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 channel_is_verified: Whether the channel is verified on the platform.
290 location: Physical location where the video was filmed.
291 subtitles: The available subtitles as a dictionary in the format
292 {tag: subformats}. "tag" is usually a language code, and
293 "subformats" is a list sorted from lower to higher
294 preference, each element is a dictionary with the "ext"
296 * "data": The subtitles file contents
297 * "url": A URL pointing to the subtitles file
298 It can optionally also have:
299 * "name": Name or description of the subtitles
300 * "http_headers": A dictionary of additional HTTP headers
301 to add to the request.
302 "ext" will be calculated from URL if missing
303 automatic_captions: Like 'subtitles'; contains automatically generated
304 captions instead of normal subtitles
305 duration: Length of the video in seconds, as an integer or float.
306 view_count: How many users have watched the video on the platform.
307 concurrent_view_count: How many users are currently watching the video on the platform.
308 like_count: Number of positive ratings of the video
309 dislike_count: Number of negative ratings of the video
310 repost_count: Number of reposts of the video
311 average_rating: Average rating give by users, the scale used depends on the webpage
312 comment_count: Number of comments on the video
313 comments: A list of comments, each with one or more of the following
314 properties (all but one of text or html optional):
315 * "author" - human-readable name of the comment author
316 * "author_id" - user ID of the comment author
317 * "author_thumbnail" - The thumbnail of the comment author
318 * "author_url" - The url to the comment author's page
319 * "author_is_verified" - Whether the author is verified
321 * "author_is_uploader" - Whether the comment is made by
324 * "html" - Comment as HTML
325 * "text" - Plain text of the comment
326 * "timestamp" - UNIX timestamp of comment
327 * "parent" - ID of the comment this one is replying to.
328 Set to "root" to indicate that this is a
329 comment to the original video.
330 * "like_count" - Number of positive ratings of the comment
331 * "dislike_count" - Number of negative ratings of the comment
332 * "is_favorited" - Whether the comment is marked as
333 favorite by the video uploader
334 * "is_pinned" - Whether the comment is pinned to
335 the top of the comments
336 age_limit: Age restriction for the video, as an integer (years)
337 webpage_url: The URL to the video webpage, if given to yt-dlp it
338 should allow to get the same result again. (It will be set
339 by YoutubeDL if it's missing)
340 categories: A list of categories that the video falls in, for example
342 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
343 cast: A list of the video cast
344 is_live: True, False, or None (=unknown). Whether this video is a
345 live stream that goes on instead of a fixed-length video.
346 was_live: True, False, or None (=unknown). Whether this video was
347 originally a live stream.
348 live_status: None (=unknown), 'is_live', 'is_upcoming', 'was_live', 'not_live',
349 or 'post_live' (was live, but VOD is not yet processed)
350 If absent, automatically set from is_live, was_live
351 start_time: Time in seconds where the reproduction should start, as
352 specified in the URL.
353 end_time: Time in seconds where the reproduction should end, as
354 specified in the URL.
355 chapters: A list of dictionaries, with the following entries:
356 * "start_time" - The start time of the chapter in seconds
357 * "end_time" - The end time of the chapter in seconds
358 * "title" (optional, string)
359 heatmap: A list of dictionaries, with the following entries:
360 * "start_time" - The start time of the data point in seconds
361 * "end_time" - The end time of the data point in seconds
362 * "value" - The normalized value of the data point (float between 0 and 1)
363 playable_in_embed: Whether this video is allowed to play in embedded
364 players on other sites. Can be True (=always allowed),
365 False (=never allowed), None (=unknown), or a string
366 specifying the criteria for embedability; e.g. 'whitelist'
367 availability: Under what condition the video is available. One of
368 'private', 'premium_only', 'subscriber_only', 'needs_auth',
369 'unlisted' or 'public'. Use 'InfoExtractor._availability'
371 _old_archive_ids: A list of old archive ids needed for backward compatibility
372 _format_sort_fields: A list of fields to use for sorting formats
373 __post_extractor: A function to be called just before the metadata is
374 written to either disk, logger or console. The function
375 must return a dict which will be added to the info_dict.
376 This is usefull for additional information that is
377 time-consuming to extract. Note that the fields thus
378 extracted will not be available to output template and
379 match_filter. So, only "comments" and "comment_count" are
380 currently allowed to be extracted via this method.
382 The following fields should only be used when the video belongs to some logical
385 chapter: Name or title of the chapter the video belongs to.
386 chapter_number: Number of the chapter the video belongs to, as an integer.
387 chapter_id: Id of the chapter the video belongs to, as a unicode string.
389 The following fields should only be used when the video is an episode of some
390 series, programme or podcast:
392 series: Title of the series or programme the video episode belongs to.
393 series_id: Id of the series or programme the video episode belongs to, as a unicode string.
394 season: Title of the season the video episode belongs to.
395 season_number: Number of the season the video episode belongs to, as an integer.
396 season_id: Id of the season the video episode belongs to, as a unicode string.
397 episode: Title of the video episode. Unlike mandatory video title field,
398 this field should denote the exact title of the video episode
399 without any kind of decoration.
400 episode_number: Number of the video episode within a season, as an integer.
401 episode_id: Id of the video episode, as a unicode string.
403 The following fields should only be used when the media is a track or a part of
406 track: Title of the track.
407 track_number: Number of the track within an album or a disc, as an integer.
408 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
410 artist: Artist(s) of the track.
411 genre: Genre(s) of the track.
412 album: Title of the album the track belongs to.
413 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
414 album_artist: List of all artists appeared on the album (e.g.
415 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
417 disc_number: Number of the disc or other physical medium the track belongs to,
419 release_year: Year (YYYY) when the album was released.
420 composer: Composer of the piece
422 The following fields should only be set for clips that should be cut from the original video:
424 section_start: Start time of the section in seconds
425 section_end: End time of the section in seconds
427 The following fields should only be set for storyboards:
428 rows: Number of rows in each storyboard fragment, as an integer
429 columns: Number of columns in each storyboard fragment, as an integer
431 Unless mentioned otherwise, the fields should be Unicode strings.
433 Unless mentioned otherwise, None is equivalent to absence of information.
436 _type "playlist" indicates multiple videos.
437 There must be a key "entries", which is a list, an iterable, or a PagedList
438 object, each element of which is a valid dictionary by this specification.
440 Additionally, playlists can have "id", "title", and any other relevant
441 attributes with the same semantics as videos (see above).
443 It can also have the following optional fields:
445 playlist_count: The total number of videos in a playlist. If not given,
446 YoutubeDL tries to calculate it from "entries"
449 _type "multi_video" indicates that there are multiple videos that
450 form a single show, for examples multiple acts of an opera or TV episode.
451 It must have an entries key like a playlist and contain all the keys
452 required for a video at the same time.
455 _type "url" indicates that the video must be extracted from another
456 location, possibly by a different extractor. Its only required key is:
457 "url" - the next URL to extract.
458 The key "ie_key" can be set to the class name (minus the trailing "IE",
459 e.g. "Youtube") if the extractor class is known in advance.
460 Additionally, the dictionary may have any properties of the resolved entity
461 known in advance, for example "title" if the title of the referred video is
465 _type "url_transparent" entities have the same specification as "url", but
466 indicate that the given additional information is more precise than the one
467 associated with the resolved URL.
468 This is useful when a site employs a video service that hosts the video and
469 its technical metadata, but that video service does not embed a useful
470 title, description etc.
473 Subclasses of this should also be added to the list of extractors and
474 should define a _VALID_URL regexp and, re-define the _real_extract() and
475 (optionally) _real_initialize() methods.
477 Subclasses may also override suitable() if necessary, but ensure the function
478 signature is preserved and that this function imports everything it needs
479 (except other extractors), so that lazy_extractors works correctly.
481 Subclasses can define a list of _EMBED_REGEX, which will be searched for in
482 the HTML of Generic webpages. It may also override _extract_embed_urls
483 or _extract_from_webpage as necessary. While these are normally classmethods,
484 _extract_from_webpage is allowed to be an instance method.
486 _extract_from_webpage may raise self.StopExtraction() to stop further
487 processing of the webpage and obtain exclusive rights to it. This is useful
488 when the extractor cannot reliably be matched using just the URL,
489 e.g. invidious/peertube instances
491 Embed-only extractors can be defined by setting _VALID_URL = False.
493 To support username + password (or netrc) login, the extractor must define a
494 _NETRC_MACHINE and re-define _perform_login(username, password) and
495 (optionally) _initialize_pre_login() methods. The _perform_login method will
496 be called between _initialize_pre_login and _real_initialize if credentials
497 are passed by the user. In cases where it is necessary to have the login
498 process as part of the extraction rather than initialization, _perform_login
499 can be left undefined.
501 _GEO_BYPASS attribute may be set to False in order to disable
502 geo restriction bypass mechanisms for a particular extractor.
503 Though it won't disable explicit geo restriction bypass based on
504 country code provided with geo_bypass_country.
506 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
507 countries for this extractor. One of these countries will be used by
508 geo restriction bypass mechanism right away in order to bypass
509 geo restriction, of course, if the mechanism is not disabled.
511 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
512 IP blocks in CIDR notation for this extractor. One of these IP blocks
513 will be used by geo restriction bypass mechanism similarly
516 The _ENABLED attribute should be set to False for IEs that
517 are disabled by default and must be explicitly enabled.
519 The _WORKING attribute should be set to False for broken IEs
520 in order to warn the users and skip the tests.
525 _x_forwarded_for_ip
= None
527 _GEO_COUNTRIES
= None
528 _GEO_IP_BLOCKS
= None
531 _NETRC_MACHINE
= None
537 def _login_hint(self
, method
=NO_DEFAULT
, netrc
=None):
538 password_hint
= f
'--username and --password, or --netrc ({netrc or self._NETRC_MACHINE}) to provide account credentials'
541 'any': f
'Use --cookies, --cookies-from-browser, {password_hint}',
542 'password': f
'Use {password_hint}',
544 'Use --cookies-from-browser or --cookies for the authentication. '
545 'See https://github.com/yt-dlp/yt-dlp/wiki/FAQ#how-do-i-pass-cookies-to-yt-dlp for how to manually pass cookies'),
546 }[method
if method
is not NO_DEFAULT
else 'any' if self
.supports_login() else 'cookies']
548 def __init__(self
, downloader
=None):
549 """Constructor. Receives an optional downloader (a YoutubeDL instance).
550 If a downloader is not passed during initialization,
551 it must be set using "set_downloader()" before "extract()" is called"""
553 self
._x
_forwarded
_for
_ip
= None
554 self
._printed
_messages
= set()
555 self
.set_downloader(downloader
)
558 def _match_valid_url(cls
, url
):
559 if cls
._VALID
_URL
is False:
561 # This does not use has/getattr intentionally - we want to know whether
562 # we have cached the regexp for *this* class, whereas getattr would also
563 # match the superclass
564 if '_VALID_URL_RE' not in cls
.__dict
__:
565 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
566 return cls
._VALID
_URL
_RE
.match(url
)
569 def suitable(cls
, url
):
570 """Receives a URL and returns True if suitable for this IE."""
571 # This function must import everything it needs (except other extractors),
572 # so that lazy_extractors works correctly
573 return cls
._match
_valid
_url
(url
) is not None
576 def _match_id(cls
, url
):
577 return cls
._match
_valid
_url
(url
).group('id')
580 def get_temp_id(cls
, url
):
582 return cls
._match
_id
(url
)
583 except (IndexError, AttributeError):
588 """Getter method for _WORKING."""
592 def supports_login(cls
):
593 return bool(cls
._NETRC
_MACHINE
)
595 def initialize(self
):
596 """Initializes an instance (authentication, etc)."""
597 self
._printed
_messages
= set()
598 self
._initialize
_geo
_bypass
({
599 'countries': self
._GEO
_COUNTRIES
,
600 'ip_blocks': self
._GEO
_IP
_BLOCKS
,
603 self
._initialize
_pre
_login
()
604 if self
.supports_login():
605 username
, password
= self
._get
_login
_info
()
607 self
._perform
_login
(username
, password
)
608 elif self
.get_param('username') and False not in (self
.IE_DESC
, self
._NETRC
_MACHINE
):
609 self
.report_warning(f
'Login with password is not supported for this website. {self._login_hint("cookies")}')
610 self
._real
_initialize
()
613 def _initialize_geo_bypass(self
, geo_bypass_context
):
615 Initialize geo restriction bypass mechanism.
617 This method is used to initialize geo bypass mechanism based on faking
618 X-Forwarded-For HTTP header. A random country from provided country list
619 is selected and a random IP belonging to this country is generated. This
620 IP will be passed as X-Forwarded-For HTTP header in all subsequent
623 This method will be used for initial geo bypass mechanism initialization
624 during the instance initialization with _GEO_COUNTRIES and
627 You may also manually call it from extractor's code if geo bypass
628 information is not available beforehand (e.g. obtained during
629 extraction) or due to some other reason. In this case you should pass
630 this information in geo bypass context passed as first argument. It may
631 contain following fields:
633 countries: List of geo unrestricted countries (similar
635 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
636 (similar to _GEO_IP_BLOCKS)
639 if not self
._x
_forwarded
_for
_ip
:
641 # Geo bypass mechanism is explicitly disabled by user
642 if not self
.get_param('geo_bypass', True):
645 if not geo_bypass_context
:
646 geo_bypass_context
= {}
648 # Backward compatibility: previously _initialize_geo_bypass
649 # expected a list of countries, some 3rd party code may still use
651 if isinstance(geo_bypass_context
, (list, tuple)):
652 geo_bypass_context
= {
653 'countries': geo_bypass_context
,
656 # The whole point of geo bypass mechanism is to fake IP
657 # as X-Forwarded-For HTTP header based on some IP block or
660 # Path 1: bypassing based on IP block in CIDR notation
662 # Explicit IP block specified by user, use it right away
663 # regardless of whether extractor is geo bypassable or not
664 ip_block
= self
.get_param('geo_bypass_ip_block', None)
666 # Otherwise use random IP block from geo bypass context but only
667 # if extractor is known as geo bypassable
669 ip_blocks
= geo_bypass_context
.get('ip_blocks')
670 if self
._GEO
_BYPASS
and ip_blocks
:
671 ip_block
= random
.choice(ip_blocks
)
674 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(ip_block
)
675 self
.write_debug(f
'Using fake IP {self._x_forwarded_for_ip} as X-Forwarded-For')
678 # Path 2: bypassing based on country code
680 # Explicit country code specified by user, use it right away
681 # regardless of whether extractor is geo bypassable or not
682 country
= self
.get_param('geo_bypass_country', None)
684 # Otherwise use random country code from geo bypass context but
685 # only if extractor is known as geo bypassable
687 countries
= geo_bypass_context
.get('countries')
688 if self
._GEO
_BYPASS
and countries
:
689 country
= random
.choice(countries
)
692 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country
)
693 self
._downloader
.write_debug(
694 f
'Using fake IP {self._x_forwarded_for_ip} ({country.upper()}) as X-Forwarded-For')
696 def extract(self
, url
):
697 """Extracts URL information and returns it in list of dicts."""
702 self
.to_screen('Extracting URL: %s' % (
703 url
if self
.get_param('verbose') else truncate_string(url
, 100, 20)))
704 ie_result
= self
._real
_extract
(url
)
705 if ie_result
is None:
707 if self
._x
_forwarded
_for
_ip
:
708 ie_result
['__x_forwarded_for_ip'] = self
._x
_forwarded
_for
_ip
709 subtitles
= ie_result
.get('subtitles') or {}
710 if 'no-live-chat' in self
.get_param('compat_opts'):
711 for lang
in ('live_chat', 'comments', 'danmaku'):
712 subtitles
.pop(lang
, None)
714 except GeoRestrictedError
as e
:
715 if self
.__maybe
_fake
_ip
_and
_retry
(e
.countries
):
718 except UnsupportedError
:
720 except ExtractorError
as e
:
721 e
.video_id
= e
.video_id
or self
.get_temp_id(url
),
722 e
.ie
= e
.ie
or self
.IE_NAME
,
723 e
.traceback
= e
.traceback
or sys
.exc_info()[2]
725 except http
.client
.IncompleteRead
as e
:
726 raise ExtractorError('A network error has occurred.', cause
=e
, expected
=True, video_id
=self
.get_temp_id(url
))
727 except (KeyError, StopIteration) as e
:
728 raise ExtractorError('An extractor error has occurred.', cause
=e
, video_id
=self
.get_temp_id(url
))
730 def __maybe_fake_ip_and_retry(self
, countries
):
731 if (not self
.get_param('geo_bypass_country', None)
733 and self
.get_param('geo_bypass', True)
734 and not self
._x
_forwarded
_for
_ip
736 country_code
= random
.choice(countries
)
737 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country_code
)
738 if self
._x
_forwarded
_for
_ip
:
740 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
741 % (self
._x
_forwarded
_for
_ip
, country_code
.upper()))
745 def set_downloader(self
, downloader
):
746 """Sets a YoutubeDL instance as the downloader for this IE."""
