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 location: Physical location where the video was filmed.
265 subtitles: The available subtitles as a dictionary in the format
266 {tag: subformats}. "tag" is usually a language code, and
267 "subformats" is a list sorted from lower to higher
268 preference, each element is a dictionary with the "ext"
270 * "data": The subtitles file contents
271 * "url": A URL pointing to the subtitles file
272 It can optionally also have:
273 * "name": Name or description of the subtitles
274 "ext" will be calculated from URL if missing
275 automatic_captions: Like 'subtitles'; contains automatically generated
276 captions instead of normal subtitles
277 duration: Length of the video in seconds, as an integer or float.
278 view_count: How many users have watched the video on the platform.
279 like_count: Number of positive ratings of the video
280 dislike_count: Number of negative ratings of the video
281 repost_count: Number of reposts of the video
282 average_rating: Average rating give by users, the scale used depends on the webpage
283 comment_count: Number of comments on the video
284 comments: A list of comments, each with one or more of the following
285 properties (all but one of text or html optional):
286 * "author" - human-readable name of the comment author
287 * "author_id" - user ID of the comment author
288 * "author_thumbnail" - The thumbnail of the comment author
290 * "html" - Comment as HTML
291 * "text" - Plain text of the comment
292 * "timestamp" - UNIX timestamp of comment
293 * "parent" - ID of the comment this one is replying to.
294 Set to "root" to indicate that this is a
295 comment to the original video.
296 * "like_count" - Number of positive ratings of the comment
297 * "dislike_count" - Number of negative ratings of the comment
298 * "is_favorited" - Whether the comment is marked as
299 favorite by the video uploader
300 * "author_is_uploader" - Whether the comment is made by
302 age_limit: Age restriction for the video, as an integer (years)
303 webpage_url: The URL to the video webpage, if given to yt-dlp it
304 should allow to get the same result again. (It will be set
305 by YoutubeDL if it's missing)
306 categories: A list of categories that the video falls in, for example
308 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
309 cast: A list of the video cast
310 is_live: True, False, or None (=unknown). Whether this video is a
311 live stream that goes on instead of a fixed-length video.
312 was_live: True, False, or None (=unknown). Whether this video was
313 originally a live stream.
314 live_status: 'is_live', 'is_upcoming', 'was_live', 'not_live' or None (=unknown)
315 If absent, automatically set from is_live, was_live
316 start_time: Time in seconds where the reproduction should start, as
317 specified in the URL.
318 end_time: Time in seconds where the reproduction should end, as
319 specified in the URL.
320 chapters: A list of dictionaries, with the following entries:
321 * "start_time" - The start time of the chapter in seconds
322 * "end_time" - The end time of the chapter in seconds
323 * "title" (optional, string)
324 playable_in_embed: Whether this video is allowed to play in embedded
325 players on other sites. Can be True (=always allowed),
326 False (=never allowed), None (=unknown), or a string
327 specifying the criteria for embedability (Eg: 'whitelist')
328 availability: Under what condition the video is available. One of
329 'private', 'premium_only', 'subscriber_only', 'needs_auth',
330 'unlisted' or 'public'. Use 'InfoExtractor._availability'
332 __post_extractor: A function to be called just before the metadata is
333 written to either disk, logger or console. The function
334 must return a dict which will be added to the info_dict.
335 This is usefull for additional information that is
336 time-consuming to extract. Note that the fields thus
337 extracted will not be available to output template and
338 match_filter. So, only "comments" and "comment_count" are
339 currently allowed to be extracted via this method.
341 The following fields should only be used when the video belongs to some logical
344 chapter: Name or title of the chapter the video belongs to.
345 chapter_number: Number of the chapter the video belongs to, as an integer.
346 chapter_id: Id of the chapter the video belongs to, as a unicode string.
348 The following fields should only be used when the video is an episode of some
349 series, programme or podcast:
351 series: Title of the series or programme the video episode belongs to.
352 series_id: Id of the series or programme the video episode belongs to, as a unicode string.
353 season: Title of the season the video episode belongs to.
354 season_number: Number of the season the video episode belongs to, as an integer.
355 season_id: Id of the season the video episode belongs to, as a unicode string.
356 episode: Title of the video episode. Unlike mandatory video title field,
357 this field should denote the exact title of the video episode
358 without any kind of decoration.
359 episode_number: Number of the video episode within a season, as an integer.
360 episode_id: Id of the video episode, as a unicode string.
362 The following fields should only be used when the media is a track or a part of
365 track: Title of the track.
366 track_number: Number of the track within an album or a disc, as an integer.
367 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
369 artist: Artist(s) of the track.
370 genre: Genre(s) of the track.
371 album: Title of the album the track belongs to.
372 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
373 album_artist: List of all artists appeared on the album (e.g.
374 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
376 disc_number: Number of the disc or other physical medium the track belongs to,
378 release_year: Year (YYYY) when the album was released.
379 composer: Composer of the piece
381 Unless mentioned otherwise, the fields should be Unicode strings.
383 Unless mentioned otherwise, None is equivalent to absence of information.
386 _type "playlist" indicates multiple videos.
387 There must be a key "entries", which is a list, an iterable, or a PagedList
388 object, each element of which is a valid dictionary by this specification.
390 Additionally, playlists can have "id", "title", and any other relevent
391 attributes with the same semantics as videos (see above).
393 It can also have the following optional fields:
395 playlist_count: The total number of videos in a playlist. If not given,
396 YoutubeDL tries to calculate it from "entries"
399 _type "multi_video" indicates that there are multiple videos that
400 form a single show, for examples multiple acts of an opera or TV episode.
401 It must have an entries key like a playlist and contain all the keys
402 required for a video at the same time.
405 _type "url" indicates that the video must be extracted from another
406 location, possibly by a different extractor. Its only required key is:
407 "url" - the next URL to extract.
408 The key "ie_key" can be set to the class name (minus the trailing "IE",
409 e.g. "Youtube") if the extractor class is known in advance.
410 Additionally, the dictionary may have any properties of the resolved entity
411 known in advance, for example "title" if the title of the referred video is
415 _type "url_transparent" entities have the same specification as "url", but
416 indicate that the given additional information is more precise than the one
417 associated with the resolved URL.
418 This is useful when a site employs a video service that hosts the video and
419 its technical metadata, but that video service does not embed a useful
420 title, description etc.
423 Subclasses of this one should re-define the _real_initialize() and
424 _real_extract() methods and define a _VALID_URL regexp.
425 Probably, they should also be added to the list of extractors.
427 Subclasses may also override suitable() if necessary, but ensure the function
428 signature is preserved and that this function imports everything it needs
429 (except other extractors), so that lazy_extractors works correctly
431 _GEO_BYPASS attribute may be set to False in order to disable
432 geo restriction bypass mechanisms for a particular extractor.
433 Though it won't disable explicit geo restriction bypass based on
434 country code provided with geo_bypass_country.
436 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
437 countries for this extractor. One of these countries will be used by
438 geo restriction bypass mechanism right away in order to bypass
439 geo restriction, of course, if the mechanism is not disabled.
441 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
442 IP blocks in CIDR notation for this extractor. One of these IP blocks
443 will be used by geo restriction bypass mechanism similarly
446 The _WORKING attribute should be set to False for broken IEs
447 in order to warn the users and skip the tests.
452 _x_forwarded_for_ip
= None
454 _GEO_COUNTRIES
= None
455 _GEO_IP_BLOCKS
= None
459 'any': 'Use --cookies, --username and --password, or --netrc to provide account credentials',
461 'Use --cookies-from-browser or --cookies for the authentication. '
462 'See https://github.com/ytdl-org/youtube-dl#how-do-i-pass-cookies-to-youtube-dl for how to manually pass cookies'),
463 'password': 'Use --username and --password, or --netrc to provide account credentials',
466 def __init__(self
, downloader
=None):
467 """Constructor. Receives an optional downloader (a YoutubeDL instance).
468 If a downloader is not passed during initialization,
469 it must be set using "set_downloader()" before "extract()" is called"""
471 self
._x
_forwarded
_for
_ip
= None
472 self
._printed
_messages
= set()
473 self
.set_downloader(downloader
)
476 def _match_valid_url(cls
, url
):
477 # This does not use has/getattr intentionally - we want to know whether
478 # we have cached the regexp for *this* class, whereas getattr would also
479 # match the superclass
480 if '_VALID_URL_RE' not in cls
.__dict
__:
481 if '_VALID_URL' not in cls
.__dict
__:
482 cls
._VALID
_URL
= cls
._make
_valid
_url
()
483 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
484 return cls
._VALID
_URL
_RE
.match(url
)
487 def suitable(cls
, url
):
488 """Receives a URL and returns True if suitable for this IE."""
489 # This function must import everything it needs (except other extractors),
490 # so that lazy_extractors works correctly
491 return cls
._match
_valid
_url
(url
) is not None
494 def _match_id(cls
, url
):
495 return cls
._match
_valid
_url
(url
).group('id')
498 def get_temp_id(cls
, url
):
500 return cls
._match
_id
(url
)
501 except (IndexError, AttributeError):
506 """Getter method for _WORKING."""
509 def initialize(self
):
510 """Initializes an instance (authentication, etc)."""
511 self
._printed
_messages
= set()
512 self
._initialize
_geo
_bypass
({
513 'countries': self
._GEO
_COUNTRIES
,
514 'ip_blocks': self
._GEO
_IP
_BLOCKS
,
517 self
._real
_initialize
()
520 def _initialize_geo_bypass(self
, geo_bypass_context
):
522 Initialize geo restriction bypass mechanism.
524 This method is used to initialize geo bypass mechanism based on faking
525 X-Forwarded-For HTTP header. A random country from provided country list
526 is selected and a random IP belonging to this country is generated. This
527 IP will be passed as X-Forwarded-For HTTP header in all subsequent
530 This method will be used for initial geo bypass mechanism initialization
531 during the instance initialization with _GEO_COUNTRIES and
534 You may also manually call it from extractor's code if geo bypass
535 information is not available beforehand (e.g. obtained during
536 extraction) or due to some other reason. In this case you should pass
537 this information in geo bypass context passed as first argument. It may
538 contain following fields:
540 countries: List of geo unrestricted countries (similar
542 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
543 (similar to _GEO_IP_BLOCKS)
546 if not self
._x
_forwarded
_for
_ip
:
548 # Geo bypass mechanism is explicitly disabled by user
549 if not self
.get_param('geo_bypass', True):
552 if not geo_bypass_context
:
553 geo_bypass_context
= {}
555 # Backward compatibility: previously _initialize_geo_bypass
556 # expected a list of countries, some 3rd party code may still use
558 if isinstance(geo_bypass_context
, (list, tuple)):
559 geo_bypass_context
= {
560 'countries': geo_bypass_context
,
563 # The whole point of geo bypass mechanism is to fake IP
564 # as X-Forwarded-For HTTP header based on some IP block or
567 # Path 1: bypassing based on IP block in CIDR notation
569 # Explicit IP block specified by user, use it right away
570 # regardless of whether extractor is geo bypassable or not
571 ip_block
= self
.get_param('geo_bypass_ip_block', None)
573 # Otherwise use random IP block from geo bypass context but only
574 # if extractor is known as geo bypassable
576 ip_blocks
= geo_bypass_context
.get('ip_blocks')
577 if self
._GEO
_BYPASS
and ip_blocks
:
578 ip_block
= random
.choice(ip_blocks
)
581 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(ip_block
)
582 self
._downloader
.write_debug(
583 '[debug] Using fake IP %s as X-Forwarded-For' % self
._x
_forwarded
_for
_ip
)
586 # Path 2: bypassing based on country code
588 # Explicit country code specified by user, use it right away
589 # regardless of whether extractor is geo bypassable or not
590 country
= self
.get_param('geo_bypass_country', None)
592 # Otherwise use random country code from geo bypass context but
593 # only if extractor is known as geo bypassable
595 countries
= geo_bypass_context
.get('countries')
596 if self
._GEO
_BYPASS
and countries
:
597 country
= random
.choice(countries
)
600 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country
)
601 self
._downloader
.write_debug(
602 'Using fake IP %s (%s) as X-Forwarded-For' % (self
._x
_forwarded
_for
_ip
, country
.upper()))
604 def extract(self
, url
):
605 """Extracts URL information and returns it in list of dicts."""
