2 from __future__
import unicode_literals
17 from ..compat
import (
18 compat_cookiejar_Cookie
,
19 compat_cookies_SimpleCookie
,
21 compat_etree_fromstring
,
28 compat_urllib_parse_unquote
,
29 compat_urllib_parse_urlencode
,
30 compat_urllib_request
,
32 compat_xml_parse_error
,
34 from ..downloader
import FileDownloader
35 from ..downloader
.f4m
import (
37 remove_encrypted_media
,
69 parse_m3u8_attributes
,
94 class InfoExtractor(object):
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.
114 Additionally, it must contain either a formats entry or a url one:
116 formats: A list of dictionaries for each format available, ordered
117 from worst to best quality.
120 * url The mandatory URL representing the media:
121 for plain file media - HTTP URL of this file,
123 for HLS - URL of the M3U8 media playlist,
124 for HDS - URL of the F4M manifest,
126 - HTTP URL to plain file media (in case of
128 - URL of the MPD manifest or base URL
129 representing the media if MPD manifest
130 is parsed from a string (in case of
132 for MSS - URL of the ISM manifest.
134 The URL of the manifest file in case of
136 for HLS - URL of the M3U8 master playlist,
137 for HDS - URL of the F4M manifest,
138 for DASH - URL of the MPD manifest,
139 for MSS - URL of the ISM manifest.
140 * ext Will be calculated from URL if missing
141 * format A human-readable description of the format
142 ("mp4 container with h264/opus").
143 Calculated from the format_id, width, height.
144 and format_note fields if missing.
145 * format_id A short description of the format
146 ("mp4_h264_opus" or "19").
147 Technically optional, but strongly recommended.
148 * format_note Additional info about the format
149 ("3D" or "DASH video")
150 * width Width of the video, if known
151 * height Height of the video, if known
152 * resolution Textual description of width and height
153 * dynamic_range The dynamic range of the video. One of:
154 "SDR" (None), "HDR10", "HDR10+, "HDR12", "HLG, "DV"
155 * tbr Average bitrate of audio and video in KBit/s
156 * abr Average audio bitrate in KBit/s
157 * acodec Name of the audio codec in use
158 * asr Audio sampling rate in Hertz
159 * vbr Average video bitrate in KBit/s
161 * vcodec Name of the video codec in use
162 * container Name of the container format
163 * filesize The number of bytes, if known in advance
164 * filesize_approx An estimate for the number of bytes
165 * player_url SWF Player URL (used for rtmpdump).
166 * protocol The protocol that will be used for the actual
167 download, lower-case. One of "http", "https" or
168 one of the protocols defined in downloader.PROTOCOL_MAP
170 Base URL for fragments. Each fragment's path
171 value (if present) will be relative to
173 * fragments A list of fragments of a fragmented media.
174 Each fragment entry must contain either an url
175 or a path. If an url is present it should be
176 considered by a client. Otherwise both path and
177 fragment_base_url must be present. Here is
178 the list of all potential fields:
179 * "url" - fragment's URL
180 * "path" - fragment's path relative to
182 * "duration" (optional, int or float)
183 * "filesize" (optional, int)
184 * is_from_start Is a live format that can be downloaded
185 from the start. Boolean
186 * preference Order number of this format. If this field is
187 present and not None, the formats get sorted
188 by this field, regardless of all other values.
189 -1 for default (order by other properties),
190 -2 or smaller for less than default.
191 < -1000 to hide the format (if there is
192 another one which is strictly better)
193 * language Language code, e.g. "de" or "en-US".
194 * language_preference Is this in the language mentioned in
196 10 if it's what the URL is about,
197 -1 for default (don't know),
198 -10 otherwise, other values reserved for now.
199 * quality Order number of the video quality of this
200 format, irrespective of the file format.
201 -1 for default (order by other properties),
202 -2 or smaller for less than default.
203 * source_preference Order number for this video source
204 (quality takes higher priority)
205 -1 for default (order by other properties),
206 -2 or smaller for less than default.
207 * http_headers A dictionary of additional HTTP headers
208 to add to the request.
209 * stretched_ratio If given and not 1, indicates that the
210 video's pixels are not square.
211 width : height ratio as float.
212 * no_resume The server does not support resuming the
213 (HTTP or RTMP) download. Boolean.
214 * has_drm The format has DRM and cannot be downloaded. Boolean
215 * downloader_options A dictionary of downloader options as
216 described in FileDownloader
217 RTMP formats can also have the additional fields: page_url,
218 app, play_path, tc_url, flash_version, rtmp_live, rtmp_conn,
219 rtmp_protocol, rtmp_real_time
221 url: Final video URL.
222 ext: Video filename extension.
223 format: The video format, defaults to ext (used for --get-format)
224 player_url: SWF Player URL (used for rtmpdump).
226 The following fields are optional:
228 alt_title: A secondary title of the video.
229 display_id An alternative identifier for the video, not necessarily
230 unique, but available before title. Typically, id is
231 something like "4234987", title "Dancing naked mole rats",
232 and display_id "dancing-naked-mole-rats"
233 thumbnails: A list of dictionaries, with the following entries:
234 * "id" (optional, string) - Thumbnail format ID
236 * "preference" (optional, int) - quality of the image
237 * "width" (optional, int)
238 * "height" (optional, int)
239 * "resolution" (optional, string "{width}x{height}",
241 * "filesize" (optional, int)
242 thumbnail: Full URL to a video thumbnail image.
243 description: Full video description.
244 uploader: Full name of the video uploader.
245 license: License name the video is licensed under.
246 creator: The creator of the video.
247 timestamp: UNIX timestamp of the moment the video was uploaded
248 upload_date: Video upload date (YYYYMMDD).
249 If not explicitly set, calculated from timestamp
250 release_timestamp: UNIX timestamp of the moment the video was released.
251 If it is not clear whether to use timestamp or this, use the former
252 release_date: The date (YYYYMMDD) when the video was released.
253 If not explicitly set, calculated from release_timestamp
254 modified_timestamp: UNIX timestamp of the moment the video was last modified.
255 modified_date: The date (YYYYMMDD) when the video was last modified.
256 If not explicitly set, calculated from modified_timestamp
257 uploader_id: Nickname or id of the video uploader.
258 uploader_url: Full URL to a personal webpage of the video uploader.
259 channel: Full name of the channel the video is uploaded on.
260 Note that channel fields may or may not repeat uploader
261 fields. This depends on a particular extractor.
262 channel_id: Id of the channel.
263 channel_url: Full URL to a channel webpage.
264 channel_follower_count: Number of followers of the channel.
265 location: Physical location where the video was filmed.
266 subtitles: The available subtitles as a dictionary in the format
267 {tag: subformats}. "tag" is usually a language code, and
268 "subformats" is a list sorted from lower to higher
269 preference, each element is a dictionary with the "ext"
271 * "data": The subtitles file contents
272 * "url": A URL pointing to the subtitles file
273 It can optionally also have:
274 * "name": Name or description of the subtitles
275 "ext" will be calculated from URL if missing
276 automatic_captions: Like 'subtitles'; contains automatically generated
277 captions instead of normal subtitles
278 duration: Length of the video in seconds, as an integer or float.
279 view_count: How many users have watched the video on the platform.
280 like_count: Number of positive ratings of the video
281 dislike_count: Number of negative ratings of the video
282 repost_count: Number of reposts of the video
283 average_rating: Average rating give by users, the scale used depends on the webpage
284 comment_count: Number of comments on the video
285 comments: A list of comments, each with one or more of the following
286 properties (all but one of text or html optional):
287 * "author" - human-readable name of the comment author
288 * "author_id" - user ID of the comment author
289 * "author_thumbnail" - The thumbnail of the comment author
291 * "html" - Comment as HTML
292 * "text" - Plain text of the comment
293 * "timestamp" - UNIX timestamp of comment
294 * "parent" - ID of the comment this one is replying to.
295 Set to "root" to indicate that this is a
296 comment to the original video.
297 * "like_count" - Number of positive ratings of the comment
298 * "dislike_count" - Number of negative ratings of the comment
299 * "is_favorited" - Whether the comment is marked as
300 favorite by the video uploader
301 * "author_is_uploader" - Whether the comment is made by
303 age_limit: Age restriction for the video, as an integer (years)
304 webpage_url: The URL to the video webpage, if given to yt-dlp it
305 should allow to get the same result again. (It will be set
306 by YoutubeDL if it's missing)
307 categories: A list of categories that the video falls in, for example
309 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
310 cast: A list of the video cast
311 is_live: True, False, or None (=unknown). Whether this video is a
312 live stream that goes on instead of a fixed-length video.
313 was_live: True, False, or None (=unknown). Whether this video was
314 originally a live stream.
315 live_status: 'is_live', 'is_upcoming', 'was_live', 'not_live' or None (=unknown)
316 If absent, automatically set from is_live, was_live
317 start_time: Time in seconds where the reproduction should start, as
318 specified in the URL.
319 end_time: Time in seconds where the reproduction should end, as
320 specified in the URL.
321 chapters: A list of dictionaries, with the following entries:
322 * "start_time" - The start time of the chapter in seconds
323 * "end_time" - The end time of the chapter in seconds
324 * "title" (optional, string)
325 playable_in_embed: Whether this video is allowed to play in embedded
326 players on other sites. Can be True (=always allowed),
327 False (=never allowed), None (=unknown), or a string
328 specifying the criteria for embedability (Eg: 'whitelist')
329 availability: Under what condition the video is available. One of
330 'private', 'premium_only', 'subscriber_only', 'needs_auth',
331 'unlisted' or 'public'. Use 'InfoExtractor._availability'
333 __post_extractor: A function to be called just before the metadata is
334 written to either disk, logger or console. The function
335 must return a dict which will be added to the info_dict.
336 This is usefull for additional information that is
337 time-consuming to extract. Note that the fields thus
338 extracted will not be available to output template and
339 match_filter. So, only "comments" and "comment_count" are
340 currently allowed to be extracted via this method.
342 The following fields should only be used when the video belongs to some logical
345 chapter: Name or title of the chapter the video belongs to.
346 chapter_number: Number of the chapter the video belongs to, as an integer.
347 chapter_id: Id of the chapter the video belongs to, as a unicode string.
349 The following fields should only be used when the video is an episode of some
350 series, programme or podcast:
352 series: Title of the series or programme the video episode belongs to.
353 series_id: Id of the series or programme the video episode belongs to, as a unicode string.
354 season: Title of the season the video episode belongs to.
355 season_number: Number of the season the video episode belongs to, as an integer.
356 season_id: Id of the season the video episode belongs to, as a unicode string.
357 episode: Title of the video episode. Unlike mandatory video title field,
358 this field should denote the exact title of the video episode
359 without any kind of decoration.
360 episode_number: Number of the video episode within a season, as an integer.
361 episode_id: Id of the video episode, as a unicode string.
363 The following fields should only be used when the media is a track or a part of
366 track: Title of the track.
367 track_number: Number of the track within an album or a disc, as an integer.
368 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
370 artist: Artist(s) of the track.
371 genre: Genre(s) of the track.
372 album: Title of the album the track belongs to.
373 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
374 album_artist: List of all artists appeared on the album (e.g.
375 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
377 disc_number: Number of the disc or other physical medium the track belongs to,
379 release_year: Year (YYYY) when the album was released.
380 composer: Composer of the piece
382 Unless mentioned otherwise, the fields should be Unicode strings.
384 Unless mentioned otherwise, None is equivalent to absence of information.
387 _type "playlist" indicates multiple videos.
388 There must be a key "entries", which is a list, an iterable, or a PagedList
389 object, each element of which is a valid dictionary by this specification.
391 Additionally, playlists can have "id", "title", and any other relevent
392 attributes with the same semantics as videos (see above).
394 It can also have the following optional fields:
396 playlist_count: The total number of videos in a playlist. If not given,
397 YoutubeDL tries to calculate it from "entries"
400 _type "multi_video" indicates that there are multiple videos that
401 form a single show, for examples multiple acts of an opera or TV episode.
402 It must have an entries key like a playlist and contain all the keys
403 required for a video at the same time.
406 _type "url" indicates that the video must be extracted from another
407 location, possibly by a different extractor. Its only required key is:
408 "url" - the next URL to extract.
409 The key "ie_key" can be set to the class name (minus the trailing "IE",
410 e.g. "Youtube") if the extractor class is known in advance.
411 Additionally, the dictionary may have any properties of the resolved entity
412 known in advance, for example "title" if the title of the referred video is
416 _type "url_transparent" entities have the same specification as "url", but
417 indicate that the given additional information is more precise than the one
418 associated with the resolved URL.
419 This is useful when a site employs a video service that hosts the video and
420 its technical metadata, but that video service does not embed a useful
421 title, description etc.
424 Subclasses of this one should re-define the _real_initialize() and
425 _real_extract() methods and define a _VALID_URL regexp.
426 Probably, they should also be added to the list of extractors.
428 Subclasses may also override suitable() if necessary, but ensure the function
429 signature is preserved and that this function imports everything it needs
430 (except other extractors), so that lazy_extractors works correctly
432 _GEO_BYPASS attribute may be set to False in order to disable
433 geo restriction bypass mechanisms for a particular extractor.
434 Though it won't disable explicit geo restriction bypass based on
435 country code provided with geo_bypass_country.
437 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
438 countries for this extractor. One of these countries will be used by
439 geo restriction bypass mechanism right away in order to bypass
440 geo restriction, of course, if the mechanism is not disabled.
442 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
443 IP blocks in CIDR notation for this extractor. One of these IP blocks
444 will be used by geo restriction bypass mechanism similarly
447 The _WORKING attribute should be set to False for broken IEs
448 in order to warn the users and skip the tests.
453 _x_forwarded_for_ip
= None
455 _GEO_COUNTRIES
= None
456 _GEO_IP_BLOCKS
= None
460 'any': 'Use --cookies, --username and --password, or --netrc to provide account credentials',
462 'Use --cookies-from-browser or --cookies for the authentication. '
463 'See https://github.com/ytdl-org/youtube-dl#how-do-i-pass-cookies-to-youtube-dl for how to manually pass cookies'),
464 'password': 'Use --username and --password, or --netrc to provide account credentials',
467 def __init__(self
, downloader
=None):
468 """Constructor. Receives an optional downloader (a YoutubeDL instance).
469 If a downloader is not passed during initialization,
470 it must be set using "set_downloader()" before "extract()" is called"""
472 self
._x
_forwarded
_for
_ip
= None
473 self
._printed
_messages
= set()
474 self
.set_downloader(downloader
)
477 def _match_valid_url(cls
, url
):
478 # This does not use has/getattr intentionally - we want to know whether
479 # we have cached the regexp for *this* class, whereas getattr would also
480 # match the superclass
481 if '_VALID_URL_RE' not in cls
.__dict
__:
482 if '_VALID_URL' not in cls
.__dict
__:
483 cls
._VALID
_URL
= cls
._make
_valid
_url
()
484 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
485 return cls
._VALID
_URL
_RE
.match(url
)
488 def suitable(cls
, url
):
489 """Receives a URL and returns True if suitable for this IE."""
490 # This function must import everything it needs (except other extractors),
491 # so that lazy_extractors works correctly
492 return cls
._match
_valid
_url
(url
) is not None
495 def _match_id(cls
, url
):
496 return cls
._match
_valid
_url
(url
).group('id')
499 def get_temp_id(cls
, url
):
501 return cls
._match
_id
(url
)
502 except (IndexError, AttributeError):
507 """Getter method for _WORKING."""
510 def initialize(self
):
511 """Initializes an instance (authentication, etc)."""