747 self
._downloader
= downloader
751 return self
._downloader
.cache
755 return self
._downloader
.cookiejar
757 def _initialize_pre_login(self
):
758 """ Initialization before login. Redefine in subclasses."""
761 def _perform_login(self
, username
, password
):
762 """ Login with username and password. Redefine in subclasses."""
765 def _real_initialize(self
):
766 """Real initialization process. Redefine in subclasses."""
769 def _real_extract(self
, url
):
770 """Real extraction process. Redefine in subclasses."""
771 raise NotImplementedError('This method must be implemented by subclasses')
775 """A string for getting the InfoExtractor with get_info_extractor"""
776 return cls
.__name
__[:-2]
780 return cls
.__name
__[:-2]
783 def __can_accept_status_code(err
, expected_status
):
784 assert isinstance(err
, urllib
.error
.HTTPError
)
785 if expected_status
is None:
787 elif callable(expected_status
):
788 return expected_status(err
.code
) is True
790 return err
.code
in variadic(expected_status
)
792 def _create_request(self
, url_or_request
, data
=None, headers
=None, query
=None):
793 if isinstance(url_or_request
, urllib
.request
.Request
):
794 return update_Request(url_or_request
, data
=data
, headers
=headers
, query
=query
)
796 url_or_request
= update_url_query(url_or_request
, query
)
797 return sanitized_Request(url_or_request
, data
, headers
or {})
799 def _request_webpage(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, data
=None, headers
=None, query
=None, expected_status
=None):
801 Return the response handle.
803 See _download_webpage docstring for arguments specification.
805 if not self
._downloader
._first
_webpage
_request
:
806 sleep_interval
= self
.get_param('sleep_interval_requests') or 0
807 if sleep_interval
> 0:
808 self
.to_screen('Sleeping %s seconds ...' % sleep_interval
)
809 time
.sleep(sleep_interval
)
811 self
._downloader
._first
_webpage
_request
= False
814 self
.report_download_webpage(video_id
)
815 elif note
is not False:
817 self
.to_screen(str(note
))
819 self
.to_screen(f
'{video_id}: {note}')
821 # Some sites check X-Forwarded-For HTTP header in order to figure out
822 # the origin of the client behind proxy. This allows bypassing geo
823 # restriction by faking this header's value to IP that belongs to some
824 # geo unrestricted country. We will do so once we encounter any
825 # geo restriction error.
826 if self
._x
_forwarded
_for
_ip
:
827 headers
= (headers
or {}).copy()
828 headers
.setdefault('X-Forwarded-For', self
._x
_forwarded
_for
_ip
)
831 return self
._downloader
.urlopen(self
._create
_request
(url_or_request
, data
, headers
, query
))
832 except network_exceptions
as err
:
833 if isinstance(err
, urllib
.error
.HTTPError
):
834 if self
.__can
_accept
_status
_code
(err
, expected_status
):
835 # Retain reference to error to prevent file object from
836 # being closed before it can be read. Works around the
837 # effects of <https://bugs.python.org/issue15002>
838 # introduced in Python 3.4.1.
845 errnote
= 'Unable to download webpage'
847 errmsg
= f
'{errnote}: {error_to_compat_str(err)}'
849 raise ExtractorError(errmsg
, cause
=err
)
851 self
.report_warning(errmsg
)
854 def _download_webpage_handle(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True,
855 encoding
=None, data
=None, headers
={}, query={}
, expected_status
=None):
857 Return a tuple (page content as string, URL handle).
860 url_or_request -- plain text URL as a string or
861 a urllib.request.Request object
862 video_id -- Video/playlist/item identifier (string)
865 note -- note printed before downloading (string)
866 errnote -- note printed in case of an error (string)
867 fatal -- flag denoting whether error should be considered fatal,
868 i.e. whether it should cause ExtractionError to be raised,
869 otherwise a warning will be reported and extraction continued
870 encoding -- encoding for a page content decoding, guessed automatically
871 when not explicitly specified
872 data -- POST data (bytes)
873 headers -- HTTP headers (dict)
874 query -- URL query (dict)
875 expected_status -- allows to accept failed HTTP requests (non 2xx
876 status code) by explicitly specifying a set of accepted status
877 codes. Can be any of the following entities:
878 - an integer type specifying an exact failed status code to
880 - a list or a tuple of integer types specifying a list of
881 failed status codes to accept
882 - a callable accepting an actual failed status code and
883 returning True if it should be accepted
884 Note that this argument does not affect success status codes (2xx)
885 which are always accepted.
888 # Strip hashes from the URL (#1038)
889 if isinstance(url_or_request
, str):
890 url_or_request
= url_or_request
.partition('#')[0]
892 urlh
= self
._request
_webpage
(url_or_request
, video_id
, note
, errnote
, fatal
, data
=data
, headers
=headers
, query
=query
, expected_status
=expected_status
)
896 content
= self
._webpage
_read
_content
(urlh
, url_or_request
, video_id
, note
, errnote
, fatal
, encoding
=encoding
)
897 return (content
, urlh
)
900 def _guess_encoding_from_content(content_type
, webpage_bytes
):
901 m
= re
.match(r
'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type
)
903 encoding
= m
.group(1)
905 m
= re
.search(br
'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
906 webpage_bytes[:1024])
908 encoding = m.group(1).decode('ascii')
909 elif webpage_bytes.startswith(b'\xff\xfe'):
916 def __check_blocked(self, content):
917 first_block = content[:512]
918 if ('<title>Access to this site is blocked</title>' in content
919 and 'Websense' in first_block):
920 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
921 blocked_iframe = self._html_search_regex(
922 r'<iframe src="([^
"]+)"', content,
923 'Websense information URL
', default=None)
925 msg += ' Visit
%s for more details
' % blocked_iframe
926 raise ExtractorError(msg, expected=True)
927 if '<title
>The URL you requested has been blocked
</title
>' in first_block:
929 'Access to this webpage has been blocked by Indian censorship
. '
930 'Use a VPN
or proxy
server (with --proxy
) to route around it
.')
931 block_msg = self._html_search_regex(
932 r'</h1
><p
>(.*?
)</p
>',
933 content, 'block message
', default=None)
935 msg += ' (Message
: "%s")' % block_msg.replace('\n', ' ')
936 raise ExtractorError(msg, expected=True)
937 if ('<title
>TTK
:: Доступ к ресурсу ограничен
</title
>' in content
938 and 'blocklist
.rkn
.gov
.ru
' in content):
939 raise ExtractorError(
940 'Access to this webpage has been blocked by decision of the Russian government
. '
941 'Visit http
://blocklist
.rkn
.gov
.ru
/ for a block reason
.',
944 def _request_dump_filename(self, url, video_id):
945 basen = f'{video_id}_{url}
'
946 trim_length = self.get_param('trim_file_name
') or 240
947 if len(basen) > trim_length:
948 h = '___
' + hashlib.md5(basen.encode('utf
-8')).hexdigest()
949 basen = basen[:trim_length - len(h)] + h
950 filename = sanitize_filename(f'{basen}
.dump
', restricted=True)
951 # Working around MAX_PATH limitation on Windows (see
952 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
953 if compat_os_name == 'nt
':
954 absfilepath = os.path.abspath(filename)
955 if len(absfilepath) > 259:
956 filename = fR'\\?\{absfilepath}
'
959 def __decode_webpage(self, webpage_bytes, encoding, headers):
961 encoding = self._guess_encoding_from_content(headers.get('Content
-Type
', ''), webpage_bytes)
963 return webpage_bytes.decode(encoding, 'replace
')
965 return webpage_bytes.decode('utf
-8', 'replace
')
967 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
968 webpage_bytes = urlh.read()
969 if prefix is not None:
970 webpage_bytes = prefix + webpage_bytes
971 if self.get_param('dump_intermediate_pages
', False):
972 self.to_screen('Dumping request to
' + urlh.geturl())
973 dump = base64.b64encode(webpage_bytes).decode('ascii
')
974 self._downloader.to_screen(dump)
975 if self.get_param('write_pages
'):
976 filename = self._request_dump_filename(urlh.geturl(), video_id)
977 self.to_screen(f'Saving request to {filename}
')
978 with open(filename, 'wb
') as outf:
979 outf.write(webpage_bytes)
981 content = self.__decode_webpage(webpage_bytes, encoding, urlh.headers)
982 self.__check_blocked(content)
986 def __print_error(self, errnote, fatal, video_id, err):
988 raise ExtractorError(f'{video_id}
: {errnote}
', cause=err)
990 self.report_warning(f'{video_id}
: {errnote}
: {err}
')
992 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True, errnote=None):
994 xml_string = transform_source(xml_string)
996 return compat_etree_fromstring(xml_string.encode('utf
-8'))
997 except xml.etree.ElementTree.ParseError as ve:
998 self.__print_error('Failed to parse XML
' if errnote is None else errnote, fatal, video_id, ve)
1000 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True, errnote=None, **parser_kwargs):
1003 json_string, cls=LenientJSONDecoder, strict=False, transform_source=transform_source, **parser_kwargs)
1004 except ValueError as ve:
1005 self.__print_error('Failed to parse JSON
' if errnote is None else errnote, fatal, video_id, ve)
1007 def _parse_socket_response_as_json(self, data, *args, **kwargs):
1008 return self._parse_json(data[data.find('{'):data.rfind('}
') + 1], *args, **kwargs)
1010 def __create_download_methods(name, parser, note, errnote, return_value):
1012 def parse(ie, content, *args, errnote=errnote, **kwargs):
1015 if errnote is False:
1016 kwargs['errnote
'] = errnote
1017 # parser is fetched by name so subclasses can override it
1018 return getattr(ie, parser)(content, *args, **kwargs)
1020 def download_handle(self, url_or_request, video_id, note=note, errnote=errnote, transform_source=None,
1021 fatal=True, encoding=None, data=None, headers={}, query={}, expected_status=None):
1022 res = self._download_webpage_handle(
1023 url_or_request, video_id, note=note, errnote=errnote, fatal=fatal, encoding=encoding,
1024 data=data, headers=headers, query=query, expected_status=expected_status)
1028 return parse(self, content, video_id, transform_source=transform_source, fatal=fatal, errnote=errnote), urlh
1030 def download_content(self, url_or_request, video_id, note=note, errnote=errnote, transform_source=None,
1031 fatal=True, encoding=None, data=None, headers={}, query={}, expected_status=None):
1032 if self.get_param('load_pages
'):
1033 url_or_request = self._create_request(url_or_request, data, headers, query)
1034 filename = self._request_dump_filename(url_or_request.full_url, video_id)
1035 self.to_screen(f'Loading request
from {filename}
')
1037 with open(filename, 'rb
') as dumpf:
1038 webpage_bytes = dumpf.read()
1039 except OSError as e:
1040 self.report_warning(f'Unable to load request
from disk
: {e}
')
1042 content = self.__decode_webpage(webpage_bytes, encoding, url_or_request.headers)
1043 return parse(self, content, video_id, transform_source=transform_source, fatal=fatal, errnote=errnote)
1047 'transform_source
': transform_source,
1049 'encoding
': encoding,
1053 'expected_status
': expected_status,
1056 kwargs.pop('transform_source
')
1057 # The method is fetched by name so subclasses can override _download_..._handle
1058 res = getattr(self, download_handle.__name__)(url_or_request, video_id, **kwargs)
1059 return res if res is False else res[0]
1061 def impersonate(func, name, return_value):
1062 func.__name__, func.__qualname__ = name, f'InfoExtractor
.{name}
'
1064 @param transform_source Apply this transformation before parsing
1065 @returns {return_value}
1067 See _download_webpage_handle docstring for other arguments specification
1070 impersonate(download_handle, f'_download_{name}_handle
', f'({return_value}
, URL handle
)')
1071 impersonate(download_content, f'_download_{name}
', f'{return_value}
')
1072 return download_handle, download_content
1074 _download_xml_handle, _download_xml = __create_download_methods(
1075 'xml
', '_parse_xml
', 'Downloading XML
', 'Unable to download XML
', 'xml
as an xml
.etree
.ElementTree
.Element
')
1076 _download_json_handle, _download_json = __create_download_methods(
1077 'json
', '_parse_json
', 'Downloading JSON metadata
', 'Unable to download JSON metadata
', 'JSON
object as a
dict')
1078 _download_socket_json_handle, _download_socket_json = __create_download_methods(
1079 'socket_json
', '_parse_socket_response_as_json
', 'Polling socket
', 'Unable to poll socket
', 'JSON
object as a
dict')
1080 __download_webpage = __create_download_methods('webpage
', None, None, None, 'data of the page
as a string
')[1]
1082 def _download_webpage(
1083 self, url_or_request, video_id, note=None, errnote=None,
1084 fatal=True, tries=1, timeout=NO_DEFAULT, *args, **kwargs):
1086 Return the data of the page as a string.
1089 tries -- number of tries
1090 timeout -- sleep interval between tries
1092 See _download_webpage_handle docstring for other arguments specification.
1095 R''' # NB: These are unused; should they be deprecated?
1097 self._downloader.deprecation_warning('tries argument
is deprecated
in InfoExtractor
._download
_webpage
')
1098 if timeout is NO_DEFAULT:
1101 self._downloader.deprecation_warning('timeout argument
is deprecated
in InfoExtractor
._download
_webpage
')
1107 return self.__download_webpage(url_or_request, video_id, note, errnote, None, fatal, *args, **kwargs)
1108 except http.client.IncompleteRead as e:
1110 if try_count >= tries:
1112 self._sleep(timeout, video_id)
1114 def report_warning(self, msg, video_id=None, *args, only_once=False, **kwargs):
1115 idstr = format_field(video_id, None, '%s: ')
1116 msg = f'[{self.IE_NAME}
] {idstr}{msg}
'
1118 if f'WARNING
: {msg}
' in self._printed_messages:
1120 self._printed_messages.add(f'WARNING
: {msg}
')
1121 self._downloader.report_warning(msg, *args, **kwargs)
1123 def to_screen(self, msg, *args, **kwargs):
1124 """Print msg to screen, prefixing it with '[ie_name
]'"""
1125 self._downloader.to_screen(f'[{self.IE_NAME}
] {msg}
', *args, **kwargs)
1127 def write_debug(self, msg, *args, **kwargs):
1128 self._downloader.write_debug(f'[{self.IE_NAME}
] {msg}
', *args, **kwargs)
1130 def get_param(self, name, default=None, *args, **kwargs):
1131 if self._downloader:
1132 return self._downloader.params.get(name, default, *args, **kwargs)
1135 def report_drm(self, video_id, partial=NO_DEFAULT):
1136 if partial is not NO_DEFAULT:
1137 self._downloader.deprecation_warning('InfoExtractor
.report_drm no longer accepts the argument partial
')
1138 self.raise_no_formats('This video
is DRM protected
', expected=True, video_id=video_id)
1140 def report_extraction(self, id_or_name):
1141 """Report information extraction."""
1142 self.to_screen('%s: Extracting information
' % id_or_name)
1144 def report_download_webpage(self, video_id):
1145 """Report webpage download."""
1146 self.to_screen('%s: Downloading webpage
' % video_id)
1148 def report_age_confirmation(self):
1149 """Report attempt to confirm age."""
1150 self.to_screen('Confirming age
')
1152 def report_login(self):
1153 """Report attempt to log in."""
1154 self.to_screen('Logging
in')
1156 def raise_login_required(
1157 self, msg='This video
is only available
for registered users
',
1158 metadata_available=False, method=NO_DEFAULT):
1159 if metadata_available and (
1160 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1161 self.report_warning(msg)
1163 msg += format_field(self._login_hint(method), None, '. %s')
1164 raise ExtractorError(msg, expected=True)
1166 def raise_geo_restricted(
1167 self, msg='This video
is not available
from your location due to geo restriction
',
1168 countries=None, metadata_available=False):
1169 if metadata_available and (
1170 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1171 self.report_warning(msg)
1173 raise GeoRestrictedError(msg, countries=countries)
1175 def raise_no_formats(self, msg, expected=False, video_id=None):
1177 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1178 self.report_warning(msg, video_id)
1179 elif isinstance(msg, ExtractorError):
1182 raise ExtractorError(msg, expected=expected, video_id=video_id)
1184 # Methods for following #608
1186 def url_result(url, ie=None, video_id=None, video_title=None, *, url_transparent=False, **kwargs):
1187 """Returns a URL that points to a page that should be processed"""
1189 kwargs['ie_key
'] = ie if isinstance(ie, str) else ie.ie_key()
1190 if video_id is not None:
1191 kwargs['id'] = video_id
1192 if video_title is not None:
1193 kwargs['title
'] = video_title
1196 '_type
': 'url_transparent
' if url_transparent else 'url
',
1201 def playlist_from_matches(cls, matches, playlist_id=None, playlist_title=None,
1202 getter=IDENTITY, ie=None, video_kwargs=None, **kwargs):
1203 return cls.playlist_result(
1204 (cls.url_result(m, ie, **(video_kwargs or {})) for m in orderedSet(map(getter, matches), lazy=True)),
1205 playlist_id, playlist_title, **kwargs)
1208 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None, *, multi_video=False, **kwargs):
1209 """Returns a playlist"""
1211 kwargs['id'] = playlist_id
1213 kwargs['title
'] = playlist_title
1214 if playlist_description is not None:
1215 kwargs['description
'] = playlist_description
1218 '_type
': 'multi_video
' if multi_video else 'playlist
',
1222 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1224 Perform a regex search on the given string, using a single or a list of
1225 patterns returning the first matching group.