610 self
.write_debug('Extracting URL: %s' % url
)
611 ie_result
= self
._real
_extract
(url
)
612 if ie_result
is None:
614 if self
._x
_forwarded
_for
_ip
:
615 ie_result
['__x_forwarded_for_ip'] = self
._x
_forwarded
_for
_ip
616 subtitles
= ie_result
.get('subtitles')
617 if (subtitles
and 'live_chat' in subtitles
618 and 'no-live-chat' in self
.get_param('compat_opts', [])):
619 del subtitles
['live_chat']
621 except GeoRestrictedError
as e
:
622 if self
.__maybe
_fake
_ip
_and
_retry
(e
.countries
):
625 except UnsupportedError
:
627 except ExtractorError
as e
:
629 'video_id': e
.video_id
or self
.get_temp_id(url
),
631 'tb': e
.traceback
or sys
.exc_info()[2],
632 'expected': e
.expected
,
635 if hasattr(e
, 'countries'):
636 kwargs
['countries'] = e
.countries
637 raise type(e
)(e
.msg
, **kwargs
)
638 except compat_http_client
.IncompleteRead
as e
:
639 raise ExtractorError('A network error has occurred.', cause
=e
, expected
=True, video_id
=self
.get_temp_id(url
))
640 except (KeyError, StopIteration) as e
:
641 raise ExtractorError('An extractor error has occurred.', cause
=e
, video_id
=self
.get_temp_id(url
))
643 def __maybe_fake_ip_and_retry(self
, countries
):
644 if (not self
.get_param('geo_bypass_country', None)
646 and self
.get_param('geo_bypass', True)
647 and not self
._x
_forwarded
_for
_ip
649 country_code
= random
.choice(countries
)
650 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country_code
)
651 if self
._x
_forwarded
_for
_ip
:
653 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
654 % (self
._x
_forwarded
_for
_ip
, country_code
.upper()))
658 def set_downloader(self
, downloader
):
659 """Sets the downloader for this IE."""
660 self
._downloader
= downloader
662 def _real_initialize(self
):
663 """Real initialization process. Redefine in subclasses."""
666 def _real_extract(self
, url
):
667 """Real extraction process. Redefine in subclasses."""
672 """A string for getting the InfoExtractor with get_info_extractor"""
673 return cls
.__name
__[:-2]
677 return compat_str(type(self
).__name
__[:-2])
680 def __can_accept_status_code(err
, expected_status
):
681 assert isinstance(err
, compat_urllib_error
.HTTPError
)
682 if expected_status
is None:
684 elif callable(expected_status
):
685 return expected_status(err
.code
) is True
687 return err
.code
in variadic(expected_status
)
689 def _request_webpage(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
, expected_status
=None):
691 Return the response handle.
693 See _download_webpage docstring for arguments specification.
695 if not self
._downloader
._first
_webpage
_request
:
696 sleep_interval
= self
.get_param('sleep_interval_requests') or 0
697 if sleep_interval
> 0:
698 self
.to_screen('Sleeping %s seconds ...' % sleep_interval
)
699 time
.sleep(sleep_interval
)
701 self
._downloader
._first
_webpage
_request
= False
704 self
.report_download_webpage(video_id
)
705 elif note
is not False:
707 self
.to_screen('%s' % (note
,))
709 self
.to_screen('%s: %s' % (video_id
, note
))
711 # Some sites check X-Forwarded-For HTTP header in order to figure out
712 # the origin of the client behind proxy. This allows bypassing geo
713 # restriction by faking this header's value to IP that belongs to some
714 # geo unrestricted country. We will do so once we encounter any
715 # geo restriction error.
716 if self
._x
_forwarded
_for
_ip
:
717 if 'X-Forwarded-For' not in headers
:
718 headers
['X-Forwarded-For'] = self
._x
_forwarded
_for
_ip
720 if isinstance(url_or_request
, compat_urllib_request
.Request
):
721 url_or_request
= update_Request(
722 url_or_request
, data
=data
, headers
=headers
, query
=query
)
725 url_or_request
= update_url_query(url_or_request
, query
)
726 if data
is not None or headers
:
727 url_or_request
= sanitized_Request(url_or_request
, data
, headers
)
729 return self
._downloader
.urlopen(url_or_request
)
730 except network_exceptions
as err
:
731 if isinstance(err
, compat_urllib_error
.HTTPError
):
732 if self
.__can
_accept
_status
_code
(err
, expected_status
):
733 # Retain reference to error to prevent file object from
734 # being closed before it can be read. Works around the
735 # effects of <https://bugs.python.org/issue15002>
736 # introduced in Python 3.4.1.
743 errnote
= 'Unable to download webpage'
745 errmsg
= '%s: %s' % (errnote
, error_to_compat_str(err
))
747 raise ExtractorError(errmsg
, sys
.exc_info()[2], cause
=err
)
749 self
.report_warning(errmsg
)
752 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):
754 Return a tuple (page content as string, URL handle).
756 See _download_webpage docstring for arguments specification.
758 # Strip hashes from the URL (#1038)
759 if isinstance(url_or_request
, (compat_str
, str)):
760 url_or_request
= url_or_request
.partition('#')[0]
762 urlh
= self
._request
_webpage
(url_or_request
, video_id
, note
, errnote
, fatal
, data
=data
, headers
=headers
, query
=query
, expected_status
=expected_status
)
766 content
= self
._webpage
_read
_content
(urlh
, url_or_request
, video_id
, note
, errnote
, fatal
, encoding
=encoding
)
767 return (content
, urlh
)
770 def _guess_encoding_from_content(content_type
, webpage_bytes
):
771 m
= re
.match(r
'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type
)
773 encoding
= m
.group(1)
775 m
= re
.search(br
'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
776 webpage_bytes[:1024])
778 encoding = m.group(1).decode('ascii')
779 elif webpage_bytes.startswith(b'\xff\xfe'):
786 def __check_blocked(self, content):
787 first_block = content[:512]
788 if ('<title>Access to this site is blocked</title>' in content
789 and 'Websense' in first_block):
790 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
791 blocked_iframe = self._html_search_regex(
792 r'<iframe src="([^
"]+)"', content,
793 'Websense information URL
', default=None)
795 msg += ' Visit
%s for more details
' % blocked_iframe
796 raise ExtractorError(msg, expected=True)
797 if '<title
>The URL you requested has been blocked
</title
>' in first_block:
799 'Access to this webpage has been blocked by Indian censorship
. '
800 'Use a VPN
or proxy
server (with --proxy
) to route around it
.')
801 block_msg = self._html_search_regex(
802 r'</h1
><p
>(.*?
)</p
>',
803 content, 'block message
', default=None)
805 msg += ' (Message
: "%s")' % block_msg.replace('\n', ' ')
806 raise ExtractorError(msg, expected=True)
807 if ('<title
>TTK
:: Доступ к ресурсу ограничен
</title
>' in content
808 and 'blocklist
.rkn
.gov
.ru
' in content):
809 raise ExtractorError(
810 'Access to this webpage has been blocked by decision of the Russian government
. '
811 'Visit http
://blocklist
.rkn
.gov
.ru
/ for a block reason
.',
814 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
815 content_type = urlh.headers.get('Content
-Type
', '')
816 webpage_bytes = urlh.read()
817 if prefix is not None:
818 webpage_bytes = prefix + webpage_bytes
820 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
821 if self.get_param('dump_intermediate_pages
', False):
822 self.to_screen('Dumping request to
' + urlh.geturl())
823 dump = base64.b64encode(webpage_bytes).decode('ascii
')
824 self._downloader.to_screen(dump)
825 if self.get_param('write_pages
', False):
826 basen = '%s_%s' % (video_id, urlh.geturl())
827 trim_length = self.get_param('trim_file_name
') or 240
828 if len(basen) > trim_length:
829 h = '___
' + hashlib.md5(basen.encode('utf
-8')).hexdigest()
830 basen = basen[:trim_length - len(h)] + h
831 raw_filename = basen + '.dump
'
832 filename = sanitize_filename(raw_filename, restricted=True)
833 self.to_screen('Saving request to
' + filename)
834 # Working around MAX_PATH limitation on Windows (see
835 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
836 if compat_os_name == 'nt
':
837 absfilepath = os.path.abspath(filename)
838 if len(absfilepath) > 259:
839 filename = '\\\\?
\\' + absfilepath
840 with open(filename, 'wb
') as outf:
841 outf.write(webpage_bytes)
844 content = webpage_bytes.decode(encoding, 'replace
')
846 content = webpage_bytes.decode('utf
-8', 'replace
')
848 self.__check_blocked(content)
852 def _download_webpage(
853 self, url_or_request, video_id, note=None, errnote=None,
854 fatal=True, tries=1, timeout=5, encoding=None, data=None,
855 headers={}, query={}, expected_status=None):
857 Return the data of the page as a string.
860 url_or_request -- plain text URL as a string or
861 a compat_urllib_request.Requestobject
862 video_id -- Video/playlist/item identifier (string)
865 note -- note printed before downloading (string)
866 errnote -- note printed in case of an error (string)
867 fatal -- flag denoting whether error should be considered fatal,
868 i.e. whether it should cause ExtractionError to be raised,
869 otherwise a warning will be reported and extraction continued
870 tries -- number of tries
871 timeout -- sleep interval between tries
872 encoding -- encoding for a page content decoding, guessed automatically
873 when not explicitly specified
874 data -- POST data (bytes)
875 headers -- HTTP headers (dict)
876 query -- URL query (dict)
877 expected_status -- allows to accept failed HTTP requests (non 2xx
878 status code) by explicitly specifying a set of accepted status
879 codes. Can be any of the following entities:
880 - an integer type specifying an exact failed status code to
882 - a list or a tuple of integer types specifying a list of
883 failed status codes to accept
884 - a callable accepting an actual failed status code and
885 returning True if it should be accepted
886 Note that this argument does not affect success status codes (2xx)
887 which are always accepted.
892 while success is False:
894 res = self._download_webpage_handle(
895 url_or_request, video_id, note, errnote, fatal,
896 encoding=encoding, data=data, headers=headers, query=query,
897 expected_status=expected_status)
899 except compat_http_client.IncompleteRead as e:
901 if try_count >= tries:
903 self._sleep(timeout, video_id)
910 def _download_xml_handle(
911 self, url_or_request, video_id, note='Downloading XML
',
912 errnote='Unable to download XML
', transform_source=None,
913 fatal=True, encoding=None, data=None, headers={}, query={},
914 expected_status=None):
916 Return a tuple (xml as an compat_etree_Element, URL handle).
918 See _download_webpage docstring for arguments specification.
920 res = self._download_webpage_handle(
921 url_or_request, video_id, note, errnote, fatal=fatal,
922 encoding=encoding, data=data, headers=headers, query=query,
923 expected_status=expected_status)
926 xml_string, urlh = res
927 return self._parse_xml(
928 xml_string, video_id, transform_source=transform_source,
932 self, url_or_request, video_id,
933 note='Downloading XML
', errnote='Unable to download XML
',
934 transform_source=None, fatal=True, encoding=None,
935 data=None, headers={}, query={}, expected_status=None):
937 Return the xml as an compat_etree_Element.
939 See _download_webpage docstring for arguments specification.
941 res = self._download_xml_handle(
942 url_or_request, video_id, note=note, errnote=errnote,
943 transform_source=transform_source, fatal=fatal, encoding=encoding,
944 data=data, headers=headers, query=query,
945 expected_status=expected_status)
946 return res if res is False else res[0]
948 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
950 xml_string = transform_source(xml_string)
952 return compat_etree_fromstring(xml_string.encode('utf
-8'))
953 except compat_xml_parse_error as ve:
954 errmsg = '%s: Failed to parse XML
' % video_id
956 raise ExtractorError(errmsg, cause=ve)
958 self.report_warning(errmsg + str(ve))
960 def _download_json_handle(
961 self, url_or_request, video_id, note='Downloading JSON metadata
',
962 errnote='Unable to download JSON metadata
', transform_source=None,
963 fatal=True, encoding=None, data=None, headers={}, query={},
964 expected_status=None):
966 Return a tuple (JSON object, URL handle).
968 See _download_webpage docstring for arguments specification.
970 res = self._download_webpage_handle(
971 url_or_request, video_id, note, errnote, fatal=fatal,
972 encoding=encoding, data=data, headers=headers, query=query,
973 expected_status=expected_status)
976 json_string, urlh = res
977 return self._parse_json(
978 json_string, video_id, transform_source=transform_source,
982 self, url_or_request, video_id, note='Downloading JSON metadata
',
983 errnote='Unable to download JSON metadata
', transform_source=None,
984 fatal=True, encoding=None, data=None, headers={}, query={},
985 expected_status=None):