512 self
._printed
_messages
= set()
513 self
._initialize
_geo
_bypass
({
514 'countries': self
._GEO
_COUNTRIES
,
515 'ip_blocks': self
._GEO
_IP
_BLOCKS
,
518 self
._real
_initialize
()
521 def _initialize_geo_bypass(self
, geo_bypass_context
):
523 Initialize geo restriction bypass mechanism.
525 This method is used to initialize geo bypass mechanism based on faking
526 X-Forwarded-For HTTP header. A random country from provided country list
527 is selected and a random IP belonging to this country is generated. This
528 IP will be passed as X-Forwarded-For HTTP header in all subsequent
531 This method will be used for initial geo bypass mechanism initialization
532 during the instance initialization with _GEO_COUNTRIES and
535 You may also manually call it from extractor's code if geo bypass
536 information is not available beforehand (e.g. obtained during
537 extraction) or due to some other reason. In this case you should pass
538 this information in geo bypass context passed as first argument. It may
539 contain following fields:
541 countries: List of geo unrestricted countries (similar
543 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
544 (similar to _GEO_IP_BLOCKS)
547 if not self
._x
_forwarded
_for
_ip
:
549 # Geo bypass mechanism is explicitly disabled by user
550 if not self
.get_param('geo_bypass', True):
553 if not geo_bypass_context
:
554 geo_bypass_context
= {}
556 # Backward compatibility: previously _initialize_geo_bypass
557 # expected a list of countries, some 3rd party code may still use
559 if isinstance(geo_bypass_context
, (list, tuple)):
560 geo_bypass_context
= {
561 'countries': geo_bypass_context
,
564 # The whole point of geo bypass mechanism is to fake IP
565 # as X-Forwarded-For HTTP header based on some IP block or
568 # Path 1: bypassing based on IP block in CIDR notation
570 # Explicit IP block specified by user, use it right away
571 # regardless of whether extractor is geo bypassable or not
572 ip_block
= self
.get_param('geo_bypass_ip_block', None)
574 # Otherwise use random IP block from geo bypass context but only
575 # if extractor is known as geo bypassable
577 ip_blocks
= geo_bypass_context
.get('ip_blocks')
578 if self
._GEO
_BYPASS
and ip_blocks
:
579 ip_block
= random
.choice(ip_blocks
)
582 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(ip_block
)
583 self
._downloader
.write_debug(
584 '[debug] Using fake IP %s as X-Forwarded-For' % self
._x
_forwarded
_for
_ip
)
587 # Path 2: bypassing based on country code
589 # Explicit country code specified by user, use it right away
590 # regardless of whether extractor is geo bypassable or not
591 country
= self
.get_param('geo_bypass_country', None)
593 # Otherwise use random country code from geo bypass context but
594 # only if extractor is known as geo bypassable
596 countries
= geo_bypass_context
.get('countries')
597 if self
._GEO
_BYPASS
and countries
:
598 country
= random
.choice(countries
)
601 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country
)
602 self
._downloader
.write_debug(
603 'Using fake IP %s (%s) as X-Forwarded-For' % (self
._x
_forwarded
_for
_ip
, country
.upper()))
605 def extract(self
, url
):
606 """Extracts URL information and returns it in list of dicts."""
611 self
.write_debug('Extracting URL: %s' % url
)
612 ie_result
= self
._real
_extract
(url
)
613 if ie_result
is None:
615 if self
._x
_forwarded
_for
_ip
:
616 ie_result
['__x_forwarded_for_ip'] = self
._x
_forwarded
_for
_ip
617 subtitles
= ie_result
.get('subtitles')
618 if (subtitles
and 'live_chat' in subtitles
619 and 'no-live-chat' in self
.get_param('compat_opts', [])):
620 del subtitles
['live_chat']
622 except GeoRestrictedError
as e
:
623 if self
.__maybe
_fake
_ip
_and
_retry
(e
.countries
):
626 except UnsupportedError
:
628 except ExtractorError
as e
:
630 'video_id': e
.video_id
or self
.get_temp_id(url
),
632 'tb': e
.traceback
or sys
.exc_info()[2],
633 'expected': e
.expected
,
636 if hasattr(e
, 'countries'):
637 kwargs
['countries'] = e
.countries
638 raise type(e
)(e
.msg
, **kwargs
)
639 except compat_http_client
.IncompleteRead
as e
:
640 raise ExtractorError('A network error has occurred.', cause
=e
, expected
=True, video_id
=self
.get_temp_id(url
))
641 except (KeyError, StopIteration) as e
:
642 raise ExtractorError('An extractor error has occurred.', cause
=e
, video_id
=self
.get_temp_id(url
))
644 def __maybe_fake_ip_and_retry(self
, countries
):
645 if (not self
.get_param('geo_bypass_country', None)
647 and self
.get_param('geo_bypass', True)
648 and not self
._x
_forwarded
_for
_ip
650 country_code
= random
.choice(countries
)
651 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country_code
)
652 if self
._x
_forwarded
_for
_ip
:
654 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
655 % (self
._x
_forwarded
_for
_ip
, country_code
.upper()))
659 def set_downloader(self
, downloader
):
660 """Sets the downloader for this IE."""
661 self
._downloader
= downloader
663 def _real_initialize(self
):
664 """Real initialization process. Redefine in subclasses."""
667 def _real_extract(self
, url
):
668 """Real extraction process. Redefine in subclasses."""
673 """A string for getting the InfoExtractor with get_info_extractor"""
674 return cls
.__name
__[:-2]
678 return compat_str(type(self
).__name
__[:-2])
681 def __can_accept_status_code(err
, expected_status
):
682 assert isinstance(err
, compat_urllib_error
.HTTPError
)
683 if expected_status
is None:
685 elif callable(expected_status
):
686 return expected_status(err
.code
) is True
688 return err
.code
in variadic(expected_status
)
690 def _request_webpage(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
, expected_status
=None):
692 Return the response handle.
694 See _download_webpage docstring for arguments specification.
696 if not self
._downloader
._first
_webpage
_request
:
697 sleep_interval
= self
.get_param('sleep_interval_requests') or 0
698 if sleep_interval
> 0:
699 self
.to_screen('Sleeping %s seconds ...' % sleep_interval
)
700 time
.sleep(sleep_interval
)
702 self
._downloader
._first
_webpage
_request
= False
705 self
.report_download_webpage(video_id
)
706 elif note
is not False:
708 self
.to_screen('%s' % (note
,))
710 self
.to_screen('%s: %s' % (video_id
, note
))
712 # Some sites check X-Forwarded-For HTTP header in order to figure out
713 # the origin of the client behind proxy. This allows bypassing geo
714 # restriction by faking this header's value to IP that belongs to some
715 # geo unrestricted country. We will do so once we encounter any
716 # geo restriction error.
717 if self
._x
_forwarded
_for
_ip
:
718 if 'X-Forwarded-For' not in headers
:
719 headers
['X-Forwarded-For'] = self
._x
_forwarded
_for
_ip
721 if isinstance(url_or_request
, compat_urllib_request
.Request
):
722 url_or_request
= update_Request(
723 url_or_request
, data
=data
, headers
=headers
, query
=query
)
726 url_or_request
= update_url_query(url_or_request
, query
)
727 if data
is not None or headers
:
728 url_or_request
= sanitized_Request(url_or_request
, data
, headers
)
730 return self
._downloader
.urlopen(url_or_request
)
731 except network_exceptions
as err
:
732 if isinstance(err
, compat_urllib_error
.HTTPError
):
733 if self
.__can
_accept
_status
_code
(err
, expected_status
):
734 # Retain reference to error to prevent file object from
735 # being closed before it can be read. Works around the
736 # effects of <https://bugs.python.org/issue15002>
737 # introduced in Python 3.4.1.
744 errnote
= 'Unable to download webpage'
746 errmsg
= '%s: %s' % (errnote
, error_to_compat_str(err
))
748 raise ExtractorError(errmsg
, sys
.exc_info()[2], cause
=err
)
750 self
.report_warning(errmsg
)
753 def _download_webpage_handle(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, encoding
=None, data
=None, headers
={}, query={}
, expected_status
=None):
755 Return a tuple (page content as string, URL handle).
757 See _download_webpage docstring for arguments specification.
759 # Strip hashes from the URL (#1038)
760 if isinstance(url_or_request
, (compat_str
, str)):
761 url_or_request
= url_or_request
.partition('#')[0]
763 urlh
= self
._request
_webpage
(url_or_request
, video_id
, note
, errnote
, fatal
, data
=data
, headers
=headers
, query
=query
, expected_status
=expected_status
)
767 content
= self
._webpage
_read
_content
(urlh
, url_or_request
, video_id
, note
, errnote
, fatal
, encoding
=encoding
)
768 return (content
, urlh
)
771 def _guess_encoding_from_content(content_type
, webpage_bytes
):
772 m
= re
.match(r
'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type
)
774 encoding
= m
.group(1)
776 m
= re
.search(br
'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
777 webpage_bytes[:1024])
779 encoding = m.group(1).decode('ascii')
780 elif webpage_bytes.startswith(b'\xff\xfe'):
787 def __check_blocked(self, content):
788 first_block = content[:512]
789 if ('<title>Access to this site is blocked</title>' in content
790 and 'Websense' in first_block):
791 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
792 blocked_iframe = self._html_search_regex(
793 r'<iframe src="([^
"]+)"', content,
794 'Websense information URL
', default=None)
796 msg += ' Visit
%s for more details
' % blocked_iframe
797 raise ExtractorError(msg, expected=True)
798 if '<title
>The URL you requested has been blocked
</title
>' in first_block:
800 'Access to this webpage has been blocked by Indian censorship
. '
801 'Use a VPN
or proxy
server (with --proxy
) to route around it
.')
802 block_msg = self._html_search_regex(
803 r'</h1
><p
>(.*?
)</p
>',
804 content, 'block message
', default=None)
806 msg += ' (Message
: "%s")' % block_msg.replace('\n', ' ')
807 raise ExtractorError(msg, expected=True)
808 if ('<title
>TTK
:: Доступ к ресурсу ограничен
</title
>' in content
809 and 'blocklist
.rkn
.gov
.ru
' in content):
810 raise ExtractorError(
811 'Access to this webpage has been blocked by decision of the Russian government
. '
812 'Visit http
://blocklist
.rkn
.gov
.ru
/ for a block reason
.',
815 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
816 content_type = urlh.headers.get('Content
-Type
', '')
817 webpage_bytes = urlh.read()
818 if prefix is not None:
819 webpage_bytes = prefix + webpage_bytes
821 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
822 if self.get_param('dump_intermediate_pages
', False):
823 self.to_screen('Dumping request to
' + urlh.geturl())
824 dump = base64.b64encode(webpage_bytes).decode('ascii
')
825 self._downloader.to_screen(dump)
826 if self.get_param('write_pages
', False):
827 basen = '%s_%s' % (video_id, urlh.geturl())
828 trim_length = self.get_param('trim_file_name
') or 240
829 if len(basen) > trim_length:
830 h = '___
' + hashlib.md5(basen.encode('utf
-8')).hexdigest()
831 basen = basen[:trim_length - len(h)] + h
832 raw_filename = basen + '.dump
'
833 filename = sanitize_filename(raw_filename, restricted=True)
834 self.to_screen('Saving request to
' + filename)
835 # Working around MAX_PATH limitation on Windows (see
836 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
837 if compat_os_name == 'nt
':
838 absfilepath = os.path.abspath(filename)
839 if len(absfilepath) > 259:
840 filename = '\\\\?
\\' + absfilepath
841 with open(filename, 'wb
') as outf:
842 outf.write(webpage_bytes)
845 content = webpage_bytes.decode(encoding, 'replace
')
847 content = webpage_bytes.decode('utf
-8', 'replace
')
849 self.__check_blocked(content)
853 def _download_webpage(
854 self, url_or_request, video_id, note=None, errnote=None,
855 fatal=True, tries=1, timeout=5, encoding=None, data=None,
856 headers={}, query={}, expected_status=None):
858 Return the data of the page as a string.
861 url_or_request -- plain text URL as a string or
862 a compat_urllib_request.Requestobject
863 video_id -- Video/playlist/item identifier (string)
866 note -- note printed before downloading (string)
867 errnote -- note printed in case of an error (string)
868 fatal -- flag denoting whether error should be considered fatal,
869 i.e. whether it should cause ExtractionError to be raised,
870 otherwise a warning will be reported and extraction continued
871 tries -- number of tries
872 timeout -- sleep interval between tries
873 encoding -- encoding for a page content decoding, guessed automatically
874 when not explicitly specified
875 data -- POST data (bytes)
876 headers -- HTTP headers (dict)
877 query -- URL query (dict)
878 expected_status -- allows to accept failed HTTP requests (non 2xx
879 status code) by explicitly specifying a set of accepted status
880 codes. Can be any of the following entities:
881 - an integer type specifying an exact failed status code to
883 - a list or a tuple of integer types specifying a list of
884 failed status codes to accept
885 - a callable accepting an actual failed status code and
886 returning True if it should be accepted
887 Note that this argument does not affect success status codes (2xx)
888 which are always accepted.
893 while success is False:
895 res = self._download_webpage_handle(
896 url_or_request, video_id, note, errnote, fatal,
897 encoding=encoding, data=data, headers=headers, query=query,
898 expected_status=expected_status)
900 except compat_http_client.IncompleteRead as e:
902 if try_count >= tries:
904 self._sleep(timeout, video_id)
911 def _download_xml_handle(
912 self, url_or_request, video_id, note='Downloading XML
',
913 errnote='Unable to download XML
', transform_source=None,
914 fatal=True, encoding=None, data=None, headers={}, query={},
915 expected_status=None):
917 Return a tuple (xml as an compat_etree_Element, URL handle).
919 See _download_webpage docstring for arguments specification.
921 res = self._download_webpage_handle(
922 url_or_request, video_id, note, errnote, fatal=fatal,
923 encoding=encoding, data=data, headers=headers, query=query,
924 expected_status=expected_status)
927 xml_string, urlh = res
928 return self._parse_xml(
929 xml_string, video_id, transform_source=transform_source,
933 self, url_or_request, video_id,
934 note='Downloading XML
', errnote='Unable to download XML
',
935 transform_source=None, fatal=True, encoding=None,
936 data=None, headers={}, query={}, expected_status=None):
938 Return the xml as an compat_etree_Element.
940 See _download_webpage docstring for arguments specification.
942 res = self._download_xml_handle(
943 url_or_request, video_id, note=note, errnote=errnote,
944 transform_source=transform_source, fatal=fatal, encoding=encoding,
945 data=data, headers=headers, query=query,
946 expected_status=expected_status)
947 return res if res is False else res[0]
949 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
951 xml_string = transform_source(xml_string)
953 return compat_etree_fromstring(xml_string.encode('utf
-8'))
954 except compat_xml_parse_error as ve:
955 errmsg = '%s: Failed to parse XML
' % video_id
957 raise ExtractorError(errmsg, cause=ve)
959 self.report_warning(errmsg + str(ve))
961 def _download_json_handle(
962 self, url_or_request, video_id, note='Downloading JSON metadata
',
963 errnote='Unable to download JSON metadata
', transform_source=None,
964 fatal=True, encoding=None, data=None, headers={}, query={},
965 expected_status=None):
967 Return a tuple (JSON object, URL handle).
969 See _download_webpage docstring for arguments specification.
971 res = self._download_webpage_handle(
972 url_or_request, video_id, note, errnote, fatal=fatal,
973 encoding=encoding, data=data, headers=headers, query=query,
974 expected_status=expected_status)
977 json_string, urlh = res
978 return self._parse_json(
979 json_string, video_id, transform_source=transform_source,
983 self, url_or_request, video_id, note='Downloading JSON metadata
',
984 errnote='Unable to download JSON metadata
', transform_source=None,
985 fatal=True, encoding=None, data=None, headers={}, query={},
986 expected_status=None):
988 Return the JSON object as a dict.
990 See _download_webpage docstring for arguments specification.
992 res = self._download_json_handle(
993 url_or_request, video_id, note=note, errnote=errnote,
994 transform_source=transform_source, fatal=fatal, encoding=encoding,
995 data=data, headers=headers, query=query,
996 expected_status=expected_status)
997 return res if res is False else res[0]
999 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
1000 if transform_source:
1001 json_string = transform_source(json_string)
1003 return json.loads(json_string)
1004 except ValueError as ve:
1005 errmsg = '%s: Failed to parse JSON
' % video_id
1007 raise ExtractorError(errmsg, cause=ve)
1009 self.report_warning(errmsg + str(ve))
1011 def _parse_socket_response_as_json(self, data, video_id, transform_source=None, fatal=True):
1012 return self._parse_json(
1013 data[data.find('{'):data.rfind('}
') + 1],
1014 video_id, transform_source, fatal)
1016 def _download_socket_json_handle(
1017 self, url_or_request, video_id, note='Polling socket
',
1018 errnote='Unable to poll socket
', transform_source=None,
1019 fatal=True, encoding=None, data=None, headers={}, query={},
1020 expected_status=None):
1022 Return a tuple (JSON object, URL handle).
1024 See _download_webpage docstring for arguments specification.
1026 res = self._download_webpage_handle(
1027 url_or_request, video_id, note, errnote, fatal=fatal,
1028 encoding=encoding, data=data, headers=headers, query=query,
1029 expected_status=expected_status)
1033 return self._parse_socket_response_as_json(
1034 webpage, video_id, transform_source=transform_source,
1037 def _download_socket_json(
1038 self, url_or_request, video_id, note='Polling socket
',
1039 errnote='Unable to poll socket
', transform_source=None,
1040 fatal=True, encoding=None, data=None, headers={}, query={},
1041 expected_status=None):
1043 Return the JSON object as a dict.
1045 See _download_webpage docstring for arguments specification.
1047 res = self._download_socket_json_handle(
1048 url_or_request, video_id, note=note, errnote=errnote,
1049 transform_source=transform_source, fatal=fatal, encoding=encoding,
1050 data=data, headers=headers, query=query,
1051 expected_status=expected_status)
1052 return res if res is False else res[0]
1054 def report_warning(self, msg, video_id=None, *args, only_once=False, **kwargs):
1055 idstr = format_field(video_id, template='%s: ')
1056 msg = f'[{self.IE_NAME}
] {idstr}{msg}
'
1058 if f'WARNING
: {msg}
' in self._printed_messages:
1060 self._printed_messages.add(f'WARNING
: {msg}
')
1061 self._downloader.report_warning(msg, *args, **kwargs)
1063 def to_screen(self, msg, *args, **kwargs):
1064 """Print msg to screen, prefixing it with '[ie_name
]'"""
1065 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg), *args, **kwargs)
1067 def write_debug(self, msg, *args, **kwargs):
1068 self._downloader.write_debug('[%s] %s' % (self.IE_NAME, msg), *args, **kwargs)
1070 def get_param(self, name, default=None, *args, **kwargs):
1071 if self._downloader:
1072 return self._downloader.params.get(name, default, *args, **kwargs)
1075 def report_drm(self, video_id, partial=False):
1076 self.raise_no_formats('This video
is DRM protected
', expected=True, video_id=video_id)
1078 def report_extraction(self, id_or_name):
1079 """Report information extraction."""
