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 return self._og_search_property('title
', html, **kargs)
1296 def _og_search_video_url(self, html, name='video url
', secure=True, **kargs):
1297 regexes = self._og_regexes('video
') + self._og_regexes('video
:url
')
1299 regexes = self._og_regexes('video
:secure_url
') + regexes
1300 return self._html_search_regex(regexes, html, name, **kargs)
1302 def _og_search_url(self, html, **kargs):
1303 return self._og_search_property('url
', html, **kargs)
1305 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1306 name = variadic(name)
1307 if display_name is None:
1308 display_name = name[0]
1309 return self._html_search_regex(
1310 [self._meta_regex(n) for n in name],
1311 html, display_name, fatal=fatal, group='content
', **kwargs)
1313 def _dc_search_uploader(self, html):
1314 return self._html_search_meta('dc
.creator
', html, 'uploader
')
1316 def _rta_search(self, html):
1317 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1318 if re.search(r'(?ix
)<meta\s
+name
="rating"\s
+'
1319 r' content
="RTA-5042-1996-1400-1577-RTA"',
1324 def _media_rating_search(self, html):
1325 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1326 rating = self._html_search_meta('rating
', html)
1338 return RATING_TABLE.get(rating.lower())
1340 def _family_friendly_search(self, html):
1341 # See http://schema.org/VideoObject
1342 family_friendly = self._html_search_meta(
1343 'isFamilyFriendly
', html, default=None)
1345 if not family_friendly:
1354 return RATING_TABLE.get(family_friendly.lower())
1356 def _twitter_search_player(self, html):
1357 return self._html_search_meta('twitter
:player
', html,
1358 'twitter card player
')
1360 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
1361 json_ld_list = list(re.finditer(JSON_LD_RE, html))
1362 default = kwargs.get('default
', NO_DEFAULT)
1363 # JSON-LD may be malformed and thus `fatal` should be respected.
1364 # At the same time `default` may be passed that assumes `fatal=False`
1365 # for _search_regex. Let's simulate the same behavior here
as well
.
1366 fatal
= kwargs
.get('fatal', True) if default
is NO_DEFAULT
else False
1368 for mobj
in json_ld_list
:
1369 json_ld_item
= self
._parse
_json
(
1370 mobj
.group('json_ld'), video_id
, fatal
=fatal
)
1371 if not json_ld_item
:
1373 if isinstance(json_ld_item
, dict):
1374 json_ld
.append(json_ld_item
)
1375 elif isinstance(json_ld_item
, (list, tuple)):
1376 json_ld
.extend(json_ld_item
)
1378 json_ld
= self
._json
_ld
(json_ld
, video_id
, fatal
=fatal
, expected_type
=expected_type
)
1381 if default
is not NO_DEFAULT
:
1384 raise RegexNotFoundError('Unable to extract JSON-LD')
1386 self
.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1389 def _json_ld(self
, json_ld
, video_id
, fatal
=True, expected_type
=None):
1390 if isinstance(json_ld
, compat_str
):
1391 json_ld
= self
._parse
_json
(json_ld
, video_id
, fatal
=fatal
)
1395 if not isinstance(json_ld
, (list, tuple, dict)):
1397 if isinstance(json_ld
, dict):
1400 INTERACTION_TYPE_MAP
= {
1401 'CommentAction': 'comment',
1402 'AgreeAction': 'like',
1403 'DisagreeAction': 'dislike',
1404 'LikeAction': 'like',
1405 'DislikeAction': 'dislike',
1406 'ListenAction': 'view',
1407 'WatchAction': 'view',
1408 'ViewAction': 'view',
1411 def extract_interaction_type(e
):
1412 interaction_type
= e
.get('interactionType')
1413 if isinstance(interaction_type
, dict):
1414 interaction_type
= interaction_type
.get('@type')
1415 return str_or_none(interaction_type
)
1417 def extract_interaction_statistic(e
):
1418 interaction_statistic
= e
.get('interactionStatistic')
1419 if isinstance(interaction_statistic
, dict):
1420 interaction_statistic
= [interaction_statistic
]
1421 if not isinstance(interaction_statistic
, list):
1423 for is_e
in interaction_statistic
:
1424 if not isinstance(is_e
, dict):
1426 if is_e
.get('@type') != 'InteractionCounter':
1428 interaction_type
= extract_interaction_type(is_e
)
1429 if not interaction_type
:
1431 # For interaction count some sites provide string instead of
1432 # an integer (as per spec) with non digit characters (e.g. ",")
1433 # so extracting count with more relaxed str_to_int
1434 interaction_count
= str_to_int(is_e
.get('userInteractionCount'))
1435 if interaction_count
is None:
1437 count_kind
= INTERACTION_TYPE_MAP
.get(interaction_type
.split('/')[-1])
1440 count_key
= '%s_count' % count_kind
1441 if info
.get(count_key
) is not None:
1443 info
[count_key
] = interaction_count
1445 def extract_chapter_information(e
):
1447 'title': part
.get('name'),
1448 'start_time': part
.get('startOffset'),
1449 'end_time': part
.get('endOffset'),
1450 } for part
in e
.get('hasPart', []) if part
.get('@type') == 'Clip']
1451 for idx
, (last_c
, current_c
, next_c
) in enumerate(zip(
1452 [{'end_time': 0}
] + chapters
, chapters
, chapters
[1:])):
1453 current_c
['end_time'] = current_c
['end_time'] or next_c
['start_time']
1454 current_c
['start_time'] = current_c
['start_time'] or last_c
['end_time']
1455 if None in current_c
.values():
1456 self
.report_warning(f
'Chapter {idx} contains broken data. Not extracting chapters')
1459 chapters
[-1]['end_time'] = chapters
[-1]['end_time'] or info
['duration']
1460 info
['chapters'] = chapters
1462 def extract_video_object(e
):
1463 assert e
['@type'] == 'VideoObject'
1464 author
= e
.get('author')
1466 'url': url_or_none(e
.get('contentUrl')),
1467 'title': unescapeHTML(e
.get('name')),
1468 'description': unescapeHTML(e
.get('description')),
1469 'thumbnails': [{'url': url_or_none(url)}
1470 for url
in variadic(traverse_obj(e
, 'thumbnailUrl', 'thumbnailURL'))],
1471 'duration': parse_duration(e
.get('duration')),
1472 'timestamp': unified_timestamp(e
.get('uploadDate')),
1473 # author can be an instance of 'Organization' or 'Person' types.
1474 # both types can have 'name' property(inherited from 'Thing' type). [1]
1475 # however some websites are using 'Text' type instead.
1476 # 1. https://schema.org/VideoObject
1477 'uploader': author
.get('name') if isinstance(author
, dict) else author
if isinstance(author
, compat_str
) else None,
1478 'filesize': float_or_none(e
.get('contentSize')),
1479 'tbr': int_or_none(e
.get('bitrate')),
1480 'width': int_or_none(e
.get('width')),
1481 'height': int_or_none(e
.get('height')),
1482 'view_count': int_or_none(e
.get('interactionCount')),
1484 extract_interaction_statistic(e
)
1485 extract_chapter_information(e
)
1487 def traverse_json_ld(json_ld
, at_top_level
=True):
1489 if at_top_level
and '@context' not in e
:
1491 if at_top_level
and set(e
.keys()) == {'@context', '@graph'}
:
1492 traverse_json_ld(variadic(e
['@graph'], allowed_types
=(dict,)), at_top_level
=False)
1494 item_type
= e
.get('@type')
1495 if expected_type
is not None and expected_type
!= item_type
:
1497 rating
= traverse_obj(e
, ('aggregateRating', 'ratingValue'), expected_type
=float_or_none
)
1498 if rating
is not None:
1499 info
['average_rating'] = rating
1500 if item_type
in ('TVEpisode', 'Episode'):
1501 episode_name
= unescapeHTML(e
.get('name'))
1503 'episode': episode_name
,
1504 'episode_number': int_or_none(e
.get('episodeNumber')),
1505 'description': unescapeHTML(e
.get('description')),
1507 if not info
.get('title') and episode_name
:
1508 info
['title'] = episode_name
1509 part_of_season
= e
.get('partOfSeason')
1510 if isinstance(part_of_season
, dict) and part_of_season
.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1512 'season': unescapeHTML(part_of_season
.get('name')),
1513 'season_number': int_or_none(part_of_season
.get('seasonNumber')),
1515 part_of_series
= e
.get('partOfSeries') or e
.get('partOfTVSeries')
1516 if isinstance(part_of_series
, dict) and part_of_series
.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1517 info
['series'] = unescapeHTML(part_of_series
.get('name'))
1518 elif item_type
== 'Movie':
1520 'title': unescapeHTML(e
.get('name')),
1521 'description': unescapeHTML(e
.get('description')),
1522 'duration': parse_duration(e
.get('duration')),
1523 'timestamp': unified_timestamp(e
.get('dateCreated')),
1525 elif item_type
in ('Article', 'NewsArticle'):
1527 'timestamp': parse_iso8601(e
.get('datePublished')),
1528 'title': unescapeHTML(e
.get('headline')),
1529 'description': unescapeHTML(e
.get('articleBody') or e
.get('description')),
1531 elif item_type
== 'VideoObject':
1532 extract_video_object(e
)
1533 if expected_type
is None:
1537 video
= e
.get('video')
1538 if isinstance(video
, dict) and video
.get('@type') == 'VideoObject':
1539 extract_video_object(video
)
1540 if expected_type
is None:
1544 traverse_json_ld(json_ld
)
1546 return dict((k
, v
) for k
, v
in info
.items() if v
is not None)
1548 def _search_nextjs_data(self
, webpage
, video_id
, *, transform_source
=None, fatal
=True, **kw
):
1549 return self
._parse
_json
(
1551 r
'(?s)<script[^>]+id=[\'"]__NEXT_DATA__[\'"][^
>]*>([^
<]+)</script
>',
1552 webpage, 'next
.js data
', fatal=fatal, **kw),
1553 video_id, transform_source=transform_source, fatal=fatal)
1555 def _search_nuxt_data(self, webpage, video_id, context_name='__NUXT__
'):
1556 ''' Parses Nuxt.js metadata. This works as long as the function __NUXT__ invokes is a pure function. '''
1557 # not all website do this, but it can be changed
1558 # https://stackoverflow.com/questions/67463109/how-to-change-or-hide-nuxt-and-nuxt-keyword-in-page-source
1559 rectx = re.escape(context_name)
1560 js, arg_keys, arg_vals = self._search_regex(
1561 (r'<script
>window\
.%s=\
(function\
((?P
<arg_keys
>.*?
)\
)\{return\s(?P<js>\{.*?\}
)\
}\
((?P
<arg_vals
>.+?
)\
)\
);?
</script
>' % rectx,
1562 r'%s\
(.*?\
(function\
((?P
<arg_keys
>.*?
)\
)\{return\s(?P<js>\{.*?\}
)\
}\
((?P
<arg_vals
>.*?
)\
)' % rectx),
1563 webpage, context_name, group=['js
', 'arg_keys
', 'arg_vals
'])
1565 args = dict(zip(arg_keys.split(','), arg_vals.split(',')))
1567 for key, val in args.items():
1568 if val in ('undefined
', 'void
0'):
1571 return self._parse_json(js_to_json(js, args), video_id)['data
'][0]
1574 def _hidden_inputs(html):
1575 html = re.sub(r'<!--(?
:(?