987 Return the JSON object as a dict.
989 See _download_webpage docstring for arguments specification.
991 res = self._download_json_handle(
992 url_or_request, video_id, note=note, errnote=errnote,
993 transform_source=transform_source, fatal=fatal, encoding=encoding,
994 data=data, headers=headers, query=query,
995 expected_status=expected_status)
996 return res if res is False else res[0]
998 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
1000 json_string = transform_source(json_string)
1002 return json.loads(json_string)
1003 except ValueError as ve:
1004 errmsg = '%s: Failed to parse JSON
' % video_id
1006 raise ExtractorError(errmsg, cause=ve)
1008 self.report_warning(errmsg + str(ve))
1010 def _parse_socket_response_as_json(self, data, video_id, transform_source=None, fatal=True):
1011 return self._parse_json(
1012 data[data.find('{'):data.rfind('}
') + 1],
1013 video_id, transform_source, fatal)
1015 def _download_socket_json_handle(
1016 self, url_or_request, video_id, note='Polling socket
',
1017 errnote='Unable to poll socket
', transform_source=None,
1018 fatal=True, encoding=None, data=None, headers={}, query={},
1019 expected_status=None):
1021 Return a tuple (JSON object, URL handle).
1023 See _download_webpage docstring for arguments specification.
1025 res = self._download_webpage_handle(
1026 url_or_request, video_id, note, errnote, fatal=fatal,
1027 encoding=encoding, data=data, headers=headers, query=query,
1028 expected_status=expected_status)
1032 return self._parse_socket_response_as_json(
1033 webpage, video_id, transform_source=transform_source,
1036 def _download_socket_json(
1037 self, url_or_request, video_id, note='Polling socket
',
1038 errnote='Unable to poll socket
', transform_source=None,
1039 fatal=True, encoding=None, data=None, headers={}, query={},
1040 expected_status=None):
1042 Return the JSON object as a dict.
1044 See _download_webpage docstring for arguments specification.
1046 res = self._download_socket_json_handle(
1047 url_or_request, video_id, note=note, errnote=errnote,
1048 transform_source=transform_source, fatal=fatal, encoding=encoding,
1049 data=data, headers=headers, query=query,
1050 expected_status=expected_status)
1051 return res if res is False else res[0]
1053 def report_warning(self, msg, video_id=None, *args, only_once=False, **kwargs):
1054 idstr = format_field(video_id, template='%s: ')
1055 msg = f'[{self.IE_NAME}
] {idstr}{msg}
'
1057 if f'WARNING
: {msg}
' in self._printed_messages:
1059 self._printed_messages.add(f'WARNING
: {msg}
')
1060 self._downloader.report_warning(msg, *args, **kwargs)
1062 def to_screen(self, msg, *args, **kwargs):
1063 """Print msg to screen, prefixing it with '[ie_name
]'"""
1064 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg), *args, **kwargs)
1066 def write_debug(self, msg, *args, **kwargs):
1067 self._downloader.write_debug('[%s] %s' % (self.IE_NAME, msg), *args, **kwargs)
1069 def get_param(self, name, default=None, *args, **kwargs):
1070 if self._downloader:
1071 return self._downloader.params.get(name, default, *args, **kwargs)
1074 def report_drm(self, video_id, partial=False):
1075 self.raise_no_formats('This video
is DRM protected
', expected=True, video_id=video_id)
1077 def report_extraction(self, id_or_name):
1078 """Report information extraction."""
1079 self.to_screen('%s: Extracting information
' % id_or_name)
1081 def report_download_webpage(self, video_id):
1082 """Report webpage download."""
1083 self.to_screen('%s: Downloading webpage
' % video_id)
1085 def report_age_confirmation(self):
1086 """Report attempt to confirm age."""
1087 self.to_screen('Confirming age
')
1089 def report_login(self):
1090 """Report attempt to log in."""
1091 self.to_screen('Logging
in')
1093 def raise_login_required(
1094 self, msg='This video
is only available
for registered users
',
1095 metadata_available=False, method='any
'):
1096 if metadata_available and (
1097 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1098 self.report_warning(msg)
1099 if method is not None:
1100 msg = '%s. %s' % (msg, self._LOGIN_HINTS[method])
1101 raise ExtractorError(msg, expected=True)
1103 def raise_geo_restricted(
1104 self, msg='This video
is not available
from your location due to geo restriction
',
1105 countries=None, metadata_available=False):
1106 if metadata_available and (
1107 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1108 self.report_warning(msg)
1110 raise GeoRestrictedError(msg, countries=countries)
1112 def raise_no_formats(self, msg, expected=False, video_id=None):
1114 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1115 self.report_warning(msg, video_id)
1116 elif isinstance(msg, ExtractorError):
1119 raise ExtractorError(msg, expected=expected, video_id=video_id)
1121 # Methods for following #608
1123 def url_result(url, ie=None, video_id=None, video_title=None, **kwargs):
1124 """Returns a URL that points to a page that should be processed"""
1125 # TODO: ie should be the class used for getting the info
1126 video_info = {'_type
': 'url
',
1129 video_info.update(kwargs)
1130 if video_id is not None:
1131 video_info['id'] = video_id
1132 if video_title is not None:
1133 video_info['title
'] = video_title
1136 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None):
1138 self.url_result(self._proto_relative_url(getter(m) if getter else m), ie)
1140 return self.playlist_result(
1141 urls, playlist_id=playlist_id, playlist_title=playlist_title)
1144 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None, **kwargs):
1145 """Returns a playlist"""
1146 video_info = {'_type
': 'playlist
',
1148 video_info.update(kwargs)
1150 video_info['id'] = playlist_id
1152 video_info['title
'] = playlist_title
1153 if playlist_description is not None:
1154 video_info['description
'] = playlist_description
1157 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1159 Perform a regex search on the given string, using a single or a list of
1160 patterns returning the first matching group.
1161 In case of failure return a default value or raise a WARNING or a
1162 RegexNotFoundError, depending on fatal, specifying the field name.
1164 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
1165 mobj = re.search(pattern, string, flags)
1168 mobj = re.search(p, string, flags)
1172 _name = self._downloader._format_err(name, self._downloader.Styles.EMPHASIS)
1176 # return the first matching group
1177 return next(g for g in mobj.groups() if g is not None)
1178 elif isinstance(group, (list, tuple)):
1179 return tuple(mobj.group(g) for g in group)
1181 return mobj.group(group)
1182 elif default is not NO_DEFAULT:
1185 raise RegexNotFoundError('Unable to extract
%s' % _name)
1187 self.report_warning('unable to extract
%s' % _name + bug_reports_message())
1190 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1192 Like _search_regex, but strips HTML tags and unescapes entities.
1194 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1196 return clean_html(res).strip()
1200 def _get_netrc_login_info(self, netrc_machine=None):
1203 netrc_machine = netrc_machine or self._NETRC_MACHINE
1205 if self.get_param('usenetrc
', False):
1207 netrc_file = compat_expanduser(self.get_param('netrc_location
') or '~
')
1208 if os.path.isdir(netrc_file):
1209 netrc_file = os.path.join(netrc_file, '.netrc
')
1210 info = netrc.netrc(file=netrc_file).authenticators(netrc_machine)
1211 if info is not None:
1215 raise netrc.NetrcParseError(
1216 'No authenticators
for %s' % netrc_machine)
1217 except (IOError, netrc.NetrcParseError) as err:
1218 self.report_warning(
1219 'parsing
.netrc
: %s' % error_to_compat_str(err))
1221 return username, password
1223 def _get_login_info(self, username_option='username
', password_option='password
', netrc_machine=None):
1225 Get the login info as (username, password)
1226 First look for the manually specified credentials using username_option
1227 and password_option as keys in params dictionary. If no such credentials
1228 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1230 If there's no info available
, return (None, None)
1233 # Attempt to use provided username and password or .netrc data
1234 username = self.get_param(username_option)
1235 if username is not None:
1236 password = self.get_param(password_option)
1238 username, password = self._get_netrc_login_info(netrc_machine)
1240 return username, password
1242 def _get_tfa_info(self, note='two-factor verification code'):
1244 Get the two
-factor authentication info
1245 TODO
- asking the user will be required
for sms
/phone verify
1246 currently just uses the command line option
1247 If there
's no info available, return None
1250 tfa = self.get_param('twofactor
')
1254 return compat_getpass('Type
%s and press
[Return
]: ' % note)
1256 # Helper functions for extracting OpenGraph info
1258 def _og_regexes(prop):
1259 content_re = r'content
=(?
:"([^"]+?
)"|\'([^\']+?)\'|\s*([^\s"\'=<>`
]+?
))'
1260 property_re = (r'(?
:name|
property)=(?
:\'og
[:-]%(prop)s\'|
"og[:-]%(prop)s"|\s
*og
[:-]%(prop)s\b)'
1261 % {'prop': re.escape(prop)})
1262 template = r'<meta
[^
>]+?
%s[^
>]+?
%s'
1264 template % (property_re, content_re),
1265 template % (content_re, property_re),
1269 def _meta_regex(prop):
1270 return r'''(?isx)<meta
1271 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1272 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1274 def _og_search_property(self, prop, html, name=None, **kargs):
1275 prop = variadic(prop)
1277 name = 'OpenGraph
%s' % prop[0]
1280 og_regexes.extend(self._og_regexes(p))
1281 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1284 return unescapeHTML(escaped)
1286 def _og_search_thumbnail(self, html, **kargs):
1287 return self._og_search_property('image
', html, 'thumbnail URL
', fatal=False, **kargs)
1289 def _og_search_description(self, html, **kargs):
1290 return self._og_search_property('description
', html, fatal=False, **kargs)
1292 def _og_search_title(self, html, **kargs):
1293 return self._og_search_property('title
', html, **kargs)
1295 def _og_search_video_url(self, html, name='video url
', secure=True, **kargs):
1296 regexes = self._og_regexes('video
') + self._og_regexes('video
:url
')
1298 regexes = self._og_regexes('video
:secure_url
') + regexes
1299 return self._html_search_regex(regexes, html, name, **kargs)
1301 def _og_search_url(self, html, **kargs):
1302 return self._og_search_property('url
', html, **kargs)
1304 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1305 name = variadic(name)
1306 if display_name is None:
1307 display_name = name[0]
1308 return self._html_search_regex(
1309 [self._meta_regex(n) for n in name],
1310 html, display_name, fatal=fatal, group='content
', **kwargs)
1312 def _dc_search_uploader(self, html):
1313 return self._html_search_meta('dc
.creator
', html, 'uploader
')
1315 def _rta_search(self, html):
1316 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1317 if re.search(r'(?ix
)<meta\s
+name
="rating"\s
+'
1318 r' content
="RTA-5042-1996-1400-1577-RTA"',
1323 def _media_rating_search(self, html):
1324 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1325 rating = self._html_search_meta('rating
', html)
1337 return RATING_TABLE.get(rating.lower())
1339 def _family_friendly_search(self, html):
1340 # See http://schema.org/VideoObject
1341 family_friendly = self._html_search_meta(
1342 'isFamilyFriendly
', html, default=None)
1344 if not family_friendly:
1353 return RATING_TABLE.get(family_friendly.lower())
1355 def _twitter_search_player(self, html):
1356 return self._html_search_meta('twitter
:player
', html,
1357 'twitter card player
')
1359 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
1360 json_ld_list = list(re.finditer(JSON_LD_RE, html))
1361 default = kwargs.get('default
', NO_DEFAULT)
1362 # JSON-LD may be malformed and thus `fatal` should be respected.
1363 # At the same time `default` may be passed that assumes `fatal=False`
1364 # for _search_regex. Let's simulate the same behavior here
as well
.