1080 self.to_screen('%s: Extracting information
' % id_or_name)
1082 def report_download_webpage(self, video_id):
1083 """Report webpage download."""
1084 self.to_screen('%s: Downloading webpage
' % video_id)
1086 def report_age_confirmation(self):
1087 """Report attempt to confirm age."""
1088 self.to_screen('Confirming age
')
1090 def report_login(self):
1091 """Report attempt to log in."""
1092 self.to_screen('Logging
in')
1094 def raise_login_required(
1095 self, msg='This video
is only available
for registered users
',
1096 metadata_available=False, method='any
'):
1097 if metadata_available and (
1098 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1099 self.report_warning(msg)
1100 if method is not None:
1101 msg = '%s. %s' % (msg, self._LOGIN_HINTS[method])
1102 raise ExtractorError(msg, expected=True)
1104 def raise_geo_restricted(
1105 self, msg='This video
is not available
from your location due to geo restriction
',
1106 countries=None, metadata_available=False):
1107 if metadata_available and (
1108 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1109 self.report_warning(msg)
1111 raise GeoRestrictedError(msg, countries=countries)
1113 def raise_no_formats(self, msg, expected=False, video_id=None):
1115 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1116 self.report_warning(msg, video_id)
1117 elif isinstance(msg, ExtractorError):
1120 raise ExtractorError(msg, expected=expected, video_id=video_id)
1122 # Methods for following #608
1124 def url_result(url, ie=None, video_id=None, video_title=None, *, url_transparent=False, **kwargs):
1125 """Returns a URL that points to a page that should be processed"""
1127 kwargs['ie_key
'] = ie if isinstance(ie, str) else ie.ie_key()
1128 if video_id is not None:
1129 kwargs['id'] = video_id
1130 if video_title is not None:
1131 kwargs['title
'] = video_title
1134 '_type
': 'url_transparent
' if url_transparent else 'url
',
1138 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None, **kwargs):
1139 urls = (self.url_result(self._proto_relative_url(m), ie)
1140 for m in orderedSet(map(getter, matches) if getter else matches))
1141 return self.playlist_result(urls, playlist_id, playlist_title, **kwargs)
1144 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None, *, multi_video=False, **kwargs):
1145 """Returns a playlist"""
1147 kwargs['id'] = playlist_id
1149 kwargs['title
'] = playlist_title
1150 if playlist_description is not None:
1151 kwargs['description
'] = playlist_description
1154 '_type
': 'multi_video
' if multi_video else 'playlist
',
1158 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1160 Perform a regex search on the given string, using a single or a list of
1161 patterns returning the first matching group.
1162 In case of failure return a default value or raise a WARNING or a
1163 RegexNotFoundError, depending on fatal, specifying the field name.
1165 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
1166 mobj = re.search(pattern, string, flags)
1169 mobj = re.search(p, string, flags)
1173 _name = self._downloader._format_err(name, self._downloader.Styles.EMPHASIS)
1177 # return the first matching group
1178 return next(g for g in mobj.groups() if g is not None)
1179 elif isinstance(group, (list, tuple)):
1180 return tuple(mobj.group(g) for g in group)
1182 return mobj.group(group)
1183 elif default is not NO_DEFAULT:
1186 raise RegexNotFoundError('Unable to extract
%s' % _name)
1188 self.report_warning('unable to extract
%s' % _name + bug_reports_message())
1191 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1193 Like _search_regex, but strips HTML tags and unescapes entities.
1195 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1197 return clean_html(res).strip()
1201 def _get_netrc_login_info(self, netrc_machine=None):
1204 netrc_machine = netrc_machine or self._NETRC_MACHINE
1206 if self.get_param('usenetrc
', False):
1208 netrc_file = compat_expanduser(self.get_param('netrc_location
') or '~
')
1209 if os.path.isdir(netrc_file):
1210 netrc_file = os.path.join(netrc_file, '.netrc
')
1211 info = netrc.netrc(file=netrc_file).authenticators(netrc_machine)
1212 if info is not None:
1216 raise netrc.NetrcParseError(
1217 'No authenticators
for %s' % netrc_machine)
1218 except (IOError, netrc.NetrcParseError) as err:
1219 self.report_warning(
1220 'parsing
.netrc
: %s' % error_to_compat_str(err))
1222 return username, password
1224 def _get_login_info(self, username_option='username
', password_option='password
', netrc_machine=None):
1226 Get the login info as (username, password)
1227 First look for the manually specified credentials using username_option
1228 and password_option as keys in params dictionary. If no such credentials
1229 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1231 If there's no info available
, return (None, None)
1234 # Attempt to use provided username and password or .netrc data
1235 username = self.get_param(username_option)
1236 if username is not None:
1237 password = self.get_param(password_option)
1239 username, password = self._get_netrc_login_info(netrc_machine)
1241 return username, password
1243 def _get_tfa_info(self, note='two-factor verification code'):
1245 Get the two
-factor authentication info
1246 TODO
- asking the user will be required
for sms
/phone verify
1247 currently just uses the command line option
1248 If there
's no info available, return None
1251 tfa = self.get_param('twofactor
')
1255 return compat_getpass('Type
%s and press
[Return
]: ' % note)
1257 # Helper functions for extracting OpenGraph info
1259 def _og_regexes(prop):
1260 content_re = r'content
=(?
:"([^"]+?
)"|\'([^\']+?)\'|\s*([^\s"\'=<>`
]+?
))'
1261 property_re = (r'(?
:name|
property)=(?
:\'og
[:-]%(prop)s\'|
"og[:-]%(prop)s"|\s
*og
[:-]%(prop)s\b)'
1262 % {'prop': re.escape(prop)})
1263 template = r'<meta
[^
>]+?
%s[^
>]+?
%s'
1265 template % (property_re, content_re),
1266 template % (content_re, property_re),
1270 def _meta_regex(prop):
1271 return r'''(?isx)<meta
1272 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1273 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1275 def _og_search_property(self, prop, html, name=None, **kargs):
1276 prop = variadic(prop)
1278 name = 'OpenGraph
%s' % prop[0]
1281 og_regexes.extend(self._og_regexes(p))
1282 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1285 return unescapeHTML(escaped)
1287 def _og_search_thumbnail(self, html, **kargs):
1288 return self._og_search_property('image
', html, 'thumbnail URL
', fatal=False, **kargs)
1290 def _og_search_description(self, html, **kargs):
1291 return self._og_search_property('description
', html, fatal=False, **kargs)
1293 def _og_search_title(self, html, **kargs):
1294 kargs.setdefault('fatal
', False)
1295 return self._og_search_property('title
', html, **kargs)
1297 def _og_search_video_url(self, html, name='video url
', secure=True, **kargs):
1298 regexes = self._og_regexes('video
') + self._og_regexes('video
:url
')
1300 regexes = self._og_regexes('video
:secure_url
') + regexes
1301 return self._html_search_regex(regexes, html, name, **kargs)
1303 def _og_search_url(self, html, **kargs):
1304 return self._og_search_property('url
', html, **kargs)
1306 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1307 name = variadic(name)
1308 if display_name is None:
1309 display_name = name[0]
1310 return self._html_search_regex(
1311 [self._meta_regex(n) for n in name],
1312 html, display_name, fatal=fatal, group='content
', **kwargs)
1314 def _dc_search_uploader(self, html):
1315 return self._html_search_meta('dc
.creator
', html, 'uploader
')
1317 def _rta_search(self, html):
1318 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1319 if re.search(r'(?ix
)<meta\s
+name
="rating"\s
+'
1320 r' content
="RTA-5042-1996-1400-1577-RTA"',
1325 def _media_rating_search(self, html):
1326 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1327 rating = self._html_search_meta('rating
', html)
1339 return RATING_TABLE.get(rating.lower())
1341 def _family_friendly_search(self, html):
1342 # See http://schema.org/VideoObject
1343 family_friendly = self._html_search_meta(
1344 'isFamilyFriendly
', html, default=None)
1346 if not family_friendly:
1355 return RATING_TABLE.get(family_friendly.lower())
1357 def _twitter_search_player(self, html):
1358 return self._html_search_meta('twitter
:player
', html,
1359 'twitter card player
')
1361 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
1362 json_ld_list = list(re.finditer(JSON_LD_RE, html))
1363 default = kwargs.get('default
', NO_DEFAULT)
1364 # JSON-LD may be malformed and thus `fatal` should be respected.
1365 # At the same time `default` may be passed that assumes `fatal=False`
1366 # for _search_regex. Let's simulate the same behavior here
as well
.
1367 fatal
= kwargs
.get('fatal', True) if default
is NO_DEFAULT
else False
1369 for mobj
in json_ld_list
:
1370 json_ld_item
= self
._parse
_json
(
1371 mobj
.group('json_ld'), video_id
, fatal
=fatal
)
1372 if not json_ld_item
:
1374 if isinstance(json_ld_item
, dict):
1375 json_ld
.append(json_ld_item
)
1376 elif isinstance(json_ld_item
, (list, tuple)):
1377 json_ld
.extend(json_ld_item
)
1379 json_ld
= self
._json
_ld
(json_ld
, video_id
, fatal
=fatal
, expected_type
=expected_type
)
1382 if default
is not NO_DEFAULT
:
1385 raise RegexNotFoundError('Unable to extract JSON-LD')
1387 self
.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1390 def _json_ld(self
, json_ld
, video_id
, fatal
=True, expected_type
=None):
1391 if isinstance(json_ld
, compat_str
):
1392 json_ld
= self
._parse
_json
(json_ld
, video_id
, fatal
=fatal
)
1396 if not isinstance(json_ld
, (list, tuple, dict)):
1398 if isinstance(json_ld
, dict):
1401 INTERACTION_TYPE_MAP
= {
1402 'CommentAction': 'comment',
1403 'AgreeAction': 'like',
1404 'DisagreeAction': 'dislike',
1405 'LikeAction': 'like',
1406 'DislikeAction': 'dislike',
1407 'ListenAction': 'view',
1408 'WatchAction': 'view',
1409 'ViewAction': 'view',
1412 def extract_interaction_type(e
):
1413 interaction_type
= e
.get('interactionType')
1414 if isinstance(interaction_type
, dict):
1415 interaction_type
= interaction_type
.get('@type')
1416 return str_or_none(interaction_type
)
1418 def extract_interaction_statistic(e
):
1419 interaction_statistic
= e
.get('interactionStatistic')
1420 if isinstance(interaction_statistic
, dict):
1421 interaction_statistic
= [interaction_statistic
]
1422 if not isinstance(interaction_statistic
, list):
1424 for is_e
in interaction_statistic
:
1425 if not isinstance(is_e
, dict):
1427 if is_e
.get('@type') != 'InteractionCounter':
1429 interaction_type
= extract_interaction_type(is_e
)
1430 if not interaction_type
:
1432 # For interaction count some sites provide string instead of
1433 # an integer (as per spec) with non digit characters (e.g. ",")
1434 # so extracting count with more relaxed str_to_int
1435 interaction_count
= str_to_int(is_e
.get('userInteractionCount'))
1436 if interaction_count
is None:
1438 count_kind
= INTERACTION_TYPE_MAP
.get(interaction_type
.split('/')[-1])
1441 count_key
= '%s_count' % count_kind
1442 if info
.get(count_key
) is not None:
1444 info
[count_key
] = interaction_count
1446 def extract_chapter_information(e
):
1448 'title': part
.get('name'),
1449 'start_time': part
.get('startOffset'),
1450 'end_time': part
.get('endOffset'),
1451 } for part
in variadic(e
.get('hasPart') or []) if part
.get('@type') == 'Clip']
1452 for idx
, (last_c
, current_c
, next_c
) in enumerate(zip(
1453 [{'end_time': 0}
] + chapters
, chapters
, chapters
[1:])):
1454 current_c
['end_time'] = current_c
['end_time'] or next_c
['start_time']
1455 current_c
['start_time'] = current_c
['start_time'] or last_c
['end_time']
1456 if None in current_c
.values():
1457 self
.report_warning(f
'Chapter {idx} contains broken data. Not extracting chapters')
1460 chapters
[-1]['end_time'] = chapters
[-1]['end_time'] or info
['duration']
1461 info
['chapters'] = chapters
1463 def extract_video_object(e
):
1464 assert e
['@type'] == 'VideoObject'
1465 author
= e
.get('author')
1467 'url': url_or_none(e
.get('contentUrl')),
1468 'title': unescapeHTML(e
.get('name')),
1469 'description': unescapeHTML(e
.get('description')),
1470 'thumbnails': [{'url': url_or_none(url)}
1471 for url
in variadic(traverse_obj(e
, 'thumbnailUrl', 'thumbnailURL'))],
1472 'duration': parse_duration(e
.get('duration')),
1473 'timestamp': unified_timestamp(e
.get('uploadDate')),
1474 # author can be an instance of 'Organization' or 'Person' types.
1475 # both types can have 'name' property(inherited from 'Thing' type). [1]
1476 # however some websites are using 'Text' type instead.
1477 # 1. https://schema.org/VideoObject
1478 'uploader': author
.get('name') if isinstance(author
, dict) else author
if isinstance(author
, compat_str
) else None,
1479 'filesize': float_or_none(e
.get('contentSize')),
1480 'tbr': int_or_none(e
.get('bitrate')),
1481 'width': int_or_none(e
.get('width')),
1482 'height': int_or_none(e
.get('height')),
1483 'view_count': int_or_none(e
.get('interactionCount')),
1485 extract_interaction_statistic(e
)
1486 extract_chapter_information(e
)
1488 def traverse_json_ld(json_ld
, at_top_level
=True):
1490 if at_top_level
and '@context' not in e
:
1492 if at_top_level
and set(e
.keys()) == {'@context', '@graph'}
:
1493 traverse_json_ld(variadic(e
['@graph'], allowed_types
=(dict,)), at_top_level
=False)
1495 item_type
= e
.get('@type')
1496 if expected_type
is not None and expected_type
!= item_type
:
1498 rating
= traverse_obj(e
, ('aggregateRating', 'ratingValue'), expected_type
=float_or_none
)
1499 if rating
is not None:
1500 info
['average_rating'] = rating
1501 if item_type
in ('TVEpisode', 'Episode'):
1502 episode_name
= unescapeHTML(e
.get('name'))
1504 'episode': episode_name
,
1505 'episode_number': int_or_none(e
.get('episodeNumber')),
1506 'description': unescapeHTML(e
.get('description')),
1508 if not info
.get('title') and episode_name
:
1509 info
['title'] = episode_name
1510 part_of_season
= e
.get('partOfSeason')
1511 if isinstance(part_of_season
, dict) and part_of_season
.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1513 'season': unescapeHTML(part_of_season
.get('name')),
1514 'season_number': int_or_none(part_of_season
.get('seasonNumber')),
1516 part_of_series
= e
.get('partOfSeries') or e
.get('partOfTVSeries')
1517 if isinstance(part_of_series
, dict) and part_of_series
.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1518 info
['series'] = unescapeHTML(part_of_series
.get('name'))
1519 elif item_type
== 'Movie':
1521 'title': unescapeHTML(e
.get('name')),
1522 'description': unescapeHTML(e
.get('description')),
1523 'duration': parse_duration(e
.get('duration')),
1524 'timestamp': unified_timestamp(e
.get('dateCreated')),
1526 elif item_type
in ('Article', 'NewsArticle'):
1528 'timestamp': parse_iso8601(e
.get('datePublished')),
1529 'title': unescapeHTML(e
.get('headline')),
1530 'description': unescapeHTML(e
.get('articleBody') or e
.get('description')),
1532 if traverse_obj(e
, ('video', 0, '@type')) == 'VideoObject':
1533 extract_video_object(e
['video'][0])
1534 elif item_type
== 'VideoObject':
1535 extract_video_object(e
)
1536 if expected_type
is None:
1540 video
= e
.get('video')
1541 if isinstance(video
, dict) and video
.get('@type') == 'VideoObject':
1542 extract_video_object(video
)
1543 if expected_type
is None:
1547 traverse_json_ld(json_ld
)
1549 return dict((k
, v
) for k
, v
in info
.items() if v
is not None)
1551 def _search_nextjs_data(self
, webpage
, video_id
, *, transform_source
=None, fatal
=True, **kw
):
1552 return self
._parse
_json
(
1554 r
'(?s)<script[^>]+id=[\'"]__NEXT_DATA__[\'"][^
>]*>([^
<]+)</script
>',
1555 webpage, 'next
.js data
', fatal=fatal, **kw),
1556 video_id, transform_source=transform_source, fatal=fatal)
1558 def _search_nuxt_data(self, webpage, video_id, context_name='__NUXT__
'):
1559 ''' Parses Nuxt.js metadata. This works as long as the function __NUXT__ invokes is a pure function. '''
1560 # not all website do this, but it can be changed
1561 # https://stackoverflow.com/questions/67463109/how-to-change-or-hide-nuxt-and-nuxt-keyword-in-page-source
1562 rectx = re.escape(context_name)
1563 js, arg_keys, arg_vals = self._search_regex(
1564 (r'<script
>window\
.%s=\
(function\
((?P
<arg_keys
>.*?
