification_proxy
= self
.get_param('geo_verification_proxy')
3649 if geo_verification_proxy
:
3650 headers
['Ytdl-request-proxy'] = geo_verification_proxy
3654 def _generic_id(url
):
3655 return urllib
.parse
.unquote(os
.path
.splitext(url
.rstrip('/').split('/')[-1])[0])
3657 def _generic_title(self
, url
='', webpage
='', *, default
=None):
3658 return (self
._og
_search
_title
(webpage
, default
=None)
3659 or self
._html
_extract
_title
(webpage
, default
=None)
3660 or urllib
.parse
.unquote(os
.path
.splitext(url_basename(url
))[0])
3663 def _extract_chapters_helper(self
, chapter_list
, start_function
, title_function
, duration
, strict
=True):
3667 'start_time': start_function(chapter
),
3668 'title': title_function(chapter
),
3669 } for chapter
in chapter_list
or []]
3671 warn
= self
.report_warning
3673 warn
= self
.write_debug
3674 chapter_list
.sort(key
=lambda c
: c
['start_time'] or 0)
3676 chapters
= [{'start_time': 0}
]
3677 for idx
, chapter
in enumerate(chapter_list
):
3678 if chapter
['start_time'] is None:
3679 warn(f
'Incomplete chapter {idx}')
3680 elif chapters
[-1]['start_time'] <= chapter
['start_time'] <= duration
:
3681 chapters
.append(chapter
)
3682 elif chapter
not in chapters
:
3683 issue
= (f
'{chapter["start_time"]} > {duration}' if chapter
['start_time'] > duration
3684 else f
'{chapter["start_time"]} < {chapters[-1]["start_time"]}')
3685 warn(f
'Invalid start time ({issue}) for chapter "{chapter["title"]}"')
3688 def _extract_chapters_from_description(self
, description
, duration
):
3689 duration_re
= r
'(?:\d+:)?\d{1,2}:\d{2}'
3690 sep_re
= r
'(?m)^\s*(%s)\b\W*\s(%s)\s*$'
3691 return self
._extract
_chapters
_helper
(
3692 re
.findall(sep_re
% (duration_re
, r
'.+?'), description
or ''),
3693 start_function
=lambda x
: parse_duration(x
[0]), title_function
=lambda x
: x
[1],
3694 duration
=duration
, strict
=False) or self
._extract
_chapters
_helper
(
3695 re
.findall(sep_re
% (r
'.+?', duration_re
), description
or ''),
3696 start_function
=lambda x
: parse_duration(x
[1]), title_function
=lambda x
: x
[0],
3697 duration
=duration
, strict
=False)
3700 def _availability(is_private
=None, needs_premium
=None, needs_subscription
=None, needs_auth
=None, is_unlisted
=None):
3701 all_known
= all(map(
3702 lambda x
: x
is not None,
3703 (is_private
, needs_premium
, needs_subscription
, needs_auth
, is_unlisted
)))
3705 'private' if is_private
3706 else 'premium_only' if needs_premium
3707 else 'subscriber_only' if needs_subscription
3708 else 'needs_auth' if needs_auth
3709 else 'unlisted' if is_unlisted
3710 else 'public' if all_known
3713 def _configuration_arg(self
, key
, default
=NO_DEFAULT
, *, ie_key
=None, casesense
=False):
3715 @returns A list of values for the extractor argument given by "key"
3716 or "default" if no such key is present
3717 @param default The default value to return when the key is not present (default: [])
3718 @param casesense When false, the values are converted to lower case
3720 ie_key
= ie_key
if isinstance(ie_key
, str) else (ie_key
or self
).ie_key()
3721 val
= traverse_obj(self
._downloader
.params
, ('extractor_args', ie_key
.lower(), key
))
3723 return [] if default
is NO_DEFAULT
else default
3724 return list(val
) if casesense
else [x
.lower() for x
in val
]
3726 def _yes_playlist(self
, playlist_id
, video_id
, smuggled_data
=None, *, playlist_label
='playlist', video_label
='video'):
3727 if not playlist_id
or not video_id
:
3730 no_playlist
= (smuggled_data
or {}).get('force_noplaylist')
3731 if no_playlist
is not None:
3732 return not no_playlist
3734 video_id
= '' if video_id
is True else f
' {video_id}'
3735 playlist_id
= '' if playlist_id
is True else f
' {playlist_id}'
3736 if self
.get_param('noplaylist'):
3737 self
.to_screen(f
'Downloading just the {video_label}{video_id} because of --no-playlist')
3739 self
.to_screen(f
'Downloading {playlist_label}{playlist_id} - add --no-playlist to download just the {video_label}{video_id}')
3742 def _error_or_warning(self
, err
, _count
=None, _retries
=0, *, fatal
=True):
3743 RetryManager
.report_retry(
3744 err
, _count
or int(fatal
), _retries
,
3745 info
=self
.to_screen
, warn
=self
.report_warning
, error
=None if fatal
else self
.report_warning
,
3746 sleep_func
=self
.get_param('retry_sleep_functions', {}).get('extractor'))
3748 def RetryManager(self