1365 fatal
= kwargs
.get('fatal', True) if default
is NO_DEFAULT
else False
1367 for mobj
in json_ld_list
:
1368 json_ld_item
= self
._parse
_json
(
1369 mobj
.group('json_ld'), video_id
, fatal
=fatal
)
1370 if not json_ld_item
:
1372 if isinstance(json_ld_item
, dict):
1373 json_ld
.append(json_ld_item
)
1374 elif isinstance(json_ld_item
, (list, tuple)):
1375 json_ld
.extend(json_ld_item
)
1377 json_ld
= self
._json
_ld
(json_ld
, video_id
, fatal
=fatal
, expected_type
=expected_type
)
1380 if default
is not NO_DEFAULT
:
1383 raise RegexNotFoundError('Unable to extract JSON-LD')
1385 self
.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1388 def _json_ld(self
, json_ld
, video_id
, fatal
=True, expected_type
=None):
1389 if isinstance(json_ld
, compat_str
):
1390 json_ld
= self
._parse
_json
(json_ld
, video_id
, fatal
=fatal
)
1394 if not isinstance(json_ld
, (list, tuple, dict)):
1396 if isinstance(json_ld
, dict):
1399 INTERACTION_TYPE_MAP
= {
1400 'CommentAction': 'comment',
1401 'AgreeAction': 'like',
1402 'DisagreeAction': 'dislike',
1403 'LikeAction': 'like',
1404 'DislikeAction': 'dislike',
1405 'ListenAction': 'view',
1406 'WatchAction': 'view',
1407 'ViewAction': 'view',
1410 def extract_interaction_type(e
):
1411 interaction_type
= e
.get('interactionType')
1412 if isinstance(interaction_type
, dict):
1413 interaction_type
= interaction_type
.get('@type')
1414 return str_or_none(interaction_type
)
1416 def extract_interaction_statistic(e
):
1417 interaction_statistic
= e
.get('interactionStatistic')
1418 if isinstance(interaction_statistic
, dict):
1419 interaction_statistic
= [interaction_statistic
]
1420 if not isinstance(interaction_statistic
, list):
1422 for is_e
in interaction_statistic
:
1423 if not isinstance(is_e
, dict):
1425 if is_e
.get('@type') != 'InteractionCounter':
1427 interaction_type
= extract_interaction_type(is_e
)
1428 if not interaction_type
:
1430 # For interaction count some sites provide string instead of
1431 # an integer (as per spec) with non digit characters (e.g. ",")
1432 # so extracting count with more relaxed str_to_int
1433 interaction_count
= str_to_int(is_e
.get('userInteractionCount'))
1434 if interaction_count
is None:
1436 count_kind
= INTERACTION_TYPE_MAP
.get(interaction_type
.split('/')[-1])
1439 count_key
= '%s_count' % count_kind
1440 if info
.get(count_key
) is not None:
1442 info
[count_key
] = interaction_count
1444 def extract_chapter_information(e
):
1446 'title': part
.get('name'),
1447 'start_time': part
.get('startOffset'),
1448 'end_time': part
.get('endOffset'),
1449 } for part
in e
.get('hasPart', []) if part
.get('@type') == 'Clip']
1450 for idx
, (last_c
, current_c
, next_c
) in enumerate(zip(
1451 [{'end_time': 0}
] + chapters
, chapters
, chapters
[1:])):
1452 current_c
['end_time'] = current_c
['end_time'] or next_c
['start_time']
1453 current_c
['start_time'] = current_c
['start_time'] or last_c
['end_time']
1454 if None in current_c
.values():
1455 self
.report_warning(f
'Chapter {idx} contains broken data. Not extracting chapters')
1458 chapters
[-1]['end_time'] = chapters
[-1]['end_time'] or info
['duration']
1459 info
['chapters'] = chapters
1461 def extract_video_object(e
):
1462 assert e
['@type'] == 'VideoObject'
1463 author
= e
.get('author')
1465 'url': url_or_none(e
.get('contentUrl')),
1466 'title': unescapeHTML(e
.get('name')),
1467 'description': unescapeHTML(e
.get('description')),
1468 'thumbnails': [{'url': url_or_none(url)}
1469 for url
in variadic(traverse_obj(e
, 'thumbnailUrl', 'thumbnailURL'))],
1470 'duration': parse_duration(e
.get('duration')),
1471 'timestamp': unified_timestamp(e
.get('uploadDate')),
1472 # author can be an instance of 'Organization' or 'Person' types.
1473 # both types can have 'name' property(inherited from 'Thing' type). [1]
1474 # however some websites are using 'Text' type instead.
1475 # 1. https://schema.org/VideoObject
1476 'uploader': author
.get('name') if isinstance(author
, dict) else author
if isinstance(author
, compat_str
) else None,
1477 'filesize': float_or_none(e
.get('contentSize')),
1478 'tbr': int_or_none(e
.get('bitrate')),
1479 'width': int_or_none(e
.get('width')),
1480 'height': int_or_none(e
.get('height')),
1481 'view_count': int_or_none(e
.get('interactionCount')),
1483 extract_interaction_statistic(e
)
1484 extract_chapter_information(e
)
1486 def traverse_json_ld(json_ld
, at_top_level
=True):
1488 if at_top_level
and '@context' not in e
:
1490 if at_top_level
and set(e
.keys()) == {'@context', '@graph'}
:
1491 traverse_json_ld(variadic(e
['@graph'], allowed_types
=(dict,)), at_top_level
=False)
1493 item_type
= e
.get('@type')
1494 if expected_type
is not None and expected_type
!= item_type
:
1496 rating
= traverse_obj(e
, ('aggregateRating', 'ratingValue'), expected_type
=float_or_none
)
1497 if rating
is not None:
1498 info
['average_rating'] = rating
1499 if item_type
in ('TVEpisode', 'Episode'):
1500 episode_name
= unescapeHTML(e
.get('name'))
1502 'episode': episode_name
,
1503 'episode_number': int_or_none(e
.get('episodeNumber')),
1504 'description': unescapeHTML(e
.get('description')),
1506 if not info
.get('title') and episode_name
:
1507 info
['title'] = episode_name
1508 part_of_season
= e
.get('partOfSeason')
1509 if isinstance(part_of_season
, dict) and part_of_season
.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1511 'season': unescapeHTML(part_of_season
.get('name')),
1512 'season_number': int_or_none(part_of_season
.get('seasonNumber')),
1514 part_of_series
= e
.get('partOfSeries') or e
.get('partOfTVSeries')
1515 if isinstance(part_of_series
, dict) and part_of_series
.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1516 info
['series'] = unescapeHTML(part_of_series
.get('name'))
1517 elif item_type
== 'Movie':
1519 'title': unescapeHTML(e
.get('name')),
1520 'description': unescapeHTML(e
.get('description')),
1521 'duration': parse_duration(e
.get('duration')),
1522 'timestamp': unified_timestamp(e
.get('dateCreated')),
1524 elif item_type
in ('Article', 'NewsArticle'):
1526 'timestamp': parse_iso8601(e
.get('datePublished')),
1527 'title': unescapeHTML(e
.get('headline')),
1528 'description': unescapeHTML(e
.get('articleBody') or e
.get('description')),
1530 elif item_type
== 'VideoObject':
1531 extract_video_object(e
)
1532 if expected_type
is None:
1536 video
= e
.get('video')
1537 if isinstance(video
, dict) and video
.get('@type') == 'VideoObject':
1538 extract_video_object(video
)
1539 if expected_type
is None:
1543 traverse_json_ld(json_ld
)
1545 return dict((k
, v
) for k
, v
in info
.items() if v
is not None)
1547 def _search_nextjs_data(self
, webpage
, video_id
, **kw
):
1548 return self
._parse
_json
(
1550 r
'(?s)<script[^>]+id=[\'"]__NEXT_DATA__[\'"][^
>]*>([^
<]+)</script
>',
1551 webpage, 'next
.js data
', **kw),
1554 def _search_nuxt_data(self, webpage, video_id, context_name='__NUXT__
'):
1555 ''' Parses Nuxt.js metadata. This works as long as the function __NUXT__ invokes is a pure function. '''
1556 # not all website do this, but it can be changed
1557 # https://stackoverflow.com/questions/67463109/how-to-change-or-hide-nuxt-and-nuxt-keyword-in-page-source
1558 rectx = re.escape(context_name)
1559 js, arg_keys, arg_vals = self._search_regex(
1560 (r'<script
>window\
.%s=\
(function\
((?P
<arg_keys
>.*?
)\
)\{return\s(?P<js>\{.*?\}
)\
}\
((?P
<arg_vals
>.+?
)\
)\
);?
</script
>' % rectx,
1561 r'%s\
(.*?\
(function\
((?P
<arg_keys
>.*?
)\
)\{return\s(?P<js>\{.*?\}
)\
}\
((?P
<arg_vals
>.*?
)\
)' % rectx),
1562 webpage, context_name, group=['js
', 'arg_keys
', 'arg_vals
'])
1564 args = dict(zip(arg_keys.split(','), arg_vals.split(',')))
1566 for key, val in args.items():
1567 if val in ('undefined
', 'void
0'):
1570 return self._parse_json(js_to_json(js, args), video_id)['data
'][0]
1573 def _hidden_inputs(html):
1574 html = re.sub(r'<!--(?
:(?
!<!--).)*-->', '', html)
1576 for input in re.findall(r'(?i
)(<input[^
>]+>)', html):
1577 attrs = extract_attributes(input)
1580 if attrs.get('type') not in ('hidden
', 'submit
'):
1582 name = attrs.get('name
') or attrs.get('id')
1583 value = attrs.get('value
')
1584 if name and value is not None:
1585 hidden_inputs[name] = value
1586 return hidden_inputs
1588 def _form_hidden_inputs(self, form_id, html):
1589 form = self._search_regex(
1590 r'(?
is)<form
[^
>]+?