)\
)\{return\s(?P<js>\{.*?\}
)\
}\
((?P
<arg_vals
>.+?
)\
)\
);?
</script
>' % rectx,
1565 r'%s\
(.*?\
(function\
((?P
<arg_keys
>.*?
)\
)\{return\s(?P<js>\{.*?\}
)\
}\
((?P
<arg_vals
>.*?
)\
)' % rectx),
1566 webpage, context_name, group=['js
', 'arg_keys
', 'arg_vals
'])
1568 args = dict(zip(arg_keys.split(','), arg_vals.split(',')))
1570 for key, val in args.items():
1571 if val in ('undefined
', 'void
0'):
1574 return self._parse_json(js_to_json(js, args), video_id)['data
'][0]
1577 def _hidden_inputs(html):
1578 html = re.sub(r'<!--(?
:(?
!<!--).)*-->', '', html)
1580 for input in re.findall(r'(?i
)(<input[^
>]+>)', html):
1581 attrs = extract_attributes(input)
1584 if attrs.get('type') not in ('hidden
', 'submit
'):
1586 name = attrs.get('name
') or attrs.get('id')
1587 value = attrs.get('value
')
1588 if name and value is not None:
1589 hidden_inputs[name] = value
1590 return hidden_inputs
1592 def _form_hidden_inputs(self, form_id, html):
1593 form = self._search_regex(
1594 r'(?
is)<form
[^
>]+?
id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1595 html, '%s form' % form_id, group='form')
1596 return self._hidden_inputs(form)
1599 regex = r' *((?P<reverse>\+)?(?P<field>[a-zA-Z0-9_]+)((?P<separator>[~:])(?P<limit>.*?))?)? *$'
1601 default = ('hidden', 'aud_or_vid', 'hasvid', 'ie_pref', 'lang', 'quality',
1602 'res', 'fps', 'hdr:12', 'codec:vp9.2', 'size', 'br', 'asr',
1603 'proto', 'ext', 'hasaud', 'source', 'id') # These must not be aliases
1604 ytdl_default = ('hasaud', 'lang', 'quality', 'tbr', 'filesize', 'vbr',
1605 'height', 'width', 'proto', 'vext', 'abr', 'aext',
1606 'fps', 'fs_approx', 'source', 'id')
1609 'vcodec': {'type': 'ordered', 'regex': True,
1610 'order': ['av0?1', 'vp0?9.2', 'vp0?9', '[hx]265|he?vc?', '[hx]264|avc', 'vp0?8', 'mp4v|h263', 'theora', '', None, 'none']},
1611 'acodec': {'type': 'ordered', 'regex': True,
1612 'order': ['[af]lac', 'wav|aiff', 'opus', 'vorbis', 'aac', 'mp?4a?', 'mp3', 'e-?a?c-?3', 'ac-?3', 'dts', '', None, 'none']},
1613 'hdr': {'type': 'ordered', 'regex': True, 'field': 'dynamic_range',
1614 'order': ['dv', '(hdr)?12', r'(hdr)?10\+', '(hdr)?10', 'hlg', '', 'sdr', None]},
1615 'proto': {'type': 'ordered', 'regex': True, 'field': 'protocol',
1616 'order': ['(ht|f)tps', '(ht|f)tp$', 'm3u8.*', '.*dash', 'websocket_frag', 'rtmpe?', '', 'mms|rtsp', 'ws|websocket', 'f4']},
1617 'vext': {'type': 'ordered', 'field': 'video_ext',
1618 'order': ('mp4', 'webm', 'flv', '', 'none'),
1619 'order_free': ('webm', 'mp4', 'flv', '', 'none')},
1620 'aext': {'type': 'ordered', 'field': 'audio_ext',
1621 'order': ('m4a', 'aac', 'mp3', 'ogg', 'opus', 'webm', '', 'none'),
1622 'order_free': ('opus', 'ogg', 'webm', 'm4a', 'mp3', 'aac', '', 'none')},
1623 'hidden': {'visible': False, 'forced': True, 'type': 'extractor', 'max': -1000},
1624 'aud_or_vid': {'visible': False, 'forced': True, 'type': 'multiple',
1625 'field': ('vcodec', 'acodec'),
1626 'function': lambda it: int(any(v != 'none' for v in it))},
1627 'ie_pref': {'priority': True, 'type': 'extractor'},
1628 'hasvid': {'priority': True, 'field': 'vcodec', 'type': 'boolean', 'not_in_list': ('none',)},
1629 'hasaud': {'field': 'acodec', 'type': 'boolean', 'not_in_list': ('none',)},
1630 'lang': {'convert': 'float', 'field': 'language_preference', 'default': -1},
1631 'quality': {'convert': 'float', 'default': -1},
1632 'filesize': {'convert': 'bytes'},
1633 'fs_approx': {'convert': 'bytes', 'field': 'filesize_approx'},
1634 'id': {'convert': 'string', 'field': 'format_id'},
1635 'height': {'convert': 'float_none'},
1636 'width': {'convert': 'float_none'},
1637 'fps': {'convert': 'float_none'},
1638 'tbr': {'convert': 'float_none'},
1639 'vbr': {'convert': 'float_none'},
1640 'abr': {'convert': 'float_none'},
1641 'asr': {'convert': 'float_none'},
1642 'source': {'convert': 'float', 'field': 'source_preference', 'default': -1},
1644 'codec': {'type': 'combined', 'field': ('vcodec', 'acodec')},
1645 'br': {'type': 'combined', 'field': ('tbr', 'vbr', 'abr'), 'same_limit': True},
1646 'size': {'type': 'combined', 'same_limit': True, 'field': ('filesize', 'fs_approx')},
1647 'ext': {'type': 'combined', 'field': ('vext', 'aext')},
1648 'res': {'type': 'multiple', 'field': ('height', 'width'),
1649 'function': lambda it: (lambda l: min(l) if l else 0)(tuple(filter(None, it)))},
1651 # For compatibility with youtube-dl
1652 'format_id': {'type': 'alias', 'field': 'id'},
1653 'preference': {'type': 'alias', 'field': 'ie_pref'},
1654 'language_preference': {'type': 'alias', 'field': 'lang'},
1657 'dimension': {'type': 'alias', 'field': 'res'},
1658 'resolution': {'type': 'alias', 'field': 'res'},
1659 'extension': {'type': 'alias', 'field': 'ext'},
1660 'bitrate': {'type': 'alias', 'field': 'br'},
1661 'total_bitrate': {'type': 'alias', 'field': 'tbr'},
1662 'video_bitrate': {'type': 'alias', 'field': 'vbr'},
1663 'audio_bitrate': {'type': 'alias', 'field': 'abr'},
1664 'framerate': {'type': 'alias', 'field': 'fps'},
1665 'protocol': {'type': 'alias', 'field': 'proto'},
1666 'source_preference': {'type': 'alias', 'field': 'source'},
1667 'filesize_approx': {'type': 'alias', 'field': 'fs_approx'},
1668 'filesize_estimate': {'type': 'alias', 'field': 'size'},
1669 'samplerate': {'type': 'alias', 'field': 'asr'},
1670 'video_ext': {'type': 'alias', 'field': 'vext'},
1671 'audio_ext': {'type': 'alias', 'field': 'aext'},
1672 'video_codec': {'type': 'alias', 'field': 'vcodec'},
1673 'audio_codec': {'type': 'alias', 'field': 'acodec'},
1674 'video': {'type': 'alias', 'field': 'hasvid'},
1675 'has_video': {'type': 'alias', 'field': 'hasvid'},
1676 'audio': {'type': 'alias', 'field': 'hasaud'},
1677 'has_audio': {'type': 'alias', 'field': 'hasaud'},
1678 'extractor': {'type': 'alias', 'field': 'ie_pref'},
1679 'extractor_preference': {'type': 'alias', 'field': 'ie_pref'},
1682 def __init__(self, ie, field_preference):
1684 self.ydl = ie._downloader
1685 self.evaluate_params(self.ydl.params, field_preference)
1686 if ie.get_param('verbose'):
1687 self.print_verbose_info(self.ydl.write_debug)
1689 def _get_field_setting(self, field, key):
1690 if field not in self.settings:
1691 if key in ('forced', 'priority'):
1693 self.ydl.deprecation_warning(
1694 f'Using arbitrary fields ({field}) for format sorting is deprecated '
1695 'and may be removed in a future version')
1696 self.settings[field] = {}
1697 propObj = self.settings[field]
1698 if key not in propObj:
1699 type = propObj.get('type')
1701 default = 'preference' if type == 'extractor' else (field,) if type in ('combined', 'multiple') else field
1702 elif key == 'convert':
1703 default = 'order' if type == 'ordered' else 'float_string' if field else 'ignore'
1705 default = {'type': 'field', 'visible': True, 'order': [], 'not_in_list': (None,)}.get(key, None)
1706 propObj[key] = default
1709 def _resolve_field_value(self, field, value, convertNone=False):
1714 value = value.lower()
1715 conversion = self._get_field_setting(field, 'convert')
1716 if conversion == 'ignore':
1718 if conversion == 'string':
1720 elif conversion == 'float_none':
1721 return float_or_none(value)
1722 elif conversion == 'bytes':
1723 return FileDownloader.parse_bytes(value)
1724 elif conversion == 'order':
1725 order_list = (self._use_free_order and self._get_field_setting(field, 'order_free')) or self._get_field_setting(field, 'order')
1726 use_regex = self._get_field_setting(field, 'regex')
1727 list_length = len(order_list)
1728 empty_pos = order_list.index('') if '' in order_list else list_length + 1
1729 if use_regex and value is not None:
1730 for i, regex in enumerate(order_list):
1731 if regex and re.match(regex, value):
1732 return list_length - i
1733 return list_length - empty_pos # not in list
1734 else: # not regex or value = None
1735 return list_length - (order_list.index(value) if value in order_list else empty_pos)
1737 if value.isnumeric():
1740 self.settings[field]['convert'] = 'string'
1743 def evaluate_params(self, params, sort_extractor):
1744 self._use_free_order = params.get('prefer_free_formats', False)
1745 self._sort_user = params.get('format_sort', [])
1746 self._sort_extractor = sort_extractor
1748 def add_item(field, reverse, closest, limit_text):
1749 field = field.lower()
1750 if field in self._order:
1752 self._order.append(field)
1753 limit = self._resolve_field_value(field, limit_text)
1756 'closest': False if limit is None else closest,
1757 'limit_text': limit_text,
1759 if field in self.settings:
1760 self.settings[field].update(data)
1762 self.settings[field] = data
1765 tuple(field for field in self.default if self._get_field_setting(field, 'forced'))
1766 + (tuple() if params.get('format_sort_force', False)
1767 else tuple(field for field in self.default if self._get_field_setting(field, 'priority')))
1768 + tuple(self._sort_user) + tuple(sort_extractor) + self.default)
1770 for item in sort_list:
1771 match = re.match(self.regex, item)
1773 raise ExtractorError('Invalid format sort string "%s" given by extractor' % item)
1774 field = match.group('field')
1777 if self._get_field_setting(field, 'type') == 'alias':
1778 alias, field = field, self._get_field_setting(field, 'field')
1779 if alias not in ('format_id', 'preference', 'language_preference'):
1780 self.ydl.deprecation_warning(
1781 f'Format sorting alias {alias} is deprecated '
1782 f'and may be removed in a future version. Please use {field} instead')
1783 reverse = match.group('reverse') is not None
1784 closest = match.group('separator') == '~'
1785 limit_text = match.group('limit')
1787 has_limit = limit_text is not None
1788 has_multiple_fields = self._get_field_setting(field, 'type') == 'combined'
1789 has_multiple_limits = has_limit and has_multiple_fields and not self._get_field_setting(field, 'same_limit')
1791 fields = self._get_field_setting(field, 'field') if has_multiple_fields else (field,)
1792 limits = limit_text.split(':') if has_multiple_limits else (limit_text,) if has_limit else tuple()
1793 limit_count = len(limits)
1794 for (i, f) in enumerate(fields):
1795 add_item(f, reverse, closest,
1796 limits[i] if i < limit_count
1797 else limits[0] if has_limit and not has_multiple_limits
1800 def print_verbose_info(self, write_debug):
1802 write_debug('Sort order given by user: %s' % ', '.join(self._sort_user))
1803 if self._sort_extractor:
1804 write_debug('Sort order given by extractor: %s' % ', '.join(self._sort_extractor))
1805 write_debug('Formats sorted by: %s' % ', '.join(['%s%s%s' % (
1806 '+' if self._get_field_setting(field, 'reverse') else '', field,
1807 '%s%s(%s)' % ('~' if self._get_field_setting(field, 'closest') else ':',
1808 self._get_field_setting(field, 'limit_text'),
1809 self._get_field_setting(field, 'limit'))
1810 if self._get_field_setting(field, 'limit_text') is not None else '')
1811 for field in self._order if self._get_field_setting(field, 'visible')]))
1813 def _calculate_field_preference_from_value(self, format, field, type, value):
1814 reverse = self._get_field_setting(field, 'reverse')
1815 closest = self._get_field_setting(field, 'closest')
1816 limit = self._get_field_setting(field, 'limit')
1818 if type == 'extractor':
1819 maximum = self._get_field_setting(field, 'max')
1820 if value is None or (maximum is not None and value >= maximum):
1822 elif type == 'boolean':
1823 in_list = self._get_field_setting(field, 'in_list')
1824 not_in_list = self._get_field_setting(field, 'not_in_list')
1825 value = 0 if ((in_list is None or value in in_list) and (not_in_list is None or value not in not_in_list)) else -1
1826 elif type == 'ordered':
1827 value = self._resolve_field_value(field, value, True)
1829 # try to convert to number
1830 val_num = float_or_none(value, default=self._get_field_setting(field, 'default'))
1831 is_num = self._get_field_setting(field, 'convert') != 'string' and val_num is not None
1835 return ((-10, 0) if value is None
1836 else (1, value, 0) if not is_num # if a field has mixed strings and numbers, strings are sorted higher
1837 else (0, -abs(value - limit), value - limit if reverse else limit - value) if closest
1838 else (0, value, 0) if not reverse and (limit is None or value <= limit)
1839 else (0, -value, 0) if limit is None or (reverse and value == limit) or value > limit
1840 else (-1, value, 0))
1842 def _calculate_field_preference(self, format, field):
1843 type = self._get_field_setting(field, 'type') # extractor, boolean, ordered, field, multiple
1844 get_value = lambda f: format.get(self._get_field_setting(f, 'field'))
1845 if type == 'multiple':
1846 type = 'field' # Only 'field' is allowed in multiple for now
1847 actual_fields = self._get_field_setting(field, 'field')
1849 value = self._get_field_setting(field, 'function')(get_value(f) for f in actual_fields)
1851 value = get_value(field)
1852 return self._calculate_field_preference_from_value(format, field, type, value)
1854 def calculate_preference(self, format):
1855 # Determine missing protocol
1856 if not format.get('protocol'):
1857 format['protocol'] = determine_protocol(format)
1859 # Determine missing ext
1860 if not format.get('ext') and 'url' in format:
1861 format['ext'] = determine_ext(format['url'])
1862 if format.get('vcodec') == 'none':
1863 format['audio_ext'] = format['ext'] if format.get('acodec') != 'none' else 'none'
1864 format['video_ext'] = 'none'
1866 format['video_ext'] = format['ext']
1867 format['audio_ext'] = 'none'
1868 # if format.get('preference') is None and format.get('ext') in ('f4f', 'f4m'): # Not supported?