1226 In case of failure return a default value or raise a WARNING or a
1227 RegexNotFoundError, depending on fatal, specifying the field name.
1231 elif isinstance(pattern, (str, re.Pattern)):
1232 mobj = re.search(pattern, string, flags)
1235 mobj = re.search(p, string, flags)
1239 _name = self._downloader._format_err(name, self._downloader.Styles.EMPHASIS)
1243 # return the first matching group
1244 return next(g for g in mobj.groups() if g is not None)
1245 elif isinstance(group, (list, tuple)):
1246 return tuple(mobj.group(g) for g in group)
1248 return mobj.group(group)
1249 elif default is not NO_DEFAULT:
1252 raise RegexNotFoundError('Unable to extract
%s' % _name)
1254 self.report_warning('unable to extract
%s' % _name + bug_reports_message())
1257 def _search_json(self, start_pattern, string, name, video_id, *, end_pattern='',
1258 contains_pattern=r'{(?s:.+)}
', fatal=True, default=NO_DEFAULT, **kwargs):
1259 """Searches string for the JSON object specified by start_pattern"""
1260 # NB: end_pattern is only used to reduce the size of the initial match
1261 if default is NO_DEFAULT:
1262 default, has_default = {}, False
1264 fatal, has_default = False, True
1266 json_string = self._search_regex(
1267 rf'(?
:{start_pattern}
)\s
*(?P
<json
>{contains_pattern}
)\s
*(?
:{end_pattern}
)',
1268 string, name, group='json
', fatal=fatal, default=None if has_default else NO_DEFAULT)
1272 _name = self._downloader._format_err(name, self._downloader.Styles.EMPHASIS)
1274 return self._parse_json(json_string, video_id, ignore_extra=True, **kwargs)
1275 except ExtractorError as e:
1277 raise ExtractorError(
1278 f'Unable to extract {_name}
- Failed to parse JSON
', cause=e.cause, video_id=video_id)
1279 elif not has_default:
1280 self.report_warning(
1281 f'Unable to extract {_name}
- Failed to parse JSON
: {e}
', video_id=video_id)
1284 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1286 Like _search_regex, but strips HTML tags and unescapes entities.
1288 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1289 if isinstance(res, tuple):
1290 return tuple(map(clean_html, res))
1291 return clean_html(res)
1293 def _get_netrc_login_info(self, netrc_machine=None):
1296 netrc_machine = netrc_machine or self._NETRC_MACHINE
1298 if self.get_param('usenetrc
', False):
1300 netrc_file = compat_expanduser(self.get_param('netrc_location
') or '~
')
1301 if os.path.isdir(netrc_file):
1302 netrc_file = os.path.join(netrc_file, '.netrc
')
1303 info = netrc.netrc(file=netrc_file).authenticators(netrc_machine)
1304 if info is not None:
1308 raise netrc.NetrcParseError(
1309 'No authenticators
for %s' % netrc_machine)
1310 except (OSError, netrc.NetrcParseError) as err:
1311 self.report_warning(
1312 'parsing
.netrc
: %s' % error_to_compat_str(err))
1314 return username, password
1316 def _get_login_info(self, username_option='username
', password_option='password
', netrc_machine=None):
1318 Get the login info as (username, password)
1319 First look for the manually specified credentials using username_option
1320 and password_option as keys in params dictionary. If no such credentials
1321 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1323 If there's no info available
, return (None, None)
1326 # Attempt to use provided username and password or .netrc data
1327 username = self.get_param(username_option)
1328 if username is not None:
1329 password = self.get_param(password_option)
1331 username, password = self._get_netrc_login_info(netrc_machine)
1333 return username, password
1335 def _get_tfa_info(self, note='two-factor verification code'):
1337 Get the two
-factor authentication info
1338 TODO
- asking the user will be required
for sms
/phone verify
1339 currently just uses the command line option
1340 If there
's no info available, return None
1343 tfa = self.get_param('twofactor
')
1347 return getpass.getpass('Type
%s and press
[Return
]: ' % note)
1349 # Helper functions for extracting OpenGraph info
1351 def _og_regexes(prop):
1352 content_re = r'content
=(?
:"([^"]+?
)"|\'([^\']+?)\'|\s*([^\s"\'=<>`
]+?
)(?
=\s|
/?
>))'
1353 property_re = (r'(?
:name|
property)=(?
:\'og
%(sep)s%(prop)s\'|
"og%(sep)s%(prop)s"|\s
*og
%(sep)s%(prop)s\b)'
1354 % {'prop': re.escape(prop), 'sep': '(?::|[:-])'})
1355 template = r'<meta
[^
>]+?
%s[^
>]+?
%s'
1357 template % (property_re, content_re),
1358 template % (content_re, property_re),
1362 def _meta_regex(prop):
1363 return r'''(?isx)<meta
1364 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1365 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1367 def _og_search_property(self, prop, html, name=None, **kargs):
1368 prop = variadic(prop)
1370 name = 'OpenGraph
%s' % prop[0]
1373 og_regexes.extend(self._og_regexes(p))
1374 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1377 return unescapeHTML(escaped)
1379 def _og_search_thumbnail(self, html, **kargs):
1380 return self._og_search_property('image
', html, 'thumbnail URL
', fatal=False, **kargs)
1382 def _og_search_description(self, html, **kargs):
1383 return self._og_search_property('description
', html, fatal=False, **kargs)
1385 def _og_search_title(self, html, *, fatal=False, **kargs):
1386 return self._og_search_property('title
', html, fatal=fatal, **kargs)
1388 def _og_search_video_url(self, html, name='video url
', secure=True, **kargs):
1389 regexes = self._og_regexes('video
') + self._og_regexes('video
:url
')
1391 regexes = self._og_regexes('video
:secure_url
') + regexes
1392 return self._html_search_regex(regexes, html, name, **kargs)
1394 def _og_search_url(self, html, **kargs):
1395 return self._og_search_property('url
', html, **kargs)
1397 def _html_extract_title(self, html, name='title
', *, fatal=False, **kwargs):
1398 return self._html_search_regex(r'(?s
)<title
\b[^
>]*>([^
<]+)</title
>', html, name, fatal=fatal, **kwargs)
1400 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1401 name = variadic(name)
1402 if display_name is None:
1403 display_name = name[0]
1404 return self._html_search_regex(
1405 [self._meta_regex(n) for n in name],
1406 html, display_name, fatal=fatal, group='content
', **kwargs)
1408 def _dc_search_uploader(self, html):
1409 return self._html_search_meta('dc
.creator
', html, 'uploader
')
1412 def _rta_search(html):
1413 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1414 if re.search(r'(?ix
)<meta\s
+name
="rating"\s
+'
1415 r' content
="RTA-5042-1996-1400-1577-RTA"',
1419 # And then there are the jokers who advertise that they use RTA, but actually don't
.
1420 AGE_LIMIT_MARKERS
= [
1421 r
'Proudly Labeled <a href="http://www\.rtalabel\.org/" title="Restricted to Adults">RTA</a>',
1422 r
'>[^<]*you acknowledge you are at least (\d+) years old',
1423 r
'>\s*(?:18\s+U(?:\.S\.C\.|SC)\s+)?(?:§+\s*)?2257\b',
1427 for marker
in AGE_LIMIT_MARKERS
:
1428 mobj
= re
.search(marker
, html
)
1430 age_limit
= max(age_limit
, int(traverse_obj(mobj
, 1, default
=18)))
1433 def _media_rating_search(self
, html
):
1434 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1435 rating
= self
._html
_search
_meta
('rating', html
)
1447 return RATING_TABLE
.get(rating
.lower())
1449 def _family_friendly_search(self
, html
):
1450 # See http://schema.org/VideoObject
1451 family_friendly
= self
._html
_search
_meta
(
1452 'isFamilyFriendly', html
, default
=None)
1454 if not family_friendly
:
1463 return RATING_TABLE
.get(family_friendly
.lower())
1465 def _twitter_search_player(self
, html
):
1466 return self
._html
_search
_meta
('twitter:player', html
,
1467 'twitter card player')
1469 def _yield_json_ld(self
, html
, video_id
, *, fatal
=True, default
=NO_DEFAULT
):
1470 """Yield all json ld objects in the html"""
1471 if default
is not NO_DEFAULT
:
1473 for mobj
in re
.finditer(JSON_LD_RE
, html
):
1474 json_ld_item
= self
._parse
_json
(mobj
.group('json_ld'), video_id
, fatal
=fatal
)
1475 for json_ld
in variadic(json_ld_item
):
1476 if isinstance(json_ld
, dict):
1479 def _search_json_ld(self
, html
, video_id
, expected_type
=None, *, fatal
=True, default
=NO_DEFAULT
):
1480 """Search for a video in any json ld in the html"""
1481 if default
is not NO_DEFAULT
:
1483 info
= self
._json
_ld
(
1484 list(self
._yield
_json
_ld
(html
, video_id
, fatal
=fatal
, default
=default
)),
1485 video_id
, fatal
=fatal
, expected_type
=expected_type
)
1488 if default
is not NO_DEFAULT
:
1491 raise RegexNotFoundError('Unable to extract JSON-LD')
1493 self
.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1496 def _json_ld(self
, json_ld
, video_id
, fatal
=True, expected_type
=None):
1497 if isinstance(json_ld
, str):
1498 json_ld
= self
._parse
_json
(json_ld
, video_id
, fatal
=fatal
)
1503 INTERACTION_TYPE_MAP
= {
1504 'CommentAction': 'comment',
1505 'AgreeAction': 'like',
1506 'DisagreeAction': 'dislike',
1507 'LikeAction': 'like',
1508 'DislikeAction': 'dislike',
1509 'ListenAction': 'view',
1510 'WatchAction': 'view',
1511 'ViewAction': 'view',
1514 def is_type(e
, *expected_types
):
1515 type = variadic(traverse_obj(e
, '@type'))
1516 return any(x
in type for x
in expected_types
)
1518 def extract_interaction_type(e
):
1519 interaction_type
= e
.get('interactionType')
1520 if isinstance(interaction_type
, dict):
1521 interaction_type
= interaction_type
.get('@type')
1522 return str_or_none(interaction_type
)
1524 def extract_interaction_statistic(e
):
1525 interaction_statistic
= e
.get('interactionStatistic')
1526 if isinstance(interaction_statistic
, dict):
1527 interaction_statistic
= [interaction_statistic
]
1528 if not isinstance(interaction_statistic
, list):
1530 for is_e
in interaction_statistic
:
1531 if not is_type(is_e
, 'InteractionCounter'):
1533 interaction_type
= extract_interaction_type(is_e
)
1534 if not interaction_type
:
1536 # For interaction count some sites provide string instead of
1537 # an integer (as per spec) with non digit characters (e.g. ",")
1538 # so extracting count with more relaxed str_to_int
1539 interaction_count
= str_to_int(is_e
.get('userInteractionCount'))
1540 if interaction_count
is None:
1542 count_kind
= INTERACTION_TYPE_MAP
.get(interaction_type
.split('/')[-1])
1545 count_key
= '%s_count' % count_kind
1546 if info
.get(count_key
) is not None:
1548 info
[count_key
] = interaction_count
1550 def extract_chapter_information(e
):
1552 'title': part
.get('name'),
1553 'start_time': part
.get('startOffset'),
1554 'end_time': part
.get('endOffset'),
1555 } for part
in variadic(e
.get('hasPart') or []) if part
.get('@type') == 'Clip']
1556 for idx
, (last_c
, current_c
, next_c
) in enumerate(zip(
1557 [{'end_time': 0}
] + chapters
, chapters
, chapters
[1:])):
1558 current_c
['end_time'] = current_c
['end_time'] or next_c
['start_time']
1559 current_c
['start_time'] = current_c
['start_time'] or last_c
['end_time']
1560 if None in current_c
.values():
1561 self
.report_warning(f
'Chapter {idx} contains broken data. Not extracting chapters')
1564 chapters
[-1]['end_time'] = chapters
[-1]['end_time'] or info
['duration']
1565 info
['chapters'] = chapters
1567 def extract_video_object(e
):
1568 author
= e
.get('author')
1570 'url': url_or_none(e
.get('contentUrl')),
1571 'ext': mimetype2ext(e
.get('encodingFormat')),
1572 'title': unescapeHTML(e
.get('name')),
1573 'description': unescapeHTML(e
.get('description')),
1574 'thumbnails': [{'url': unescapeHTML(url)}
1575 for url
in variadic(traverse_obj(e
, 'thumbnailUrl', 'thumbnailURL'))
1576 if url_or_none(url
)],
1577 'duration': parse_duration(e
.get('duration')),
1578 'timestamp': unified_timestamp(e
.get('uploadDate')),
1579 # author can be an instance of 'Organization' or 'Person' types.
1580 # both types can have 'name' property(inherited from 'Thing' type). [1]
1581 # however some websites are using 'Text' type instead.
1582 # 1. https://schema.org/VideoObject
1583 'uploader': author
.get('name') if isinstance(author
, dict) else author
if isinstance(author
, str) else None,
1584 'artist': traverse_obj(e
, ('byArtist', 'name'), expected_type
=str),
1585 'filesize': int_or_none(float_or_none(e
.get('contentSize'))),
1586 'tbr': int_or_none(e
.get('bitrate')),
1587 'width': int_or_none(e
.get('width')),
1588 'height': int_or_none(e
.get('height')),
1589 'view_count': int_or_none(e
.get('interactionCount')),
1590 'tags': try_call(lambda: e
.get('keywords').split(',')),
1592 if is_type(e
, 'AudioObject'):
1595 'abr': int_or_none(e
.get('bitrate')),
1597 extract_interaction_statistic(e
)
1598 extract_chapter_information(e
)
1600 def traverse_json_ld(json_ld
, at_top_level
=True):
1601 for e
in variadic(json_ld
):
1602 if not isinstance(e
, dict):
1604 if at_top_level
and '@context' not in e
:
1606 if at_top_level
and set(e
.keys()) == {'@context', '@graph'}
:
1607 traverse_json_ld(e
['@graph'], at_top_level
=False)
1609 if expected_type
is not None and not is_type(e
, expected_type
):
1611 rating
= traverse_obj(e
, ('aggregateRating', 'ratingValue'), expected_type
=float_or_none
)
1612 if rating
is not None:
1613 info
['average_rating'] = rating
1614 if is_type(e
, 'TVEpisode', 'Episode'):
1615 episode_name
= unescapeHTML(e
.get('name'))
1617 'episode': episode_name
,
1618 'episode_number': int_or_none(e
.get('episodeNumber')),
1619 'description': unescapeHTML(e
.get('description')),
1621 if not info
.get('title') and episode_name
:
1622 info
['title'] = episode_name
1623 part_of_season
= e
.get('partOfSeason')
1624 if is_type(part_of_season
, 'TVSeason', 'Season', 'CreativeWorkSeason'):
1626 'season': unescapeHTML(part_of_season
.get('name')),
1627 'season_number': int_or_none(part_of_season
.get('seasonNumber')),
1629 part_of_series
= e
.get('partOfSeries') or e
.get('partOfTVSeries')
1630 if is_type(part_of_series
, 'TVSeries', 'Series', 'CreativeWorkSeries'):
1631 info
['series'] = unescapeHTML(part_of_series
.get('name'))
1632 elif is_type(e
, 'Movie'):
1634 'title': unescapeHTML(e
.get('name')),
1635 'description': unescapeHTML(e
.get('description')),
1636 'duration': parse_duration(e
.get('duration')),
1637 'timestamp': unified_timestamp(e
.get('dateCreated')),
1639 elif is_type(e
, 'Article', 'NewsArticle'):
1641 'timestamp': parse_iso8601(e
.get('datePublished')),
1642 'title': unescapeHTML(e
.get('headline')),
1643 'description': unescapeHTML(e
.get('articleBody') or e
.get('description')),
1645 if is_type(traverse_obj(e
, ('video', 0)), 'VideoObject'):
1646 extract_video_object(e
['video'][0])
1647 elif is_type(traverse_obj(e
, ('subjectOf', 0)), 'VideoObject'):
1648 extract_video_object(e
['subjectOf'][0])
1649 elif is_type(e
, 'VideoObject', 'AudioObject'):
1650 extract_video_object(e
)
1651 if expected_type
is None:
1655 video
= e
.get('video')
1656 if is_type(video
, 'VideoObject'):
1657 extract_video_object(video
)
1658 if expected_type
is None:
1663 traverse_json_ld(json_ld
)
1664 return filter_dict(info
)
1666 def _search_nextjs_data(self
, webpage
, video_id
, *, transform_source
=None, fatal
=True, **kw
):
1667 return self
._parse
_json
(
1669 r
'(?s)<script[^>]+id=[\'"]__NEXT_DATA__[\'"][^
>]*>([^
<]+)</script
>',
1670 webpage, 'next
.js data
', fatal=fatal, **kw),
1671 video_id, transform_source=transform_source, fatal=fatal)
1673 def _search_nuxt_data(self, webpage, video_id, context_name='__NUXT__
', *, fatal=True, traverse=('data
', 0)):
1674 """Parses Nuxt.js metadata. This works as long as the function __NUXT__ invokes is a pure function"""
1675 rectx = re.escape(context_name)
1676 FUNCTION_RE = r'\
(function\
((?P
<arg_keys
>.*?