id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1591 html, '%s form' % form_id, group='form')
1592 return self._hidden_inputs(form)
1595 regex = r' *((?P<reverse>\+)?(?P<field>[a-zA-Z0-9_]+)((?P<separator>[~:])(?P<limit>.*?))?)? *$'
1597 default = ('hidden', 'aud_or_vid', 'hasvid', 'ie_pref', 'lang', 'quality',
1598 'res', 'fps', 'hdr:12', 'codec:vp9.2', 'size', 'br', 'asr',
1599 'proto', 'ext', 'hasaud', 'source', 'id') # These must not be aliases
1600 ytdl_default = ('hasaud', 'lang', 'quality', 'tbr', 'filesize', 'vbr',
1601 'height', 'width', 'proto', 'vext', 'abr', 'aext',
1602 'fps', 'fs_approx', 'source', 'id')
1605 'vcodec': {'type': 'ordered', 'regex': True,
1606 'order': ['av0?1', 'vp0?9.2', 'vp0?9', '[hx]265|he?vc?', '[hx]264|avc', 'vp0?8', 'mp4v|h263', 'theora', '', None, 'none']},
1607 'acodec': {'type': 'ordered', 'regex': True,
1608 'order': ['[af]lac', 'wav|aiff', 'opus', 'vorbis', 'aac', 'mp?4a?', 'mp3', 'e-?a?c-?3', 'ac-?3', 'dts', '', None, 'none']},
1609 'hdr': {'type': 'ordered', 'regex': True, 'field': 'dynamic_range',
1610 'order': ['dv', '(hdr)?12', r'(hdr)?10\+', '(hdr)?10', 'hlg', '', 'sdr', None]},
1611 'proto': {'type': 'ordered', 'regex': True, 'field': 'protocol',
1612 'order': ['(ht|f)tps', '(ht|f)tp$', 'm3u8.*', '.*dash', 'websocket_frag', 'rtmpe?', '', 'mms|rtsp', 'ws|websocket', 'f4']},
1613 'vext': {'type': 'ordered', 'field': 'video_ext',
1614 'order': ('mp4', 'webm', 'flv', '', 'none'),
1615 'order_free': ('webm', 'mp4', 'flv', '', 'none')},
1616 'aext': {'type': 'ordered', 'field': 'audio_ext',
1617 'order': ('m4a', 'aac', 'mp3', 'ogg', 'opus', 'webm', '', 'none'),
1618 'order_free': ('opus', 'ogg', 'webm', 'm4a', 'mp3', 'aac', '', 'none')},
1619 'hidden': {'visible': False, 'forced': True, 'type': 'extractor', 'max': -1000},
1620 'aud_or_vid': {'visible': False, 'forced': True, 'type': 'multiple',
1621 'field': ('vcodec', 'acodec'),
1622 'function': lambda it: int(any(v != 'none' for v in it))},
1623 'ie_pref': {'priority': True, 'type': 'extractor'},
1624 'hasvid': {'priority': True, 'field': 'vcodec', 'type': 'boolean', 'not_in_list': ('none',)},
1625 'hasaud': {'field': 'acodec', 'type': 'boolean', 'not_in_list': ('none',)},
1626 'lang': {'convert': 'float', 'field': 'language_preference', 'default': -1},
1627 'quality': {'convert': 'float', 'default': -1},
1628 'filesize': {'convert': 'bytes'},
1629 'fs_approx': {'convert': 'bytes', 'field': 'filesize_approx'},
1630 'id': {'convert': 'string', 'field': 'format_id'},
1631 'height': {'convert': 'float_none'},
1632 'width': {'convert': 'float_none'},
1633 'fps': {'convert': 'float_none'},
1634 'tbr': {'convert': 'float_none'},
1635 'vbr': {'convert': 'float_none'},
1636 'abr': {'convert': 'float_none'},
1637 'asr': {'convert': 'float_none'},
1638 'source': {'convert': 'float', 'field': 'source_preference', 'default': -1},
1640 'codec': {'type': 'combined', 'field': ('vcodec', 'acodec')},
1641 'br': {'type': 'combined', 'field': ('tbr', 'vbr', 'abr'), 'same_limit': True},
1642 'size': {'type': 'combined', 'same_limit': True, 'field': ('filesize', 'fs_approx')},
1643 'ext': {'type': 'combined', 'field': ('vext', 'aext')},
1644 'res': {'type': 'multiple', 'field': ('height', 'width'),
1645 'function': lambda it: (lambda l: min(l) if l else 0)(tuple(filter(None, it)))},
1647 # For compatibility with youtube-dl
1648 'format_id': {'type': 'alias', 'field': 'id'},
1649 'preference': {'type': 'alias', 'field': 'ie_pref'},
1650 'language_preference': {'type': 'alias', 'field': 'lang'},
1653 'dimension': {'type': 'alias', 'field': 'res'},
1654 'resolution': {'type': 'alias', 'field': 'res'},
1655 'extension': {'type': 'alias', 'field': 'ext'},
1656 'bitrate': {'type': 'alias', 'field': 'br'},
1657 'total_bitrate': {'type': 'alias', 'field': 'tbr'},
1658 'video_bitrate': {'type': 'alias', 'field': 'vbr'},
1659 'audio_bitrate': {'type': 'alias', 'field': 'abr'},
1660 'framerate': {'type': 'alias', 'field': 'fps'},
1661 'protocol': {'type': 'alias', 'field': 'proto'},
1662 'source_preference': {'type': 'alias', 'field': 'source'},
1663 'filesize_approx': {'type': 'alias', 'field': 'fs_approx'},
1664 'filesize_estimate': {'type': 'alias', 'field': 'size'},
1665 'samplerate': {'type': 'alias', 'field': 'asr'},
1666 'video_ext': {'type': 'alias', 'field': 'vext'},
1667 'audio_ext': {'type': 'alias', 'field': 'aext'},
1668 'video_codec': {'type': 'alias', 'field': 'vcodec'},
1669 'audio_codec': {'type': 'alias', 'field': 'acodec'},
1670 'video': {'type': 'alias', 'field': 'hasvid'},
1671 'has_video': {'type': 'alias', 'field': 'hasvid'},
1672 'audio': {'type': 'alias', 'field': 'hasaud'},
1673 'has_audio': {'type': 'alias', 'field': 'hasaud'},
1674 'extractor': {'type': 'alias', 'field': 'ie_pref'},
1675 'extractor_preference': {'type': 'alias', 'field': 'ie_pref'},
1678 def __init__(self, ie, field_preference):
1680 self.ydl = ie._downloader
1681 self.evaluate_params(self.ydl.params, field_preference)
1682 if ie.get_param('verbose'):
1683 self.print_verbose_info(self.ydl.write_debug)
1685 def _get_field_setting(self, field, key):
1686 if field not in self.settings:
1687 if key in ('forced', 'priority'):
1689 self.ydl.deprecation_warning(
1690 f'Using arbitrary fields ({field}) for format sorting is deprecated '
1691 'and may be removed in a future version')
1692 self.settings[field] = {}
1693 propObj = self.settings[field]
1694 if key not in propObj:
1695 type = propObj.get('type')
1697 default = 'preference' if type == 'extractor' else (field,) if type in ('combined', 'multiple') else field
1698 elif key == 'convert':
1699 default = 'order' if type == 'ordered' else 'float_string' if field else 'ignore'
1701 default = {'type': 'field', 'visible': True, 'order': [], 'not_in_list': (None,)}.get(key, None)
1702 propObj[key] = default
1705 def _resolve_field_value(self, field, value, convertNone=False):
1710 value = value.lower()
1711 conversion = self._get_field_setting(field, 'convert')
1712 if conversion == 'ignore':
1714 if conversion == 'string':
1716 elif conversion == 'float_none':
1717 return float_or_none(value)
1718 elif conversion == 'bytes':
1719 return FileDownloader.parse_bytes(value)
1720 elif conversion == 'order':
1721 order_list = (self._use_free_order and self._get_field_setting(field, 'order_free')) or self._get_field_setting(field, 'order')
1722 use_regex = self._get_field_setting(field, 'regex')
1723 list_length = len(order_list)
1724 empty_pos = order_list.index('') if '' in order_list else list_length + 1
1725 if use_regex and value is not None:
1726 for i, regex in enumerate(order_list):
1727 if regex and re.match(regex, value):
1728 return list_length - i
1729 return list_length - empty_pos # not in list
1730 else: # not regex or value = None
1731 return list_length - (order_list.index(value) if value in order_list else empty_pos)
1733 if value.isnumeric():
1736 self.settings[field]['convert'] = 'string'
1739 def evaluate_params(self, params, sort_extractor):
1740 self._use_free_order = params.get('prefer_free_formats', False)
1741 self._sort_user = params.get('format_sort', [])
1742 self._sort_extractor = sort_extractor
1744 def add_item(field, reverse, closest, limit_text):
1745 field = field.lower()
1746 if field in self._order:
1748 self._order.append(field)
1749 limit = self._resolve_field_value(field, limit_text)
1752 'closest': False if limit is None else closest,
1753 'limit_text': limit_text,
1755 if field in self.settings:
1756 self.settings[field].update(data)
1758 self.settings[field] = data
1761 tuple(field for field in self.default if self._get_field_setting(field, 'forced'))
1762 + (tuple() if params.get('format_sort_force', False)
1763 else tuple(field for field in self.default if self._get_field_setting(field, 'priority')))
1764 + tuple(self._sort_user) + tuple(sort_extractor) + self.default)
1766 for item in sort_list:
1767 match = re.match(self.regex, item)
1769 raise ExtractorError('Invalid format sort string "%s" given by extractor' % item)
1770 field = match.group('field')
1773 if self._get_field_setting(field, 'type') == 'alias':
1774 alias, field = field, self._get_field_setting(field, 'field')
1775 if alias not in ('format_id', 'preference', 'language_preference'):
1776 self.ydl.deprecation_warning(
1777 f'Format sorting alias {alias} is deprecated '
1778 f'and may be removed in a future version. Please use {field} instead')
1779 reverse = match.group('reverse') is not None
1780 closest = match.group('separator') == '~'
1781 limit_text = match.group('limit')
1783 has_limit = limit_text is not None
1784 has_multiple_fields = self._get_field_setting(field, 'type') == 'combined'
1785 has_multiple_limits = has_limit and has_multiple_fields and not self._get_field_setting(field, 'same_limit')
1787 fields = self._get_field_setting(field, 'field') if has_multiple_fields else (field,)
1788 limits = limit_text.split(':') if has_multiple_limits else (limit_text,) if has_limit else tuple()
1789 limit_count = len(limits)
1790 for (i, f) in enumerate(fields):
1791 add_item(f, reverse, closest,
1792 limits[i] if i < limit_count
1793 else limits[0] if has_limit and not has_multiple_limits
1796 def print_verbose_info(self, write_debug):
1798 write_debug('Sort order given by user: %s' % ', '.join(self._sort_user))
1799 if self._sort_extractor:
1800 write_debug('Sort order given by extractor: %s' % ', '.join(self._sort_extractor))
1801 write_debug('Formats sorted by: %s' % ', '.join(['%s%s%s' % (
1802 '+' if self._get_field_setting(field, 'reverse') else '', field,
1803 '%s%s(%s)' % ('~' if self._get_field_setting(field, 'closest') else ':',
1804 self._get_field_setting(field, 'limit_text'),
1805 self._get_field_setting(field, 'limit'))
1806 if self._get_field_setting(field, 'limit_text') is not None else '')
1807 for field in self._order if self._get_field_setting(field, 'visible')]))
1809 def _calculate_field_preference_from_value(self, format, field, type, value):
1810 reverse = self._get_field_setting(field, 'reverse')
1811 closest = self._get_field_setting(field, 'closest')
1812 limit = self._get_field_setting(field, 'limit')
1814 if type == 'extractor':
1815 maximum = self._get_field_setting(field, 'max')
1816 if value is None or (maximum is not None and value >= maximum):
1818 elif type == 'boolean':
1819 in_list = self._get_field_setting(field, 'in_list')
1820 not_in_list = self._get_field_setting(field, 'not_in_list')
1821 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
1822 elif type == 'ordered':
1823 value = self._resolve_field_value(field, value, True)
1825 # try to convert to number
1826 val_num = float_or_none(value, default=self._get_field_setting(field, 'default'))
1827 is_num = self._get_field_setting(field, 'convert') != 'string' and val_num is not None
1831 return ((-10, 0) if value is None
1832 else (1, value, 0) if not is_num # if a field has mixed strings and numbers, strings are sorted higher
1833 else (0, -abs(value - limit), value - limit if reverse else limit - value) if closest
1834 else (0, value, 0) if not reverse and (limit is None or value <= limit)
1835 else (0, -value, 0) if limit is None or (reverse and value == limit) or value > limit
1836 else (-1, value, 0))
1838 def _calculate_field_preference(self, format, field):
1839 type = self._get_field_setting(field, 'type') # extractor, boolean, ordered, field, multiple
1840 get_value = lambda f: format.get(self._get_field_setting(f, 'field'))
1841 if type == 'multiple':
1842 type = 'field' # Only 'field' is allowed in multiple for now
1843 actual_fields = self._get_field_setting(field, 'field')
1845 value = self._get_field_setting(field, 'function')(get_value(f) for f in actual_fields)
1847 value = get_value(field)
1848 return self._calculate_field_preference_from_value(format, field, type, value)
1850 def calculate_preference(self, format):
1851 # Determine missing protocol
1852 if not format.get('protocol'):
1853 format['protocol'] = determine_protocol(format)
1855 # Determine missing ext
1856 if not format.get('ext') and 'url' in format:
1857 format['ext'] = determine_ext(format['url'])
1858 if format.get('vcodec') == 'none':
1859 format['audio_ext'] = format['ext'] if format.get('acodec') != 'none' else 'none'
1860 format['video_ext'] = 'none'
1862 format['video_ext'] = format['ext']
1863 format['audio_ext'] = 'none'
1864 # if format.get('preference') is None and format.get('ext') in ('f4f', 'f4m'): # Not supported?