1869 # format['preference'] = -1000
1871 # Determine missing bitrates
1872 if format.get('tbr') is None:
1873 if format.get('vbr') is not None and format.get('abr') is not None:
1874 format['tbr'] = format.get('vbr', 0) + format.get('abr', 0)
1876 if format.get('vcodec') != 'none' and format.get('vbr') is None:
1877 format['vbr'] = format.get('tbr') - format.get('abr', 0)
1878 if format.get('acodec') != 'none' and format.get('abr') is None:
1879 format['abr'] = format.get('tbr') - format.get('vbr', 0)
1881 return tuple(self._calculate_field_preference(format, field) for field in self._order)
1883 def _sort_formats(self, formats, field_preference=[]):
1886 format_sort = self.FormatSort(self, field_preference)
1887 formats.sort(key=lambda f: format_sort.calculate_preference(f))
1889 def _check_formats(self, formats, video_id):
1891 formats[:] = filter(
1892 lambda f: self._is_valid_url(
1894 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1898 def _remove_duplicate_formats(formats):
1902 if f['url'] not in format_urls:
1903 format_urls.add(f['url'])
1904 unique_formats.append(f)
1905 formats[:] = unique_formats
1907 def _is_valid_url(self, url, video_id, item='video', headers={}):
1908 url = self._proto_relative_url(url, scheme='http:')
1909 # For now assume non HTTP(S) URLs always valid
1910 if not (url.startswith('http://') or url.startswith('https://')):
1913 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1915 except ExtractorError as e:
1917 '%s: %s URL is invalid, skipping: %s'
1918 % (video_id, item, error_to_compat_str(e.cause)))
1921 def http_scheme(self):
1922 """ Either "http
:" or "https
:", depending on the user's preferences """
1925 if self.get_param('prefer_insecure', False)
1928 def _proto_relative_url(self, url, scheme=None):
1931 if url.startswith('//'):
1933 scheme = self.http_scheme()
1938 def _sleep(self, timeout, video_id, msg_template=None):
1939 if msg_template is None:
1940 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1941 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1945 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1946 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1947 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
1948 manifest = self._download_xml(
1949 manifest_url, video_id, 'Downloading f4m manifest',
1950 'Unable to download f4m manifest',
1951 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1952 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
1953 transform_source=transform_source,
1954 fatal=fatal, data=data, headers=headers, query=query)
1956 if manifest is False:
1959 return self._parse_f4m_formats(
1960 manifest, manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1961 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1963 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1964 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1965 fatal=True, m3u8_id=None):
1966 if not isinstance(manifest, compat_etree_Element) and not fatal:
1969 # currently yt-dlp cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1970 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1971 if akamai_pv is not None and ';' in akamai_pv.text:
1972 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1973 if playerVerificationChallenge.strip() != '':
1977 manifest_version = '1.0'
1978 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1980 manifest_version = '2.0'
1981 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1982 # Remove unsupported DRM protected media from final formats
1983 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
1984 media_nodes = remove_encrypted_media(media_nodes)
1988 manifest_base_url = get_base_url(manifest)
1990 bootstrap_info = xpath_element(
1991 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1992 'bootstrap info', default=None)
1995 mime_type = xpath_text(
1996 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1997 'base URL', default=None)
1998 if mime_type and mime_type.startswith('audio/'):
2001 for i, media_el in enumerate(media_nodes):
2002 tbr = int_or_none(media_el.attrib.get('bitrate'))
2003 width = int_or_none(media_el.attrib.get('width'))
2004 height = int_or_none(media_el.attrib.get('height'))
2005 format_id = join_nonempty(f4m_id, tbr or i)
2006 # If <bootstrapInfo> is present, the specified f4m is a
2007 # stream-level manifest, and only set-level manifests may refer to
2008 # external resources. See section 11.4 and section 4 of F4M spec
2009 if bootstrap_info is None:
2011 # @href is introduced in 2.0, see section 11.6 of F4M spec
2012 if manifest_version == '2.0':
2013 media_url = media_el.attrib.get('href')
2014 if media_url is None:
2015 media_url = media_el.attrib.get('url')
2019 media_url if media_url.startswith('http://') or media_url.startswith('https://')
2020 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
2021 # If media_url is itself a f4m manifest do the recursive extraction
2022 # since bitrates in parent manifest (this one) and media_url manifest
2023 # may differ leading to inability to resolve the format by requested
2024 # bitrate in f4m downloader
2025 ext = determine_ext(manifest_url)
2027 f4m_formats = self._extract_f4m_formats(
2028 manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
2029 transform_source=transform_source, fatal=fatal)
2030 # Sometimes stream-level manifest contains single media entry that
2031 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
2032 # At the same time parent's media entry in set-level manifest may
2033 # contain it. We will copy it from parent in such cases.
2034 if len(f4m_formats) == 1:
2037 'tbr': f.get('tbr') or tbr,
2038 'width': f.get('width') or width,
2039 'height': f.get('height') or height,
2040 'format_id': f.get('format_id') if not tbr else format_id,
2043 formats.extend(f4m_formats)
2046 formats.extend(self._extract_m3u8_formats(
2047 manifest_url, video_id, 'mp4', preference=preference,
2048 quality=quality, m3u8_id=m3u8_id, fatal=fatal))
2051 'format_id': format_id,
2052 'url': manifest_url,
2053 'manifest_url': manifest_url,
2054 'ext': 'flv' if bootstrap_info is not None else None,
2060 'preference': preference,
2065 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, quality=None, m3u8_id=None):
2067 'format_id': join_nonempty(m3u8_id, 'meta'),
2071 'preference': preference - 100 if preference else -100,
2073 'resolution': 'multiple',
2074 'format_note': 'Quality selection URL',
2077 def _report_ignoring_subs(self, name):
2078 self.report_warning(bug_reports_message(
2079 f'Ignoring subtitle tracks found in the {name} manifest; '
2080 'if any subtitle tracks are missing,'
2083 def _extract_m3u8_formats(self, *args, **kwargs):
2084 fmts, subs = self._extract_m3u8_formats_and_subtitles(*args, **kwargs)
2086 self._report_ignoring_subs('HLS')
2089 def _extract_m3u8_formats_and_subtitles(
2090 self, m3u8_url, video_id, ext=None, entry_protocol='m3u8_native',
2091 preference=None, quality=None, m3u8_id=None, note=None,
2092 errnote=None, fatal=True, live=False, data=None, headers={},
2095 res = self._download_webpage_handle(
2097 note='Downloading m3u8 information' if note is None else note,
2098 errnote='Failed to download m3u8 information' if errnote is None else errnote,
2099 fatal=fatal, data=data, headers=headers, query=query)
2104 m3u8_doc, urlh = res
2105 m3u8_url = urlh.geturl()
2107 return self._parse_m3u8_formats_and_subtitles(
2108 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
2109 preference=preference, quality=quality, m3u8_id=m3u8_id,
2110 note=note, errnote=errnote, fatal=fatal, live=live, data=data,
2111 headers=headers, query=query, video_id=video_id)
2113 def _parse_m3u8_formats_and_subtitles(
2114 self, m3u8_doc, m3u8_url=None, ext=None, entry_protocol='m3u8_native',
2115 preference=None, quality=None, m3u8_id=None, live=False, note=None,
2116 errnote=None, fatal=True, data=None, headers={}, query={},
2118 formats, subtitles = [], {}
2120 has_drm = re.search('|'.join([
2121 r'#EXT-X-FAXS-CM:', # Adobe Flash Access
2122 r'#EXT-X-(?:SESSION-)?KEY:.*?URI="skd
://', # Apple FairPlay
2125 def format_url(url):
2126 return url if re.match(r'^https?
://', url) else compat_urlparse.urljoin(m3u8_url, url)
2128 if self.get_param('hls_split_discontinuity
', False):
2129 def _extract_m3u8_playlist_indices(manifest_url=None, m3u8_doc=None):
2131 if not manifest_url:
2133 m3u8_doc = self._download_webpage(
2134 manifest_url, video_id, fatal=fatal, data=data, headers=headers,
2135 note=False, errnote='Failed to download m3u8 playlist information
')
2136 if m3u8_doc is False:
2138 return range(1 + sum(line.startswith('#EXT-X-DISCONTINUITY') for line in m3u8_doc.splitlines()))
2141 def _extract_m3u8_playlist_indices(*args
, **kwargs
):
2145 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
2146 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
2147 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
2149 # We should try extracting formats only from master playlists [1, 4.3.4],
2150 # i.e. playlists that describe available qualities. On the other hand
2151 # media playlists [1, 4.3.3] should be returned as is since they contain
2152 # just the media without qualities renditions.
2153 # Fortunately, master playlist can be easily distinguished from media
2154 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
2155 # master playlist tags MUST NOT appear in a media playlist and vice versa.
2156 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
2157 # media playlist and MUST NOT appear in master playlist thus we can
2158 # clearly detect media playlist with this criterion.
2160 if '#EXT-X-TARGETDURATION' in m3u8_doc
: # media playlist, return as is
2162 'format_id': join_nonempty(m3u8_id
, idx
),
2163 'format_index': idx
,
2164 'url': m3u8_url
or encode_data_uri(m3u8_doc
.encode('utf-8'), 'application/x-mpegurl'),
2166 'protocol': entry_protocol
,
2167 'preference': preference
,
2170 } for idx
in _extract_m3u8_playlist_indices(m3u8_doc
=m3u8_doc
)]
2172 return formats
, subtitles
2175 last_stream_inf
= {}
2177 def extract_media(x_media_line
):
2178 media
= parse_m3u8_attributes(x_media_line
)
2179 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
2180 media_type
, group_id
, name
= media
.get('TYPE'), media
.get('GROUP-ID'), media
.get('NAME')
2181 if not (media_type
and group_id
and name
):
2183 groups
.setdefault(group_id
, []).append(media
)
2184 # <https://tools.ietf.org/html/rfc8216#section-4.3.4.1>
2185 if media_type
== 'SUBTITLES':
2186 # According to RFC 8216 §4.3.4.2.1, URI is REQUIRED in the
2187 # EXT-X-MEDIA tag if the media type is SUBTITLES.
2188 # However, lack of URI has been spotted in the wild.
2189 # e.g. NebulaIE; see https://github.com/yt-dlp/yt-dlp/issues/339
2190 if not media
.get('URI'):
2192 url
= format_url(media
['URI'])
2195 'ext': determine_ext(url
),
2197 if sub_info
['ext'] == 'm3u8':
2198 # Per RFC 8216 §3.1, the only possible subtitle format m3u8
2199 # files may contain is WebVTT:
2200 # <https://tools.ietf.org/html/rfc8216#section-3.1>
2201 sub_info
['ext'] = 'vtt'
2202 sub_info
['protocol'] = 'm3u8_native'
2203 lang
= media
.get('LANGUAGE') or 'und'
2204 subtitles
.setdefault(lang
, []).append(sub_info
)
2205 if media_type
not in ('VIDEO', 'AUDIO'):
2207 media_url
= media
.get('URI')
2209 manifest_url
= format_url(media_url
)
2211 'format_id': join_nonempty(m3u8_id
, group_id
, name
, idx
),
2212 'format_note': name
,
2213 'format_index': idx
,
2214 'url': manifest_url
,
2215 'manifest_url': m3u8_url
,
2216 'language': media
.get('LANGUAGE'),
2218 'protocol': entry_protocol
,
2219 'preference': preference
,
2221 'vcodec': 'none' if media_type
== 'AUDIO' else None,
2222 } for idx
in _extract_m3u8_playlist_indices(manifest_url
))
2224 def build_stream_name():
2225 # Despite specification does not mention NAME attribute for
2226 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
2227 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
2228 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
2229 stream_name
= last_stream_inf
.get('NAME')
2232 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
2233 # from corresponding rendition group
2234 stream_group_id
= last_stream_inf
.get('VIDEO')
2235 if not stream_group_id
:
2237 stream_group
= groups
.get(stream_group_id
)
2238 if not stream_group
:
2239 return stream_group_id
2240 rendition
= stream_group
[0]
2241 return rendition
.get('NAME') or stream_group_id
2243 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
2244 # chance to detect video only formats when EXT-X-STREAM-INF tags
2245 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
2246 for line
in m3u8_doc
.splitlines():
2247 if line
.startswith('#EXT-X-MEDIA:'):
2250 for line
in m3u8_doc
.splitlines():
2251 if line
.startswith('#EXT-X-STREAM-INF:'):
2252 last_stream_inf
= parse_m3u8_attributes(line
)
2253 elif line
.startswith('#') or not line
.strip():
2256 tbr
= float_or_none(
2257 last_stream_inf
.get('AVERAGE-BANDWIDTH')
2258 or last_stream_inf
.get('BANDWIDTH'), scale
=1000)
2259 manifest_url
= format_url(line
.strip())
2261 for idx
in _extract_m3u8_playlist_indices(manifest_url
):
2262 format_id
= [m3u8_id
, None, idx
]
2263 # Bandwidth of live streams may differ over time thus making
2264 # format_id unpredictable. So it's better to keep provided
2267 stream_name
= build_stream_name()
2268 format_id
[1] = stream_name
or '%d' % (tbr
or len(formats
))
2270 'format_id': join_nonempty(*format_id
),
2271 'format_index': idx
,
2272 'url': manifest_url
,
2273 'manifest_url': m3u8_url
,
2276 'fps': float_or_none(last_stream_inf
.get('FRAME-RATE')),
2277 'protocol': entry_protocol
,
2278 'preference': preference
,
2281 resolution
= last_stream_inf
.get('RESOLUTION')
2283 mobj
= re
.search(r
'(?P<width>\d+)[xX](?P<height>\d+)', resolution
)
2285 f
['width'] = int(mobj
.group('width'))
2286 f
['height'] = int(mobj
.group('height'))
2287 # Unified Streaming Platform
2289 r
'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f
['url'])
2291 abr
, vbr
= mobj
.groups()
2292 abr
, vbr
= float_or_none(abr
, 1000), float_or_none(vbr
, 1000)
2297 codecs
= parse_codecs(last_stream_inf
.get('CODECS'))
2299 audio_group_id
= last_stream_inf
.get('AUDIO')
2300 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
2301 # references a rendition group MUST have a CODECS attribute.
2302 # However, this is not always respected, for example, [2]
2303 # contains EXT-X-STREAM-INF tag which references AUDIO
2304 # rendition group but does not have CODECS and despite
2305 # referencing an audio group it represents a complete
2306 # (with audio and video) format. So, for such cases we will
2307 # ignore references to rendition groups and treat them
2308 # as complete formats.