)\
){return\s+(?P<js>{.*?}
)\s
*;?\s
*}\
((?P
<arg_vals
>.*?
)\
)'
1677 js, arg_keys, arg_vals = self._search_regex(
1678 (rf'<script
>\s
*window\
.{rectx}
={FUNCTION_RE}\s
*\
)\s
*;?\s
*</script
>', rf'{rectx}\
(.*?{FUNCTION_RE}
'),
1679 webpage, context_name, group=('js
', 'arg_keys
', 'arg_vals
'),
1680 default=NO_DEFAULT if fatal else (None, None, None))
1684 args = dict(zip(arg_keys.split(','), map(json.dumps, self._parse_json(
1685 f'[{arg_vals}
]', video_id, transform_source=js_to_json, fatal=fatal) or ())))
1687 ret = self._parse_json(js, video_id, transform_source=functools.partial(js_to_json, vars=args), fatal=fatal)
1688 return traverse_obj(ret, traverse) or {}
1691 def _hidden_inputs(html):
1692 html = re.sub(r'<!--(?
:(?
!<!--).)*-->', '', html)
1694 for input in re.findall(r'(?i
)(<input[^
>]+>)', html):
1695 attrs = extract_attributes(input)
1698 if attrs.get('type') not in ('hidden
', 'submit
'):
1700 name = attrs.get('name
') or attrs.get('id')
1701 value = attrs.get('value
')
1702 if name and value is not None:
1703 hidden_inputs[name] = value
1704 return hidden_inputs
1706 def _form_hidden_inputs(self, form_id, html):
1707 form = self._search_regex(
1708 r'(?
is)<form
[^
>]+?
id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1709 html, '%s form' % form_id, group='form')
1710 return self._hidden_inputs(form)
1712 @classproperty(cache=True)
1713 def FormatSort(cls):
1714 class FormatSort(FormatSorter):
1715 def __init__(ie, *args, **kwargs):
1716 super().__init__(ie._downloader, *args, **kwargs)
1718 deprecation_warning(
1719 'yt_dlp.InfoExtractor.FormatSort is deprecated and may be removed in the future. '
1720 'Use yt_dlp.utils.FormatSorter instead')
1723 def _sort_formats(self, formats, field_preference=[]):
1724 if not field_preference:
1725 self._downloader.deprecation_warning(
1726 'yt_dlp.InfoExtractor._sort_formats is deprecated and is no longer required')
1728 self._downloader.deprecation_warning(
1729 'yt_dlp.InfoExtractor._sort_formats is deprecated and no longer works as expected. '
1730 'Return _format_sort_fields in the info_dict instead')
1732 formats[0]['__sort_fields'] = field_preference
1734 def _check_formats(self, formats, video_id):
1736 formats[:] = filter(
1737 lambda f: self._is_valid_url(
1739 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1743 def _remove_duplicate_formats(formats):
1747 if f['url'] not in format_urls:
1748 format_urls.add(f['url'])
1749 unique_formats.append(f)
1750 formats[:] = unique_formats
1752 def _is_valid_url(self, url, video_id, item='video', headers={}):
1753 url = self._proto_relative_url(url, scheme='http:')
1754 # For now assume non HTTP(S) URLs always valid
1755 if not (url.startswith('http://') or url.startswith('https://')):
1758 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1760 except ExtractorError as e:
1762 '%s: %s URL is invalid, skipping: %s'
1763 % (video_id, item, error_to_compat_str(e.cause)))
1766 def http_scheme(self):
1767 """ Either "http
:" or "https
:", depending on the user's preferences """
1770 if self.get_param('prefer_insecure', False)
1773 def _proto_relative_url(self, url, scheme=None):
1774 scheme = scheme or self.http_scheme()
1775 assert scheme.endswith(':')
1776 return sanitize_url(url, scheme=scheme[:-1])
1778 def _sleep(self, timeout, video_id, msg_template=None):
1779 if msg_template is None:
1780 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1781 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1785 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1786 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1787 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
1788 if self.get_param('ignore_no_formats_error'):
1791 res = self._download_xml_handle(
1792 manifest_url, video_id, 'Downloading f4m manifest',
1793 'Unable to download f4m manifest',
1794 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1795 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
1796 transform_source=transform_source,
1797 fatal=fatal, data=data, headers=headers, query=query)
1801 manifest, urlh = res
1802 manifest_url = urlh.geturl()
1804 return self._parse_f4m_formats(
1805 manifest, manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1806 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1808 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1809 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1810 fatal=True, m3u8_id=None):
1811 if not isinstance(manifest, xml.etree.ElementTree.Element) and not fatal:
1814 # currently yt-dlp cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1815 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1816 if akamai_pv is not None and ';' in akamai_pv.text:
1817 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1818 if playerVerificationChallenge.strip() != '':
1822 manifest_version = '1.0'
1823 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1825 manifest_version = '2.0'
1826 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1827 # Remove unsupported DRM protected media from final formats
1828 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
1829 media_nodes = remove_encrypted_media(media_nodes)
1833 manifest_base_url = get_base_url(manifest)
1835 bootstrap_info = xpath_element(
1836 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1837 'bootstrap info', default=None)
1840 mime_type = xpath_text(
1841 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1842 'base URL', default=None)
1843 if mime_type and mime_type.startswith('audio/'):
1846 for i, media_el in enumerate(media_nodes):
1847 tbr = int_or_none(media_el.attrib.get('bitrate'))
1848 width = int_or_none(media_el.attrib.get('width'))
1849 height = int_or_none(media_el.attrib.get('height'))
1850 format_id = join_nonempty(f4m_id, tbr or i)
1851 # If <bootstrapInfo> is present, the specified f4m is a
1852 # stream-level manifest, and only set-level manifests may refer to
1853 # external resources. See section 11.4 and section 4 of F4M spec
1854 if bootstrap_info is None:
1856 # @href is introduced in 2.0, see section 11.6 of F4M spec
1857 if manifest_version == '2.0':
1858 media_url = media_el.attrib.get('href')
1859 if media_url is None:
1860 media_url = media_el.attrib.get('url')
1864 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1865 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1866 # If media_url is itself a f4m manifest do the recursive extraction
1867 # since bitrates in parent manifest (this one) and media_url manifest
1868 # may differ leading to inability to resolve the format by requested
1869 # bitrate in f4m downloader
1870 ext = determine_ext(manifest_url)
1872 f4m_formats = self._extract_f4m_formats(
1873 manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1874 transform_source=transform_source, fatal=fatal)
1875 # Sometimes stream-level manifest contains single media entry that
1876 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1877 # At the same time parent's media entry in set-level manifest may
1878 # contain it. We will copy it from parent in such cases.
1879 if len(f4m_formats) == 1:
1882 'tbr': f.get('tbr') or tbr,
1883 'width': f.get('width') or width,
1884 'height': f.get('height') or height,
1885 'format_id': f.get('format_id') if not tbr else format_id,
1888 formats.extend(f4m_formats)
1891 formats.extend(self._extract_m3u8_formats(
1892 manifest_url, video_id, 'mp4', preference=preference,
1893 quality=quality, m3u8_id=m3u8_id, fatal=fatal))
1896 'format_id': format_id,
1897 'url': manifest_url,
1898 'manifest_url': manifest_url,
1899 'ext': 'flv' if bootstrap_info is not None else None,
1905 'preference': preference,
1910 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, quality=None, m3u8_id=None):
1912 'format_id': join_nonempty(m3u8_id, 'meta'),
1916 'preference': preference - 100 if preference else -100,
1918 'resolution': 'multiple',
1919 'format_note': 'Quality selection URL',
1922 def _report_ignoring_subs(self, name):
1923 self.report_warning(bug_reports_message(
1924 f'Ignoring subtitle tracks found in the {name} manifest; '
1925 'if any subtitle tracks are missing,'
1928 def _extract_m3u8_formats(self, *args, **kwargs):
1929 fmts, subs = self._extract_m3u8_formats_and_subtitles(*args, **kwargs)
1931 self._report_ignoring_subs('HLS')
1934 def _extract_m3u8_formats_and_subtitles(
1935 self, m3u8_url, video_id, ext=None, entry_protocol='m3u8_native',
1936 preference=None, quality=None, m3u8_id=None, note=None,
1937 errnote=None, fatal=True, live=False, data=None, headers={},
1940 if self.get_param('ignore_no_formats_error'):
1944 if errnote is not False:
1945 errnote = errnote or 'Failed to obtain m3u8 URL'
1947 raise ExtractorError(errnote, video_id=video_id)
1948 self.report_warning(f'{errnote}{bug_reports_message()}')
1951 res = self._download_webpage_handle(
1953 note='Downloading m3u8 information' if note is None else note,
1954 errnote='Failed to download m3u8 information' if errnote is None else errnote,
1955 fatal=fatal, data=data, headers=headers, query=query)
1960 m3u8_doc, urlh = res
1961 m3u8_url = urlh.geturl()
1963 return self._parse_m3u8_formats_and_subtitles(
1964 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
1965 preference=preference, quality=quality, m3u8_id=m3u8_id,
1966 note=note, errnote=errnote, fatal=fatal, live=live, data=data,
1967 headers=headers, query=query, video_id=video_id)
1969 def _parse_m3u8_formats_and_subtitles(
1970 self, m3u8_doc, m3u8_url=None, ext=None, entry_protocol='m3u8_native',
1971 preference=None, quality=None, m3u8_id=None, live=False, note=None,
1972 errnote=None, fatal=True, data=None, headers={}, query={},
1974 formats, subtitles = [], {}
1976 has_drm = re.search('|'.join([
1977 r'#EXT-X-FAXS-CM:', # Adobe Flash Access
1978 r'#EXT-X-(?:SESSION-)?KEY:.*?URI="skd
://', # Apple FairPlay
1981 def format_url(url):
1982 return url if re.match(r'^https?
://', url) else urllib.parse.urljoin(m3u8_url, url)
1984 if self.get_param('hls_split_discontinuity
', False):
1985 def _extract_m3u8_playlist_indices(manifest_url=None, m3u8_doc=None):
1987 if not manifest_url:
1989 m3u8_doc = self._download_webpage(
1990 manifest_url, video_id, fatal=fatal, data=data, headers=headers,
1991 note=False, errnote='Failed to download m3u8 playlist information
')
1992 if m3u8_doc is False:
1994 return range(1 + sum(line.startswith('#EXT-X-DISCONTINUITY') for line in m3u8_doc.splitlines()))
1997 def _extract_m3u8_playlist_indices(*args
, **kwargs
):
2001 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
2002 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
2003 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
2005 # We should try extracting formats only from master playlists [1, 4.3.4],
2006 # i.e. playlists that describe available qualities. On the other hand
2007 # media playlists [1, 4.3.3] should be returned as is since they contain
2008 # just the media without qualities renditions.
2009 # Fortunately, master playlist can be easily distinguished from media
2010 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
2011 # master playlist tags MUST NOT appear in a media playlist and vice versa.
2012 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
2013 # media playlist and MUST NOT appear in master playlist thus we can
2014 # clearly detect media playlist with this criterion.
2016 if '#EXT-X-TARGETDURATION' in m3u8_doc
: # media playlist, return as is
2018 'format_id': join_nonempty(m3u8_id
, idx
),
2019 'format_index': idx
,
2020 'url': m3u8_url
or encode_data_uri(m3u8_doc
.encode('utf-8'), 'application/x-mpegurl'),
2022 'protocol': entry_protocol
,
2023 'preference': preference
,
2026 } for idx
in _extract_m3u8_playlist_indices(m3u8_doc
=m3u8_doc
)]
2028 return formats
, subtitles
2031 last_stream_inf
= {}
2033 def extract_media(x_media_line
):
2034 media
= parse_m3u8_attributes(x_media_line
)
2035 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
2036 media_type
, group_id
, name
= media
.get('TYPE'), media
.get('GROUP-ID'), media
.get('NAME')
2037 if not (media_type
and group_id
and name
):
2039 groups
.setdefault(group_id
, []).append(media
)
2040 # <https://tools.ietf.org/html/rfc8216#section-4.3.4.1>
2041 if media_type
== 'SUBTITLES':
2042 # According to RFC 8216 §4.3.4.2.1, URI is REQUIRED in the
2043 # EXT-X-MEDIA tag if the media type is SUBTITLES.
2044 # However, lack of URI has been spotted in the wild.
2045 # e.g. NebulaIE; see https://github.com/yt-dlp/yt-dlp/issues/339
2046 if not media
.get('URI'):
2048 url
= format_url(media
['URI'])
2051 'ext': determine_ext(url
),
2053 if sub_info
['ext'] == 'm3u8':
2054 # Per RFC 8216 §3.1, the only possible subtitle format m3u8
2055 # files may contain is WebVTT:
2056 # <https://tools.ietf.org/html/rfc8216#section-3.1>
2057 sub_info
['ext'] = 'vtt'
2058 sub_info
['protocol'] = 'm3u8_native'
2059 lang
= media
.get('LANGUAGE') or 'und'
2060 subtitles
.setdefault(lang
, []).append(sub_info
)
2061 if media_type
not in ('VIDEO', 'AUDIO'):
2063 media_url
= media
.get('URI')
2065 manifest_url
= format_url(media_url
)
2067 'format_id': join_nonempty(m3u8_id
, group_id
, name
, idx
),
2068 'format_note': name
,
2069 'format_index': idx
,
2070 'url': manifest_url
,
2071 'manifest_url': m3u8_url
,
2072 'language': media
.get('LANGUAGE'),
2074 'protocol': entry_protocol
,
2075 'preference': preference
,
2078 'vcodec': 'none' if media_type
== 'AUDIO' else None,
2079 } for idx
in _extract_m3u8_playlist_indices(manifest_url
))
2081 def build_stream_name():
2082 # Despite specification does not mention NAME attribute for
2083 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
2084 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
2085 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
2086 stream_name
= last_stream_inf
.get('NAME')
2089 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
2090 # from corresponding rendition group
2091 stream_group_id
= last_stream_inf
.get('VIDEO')
2092 if not stream_group_id
:
2094 stream_group
= groups
.get(stream_group_id
)
2095 if not stream_group
:
2096 return stream_group_id
2097 rendition
= stream_group
[0]
2098 return rendition
.get('NAME') or stream_group_id
2100 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
2101 # chance to detect video only formats when EXT-X-STREAM-INF tags
2102 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
2103 for line
in m3u8_doc
.splitlines():
2104 if line
.startswith('#EXT-X-MEDIA:'):
2107 for line
in m3u8_doc
.splitlines():
2108 if line
.startswith('#EXT-X-STREAM-INF:'):
2109 last_stream_inf
= parse_m3u8_attributes(line
)
2110 elif line
.startswith('#') or not line
.strip():
2113 tbr
= float_or_none(
2114 last_stream_inf
.get('AVERAGE-BANDWIDTH')
2115 or last_stream_inf
.get('BANDWIDTH'), scale
=1000)
2116 manifest_url
= format_url(line
.strip())
2118 for idx
in _extract_m3u8_playlist_indices(manifest_url
):
2119 format_id
= [m3u8_id
, None, idx
]
2120 # Bandwidth of live streams may differ over time thus making
2121 # format_id unpredictable. So it's better to keep provided
2124 stream_name
= build_stream_name()
2125 format_id
[1] = stream_name
or '%d' % (tbr
or len(formats
))
2127 'format_id': join_nonempty(*format_id
),
2128 'format_index': idx
,
2129 'url': manifest_url
,
2130 'manifest_url': m3u8_url
,
2133 'fps': float_or_none(last_stream_inf
.get('FRAME-RATE')),
2134 'protocol': entry_protocol
,
2135 'preference': preference
,
2139 resolution
= last_stream_inf
.get('RESOLUTION')
2141 mobj
= re
.search(r
'(?P<width>\d+)[xX](?P<height>\d+)', resolution
)
2143 f
['width'] = int(mobj
.group('width'))
2144 f
['height'] = int(mobj
.group('height'))
2145 # Unified Streaming Platform
2147 r
'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f
['url'])
2149 abr
, vbr
= mobj
.groups()
2150 abr
, vbr
= float_or_none(abr
, 1000), float_or_none(vbr
, 1000)
2155 codecs
= parse_codecs(last_stream_inf
.get('CODECS'))
2157 audio_group_id
= last_stream_inf
.get('AUDIO')
2158 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
2159 # references a rendition group MUST have a CODECS attribute.