1865 # format['preference'] = -1000
1867 # Determine missing bitrates
1868 if format.get('tbr') is None:
1869 if format.get('vbr') is not None and format.get('abr') is not None:
1870 format['tbr'] = format.get('vbr', 0) + format.get('abr', 0)
1872 if format.get('vcodec') != 'none' and format.get('vbr') is None:
1873 format['vbr'] = format.get('tbr') - format.get('abr', 0)
1874 if format.get('acodec') != 'none' and format.get('abr') is None:
1875 format['abr'] = format.get('tbr') - format.get('vbr', 0)
1877 return tuple(self._calculate_field_preference(format, field) for field in self._order)
1879 def _sort_formats(self, formats, field_preference=[]):
1882 format_sort = self.FormatSort(self, field_preference)
1883 formats.sort(key=lambda f: format_sort.calculate_preference(f))
1885 def _check_formats(self, formats, video_id):
1887 formats[:] = filter(
1888 lambda f: self._is_valid_url(
1890 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1894 def _remove_duplicate_formats(formats):
1898 if f['url'] not in format_urls:
1899 format_urls.add(f['url'])
1900 unique_formats.append(f)
1901 formats[:] = unique_formats
1903 def _is_valid_url(self, url, video_id, item='video', headers={}):
1904 url = self._proto_relative_url(url, scheme='http:')
1905 # For now assume non HTTP(S) URLs always valid
1906 if not (url.startswith('http://') or url.startswith('https://')):
1909 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1911 except ExtractorError as e:
1913 '%s: %s URL is invalid, skipping: %s'
1914 % (video_id, item, error_to_compat_str(e.cause)))
1917 def http_scheme(self):
1918 """ Either "http
:" or "https
:", depending on the user's preferences """
1921 if self.get_param('prefer_insecure', False)
1924 def _proto_relative_url(self, url, scheme=None):
1927 if url.startswith('//'):
1929 scheme = self.http_scheme()
1934 def _sleep(self, timeout, video_id, msg_template=None):
1935 if msg_template is None:
1936 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1937 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1941 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1942 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1943 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
1944 manifest = self._download_xml(
1945 manifest_url, video_id, 'Downloading f4m manifest',
1946 'Unable to download f4m manifest',
1947 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1948 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
1949 transform_source=transform_source,
1950 fatal=fatal, data=data, headers=headers, query=query)
1952 if manifest is False:
1955 return self._parse_f4m_formats(
1956 manifest, manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1957 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1959 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1960 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1961 fatal=True, m3u8_id=None):
1962 if not isinstance(manifest, compat_etree_Element) and not fatal:
1965 # currently yt-dlp cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1966 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1967 if akamai_pv is not None and ';' in akamai_pv.text:
1968 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1969 if playerVerificationChallenge.strip() != '':
1973 manifest_version = '1.0'
1974 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1976 manifest_version = '2.0'
1977 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1978 # Remove unsupported DRM protected media from final formats
1979 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
1980 media_nodes = remove_encrypted_media(media_nodes)
1984 manifest_base_url = get_base_url(manifest)
1986 bootstrap_info = xpath_element(
1987 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1988 'bootstrap info', default=None)
1991 mime_type = xpath_text(
1992 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1993 'base URL', default=None)
1994 if mime_type and mime_type.startswith('audio/'):
1997 for i, media_el in enumerate(media_nodes):
1998 tbr = int_or_none(media_el.attrib.get('bitrate'))
1999 width = int_or_none(media_el.attrib.get('width'))
2000 height = int_or_none(media_el.attrib.get('height'))
2001 format_id = join_nonempty(f4m_id, tbr or i)
2002 # If <bootstrapInfo> is present, the specified f4m is a
2003 # stream-level manifest, and only set-level manifests may refer to
2004 # external resources. See section 11.4 and section 4 of F4M spec
2005 if bootstrap_info is None:
2007 # @href is introduced in 2.0, see section 11.6 of F4M spec
2008 if manifest_version == '2.0':
2009 media_url = media_el.attrib.get('href')
2010 if media_url is None:
2011 media_url = media_el.attrib.get('url')
2015 media_url if media_url.startswith('http://') or media_url.startswith('https://')
2016 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
2017 # If media_url is itself a f4m manifest do the recursive extraction
2018 # since bitrates in parent manifest (this one) and media_url manifest
2019 # may differ leading to inability to resolve the format by requested
2020 # bitrate in f4m downloader
2021 ext = determine_ext(manifest_url)
2023 f4m_formats = self._extract_f4m_formats(
2024 manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
2025 transform_source=transform_source, fatal=fatal)
2026 # Sometimes stream-level manifest contains single media entry that
2027 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
2028 # At the same time parent's media entry in set-level manifest may
2029 # contain it. We will copy it from parent in such cases.
2030 if len(f4m_formats) == 1:
2033 'tbr': f.get('tbr') or tbr,
2034 'width': f.get('width') or width,
2035 'height': f.get('height') or height,
2036 'format_id': f.get('format_id') if not tbr else format_id,
2039 formats.extend(f4m_formats)
2042 formats.extend(self._extract_m3u8_formats(
2043 manifest_url, video_id, 'mp4', preference=preference,
2044 quality=quality, m3u8_id=m3u8_id, fatal=fatal))
2047 'format_id': format_id,
2048 'url': manifest_url,
2049 'manifest_url': manifest_url,
2050 'ext': 'flv' if bootstrap_info is not None else None,
2056 'preference': preference,
2061 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, quality=None, m3u8_id=None):
2063 'format_id': join_nonempty(m3u8_id, 'meta'),
2067 'preference': preference - 100 if preference else -100,
2069 'resolution': 'multiple',
2070 'format_note': 'Quality selection URL',
2073 def _report_ignoring_subs(self, name):
2074 self.report_warning(bug_reports_message(
2075 f'Ignoring subtitle tracks found in the {name} manifest; '
2076 'if any subtitle tracks are missing,'
2079 def _extract_m3u8_formats(self, *args, **kwargs):
2080 fmts, subs = self._extract_m3u8_formats_and_subtitles(*args, **kwargs)
2082 self._report_ignoring_subs('HLS')
2085 def _extract_m3u8_formats_and_subtitles(
2086 self, m3u8_url, video_id, ext=None, entry_protocol='m3u8_native',
2087 preference=None, quality=None, m3u8_id=None, note=None,
2088 errnote=None, fatal=True, live=False, data=None, headers={},
2091 res = self._download_webpage_handle(
2093 note='Downloading m3u8 information' if note is None else note,
2094 errnote='Failed to download m3u8 information' if errnote is None else errnote,
2095 fatal=fatal, data=data, headers=headers, query=query)
2100 m3u8_doc, urlh = res
2101 m3u8_url = urlh.geturl()
2103 return self._parse_m3u8_formats_and_subtitles(
2104 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
2105 preference=preference, quality=quality, m3u8_id=m3u8_id,
2106 note=note, errnote=errnote, fatal=fatal, live=live, data=data,
2107 headers=headers, query=query, video_id=video_id)
2109 def _parse_m3u8_formats_and_subtitles(
2110 self, m3u8_doc, m3u8_url=None, ext=None, entry_protocol='m3u8_native',
2111 preference=None, quality=None, m3u8_id=None, live=False, note=None,
2112 errnote=None, fatal=True, data=None, headers={}, query={},
2114 formats, subtitles = [], {}
2116 has_drm = re.search('|'.join([
2117 r'#EXT-X-FAXS-CM:', # Adobe Flash Access
2118 r'#EXT-X-(?:SESSION-)?KEY:.*?URI="skd
://', # Apple FairPlay
2121 def format_url(url):
2122 return url if re.match(r'^https?
://', url) else compat_urlparse.urljoin(m3u8_url, url)
2124 if self.get_param('hls_split_discontinuity
', False):
2125 def _extract_m3u8_playlist_indices(manifest_url=None, m3u8_doc=None):
2127 if not manifest_url:
2129 m3u8_doc = self._download_webpage(
2130 manifest_url, video_id, fatal=fatal, data=data, headers=headers,
2131 note=False, errnote='Failed to download m3u8 playlist information
')
2132 if m3u8_doc is False:
2134 return range(1 + sum(line.startswith('#EXT-X-DISCONTINUITY') for line in m3u8_doc.splitlines()))
2137 def _extract_m3u8_playlist_indices(*args
, **kwargs
):
2141 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
2142 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
2143 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
2145 # We should try extracting formats only from master playlists [1, 4.3.4],
2146 # i.e. playlists that describe available qualities. On the other hand
2147 # media playlists [1, 4.3.3] should be returned as is since they contain
2148 # just the media without qualities renditions.
2149 # Fortunately, master playlist can be easily distinguished from media
2150 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
2151 # master playlist tags MUST NOT appear in a media playlist and vice versa.
2152 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
2153 # media playlist and MUST NOT appear in master playlist thus we can
2154 # clearly detect media playlist with this criterion.
2156 if '#EXT-X-TARGETDURATION' in m3u8_doc
: # media playlist, return as is
2158 'format_id': join_nonempty(m3u8_id
, idx
),
2159 'format_index': idx
,
2160 'url': m3u8_url
or encode_data_uri(m3u8_doc
.encode('utf-8'), 'application/x-mpegurl'),
2162 'protocol': entry_protocol
,
2163 'preference': preference
,
2166 } for idx
in _extract_m3u8_playlist_indices(m3u8_doc
=m3u8_doc
)]
2168 return formats
, subtitles
2171 last_stream_inf
= {}
2173 def extract_media(x_media_line
):
2174 media
= parse_m3u8_attributes(x_media_line
)
2175 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
2176 media_type
, group_id
, name
= media
.get('TYPE'), media
.get('GROUP-ID'), media
.get('NAME')
2177 if not (media_type
and group_id
and name
):
2179 groups
.setdefault(group_id
, []).append(media
)
2180 # <https://tools.ietf.org/html/rfc8216#section-4.3.4.1>
2181 if media_type
== 'SUBTITLES':
2182 # According to RFC 8216 §4.3.4.2.1, URI is REQUIRED in the
2183 # EXT-X-MEDIA tag if the media type is SUBTITLES.
2184 # However, lack of URI has been spotted in the wild.
2185 # e.g. NebulaIE; see https://github.com/yt-dlp/yt-dlp/issues/339
2186 if not media
.get('URI'):
2188 url
= format_url(media
['URI'])
2191 'ext': determine_ext(url
),
2193 if sub_info
['ext'] == 'm3u8':
2194 # Per RFC 8216 §3.1, the only possible subtitle format m3u8
2195 # files may contain is WebVTT:
2196 # <https://tools.ietf.org/html/rfc8216#section-3.1>
2197 sub_info
['ext'] = 'vtt'
2198 sub_info
['protocol'] = 'm3u8_native'
2199 lang
= media
.get('LANGUAGE') or 'und'
2200 subtitles
.setdefault(lang
, []).append(sub_info
)
2201 if media_type
not in ('VIDEO', 'AUDIO'):
2203 media_url
= media
.get('URI')
2205 manifest_url
= format_url(media_url
)
2207 'format_id': join_nonempty(m3u8_id
, group_id
, name
, idx
),
2208 'format_note': name
,
2209 'format_index': idx
,
2210 'url': manifest_url
,
2211 'manifest_url': m3u8_url
,
2212 'language': media
.get('LANGUAGE'),
2214 'protocol': entry_protocol
,
2215 'preference': preference
,
2217 'vcodec': 'none' if media_type
== 'AUDIO' else None,
2218 } for idx
in _extract_m3u8_playlist_indices(manifest_url
))
2220 def build_stream_name():
2221 # Despite specification does not mention NAME attribute for
2222 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
2223 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
2224 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
2225 stream_name
= last_stream_inf
.get('NAME')
2228 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
2229 # from corresponding rendition group
2230 stream_group_id
= last_stream_inf
.get('VIDEO')
2231 if not stream_group_id
:
2233 stream_group
= groups
.get(stream_group_id
)
2234 if not stream_group
:
2235 return stream_group_id
2236 rendition
= stream_group
[0]
2237 return rendition
.get('NAME') or stream_group_id
2239 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
2240 # chance to detect video only formats when EXT-X-STREAM-INF tags
2241 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
2242 for line
in m3u8_doc
.splitlines():
2243 if line
.startswith('#EXT-X-MEDIA:'):
2246 for line
in m3u8_doc
.splitlines():
2247 if line
.startswith('#EXT-X-STREAM-INF:'):
2248 last_stream_inf
= parse_m3u8_attributes(line
)
2249 elif line
.startswith('#') or not line
.strip():
2252 tbr
= float_or_none(
2253 last_stream_inf
.get('AVERAGE-BANDWIDTH')
2254 or last_stream_inf
.get('BANDWIDTH'), scale
=1000)
2255 manifest_url
= format_url(line
.strip())
2257 for idx
in _extract_m3u8_playlist_indices(manifest_url
):
2258 format_id
= [m3u8_id
, None, idx
]
2259 # Bandwidth of live streams may differ over time thus making
2260 # format_id unpredictable. So it's better to keep provided
2263 stream_name
= build_stream_name()
2264 format_id
[1] = stream_name
or '%d' % (tbr
or len(formats
))
2266 'format_id': join_nonempty(*format_id
),
2267 'format_index': idx
,
2268 'url': manifest_url
,
2269 'manifest_url': m3u8_url
,
2272 'fps': float_or_none(last_stream_inf
.get('FRAME-RATE')),
2273 'protocol': entry_protocol
,
2274 'preference': preference
,
2277 resolution
= last_stream_inf
.get('RESOLUTION')
2279 mobj
= re
.search(r
'(?P<width>\d+)[xX](?P<height>\d+)', resolution
)
2281 f
['width'] = int(mobj
.group('width'))
2282 f
['height'] = int(mobj
.group('height'))
2283 # Unified Streaming Platform
2285 r
'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f
['url'])
2287 abr
, vbr
= mobj
.groups()
2288 abr
, vbr
= float_or_none(abr
, 1000), float_or_none(vbr
, 1000)
2293 codecs
= parse_codecs(last_stream_inf
.get('CODECS'))
2295 audio_group_id
= last_stream_inf
.get('AUDIO')
2296 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
2297 # references a rendition group MUST have a CODECS attribute.