!<!--).)*-->', '', html)
1577 for input in re.findall(r'(?i
)(<input[^
>]+>)', html):
1578 attrs = extract_attributes(input)
1581 if attrs.get('type') not in ('hidden
', 'submit
'):
1583 name = attrs.get('name
') or attrs.get('id')
1584 value = attrs.get('value
')
1585 if name and value is not None:
1586 hidden_inputs[name] = value
1587 return hidden_inputs
1589 def _form_hidden_inputs(self, form_id, html):
1590 form = self._search_regex(
1591 r'(?
is)<form
[^
>]+?
id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1592 html, '%s form' % form_id, group='form')
1593 return self._hidden_inputs(form)
1596 regex = r' *((?P<reverse>\+)?(?P<field>[a-zA-Z0-9_]+)((?P<separator>[~:])(?P<limit>.*?))?)? *$'
1598 default = ('hidden', 'aud_or_vid', 'hasvid', 'ie_pref', 'lang', 'quality',
1599 'res', 'fps', 'hdr:12', 'codec:vp9.2', 'size', 'br', 'asr',
1600 'proto', 'ext', 'hasaud', 'source', 'id') # These must not be aliases
1601 ytdl_default = ('hasaud', 'lang', 'quality', 'tbr', 'filesize', 'vbr',
1602 'height', 'width', 'proto', 'vext', 'abr', 'aext',
1603 'fps', 'fs_approx', 'source', 'id')
1606 'vcodec': {'type': 'ordered', 'regex': True,
1607 'order': ['av0?1', 'vp0?9.2', 'vp0?9', '[hx]265|he?vc?', '[hx]264|avc', 'vp0?8', 'mp4v|h263', 'theora', '', None, 'none']},
1608 'acodec': {'type': 'ordered', 'regex': True,
1609 'order': ['[af]lac', 'wav|aiff', 'opus', 'vorbis', 'aac', 'mp?4a?', 'mp3', 'e-?a?c-?3', 'ac-?3', 'dts', '', None, 'none']},
1610 'hdr': {'type': 'ordered', 'regex': True, 'field': 'dynamic_range',
1611 'order': ['dv', '(hdr)?12', r'(hdr)?10\+', '(hdr)?10', 'hlg', '', 'sdr', None]},
1612 'proto': {'type': 'ordered', 'regex': True, 'field': 'protocol',
1613 'order': ['(ht|f)tps', '(ht|f)tp$', 'm3u8.*', '.*dash', 'websocket_frag', 'rtmpe?', '', 'mms|rtsp', 'ws|websocket', 'f4']},
1614 'vext': {'type': 'ordered', 'field': 'video_ext',
1615 'order': ('mp4', 'webm', 'flv', '', 'none'),
1616 'order_free': ('webm', 'mp4', 'flv', '', 'none')},
1617 'aext': {'type': 'ordered', 'field': 'audio_ext',
1618 'order': ('m4a', 'aac', 'mp3', 'ogg', 'opus', 'webm', '', 'none'),
1619 'order_free': ('opus', 'ogg', 'webm', 'm4a', 'mp3', 'aac', '', 'none')},
1620 'hidden': {'visible': False, 'forced': True, 'type': 'extractor', 'max': -1000},
1621 'aud_or_vid': {'visible': False, 'forced': True, 'type': 'multiple',
1622 'field': ('vcodec', 'acodec'),
1623 'function': lambda it: int(any(v != 'none' for v in it))},
1624 'ie_pref': {'priority': True, 'type': 'extractor'},
1625 'hasvid': {'priority': True, 'field': 'vcodec', 'type': 'boolean', 'not_in_list': ('none',)},
1626 'hasaud': {'field': 'acodec', 'type': 'boolean', 'not_in_list': ('none',)},
1627 'lang': {'convert': 'float', 'field': 'language_preference', 'default': -1},
1628 'quality': {'convert': 'float', 'default': -1},
1629 'filesize': {'convert': 'bytes'},
1630 'fs_approx': {'convert': 'bytes', 'field': 'filesize_approx'},
1631 'id': {'convert': 'string', 'field': 'format_id'},
1632 'height': {'convert': 'float_none'},
1633 'width': {'convert': 'float_none'},
1634 'fps': {'convert': 'float_none'},
1635 'tbr': {'convert': 'float_none'},
1636 'vbr': {'convert': 'float_none'},
1637 'abr': {'convert': 'float_none'},
1638 'asr': {'convert': 'float_none'},
1639 'source': {'convert': 'float', 'field': 'source_preference', 'default': -1},
1641 'codec': {'type': 'combined', 'field': ('vcodec', 'acodec')},
1642 'br': {'type': 'combined', 'field': ('tbr', 'vbr', 'abr'), 'same_limit': True},
1643 'size': {'type': 'combined', 'same_limit': True, 'field': ('filesize', 'fs_approx')},
1644 'ext': {'type': 'combined', 'field': ('vext', 'aext')},
1645 'res': {'type': 'multiple', 'field': ('height', 'width'),
1646 'function': lambda it: (lambda l: min(l) if l else 0)(tuple(filter(None, it)))},
1648 # For compatibility with youtube-dl
1649 'format_id': {'type': 'alias', 'field': 'id'},
1650 'preference': {'type': 'alias', 'field': 'ie_pref'},
1651 'language_preference': {'type': 'alias', 'field': 'lang'},
1654 'dimension': {'type': 'alias', 'field': 'res'},
1655 'resolution': {'type': 'alias', 'field': 'res'},
1656 'extension': {'type': 'alias', 'field': 'ext'},
1657 'bitrate': {'type': 'alias', 'field': 'br'},
1658 'total_bitrate': {'type': 'alias', 'field': 'tbr'},
1659 'video_bitrate': {'type': 'alias', 'field': 'vbr'},
1660 'audio_bitrate': {'type': 'alias', 'field': 'abr'},
1661 'framerate': {'type': 'alias', 'field': 'fps'},
1662 'protocol': {'type': 'alias', 'field': 'proto'},
1663 'source_preference': {'type': 'alias', 'field': 'source'},
1664 'filesize_approx': {'type': 'alias', 'field': 'fs_approx'},
1665 'filesize_estimate': {'type': 'alias', 'field': 'size'},
1666 'samplerate': {'type': 'alias', 'field': 'asr'},
1667 'video_ext': {'type': 'alias', 'field': 'vext'},
1668 'audio_ext': {'type': 'alias', 'field': 'aext'},
1669 'video_codec': {'type': 'alias', 'field': 'vcodec'},
1670 'audio_codec': {'type': 'alias', 'field': 'acodec'},
1671 'video': {'type': 'alias', 'field': 'hasvid'},
1672 'has_video': {'type': 'alias', 'field': 'hasvid'},
1673 'audio': {'type': 'alias', 'field': 'hasaud'},
1674 'has_audio': {'type': 'alias', 'field': 'hasaud'},
1675 'extractor': {'type': 'alias', 'field': 'ie_pref'},
1676 'extractor_preference': {'type': 'alias', 'field': 'ie_pref'},
1679 def __init__(self, ie, field_preference):
1681 self.ydl = ie._downloader
1682 self.evaluate_params(self.ydl.params, field_preference)
1683 if ie.get_param('verbose'):
1684 self.print_verbose_info(self.ydl.write_debug)
1686 def _get_field_setting(self, field, key):
1687 if field not in self.settings:
1688 if key in ('forced', 'priority'):
1690 self.ydl.deprecation_warning(
1691 f'Using arbitrary fields ({field}) for format sorting is deprecated '
1692 'and may be removed in a future version')
1693 self.settings[field] = {}
1694 propObj = self.settings[field]
1695 if key not in propObj:
1696 type = propObj.get('type')
1698 default = 'preference' if type == 'extractor' else (field,) if type in ('combined', 'multiple') else field
1699 elif key == 'convert':
1700 default = 'order' if type == 'ordered' else 'float_string' if field else 'ignore'
1702 default = {'type': 'field', 'visible': True, 'order': [], 'not_in_list': (None,)}.get(key, None)
1703 propObj[key] = default
1706 def _resolve_field_value(self, field, value, convertNone=False):
1711 value = value.lower()
1712 conversion = self._get_field_setting(field, 'convert')
1713 if conversion == 'ignore':
1715 if conversion == 'string':
1717 elif conversion == 'float_none':
1718 return float_or_none(value)
1719 elif conversion == 'bytes':
1720 return FileDownloader.parse_bytes(value)
1721 elif conversion == 'order':
1722 order_list = (self._use_free_order and self._get_field_setting(field, 'order_free')) or self._get_field_setting(field, 'order')
1723 use_regex = self._get_field_setting(field, 'regex')
1724 list_length = len(order_list)
1725 empty_pos = order_list.index('') if '' in order_list else list_length + 1
1726 if use_regex and value is not None:
1727 for i, regex in enumerate(order_list):
1728 if regex and re.match(regex, value):
1729 return list_length - i
1730 return list_length - empty_pos # not in list
1731 else: # not regex or value = None
1732 return list_length - (order_list.index(value) if value in order_list else empty_pos)
1734 if value.isnumeric():
1737 self.settings[field]['convert'] = 'string'
1740 def evaluate_params(self, params, sort_extractor):
1741 self._use_free_order = params.get('prefer_free_formats', False)
1742 self._sort_user = params.get('format_sort', [])
1743 self._sort_extractor = sort_extractor
1745 def add_item(field, reverse, closest, limit_text):
1746 field = field.lower()
1747 if field in self._order:
1749 self._order.append(field)
1750 limit = self._resolve_field_value(field, limit_text)
1753 'closest': False if limit is None else closest,
1754 'limit_text': limit_text,
1756 if field in self.settings:
1757 self.settings[field].update(data)
1759 self.settings[field] = data
1762 tuple(field for field in self.default if self._get_field_setting(field, 'forced'))
1763 + (tuple() if params.get('format_sort_force', False)
1764 else tuple(field for field in self.default if self._get_field_setting(field, 'priority')))
1765 + tuple(self._sort_user) + tuple(sort_extractor) + self.default)
1767 for item in sort_list:
1768 match = re.match(self.regex, item)
1770 raise ExtractorError('Invalid format sort string "%s" given by extractor' % item)
1771 field = match.group('field')
1774 if self._get_field_setting(field, 'type') == 'alias':
1775 alias, field = field, self._get_field_setting(field, 'field')
1776 if alias not in ('format_id', 'preference', 'language_preference'):
1777 self.ydl.deprecation_warning(
1778 f'Format sorting alias {alias} is deprecated '
1779 f'and may be removed in a future version. Please use {field} instead')
1780 reverse = match.group('reverse') is not None
1781 closest = match.group('separator') == '~'
1782 limit_text = match.group('limit')
1784 has_limit = limit_text is not None
1785 has_multiple_fields = self._get_field_setting(field, 'type') == 'combined'
1786 has_multiple_limits = has_limit and has_multiple_fields and not self._get_field_setting(field, 'same_limit')
1788 fields = self._get_field_setting(field, 'field') if has_multiple_fields else (field,)
1789 limits = limit_text.split(':') if has_multiple_limits else (limit_text,) if has_limit else tuple()
1790 limit_count = len(limits)
1791 for (i, f) in enumerate(fields):
1792 add_item(f, reverse, closest,
1793 limits[i] if i < limit_count
1794 else limits[0] if has_limit and not has_multiple_limits
1797 def print_verbose_info(self, write_debug):
1799 write_debug('Sort order given by user: %s' % ', '.join(self._sort_user))
1800 if self._sort_extractor:
1801 write_debug('Sort order given by extractor: %s' % ', '.join(self._sort_extractor))
1802 write_debug('Formats sorted by: %s' % ', '.join(['%s%s%s' % (
1803 '+' if self._get_field_setting(field, 'reverse') else '', field,
1804 '%s%s(%s)' % ('~' if self._get_field_setting(field, 'closest') else ':',
1805 self._get_field_setting(field, 'limit_text'),
1806 self._get_field_setting(field, 'limit'))
1807 if self._get_field_setting(field, 'limit_text') is not None else '')
1808 for field in self._order if self._get_field_setting(field, 'visible')]))
1810 def _calculate_field_preference_from_value(self, format, field, type, value):
1811 reverse = self._get_field_setting(field, 'reverse')
1812 closest = self._get_field_setting(field, 'closest')
1813 limit = self._get_field_setting(field, 'limit')
1815 if type == 'extractor':
1816 maximum = self._get_field_setting(field, 'max')
1817 if value is None or (maximum is not None and value >= maximum):
1819 elif type == 'boolean':
1820 in_list = self._get_field_setting(field, 'in_list')
1821 not_in_list = self._get_field_setting(field, 'not_in_list')
1822 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
1823 elif type == 'ordered':
1824 value = self._resolve_field_value(field, value, True)
1826 # try to convert to number
1827 val_num = float_or_none(value, default=self._get_field_setting(field, 'default'))
1828 is_num = self._get_field_setting(field, 'convert') != 'string' and val_num is not None
1832 return ((-10, 0) if value is None
1833 else (1, value, 0) if not is_num # if a field has mixed strings and numbers, strings are sorted higher
1834 else (0, -abs(value - limit), value - limit if reverse else limit - value) if closest
1835 else (0, value, 0) if not reverse and (limit is None or value <= limit)
1836 else (0, -value, 0) if limit is None or (reverse and value == limit) or value > limit
1837 else (-1, value, 0))
1839 def _calculate_field_preference(self, format, field):
1840 type = self._get_field_setting(field, 'type') # extractor, boolean, ordered, field, multiple
1841 get_value = lambda f: format.get(self._get_field_setting(f, 'field'))
1842 if type == 'multiple':
1843 type = 'field' # Only 'field' is allowed in multiple for now
1844 actual_fields = self._get_field_setting(field, 'field')
1846 value = self._get_field_setting(field, 'function')(get_value(f) for f in actual_fields)
1848 value = get_value(field)
1849 return self._calculate_field_preference_from_value(format, field, type, value)
1851 def calculate_preference(self, format):
1852 # Determine missing protocol
1853 if not format.get('protocol'):
1854 format['protocol'] = determine_protocol(format)
1856 # Determine missing ext
1857 if not format.get('ext') and 'url' in format:
1858 format['ext'] = determine_ext(format['url'])
1859 if format.get('vcodec') == 'none':
1860 format['audio_ext'] = format['ext'] if format.get('acodec') != 'none' else 'none'
1861 format['video_ext'] = 'none'
1863 format['video_ext'] = format['ext']
1864 format['audio_ext'] = 'none'
1865 # if format.get('preference') is None and format.get('ext') in ('f4f', 'f4m'): # Not supported?