2160 # However, this is not always respected. E.g. [2]
2161 # contains EXT-X-STREAM-INF tag which references AUDIO
2162 # rendition group but does not have CODECS and despite
2163 # referencing an audio group it represents a complete
2164 # (with audio and video) format. So, for such cases we will
2165 # ignore references to rendition groups and treat them
2166 # as complete formats.
2167 if audio_group_id
and codecs
and f
.get('vcodec') != 'none':
2168 audio_group
= groups
.get(audio_group_id
)
2169 if audio_group
and audio_group
[0].get('URI'):
2170 # TODO: update acodec for audio only formats with
2172 f
['acodec'] = 'none'
2173 if not f
.get('ext'):
2174 f
['ext'] = 'm4a' if f
.get('vcodec') == 'none' else 'mp4'
2178 progressive_uri
= last_stream_inf
.get('PROGRESSIVE-URI')
2181 del http_f
['manifest_url']
2183 'format_id': f
['format_id'].replace('hls-', 'http-'),
2185 'url': progressive_uri
,
2187 formats
.append(http_f
)
2189 last_stream_inf
= {}
2190 return formats
, subtitles
2192 def _extract_m3u8_vod_duration(
2193 self
, m3u8_vod_url
, video_id
, note
=None, errnote
=None, data
=None, headers
={}, query={}
):
2195 m3u8_vod
= self
._download
_webpage
(
2196 m3u8_vod_url
, video_id
,
2197 note
='Downloading m3u8 VOD manifest' if note
is None else note
,
2198 errnote
='Failed to download VOD manifest' if errnote
is None else errnote
,
2199 fatal
=False, data
=data
, headers
=headers
, query
=query
)
2201 return self
._parse
_m
3u8_vod
_duration
(m3u8_vod
or '', video_id
)
2203 def _parse_m3u8_vod_duration(self
, m3u8_vod
, video_id
):
2204 if '#EXT-X-ENDLIST' not in m3u8_vod
:
2208 float(line
[len('#EXTINF:'):].split(',')[0])
2209 for line
in m3u8_vod
.splitlines() if line
.startswith('#EXTINF:'))) or None
2211 def _extract_mpd_vod_duration(
2212 self
, mpd_url
, video_id
, note
=None, errnote
=None, data
=None, headers
={}, query={}
):
2214 mpd_doc
= self
._download
_xml
(
2216 note
='Downloading MPD VOD manifest' if note
is None else note
,
2217 errnote
='Failed to download VOD manifest' if errnote
is None else errnote
,
2218 fatal
=False, data
=data
, headers
=headers
, query
=query
) or {}
2219 return int_or_none(parse_duration(mpd_doc
.get('mediaPresentationDuration')))
2222 def _xpath_ns(path
, namespace
=None):
2226 for c
in path
.split('/'):
2227 if not c
or c
== '.':
2230 out
.append('{%s}%s' % (namespace
, c
))
2231 return '/'.join(out
)
2233 def _extract_smil_formats_and_subtitles(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None, transform_source
=None):
2234 if self
.get_param('ignore_no_formats_error'):
2237 res
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
, transform_source
=transform_source
)
2243 smil_url
= urlh
.geturl()
2245 namespace
= self
._parse
_smil
_namespace
(smil
)
2247 fmts
= self
._parse
_smil
_formats
(
2248 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2249 subs
= self
._parse
_smil
_subtitles
(
2250 smil
, namespace
=namespace
)
2254 def _extract_smil_formats(self
, *args
, **kwargs
):
2255 fmts
, subs
= self
._extract
_smil
_formats
_and
_subtitles
(*args
, **kwargs
)
2257 self
._report
_ignoring
_subs
('SMIL')
2260 def _extract_smil_info(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None):
2261 res
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
)
2266 smil_url
= urlh
.geturl()
2268 return self
._parse
_smil
(smil
, smil_url
, video_id
, f4m_params
=f4m_params
)
2270 def _download_smil(self
, smil_url
, video_id
, fatal
=True, transform_source
=None):
2271 return self
._download
_xml
_handle
(
2272 smil_url
, video_id
, 'Downloading SMIL file',
2273 'Unable to download SMIL file', fatal
=fatal
, transform_source
=transform_source
)
2275 def _parse_smil(self
, smil
, smil_url
, video_id
, f4m_params
=None):
2276 namespace
= self
._parse
_smil
_namespace
(smil
)
2278 formats
= self
._parse
_smil
_formats
(
2279 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2280 subtitles
= self
._parse
_smil
_subtitles
(smil
, namespace
=namespace
)
2282 video_id
= os
.path
.splitext(url_basename(smil_url
))[0]
2286 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2287 name
= meta
.attrib
.get('name')
2288 content
= meta
.attrib
.get('content')
2289 if not name
or not content
:
2291 if not title
and name
== 'title':
2293 elif not description
and name
in ('description', 'abstract'):
2294 description
= content
2295 elif not upload_date
and name
== 'date':
2296 upload_date
= unified_strdate(content
)
2299 'id': image
.get('type'),
2300 'url': image
.get('src'),
2301 'width': int_or_none(image
.get('width')),
2302 'height': int_or_none(image
.get('height')),
2303 } for image
in smil
.findall(self
._xpath
_ns
('.//image', namespace
)) if image
.get('src')]
2307 'title': title
or video_id
,
2308 'description': description
,
2309 'upload_date': upload_date
,
2310 'thumbnails': thumbnails
,
2312 'subtitles': subtitles
,
2315 def _parse_smil_namespace(self
, smil
):
2316 return self
._search
_regex
(
2317 r
'(?i)^{([^}]+)?}smil$', smil
.tag
, 'namespace', default
=None)
2319 def _parse_smil_formats(self
, smil
, smil_url
, video_id
, namespace
=None, f4m_params
=None, transform_rtmp_url
=None):
2321 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2322 b
= meta
.get('base') or meta
.get('httpBase')
2334 media
= smil
.findall(self
._xpath
_ns
('.//video', namespace
)) + smil
.findall(self
._xpath
_ns
('.//audio', namespace
))
2335 for medium
in media
:
2336 src
= medium
.get('src')
2337 if not src
or src
in srcs
:
2341 bitrate
= float_or_none(medium
.get('system-bitrate') or medium
.get('systemBitrate'), 1000)
2342 filesize
= int_or_none(medium
.get('size') or medium
.get('fileSize'))
2343 width
= int_or_none(medium
.get('width'))
2344 height
= int_or_none(medium
.get('height'))
2345 proto
= medium
.get('proto')
2346 ext
= medium
.get('ext')
2347 src_ext
= determine_ext(src
, default_ext
=None) or ext
or urlhandle_detect_ext(
2348 self
._request
_webpage
(HEADRequest(src
), video_id
, note
='Requesting extension info', fatal
=False))
2349 streamer
= medium
.get('streamer') or base
2351 if proto
== 'rtmp' or streamer
.startswith('rtmp'):
2357 'format_id': 'rtmp-%d' % (rtmp_count
if bitrate
is None else bitrate
),
2359 'filesize': filesize
,
2363 if transform_rtmp_url
:
2364 streamer
, src
= transform_rtmp_url(streamer
, src
)
2365 formats
[-1].update({
2371 src_url
= src
if src
.startswith('http') else urllib
.parse
.urljoin(base
, src
)
2372 src_url
= src_url
.strip()
2374 if proto
== 'm3u8' or src_ext
== 'm3u8':
2375 m3u8_formats
= self
._extract
_m
3u8_formats
(
2376 src_url
, video_id
, ext
or 'mp4', m3u8_id
='hls', fatal
=False)
2377 if len(m3u8_formats
) == 1:
2379 m3u8_formats
[0].update({
2380 'format_id': 'hls-%d' % (m3u8_count
if bitrate
is None else bitrate
),
2385 formats
.extend(m3u8_formats
)
2386 elif src_ext
== 'f4m':
2391 'plugin': 'flowplayer-3.2.0.1',
2393 f4m_url
+= '&' if '?' in f4m_url
else '?'
2394 f4m_url
+= urllib
.parse
.urlencode(f4m_params
)
2395 formats
.extend(self
._extract
_f
4m
_formats
(f4m_url
, video_id
, f4m_id
='hds', fatal
=False))
2396 elif src_ext
== 'mpd':
2397 formats
.extend(self
._extract
_mpd
_formats
(
2398 src_url
, video_id
, mpd_id
='dash', fatal
=False))
2399 elif re
.search(r
'\.ism/[Mm]anifest', src_url
):
2400 formats
.extend(self
._extract
_ism
_formats
(
2401 src_url
, video_id
, ism_id
='mss', fatal
=False))
2402 elif src_url
.startswith('http') and self
._is
_valid
_url
(src
, video_id
):
2406 'ext': ext
or src_ext
or 'flv',
2407 'format_id': 'http-%d' % (bitrate
or http_count
),
2409 'filesize': filesize
,
2414 for medium
in smil
.findall(self
._xpath
_ns
('.//imagestream', namespace
)):
2415 src
= medium
.get('src')
2416 if not src
or src
in srcs
:
2422 'format_id': 'imagestream-%d' % (imgs_count
),
2424 'ext': mimetype2ext(medium
.get('type')),
2427 'width': int_or_none(medium
.get('width')),
2428 'height': int_or_none(medium
.get('height')),
2429 'format_note': 'SMIL storyboards',
2434 def _parse_smil_subtitles(self
, smil
, namespace
=None, subtitles_lang
='en'):
2437 for num
, textstream
in enumerate(smil
.findall(self
._xpath
_ns
('.//textstream', namespace
))):
2438 src
= textstream
.get('src')
2439 if not src
or src
in urls
:
2442 ext
= textstream
.get('ext') or mimetype2ext(textstream
.get('type')) or determine_ext(src
)
2443 lang
= textstream
.get('systemLanguage') or textstream
.get('systemLanguageName') or textstream
.get('lang') or subtitles_lang
2444 subtitles
.setdefault(lang
, []).append({
2450 def _extract_xspf_playlist(self
, xspf_url
, playlist_id
, fatal
=True):
2451 res
= self
._download
_xml
_handle
(
2452 xspf_url
, playlist_id
, 'Downloading xpsf playlist',
2453 'Unable to download xspf manifest', fatal
=fatal
)
2458 xspf_url
= urlh
.geturl()
2460 return self
._parse
_xspf
(
2461 xspf
, playlist_id
, xspf_url
=xspf_url
,
2462 xspf_base_url
=base_url(xspf_url
))
2464 def _parse_xspf(self
, xspf_doc
, playlist_id
, xspf_url
=None, xspf_base_url
=None):
2466 'xspf': 'http://xspf.org/ns/0/',
2467 's1': 'http://static.streamone.nl/player/ns/0',
2471 for track
in xspf_doc
.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP
)):
2473 track
, xpath_with_ns('./xspf:title', NS_MAP
), 'title', default
=playlist_id
)
2474 description
= xpath_text(
2475 track
, xpath_with_ns('./xspf:annotation', NS_MAP
), 'description')
2476 thumbnail
= xpath_text(
2477 track
, xpath_with_ns('./xspf:image', NS_MAP
), 'thumbnail')
2478 duration
= float_or_none(
2479 xpath_text(track
, xpath_with_ns('./xspf:duration', NS_MAP
), 'duration'), 1000)
2482 for location
in track
.findall(xpath_with_ns('./xspf:location', NS_MAP
)):
2483 format_url
= urljoin(xspf_base_url
, location
.text
)
2488 'manifest_url': xspf_url
,
2489 'format_id': location
.get(xpath_with_ns('s1:label', NS_MAP
)),
2490 'width': int_or_none(location
.get(xpath_with_ns('s1:width', NS_MAP
))),
2491 'height': int_or_none(location
.get(xpath_with_ns('s1:height', NS_MAP
))),
2497 'description': description
,
2498 'thumbnail': thumbnail
,
2499 'duration': duration
,
2504 def _extract_mpd_formats(self
, *args
, **kwargs
):
2505 fmts
, subs
= self
._extract
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2507 self
._report
_ignoring
_subs
('DASH')
2510 def _extract_mpd_formats_and_subtitles(
2511 self
, mpd_url
, video_id
, mpd_id
=None, note
=None, errnote
=None,
2512 fatal
=True, data
=None, headers
={}, query={}
):
2514 if self
.get_param('ignore_no_formats_error'):
2517 res
= self
._download
_xml
_handle
(
2519 note
='Downloading MPD manifest' if note
is None else note
,
2520 errnote
='Failed to download MPD manifest' if errnote
is None else errnote
,
2521 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
2528 # We could have been redirected to a new url when we retrieved our mpd file.
2529 mpd_url
= urlh
.geturl()
2530 mpd_base_url
= base_url(mpd_url
)
2532 return self
._parse
_mpd
_formats
_and
_subtitles
(
2533 mpd_doc
, mpd_id
, mpd_base_url
, mpd_url
)
2535 def _parse_mpd_formats(self
, *args
, **kwargs
):
2536 fmts
, subs
= self
._parse
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2538 self
._report
_ignoring
_subs
('DASH')
2541 def _parse_mpd_formats_and_subtitles(
2542 self
, mpd_doc
, mpd_id
=None, mpd_base_url
='', mpd_url
=None):