2309 if audio_group_id
and codecs
and f
.get('vcodec') != 'none':
2310 audio_group
= groups
.get(audio_group_id
)
2311 if audio_group
and audio_group
[0].get('URI'):
2312 # TODO: update acodec for audio only formats with
2314 f
['acodec'] = 'none'
2315 if not f
.get('ext'):
2316 f
['ext'] = 'm4a' if f
.get('vcodec') == 'none' else 'mp4'
2320 progressive_uri
= last_stream_inf
.get('PROGRESSIVE-URI')
2323 del http_f
['manifest_url']
2325 'format_id': f
['format_id'].replace('hls-', 'http-'),
2327 'url': progressive_uri
,
2329 formats
.append(http_f
)
2331 last_stream_inf
= {}
2332 return formats
, subtitles
2334 def _extract_m3u8_vod_duration(
2335 self
, m3u8_vod_url
, video_id
, note
=None, errnote
=None, data
=None, headers
={}, query={}
):
2337 m3u8_vod
= self
._download
_webpage
(
2338 m3u8_vod_url
, video_id
,
2339 note
='Downloading m3u8 VOD manifest' if note
is None else note
,
2340 errnote
='Failed to download VOD manifest' if errnote
is None else errnote
,
2341 fatal
=False, data
=data
, headers
=headers
, query
=query
)
2343 return self
._parse
_m
3u8_vod
_duration
(m3u8_vod
or '', video_id
)
2345 def _parse_m3u8_vod_duration(self
, m3u8_vod
, video_id
):
2346 if '#EXT-X-PLAYLIST-TYPE:VOD' not in m3u8_vod
:
2350 float(line
[len('#EXTINF:'):].split(',')[0])
2351 for line
in m3u8_vod
.splitlines() if line
.startswith('#EXTINF:'))) or None
2354 def _xpath_ns(path
, namespace
=None):
2358 for c
in path
.split('/'):
2359 if not c
or c
== '.':
2362 out
.append('{%s}%s' % (namespace
, c
))
2363 return '/'.join(out
)
2365 def _extract_smil_formats_and_subtitles(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None, transform_source
=None):
2366 smil
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
, transform_source
=transform_source
)
2372 namespace
= self
._parse
_smil
_namespace
(smil
)
2374 fmts
= self
._parse
_smil
_formats
(
2375 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2376 subs
= self
._parse
_smil
_subtitles
(
2377 smil
, namespace
=namespace
)
2381 def _extract_smil_formats(self
, *args
, **kwargs
):
2382 fmts
, subs
= self
._extract
_smil
_formats
_and
_subtitles
(*args
, **kwargs
)
2384 self
._report
_ignoring
_subs
('SMIL')
2387 def _extract_smil_info(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None):
2388 smil
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
)
2391 return self
._parse
_smil
(smil
, smil_url
, video_id
, f4m_params
=f4m_params
)
2393 def _download_smil(self
, smil_url
, video_id
, fatal
=True, transform_source
=None):
2394 return self
._download
_xml
(
2395 smil_url
, video_id
, 'Downloading SMIL file',
2396 'Unable to download SMIL file', fatal
=fatal
, transform_source
=transform_source
)
2398 def _parse_smil(self
, smil
, smil_url
, video_id
, f4m_params
=None):
2399 namespace
= self
._parse
_smil
_namespace
(smil
)
2401 formats
= self
._parse
_smil
_formats
(
2402 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2403 subtitles
= self
._parse
_smil
_subtitles
(smil
, namespace
=namespace
)
2405 video_id
= os
.path
.splitext(url_basename(smil_url
))[0]
2409 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2410 name
= meta
.attrib
.get('name')
2411 content
= meta
.attrib
.get('content')
2412 if not name
or not content
:
2414 if not title
and name
== 'title':
2416 elif not description
and name
in ('description', 'abstract'):
2417 description
= content
2418 elif not upload_date
and name
== 'date':
2419 upload_date
= unified_strdate(content
)
2422 'id': image
.get('type'),
2423 'url': image
.get('src'),
2424 'width': int_or_none(image
.get('width')),
2425 'height': int_or_none(image
.get('height')),
2426 } for image
in smil
.findall(self
._xpath
_ns
('.//image', namespace
)) if image
.get('src')]
2430 'title': title
or video_id
,
2431 'description': description
,
2432 'upload_date': upload_date
,
2433 'thumbnails': thumbnails
,
2435 'subtitles': subtitles
,
2438 def _parse_smil_namespace(self
, smil
):
2439 return self
._search
_regex
(
2440 r
'(?i)^{([^}]+)?}smil$', smil
.tag
, 'namespace', default
=None)
2442 def _parse_smil_formats(self
, smil
, smil_url
, video_id
, namespace
=None, f4m_params
=None, transform_rtmp_url
=None):
2444 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2445 b
= meta
.get('base') or meta
.get('httpBase')
2457 media
= smil
.findall(self
._xpath
_ns
('.//video', namespace
)) + smil
.findall(self
._xpath
_ns
('.//audio', namespace
))
2458 for medium
in media
:
2459 src
= medium
.get('src')
2460 if not src
or src
in srcs
:
2464 bitrate
= float_or_none(medium
.get('system-bitrate') or medium
.get('systemBitrate'), 1000)
2465 filesize
= int_or_none(medium
.get('size') or medium
.get('fileSize'))
2466 width
= int_or_none(medium
.get('width'))
2467 height
= int_or_none(medium
.get('height'))
2468 proto
= medium
.get('proto')
2469 ext
= medium
.get('ext')
2470 src_ext
= determine_ext(src
)
2471 streamer
= medium
.get('streamer') or base
2473 if proto
== 'rtmp' or streamer
.startswith('rtmp'):
2479 'format_id': 'rtmp-%d' % (rtmp_count
if bitrate
is None else bitrate
),
2481 'filesize': filesize
,
2485 if transform_rtmp_url
:
2486 streamer
, src
= transform_rtmp_url(streamer
, src
)
2487 formats
[-1].update({
2493 src_url
= src
if src
.startswith('http') else compat_urlparse
.urljoin(base
, src
)
2494 src_url
= src_url
.strip()
2496 if proto
== 'm3u8' or src_ext
== 'm3u8':
2497 m3u8_formats
= self
._extract
_m
3u8_formats
(
2498 src_url
, video_id
, ext
or 'mp4', m3u8_id
='hls', fatal
=False)
2499 if len(m3u8_formats
) == 1:
2501 m3u8_formats
[0].update({
2502 'format_id': 'hls-%d' % (m3u8_count
if bitrate
is None else bitrate
),
2507 formats
.extend(m3u8_formats
)
2508 elif src_ext
== 'f4m':
2513 'plugin': 'flowplayer-3.2.0.1',
2515 f4m_url
+= '&' if '?' in f4m_url
else '?'
2516 f4m_url
+= compat_urllib_parse_urlencode(f4m_params
)
2517 formats
.extend(self
._extract
_f
4m
_formats
(f4m_url
, video_id
, f4m_id
='hds', fatal
=False))
2518 elif src_ext
== 'mpd':
2519 formats
.extend(self
._extract
_mpd
_formats
(
2520 src_url
, video_id
, mpd_id
='dash', fatal
=False))
2521 elif re
.search(r
'\.ism/[Mm]anifest', src_url
):
2522 formats
.extend(self
._extract
_ism
_formats
(
2523 src_url
, video_id
, ism_id
='mss', fatal
=False))
2524 elif src_url
.startswith('http') and self
._is
_valid
_url
(src
, video_id
):
2528 'ext': ext
or src_ext
or 'flv',
2529 'format_id': 'http-%d' % (bitrate
or http_count
),
2531 'filesize': filesize
,
2536 for medium
in smil
.findall(self
._xpath
_ns
('.//imagestream', namespace
)):
2537 src
= medium
.get('src')
2538 if not src
or src
in srcs
:
2544 'format_id': 'imagestream-%d' % (imgs_count
),
2546 'ext': mimetype2ext(medium
.get('type')),
2549 'width': int_or_none(medium
.get('width')),
2550 'height': int_or_none(medium
.get('height')),
2551 'format_note': 'SMIL storyboards',
2556 def _parse_smil_subtitles(self
, smil
, namespace
=None, subtitles_lang
='en'):
2559 for num
, textstream
in enumerate(smil
.findall(self
._xpath
_ns
('.//textstream', namespace
))):
2560 src
= textstream
.get('src')
2561 if not src
or src
in urls
:
2564 ext
= textstream
.get('ext') or mimetype2ext(textstream
.get('type')) or determine_ext(src
)
2565 lang
= textstream
.get('systemLanguage') or textstream
.get('systemLanguageName') or textstream
.get('lang') or subtitles_lang
2566 subtitles
.setdefault(lang
, []).append({
2572 def _extract_xspf_playlist(self
, xspf_url
, playlist_id
, fatal
=True):
2573 xspf
= self
._download
_xml
(
2574 xspf_url
, playlist_id
, 'Downloading xpsf playlist',
2575 'Unable to download xspf manifest', fatal
=fatal
)
2578 return self
._parse
_xspf
(
2579 xspf
, playlist_id
, xspf_url
=xspf_url
,
2580 xspf_base_url
=base_url(xspf_url
))
2582 def _parse_xspf(self
, xspf_doc
, playlist_id
, xspf_url
=None, xspf_base_url
=None):
2584 'xspf': 'http://xspf.org/ns/0/',
2585 's1': 'http://static.streamone.nl/player/ns/0',
2589 for track
in xspf_doc
.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP
)):
2591 track
, xpath_with_ns('./xspf:title', NS_MAP
), 'title', default
=playlist_id
)
2592 description
= xpath_text(
2593 track
, xpath_with_ns('./xspf:annotation', NS_MAP
), 'description')
2594 thumbnail
= xpath_text(
2595 track
, xpath_with_ns('./xspf:image', NS_MAP
), 'thumbnail')
2596 duration
= float_or_none(
2597 xpath_text(track
, xpath_with_ns('./xspf:duration', NS_MAP
), 'duration'), 1000)
2600 for location
in track
.findall(xpath_with_ns('./xspf:location', NS_MAP
)):
2601 format_url
= urljoin(xspf_base_url
, location
.text
)
2606 'manifest_url': xspf_url
,
2607 'format_id': location
.get(xpath_with_ns('s1:label', NS_MAP
)),
2608 'width': int_or_none(location
.get(xpath_with_ns('s1:width', NS_MAP
))),
2609 'height': int_or_none(location
.get(xpath_with_ns('s1:height', NS_MAP
))),
2611 self
._sort
_formats
(formats
)
2616 'description': description
,
2617 'thumbnail': thumbnail
,
2618 'duration': duration
,
2623 def _extract_mpd_formats(self
, *args
, **kwargs
):
2624 fmts
, subs
= self
._extract
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2626 self
._report
_ignoring
_subs
('DASH')
2629 def _extract_mpd_formats_and_subtitles(
2630 self
, mpd_url
, video_id
, mpd_id
=None, note
=None, errnote
=None,
2631 fatal
=True, data
=None, headers
={}, query={}
):
2632 res
= self
._download
_xml
_handle
(
2634 note
='Downloading MPD manifest' if note
is None else note
,
2635 errnote
='Failed to download MPD manifest' if errnote
is None else errnote
,
2636 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
2642 mpd_base_url
= base_url(urlh
.geturl())
2644 return self
._parse
_mpd
_formats
_and
_subtitles
(
2645 mpd_doc
, mpd_id
, mpd_base_url
, mpd_url
)
2647 def _parse_mpd_formats(self
, *args
, **kwargs
):
2648 fmts
, subs
= self
._parse
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2650 self
._report
_ignoring
_subs
('DASH')
2653 def _parse_mpd_formats_and_subtitles(
2654 self
, mpd_doc
, mpd_id
=None, mpd_base_url
='', mpd_url
=None):
2656 Parse formats from MPD manifest.
2658 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
2659 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
2660 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
2662 if not self
.get_param('dynamic_mpd', True):
2663 if mpd_doc
.get('type') == 'dynamic':
2666 namespace
= self
._search
_regex
(r
'(?i)^{([^}]+)?}MPD$', mpd_doc
.tag
, 'namespace', default
=None)
2669 return self
._xpath
_ns
(path
, namespace
)
2671 def is_drm_protected(element
):
2672 return element
.find(_add_ns('ContentProtection')) is not None
2674 def extract_multisegment_info(element
, ms_parent_info
):
2675 ms_info
= ms_parent_info
.copy()
2677 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2678 # common attributes and elements. We will only extract relevant
2680 def extract_common(source
):
2681 segment_timeline
= source
.find(_add_ns('SegmentTimeline'))
2682 if segment_timeline
is not None:
2683 s_e
= segment_timeline
.findall(_add_ns('S'))
2685 ms_info
['total_number'] = 0
2688 r
= int(s
.get('r', 0))
2689 ms_info
['total_number'] += 1 + r
2690 ms_info
['s'].append({
2691 't': int(s
.get('t', 0)),
2692 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2693 'd': int(s
.attrib
['d']),
2696 start_number
= source
.get('startNumber')
2698 ms_info
['start_number'] = int(start_number
)
2699 timescale
= source
.get('timescale')
2701 ms_info
['timescale'] = int(timescale
)
2702 segment_duration
= source
.get('duration')
2703 if segment_duration
:
2704 ms_info
['segment_duration'] = float(segment_duration
)
2706 def extract_Initialization(source
):
2707 initialization
= source
.find(_add_ns('Initialization'))
2708 if initialization
is not None:
2709 ms_info
['initialization_url'] = initialization
.attrib
['sourceURL']
2711 segment_list
= element
.find(_add_ns('SegmentList'))
2712 if segment_list
is not None:
2713 extract_common(segment_list
)
2714 extract_Initialization(segment_list
)
2715 segment_urls_e
= segment_list
.findall(_add_ns('SegmentURL'))
2717 ms_info
['segment_urls'] = [segment
.attrib
['media'] for segment
in segment_urls_e
]
2719 segment_template
= element
.find(_add_ns('SegmentTemplate'))
2720 if segment_template
is not None:
2721 extract_common(segment_template
)
2722 media
= segment_template
.get('media')
2724 ms_info
['media'] = media
2725 initialization
= segment_template
.get('initialization')
2727 ms_info
['initialization'] = initialization
2729 extract_Initialization(segment_template
)
2732 mpd_duration
= parse_duration(mpd_doc
.get('mediaPresentationDuration'))
2733 formats
, subtitles
= [], {}
2734 stream_numbers
= collections
.defaultdict(int)
2735 for period
in mpd_doc
.findall(_add_ns('Period')):
2736 period_duration
= parse_duration(period
.get('duration')) or mpd_duration
2737 period_ms_info
= extract_multisegment_info(period
, {
2741 for adaptation_set
in period
.findall(_add_ns('AdaptationSet')):
2742 adaption_set_ms_info
= extract_multisegment_info(adaptation_set
, period_ms_info
)
2743 for representation
in adaptation_set
.findall(_add_ns('Representation')):
2744 representation_attrib
= adaptation_set
.attrib
.copy()
2745 representation_attrib
.update(representation
.attrib
)
2746 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2747 mime_type
= representation_attrib
['mimeType']
2748 content_type
= representation_attrib
.get('contentType', mime_type
.split('/')[0])
2750 codecs
= parse_codecs(representation_attrib
.get('codecs', ''))
2751 if content_type
not in ('video', 'audio', 'text'):
2752 if mime_type
== 'image/jpeg':
2753 content_type
= mime_type
2754 elif codecs
['vcodec'] != 'none':
2755 content_type
= 'video'
2756 elif codecs
['acodec'] != 'none':
2757 content_type
= 'audio'
2758 elif codecs
.get('tcodec', 'none') != 'none':
2759 content_type
= 'text'
2760 elif mimetype2ext(mime_type
) in ('tt', 'dfxp', 'ttml', 'xml', 'json'):
2761 content_type
= 'text'
2763 self
.report_warning('Unknown MIME type %s in DASH manifest' % mime_type
)
2767 for element
in (representation
, adaptation_set
, period
, mpd_doc
):
2768 base_url_e
= element
.find(_add_ns('BaseURL'))
2769 if base_url_e
is not None:
2770 base_url
= base_url_e
.text
+ base_url
2771 if re
.match(r
'^https?://', base_url
):
2773 if mpd_base_url
and base_url
.startswith('/'):
2774 base_url
= compat_urlparse
.urljoin(mpd_base_url
, base_url
)
2775 elif mpd_base_url
and not re
.match(r
'^https?://', base_url
):
2776 if not mpd_base_url
.endswith('/'):
2778 base_url
= mpd_base_url
+ base_url
2779 representation_id
= representation_attrib
.get('id')
2780 lang
= representation_attrib
.get('lang')
2781 url_el
= representation
.find(_add_ns('BaseURL'))
2782 filesize
= int_or_none(url_el
.attrib
.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el
is not None else None)
2783 bandwidth
= int_or_none(representation_attrib
.get('bandwidth'))
2784 if representation_id
is not None:
2785 format_id
= representation_id
2787 format_id
= content_type
2789 format_id
= mpd_id
+ '-' + format_id
2790 if content_type
in ('video', 'audio'):
2792 'format_id': format_id
,
2793 'manifest_url': mpd_url
,
2794 'ext': mimetype2ext(mime_type
),
2795 'width': int_or_none(representation_attrib
.get('width')),
2796 'height': int_or_none(representation_attrib
.get('height')),
2797 'tbr': float_or_none(bandwidth
, 1000),
2798 'asr': int_or_none(representation_attrib
.get('audioSamplingRate')),
2799 'fps': int_or_none(representation_attrib
.get('frameRate')),
2800 'language': lang
if lang
not in ('mul', 'und', 'zxx', 'mis') else None,
2801 'format_note': 'DASH %s' % content_type
,
2802 'filesize': filesize
,
2803 'container': mimetype2ext(mime_type
) + '_dash',
2806 elif content_type
== 'text':
2808 'ext': mimetype2ext(mime_type
),
2809 'manifest_url': mpd_url
,
2810 'filesize': filesize
,
2812 elif content_type
== 'image/jpeg':
2813 # See test case in VikiIE
2814 # https://www.viki.com/videos/1175236v-choosing-spouse-by-lottery-episode-1
2816 'format_id': format_id
,
2818 'manifest_url': mpd_url
,
2819 'format_note': 'DASH storyboards (jpeg)',
2823 if is_drm_protected(adaptation_set
) or is_drm_protected(representation
):
2825 representation_ms_info
= extract_multisegment_info(representation
, adaption_set_ms_info
)
2827 def prepare_template(template_name
, identifiers
):
2828 tmpl
= representation_ms_info
[template_name
]
2829 # First of, % characters outside $...$ templates
2830 # must be escaped by doubling for proper processing
2831 # by % operator string formatting used further (see
2832 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2838 in_template
= not in_template
2839 elif c
== '%' and not in_template
:
2841 # Next, $...$ templates are translated to their
2842 # %(...) counterparts to be used with % operator
2843 if representation_id
is not None:
2844 t
= t
.replace('$RepresentationID$', representation_id
)
2845 t
= re
.sub(r
'\$(%s)\$' % '|'.join(identifiers
), r
'%(\1)d', t
)
2846 t
= re
.sub(r
'\$(%s)%%([^$]+)\$' % '|'.join(identifiers
), r
'%(\1)\2', t
)
2847 t
.replace('$$', '$')
2850 # @initialization is a regular template like @media one
2851 # so it should be handled just the same way (see
2852 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2853 if 'initialization' in representation_ms_info
:
2854 initialization_template
= prepare_template(
2856 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2857 # $Time$ shall not be included for @initialization thus
2858 # only $Bandwidth$ remains
2860 representation_ms_info
['initialization_url'] = initialization_template
% {
2861 'Bandwidth': bandwidth
,
2864 def location_key(location
):
2865 return 'url' if re
.match(r
'^https?://', location
) else 'path'
2867 if 'segment_urls' not in representation_ms_info
and 'media' in representation_ms_info
:
2869 media_template
= prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2870 media_location_key
= location_key(media_template
)
2872 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2873 # can't be used at the same time
2874 if '%(Number' in media_template
and 's' not in representation_ms_info
:
2875 segment_duration
= None
2876 if 'total_number' not in representation_ms_info
and 'segment_duration' in representation_ms_info
:
2877 segment_duration
= float_or_none(representation_ms_info
['segment_duration'], representation_ms_info
['timescale'])
2878 representation_ms_info
['total_number'] = int(math
.ceil(float(period_duration
) / segment_duration
))
2879 representation_ms_info
['fragments'] = [{
2880 media_location_key
: media_template
% {
2881 'Number': segment_number
,
2882 'Bandwidth': bandwidth
,
2884 'duration': segment_duration
,
2885 } for segment_number
in range(
2886 representation_ms_info
['start_number'],
2887 representation_ms_info
['total_number'] + representation_ms_info
['start_number'])]
2889 # $Number*$ or $Time$ in media template with S list available
2890 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2891 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2892 representation_ms_info
['fragments'] = []
2895 segment_number
= representation_ms_info
['start_number']
2897 def add_segment_url():
2898 segment_url
= media_template
% {
2899 'Time': segment_time
,
2900 'Bandwidth': bandwidth
,
2901 'Number': segment_number
,
2903 representation_ms_info
['fragments'].append({
2904 media_location_key
: segment_url
,
2905 'duration': float_or_none(segment_d
, representation_ms_info
['timescale']),
2908 for num
, s
in enumerate(representation_ms_info
['s']):
2909 segment_time
= s
.get('t') or segment_time
2913 for r
in range(s
.get('r', 0)):
2914 segment_time
+= segment_d
2917 segment_time
+= segment_d
2918 elif 'segment_urls' in representation_ms_info
and 's' in representation_ms_info
:
2920 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2921 # or any YouTube dashsegments video
2924 timescale
= representation_ms_info
['timescale']
2925 for s
in representation_ms_info
['s']:
2926 duration
= float_or_none(s
['d'], timescale
)
2927 for r
in range(s
.get('r', 0) + 1):
2928 segment_uri
= representation_ms_info
['segment_urls'][segment_index
]
2930 location_key(segment_uri
): segment_uri
,
2931 'duration': duration
,
2934 representation_ms_info
['fragments'] = fragments
2935 elif 'segment_urls' in representation_ms_info
:
2936 # Segment URLs with no SegmentTimeline
2937 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2938 # https://github.com/ytdl-org/youtube-dl/pull/14844
2940 segment_duration
= float_or_none(
2941 representation_ms_info
['segment_duration'],
2942 representation_ms_info
['timescale']) if 'segment_duration' in representation_ms_info
else None
2943 for segment_url
in representation_ms_info
['segment_urls']:
2945 location_key(segment_url
): segment_url
,
2947 if segment_duration
:
2948 fragment
['duration'] = segment_duration
2949 fragments
.append(fragment
)
2950 representation_ms_info
['fragments'] = fragments
2951 # If there is a fragments key available then we correctly recognized fragmented media.