1866 # format['preference'] = -1000
1868 # Determine missing bitrates
1869 if format.get('tbr') is None:
1870 if format.get('vbr') is not None and format.get('abr') is not None:
1871 format['tbr'] = format.get('vbr', 0) + format.get('abr', 0)
1873 if format.get('vcodec') != 'none' and format.get('vbr') is None:
1874 format['vbr'] = format.get('tbr') - format.get('abr', 0)
1875 if format.get('acodec') != 'none' and format.get('abr') is None:
1876 format['abr'] = format.get('tbr') - format.get('vbr', 0)
1878 return tuple(self._calculate_field_preference(format, field) for field in self._order)
1880 def _sort_formats(self, formats, field_preference=[]):
1883 format_sort = self.FormatSort(self, field_preference)
1884 formats.sort(key=lambda f: format_sort.calculate_preference(f))
1886 def _check_formats(self, formats, video_id):
1888 formats[:] = filter(
1889 lambda f: self._is_valid_url(
1891 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1895 def _remove_duplicate_formats(formats):
1899 if f['url'] not in format_urls:
1900 format_urls.add(f['url'])
1901 unique_formats.append(f)
1902 formats[:] = unique_formats
1904 def _is_valid_url(self, url, video_id, item='video', headers={}):
1905 url = self._proto_relative_url(url, scheme='http:')
1906 # For now assume non HTTP(S) URLs always valid
1907 if not (url.startswith('http://') or url.startswith('https://')):
1910 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1912 except ExtractorError as e:
1914 '%s: %s URL is invalid, skipping: %s'
1915 % (video_id, item, error_to_compat_str(e.cause)))
1918 def http_scheme(self):
1919 """ Either "http
:" or "https
:", depending on the user's preferences """
1922 if self.get_param('prefer_insecure', False)
1925 def _proto_relative_url(self, url, scheme=None):
1928 if url.startswith('//'):
1930 scheme = self.http_scheme()
1935 def _sleep(self, timeout, video_id, msg_template=None):
1936 if msg_template is None:
1937 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1938 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1942 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1943 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1944 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
1945 manifest = self._download_xml(
1946 manifest_url, video_id, 'Downloading f4m manifest',
1947 'Unable to download f4m manifest',
1948 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1949 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
1950 transform_source=transform_source,
1951 fatal=fatal, data=data, headers=headers, query=query)
1953 if manifest is False:
1956 return self._parse_f4m_formats(
1957 manifest, manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1958 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1960 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1961 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1962 fatal=True, m3u8_id=None):
1963 if not isinstance(manifest, compat_etree_Element) and not fatal:
1966 # currently yt-dlp cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1967 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1968 if akamai_pv is not None and ';' in akamai_pv.text:
1969 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1970 if playerVerificationChallenge.strip() != '':
1974 manifest_version = '1.0'
1975 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1977 manifest_version = '2.0'
1978 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1979 # Remove unsupported DRM protected media from final formats
1980 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
1981 media_nodes = remove_encrypted_media(media_nodes)
1985 manifest_base_url = get_base_url(manifest)
1987 bootstrap_info = xpath_element(
1988 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1989 'bootstrap info', default=None)
1992 mime_type = xpath_text(
1993 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1994 'base URL', default=None)
1995 if mime_type and mime_type.startswith('audio/'):
1998 for i, media_el in enumerate(media_nodes):
1999 tbr = int_or_none(media_el.attrib.get('bitrate'))
2000 width = int_or_none(media_el.attrib.get('width'))
2001 height = int_or_none(media_el.attrib.get('height'))
2002 format_id = join_nonempty(f4m_id, tbr or i)
2003 # If <bootstrapInfo> is present, the specified f4m is a
2004 # stream-level manifest, and only set-level manifests may refer to
2005 # external resources. See section 11.4 and section 4 of F4M spec
2006 if bootstrap_info is None:
2008 # @href is introduced in 2.0, see section 11.6 of F4M spec
2009 if manifest_version == '2.0':
2010 media_url = media_el.attrib.get('href')
2011 if media_url is None:
2012 media_url = media_el.attrib.get('url')
2016 media_url if media_url.startswith('http://') or media_url.startswith('https://')
2017 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
2018 # If media_url is itself a f4m manifest do the recursive extraction
2019 # since bitrates in parent manifest (this one) and media_url manifest
2020 # may differ leading to inability to resolve the format by requested
2021 # bitrate in f4m downloader
2022 ext = determine_ext(manifest_url)
2024 f4m_formats = self._extract_f4m_formats(
2025 manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
2026 transform_source=transform_source, fatal=fatal)
2027 # Sometimes stream-level manifest contains single media entry that
2028 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
2029 # At the same time parent's media entry in set-level manifest may
2030 # contain it. We will copy it from parent in such cases.
2031 if len(f4m_formats) == 1:
2034 'tbr': f.get('tbr') or tbr,
2035 'width': f.get('width') or width,
2036 'height': f.get('height') or height,
2037 'format_id': f.get('format_id') if not tbr else format_id,
2040 formats.extend(f4m_formats)
2043 formats.extend(self._extract_m3u8_formats(
2044 manifest_url, video_id, 'mp4', preference=preference,
2045 quality=quality, m3u8_id=m3u8_id, fatal=fatal))
2048 'format_id': format_id,
2049 'url': manifest_url,
2050 'manifest_url': manifest_url,
2051 'ext': 'flv' if bootstrap_info is not None else None,
2057 'preference': preference,
2062 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, quality=None, m3u8_id=None):
2064 'format_id': join_nonempty(m3u8_id, 'meta'),
2068 'preference': preference - 100 if preference else -100,
2070 'resolution': 'multiple',
2071 'format_note': 'Quality selection URL',
2074 def _report_ignoring_subs(self, name):
2075 self.report_warning(bug_reports_message(
2076 f'Ignoring subtitle tracks found in the {name} manifest; '
2077 'if any subtitle tracks are missing,'
2080 def _extract_m3u8_formats(self, *args, **kwargs):
2081 fmts, subs = self._extract_m3u8_formats_and_subtitles(*args, **kwargs)
2083 self._report_ignoring_subs('HLS')
2086 def _extract_m3u8_formats_and_subtitles(
2087 self, m3u8_url, video_id, ext=None, entry_protocol='m3u8_native',
2088 preference=None, quality=None, m3u8_id=None, note=None,
2089 errnote=None, fatal=True, live=False, data=None, headers={},
2092 res = self._download_webpage_handle(
2094 note='Downloading m3u8 information' if note is None else note,
2095 errnote='Failed to download m3u8 information' if errnote is None else errnote,
2096 fatal=fatal, data=data, headers=headers, query=query)
2101 m3u8_doc, urlh = res
2102 m3u8_url = urlh.geturl()
2104 return self._parse_m3u8_formats_and_subtitles(
2105 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
2106 preference=preference, quality=quality, m3u8_id=m3u8_id,
2107 note=note, errnote=errnote, fatal=fatal, live=live, data=data,
2108 headers=headers, query=query, video_id=video_id)
2110 def _parse_m3u8_formats_and_subtitles(
2111 self, m3u8_doc, m3u8_url=None, ext=None, entry_protocol='m3u8_native',
2112 preference=None, quality=None, m3u8_id=None, live=False, note=None,
2113 errnote=None, fatal=True, data=None, headers={}, query={},
2115 formats, subtitles = [], {}
2117 has_drm = re.search('|'.join([
2118 r'#EXT-X-FAXS-CM:', # Adobe Flash Access
2119 r'#EXT-X-(?:SESSION-)?KEY:.*?URI="skd
://', # Apple FairPlay
2122 def format_url(url):
2123 return url if re.match(r'^https?
://', url) else compat_urlparse.urljoin(m3u8_url, url)
2125 if self.get_param('hls_split_discontinuity
', False):
2126 def _extract_m3u8_playlist_indices(manifest_url=None, m3u8_doc=None):
2128 if not manifest_url:
2130 m3u8_doc = self._download_webpage(
2131 manifest_url, video_id, fatal=fatal, data=data, headers=headers,
2132 note=False, errnote='Failed to download m3u8 playlist information
')
2133 if m3u8_doc is False:
2135 return range(1 + sum(line.startswith('#EXT-X-DISCONTINUITY') for line in m3u8_doc.splitlines()))
2138 def _extract_m3u8_playlist_indices(*args
, **kwargs
):
2142 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
2143 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
2144 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
2146 # We should try extracting formats only from master playlists [1, 4.3.4],
2147 # i.e. playlists that describe available qualities. On the other hand
2148 # media playlists [1, 4.3.3] should be returned as is since they contain
2149 # just the media without qualities renditions.
2150 # Fortunately, master playlist can be easily distinguished from media
2151 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
2152 # master playlist tags MUST NOT appear in a media playlist and vice versa.
2153 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
2154 # media playlist and MUST NOT appear in master playlist thus we can
2155 # clearly detect media playlist with this criterion.