2544 Parse formats from MPD manifest.
2546 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
2547 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
2548 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
2550 if not self
.get_param('dynamic_mpd', True):
2551 if mpd_doc
.get('type') == 'dynamic':
2554 namespace
= self
._search
_regex
(r
'(?i)^{([^}]+)?}MPD$', mpd_doc
.tag
, 'namespace', default
=None)
2557 return self
._xpath
_ns
(path
, namespace
)
2559 def is_drm_protected(element
):
2560 return element
.find(_add_ns('ContentProtection')) is not None
2562 def extract_multisegment_info(element
, ms_parent_info
):
2563 ms_info
= ms_parent_info
.copy()
2565 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2566 # common attributes and elements. We will only extract relevant
2568 def extract_common(source
):
2569 segment_timeline
= source
.find(_add_ns('SegmentTimeline'))
2570 if segment_timeline
is not None:
2571 s_e
= segment_timeline
.findall(_add_ns('S'))
2573 ms_info
['total_number'] = 0
2576 r
= int(s
.get('r', 0))
2577 ms_info
['total_number'] += 1 + r
2578 ms_info
['s'].append({
2579 't': int(s
.get('t', 0)),
2580 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2581 'd': int(s
.attrib
['d']),
2584 start_number
= source
.get('startNumber')
2586 ms_info
['start_number'] = int(start_number
)
2587 timescale
= source
.get('timescale')
2589 ms_info
['timescale'] = int(timescale
)
2590 segment_duration
= source
.get('duration')
2591 if segment_duration
:
2592 ms_info
['segment_duration'] = float(segment_duration
)
2594 def extract_Initialization(source
):
2595 initialization
= source
.find(_add_ns('Initialization'))
2596 if initialization
is not None:
2597 ms_info
['initialization_url'] = initialization
.attrib
['sourceURL']
2599 segment_list
= element
.find(_add_ns('SegmentList'))
2600 if segment_list
is not None:
2601 extract_common(segment_list
)
2602 extract_Initialization(segment_list
)
2603 segment_urls_e
= segment_list
.findall(_add_ns('SegmentURL'))
2605 ms_info
['segment_urls'] = [segment
.attrib
['media'] for segment
in segment_urls_e
]
2607 segment_template
= element
.find(_add_ns('SegmentTemplate'))
2608 if segment_template
is not None:
2609 extract_common(segment_template
)
2610 media
= segment_template
.get('media')
2612 ms_info
['media'] = media
2613 initialization
= segment_template
.get('initialization')
2615 ms_info
['initialization'] = initialization
2617 extract_Initialization(segment_template
)
2620 mpd_duration
= parse_duration(mpd_doc
.get('mediaPresentationDuration'))
2621 formats
, subtitles
= [], {}
2622 stream_numbers
= collections
.defaultdict(int)
2623 for period
in mpd_doc
.findall(_add_ns('Period')):
2624 period_duration
= parse_duration(period
.get('duration')) or mpd_duration
2625 period_ms_info
= extract_multisegment_info(period
, {
2629 for adaptation_set
in period
.findall(_add_ns('AdaptationSet')):
2630 adaption_set_ms_info
= extract_multisegment_info(adaptation_set
, period_ms_info
)
2631 for representation
in adaptation_set
.findall(_add_ns('Representation')):
2632 representation_attrib
= adaptation_set
.attrib
.copy()
2633 representation_attrib
.update(representation
.attrib
)
2634 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2635 mime_type
= representation_attrib
['mimeType']
2636 content_type
= representation_attrib
.get('contentType', mime_type
.split('/')[0])
2638 codec_str
= representation_attrib
.get('codecs', '')
2639 # Some kind of binary subtitle found in some youtube livestreams
2640 if mime_type
== 'application/x-rawcc':
2641 codecs
= {'scodec': codec_str}
2643 codecs
= parse_codecs(codec_str
)
2644 if content_type
not in ('video', 'audio', 'text'):
2645 if mime_type
== 'image/jpeg':
2646 content_type
= mime_type
2647 elif codecs
.get('vcodec', 'none') != 'none':
2648 content_type
= 'video'
2649 elif codecs
.get('acodec', 'none') != 'none':
2650 content_type
= 'audio'
2651 elif codecs
.get('scodec', 'none') != 'none':
2652 content_type
= 'text'
2653 elif mimetype2ext(mime_type
) in ('tt', 'dfxp', 'ttml', 'xml', 'json'):
2654 content_type
= 'text'
2656 self
.report_warning('Unknown MIME type %s in DASH manifest' % mime_type
)
2660 for element
in (representation
, adaptation_set
, period
, mpd_doc
):
2661 base_url_e
= element
.find(_add_ns('BaseURL'))
2662 if try_call(lambda: base_url_e
.text
) is not None:
2663 base_url
= base_url_e
.text
+ base_url
2664 if re
.match(r
'^https?://', base_url
):
2666 if mpd_base_url
and base_url
.startswith('/'):
2667 base_url
= urllib
.parse
.urljoin(mpd_base_url
, base_url
)
2668 elif mpd_base_url
and not re
.match(r
'^https?://', base_url
):
2669 if not mpd_base_url
.endswith('/'):
2671 base_url
= mpd_base_url
+ base_url
2672 representation_id
= representation_attrib
.get('id')
2673 lang
= representation_attrib
.get('lang')
2674 url_el
= representation
.find(_add_ns('BaseURL'))
2675 filesize
= int_or_none(url_el
.attrib
.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el
is not None else None)
2676 bandwidth
= int_or_none(representation_attrib
.get('bandwidth'))
2677 if representation_id
is not None:
2678 format_id
= representation_id
2680 format_id
= content_type
2682 format_id
= mpd_id
+ '-' + format_id
2683 if content_type
in ('video', 'audio'):
2685 'format_id': format_id
,
2686 'manifest_url': mpd_url
,
2687 'ext': mimetype2ext(mime_type
),
2688 'width': int_or_none(representation_attrib
.get('width')),
2689 'height': int_or_none(representation_attrib
.get('height')),
2690 'tbr': float_or_none(bandwidth
, 1000),
2691 'asr': int_or_none(representation_attrib
.get('audioSamplingRate')),
2692 'fps': int_or_none(representation_attrib
.get('frameRate')),
2693 'language': lang
if lang
not in ('mul', 'und', 'zxx', 'mis') else None,
2694 'format_note': 'DASH %s' % content_type
,
2695 'filesize': filesize
,
2696 'container': mimetype2ext(mime_type
) + '_dash',
2699 elif content_type
== 'text':
2701 'ext': mimetype2ext(mime_type
),
2702 'manifest_url': mpd_url
,
2703 'filesize': filesize
,
2705 elif content_type
== 'image/jpeg':
2706 # See test case in VikiIE
2707 # https://www.viki.com/videos/1175236v-choosing-spouse-by-lottery-episode-1
2709 'format_id': format_id
,
2711 'manifest_url': mpd_url
,
2712 'format_note': 'DASH storyboards (jpeg)',
2716 if is_drm_protected(adaptation_set
) or is_drm_protected(representation
):
2718 representation_ms_info
= extract_multisegment_info(representation
, adaption_set_ms_info
)
2720 def prepare_template(template_name
, identifiers
):
2721 tmpl
= representation_ms_info
[template_name
]
2722 if representation_id
is not None:
2723 tmpl
= tmpl
.replace('$RepresentationID$', representation_id
)
2724 # First of, % characters outside $...$ templates
2725 # must be escaped by doubling for proper processing
2726 # by % operator string formatting used further (see
2727 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2733 in_template
= not in_template
2734 elif c
== '%' and not in_template
:
2736 # Next, $...$ templates are translated to their
2737 # %(...) counterparts to be used with % operator
2738 t
= re
.sub(r
'\$(%s)\$' % '|'.join(identifiers
), r
'%(\1)d', t
)
2739 t
= re
.sub(r
'\$(%s)%%([^$]+)\$' % '|'.join(identifiers
), r
'%(\1)\2', t
)
2740 t
.replace('$$', '$')
2743 # @initialization is a regular template like @media one
2744 # so it should be handled just the same way (see
2745 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2746 if 'initialization' in representation_ms_info
:
2747 initialization_template
= prepare_template(
2749 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2750 # $Time$ shall not be included for @initialization thus
2751 # only $Bandwidth$ remains
2753 representation_ms_info
['initialization_url'] = initialization_template
% {
2754 'Bandwidth': bandwidth
,
2757 def location_key(location
):
2758 return 'url' if re
.match(r
'^https?://', location
) else 'path'
2760 if 'segment_urls' not in representation_ms_info
and 'media' in representation_ms_info
:
2762 media_template
= prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2763 media_location_key
= location_key(media_template
)
2765 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2766 # can't be used at the same time
2767 if '%(Number' in media_template
and 's' not in representation_ms_info
:
2768 segment_duration
= None
2769 if 'total_number' not in representation_ms_info
and 'segment_duration' in representation_ms_info
:
2770 segment_duration
= float_or_none(representation_ms_info
['segment_duration'], representation_ms_info
['timescale'])
2771 representation_ms_info
['total_number'] = int(math
.ceil(
2772 float_or_none(period_duration
, segment_duration
, default
=0)))
2773 representation_ms_info
['fragments'] = [{
2774 media_location_key
: media_template
% {
2775 'Number': segment_number
,
2776 'Bandwidth': bandwidth
,
2778 'duration': segment_duration
,
2779 } for segment_number
in range(
2780 representation_ms_info
['start_number'],
2781 representation_ms_info
['total_number'] + representation_ms_info
['start_number'])]
2783 # $Number*$ or $Time$ in media template with S list available
2784 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2785 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2786 representation_ms_info
['fragments'] = []
2789 segment_number
= representation_ms_info
['start_number']
2791 def add_segment_url():
2792 segment_url
= media_template
% {
2793 'Time': segment_time
,
2794 'Bandwidth': bandwidth
,
2795 'Number': segment_number
,
2797 representation_ms_info
['fragments'].append({
2798 media_location_key
: segment_url
,
2799 'duration': float_or_none(segment_d
, representation_ms_info
['timescale']),
2802 for num
, s
in enumerate(representation_ms_info
['s']):
2803 segment_time
= s
.get('t') or segment_time
2807 for r
in range(s
.get('r', 0)):
2808 segment_time
+= segment_d
2811 segment_time
+= segment_d
2812 elif 'segment_urls' in representation_ms_info
and 's' in representation_ms_info
:
2813 # No media template,
2814 # e.g. https://www.youtube.com/watch?v=iXZV5uAYMJI
2815 # or any YouTube dashsegments video
2818 timescale
= representation_ms_info
['timescale']
2819 for s
in representation_ms_info
['s']:
2820 duration
= float_or_none(s
['d'], timescale
)
2821 for r
in range(s
.get('r', 0) + 1):
2822 segment_uri
= representation_ms_info
['segment_urls'][segment_index
]
2824 location_key(segment_uri
): segment_uri
,
2825 'duration': duration
,
2828 representation_ms_info
['fragments'] = fragments
2829 elif 'segment_urls' in representation_ms_info
:
2830 # Segment URLs with no SegmentTimeline
2831 # E.g. https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2832 # https://github.com/ytdl-org/youtube-dl/pull/14844
2834 segment_duration
= float_or_none(
2835 representation_ms_info
['segment_duration'],
2836 representation_ms_info
['timescale']) if 'segment_duration' in representation_ms_info
else None
2837 for segment_url
in representation_ms_info
['segment_urls']:
2839 location_key(segment_url
): segment_url
,
2841 if segment_duration
:
2842 fragment
['duration'] = segment_duration
2843 fragments
.append(fragment
)
2844 representation_ms_info
['fragments'] = fragments
2845 # If there is a fragments key available then we correctly recognized fragmented media.
2846 # Otherwise we will assume unfragmented media with direct access. Technically, such
2847 # assumption is not necessarily correct since we may simply have no support for
2848 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2849 if 'fragments' in representation_ms_info
:
2851 # NB: mpd_url may be empty when MPD manifest is parsed from a string
2852 'url': mpd_url
or base_url
,
2853 'fragment_base_url': base_url
,
2855 'protocol': 'http_dash_segments' if mime_type
!= 'image/jpeg' else 'mhtml',
2857 if 'initialization_url' in representation_ms_info
:
2858 initialization_url
= representation_ms_info
['initialization_url']
2859 if not f
.get('url'):
2860 f
['url'] = initialization_url
2861 f
['fragments'].append({location_key(initialization_url): initialization_url}
)
2862 f
['fragments'].extend(representation_ms_info
['fragments'])
2863 if not period_duration
:
2864 period_duration
= try_get(
2865 representation_ms_info
,
2866 lambda r
: sum(frag
['duration'] for frag
in r
['fragments']), float)
2868 # Assuming direct URL to unfragmented media.
2870 if content_type
in ('video', 'audio', 'image/jpeg'):
2871 f
['manifest_stream_number'] = stream_numbers
[f
['url']]
2872 stream_numbers
[f
['url']] += 1
2874 elif content_type
== 'text':
2875 subtitles
.setdefault(lang
or 'und', []).append(f
)
2877 return formats
, subtitles
2879 def _extract_ism_formats(self
, *args
, **kwargs
):
2880 fmts
, subs
= self
._extract
_ism
_formats
_and
_subtitles
(*args
, **kwargs
)
2882 self
._report
_ignoring
_subs
('ISM')
2885 def _extract_ism_formats_and_subtitles(self
, ism_url
, video_id
, ism_id
=None, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
):
2886 if self
.get_param('ignore_no_formats_error'):
2889 res
= self
._download
_xml
_handle
(
2891 note
='Downloading ISM manifest' if note
is None else note
,
2892 errnote
='Failed to download ISM manifest' if errnote
is None else errnote
,
2893 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
2900 return self
._parse
_ism
_formats
_and
_subtitles
(ism_doc
, urlh
.geturl(), ism_id
)
2902 def _parse_ism_formats_and_subtitles(self
, ism_doc
, ism_url
, ism_id
=None):
2904 Parse formats from ISM manifest.
2906 1. [MS-SSTR]: Smooth Streaming Protocol,
2907 https://msdn.microsoft.com/en-us/library/ff469518.aspx
2909 if ism_doc
.get('IsLive') == 'TRUE':
2912 duration
= int(ism_doc
.attrib
['Duration'])
2913 timescale
= int_or_none(ism_doc
.get('TimeScale')) or 10000000
2917 for stream
in ism_doc
.findall('StreamIndex'):
2918 stream_type
= stream
.get('Type')
2919 if stream_type
not in ('video', 'audio', 'text'):
2921 url_pattern
= stream
.attrib
['Url']
2922 stream_timescale
= int_or_none(stream
.get('TimeScale')) or timescale
2923 stream_name
= stream
.get('Name')
2924 stream_language
= stream
.get('Language', 'und')
2925 for track
in stream
.findall('QualityLevel'):
2926 KNOWN_TAGS
= {'255': 'AACL', '65534': 'EC-3'}
2927 fourcc
= track
.get('FourCC') or KNOWN_TAGS
.get(track
.get('AudioTag'))
2928 # TODO: add support for WVC1 and WMAP
2929 if fourcc
not in ('H264', 'AVC1', 'AACL', 'TTML', 'EC-3'):
2930 self
.report_warning('%s is not a supported codec' % fourcc
)
2932 tbr
= int(track
.attrib
['Bitrate']) // 1000
2933 # [1] does not mention Width and Height attributes. However,
2934 # they're often present while MaxWidth and MaxHeight are
2935 # missing, so should be used as fallbacks
2936 width
= int_or_none(track
.get('MaxWidth') or track
.get('Width'))
2937 height
= int_or_none(track
.get('MaxHeight') or track
.get('Height'))
2938 sampling_rate
= int_or_none(track
.get('SamplingRate'))
2940 track_url_pattern
= re
.sub(r
'{[Bb]itrate}', track
.attrib
['Bitrate'], url_pattern
)
2941 track_url_pattern
= urllib
.parse
.urljoin(ism_url
, track_url_pattern
)
2947 stream_fragments
= stream
.findall('c')
2948 for stream_fragment_index
, stream_fragment
in enumerate(stream_fragments
):
2949 fragment_ctx
['time'] = int_or_none(stream_fragment
.get('t')) or fragment_ctx
['time']
2950 fragment_repeat
= int_or_none(stream_fragment
.get('r')) or 1
2951 fragment_ctx
['duration'] = int_or_none(stream_fragment
.get('d'))
2952 if not fragment_ctx
['duration']:
2954 next_fragment_time
= int(stream_fragment
[stream_fragment_index
+ 1].attrib
['t'])
2956 next_fragment_time
= duration
2957 fragment_ctx
['duration'] = (next_fragment_time
- fragment_ctx
['time']) / fragment_repeat
2958 for _
in range(fragment_repeat
):
2960 'url': re
.sub(r
'{start[ _]time}', str(fragment_ctx
['time']), track_url_pattern
),
2961 'duration': fragment_ctx
['duration'] / stream_timescale
,
2963 fragment_ctx
['time'] += fragment_ctx
['duration']
2965 if stream_type
== 'text':
2966 subtitles
.setdefault(stream_language
, []).append({
2970 'manifest_url': ism_url
,
2971 'fragments': fragments
,
2972 '_download_params': {
2973 'stream_type': stream_type
,
2974 'duration': duration
,
2975 'timescale': stream_timescale
,
2977 'language': stream_language
,
2978 'codec_private_data': track
.get('CodecPrivateData'),
2981 elif stream_type
in ('video', 'audio'):
2983 'format_id': join_nonempty(ism_id
, stream_name
, tbr
),
2985 'manifest_url': ism_url
,
2986 'ext': 'ismv' if stream_type
== 'video' else 'isma',
2990 'asr': sampling_rate
,
2991 'vcodec': 'none' if stream_type
== 'audio' else fourcc
,
2992 'acodec': 'none' if stream_type
== 'video' else fourcc
,
2994 'fragments': fragments
,
2995 'has_drm': ism_doc
.find('Protection') is not None,
2996 'language': stream_language
,
2997 'audio_channels': int_or_none(track
.get('Channels')),
2998 '_download_params': {
2999 'stream_type': stream_type
,
3000 'duration': duration
,
3001 'timescale': stream_timescale
,
3002 'width': width
or 0,
3003 'height': height
or 0,
3005 'language': stream_language
,
3006 'codec_private_data': track
.get('CodecPrivateData'),
3007 'sampling_rate': sampling_rate
,
3008 'channels': int_or_none(track
.get('Channels', 2)),
3009 'bits_per_sample': int_or_none(track
.get('BitsPerSample', 16)),
3010 'nal_unit_length_field': int_or_none(track
.get('NALUnitLengthField', 4)),
3013 return formats
, subtitles
3015 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):
3016 def absolute_url(item_url
):
3017 return urljoin(base_url
, item_url
)
3019 def parse_content_type(content_type
):
3020 if not content_type
:
3022 ctr
= re
.search(r
'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type
)
3024 mimetype
, codecs
= ctr
.groups()
3025 f
= parse_codecs(codecs
)
3026 f
['ext'] = mimetype2ext(mimetype
)
3030 def _media_formats(src
, cur_media_type
, type_info
=None):
3031 type_info
= type_info
or {}
3032 full_url
= absolute_url(src
)
3033 ext
= type_info
.get('ext') or determine_ext(full_url
)
3035 is_plain_url
= False
3036 formats
= self
._extract
_m
3u8_formats
(
3037 full_url
, video_id
, ext
='mp4',
3038 entry_protocol
=m3u8_entry_protocol
, m3u8_id
=m3u8_id
,
3039 preference
=preference
, quality
=quality
, fatal
=False)
3041 is_plain_url
= False
3042 formats
= self
._extract
_mpd
_formats
(
3043 full_url
, video_id
, mpd_id
=mpd_id
, fatal
=False)
3048 'vcodec': 'none' if cur_media_type
== 'audio' else None,
3051 return is_plain_url
, formats
3054 # amp-video and amp-audio are very similar to their HTML5 counterparts
3055 # so we will include them right here (see
3056 # https://www.ampproject.org/docs/reference/components/amp-video)
3057 # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/
3058 _MEDIA_TAG_NAME_RE
= r
'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)'
3059 media_tags
= [(media_tag
, media_tag_name
, media_type
, '')
3060 for media_tag
, media_tag_name
, media_type
3061 in re
.findall(r
'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE
, webpage
)]
3062 media_tags
.extend(re
.findall(
3063 # We only allow video|audio followed by a whitespace or '>'.