2952 # Otherwise we will assume unfragmented media with direct access. Technically, such
2953 # assumption is not necessarily correct since we may simply have no support for
2954 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2955 if 'fragments' in representation_ms_info
:
2957 # NB: mpd_url may be empty when MPD manifest is parsed from a string
2958 'url': mpd_url
or base_url
,
2959 'fragment_base_url': base_url
,
2961 'protocol': 'http_dash_segments' if mime_type
!= 'image/jpeg' else 'mhtml',
2963 if 'initialization_url' in representation_ms_info
:
2964 initialization_url
= representation_ms_info
['initialization_url']
2965 if not f
.get('url'):
2966 f
['url'] = initialization_url
2967 f
['fragments'].append({location_key(initialization_url): initialization_url}
)
2968 f
['fragments'].extend(representation_ms_info
['fragments'])
2970 # Assuming direct URL to unfragmented media.
2972 if content_type
in ('video', 'audio', 'image/jpeg'):
2973 f
['manifest_stream_number'] = stream_numbers
[f
['url']]
2974 stream_numbers
[f
['url']] += 1
2976 elif content_type
== 'text':
2977 subtitles
.setdefault(lang
or 'und', []).append(f
)
2979 return formats
, subtitles
2981 def _extract_ism_formats(self
, *args
, **kwargs
):
2982 fmts
, subs
= self
._extract
_ism
_formats
_and
_subtitles
(*args
, **kwargs
)
2984 self
._report
_ignoring
_subs
('ISM')
2987 def _extract_ism_formats_and_subtitles(self
, ism_url
, video_id
, ism_id
=None, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
):
2988 res
= self
._download
_xml
_handle
(
2990 note
='Downloading ISM manifest' if note
is None else note
,
2991 errnote
='Failed to download ISM manifest' if errnote
is None else errnote
,
2992 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
2999 return self
._parse
_ism
_formats
_and
_subtitles
(ism_doc
, urlh
.geturl(), ism_id
)
3001 def _parse_ism_formats_and_subtitles(self
, ism_doc
, ism_url
, ism_id
=None):
3003 Parse formats from ISM manifest.
3005 1. [MS-SSTR]: Smooth Streaming Protocol,
3006 https://msdn.microsoft.com/en-us/library/ff469518.aspx
3008 if ism_doc
.get('IsLive') == 'TRUE':
3011 duration
= int(ism_doc
.attrib
['Duration'])
3012 timescale
= int_or_none(ism_doc
.get('TimeScale')) or 10000000
3016 for stream
in ism_doc
.findall('StreamIndex'):
3017 stream_type
= stream
.get('Type')
3018 if stream_type
not in ('video', 'audio', 'text'):
3020 url_pattern
= stream
.attrib
['Url']
3021 stream_timescale
= int_or_none(stream
.get('TimeScale')) or timescale
3022 stream_name
= stream
.get('Name')
3023 stream_language
= stream
.get('Language', 'und')
3024 for track
in stream
.findall('QualityLevel'):
3025 fourcc
= track
.get('FourCC') or ('AACL' if track
.get('AudioTag') == '255' else None)
3026 # TODO: add support for WVC1 and WMAP
3027 if fourcc
not in ('H264', 'AVC1', 'AACL', 'TTML'):
3028 self
.report_warning('%s is not a supported codec' % fourcc
)
3030 tbr
= int(track
.attrib
['Bitrate']) // 1000
3031 # [1] does not mention Width and Height attributes. However,
3032 # they're often present while MaxWidth and MaxHeight are
3033 # missing, so should be used as fallbacks
3034 width
= int_or_none(track
.get('MaxWidth') or track
.get('Width'))
3035 height
= int_or_none(track
.get('MaxHeight') or track
.get('Height'))
3036 sampling_rate
= int_or_none(track
.get('SamplingRate'))
3038 track_url_pattern
= re
.sub(r
'{[Bb]itrate}', track
.attrib
['Bitrate'], url_pattern
)
3039 track_url_pattern
= compat_urlparse
.urljoin(ism_url
, track_url_pattern
)
3045 stream_fragments
= stream
.findall('c')
3046 for stream_fragment_index
, stream_fragment
in enumerate(stream_fragments
):
3047 fragment_ctx
['time'] = int_or_none(stream_fragment
.get('t')) or fragment_ctx
['time']
3048 fragment_repeat
= int_or_none(stream_fragment
.get('r')) or 1
3049 fragment_ctx
['duration'] = int_or_none(stream_fragment
.get('d'))
3050 if not fragment_ctx
['duration']:
3052 next_fragment_time
= int(stream_fragment
[stream_fragment_index
+ 1].attrib
['t'])
3054 next_fragment_time
= duration
3055 fragment_ctx
['duration'] = (next_fragment_time
- fragment_ctx
['time']) / fragment_repeat
3056 for _
in range(fragment_repeat
):
3058 'url': re
.sub(r
'{start[ _]time}', compat_str(fragment_ctx
['time']), track_url_pattern
),
3059 'duration': fragment_ctx
['duration'] / stream_timescale
,
3061 fragment_ctx
['time'] += fragment_ctx
['duration']
3063 if stream_type
== 'text':
3064 subtitles
.setdefault(stream_language
, []).append({
3068 'manifest_url': ism_url
,
3069 'fragments': fragments
,
3070 '_download_params': {
3071 'stream_type': stream_type
,
3072 'duration': duration
,
3073 'timescale': stream_timescale
,
3075 'language': stream_language
,
3076 'codec_private_data': track
.get('CodecPrivateData'),
3079 elif stream_type
in ('video', 'audio'):
3081 'format_id': join_nonempty(ism_id
, stream_name
, tbr
),
3083 'manifest_url': ism_url
,
3084 'ext': 'ismv' if stream_type
== 'video' else 'isma',
3088 'asr': sampling_rate
,
3089 'vcodec': 'none' if stream_type
== 'audio' else fourcc
,
3090 'acodec': 'none' if stream_type
== 'video' else fourcc
,
3092 'fragments': fragments
,
3093 'has_drm': ism_doc
.find('Protection') is not None,
3094 '_download_params': {
3095 'stream_type': stream_type
,
3096 'duration': duration
,
3097 'timescale': stream_timescale
,
3098 'width': width
or 0,
3099 'height': height
or 0,
3101 'language': stream_language
,
3102 'codec_private_data': track
.get('CodecPrivateData'),
3103 'sampling_rate': sampling_rate
,
3104 'channels': int_or_none(track
.get('Channels', 2)),
3105 'bits_per_sample': int_or_none(track
.get('BitsPerSample', 16)),
3106 'nal_unit_length_field': int_or_none(track
.get('NALUnitLengthField', 4)),
3109 return formats
, subtitles
3111 def _parse_html5_media_entries(self
, base_url
, webpage
, video_id
, m3u8_id
=None, m3u8_entry_protocol
='m3u8', mpd_id
=None, preference
=None, quality
=None):
3112 def absolute_url(item_url
):
3113 return urljoin(base_url
, item_url
)
3115 def parse_content_type(content_type
):
3116 if not content_type
:
3118 ctr
= re
.search(r
'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type
)
3120 mimetype
, codecs
= ctr
.groups()
3121 f
= parse_codecs(codecs
)
3122 f
['ext'] = mimetype2ext(mimetype
)
3126 def _media_formats(src
, cur_media_type
, type_info
={}):
3127 full_url
= absolute_url(src
)
3128 ext
= type_info
.get('ext') or determine_ext(full_url
)
3130 is_plain_url
= False
3131 formats
= self
._extract
_m
3u8_formats
(
3132 full_url
, video_id
, ext
='mp4',
3133 entry_protocol
=m3u8_entry_protocol
, m3u8_id
=m3u8_id
,
3134 preference
=preference
, quality
=quality
, fatal
=False)
3136 is_plain_url
= False
3137 formats
= self
._extract
_mpd
_formats
(
3138 full_url
, video_id
, mpd_id
=mpd_id
, fatal
=False)
3143 'vcodec': 'none' if cur_media_type
== 'audio' else None,
3145 return is_plain_url
, formats
3148 # amp-video and amp-audio are very similar to their HTML5 counterparts
3149 # so we wll include them right here (see
3150 # https://www.ampproject.org/docs/reference/components/amp-video)
3151 # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/
3152 _MEDIA_TAG_NAME_RE
= r
'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)'
3153 media_tags
= [(media_tag
, media_tag_name
, media_type
, '')
3154 for media_tag
, media_tag_name
, media_type
3155 in re
.findall(r
'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE
, webpage
)]
3156 media_tags
.extend(re
.findall(
3157 # We only allow video|audio followed by a whitespace or '>'.
3158 # Allowing more characters may end up in significant slow down (see
3159 # https://github.com/ytdl-org/youtube-dl/issues/11979, example URL:
3160 # http://www.porntrex.com/maps/videositemap.xml).
3161 r
'(?s)(<(?P<tag>%s)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>' % _MEDIA_TAG_NAME_RE
, webpage
))
3162 for media_tag
, _
, media_type
, media_content
in media_tags
:
3167 media_attributes
= extract_attributes(media_tag
)
3168 src
= strip_or_none(media_attributes
.get('src'))
3170 _
, formats
= _media_formats(src
, media_type
)
3171 media_info
['formats'].extend(formats
)
3172 media_info
['thumbnail'] = absolute_url(media_attributes
.get('poster'))
3174 for source_tag
in re
.findall(r
'<source[^>]+>', media_content
):
3175 s_attr
= extract_attributes(source_tag
)
3176 # data-video-src and data-src are non standard but seen
3177 # several times in the wild
3178 src
= strip_or_none(dict_get(s_attr
, ('src', 'data-video-src', 'data-src')))
3181 f
= parse_content_type(s_attr
.get('type'))
3182 is_plain_url
, formats
= _media_formats(src
, media_type
, f
)
3184 # width, height, res, label and title attributes are
3185 # all not standard but seen several times in the wild
3188 for lbl
in ('label', 'title')
3189 if str_or_none(s_attr
.get(lbl
))
3191 width
= int_or_none(s_attr
.get('width'))
3192 height
= (int_or_none(s_attr
.get('height'))
3193 or int_or_none(s_attr
.get('res')))
3194 if not width
or not height
:
3196 resolution
= parse_resolution(lbl
)
3199 width
= width
or resolution
.get('width')
3200 height
= height
or resolution
.get('height')
3202 tbr
= parse_bitrate(lbl
)
3211 'format_id': s_attr
.get('label') or s_attr
.get('title'),
3213 f
.update(formats
[0])
3214 media_info
['formats'].append(f
)
3216 media_info
['formats'].extend(formats
)
3217 for track_tag
in re
.findall(r
'<track[^>]+>', media_content
):
3218 track_attributes
= extract_attributes(track_tag
)
3219 kind
= track_attributes
.get('kind')
3220 if not kind
or kind
in ('subtitles', 'captions'):
3221 src
= strip_or_none(track_attributes
.get('src'))
3224 lang
= track_attributes
.get('srclang') or track_attributes
.get('lang') or track_attributes
.get('label')
3225 media_info
['subtitles'].setdefault(lang
, []).append({
3226 'url': absolute_url(src
),
3228 for f
in media_info
['formats']:
3229 f
.setdefault('http_headers', {})['Referer'] = base_url
3230 if media_info
['formats'] or media_info
['subtitles']:
3231 entries
.append(media_info
)
3234 def _extract_akamai_formats(self
, *args
, **kwargs
):
3235 fmts
, subs
= self
._extract
_akamai
_formats
_and
_subtitles
(*args
, **kwargs
)
3237 self
._report
_ignoring
_subs
('akamai')
3240 def _extract_akamai_formats_and_subtitles(self
, manifest_url
, video_id
, hosts
={}):
3241 signed
= 'hdnea=' in manifest_url
3243 # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
3244 manifest_url
= re
.sub(
3245 r
'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
3246 '', manifest_url
).strip('?')