2157 if '#EXT-X-TARGETDURATION' in m3u8_doc
: # media playlist, return as is
2159 'format_id': join_nonempty(m3u8_id
, idx
),
2160 'format_index': idx
,
2161 'url': m3u8_url
or encode_data_uri(m3u8_doc
.encode('utf-8'), 'application/x-mpegurl'),
2163 'protocol': entry_protocol
,
2164 'preference': preference
,
2167 } for idx
in _extract_m3u8_playlist_indices(m3u8_doc
=m3u8_doc
)]
2169 return formats
, subtitles
2172 last_stream_inf
= {}
2174 def extract_media(x_media_line
):
2175 media
= parse_m3u8_attributes(x_media_line
)
2176 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
2177 media_type
, group_id
, name
= media
.get('TYPE'), media
.get('GROUP-ID'), media
.get('NAME')
2178 if not (media_type
and group_id
and name
):
2180 groups
.setdefault(group_id
, []).append(media
)
2181 # <https://tools.ietf.org/html/rfc8216#section-4.3.4.1>
2182 if media_type
== 'SUBTITLES':
2183 # According to RFC 8216 §4.3.4.2.1, URI is REQUIRED in the
2184 # EXT-X-MEDIA tag if the media type is SUBTITLES.
2185 # However, lack of URI has been spotted in the wild.
2186 # e.g. NebulaIE; see https://github.com/yt-dlp/yt-dlp/issues/339
2187 if not media
.get('URI'):
2189 url
= format_url(media
['URI'])
2192 'ext': determine_ext(url
),
2194 if sub_info
['ext'] == 'm3u8':
2195 # Per RFC 8216 §3.1, the only possible subtitle format m3u8
2196 # files may contain is WebVTT:
2197 # <https://tools.ietf.org/html/rfc8216#section-3.1>
2198 sub_info
['ext'] = 'vtt'
2199 sub_info
['protocol'] = 'm3u8_native'
2200 lang
= media
.get('LANGUAGE') or 'und'
2201 subtitles
.setdefault(lang
, []).append(sub_info
)
2202 if media_type
not in ('VIDEO', 'AUDIO'):
2204 media_url
= media
.get('URI')
2206 manifest_url
= format_url(media_url
)
2208 'format_id': join_nonempty(m3u8_id
, group_id
, name
, idx
),
2209 'format_note': name
,
2210 'format_index': idx
,
2211 'url': manifest_url
,
2212 'manifest_url': m3u8_url
,
2213 'language': media
.get('LANGUAGE'),
2215 'protocol': entry_protocol
,
2216 'preference': preference
,
2218 'vcodec': 'none' if media_type
== 'AUDIO' else None,
2219 } for idx
in _extract_m3u8_playlist_indices(manifest_url
))
2221 def build_stream_name():
2222 # Despite specification does not mention NAME attribute for
2223 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
2224 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
2225 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
2226 stream_name
= last_stream_inf
.get('NAME')
2229 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
2230 # from corresponding rendition group
2231 stream_group_id
= last_stream_inf
.get('VIDEO')
2232 if not stream_group_id
:
2234 stream_group
= groups
.get(stream_group_id
)
2235 if not stream_group
:
2236 return stream_group_id
2237 rendition
= stream_group
[0]
2238 return rendition
.get('NAME') or stream_group_id
2240 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
2241 # chance to detect video only formats when EXT-X-STREAM-INF tags
2242 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
2243 for line
in m3u8_doc
.splitlines():
2244 if line
.startswith('#EXT-X-MEDIA:'):
2247 for line
in m3u8_doc
.splitlines():
2248 if line
.startswith('#EXT-X-STREAM-INF:'):
2249 last_stream_inf
= parse_m3u8_attributes(line
)
2250 elif line
.startswith('#') or not line
.strip():
2253 tbr
= float_or_none(
2254 last_stream_inf
.get('AVERAGE-BANDWIDTH')
2255 or last_stream_inf
.get('BANDWIDTH'), scale
=1000)
2256 manifest_url
= format_url(line
.strip())
2258 for idx
in _extract_m3u8_playlist_indices(manifest_url
):
2259 format_id
= [m3u8_id
, None, idx
]
2260 # Bandwidth of live streams may differ over time thus making
2261 # format_id unpredictable. So it's better to keep provided
2264 stream_name
= build_stream_name()
2265 format_id
[1] = stream_name
or '%d' % (tbr
or len(formats
))
2267 'format_id': join_nonempty(*format_id
),
2268 'format_index': idx
,
2269 'url': manifest_url
,
2270 'manifest_url': m3u8_url
,
2273 'fps': float_or_none(last_stream_inf
.get('FRAME-RATE')),
2274 'protocol': entry_protocol
,
2275 'preference': preference
,
2278 resolution
= last_stream_inf
.get('RESOLUTION')
2280 mobj
= re
.search(r
'(?P<width>\d+)[xX](?P<height>\d+)', resolution
)
2282 f
['width'] = int(mobj
.group('width'))
2283 f
['height'] = int(mobj
.group('height'))
2284 # Unified Streaming Platform
2286 r
'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f
['url'])
2288 abr
, vbr
= mobj
.groups()
2289 abr
, vbr
= float_or_none(abr
, 1000), float_or_none(vbr
, 1000)
2294 codecs
= parse_codecs(last_stream_inf
.get('CODECS'))
2296 audio_group_id
= last_stream_inf
.get('AUDIO')
2297 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
2298 # references a rendition group MUST have a CODECS attribute.
2299 # However, this is not always respected, for example, [2]
2300 # contains EXT-X-STREAM-INF tag which references AUDIO
2301 # rendition group but does not have CODECS and despite
2302 # referencing an audio group it represents a complete
2303 # (with audio and video) format. So, for such cases we will
2304 # ignore references to rendition groups and treat them
2305 # as complete formats.
2306 if audio_group_id
and codecs
and f
.get('vcodec') != 'none':
2307 audio_group
= groups
.get(audio_group_id
)
2308 if audio_group
and audio_group
[0].get('URI'):
2309 # TODO: update acodec for audio only formats with
2311 f
['acodec'] = 'none'
2312 if not f
.get('ext'):
2313 f
['ext'] = 'm4a' if f
.get('vcodec') == 'none' else 'mp4'
2317 progressive_uri
= last_stream_inf
.get('PROGRESSIVE-URI')
2320 del http_f
['manifest_url']
2322 'format_id': f
['format_id'].replace('hls-', 'http-'),
2324 'url': progressive_uri
,
2326 formats
.append(http_f
)
2328 last_stream_inf
= {}
2329 return formats
, subtitles
2331 def _extract_m3u8_vod_duration(
2332 self
, m3u8_vod_url
, video_id
, note
=None, errnote
=None, data
=None, headers
={}, query={}
):
2334 m3u8_vod
= self
._download
_webpage
(
2335 m3u8_vod_url
, video_id
,
2336 note
='Downloading m3u8 VOD manifest' if note
is None else note
,
2337 errnote
='Failed to download VOD manifest' if errnote
is None else errnote
,
2338 fatal
=False, data
=data
, headers
=headers
, query
=query
)
2340 return self
._parse
_m
3u8_vod
_duration
(m3u8_vod
or '', video_id
)
2342 def _parse_m3u8_vod_duration(self
, m3u8_vod
, video_id
):
2343 if '#EXT-X-PLAYLIST-TYPE:VOD' not in m3u8_vod
:
2347 float(line
[len('#EXTINF:'):].split(',')[0])
2348 for line
in m3u8_vod
.splitlines() if line
.startswith('#EXTINF:'))) or None
2351 def _xpath_ns(path
, namespace
=None):
2355 for c
in path
.split('/'):
2356 if not c
or c
== '.':
2359 out
.append('{%s}%s' % (namespace
, c
))
2360 return '/'.join(out
)
2362 def _extract_smil_formats_and_subtitles(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None, transform_source
=None):
2363 smil
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
, transform_source
=transform_source
)
2369 namespace
= self
._parse
_smil
_namespace
(smil
)
2371 fmts
= self
._parse
_smil
_formats
(
2372 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2373 subs
= self
._parse
_smil
_subtitles
(
2374 smil
, namespace
=namespace
)
2378 def _extract_smil_formats(self
, *args
, **kwargs
):
2379 fmts
, subs
= self
._extract
_smil
_formats
_and
_subtitles
(*args
, **kwargs
)
2381 self
._report
_ignoring
_subs
('SMIL')
2384 def _extract_smil_info(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None):
2385 smil
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
)
2388 return self
._parse
_smil
(smil
, smil_url
, video_id
, f4m_params
=f4m_params
)
2390 def _download_smil(self
, smil_url
, video_id
, fatal
=True, transform_source
=None):
2391 return self
._download
_xml
(
2392 smil_url
, video_id
, 'Downloading SMIL file',
2393 'Unable to download SMIL file', fatal
=fatal
, transform_source
=transform_source
)
2395 def _parse_smil(self
, smil
, smil_url
, video_id
, f4m_params
=None):
2396 namespace
= self
._parse
_smil
_namespace
(smil
)
2398 formats
= self
._parse
_smil
_formats
(
2399 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2400 subtitles
= self
._parse
_smil
_subtitles
(smil
, namespace
=namespace
)
2402 video_id
= os
.path
.splitext(url_basename(smil_url
))[0]
2406 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2407 name
= meta
.attrib
.get('name')
2408 content
= meta
.attrib
.get('content')
2409 if not name
or not content
:
2411 if not title
and name
== 'title':
2413 elif not description
and name
in ('description', 'abstract'):
2414 description
= content
2415 elif not upload_date
and name
== 'date':
2416 upload_date
= unified_strdate(content
)
2419 'id': image
.get('type'),
2420 'url': image
.get('src'),
2421 'width': int_or_none(image
.get('width')),
2422 'height': int_or_none(image
.get('height')),
2423 } for image
in smil
.findall(self
._xpath
_ns
('.//image', namespace
)) if image
.get('src')]
2427 'title': title
or video_id
,
2428 'description': description
,
2429 'upload_date': upload_date
,
2430 'thumbnails': thumbnails
,
2432 'subtitles': subtitles
,
2435 def _parse_smil_namespace(self
, smil
):
2436 return self
._search
_regex
(
2437 r
'(?i)^{([^}]+)?}smil$', smil
.tag
, 'namespace', default
=None)
2439 def _parse_smil_formats(self
, smil
, smil_url
, video_id
, namespace
=None, f4m_params
=None, transform_rtmp_url
=None):
2441 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2442 b
= meta
.get('base') or meta
.get('httpBase')
2454 media
= smil
.findall(self
._xpath
_ns
('.//video', namespace
)) + smil
.findall(self
._xpath
_ns
('.//audio', namespace
))
2455 for medium
in media
:
2456 src
= medium
.get('src')
2457 if not src
or src
in srcs
:
2461 bitrate
= float_or_none(medium
.get('system-bitrate') or medium
.get('systemBitrate'), 1000)
2462 filesize
= int_or_none(medium
.get('size') or medium
.get('fileSize'))
2463 width
= int_or_none(medium
.get('width'))
2464 height
= int_or_none(medium
.get('height'))
2465 proto
= medium
.get('proto')
2466 ext
= medium
.get('ext')
2467 src_ext
= determine_ext(src
)
2468 streamer
= medium
.get('streamer') or base
2470 if proto
== 'rtmp' or streamer
.startswith('rtmp'):
2476 'format_id': 'rtmp-%d' % (rtmp_count
if bitrate
is None else bitrate
),
2478 'filesize': filesize
,
2482 if transform_rtmp_url
:
2483 streamer
, src
= transform_rtmp_url(streamer
, src
)
2484 formats
[-1].update({
2490 src_url
= src
if src
.startswith('http') else compat_urlparse
.urljoin(base
, src
)
2491 src_url
= src_url
.strip()
2493 if proto
== 'm3u8' or src_ext
== 'm3u8':
2494 m3u8_formats
= self
._extract
_m
3u8_formats
(
2495 src_url
, video_id
, ext
or 'mp4', m3u8_id
='hls', fatal
=False)
2496 if len(m3u8_formats
) == 1:
2498 m3u8_formats
[0].update({
2499 'format_id': 'hls-%d' % (m3u8_count
if bitrate
is None else bitrate
),
2504 formats
.extend(m3u8_formats
)
2505 elif src_ext
== 'f4m':
2510 'plugin': 'flowplayer-3.2.0.1',
2512 f4m_url
+= '&' if '?' in f4m_url
else '?'