3064 # Allowing more characters may end up in significant slow down (see
3065 # https://github.com/ytdl-org/youtube-dl/issues/11979,
3066 # e.g. http://www.porntrex.com/maps/videositemap.xml).
3067 r
'(?s)(<(?P<tag>%s)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>' % _MEDIA_TAG_NAME_RE
, webpage
))
3068 for media_tag
, _
, media_type
, media_content
in media_tags
:
3073 media_attributes
= extract_attributes(media_tag
)
3074 src
= strip_or_none(dict_get(media_attributes
, ('src', 'data-video-src', 'data-src', 'data-source')))
3076 f
= parse_content_type(media_attributes
.get('type'))
3077 _
, formats
= _media_formats(src
, media_type
, f
)
3078 media_info
['formats'].extend(formats
)
3079 media_info
['thumbnail'] = absolute_url(media_attributes
.get('poster'))
3081 for source_tag
in re
.findall(r
'<source[^>]+>', media_content
):
3082 s_attr
= extract_attributes(source_tag
)
3083 # data-video-src and data-src are non standard but seen
3084 # several times in the wild
3085 src
= strip_or_none(dict_get(s_attr
, ('src', 'data-video-src', 'data-src', 'data-source')))
3088 f
= parse_content_type(s_attr
.get('type'))
3089 is_plain_url
, formats
= _media_formats(src
, media_type
, f
)
3091 # width, height, res, label and title attributes are
3092 # all not standard but seen several times in the wild
3095 for lbl
in ('label', 'title')
3096 if str_or_none(s_attr
.get(lbl
))
3098 width
= int_or_none(s_attr
.get('width'))
3099 height
= (int_or_none(s_attr
.get('height'))
3100 or int_or_none(s_attr
.get('res')))
3101 if not width
or not height
:
3103 resolution
= parse_resolution(lbl
)
3106 width
= width
or resolution
.get('width')
3107 height
= height
or resolution
.get('height')
3109 tbr
= parse_bitrate(lbl
)
3118 'format_id': s_attr
.get('label') or s_attr
.get('title'),
3120 f
.update(formats
[0])
3121 media_info
['formats'].append(f
)
3123 media_info
['formats'].extend(formats
)
3124 for track_tag
in re
.findall(r
'<track[^>]+>', media_content
):
3125 track_attributes
= extract_attributes(track_tag
)
3126 kind
= track_attributes
.get('kind')
3127 if not kind
or kind
in ('subtitles', 'captions'):
3128 src
= strip_or_none(track_attributes
.get('src'))
3131 lang
= track_attributes
.get('srclang') or track_attributes
.get('lang') or track_attributes
.get('label')
3132 media_info
['subtitles'].setdefault(lang
, []).append({
3133 'url': absolute_url(src
),
3135 for f
in media_info
['formats']:
3136 f
.setdefault('http_headers', {})['Referer'] = base_url
3137 if media_info
['formats'] or media_info
['subtitles']:
3138 entries
.append(media_info
)
3141 def _extract_akamai_formats(self
, *args
, **kwargs
):
3142 fmts
, subs
= self
._extract
_akamai
_formats
_and
_subtitles
(*args
, **kwargs
)
3144 self
._report
_ignoring
_subs
('akamai')
3147 def _extract_akamai_formats_and_subtitles(self
, manifest_url
, video_id
, hosts
={}):
3148 signed
= 'hdnea=' in manifest_url
3150 # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
3151 manifest_url
= re
.sub(
3152 r
'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
3153 '', manifest_url
).strip('?')
3158 hdcore_sign
= 'hdcore=3.7.0'
3159 f4m_url
= re
.sub(r
'(https?://[^/]+)/i/', r
'\1/z/', manifest_url
).replace('/master.m3u8', '/manifest.f4m')
3160 hds_host
= hosts
.get('hds')
3162 f4m_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hds_host
, f4m_url
)
3163 if 'hdcore=' not in f4m_url
:
3164 f4m_url
+= ('&' if '?' in f4m_url
else '?') + hdcore_sign
3165 f4m_formats
= self
._extract
_f
4m
_formats
(
3166 f4m_url
, video_id
, f4m_id
='hds', fatal
=False)
3167 for entry
in f4m_formats
:
3168 entry
.update({'extra_param_to_segment_url': hdcore_sign}
)
3169 formats
.extend(f4m_formats
)
3171 m3u8_url
= re
.sub(r
'(https?://[^/]+)/z/', r
'\1/i/', manifest_url
).replace('/manifest.f4m', '/master.m3u8')
3172 hls_host
= hosts
.get('hls')
3174 m3u8_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hls_host
, m3u8_url
)
3175 m3u8_formats
, m3u8_subtitles
= self
._extract
_m
3u8_formats
_and
_subtitles
(
3176 m3u8_url
, video_id
, 'mp4', 'm3u8_native',
3177 m3u8_id
='hls', fatal
=False)
3178 formats
.extend(m3u8_formats
)
3179 subtitles
= self
._merge
_subtitles
(subtitles
, m3u8_subtitles
)
3181 http_host
= hosts
.get('http')
3182 if http_host
and m3u8_formats
and not signed
:
3183 REPL_REGEX
= r
'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
3184 qualities
= re
.match(REPL_REGEX
, m3u8_url
).group(2).split(',')
3185 qualities_length
= len(qualities
)
3186 if len(m3u8_formats
) in (qualities_length
, qualities_length
+ 1):
3188 for f
in m3u8_formats
:
3189 if f
['vcodec'] != 'none':
3190 for protocol
in ('http', 'https'):
3192 del http_f
['manifest_url']
3194 REPL_REGEX
, protocol
+ fr
'://{http_host}/\g<1>{qualities[i]}\3', f
['url'])
3196 'format_id': http_f
['format_id'].replace('hls-', protocol
+ '-'),
3198 'protocol': protocol
,
3200 formats
.append(http_f
)
3203 return formats
, subtitles
3205 def _extract_wowza_formats(self
, url
, video_id
, m3u8_entry_protocol
='m3u8_native', skip_protocols
=[]):
3206 query
= urllib
.parse
.urlparse(url
).query
3207 url
= re
.sub(r
'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url
)
3209 r
'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url
)
3210 url_base
= mobj
.group('url')
3211 http_base_url
= '%s%s:%s' % ('http', mobj
.group('s') or '', url_base
)
3214 def manifest_url(manifest
):
3215 m_url
= f
'{http_base_url}/{manifest}'
3217 m_url
+= '?%s' % query
3220 if 'm3u8' not in skip_protocols
:
3221 formats
.extend(self
._extract
_m
3u8_formats
(
3222 manifest_url('playlist.m3u8'), video_id
, 'mp4',
3223 m3u8_entry_protocol
, m3u8_id
='hls', fatal
=False))
3224 if 'f4m' not in skip_protocols
:
3225 formats
.extend(self
._extract
_f
4m
_formats
(
3226 manifest_url('manifest.f4m'),
3227 video_id
, f4m_id
='hds', fatal
=False))
3228 if 'dash' not in skip_protocols
:
3229 formats
.extend(self
._extract
_mpd
_formats
(
3230 manifest_url('manifest.mpd'),
3231 video_id
, mpd_id
='dash', fatal
=False))
3232 if re
.search(r
'(?:/smil:|\.smil)', url_base
):
3233 if 'smil' not in skip_protocols
:
3234 rtmp_formats
= self
._extract
_smil
_formats
(
3235 manifest_url('jwplayer.smil'),
3236 video_id
, fatal
=False)
3237 for rtmp_format
in rtmp_formats
:
3238 rtsp_format
= rtmp_format
.copy()
3239 rtsp_format
['url'] = '%s/%s' % (rtmp_format
['url'], rtmp_format
['play_path'])
3240 del rtsp_format
['play_path']
3241 del rtsp_format
['ext']
3242 rtsp_format
.update({
3243 'url': rtsp_format
['url'].replace('rtmp://', 'rtsp://'),
3244 'format_id': rtmp_format
['format_id'].replace('rtmp', 'rtsp'),
3247 formats
.extend([rtmp_format
, rtsp_format
])
3249 for protocol
in ('rtmp', 'rtsp'):
3250 if protocol
not in skip_protocols
:
3252 'url': f
'{protocol}:{url_base}',
3253 'format_id': protocol
,
3254 'protocol': protocol
,
3258 def _find_jwplayer_data(self
, webpage
, video_id
=None, transform_source
=js_to_json
):
3260 r
'''(?s)jwplayer\s*\(\s*(?P<q>'|")(?!(?P=q)).+(?P=q)\s*\)(?!</script>).*?\.\s*setup\s*\(\s*(?P<options>(?:\([^)]*\)|[^)])+)\s*\)''',
3264 jwplayer_data
= self
._parse
_json
(mobj
.group('options'),
3266 transform_source
=transform_source
)
3267 except ExtractorError
:
3270 if isinstance(jwplayer_data
, dict):
3271 return jwplayer_data
3273 def _extract_jwplayer_data(self
, webpage
, video_id
, *args
, **kwargs
):
3274 jwplayer_data
= self
._find
_jwplayer
_data
(
3275 webpage
, video_id
, transform_source
=js_to_json
)
3276 return self
._parse
_jwplayer
_data
(
3277 jwplayer_data
, video_id
, *args
, **kwargs
)
3279 def _parse_jwplayer_data(self
, jwplayer_data
, video_id
=None, require_title
=True,
3280 m3u8_id
=None, mpd_id
=None, rtmp_params
=None, base_url
=None):
3282 if not isinstance(jwplayer_data
, dict):
3285 playlist_items
= jwplayer_data
.get('playlist')
3286 # JWPlayer backward compatibility: single playlist item/flattened playlists
3287 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
3288 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
3289 if not isinstance(playlist_items
, list):
3290 playlist_items
= (playlist_items
or jwplayer_data
, )
3292 for video_data
in playlist_items
:
3293 if not isinstance(video_data
, dict):
3295 # JWPlayer backward compatibility: flattened sources
3296 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
3297 if 'sources' not in video_data
:
3298 video_data
['sources'] = [video_data
]
3300 this_video_id
= video_id
or video_data
['mediaid']
3302 formats
= self
._parse
_jwplayer
_formats
(
3303 video_data
['sources'], video_id
=this_video_id
, m3u8_id
=m3u8_id
,
3304 mpd_id
=mpd_id
, rtmp_params
=rtmp_params
, base_url
=base_url
)
3307 tracks
= video_data
.get('tracks')
3308 if tracks
and isinstance(tracks
, list):
3309 for track
in tracks
:
3310 if not isinstance(track
, dict):
3312 track_kind
= track
.get('kind')
3313 if not track_kind
or not isinstance(track_kind
, str):
3315 if track_kind
.lower() not in ('captions', 'subtitles'):
3317 track_url
= urljoin(base_url
, track
.get('file'))
3320 subtitles
.setdefault(track
.get('label') or 'en', []).append({
3321 'url': self
._proto
_relative
_url
(track_url
)
3325 'id': this_video_id
,
3326 'title': unescapeHTML(video_data
['title'] if require_title
else video_data
.get('title')),
3327 'description': clean_html(video_data
.get('description')),
3328 'thumbnail': urljoin(base_url
, self
._proto
_relative
_url
(video_data
.get('image'))),
3329 'timestamp': int_or_none(video_data
.get('pubdate')),
3330 'duration': float_or_none(jwplayer_data
.get('duration') or video_data
.get('duration')),
3331 'subtitles': subtitles
,
3332 'alt_title': clean_html(video_data
.get('subtitle')), # attributes used e.g. by Tele5 ...
3333 'genre': clean_html(video_data
.get('genre')),
3334 'channel': clean_html(dict_get(video_data
, ('category', 'channel'))),
3335 'season_number': int_or_none(video_data
.get('season')),
3336 'episode_number': int_or_none(video_data
.get('episode')),
3337 'release_year': int_or_none(video_data
.get('releasedate')),
3338 'age_limit': int_or_none(video_data
.get('age_restriction')),
3340 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
3341 if len(formats
) == 1 and re
.search(r
'^(?:http|//).*(?:youtube\.com|youtu\.be)/.+', formats
[0]['url']):
3343 '_type': 'url_transparent',
3344 'url': formats
[0]['url'],
3347 entry
['formats'] = formats
3348 entries
.append(entry
)
3349 if len(entries
) == 1:
3352 return self
.playlist_result(entries
)
3354 def _parse_jwplayer_formats(self
, jwplayer_sources_data
, video_id
=None,
3355 m3u8_id
=None, mpd_id
=None, rtmp_params
=None, base_url
=None):
3358 for source
in jwplayer_sources_data
:
3359 if not isinstance(source
, dict):
3361 source_url
= urljoin(
3362 base_url
, self
._proto
_relative
_url
(source
.get('file')))
3363 if not source_url
or source_url
in urls
:
3365 urls
.add(source_url
)
3366 source_type
= source
.get('type') or ''
3367 ext
= mimetype2ext(source_type
) or determine_ext(source_url
)
3368 if source_type
== 'hls' or ext
== 'm3u8' or 'format=m3u8-aapl' in source_url
:
3369 formats
.extend(self
._extract
_m
3u8_formats
(
3370 source_url
, video_id
, 'mp4', entry_protocol
='m3u8_native',
3371 m3u8_id
=m3u8_id
, fatal
=False))
3372 elif source_type
== 'dash' or ext
== 'mpd' or 'format=mpd-time-csf' in source_url
:
3373 formats
.extend(self
._extract
_mpd
_formats
(
3374 source_url
, video_id
, mpd_id
=mpd_id
, fatal
=False))
3376 formats
.extend(self
._extract
_smil
_formats
(
3377 source_url
, video_id
, fatal
=False))
3378 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3379 elif source_type
.startswith('audio') or ext
in (
3380 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
3387 format_id
= str_or_none(source
.get('label'))
3388 height
= int_or_none(source
.get('height'))
3389 if height
is None and format_id
:
3390 # Often no height is provided but there is a label in
3391 # format like "1080p", "720p SD", or 1080.
3392 height
= parse_resolution(format_id
).get('height')
3395 'width': int_or_none(source
.get('width')),
3397 'tbr': int_or_none(source
.get('bitrate'), scale
=1000),
3398 'filesize': int_or_none(source
.get('filesize')),
3400 'format_id': format_id
3402 if source_url
.startswith('rtmp'):
3403 a_format
['ext'] = 'flv'
3404 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3405 # of jwplayer.flash.swf
3406 rtmp_url_parts
= re
.split(
3407 r
'((?:mp4|mp3|flv):)', source_url
, 1)
3408 if len(rtmp_url_parts
) == 3:
3409 rtmp_url
, prefix
, play_path
= rtmp_url_parts
3412 'play_path': prefix
+ play_path
,
3415 a_format
.update(rtmp_params
)
3416 formats
.append(a_format
)
3419 def _live_title(self
, name
):
3420 self
._downloader
.deprecation_warning('yt_dlp.InfoExtractor._live_title is deprecated and does not work as expected')
3423 def _int(self
, v
, name
, fatal
=False, **kwargs
):
3424 res
= int_or_none(v
, **kwargs
)
3426 msg
= f
'Failed to extract {name}: Could not parse value {v!r}'
3428 raise ExtractorError(msg
)
3430 self
.report_warning(msg
)
3433 def _float(self
, v
, name
, fatal
=False, **kwargs
):
3434 res
= float_or_none(v
, **kwargs
)
3436 msg
= f
'Failed to extract {name}: Could not parse value {v!r}'
3438 raise ExtractorError(msg
)
3440 self
.report_warning(msg
)
3443 def _set_cookie(self
, domain
, name
, value
, expire_time
=None, port
=None,
3444 path
='/', secure
=False, discard
=False, rest
={}, **kwargs
):
3445 cookie
= http
.cookiejar
.Cookie(
3446 0, name
, value
, port
, port
is not None, domain
, True,
3447 domain
.startswith('.'), path
, True, secure
, expire_time
,
3448 discard
, None, None, rest
)
3449 self
.cookiejar
.set_cookie(cookie
)
3451 def _get_cookies(self
, url
):
3452 """ Return a http.cookies.SimpleCookie with the cookies for the url """
3453 return LenientSimpleCookie(self
._downloader
.cookiejar
.get_cookie_header(url
))
3455 def _apply_first_set_cookie_header(self
, url_handle
, cookie
):
3457 Apply first Set-Cookie header instead of the last. Experimental.
3459 Some sites (e.g. [1-3]) may serve two cookies under the same name
3460 in Set-Cookie header and expect the first (old) one to be set rather
3461 than second (new). However, as of RFC6265 the newer one cookie
3462 should be set into cookie store what actually happens.
3463 We will workaround this issue by resetting the cookie to
3464 the first one manually.