2298 # However, this is not always respected, for example, [2]
2299 # contains EXT-X-STREAM-INF tag which references AUDIO
2300 # rendition group but does not have CODECS and despite
2301 # referencing an audio group it represents a complete
2302 # (with audio and video) format. So, for such cases we will
2303 # ignore references to rendition groups and treat them
2304 # as complete formats.
2305 if audio_group_id
and codecs
and f
.get('vcodec') != 'none':
2306 audio_group
= groups
.get(audio_group_id
)
2307 if audio_group
and audio_group
[0].get('URI'):
2308 # TODO: update acodec for audio only formats with
2310 f
['acodec'] = 'none'
2311 if not f
.get('ext'):
2312 f
['ext'] = 'm4a' if f
.get('vcodec') == 'none' else 'mp4'
2316 progressive_uri
= last_stream_inf
.get('PROGRESSIVE-URI')
2319 del http_f
['manifest_url']
2321 'format_id': f
['format_id'].replace('hls-', 'http-'),
2323 'url': progressive_uri
,
2325 formats
.append(http_f
)
2327 last_stream_inf
= {}
2328 return formats
, subtitles
2330 def _extract_m3u8_vod_duration(
2331 self
, m3u8_vod_url
, video_id
, note
=None, errnote
=None, data
=None, headers
={}, query={}
):
2333 m3u8_vod
= self
._download
_webpage
(
2334 m3u8_vod_url
, video_id
,
2335 note
='Downloading m3u8 VOD manifest' if note
is None else note
,
2336 errnote
='Failed to download VOD manifest' if errnote
is None else errnote
,
2337 fatal
=False, data
=data
, headers
=headers
, query
=query
)
2339 return self
._parse
_m
3u8_vod
_duration
(m3u8_vod
or '', video_id
)
2341 def _parse_m3u8_vod_duration(self
, m3u8_vod
, video_id
):
2342 if '#EXT-X-PLAYLIST-TYPE:VOD' not in m3u8_vod
:
2346 float(line
[len('#EXTINF:'):].split(',')[0])
2347 for line
in m3u8_vod
.splitlines() if line
.startswith('#EXTINF:'))) or None
2350 def _xpath_ns(path
, namespace
=None):
2354 for c
in path
.split('/'):
2355 if not c
or c
== '.':
2358 out
.append('{%s}%s' % (namespace
, c
))
2359 return '/'.join(out
)
2361 def _extract_smil_formats_and_subtitles(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None, transform_source
=None):
2362 smil
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
, transform_source
=transform_source
)
2368 namespace
= self
._parse
_smil
_namespace
(smil
)
2370 fmts
= self
._parse
_smil
_formats
(
2371 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2372 subs
= self
._parse
_smil
_subtitles
(
2373 smil
, namespace
=namespace
)
2377 def _extract_smil_formats(self
, *args
, **kwargs
):
2378 fmts
, subs
= self
._extract
_smil
_formats
_and
_subtitles
(*args
, **kwargs
)
2380 self
._report
_ignoring
_subs
('SMIL')
2383 def _extract_smil_info(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None):
2384 smil
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
)
2387 return self
._parse
_smil
(smil
, smil_url
, video_id
, f4m_params
=f4m_params
)
2389 def _download_smil(self
, smil_url
, video_id
, fatal
=True, transform_source
=None):
2390 return self
._download
_xml
(
2391 smil_url
, video_id
, 'Downloading SMIL file',
2392 'Unable to download SMIL file', fatal
=fatal
, transform_source
=transform_source
)
2394 def _parse_smil(self
, smil
, smil_url
, video_id
, f4m_params
=None):
2395 namespace
= self
._parse
_smil
_namespace
(smil
)
2397 formats
= self
._parse
_smil
_formats
(
2398 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2399 subtitles
= self
._parse
_smil
_subtitles
(smil
, namespace
=namespace
)
2401 video_id
= os
.path
.splitext(url_basename(smil_url
))[0]
2405 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2406 name
= meta
.attrib
.get('name')
2407 content
= meta
.attrib
.get('content')
2408 if not name
or not content
:
2410 if not title
and name
== 'title':
2412 elif not description
and name
in ('description', 'abstract'):
2413 description
= content
2414 elif not upload_date
and name
== 'date':
2415 upload_date
= unified_strdate(content
)
2418 'id': image
.get('type'),
2419 'url': image
.get('src'),
2420 'width': int_or_none(image
.get('width')),
2421 'height': int_or_none(image
.get('height')),
2422 } for image
in smil
.findall(self
._xpath
_ns
('.//image', namespace
)) if image
.get('src')]
2426 'title': title
or video_id
,
2427 'description': description
,
2428 'upload_date': upload_date
,
2429 'thumbnails': thumbnails
,
2431 'subtitles': subtitles
,
2434 def _parse_smil_namespace(self
, smil
):
2435 return self
._search
_regex
(
2436 r
'(?i)^{([^}]+)?}smil$', smil
.tag
, 'namespace', default
=None)
2438 def _parse_smil_formats(self
, smil
, smil_url
, video_id
, namespace
=None, f4m_params
=None, transform_rtmp_url
=None):
2440 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2441 b
= meta
.get('base') or meta
.get('httpBase')
2453 media
= smil
.findall(self
._xpath
_ns
('.//video', namespace
)) + smil
.findall(self
._xpath
_ns
('.//audio', namespace
))
2454 for medium
in media
:
2455 src
= medium
.get('src')
2456 if not src
or src
in srcs
:
2460 bitrate
= float_or_none(medium
.get('system-bitrate') or medium
.get('systemBitrate'), 1000)
2461 filesize
= int_or_none(medium
.get('size') or medium
.get('fileSize'))
2462 width
= int_or_none(medium
.get('width'))
2463 height
= int_or_none(medium
.get('height'))
2464 proto
= medium
.get('proto')
2465 ext
= medium
.get('ext')
2466 src_ext
= determine_ext(src
)
2467 streamer
= medium
.get('streamer') or base
2469 if proto
== 'rtmp' or streamer
.startswith('rtmp'):
2475 'format_id': 'rtmp-%d' % (rtmp_count
if bitrate
is None else bitrate
),
2477 'filesize': filesize
,
2481 if transform_rtmp_url
:
2482 streamer
, src
= transform_rtmp_url(streamer
, src
)
2483 formats
[-1].update({
2489 src_url
= src
if src
.startswith('http') else compat_urlparse
.urljoin(base
, src
)
2490 src_url
= src_url
.strip()
2492 if proto
== 'm3u8' or src_ext
== 'm3u8':
2493 m3u8_formats
= self
._extract
_m
3u8_formats
(
2494 src_url
, video_id
, ext
or 'mp4', m3u8_id
='hls', fatal
=False)
2495 if len(m3u8_formats
) == 1:
2497 m3u8_formats
[0].update({
2498 'format_id': 'hls-%d' % (m3u8_count
if bitrate
is None else bitrate
),
2503 formats
.extend(m3u8_formats
)
2504 elif src_ext
== 'f4m':
2509 'plugin': 'flowplayer-3.2.0.1',
2511 f4m_url
+= '&' if '?' in f4m_url
else '?'
2512 f4m_url
+= compat_urllib_parse_urlencode(f4m_params
)
2513 formats
.extend(self
._extract
_f
4m
_formats
(f4m_url
, video_id
, f4m_id
='hds', fatal
=False))
2514 elif src_ext
== 'mpd':
2515 formats
.extend(self
._extract
_mpd
_formats
(
2516 src_url
, video_id
, mpd_id
='dash', fatal
=False))
2517 elif re
.search(r
'\.ism/[Mm]anifest', src_url
):
2518 formats
.extend(self
._extract
_ism
_formats
(
2519 src_url
, video_id
, ism_id
='mss', fatal
=False))
2520 elif src_url
.startswith('http') and self
._is
_valid
_url
(src
, video_id
):
2524 'ext': ext
or src_ext
or 'flv',
2525 'format_id': 'http-%d' % (bitrate
or http_count
),
2527 'filesize': filesize
,
2532 for medium
in smil
.findall(self
._xpath
_ns
('.//imagestream', namespace
)):
2533 src
= medium
.get('src')
2534 if not src
or src
in srcs
:
2540 'format_id': 'imagestream-%d' % (imgs_count
),
2542 'ext': mimetype2ext(medium
.get('type')),
2545 'width': int_or_none(medium
.get('width')),
2546 'height': int_or_none(medium
.get('height')),
2547 'format_note': 'SMIL storyboards',
2552 def _parse_smil_subtitles(self
, smil
, namespace
=None, subtitles_lang
='en'):
2555 for num
, textstream
in enumerate(smil
.findall(self
._xpath
_ns
('.//textstream', namespace
))):
2556 src
= textstream
.get('src')
2557 if not src
or src
in urls
:
2560 ext
= textstream
.get('ext') or mimetype2ext(textstream
.get('type')) or determine_ext(src
)
2561 lang
= textstream
.get('systemLanguage') or textstream
.get('systemLanguageName') or textstream
.get('lang') or subtitles_lang
2562 subtitles
.setdefault(lang
, []).append({
2568 def _extract_xspf_playlist(self
, xspf_url
, playlist_id
, fatal
=True):
2569 xspf
= self
._download
_xml
(
2570 xspf_url
, playlist_id
, 'Downloading xpsf playlist',
2571 'Unable to download xspf manifest', fatal
=fatal
)
2574 return self
._parse
_xspf
(
2575 xspf
, playlist_id
, xspf_url
=xspf_url
,
2576 xspf_base_url
=base_url(xspf_url
))
2578 def _parse_xspf(self
, xspf_doc
, playlist_id
, xspf_url
=None, xspf_base_url
=None):
2580 'xspf': 'http://xspf.org/ns/0/',
2581 's1': 'http://static.streamone.nl/player/ns/0',
2585 for track
in xspf_doc
.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP
)):
2587 track
, xpath_with_ns('./xspf:title', NS_MAP
), 'title', default
=playlist_id
)
2588 description
= xpath_text(
2589 track
, xpath_with_ns('./xspf:annotation', NS_MAP
), 'description')
2590 thumbnail
= xpath_text(
2591 track
, xpath_with_ns('./xspf:image', NS_MAP
), 'thumbnail')
2592 duration
= float_or_none(
2593 xpath_text(track
, xpath_with_ns('./xspf:duration', NS_MAP
), 'duration'), 1000)
2596 for location
in track
.findall(xpath_with_ns('./xspf:location', NS_MAP
)):
2597 format_url
= urljoin(xspf_base_url
, location
.text
)
2602 'manifest_url': xspf_url
,
2603 'format_id': location
.get(xpath_with_ns('s1:label', NS_MAP
)),
2604 'width': int_or_none(location
.get(xpath_with_ns('s1:width', NS_MAP
))),
2605 'height': int_or_none(location
.get(xpath_with_ns('s1:height', NS_MAP
))),
2607 self
._sort
_formats
(formats
)
2612 'description': description
,
2613 'thumbnail': thumbnail
,
2614 'duration': duration
,
2619 def _extract_mpd_formats(self
, *args
, **kwargs
):
2620 fmts
, subs
= self
._extract
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2622 self
._report
_ignoring
_subs
('DASH')
2625 def _extract_mpd_formats_and_subtitles(
2626 self
, mpd_url
, video_id
, mpd_id
=None, note
=None, errnote
=None,
2627 fatal
=True, data
=None, headers
={}, query={}
):
2628 res
= self
._download
_xml
_handle
(
2630 note
='Downloading MPD manifest' if note
is None else note
,
2631 errnote
='Failed to download MPD manifest' if errnote
is None else errnote
,
2632 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
2638 mpd_base_url
= base_url(urlh
.geturl())
2640 return self
._parse
_mpd
_formats
_and
_subtitles
(
2641 mpd_doc
, mpd_id
, mpd_base_url
, mpd_url
)
2643 def _parse_mpd_formats(self
, *args
, **kwargs
):
2644 fmts
, subs
= self
._parse
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2646 self
._report
_ignoring
_subs
('DASH')
2649 def _parse_mpd_formats_and_subtitles(
2650 self
, mpd_doc
, mpd_id
=None, mpd_base_url
='', mpd_url
=None):