2513 f4m_url
+= compat_urllib_parse_urlencode(f4m_params
)
2514 formats
.extend(self
._extract
_f
4m
_formats
(f4m_url
, video_id
, f4m_id
='hds', fatal
=False))
2515 elif src_ext
== 'mpd':
2516 formats
.extend(self
._extract
_mpd
_formats
(
2517 src_url
, video_id
, mpd_id
='dash', fatal
=False))
2518 elif re
.search(r
'\.ism/[Mm]anifest', src_url
):
2519 formats
.extend(self
._extract
_ism
_formats
(
2520 src_url
, video_id
, ism_id
='mss', fatal
=False))
2521 elif src_url
.startswith('http') and self
._is
_valid
_url
(src
, video_id
):
2525 'ext': ext
or src_ext
or 'flv',
2526 'format_id': 'http-%d' % (bitrate
or http_count
),
2528 'filesize': filesize
,
2533 for medium
in smil
.findall(self
._xpath
_ns
('.//imagestream', namespace
)):
2534 src
= medium
.get('src')
2535 if not src
or src
in srcs
:
2541 'format_id': 'imagestream-%d' % (imgs_count
),
2543 'ext': mimetype2ext(medium
.get('type')),
2546 'width': int_or_none(medium
.get('width')),
2547 'height': int_or_none(medium
.get('height')),
2548 'format_note': 'SMIL storyboards',
2553 def _parse_smil_subtitles(self
, smil
, namespace
=None, subtitles_lang
='en'):
2556 for num
, textstream
in enumerate(smil
.findall(self
._xpath
_ns
('.//textstream', namespace
))):
2557 src
= textstream
.get('src')
2558 if not src
or src
in urls
:
2561 ext
= textstream
.get('ext') or mimetype2ext(textstream
.get('type')) or determine_ext(src
)
2562 lang
= textstream
.get('systemLanguage') or textstream
.get('systemLanguageName') or textstream
.get('lang') or subtitles_lang
2563 subtitles
.setdefault(lang
, []).append({
2569 def _extract_xspf_playlist(self
, xspf_url
, playlist_id
, fatal
=True):
2570 xspf
= self
._download
_xml
(
2571 xspf_url
, playlist_id
, 'Downloading xpsf playlist',
2572 'Unable to download xspf manifest', fatal
=fatal
)
2575 return self
._parse
_xspf
(
2576 xspf
, playlist_id
, xspf_url
=xspf_url
,
2577 xspf_base_url
=base_url(xspf_url
))
2579 def _parse_xspf(self
, xspf_doc
, playlist_id
, xspf_url
=None, xspf_base_url
=None):
2581 'xspf': 'http://xspf.org/ns/0/',
2582 's1': 'http://static.streamone.nl/player/ns/0',
2586 for track
in xspf_doc
.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP
)):
2588 track
, xpath_with_ns('./xspf:title', NS_MAP
), 'title', default
=playlist_id
)
2589 description
= xpath_text(
2590 track
, xpath_with_ns('./xspf:annotation', NS_MAP
), 'description')
2591 thumbnail
= xpath_text(
2592 track
, xpath_with_ns('./xspf:image', NS_MAP
), 'thumbnail')
2593 duration
= float_or_none(
2594 xpath_text(track
, xpath_with_ns('./xspf:duration', NS_MAP
), 'duration'), 1000)
2597 for location
in track
.findall(xpath_with_ns('./xspf:location', NS_MAP
)):
2598 format_url
= urljoin(xspf_base_url
, location
.text
)
2603 'manifest_url': xspf_url
,
2604 'format_id': location
.get(xpath_with_ns('s1:label', NS_MAP
)),
2605 'width': int_or_none(location
.get(xpath_with_ns('s1:width', NS_MAP
))),
2606 'height': int_or_none(location
.get(xpath_with_ns('s1:height', NS_MAP
))),
2608 self
._sort
_formats
(formats
)
2613 'description': description
,
2614 'thumbnail': thumbnail
,
2615 'duration': duration
,
2620 def _extract_mpd_formats(self
, *args
, **kwargs
):
2621 fmts
, subs
= self
._extract
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2623 self
._report
_ignoring
_subs
('DASH')
2626 def _extract_mpd_formats_and_subtitles(
2627 self
, mpd_url
, video_id
, mpd_id
=None, note
=None, errnote
=None,
2628 fatal
=True, data
=None, headers
={}, query={}
):
2629 res
= self
._download
_xml
_handle
(
2631 note
='Downloading MPD manifest' if note
is None else note
,
2632 errnote
='Failed to download MPD manifest' if errnote
is None else errnote
,
2633 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
2639 mpd_base_url
= base_url(urlh
.geturl())
2641 return self
._parse
_mpd
_formats
_and
_subtitles
(
2642 mpd_doc
, mpd_id
, mpd_base_url
, mpd_url
)
2644 def _parse_mpd_formats(self
, *args
, **kwargs
):
2645 fmts
, subs
= self
._parse
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2647 self
._report
_ignoring
_subs
('DASH')
2650 def _parse_mpd_formats_and_subtitles(
2651 self
, mpd_doc
, mpd_id
=None, mpd_base_url
='', mpd_url
=None):
2653 Parse formats from MPD manifest.
2655 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
2656 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
2657 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
2659 if not self
.get_param('dynamic_mpd', True):
2660 if mpd_doc
.get('type') == 'dynamic':
2663 namespace
= self
._search
_regex
(r
'(?i)^{([^}]+)?}MPD$', mpd_doc
.tag
, 'namespace', default
=None)
2666 return self
._xpath
_ns
(path
, namespace
)
2668 def is_drm_protected(element
):
2669 return element
.find(_add_ns('ContentProtection')) is not None
2671 def extract_multisegment_info(element
, ms_parent_info
):
2672 ms_info
= ms_parent_info
.copy()
2674 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2675 # common attributes and elements. We will only extract relevant
2677 def extract_common(source
):
2678 segment_timeline
= source
.find(_add_ns('SegmentTimeline'))
2679 if segment_timeline
is not None:
2680 s_e
= segment_timeline
.findall(_add_ns('S'))
2682 ms_info
['total_number'] = 0
2685 r
= int(s
.get('r', 0))
2686 ms_info
['total_number'] += 1 + r
2687 ms_info
['s'].append({
2688 't': int(s
.get('t', 0)),
2689 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2690 'd': int(s
.attrib
['d']),
2693 start_number
= source
.get('startNumber')
2695 ms_info
['start_number'] = int(start_number
)
2696 timescale
= source
.get('timescale')
2698 ms_info
['timescale'] = int(timescale
)
2699 segment_duration
= source
.get('duration')
2700 if segment_duration
:
2701 ms_info
['segment_duration'] = float(segment_duration
)
2703 def extract_Initialization(source
):
2704 initialization
= source
.find(_add_ns('Initialization'))
2705 if initialization
is not None:
2706 ms_info
['initialization_url'] = initialization
.attrib
['sourceURL']
2708 segment_list
= element
.find(_add_ns('SegmentList'))
2709 if segment_list
is not None:
2710 extract_common(segment_list
)
2711 extract_Initialization(segment_list
)
2712 segment_urls_e
= segment_list
.findall(_add_ns('SegmentURL'))
2714 ms_info
['segment_urls'] = [segment
.attrib
['media'] for segment
in segment_urls_e
]
2716 segment_template
= element
.find(_add_ns('SegmentTemplate'))
2717 if segment_template
is not None:
2718 extract_common(segment_template
)
2719 media
= segment_template
.get('media')
2721 ms_info
['media'] = media
2722 initialization
= segment_template
.get('initialization')
2724 ms_info
['initialization'] = initialization
2726 extract_Initialization(segment_template
)
2729 mpd_duration
= parse_duration(mpd_doc
.get('mediaPresentationDuration'))
2730 formats
, subtitles
= [], {}
2731 stream_numbers
= collections
.defaultdict(int)
2732 for period
in mpd_doc
.findall(_add_ns('Period')):
2733 period_duration
= parse_duration(period
.get('duration')) or mpd_duration
2734 period_ms_info
= extract_multisegment_info(period
, {
2738 for adaptation_set
in period
.findall(_add_ns('AdaptationSet')):
2739 adaption_set_ms_info
= extract_multisegment_info(adaptation_set
, period_ms_info
)
2740 for representation
in adaptation_set
.findall(_add_ns('Representation')):
2741 representation_attrib
= adaptation_set
.attrib
.copy()
2742 representation_attrib
.update(representation
.attrib
)
2743 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2744 mime_type
= representation_attrib
['mimeType']
2745 content_type
= representation_attrib
.get('contentType', mime_type
.split('/')[0])
2747 codecs
= parse_codecs(representation_attrib
.get('codecs', ''))
2748 if content_type
not in ('video', 'audio', 'text'):
2749 if mime_type
== 'image/jpeg':
2750 content_type
= mime_type
2751 elif codecs
['vcodec'] != 'none':
2752 content_type
= 'video'
2753 elif codecs
['acodec'] != 'none':
2754 content_type
= 'audio'
2755 elif codecs
.get('tcodec', 'none') != 'none':
2756 content_type
= 'text'
2757 elif mimetype2ext(mime_type
) in ('tt', 'dfxp', 'ttml', 'xml', 'json'):
2758 content_type
= 'text'
2760 self
.report_warning('Unknown MIME type %s in DASH manifest' % mime_type
)
2764 for element
in (representation
, adaptation_set
, period
, mpd_doc
):
2765 base_url_e
= element
.find(_add_ns('BaseURL'))
2766 if base_url_e
is not None:
2767 base_url
= base_url_e
.text
+ base_url
2768 if re
.match(r
'^https?://', base_url
):
2770 if mpd_base_url
and base_url
.startswith('/'):
2771 base_url
= compat_urlparse
.urljoin(mpd_base_url
, base_url
)
2772 elif mpd_base_url
and not re
.match(r
'^https?://', base_url
):
2773 if not mpd_base_url
.endswith('/'):
2775 base_url
= mpd_base_url
+ base_url
2776 representation_id
= representation_attrib
.get('id')
2777 lang
= representation_attrib
.get('lang')
2778 url_el
= representation
.find(_add_ns('BaseURL'))
2779 filesize
= int_or_none(url_el
.attrib
.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el
is not None else None)
2780 bandwidth
= int_or_none(representation_attrib
.get('bandwidth'))
2781 if representation_id
is not None:
2782 format_id
= representation_id
2784 format_id
= content_type
2786 format_id
= mpd_id
+ '-' + format_id
2787 if content_type
in ('video', 'audio'):
2789 'format_id': format_id
,
2790 'manifest_url': mpd_url
,
2791 'ext': mimetype2ext(mime_type
),
2792 'width': int_or_none(representation_attrib
.get('width')),
2793 'height': int_or_none(representation_attrib
.get('height')),
2794 'tbr': float_or_none(bandwidth
, 1000),
2795 'asr': int_or_none(representation_attrib
.get('audioSamplingRate')),
2796 'fps': int_or_none(representation_attrib
.get('frameRate')),
2797 'language': lang
if lang
not in ('mul', 'und', 'zxx', 'mis') else None,
2798 'format_note': 'DASH %s' % content_type
,
2799 'filesize': filesize
,
2800 'container': mimetype2ext(mime_type
) + '_dash',
2803 elif content_type
== 'text':
2805 'ext': mimetype2ext(mime_type
),
2806 'manifest_url': mpd_url
,
2807 'filesize': filesize
,
2809 elif content_type
== 'image/jpeg':
2810 # See test case in VikiIE
2811 # https://www.viki.com/videos/1175236v-choosing-spouse-by-lottery-episode-1
2813 'format_id': format_id
,
2815 'manifest_url': mpd_url
,
2816 'format_note': 'DASH storyboards (jpeg)',
2820 if is_drm_protected(adaptation_set
) or is_drm_protected(representation
):
2822 representation_ms_info
= extract_multisegment_info(representation
, adaption_set_ms_info
)
2824 def prepare_template(template_name
, identifiers
):
2825 tmpl
= representation_ms_info
[template_name
]
2826 # First of, % characters outside $...$ templates
2827 # must be escaped by doubling for proper processing
2828 # by % operator string formatting used further (see
2829 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2835 in_template
= not in_template
2836 elif c
== '%' and not in_template
:
2838 # Next, $...$ templates are translated to their
2839 # %(...) counterparts to be used with % operator
2840 if representation_id
is not None:
2841 t
= t
.replace('$RepresentationID$', representation_id
)
2842 t
= re
.sub(r
'\$(%s)\$' % '|'.join(identifiers
), r
'%(\1)d', t
)
2843 t
= re
.sub(r
'\$(%s)%%([^$]+)\$' % '|'.join(identifiers
), r
'%(\1)\2', t
)
2844 t
.replace('$$', '$')
2847 # @initialization is a regular template like @media one