3251 hdcore_sign
= 'hdcore=3.7.0'
3252 f4m_url
= re
.sub(r
'(https?://[^/]+)/i/', r
'\1/z/', manifest_url
).replace('/master.m3u8', '/manifest.f4m')
3253 hds_host
= hosts
.get('hds')
3255 f4m_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hds_host
, f4m_url
)
3256 if 'hdcore=' not in f4m_url
:
3257 f4m_url
+= ('&' if '?' in f4m_url
else '?') + hdcore_sign
3258 f4m_formats
= self
._extract
_f
4m
_formats
(
3259 f4m_url
, video_id
, f4m_id
='hds', fatal
=False)
3260 for entry
in f4m_formats
:
3261 entry
.update({'extra_param_to_segment_url': hdcore_sign}
)
3262 formats
.extend(f4m_formats
)
3264 m3u8_url
= re
.sub(r
'(https?://[^/]+)/z/', r
'\1/i/', manifest_url
).replace('/manifest.f4m', '/master.m3u8')
3265 hls_host
= hosts
.get('hls')
3267 m3u8_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hls_host
, m3u8_url
)
3268 m3u8_formats
, m3u8_subtitles
= self
._extract
_m
3u8_formats
_and
_subtitles
(
3269 m3u8_url
, video_id
, 'mp4', 'm3u8_native',
3270 m3u8_id
='hls', fatal
=False)
3271 formats
.extend(m3u8_formats
)
3272 subtitles
= self
._merge
_subtitles
(subtitles
, m3u8_subtitles
)
3274 http_host
= hosts
.get('http')
3275 if http_host
and m3u8_formats
and not signed
:
3276 REPL_REGEX
= r
'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
3277 qualities
= re
.match(REPL_REGEX
, m3u8_url
).group(2).split(',')
3278 qualities_length
= len(qualities
)
3279 if len(m3u8_formats
) in (qualities_length
, qualities_length
+ 1):
3281 for f
in m3u8_formats
:
3282 if f
['vcodec'] != 'none':
3283 for protocol
in ('http', 'https'):
3285 del http_f
['manifest_url']
3287 REPL_REGEX
, protocol
+ r
'://%s/\g<1>%s\3' % (http_host
, qualities
[i
]), f
['url'])
3289 'format_id': http_f
['format_id'].replace('hls-', protocol
+ '-'),
3291 'protocol': protocol
,
3293 formats
.append(http_f
)
3296 return formats
, subtitles
3298 def _extract_wowza_formats(self
, url
, video_id
, m3u8_entry_protocol
='m3u8_native', skip_protocols
=[]):
3299 query
= compat_urlparse
.urlparse(url
).query
3300 url
= re
.sub(r
'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url
)
3302 r
'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url
)
3303 url_base
= mobj
.group('url')
3304 http_base_url
= '%s%s:%s' % ('http', mobj
.group('s') or '', url_base
)
3307 def manifest_url(manifest
):
3308 m_url
= '%s/%s' % (http_base_url
, manifest
)
3310 m_url
+= '?%s' % query
3313 if 'm3u8' not in skip_protocols
:
3314 formats
.extend(self
._extract
_m
3u8_formats
(
3315 manifest_url('playlist.m3u8'), video_id
, 'mp4',
3316 m3u8_entry_protocol
, m3u8_id
='hls', fatal
=False))
3317 if 'f4m' not in skip_protocols
:
3318 formats
.extend(self
._extract
_f
4m
_formats
(
3319 manifest_url('manifest.f4m'),
3320 video_id
, f4m_id
='hds', fatal
=False))
3321 if 'dash' not in skip_protocols
:
3322 formats
.extend(self
._extract
_mpd
_formats
(
3323 manifest_url('manifest.mpd'),
3324 video_id
, mpd_id
='dash', fatal
=False))
3325 if re
.search(r
'(?:/smil:|\.smil)', url_base
):
3326 if 'smil' not in skip_protocols
:
3327 rtmp_formats
= self
._extract
_smil
_formats
(
3328 manifest_url('jwplayer.smil'),
3329 video_id
, fatal
=False)
3330 for rtmp_format
in rtmp_formats
:
3331 rtsp_format
= rtmp_format
.copy()
3332 rtsp_format
['url'] = '%s/%s' % (rtmp_format
['url'], rtmp_format
['play_path'])
3333 del rtsp_format
['play_path']
3334 del rtsp_format
['ext']
3335 rtsp_format
.update({
3336 'url': rtsp_format
['url'].replace('rtmp://', 'rtsp://'),
3337 'format_id': rtmp_format
['format_id'].replace('rtmp', 'rtsp'),
3340 formats
.extend([rtmp_format
, rtsp_format
])
3342 for protocol
in ('rtmp', 'rtsp'):
3343 if protocol
not in skip_protocols
:
3345 'url': '%s:%s' % (protocol
, url_base
),
3346 'format_id': protocol
,
3347 'protocol': protocol
,
3351 def _find_jwplayer_data(self
, webpage
, video_id
=None, transform_source
=js_to_json
):
3353 r
'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P
=quote
)\
)(?
!</script
>).*?\
.setup\s
*\
((?P
<options
>[^
)]+)\
)',
3357 jwplayer_data = self._parse_json(mobj.group('options
'),
3359 transform_source=transform_source)
3360 except ExtractorError:
3363 if isinstance(jwplayer_data, dict):
3364 return jwplayer_data
3366 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
3367 jwplayer_data = self._find_jwplayer_data(
3368 webpage, video_id, transform_source=js_to_json)
3369 return self._parse_jwplayer_data(
3370 jwplayer_data, video_id, *args, **kwargs)
3372 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
3373 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3374 # JWPlayer backward compatibility: flattened playlists
3375 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
3376 if 'playlist
' not in jwplayer_data:
3377 jwplayer_data = {'playlist': [jwplayer_data]}
3381 # JWPlayer backward compatibility: single playlist item
3382 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
3383 if not isinstance(jwplayer_data['playlist
'], list):
3384 jwplayer_data['playlist
'] = [jwplayer_data['playlist
']]
3386 for video_data in jwplayer_data['playlist
']:
3387 # JWPlayer backward compatibility: flattened sources
3388 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
3389 if 'sources
' not in video_data:
3390 video_data['sources
'] = [video_data]
3392 this_video_id = video_id or video_data['mediaid
']
3394 formats = self._parse_jwplayer_formats(
3395 video_data['sources
'], video_id=this_video_id, m3u8_id=m3u8_id,
3396 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
3399 tracks = video_data.get('tracks
')
3400 if tracks and isinstance(tracks, list):
3401 for track in tracks:
3402 if not isinstance(track, dict):
3404 track_kind = track.get('kind
')
3405 if not track_kind or not isinstance(track_kind, compat_str):
3407 if track_kind.lower() not in ('captions
', 'subtitles
'):
3409 track_url = urljoin(base_url, track.get('file'))
3412 subtitles.setdefault(track.get('label
') or 'en
', []).append({
3413 'url
': self._proto_relative_url(track_url)
3417 'id': this_video_id,
3418 'title
': unescapeHTML(video_data['title
'] if require_title else video_data.get('title
')),
3419 'description
': clean_html(video_data.get('description
')),
3420 'thumbnail
': urljoin(base_url, self._proto_relative_url(video_data.get('image
'))),
3421 'timestamp
': int_or_none(video_data.get('pubdate
')),
3422 'duration
': float_or_none(jwplayer_data.get('duration
') or video_data.get('duration
')),
3423 'subtitles
': subtitles,
3425 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
3426 if len(formats) == 1 and re.search(r'^
(?
:http|
//).*(?
:youtube\
.com|youtu\
.be
)/.+', formats[0]['url
']):
3428 '_type
': 'url_transparent
',
3429 'url
': formats[0]['url
'],
3432 self._sort_formats(formats)
3433 entry['formats
'] = formats
3434 entries.append(entry)
3435 if len(entries) == 1:
3438 return self.playlist_result(entries)
3440 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
3441 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3444 for source in jwplayer_sources_data:
3445 if not isinstance(source, dict):
3447 source_url = urljoin(
3448 base_url, self._proto_relative_url(source.get('file')))
3449 if not source_url or source_url in urls:
3451 urls.append(source_url)
3452 source_type = source.get('type') or ''
3453 ext = mimetype2ext(source_type) or determine_ext(source_url)
3454 if source_type == 'hls
' or ext == 'm3u8
':
3455 formats.extend(self._extract_m3u8_formats(
3456 source_url, video_id, 'mp4
', entry_protocol='m3u8_native
',
3457 m3u8_id=m3u8_id, fatal=False))
3458 elif source_type == 'dash
' or ext == 'mpd
':
3459 formats.extend(self._extract_mpd_formats(
3460 source_url, video_id, mpd_id=mpd_id, fatal=False))
3462 formats.extend(self._extract_smil_formats(
3463 source_url, video_id, fatal=False))
3464 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3465 elif source_type.startswith('audio
') or ext in (
3466 'oga
', 'aac
', 'mp3
', 'mpeg
', 'vorbis
'):
3473 height = int_or_none(source.get('height
'))
3475 # Often no height is provided but there is a label in
3476 # format like "1080p", "720p SD", or 1080.
3477 height = int_or_none(self._search_regex(
3478 r'^
(\d{3,4}
)[pP
]?
(?
:\b|$
)', compat_str(source.get('label
') or ''),
3479 'height
', default=None))
3482 'width
': int_or_none(source.get('width
')),
3484 'tbr
': int_or_none(source.get('bitrate
')),
3487 if source_url.startswith('rtmp
'):
3488 a_format['ext
'] = 'flv
'
3489 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3490 # of jwplayer.flash.swf
3491 rtmp_url_parts = re.split(
3492 r'((?
:mp4|mp3|flv
):)', source_url, 1)
3493 if len(rtmp_url_parts) == 3:
3494 rtmp_url, prefix, play_path = rtmp_url_parts
3497 'play_path
': prefix + play_path,
3500 a_format.update(rtmp_params)
3501 formats.append(a_format)
3504 def _live_title(self, name):
3505 self._downloader.deprecation_warning('yt_dlp
.InfoExtractor
._live
_title
is deprecated
and does
not work
as expected
')
3508 def _int(self, v, name, fatal=False, **kwargs):
3509 res = int_or_none(v, **kwargs)
3511 msg = 'Failed to extract
%s: Could
not parse value
%r' % (name, v)
3513 raise ExtractorError(msg)
3515 self.report_warning(msg)
3518 def _float(self, v, name, fatal=False, **kwargs):
3519 res = float_or_none(v, **kwargs)
3521 msg = 'Failed to extract
%s: Could
not parse value
%r' % (name, v)
3523 raise ExtractorError(msg)
3525 self.report_warning(msg)
3528 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
3529 path='/', secure=False, discard=False, rest={}, **kwargs):
3530 cookie = compat_cookiejar_Cookie(
3531 0, name, value, port, port is not None, domain, True,
3532 domain.startswith('.'), path, True, secure, expire_time,
3533 discard, None, None, rest)
3534 self._downloader.cookiejar.set_cookie(cookie)
3536 def _get_cookies(self, url):
3537 """ Return a compat_cookies_SimpleCookie with the cookies for the url """
3538 req = sanitized_Request(url)
3539 self._downloader.cookiejar.add_cookie_header(req)
3540 return compat_cookies_SimpleCookie(req.get_header('Cookie
'))
3542 def _apply_first_set_cookie_header(self, url_handle, cookie):
3544 Apply first Set-Cookie header instead of the last. Experimental.
3546 Some sites (e.g. [1-3]) may serve two cookies under the same name
3547 in Set-Cookie header and expect the first (old) one to be set rather
3548 than second (new). However, as of RFC6265 the newer one cookie
3549 should be set into cookie store what actually happens.
3550 We will workaround this issue by resetting the cookie to
3551 the first one manually.
3552 1. https://new.vk.com/
3553 2. https://github.com/ytdl-org/youtube-dl/issues/9841#issuecomment-227871201
3554 3. https://learning.oreilly.com/
3556 for header, cookies in url_handle.headers.items():
3557 if header.lower() != 'set-cookie
':
3559 if sys.version_info[0] >= 3:
3560 cookies = cookies.encode('iso
-8859-1')
3561 cookies = cookies.decode('utf
-8')
3562 cookie_value = re.search(
3563 r'%s=(.+?
);.*?
\b[Dd
]omain
=(.+?
)(?
:[,;]|$
)' % cookie, cookies)
3565 value, domain = cookie_value.groups()
3566 self._set_cookie(domain, cookie, value)
3569 def get_testcases(self, include_onlymatching=False):
3570 t = getattr(self, '_TEST
', None)
3572 assert not hasattr(self, '_TESTS
'), \
3573 '%s has _TEST
and _TESTS
' % type(self).__name__
3576 tests = getattr(self, '_TESTS
', [])
3578 if not include_onlymatching and t.get('only_matching
', False):
3580 t['name
'] = type(self).__name__[:-len('IE
')]
3583 def is_suitable(self, age_limit):
3584 """ Test whether the extractor is generally suitable for the given
3585 age limit (i.e. pornographic sites are not, all others usually are) """
3587 any_restricted = False
3588 for tc in self.get_testcases(include_onlymatching=False):
3589 if tc.get('playlist
', []):
3590 tc = tc['playlist
'][0]
3591 is_restricted = age_restricted(
3592 tc.get('info_dict
', {}).get('age_limit
'), age_limit)
3593 if not is_restricted:
3595 any_restricted = any_restricted or is_restricted
3596 return not any_restricted
3598 def extract_subtitles(self, *args, **kwargs):
3599 if (self.get_param('writesubtitles
', False)
3600 or self.get_param('listsubtitles
')):
3601 return self._get_subtitles(*args, **kwargs)
3604 def _get_subtitles(self, *args, **kwargs):
3605 raise NotImplementedError('This method must be implemented by subclasses
')
3607 def extract_comments(self, *args, **kwargs):
3608 if not self.get_param('getcomments
'):
3610 generator = self._get_comments(*args, **kwargs)
3617 comments.append(next(generator))
3618 except StopIteration:
3620 except KeyboardInterrupt:
3621 self.to_screen('Interrupted by user
')
3622 except Exception as e:
3623 if self.get_param('ignoreerrors
') is not True:
3625 self._downloader.report_error(e)
3626 comment_count = len(comments)
3627 self.to_screen(f'Extracted {comment_count} comments
')
3629 'comments
': comments,
3630 'comment_count
': None if interrupted else comment_count
3634 def _get_comments(self, *args, **kwargs):
3635 raise NotImplementedError('This method must be implemented by subclasses
')
3638 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
3639 """ Merge subtitle items for one language. Items with duplicated URLs
3640 will be dropped. """
3641 list1_urls = set([item['url
'] for item in subtitle_list1])
3642 ret = list(subtitle_list1)
3643 ret.extend([item for item in subtitle_list2 if item['url
'] not in list1_urls])
3647 def _merge_subtitles(cls, *dicts, target=None):
3648 """ Merge subtitle dictionaries, language by language. """
3652 for lang, subs in d.items():
3653 target[lang] = cls._merge_subtitle_items(target.get(lang, []), subs)
3656 def extract_automatic_captions(self, *args, **kwargs):
3657 if (self.get_param('writeautomaticsub
', False)
3658 or self.get_param('listsubtitles
')):
3659 return self._get_automatic_captions(*args, **kwargs)
3662 def _get_automatic_captions(self, *args, **kwargs):
3663 raise NotImplementedError('This method must be implemented by subclasses
')
3665 def mark_watched(self, *args, **kwargs):
3666 if not self.get_param('mark_watched
', False):
3668 if (self._get_login_info()[0] is not None
3669 or self.get_param('cookiefile
')
3670 or self.get_param('cookiesfrombrowser
')):
3671 self._mark_watched(*args, **kwargs)
3673 def _mark_watched(self, *args, **kwargs):
3674 raise NotImplementedError('This method must be implemented by subclasses
')
3676 def geo_verification_headers(self):
3678 geo_verification_proxy = self.get_param('geo_verification_proxy
')
3679 if geo_verification_proxy:
3680 headers['Ytdl
-request
-proxy
'] = geo_verification_proxy
3683 def _generic_id(self, url):
3684 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
3686 def _generic_title(self, url):
3687 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
3690 def _availability(is_private=None, needs_premium=None, needs_subscription=None, needs_auth=None, is_unlisted=None):
3691 all_known = all(map(
3692 lambda x: x is not None,
3693 (is_private, needs_premium, needs_subscription, needs_auth, is_unlisted)))
3695 'private
' if is_private
3696 else 'premium_only
' if needs_premium
3697 else 'subscriber_only
' if needs_subscription
3698 else 'needs_auth
' if needs_auth
3699 else 'unlisted
' if is_unlisted
3700 else 'public
' if all_known
3703 def _configuration_arg(self, key, default=NO_DEFAULT, *, ie_key=None, casesense=False):
3705 @returns A list of values for the extractor argument given by "key"
3706 or "default" if no such key is present
3707 @param default The default value to return when the key is not present (default: [])
3708 @param casesense When false, the values are converted to lower case
3711 self._downloader.params, ('extractor_args
', (ie_key or self.ie_key()).lower(), key))
3713 return [] if default is NO_DEFAULT else default
3714 return list(val) if casesense else [x.lower() for x in val]
3716 def _yes_playlist(self, playlist_id, video_id, smuggled_data=None, *, playlist_label='playlist
', video_label='video
'):
3717 if not playlist_id or not video_id:
3720 no_playlist = (smuggled_data or {}).get('force_noplaylist
')
3721 if no_playlist is not None:
3722 return not no_playlist
3724 video_id = '' if video_id is True else f' {video_id}
'
3725 playlist_id = '' if playlist_id is True else f' {playlist_id}
'
3726 if self.get_param('noplaylist
'):
3727 self.to_screen(f'Downloading just the {video_label}{video_id} because of
--no
-playlist
')
3729 self.to_screen(f'Downloading {playlist_label}{playlist_id}
- add
--no
-playlist to download just the {video_label}{video_id}
')
3733 class SearchInfoExtractor(InfoExtractor):
3735 Base class for paged search queries extractors.
3736 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
3737 Instances should define _SEARCH_KEY and optionally _MAX_RESULTS
3740 _MAX_RESULTS = float('inf
')
3743 def _make_valid_url(cls):
3744 return r'%s(?P
<prefix
>|
[1-9][0-9]*|all
):(?P
<query
>[\s\S
]+)' % cls._SEARCH_KEY
3746 def _real_extract(self, query):
3747 prefix, query = self._match_valid_url(query).group('prefix
', 'query
')
3749 return self._get_n_results(query, 1)
3750 elif prefix == 'all
':
3751 return self._get_n_results(query, self._MAX_RESULTS)
3755 raise ExtractorError('invalid download number
%s for query
"%s"' % (n, query))
3756 elif n > self._MAX_RESULTS:
3757 self.report_warning('%s returns
max %i results (you requested
%i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3758 n = self._MAX_RESULTS
3759 return self._get_n_results(query, n)
3761 def _get_n_results(self, query, n):
3762 """Get a specified number of results for a query.
3763 Either this function or _search_results must be overridden by subclasses """
3764 return self.playlist_result(
3765 itertools.islice(self._search_results(query), 0, None if n == float('inf
') else n),
3768 def _search_results(self, query):
3769 """Returns an iterator of search results"""
3770 raise NotImplementedError('This method must be implemented by subclasses
')
3773 def SEARCH_KEY(self):
3774 return self._SEARCH_KEY