3465 1. https://new.vk.com/
3466 2. https://github.com/ytdl-org/youtube-dl/issues/9841#issuecomment-227871201
3467 3. https://learning.oreilly.com/
3469 for header
, cookies
in url_handle
.headers
.items():
3470 if header
.lower() != 'set-cookie':
3472 cookies
= cookies
.encode('iso-8859-1').decode('utf-8')
3473 cookie_value
= re
.search(
3474 r
'%s=(.+?);.*?\b[Dd]omain=(.+?)(?:[,;]|$)' % cookie
, cookies
)
3476 value
, domain
= cookie_value
.groups()
3477 self
._set
_cookie
(domain
, cookie
, value
)
3481 def get_testcases(cls
, include_onlymatching
=False):
3482 # Do not look in super classes
3483 t
= vars(cls
).get('_TEST')
3485 assert not hasattr(cls
, '_TESTS'), f
'{cls.ie_key()}IE has _TEST and _TESTS'
3488 tests
= vars(cls
).get('_TESTS', [])
3490 if not include_onlymatching
and t
.get('only_matching', False):
3492 t
['name'] = cls
.ie_key()
3494 if getattr(cls
, '__wrapped__', None):
3495 yield from cls
.__wrapped
__.get_testcases(include_onlymatching
)
3498 def get_webpage_testcases(cls
):
3499 tests
= vars(cls
).get('_WEBPAGE_TESTS', [])
3501 t
['name'] = cls
.ie_key()
3503 if getattr(cls
, '__wrapped__', None):
3504 yield from cls
.__wrapped
__.get_webpage_testcases()
3506 @classproperty(cache
=True)
3508 """Get age limit from the testcases"""
3509 return max(traverse_obj(
3510 (*cls
.get_testcases(include_onlymatching
=False), *cls
.get_webpage_testcases()),
3511 (..., (('playlist', 0), None), 'info_dict', 'age_limit')) or [0])
3513 @classproperty(cache
=True)
3514 def _RETURN_TYPE(cls
):
3515 """What the extractor returns: "video", "playlist", "any", or None (Unknown)"""
3516 tests
= tuple(cls
.get_testcases(include_onlymatching
=False))
3519 elif not any(k
.startswith('playlist') for test
in tests
for k
in test
):
3521 elif all(any(k
.startswith('playlist') for k
in test
) for test
in tests
):
3526 def is_single_video(cls
, url
):
3527 """Returns whether the URL is of a single video, None if unknown"""
3528 if cls
.suitable(url
):
3529 return {'video': True, 'playlist': False}
.get(cls
._RETURN
_TYPE
)
3532 def is_suitable(cls
, age_limit
):
3533 """Test whether the extractor is generally suitable for the given age limit"""
3534 return not age_restricted(cls
.age_limit
, age_limit
)
3537 def description(cls
, *, markdown
=True, search_examples
=None):
3538 """Description of the extractor"""
3540 if cls
._NETRC
_MACHINE
:
3542 desc
+= f
' [*{cls._NETRC_MACHINE}*](## "netrc machine")'
3544 desc
+= f
' [{cls._NETRC_MACHINE}]'
3545 if cls
.IE_DESC
is False:
3548 desc
+= f
' {cls.IE_DESC}'
3550 desc
+= f
'{";" if cls.IE_DESC else ""} "{cls.SEARCH_KEY}:" prefix'
3552 _COUNTS
= ('', '5', '10', 'all')
3553 desc
+= f
' (e.g. "{cls.SEARCH_KEY}{random.choice(_COUNTS)}:{random.choice(search_examples)}")'
3554 if not cls
.working():
3555 desc
+= ' (**Currently broken**)' if markdown
else ' (Currently broken)'
3557 # Escape emojis. Ref: https://github.com/github/markup/issues/1153
3558 name
= (' - **%s**' % re
.sub(r
':(\w+:)', ':\u200B\\g<1>', cls
.IE_NAME
)) if markdown
else cls
.IE_NAME
3559 return f
'{name}:{desc}' if desc
else name
3561 def extract_subtitles(self
, *args
, **kwargs
):
3562 if (self
.get_param('writesubtitles', False)
3563 or self
.get_param('listsubtitles')):
3564 return self
._get
_subtitles
(*args
, **kwargs
)
3567 def _get_subtitles(self
, *args
, **kwargs
):
3568 raise NotImplementedError('This method must be implemented by subclasses')
3570 class CommentsDisabled(Exception):
3571 """Raise in _get_comments if comments are disabled for the video"""
3573 def extract_comments(self
, *args
, **kwargs
):
3574 if not self
.get_param('getcomments'):
3576 generator
= self
._get
_comments
(*args
, **kwargs
)
3583 comments
.append(next(generator
))
3584 except StopIteration:
3586 except KeyboardInterrupt:
3587 self
.to_screen('Interrupted by user')
3588 except self
.CommentsDisabled
:
3589 return {'comments': None, 'comment_count': None}
3590 except Exception as e
:
3591 if self
.get_param('ignoreerrors') is not True:
3593 self
._downloader
.report_error(e
)
3594 comment_count
= len(comments
)
3595 self
.to_screen(f
'Extracted {comment_count} comments')
3597 'comments': comments
,
3598 'comment_count': None if interrupted
else comment_count
3602 def _get_comments(self
, *args
, **kwargs
):
3603 raise NotImplementedError('This method must be implemented by subclasses')
3606 def _merge_subtitle_items(subtitle_list1
, subtitle_list2
):
3607 """ Merge subtitle items for one language. Items with duplicated URLs/data
3608 will be dropped. """
3609 list1_data
= {(item.get('url'), item.get('data')) for item in subtitle_list1}
3610 ret
= list(subtitle_list1
)
3611 ret
.extend(item
for item
in subtitle_list2
if (item
.get('url'), item
.get('data')) not in list1_data
)
3615 def _merge_subtitles(cls
, *dicts
, target
=None):
3616 """ Merge subtitle dictionaries, language by language. """
3620 for lang
, subs
in d
.items():
3621 target
[lang
] = cls
._merge
_subtitle
_items
(target
.get(lang
, []), subs
)
3624 def extract_automatic_captions(self
, *args
, **kwargs
):
3625 if (self
.get_param('writeautomaticsub', False)
3626 or self
.get_param('listsubtitles')):
3627 return self
._get
_automatic
_captions
(*args
, **kwargs
)
3630 def _get_automatic_captions(self
, *args
, **kwargs
):
3631 raise NotImplementedError('This method must be implemented by subclasses')
3633 @functools.cached_property
3634 def _cookies_passed(self
):
3635 """Whether cookies have been passed to YoutubeDL"""
3636 return self
.get_param('cookiefile') is not None or self
.get_param('cookiesfrombrowser') is not None
3638 def mark_watched(self
, *args
, **kwargs
):
3639 if not self
.get_param('mark_watched', False):
3641 if self
.supports_login() and self
._get
_login
_info
()[0] is not None or self
._cookies
_passed
:
3642 self
._mark
_watched
(*args
, **kwargs
)
3644 def _mark_watched(self
, *args
, **kwargs
):
3645 raise NotImplementedError('This method must be implemented by subclasses')
3647 def geo_verification_headers(self
):
3649 geo_verification_proxy
= self
.get_param('geo_verification_proxy')
3650 if geo_verification_proxy
:
3651 headers
['Ytdl-request-proxy'] = geo_verification_proxy
3655 def _generic_id(url
):
3656 return urllib
.parse
.unquote(os
.path
.splitext(url
.rstrip('/').split('/')[-1])[0])
3658 def _generic_title(self
, url
='', webpage
='', *, default
=None):
3659 return (self
._og
_search
_title
(webpage
, default
=None)
3660 or self
._html
_extract
_title
(webpage
, default
=None)
3661 or urllib
.parse
.unquote(os
.path
.splitext(url_basename(url
))[0])
3664 def _extract_chapters_helper(self
, chapter_list
, start_function
, title_function
, duration
, strict
=True):
3668 'start_time': start_function(chapter
),
3669 'title': title_function(chapter
),
3670 } for chapter
in chapter_list
or []]
3672 warn
= self
.report_warning
3674 warn
= self
.write_debug
3675 chapter_list
.sort(key
=lambda c
: c
['start_time'] or 0)
3677 chapters
= [{'start_time': 0}
]
3678 for idx
, chapter
in enumerate(chapter_list
):
3679 if chapter
['start_time'] is None:
3680 warn(f
'Incomplete chapter {idx}')
3681 elif chapters
[-1]['start_time'] <= chapter
['start_time'] <= duration
:
3682 chapters
.append(chapter
)
3683 elif chapter
not in chapters
:
3684 issue
= (f
'{chapter["start_time"]} > {duration}' if chapter
['start_time'] > duration
3685 else f
'{chapter["start_time"]} < {chapters[-1]["start_time"]}')
3686 warn(f
'Invalid start time ({issue}) for chapter "{chapter["title"]}"')
3689 def _extract_chapters_from_description(self
, description
, duration
):
3690 duration_re
= r
'(?:\d+:)?\d{1,2}:\d{2}'
3691 sep_re
= r
'(?m)^\s*(%s)\b\W*\s(%s)\s*$'
3692 return self
._extract
_chapters
_helper
(
3693 re
.findall(sep_re
% (duration_re
, r
'.+?'), description
or ''),
3694 start_function
=lambda x
: parse_duration(x
[0]), title_function
=lambda x
: x
[1],
3695 duration
=duration
, strict
=False) or self
._extract
_chapters
_helper
(
3696 re
.findall(sep_re
% (r
'.+?', duration_re
), description
or ''),
3697 start_function
=lambda x
: parse_duration(x
[1]), title_function
=lambda x
: x
[0],
3698 duration
=duration
, strict
=False)
3701 def _availability(is_private
=None, needs_premium
=None, needs_subscription
=None, needs_auth
=None, is_unlisted
=None):
3702 all_known
= all(map(
3703 lambda x
: x
is not None,
3704 (is_private
, needs_premium
, needs_subscription
, needs_auth
, is_unlisted
)))
3706 'private' if is_private
3707 else 'premium_only' if needs_premium
3708 else 'subscriber_only' if needs_subscription
3709 else 'needs_auth' if needs_auth
3710 else 'unlisted' if is_unlisted
3711 else 'public' if all_known
3714 def _configuration_arg(self
, key
, default
=NO_DEFAULT
, *, ie_key
=None, casesense
=False):
3716 @returns A list of values for the extractor argument given by "key"
3717 or "default" if no such key is present
3718 @param default The default value to return when the key is not present (default: [])
3719 @param casesense When false, the values are converted to lower case
3721 ie_key
= ie_key
if isinstance(ie_key
, str) else (ie_key
or self
).ie_key()
3722 val
= traverse_obj(self
._downloader
.params
, ('extractor_args', ie_key
.lower(), key
))
3724 return [] if default
is NO_DEFAULT
else default
3725 return list(val
) if casesense
else [x
.lower() for x
in val
]
3727 def _yes_playlist(self
, playlist_id
, video_id
, smuggled_data
=None, *, playlist_label
='playlist', video_label
='video'):
3728 if not playlist_id
or not video_id
:
3731 no_playlist
= (smuggled_data
or {}).get('force_noplaylist')
3732 if no_playlist
is not None:
3733 return not no_playlist
3735 video_id
= '' if video_id
is True else f
' {video_id}'
3736 playlist_id
= '' if playlist_id
is True else f
' {playlist_id}'
3737 if self
.get_param('noplaylist'):
3738 self
.to_screen(f
'Downloading just the {video_label}{video_id} because of --no-playlist')
3740 self
.to_screen(f
'Downloading {playlist_label}{playlist_id} - add --no-playlist to download just the {video_label}{video_id}')
3743 def _error_or_warning(self
, err
, _count
=None, _retries
=0, *, fatal
=True):
3744 RetryManager
.report_retry(
3745 err
, _count
or int(fatal
), _retries
,
3746 info
=self
.to_screen
, warn
=self
.report_warning
, error
=None if fatal
else self
.report_warning
,
3747 sleep_func
=self
.get_param('retry_sleep_functions', {}).get('extractor'))
3749 def RetryManager(self
, **kwargs
):
3750 return RetryManager(self
.get_param('extractor_retries', 3), self
._error
_or
_warning
, **kwargs
)
3752 def _extract_generic_embeds(self
, url
, *args
, info_dict
={}, note
='Extracting generic embeds', **kwargs
):
3753 display_id
= traverse_obj(info_dict
, 'display_id', 'id')
3754 self
.to_screen(f
'{format_field(display_id, None, "%s: ")}{note}')
3755 return self
._downloader
.get_info_extractor('Generic')._extract
_embeds
(
3756 smuggle_url(url
, {'block_ies': [self.ie_key()]}
), *args
, **kwargs
)
3759 def extract_from_webpage(cls
, ydl
, url
, webpage
):
3760 ie
= (cls
if isinstance(cls
._extract
_from
_webpage
, types
.MethodType
)
3761 else ydl
.get_info_extractor(cls
.ie_key()))
3762 for info
in ie
._extract
_from
_webpage
(url
, webpage
) or []:
3763 # url = None since we do not want to set (webpage/original)_url
3764 ydl
.add_default_extra_info(info
, ie
, None)
3768 def _extract_from_webpage(cls
, url
, webpage
):
3769 for embed_url
in orderedSet(
3770 cls
._extract
_embed
_urls
(url
, webpage
) or [], lazy
=True):
3771 yield cls
.url_result(embed_url
, None if cls
._VALID
_URL
is False else cls
)
3774 def _extract_embed_urls(cls
, url
, webpage
):
3775 """@returns all the embed urls on the webpage"""
3776 if '_EMBED_URL_RE' not in cls
.__dict
__:
3777 assert isinstance(cls
._EMBED
_REGEX
, (list, tuple))
3778 for idx
, regex
in enumerate(cls
._EMBED
_REGEX
):
3779 assert regex
.count('(?P<url>') == 1, \
3780 f
'{cls.__name__}._EMBED_REGEX[{idx}] must have exactly 1 url group\n\t{regex}'
3781 cls
._EMBED
_URL
_RE
= tuple(map(re
.compile, cls
._EMBED
_REGEX
))
3783 for regex
in cls
._EMBED
_URL
_RE
:
3784 for mobj
in regex
.finditer(webpage
):
3785 embed_url
= urllib
.parse
.urljoin(url
, unescapeHTML(mobj
.group('url')))
3786 if cls
._VALID
_URL
is False or cls
.suitable(embed_url
):
3789 class StopExtraction(Exception):
3793 def _extract_url(cls
, webpage
): # TODO: Remove
3794 """Only for compatibility with some older extractors"""
3795 return next(iter(cls
._extract
_embed
_urls
(None, webpage
) or []), None)
3798 def __init_subclass__(cls
, *, plugin_name
=None, **kwargs
):
3800 mro
= inspect
.getmro(cls
)
3801 super_class
= cls
.__wrapped
__ = mro
[mro
.index(cls
) + 1]
3802 cls
.PLUGIN_NAME
, cls
.ie_key
= plugin_name
, super_class
.ie_key
3803 cls
.IE_NAME
= f
'{super_class.IE_NAME}+{plugin_name}'
3804 while getattr(super_class
, '__wrapped__', None):
3805 super_class
= super_class
.__wrapped
__
3806 setattr(sys
.modules
[super_class
.__module
__], super_class
.__name
__, cls
)
3807 _PLUGIN_OVERRIDES
[super_class
].append(cls
)
3809 return super().__init
_subclass
__(**kwargs
)
3812 class SearchInfoExtractor(InfoExtractor
):
3814 Base class for paged search queries extractors.
3815 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
3816 Instances should define _SEARCH_KEY and optionally _MAX_RESULTS
3819 _MAX_RESULTS
= float('inf')
3820 _RETURN_TYPE
= 'playlist'
3823 def _VALID_URL(cls
):
3824 return r
'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls
._SEARCH
_KEY
3826 def _real_extract(self
, query
):
3827 prefix
, query
= self
._match
_valid
_url
(query
).group('prefix', 'query')
3829 return self
._get
_n
_results
(query
, 1)
3830 elif prefix
== 'all':
3831 return self
._get
_n
_results
(query
, self
._MAX
_RESULTS
)
3835 raise ExtractorError(f
'invalid download number {n} for query "{query}"')
3836 elif n
> self
._MAX
_RESULTS
:
3837 self
.report_warning('%s returns max %i results (you requested %i)' % (self
._SEARCH
_KEY
, self
._MAX
_RESULTS
, n
))
3838 n
= self
._MAX
_RESULTS
3839 return self
._get
_n
_results
(query
, n
)
3841 def _get_n_results(self
, query
, n
):
3842 """Get a specified number of results for a query.
3843 Either this function or _search_results must be overridden by subclasses """
3844 return self
.playlist_result(
3845 itertools
.islice(self
._search
_results
(query
), 0, None if n
== float('inf') else n
),
3848 def _search_results(self
, query
):
3849 """Returns an iterator of search results"""
3850 raise NotImplementedError('This method must be implemented by subclasses')
3853 def SEARCH_KEY(cls
):
3854 return cls
._SEARCH
_KEY
3857 class UnsupportedURLIE(InfoExtractor
):
3862 def _real_extract(self
, url
):
3863 raise UnsupportedError(url
)
3866 _PLUGIN_OVERRIDES
= collections
.defaultdict(list)