2848 # so it should be handled just the same way (see
2849 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2850 if 'initialization' in representation_ms_info
:
2851 initialization_template
= prepare_template(
2853 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2854 # $Time$ shall not be included for @initialization thus
2855 # only $Bandwidth$ remains
2857 representation_ms_info
['initialization_url'] = initialization_template
% {
2858 'Bandwidth': bandwidth
,
2861 def location_key(location
):
2862 return 'url' if re
.match(r
'^https?://', location
) else 'path'
2864 if 'segment_urls' not in representation_ms_info
and 'media' in representation_ms_info
:
2866 media_template
= prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2867 media_location_key
= location_key(media_template
)
2869 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2870 # can't be used at the same time
2871 if '%(Number' in media_template
and 's' not in representation_ms_info
:
2872 segment_duration
= None
2873 if 'total_number' not in representation_ms_info
and 'segment_duration' in representation_ms_info
:
2874 segment_duration
= float_or_none(representation_ms_info
['segment_duration'], representation_ms_info
['timescale'])
2875 representation_ms_info
['total_number'] = int(math
.ceil(float(period_duration
) / segment_duration
))
2876 representation_ms_info
['fragments'] = [{
2877 media_location_key
: media_template
% {
2878 'Number': segment_number
,
2879 'Bandwidth': bandwidth
,
2881 'duration': segment_duration
,
2882 } for segment_number
in range(
2883 representation_ms_info
['start_number'],
2884 representation_ms_info
['total_number'] + representation_ms_info
['start_number'])]
2886 # $Number*$ or $Time$ in media template with S list available
2887 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2888 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2889 representation_ms_info
['fragments'] = []
2892 segment_number
= representation_ms_info
['start_number']
2894 def add_segment_url():
2895 segment_url
= media_template
% {
2896 'Time': segment_time
,
2897 'Bandwidth': bandwidth
,
2898 'Number': segment_number
,
2900 representation_ms_info
['fragments'].append({
2901 media_location_key
: segment_url
,
2902 'duration': float_or_none(segment_d
, representation_ms_info
['timescale']),
2905 for num
, s
in enumerate(representation_ms_info
['s']):
2906 segment_time
= s
.get('t') or segment_time
2910 for r
in range(s
.get('r', 0)):
2911 segment_time
+= segment_d
2914 segment_time
+= segment_d
2915 elif 'segment_urls' in representation_ms_info
and 's' in representation_ms_info
:
2917 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2918 # or any YouTube dashsegments video
2921 timescale
= representation_ms_info
['timescale']
2922 for s
in representation_ms_info
['s']:
2923 duration
= float_or_none(s
['d'], timescale
)
2924 for r
in range(s
.get('r', 0) + 1):
2925 segment_uri
= representation_ms_info
['segment_urls'][segment_index
]
2927 location_key(segment_uri
): segment_uri
,
2928 'duration': duration
,
2931 representation_ms_info
['fragments'] = fragments
2932 elif 'segment_urls' in representation_ms_info
:
2933 # Segment URLs with no SegmentTimeline
2934 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2935 # https://github.com/ytdl-org/youtube-dl/pull/14844
2937 segment_duration
= float_or_none(
2938 representation_ms_info
['segment_duration'],
2939 representation_ms_info
['timescale']) if 'segment_duration' in representation_ms_info
else None
2940 for segment_url
in representation_ms_info
['segment_urls']:
2942 location_key(segment_url
): segment_url
,
2944 if segment_duration
:
2945 fragment
['duration'] = segment_duration
2946 fragments
.append(fragment
)
2947 representation_ms_info
['fragments'] = fragments
2948 # If there is a fragments key available then we correctly recognized fragmented media.
2949 # Otherwise we will assume unfragmented media with direct access. Technically, such
2950 # assumption is not necessarily correct since we may simply have no support for
2951 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2952 if 'fragments' in representation_ms_info
:
2954 # NB: mpd_url may be empty when MPD manifest is parsed from a string
2955 'url': mpd_url
or base_url
,
2956 'fragment_base_url': base_url
,
2958 'protocol': 'http_dash_segments' if mime_type
!= 'image/jpeg' else 'mhtml',
2960 if 'initialization_url' in representation_ms_info
:
2961 initialization_url
= representation_ms_info
['initialization_url']
2962 if not f
.get('url'):
2963 f
['url'] = initialization_url
2964 f
['fragments'].append({location_key(initialization_url): initialization_url}
)
2965 f
['fragments'].extend(representation_ms_info
['fragments'])
2967 # Assuming direct URL to unfragmented media.
2969 if content_type
in ('video', 'audio', 'image/jpeg'):
2970 f
['manifest_stream_number'] = stream_numbers
[f
['url']]
2971 stream_numbers
[f
['url']] += 1
2973 elif content_type
== 'text':
2974 subtitles
.setdefault(lang
or 'und', []).append(f
)
2976 return formats
, subtitles
2978 def _extract_ism_formats(self
, *args
, **kwargs
):
2979 fmts
, subs
= self
._extract
_ism
_formats
_and
_subtitles
(*args
, **kwargs
)
2981 self
._report
_ignoring
_subs
('ISM')
2984 def _extract_ism_formats_and_subtitles(self
, ism_url
, video_id
, ism_id
=None, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
):
2985 res
= self
._download
_xml
_handle
(
2987 note
='Downloading ISM manifest' if note
is None else note
,
2988 errnote
='Failed to download ISM manifest' if errnote
is None else errnote
,
2989 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
2996 return self
._parse
_ism
_formats
_and
_subtitles
(ism_doc
, urlh
.geturl(), ism_id
)
2998 def _parse_ism_formats_and_subtitles(self
, ism_doc
, ism_url
, ism_id
=None):
3000 Parse formats from ISM manifest.
3002 1. [MS-SSTR]: Smooth Streaming Protocol,
3003 https://msdn.microsoft.com/en-us/library/ff469518.aspx
3005 if ism_doc
.get('IsLive') == 'TRUE':
3008 duration
= int(ism_doc
.attrib
['Duration'])
3009 timescale
= int_or_none(ism_doc
.get('TimeScale')) or 10000000
3013 for stream
in ism_doc
.findall('StreamIndex'):
3014 stream_type
= stream
.get('Type')
3015 if stream_type
not in ('video', 'audio', 'text'):
3017 url_pattern
= stream
.attrib
['Url']
3018 stream_timescale
= int_or_none(stream
.get('TimeScale')) or timescale
3019 stream_name
= stream
.get('Name')
3020 stream_language
= stream
.get('Language', 'und')
3021 for track
in stream
.findall('QualityLevel'):
3022 fourcc
= track
.get('FourCC') or ('AACL' if track
.get('AudioTag') == '255' else None)
3023 # TODO: add support for WVC1 and WMAP
3024 if fourcc
not in ('H264', 'AVC1', 'AACL', 'TTML'):
3025 self
.report_warning('%s is not a supported codec' % fourcc
)
3027 tbr
= int(track
.attrib
['Bitrate']) // 1000
3028 # [1] does not mention Width and Height attributes. However,
3029 # they're often present while MaxWidth and MaxHeight are
3030 # missing, so should be used as fallbacks
3031 width
= int_or_none(track
.get('MaxWidth') or track
.get('Width'))
3032 height
= int_or_none(track
.get('MaxHeight') or track
.get('Height'))
3033 sampling_rate
= int_or_none(track
.get('SamplingRate'))
3035 track_url_pattern
= re
.sub(r
'{[Bb]itrate}', track
.attrib
['Bitrate'], url_pattern
)
3036 track_url_pattern
= compat_urlparse
.urljoin(ism_url
, track_url_pattern
)
3042 stream_fragments
= stream
.findall('c')
3043 for stream_fragment_index
, stream_fragment
in enumerate(stream_fragments
):
3044 fragment_ctx
['time'] = int_or_none(stream_fragment
.get('t')) or fragment_ctx
['time']
3045 fragment_repeat
= int_or_none(stream_fragment
.get('r')) or 1
3046 fragment_ctx
['duration'] = int_or_none(stream_fragment
.get('d'))
3047 if not fragment_ctx
['duration']:
3049 next_fragment_time
= int(stream_fragment
[stream_fragment_index
+ 1].attrib
['t'])
3051 next_fragment_time
= duration
3052 fragment_ctx
['duration'] = (next_fragment_time
- fragment_ctx
['time']) / fragment_repeat
3053 for _
in range(fragment_repeat
):
3055 'url': re
.sub(r
'{start[ _]time}', compat_str(fragment_ctx
['time']), track_url_pattern
),
3056 'duration': fragment_ctx
['duration'] / stream_timescale
,
3058 fragment_ctx
['time'] += fragment_ctx
['duration']
3060 if stream_type
== 'text':
3061 subtitles
.setdefault(stream_language
, []).append({
3065 'manifest_url': ism_url
,
3066 'fragments': fragments
,
3067 '_download_params': {
3068 'stream_type': stream_type
,
3069 'duration': duration
,
3070 'timescale': stream_timescale
,
3072 'language': stream_language
,
3073 'codec_private_data': track
.get('CodecPrivateData'),
3076 elif stream_type
in ('video', 'audio'):
3078 'format_id': join_nonempty(ism_id
, stream_name
, tbr
),
3080 'manifest_url': ism_url
,
3081 'ext': 'ismv' if stream_type
== 'video' else 'isma',
3085 'asr': sampling_rate
,
3086 'vcodec': 'none' if stream_type
== 'audio' else fourcc
,
3087 'acodec': 'none' if stream_type
== 'video' else fourcc
,
3089 'fragments': fragments
,
3090 'has_drm': ism_doc
.find('Protection') is not None,
3091 '_download_params': {
3092 'stream_type': stream_type
,
3093 'duration': duration
,
3094 'timescale': stream_timescale
,
3095 'width': width
or 0,
3096 'height': height
or 0,
3098 'language': stream_language
,
3099 'codec_private_data': track
.get('CodecPrivateData'),
3100 'sampling_rate': sampling_rate
,
3101 'channels': int_or_none(track
.get('Channels', 2)),
3102 'bits_per_sample': int_or_none(track
.get('BitsPerSample', 16)),
3103 'nal_unit_length_field': int_or_none(track
.get('NALUnitLengthField', 4)),
3106 return formats
, subtitles
3108 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):
3109 def absolute_url(item_url
):
3110 return urljoin(base_url
, item_url
)
3112 def parse_content_type(content_type
):
3113 if not content_type
:
3115 ctr
= re
.search(r
'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type
)
3117 mimetype
, codecs
= ctr
.groups()
3118 f
= parse_codecs(codecs
)
3119 f
['ext'] = mimetype2ext(mimetype
)
3123 def _media_formats(src
, cur_media_type
, type_info
={}):
3124 full_url
= absolute_url(src
)
3125 ext
= type_info
.get('ext') or determine_ext(full_url
)
3127 is_plain_url
= False
3128 formats
= self
._extract
_m
3u8_formats
(
3129 full_url
, video_id
, ext
='mp4',
3130 entry_protocol
=m3u8_entry_protocol
, m3u8_id
=m3u8_id
,
3131 preference
=preference
, quality
=quality
, fatal
=False)
3133 is_plain_url
= False
3134 formats
= self
._extract
_mpd
_formats
(
3135 full_url
, video_id
, mpd_id
=mpd_id
, fatal
=False)
3140 'vcodec': 'none' if cur_media_type
== 'audio' else None,
3142 return is_plain_url
, formats
3145 # amp-video and amp-audio are very similar to their HTML5 counterparts
3146 # so we wll include them right here (see
3147 # https://www.ampproject.org/docs/reference/components/amp-video)
3148 # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/
3149 _MEDIA_TAG_NAME_RE
= r
'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)'
3150 media_tags
= [(media_tag
, media_tag_name
, media_type
, '')
3151 for media_tag
, media_tag_name
, media_type
3152 in re
.findall(r
'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE
, webpage
)]
3153 media_tags
.extend(re
.findall(
3154 # We only allow video|audio followed by a whitespace or '>'.