, **kwargs
):
3749 return RetryManager(self
.get_param('extractor_retries', 3), self
._error
_or
_warning
, **kwargs
)
3751 def _extract_generic_embeds(self
, url
, *args
, info_dict
={}, note
='Extracting generic embeds', **kwargs
):
3752 display_id
= traverse_obj(info_dict
, 'display_id', 'id')
3753 self
.to_screen(f
'{format_field(display_id, None, "%s: ")}{note}')
3754 return self
._downloader
.get_info_extractor('Generic')._extract
_embeds
(
3755 smuggle_url(url
, {'block_ies': [self.ie_key()]}
), *args
, **kwargs
)
3758 def extract_from_webpage(cls
, ydl
, url
, webpage
):
3759 ie
= (cls
if isinstance(cls
._extract
_from
_webpage
, types
.MethodType
)
3760 else ydl
.get_info_extractor(cls
.ie_key()))
3761 for info
in ie
._extract
_from
_webpage
(url
, webpage
) or []:
3762 # url = None since we do not want to set (webpage/original)_url
3763 ydl
.add_default_extra_info(info
, ie
, None)
3767 def _extract_from_webpage(cls
, url
, webpage
):
3768 for embed_url
in orderedSet(
3769 cls
._extract
_embed
_urls
(url
, webpage
) or [], lazy
=True):
3770 yield cls
.url_result(embed_url
, None if cls
._VALID
_URL
is False else cls
)
3773 def _extract_embed_urls(cls
, url
, webpage
):
3774 """@returns all the embed urls on the webpage"""
3775 if '_EMBED_URL_RE' not in cls
.__dict
__:
3776 assert isinstance(cls
._EMBED
_REGEX
, (list, tuple))
3777 for idx
, regex
in enumerate(cls
._EMBED
_REGEX
):
3778 assert regex
.count('(?P<url>') == 1, \
3779 f
'{cls.__name__}._EMBED_REGEX[{idx}] must have exactly 1 url group\n\t{regex}'
3780 cls
._EMBED
_URL
_RE
= tuple(map(re
.compile, cls
._EMBED
_REGEX
))
3782 for regex
in cls
._EMBED
_URL
_RE
:
3783 for mobj
in regex
.finditer(webpage
):
3784 embed_url
= urllib
.parse
.urljoin(url
, unescapeHTML(mobj
.group('url')))
3785 if cls
._VALID
_URL
is False or cls
.suitable(embed_url
):
3788 class StopExtraction(Exception):
3792 def _extract_url(cls
, webpage
): # TODO: Remove
3793 """Only for compatibility with some older extractors"""
3794 return next(iter(cls
._extract
_embed
_urls
(None, webpage
) or []), None)
3797 def __init_subclass__(cls
, *, plugin_name
=None, **kwargs
):
3799 mro
= inspect
.getmro(cls
)
3800 super_class
= cls
.__wrapped
__ = mro
[mro
.index(cls
) + 1]
3801 cls
.PLUGIN_NAME
, cls
.ie_key
= plugin_name
, super_class
.ie_key
3802 cls
.IE_NAME
= f
'{super_class.IE_NAME}+{plugin_name}'
3803 while getattr(super_class
, '__wrapped__', None):
3804 super_class
= super_class
.__wrapped
__
3805 setattr(sys
.modules
[super_class
.__module
__], super_class
.__name
__, cls
)
3806 _PLUGIN_OVERRIDES
[super_class
].append(cls
)
3808 return super().__init
_subclass
__(**kwargs
)
3811 class SearchInfoExtractor(InfoExtractor
):
3813 Base class for paged search queries extractors.
3814 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
3815 Instances should define _SEARCH_KEY and optionally _MAX_RESULTS
3818 _MAX_RESULTS
= float('inf')
3819 _RETURN_TYPE
= 'playlist'
3822 def _VALID_URL(cls
):
3823 return r
'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls
._SEARCH
_KEY
3825 def _real_extract(self
, query
):
3826 prefix
, query
= self
._match
_valid
_url
(query
).group('prefix', 'query')
3828 return self
._get
_n
_results
(query
, 1)
3829 elif prefix
== 'all':
3830 return self
._get
_n
_results
(query
, self
._MAX
_RESULTS
)
3834 raise ExtractorError(f
'invalid download number {n} for query "{query}"')
3835 elif n
> self
._MAX
_RESULTS
:
3836 self
.report_warning('%s returns max %i results (you requested %i)' % (self
._SEARCH
_KEY
, self
._MAX
_RESULTS
, n
))
3837 n
= self
._MAX
_RESULTS
3838 return self
._get
_n
_results
(query
, n
)
3840 def _get_n_results(self
, query
, n
):
3841 """Get a specified number of results for a query.
3842 Either this function or _search_results must be overridden by subclasses """
3843 return self
.playlist_result(
3844 itertools
.islice(self
._search
_results
(query
), 0, None if n
== float('inf') else n
),
3847 def _search_results(self
, query
):
3848 """Returns an iterator of search results"""
3849 raise NotImplementedError('This method must be implemented by subclasses')
3852 def SEARCH_KEY(cls
):
3853 return cls
._SEARCH
_KEY
3856 class UnsupportedURLIE(InfoExtractor
):
3861 def _real_extract(self
, url
):
3862 raise UnsupportedError(url
)
3865 _PLUGIN_OVERRIDES
= collections
.defaultdict(list)