3155 # Allowing more characters may end up in significant slow down (see
3156 # https://github.com/ytdl-org/youtube-dl/issues/11979, example URL:
3157 # http://www.porntrex.com/maps/videositemap.xml).
3158 r
'(?s)(<(?P<tag>%s)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>' % _MEDIA_TAG_NAME_RE
, webpage
))
3159 for media_tag
, _
, media_type
, media_content
in media_tags
:
3164 media_attributes
= extract_attributes(media_tag
)
3165 src
= strip_or_none(media_attributes
.get('src'))
3167 _
, formats
= _media_formats(src
, media_type
)
3168 media_info
['formats'].extend(formats
)
3169 media_info
['thumbnail'] = absolute_url(media_attributes
.get('poster'))
3171 for source_tag
in re
.findall(r
'<source[^>]+>', media_content
):
3172 s_attr
= extract_attributes(source_tag
)
3173 # data-video-src and data-src are non standard but seen
3174 # several times in the wild
3175 src
= strip_or_none(dict_get(s_attr
, ('src', 'data-video-src', 'data-src')))
3178 f
= parse_content_type(s_attr
.get('type'))
3179 is_plain_url
, formats
= _media_formats(src
, media_type
, f
)
3181 # width, height, res, label and title attributes are
3182 # all not standard but seen several times in the wild
3185 for lbl
in ('label', 'title')
3186 if str_or_none(s_attr
.get(lbl
))
3188 width
= int_or_none(s_attr
.get('width'))
3189 height
= (int_or_none(s_attr
.get('height'))
3190 or int_or_none(s_attr
.get('res')))
3191 if not width
or not height
:
3193 resolution
= parse_resolution(lbl
)
3196 width
= width
or resolution
.get('width')
3197 height
= height
or resolution
.get('height')
3199 tbr
= parse_bitrate(lbl
)
3208 'format_id': s_attr
.get('label') or s_attr
.get('title'),
3210 f
.update(formats
[0])
3211 media_info
['formats'].append(f
)
3213 media_info
['formats'].extend(formats
)
3214 for track_tag
in re
.findall(r
'<track[^>]+>', media_content
):
3215 track_attributes
= extract_attributes(track_tag
)
3216 kind
= track_attributes
.get('kind')
3217 if not kind
or kind
in ('subtitles', 'captions'):
3218 src
= strip_or_none(track_attributes
.get('src'))
3221 lang
= track_attributes
.get('srclang') or track_attributes
.get('lang') or track_attributes
.get('label')
3222 media_info
['subtitles'].setdefault(lang
, []).append({
3223 'url': absolute_url(src
),
3225 for f
in media_info
['formats']:
3226 f
.setdefault('http_headers', {})['Referer'] = base_url
3227 if media_info
['formats'] or media_info
['subtitles']:
3228 entries
.append(media_info
)
3231 def _extract_akamai_formats(self
, *args
, **kwargs
):
3232 fmts
, subs
= self
._extract
_akamai
_formats
_and
_subtitles
(*args
, **kwargs
)
3234 self
._report
_ignoring
_subs
('akamai')
3237 def _extract_akamai_formats_and_subtitles(self
, manifest_url
, video_id
, hosts
={}):
3238 signed
= 'hdnea=' in manifest_url
3240 # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
3241 manifest_url
= re
.sub(
3242 r
'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
3243 '', manifest_url
).strip('?')
3248 hdcore_sign
= 'hdcore=3.7.0'
3249 f4m_url
= re
.sub(r
'(https?://[^/]+)/i/', r
'\1/z/', manifest_url
).replace('/master.m3u8', '/manifest.f4m')
3250 hds_host
= hosts
.get('hds')
3252 f4m_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hds_host
, f4m_url
)
3253 if 'hdcore=' not in f4m_url
:
3254 f4m_url
+= ('&' if '?' in f4m_url
else '?') + hdcore_sign
3255 f4m_formats
= self
._extract
_f
4m
_formats
(
3256 f4m_url
, video_id
, f4m_id
='hds', fatal
=False)
3257 for entry
in f4m_formats
:
3258 entry
.update({'extra_param_to_segment_url': hdcore_sign}
)
3259 formats
.extend(f4m_formats
)
3261 m3u8_url
= re
.sub(r
'(https?://[^/]+)/z/', r
'\1/i/', manifest_url
).replace('/manifest.f4m', '/master.m3u8')
3262 hls_host
= hosts
.get('hls')
3264 m3u8_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hls_host
, m3u8_url
)
3265 m3u8_formats
, m3u8_subtitles
= self
._extract
_m
3u8_formats
_and
_subtitles
(
3266 m3u8_url
, video_id
, 'mp4', 'm3u8_native',
3267 m3u8_id
='hls', fatal
=False)
3268 formats
.extend(m3u8_formats
)
3269 subtitles
= self
._merge
_subtitles
(subtitles
, m3u8_subtitles
)
3271 http_host
= hosts
.get('http')
3272 if http_host
and m3u8_formats
and not signed
:
3273 REPL_REGEX
= r
'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
3274 qualities
= re
.match(REPL_REGEX
, m3u8_url
).group(2).split(',')
3275 qualities_length
= len(qualities
)
3276 if len(m3u8_formats
) in (qualities_length
, qualities_length
+ 1):
3278 for f
in m3u8_formats
:
3279 if f
['vcodec'] != 'none':
3280 for protocol
in ('http', 'https'):
3282 del http_f
['manifest_url']
3284 REPL_REGEX
, protocol
+ r
'://%s/\g<1>%s\3' % (http_host
, qualities
[i
]), f
['url'])
3286 'format_id': http_f
['format_id'].replace('hls-', protocol
+ '-'),
3288 'protocol': protocol
,
3290 formats
.append(http_f
)
3293 return formats
, subtitles
3295 def _extract_wowza_formats(self
, url
, video_id
, m3u8_entry_protocol
='m3u8_native', skip_protocols
=[]):
3296 query
= compat_urlparse
.urlparse(url
).query
3297 url
= re
.sub(r
'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url
)
3299 r
'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url
)
3300 url_base
= mobj
.group('url')
3301 http_base_url
= '%s%s:%s' % ('http', mobj
.group('s') or '', url_base
)
3304 def manifest_url(manifest
):
3305 m_url
= '%s/%s' % (http_base_url
, manifest
)
3307 m_url
+= '?%s' % query
3310 if 'm3u8' not in skip_protocols
:
3311 formats
.extend(self
._extract
_m
3u8_formats
(
3312 manifest_url('playlist.m3u8'), video_id
, 'mp4',
3313 m3u8_entry_protocol
, m3u8_id
='hls', fatal
=False))
3314 if 'f4m' not in skip_protocols
:
3315 formats
.extend(self
._extract
_f
4m
_formats
(
3316 manifest_url('manifest.f4m'),
3317 video_id
, f4m_id
='hds', fatal
=False))
3318 if 'dash' not in skip_protocols
:
3319 formats
.extend(self
._extract
_mpd
_formats
(
3320 manifest_url('manifest.mpd'),
3321 video_id
, mpd_id
='dash', fatal
=False))
3322 if re
.search(r
'(?:/smil:|\.smil)', url_base
):
3323 if 'smil' not in skip_protocols
:
3324 rtmp_formats
= self
._extract
_smil
_formats
(
3325 manifest_url('jwplayer.smil'),
3326 video_id
, fatal
=False)
3327 for rtmp_format
in rtmp_formats
:
3328 rtsp_format
= rtmp_format
.copy()
3329 rtsp_format
['url'] = '%s/%s' % (rtmp_format
['url'], rtmp_format
['play_path'])
3330 del rtsp_format
['play_path']
3331 del rtsp_format
['ext']
3332 rtsp_format
.update({
3333 'url': rtsp_format
['url'].replace('rtmp://', 'rtsp://'),
3334 'format_id': rtmp_format
['format_id'].replace('rtmp', 'rtsp'),
3337 formats
.extend([rtmp_format
, rtsp_format
])
3339 for protocol
in ('rtmp', 'rtsp'):
3340 if protocol
not in skip_protocols
:
3342 'url': '%s:%s' % (protocol
, url_base
),
3343 'format_id': protocol
,
3344 'protocol': protocol
,
3348 def _find_jwplayer_data(self
, webpage
, video_id
=None, transform_source
=js_to_json
):
3350 r
'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P
=quote
)\
)(?
!</script
>).*?\
.setup\s
*\
((?P
<options
>[^
)]+)\
)',
3354 jwplayer_data = self._parse_json(mobj.group('options
'),
3356 transform_source=transform_source)
3357 except ExtractorError:
3360 if isinstance(jwplayer_data, dict):
3361 return jwplayer_data
3363 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
3364 jwplayer_data = self._find_jwplayer_data(
3365 webpage, video_id, transform_source=js_to_json)
3366 return self._parse_jwplayer_data(
3367 jwplayer_data, video_id, *args, **kwargs)
3369 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
3370 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3371 # JWPlayer backward compatibility: flattened playlists
3372 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
3373 if 'playlist
' not in jwplayer_data:
3374 jwplayer_data = {'playlist': [jwplayer_data]}
3378 # JWPlayer backward compatibility: single playlist item
3379 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
3380 if not isinstance(jwplayer_data['playlist
'], list):
3381 jwplayer_data['playlist
'] = [jwplayer_data['playlist
']]
3383 for video_data in jwplayer_data['playlist
']:
3384 # JWPlayer backward compatibility: flattened sources
3385 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
3386 if 'sources
' not in video_data:
3387 video_data['sources
'] = [video_data]
3389 this_video_id = video_id or video_data['mediaid
']
3391 formats = self._parse_jwplayer_formats(
3392 video_data['sources
'], video_id=this_video_id, m3u8_id=m3u8_id,
3393 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
3396 tracks = video_data.get('tracks
')
3397 if tracks and isinstance(tracks, list):
3398 for track in tracks:
3399 if not isinstance(track, dict):
3401 track_kind = track.get('kind
')
3402 if not track_kind or not isinstance(track_kind, compat_str):
3404 if track_kind.lower() not in ('captions
', 'subtitles
'):
3406 track_url = urljoin(base_url, track.get('file'))
3409 subtitles.setdefault(track.get('label
') or 'en
', []).append({
3410 'url
': self._proto_relative_url(track_url)
3414 'id': this_video_id,
3415 'title
': unescapeHTML(video_data['title
'] if require_title else video_data.get('title
')),
3416 'description
': clean_html(video_data.get('description
')),
3417 'thumbnail
': urljoin(base_url, self._proto_relative_url(video_data.get('image
'))),
3418 'timestamp
': int_or_none(video_data.get('pubdate
')),
3419 'duration
': float_or_none(jwplayer_data.get('duration
') or video_data.get('duration
')),
3420 'subtitles
': subtitles,
3422 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
3423 if len(formats) == 1 and re.search(r'^
(?
:http|
//).*(?
:youtube\
.com|youtu\
.be
)/.+', formats[0]['url
']):
3425 '_type
': 'url_transparent
',
3426 'url
': formats[0]['url
'],
3429 self._sort_formats(formats)
3430 entry['formats
'] = formats
3431 entries.append(entry)
3432 if len(entries) == 1:
3435 return self.playlist_result(entries)
3437 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
3438 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3441 for source in jwplayer_sources_data:
3442 if not isinstance(source, dict):
3444 source_url = urljoin(
3445 base_url, self._proto_relative_url(source.get('file')))
3446 if not source_url or source_url in urls:
3448 urls.append(source_url)
3449 source_type = source.get('type') or ''
3450 ext = mimetype2ext(source_type) or determine_ext(source_url)
3451 if source_type == 'hls
' or ext == 'm3u8
':
3452 formats.extend(self._extract_m3u8_formats(
3453 source_url, video_id, 'mp4
', entry_protocol='m3u8_native
',
3454 m3u8_id=m3u8_id, fatal=False))
3455 elif source_type == 'dash
' or ext == 'mpd
':
3456 formats.extend(self._extract_mpd_formats(
3457 source_url, video_id, mpd_id=mpd_id, fatal=False))
3459 formats.extend(self._extract_smil_formats(
3460 source_url, video_id, fatal=False))
3461 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3462 elif source_type.startswith('audio
') or ext in (
3463 'oga
', 'aac
', 'mp3
', 'mpeg
', 'vorbis
'):
3470 height = int_or_none(source.get('height
'))
3472 # Often no height is provided but there is a label in
3473 # format like "1080p", "720p SD", or 1080.
3474 height = int_or_none(self._search_regex(
3475 r'^
(\d{3,4}
)[pP
]?
(?
:\b|$
)', compat_str(source.get('label
') or ''),
3476 'height
', default=None))
3479 'width
': int_or_none(source.get('width
')),
3481 'tbr
': int_or_none(source.get('bitrate
')),
3484 if source_url.startswith('rtmp
'):
3485 a_format['ext
'] = 'flv
'
3486 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3487 # of jwplayer.flash.swf
3488 rtmp_url_parts = re.split(
3489 r'((?
:mp4|mp3|flv
):)', source_url, 1)
3490 if len(rtmp_url_parts) == 3:
3491 rtmp_url, prefix, play_path = rtmp_url_parts
3494 'play_path
': prefix + play_path,
3497 a_format.update(rtmp_params)
3498 formats.append(a_format)
3501 def _live_title(self, name):
3502 self._downloader.deprecation_warning('yt_dlp
.InfoExtractor
._live
_title
is deprecated
and does
not work
as expected
')
3505 def _int(self, v, name, fatal=False, **kwargs):
3506 res = int_or_none(v, **kwargs)
3507 if 'get_attr
' in kwargs:
3508 print(getattr(v, kwargs['get_attr
']))
3510 msg = 'Failed to extract
%s: Could
not parse value
%r' % (name, v)
3512 raise ExtractorError(msg)
3514 self.report_warning(msg)
3517 def _float(self, v, name, fatal=False, **kwargs):
3518 res = float_or_none(v, **kwargs)
3520 msg = 'Failed to extract
%s: Could
not parse value
%r' % (name, v)
3522 raise ExtractorError(msg)
3524 self.report_warning(msg)
3527 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
3528 path='/', secure=False, discard=False, rest={}, **kwargs):
3529 cookie = compat_cookiejar_Cookie(
3530 0, name, value, port, port is not None, domain, True,
3531 domain.startswith('.'), path, True, secure, expire_time,
3532 discard, None, None, rest)
3533 self._downloader.cookiejar.set_cookie(cookie)
3535 def _get_cookies(self, url):
3536 """ Return a compat_cookies_SimpleCookie with the cookies for the url """
3537 req = sanitized_Request(url)
3538 self._downloader.cookiejar.add_cookie_header(req)
3539 return compat_cookies_SimpleCookie(req.get_header('Cookie
'))
3541 def _apply_first_set_cookie_header(self, url_handle, cookie):
3543 Apply first Set-Cookie header instead of the last. Experimental.
3545 Some sites (e.g. [1-3]) may serve two cookies under the same name
3546 in Set-Cookie header and expect the first (old) one to be set rather
3547 than second (new). However, as of RFC6265 the newer one cookie
3548 should be set into cookie store what actually happens.
3549 We will workaround this issue by resetting the cookie to
3550 the first one manually.
3551 1. https://new.vk.com/
3552 2. https://github.com/ytdl-org/youtube-dl/issues/9841#issuecomment-227871201
3553 3. https://learning.oreilly.com/
3555 for header, cookies in url_handle.headers.items():
3556 if header.lower() != 'set-cookie
':
3558 if sys.version_info[0] >= 3:
3559 cookies = cookies.encode('iso
-8859-1')
3560 cookies = cookies.decode('utf
-8')
3561 cookie_value = re.search(
3562 r'%s=(.+?
);.*?
\b[Dd
]omain
=(.+?
)(?
:[,;]|$
)' % cookie, cookies)
3564 value, domain = cookie_value.groups()
3565 self._set_cookie(domain, cookie, value)
3568 def get_testcases(self, include_onlymatching=False):
3569 t = getattr(self, '_TEST
', None)
3571 assert not hasattr(self, '_TESTS
'), \
3572 '%s has _TEST
and _TESTS
' % type(self).__name__
3575 tests = getattr(self, '_TESTS
', [])
3577 if not include_onlymatching and t.get('only_matching
', False):
3579 t['name
'] = type(self).__name__[:-len('IE
')]
3582 def is_suitable(self, age_limit):
3583 """ Test whether the extractor is generally suitable for the given
3584 age limit (i.e. pornographic sites are not, all others usually are) """
3586 any_restricted = False
3587 for tc in self.get_testcases(include_onlymatching=False):
3588 if tc.get('playlist
', []):
3589 tc = tc['playlist
'][0]
3590 is_restricted = age_restricted(
3591 tc.get('info_dict
', {}).get('age_limit
'), age_limit)
3592 if not is_restricted:
3594 any_restricted = any_restricted or is_restricted
3595 return not any_restricted
3597 def extract_subtitles(self, *args, **kwargs):
3598 if (self.get_param('writesubtitles
', False)
3599 or self.get_param('listsubtitles
')):
3600 return self._get_subtitles(*args, **kwargs)
3603 def _get_subtitles(self, *args, **kwargs):
3604 raise NotImplementedError('This method must be implemented by subclasses
')
3606 def extract_comments(self, *args, **kwargs):
3607 if not self.get_param('getcomments
'):
3609 generator = self._get_comments(*args, **kwargs)
3616 comments.append(next(generator))
3617 except StopIteration:
3619 except KeyboardInterrupt:
3620 self.to_screen('Interrupted by user
')
3621 except Exception as e:
3622 if self.get_param('ignoreerrors
') is not True:
3624 self._downloader.report_error(e)
3625 comment_count = len(comments)
3626 self.to_screen(f'Extracted {comment_count} comments
')
3628 'comments
': comments,
3629 'comment_count
': None if interrupted else comment_count
3633 def _get_comments(self, *args, **kwargs):
3634 raise NotImplementedError('This method must be implemented by subclasses
')
3637 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
3638 """ Merge subtitle items for one language. Items with duplicated URLs
3639 will be dropped. """
3640 list1_urls = set([item['url
'] for item in subtitle_list1])
3641 ret = list(subtitle_list1)
3642 ret.extend([item for item in subtitle_list2 if item['url
'] not in list1_urls])
3646 def _merge_subtitles(cls, *dicts, target=None):
3647 """ Merge subtitle dictionaries, language by language. """
3651 for lang, subs in d.items():
3652 target[lang] = cls._merge_subtitle_items(target.get(lang, []), subs)
3655 def extract_automatic_captions(self, *args, **kwargs):
3656 if (self.get_param('writeautomaticsub
', False)
3657 or self.get_param('listsubtitles
')):
3658 return self._get_automatic_captions(*args, **kwargs)
3661 def _get_automatic_captions(self, *args, **kwargs):
3662 raise NotImplementedError('This method must be implemented by subclasses
')
3664 def mark_watched(self, *args, **kwargs):
3665 if not self.get_param('mark_watched
', False):
3667 if (self._get_login_info()[0] is not None
3668 or self.get_param('cookiefile
')
3669 or self.get_param('cookiesfrombrowser
')):
3670 self._mark_watched(*args, **kwargs)
3672 def _mark_watched(self, *args, **kwargs):
3673 raise NotImplementedError('This method must be implemented by subclasses
')
3675 def geo_verification_headers(self):
3677 geo_verification_proxy = self.get_param('geo_verification_proxy
')
3678 if geo_verification_proxy:
3679 headers['Ytdl
-request
-proxy
'] = geo_verification_proxy
3682 def _generic_id(self, url):
3683 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
3685 def _generic_title(self, url):
3686 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
3689 def _availability(is_private=None, needs_premium=None, needs_subscription=None, needs_auth=None, is_unlisted=None):
3690 all_known = all(map(
3691 lambda x: x is not None,
3692 (is_private, needs_premium, needs_subscription, needs_auth, is_unlisted)))
3694 'private
' if is_private
3695 else 'premium_only
' if needs_premium
3696 else 'subscriber_only
' if needs_subscription
3697 else 'needs_auth
' if needs_auth
3698 else 'unlisted
' if is_unlisted
3699 else 'public
' if all_known
3702 def _configuration_arg(self, key, default=NO_DEFAULT, *, ie_key=None, casesense=False):
3704 @returns A list of values for the extractor argument given by "key"
3705 or "default" if no such key is present
3706 @param default The default value to return when the key is not present (default: [])
3707 @param casesense When false, the values are converted to lower case
3710 self._downloader.params, ('extractor_args
', (ie_key or self.ie_key()).lower(), key))
3712 return [] if default is NO_DEFAULT else default
3713 return list(val) if casesense else [x.lower() for x in val]
3716 class SearchInfoExtractor(InfoExtractor):
3718 Base class for paged search queries extractors.
3719 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
3720 Instances should define _SEARCH_KEY and optionally _MAX_RESULTS
3723 _MAX_RESULTS = float('inf
')
3726 def _make_valid_url(cls):
3727 return r'%s(?P
<prefix
>|
[1-9][0-9]*|all
):(?P
<query
>[\s\S
]+)' % cls._SEARCH_KEY
3729 def _real_extract(self, query):
3730 prefix, query = self._match_valid_url(query).group('prefix
', 'query
')
3732 return self._get_n_results(query, 1)
3733 elif prefix == 'all
':
3734 return self._get_n_results(query, self._MAX_RESULTS)
3738 raise ExtractorError('invalid download number
%s for query
"%s"' % (n, query))
3739 elif n > self._MAX_RESULTS:
3740 self.report_warning('%s returns
max %i results (you requested
%i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3741 n = self._MAX_RESULTS
3742 return self._get_n_results(query, n)
3744 def _get_n_results(self, query, n):
3745 """Get a specified number of results for a query.
3746 Either this function or _search_results must be overridden by subclasses """
3747 return self.playlist_result(
3748 itertools.islice(self._search_results(query), 0, None if n == float('inf
') else n),
3751 def _search_results(self, query):
3752 """Returns an iterator of search results"""
3753 raise NotImplementedError('This method must be implemented by subclasses
')
3756 def SEARCH_KEY(self):
3757 return self._SEARCH_KEY