2652 Parse formats from MPD manifest.
2654 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
2655 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
2656 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
2658 if not self
.get_param('dynamic_mpd', True):
2659 if mpd_doc
.get('type') == 'dynamic':
2662 namespace
= self
._search
_regex
(r
'(?i)^{([^}]+)?}MPD$', mpd_doc
.tag
, 'namespace', default
=None)
2665 return self
._xpath
_ns
(path
, namespace
)
2667 def is_drm_protected(element
):
2668 return element
.find(_add_ns('ContentProtection')) is not None
2670 def extract_multisegment_info(element
, ms_parent_info
):
2671 ms_info
= ms_parent_info
.copy()
2673 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2674 # common attributes and elements. We will only extract relevant
2676 def extract_common(source
):
2677 segment_timeline
= source
.find(_add_ns('SegmentTimeline'))
2678 if segment_timeline
is not None:
2679 s_e
= segment_timeline
.findall(_add_ns('S'))
2681 ms_info
['total_number'] = 0
2684 r
= int(s
.get('r', 0))
2685 ms_info
['total_number'] += 1 + r
2686 ms_info
['s'].append({
2687 't': int(s
.get('t', 0)),
2688 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2689 'd': int(s
.attrib
['d']),
2692 start_number
= source
.get('startNumber')
2694 ms_info
['start_number'] = int(start_number
)
2695 timescale
= source
.get('timescale')
2697 ms_info
['timescale'] = int(timescale
)
2698 segment_duration
= source
.get('duration')
2699 if segment_duration
:
2700 ms_info
['segment_duration'] = float(segment_duration
)
2702 def extract_Initialization(source
):
2703 initialization
= source
.find(_add_ns('Initialization'))
2704 if initialization
is not None:
2705 ms_info
['initialization_url'] = initialization
.attrib
['sourceURL']
2707 segment_list
= element
.find(_add_ns('SegmentList'))
2708 if segment_list
is not None:
2709 extract_common(segment_list
)
2710 extract_Initialization(segment_list
)
2711 segment_urls_e
= segment_list
.findall(_add_ns('SegmentURL'))
2713 ms_info
['segment_urls'] = [segment
.attrib
['media'] for segment
in segment_urls_e
]
2715 segment_template
= element
.find(_add_ns('SegmentTemplate'))
2716 if segment_template
is not None:
2717 extract_common(segment_template
)
2718 media
= segment_template
.get('media')
2720 ms_info
['media'] = media
2721 initialization
= segment_template
.get('initialization')
2723 ms_info
['initialization'] = initialization
2725 extract_Initialization(segment_template
)
2728 mpd_duration
= parse_duration(mpd_doc
.get('mediaPresentationDuration'))
2729 formats
, subtitles
= [], {}
2730 stream_numbers
= collections
.defaultdict(int)
2731 for period
in mpd_doc
.findall(_add_ns('Period')):
2732 period_duration
= parse_duration(period
.get('duration')) or mpd_duration
2733 period_ms_info
= extract_multisegment_info(period
, {
2737 for adaptation_set
in period
.findall(_add_ns('AdaptationSet')):
2738 adaption_set_ms_info
= extract_multisegment_info(adaptation_set
, period_ms_info
)
2739 for representation
in adaptation_set
.findall(_add_ns('Representation')):
2740 representation_attrib
= adaptation_set
.attrib
.copy()
2741 representation_attrib
.update(representation
.attrib
)
2742 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2743 mime_type
= representation_attrib
['mimeType']
2744 content_type
= representation_attrib
.get('contentType', mime_type
.split('/')[0])
2746 codecs
= parse_codecs(representation_attrib
.get('codecs', ''))
2747 if content_type
not in ('video', 'audio', 'text'):
2748 if mime_type
== 'image/jpeg':
2749 content_type
= mime_type
2750 elif codecs
['vcodec'] != 'none':
2751 content_type
= 'video'
2752 elif codecs
['acodec'] != 'none':
2753 content_type
= 'audio'
2754 elif codecs
.get('tcodec', 'none') != 'none':
2755 content_type
= 'text'
2756 elif mimetype2ext(mime_type
) in ('tt', 'dfxp', 'ttml', 'xml', 'json'):
2757 content_type
= 'text'
2759 self
.report_warning('Unknown MIME type %s in DASH manifest' % mime_type
)
2763 for element
in (representation
, adaptation_set
, period
, mpd_doc
):
2764 base_url_e
= element
.find(_add_ns('BaseURL'))
2765 if base_url_e
is not None:
2766 base_url
= base_url_e
.text
+ base_url
2767 if re
.match(r
'^https?://', base_url
):
2769 if mpd_base_url
and base_url
.startswith('/'):
2770 base_url
= compat_urlparse
.urljoin(mpd_base_url
, base_url
)
2771 elif mpd_base_url
and not re
.match(r
'^https?://', base_url
):
2772 if not mpd_base_url
.endswith('/'):
2774 base_url
= mpd_base_url
+ base_url
2775 representation_id
= representation_attrib
.get('id')
2776 lang
= representation_attrib
.get('lang')
2777 url_el
= representation
.find(_add_ns('BaseURL'))
2778 filesize
= int_or_none(url_el
.attrib
.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el
is not None else None)
2779 bandwidth
= int_or_none(representation_attrib
.get('bandwidth'))
2780 if representation_id
is not None:
2781 format_id
= representation_id
2783 format_id
= content_type
2785 format_id
= mpd_id
+ '-' + format_id
2786 if content_type
in ('video', 'audio'):
2788 'format_id': format_id
,
2789 'manifest_url': mpd_url
,
2790 'ext': mimetype2ext(mime_type
),
2791 'width': int_or_none(representation_attrib
.get('width')),
2792 'height': int_or_none(representation_attrib
.get('height')),
2793 'tbr': float_or_none(bandwidth
, 1000),
2794 'asr': int_or_none(representation_attrib
.get('audioSamplingRate')),
2795 'fps': int_or_none(representation_attrib
.get('frameRate')),
2796 'language': lang
if lang
not in ('mul', 'und', 'zxx', 'mis') else None,
2797 'format_note': 'DASH %s' % content_type
,
2798 'filesize': filesize
,
2799 'container': mimetype2ext(mime_type
) + '_dash',
2802 elif content_type
== 'text':
2804 'ext': mimetype2ext(mime_type
),
2805 'manifest_url': mpd_url
,
2806 'filesize': filesize
,
2808 elif content_type
== 'image/jpeg':
2809 # See test case in VikiIE
2810 # https://www.viki.com/videos/1175236v-choosing-spouse-by-lottery-episode-1
2812 'format_id': format_id
,
2814 'manifest_url': mpd_url
,
2815 'format_note': 'DASH storyboards (jpeg)',
2819 if is_drm_protected(adaptation_set
) or is_drm_protected(representation
):
2821 representation_ms_info
= extract_multisegment_info(representation
, adaption_set_ms_info
)
2823 def prepare_template(template_name
, identifiers
):
2824 tmpl
= representation_ms_info
[template_name
]
2825 # First of, % characters outside $...$ templates
2826 # must be escaped by doubling for proper processing
2827 # by % operator string formatting used further (see
2828 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2834 in_template
= not in_template
2835 elif c
== '%' and not in_template
:
2837 # Next, $...$ templates are translated to their
2838 # %(...) counterparts to be used with % operator
2839 if representation_id
is not None:
2840 t
= t
.replace('$RepresentationID$', representation_id
)
2841 t
= re
.sub(r
'\$(%s)\$' % '|'.join(identifiers
), r
'%(\1)d', t
)
2842 t
= re
.sub(r
'\$(%s)%%([^$]+)\$' % '|'.join(identifiers
), r
'%(\1)\2', t
)
2843 t
.replace('$$', '$')
2846 # @initialization is a regular template like @media one
2847 # so it should be handled just the same way (see
2848 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2849 if 'initialization' in representation_ms_info
:
2850 initialization_template
= prepare_template(
2852 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2853 # $Time$ shall not be included for @initialization thus
2854 # only $Bandwidth$ remains
2856 representation_ms_info
['initialization_url'] = initialization_template
% {
2857 'Bandwidth': bandwidth
,
2860 def location_key(location
):
2861 return 'url' if re
.match(r
'^https?://', location
) else 'path'
2863 if 'segment_urls' not in representation_ms_info
and 'media' in representation_ms_info
:
2865 media_template
= prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2866 media_location_key
= location_key(media_template
)
2868 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2869 # can't be used at the same time
2870 if '%(Number' in media_template
and 's' not in representation_ms_info
:
2871 segment_duration
= None
2872 if 'total_number' not in representation_ms_info
and 'segment_duration' in representation_ms_info
:
2873 segment_duration
= float_or_none(representation_ms_info
['segment_duration'], representation_ms_info
['timescale'])
2874 representation_ms_info
['total_number'] = int(math
.ceil(float(period_duration
) / segment_duration
))
2875 representation_ms_info
['fragments'] = [{
2876 media_location_key
: media_template
% {
2877 'Number': segment_number
,
2878 'Bandwidth': bandwidth
,
2880 'duration': segment_duration
,
2881 } for segment_number
in range(
2882 representation_ms_info
['start_number'],
2883 representation_ms_info
['total_number'] + representation_ms_info
['start_number'])]
2885 # $Number*$ or $Time$ in media template with S list available
2886 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2887 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2888 representation_ms_info
['fragments'] = []
2891 segment_number
= representation_ms_info
['start_number']
2893 def add_segment_url():
2894 segment_url
= media_template
% {
2895 'Time': segment_time
,
2896 'Bandwidth': bandwidth
,
2897 'Number': segment_number
,
2899 representation_ms_info
['fragments'].append({
2900 media_location_key
: segment_url
,
2901 'duration': float_or_none(segment_d
, representation_ms_info
['timescale']),
2904 for num
, s
in enumerate(representation_ms_info
['s']):
2905 segment_time
= s
.get('t') or segment_time
2909 for r
in range(s
.get('r', 0)):
2910 segment_time
+= segment_d
2913 segment_time
+= segment_d
2914 elif 'segment_urls' in representation_ms_info
and 's' in representation_ms_info
:
2916 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2917 # or any YouTube dashsegments video
2920 timescale
= representation_ms_info
['timescale']
2921 for s
in representation_ms_info
['s']:
2922 duration
= float_or_none(s
['d'], timescale
)
2923 for r
in range(s
.get('r', 0) + 1):
2924 segment_uri
= representation_ms_info
['segment_urls'][segment_index
]
2926 location_key(segment_uri
): segment_uri
,
2927 'duration': duration
,
2930 representation_ms_info
['fragments'] = fragments
2931 elif 'segment_urls' in representation_ms_info
:
2932 # Segment URLs with no SegmentTimeline
2933 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2934 # https://github.com/ytdl-org/youtube-dl/pull/14844
2936 segment_duration
= float_or_none(
2937 representation_ms_info
['segment_duration'],
2938 representation_ms_info
['timescale']) if 'segment_duration' in representation_ms_info
else None
2939 for segment_url
in representation_ms_info
['segment_urls']:
2941 location_key(segment_url
): segment_url
,
2943 if segment_duration
:
2944 fragment
['duration'] = segment_duration
2945 fragments
.append(fragment
)
2946 representation_ms_info
['fragments'] = fragments
2947 # If there is a fragments key available then we correctly recognized fragmented media.
2948 # Otherwise we will assume unfragmented media with direct access. Technically, such
2949 # assumption is not necessarily correct since we may simply have no support for
2950 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2951 if 'fragments' in representation_ms_info
:
2953 # NB: mpd_url may be empty when MPD manifest is parsed from a string
2954 'url': mpd_url
or base_url
,
2955 'fragment_base_url': base_url
,
2957 'protocol': 'http_dash_segments' if mime_type
!= 'image/jpeg' else 'mhtml',
2959 if 'initialization_url' in representation_ms_info
:
2960 initialization_url
= representation_ms_info
['initialization_url']
2961 if not f
.get('url'):
2962 f
['url'] = initialization_url
2963 f
['fragments'].append({location_key(initialization_url): initialization_url}
)
2964 f
['fragments'].extend(representation_ms_info
['fragments'])
2966 # Assuming direct URL to unfragmented media.
2968 if content_type
in ('video', 'audio', 'image/jpeg'):
2969 f
['manifest_stream_number'] = stream_numbers
[f
['url']]
2970 stream_numbers
[f
['url']] += 1
2972 elif content_type
== 'text':
2973 subtitles
.setdefault(lang
or 'und', []).append(f
)
2975 return formats
, subtitles
2977 def _extract_ism_formats(self
, *args
, **kwargs
):
2978 fmts
, subs
= self
._extract
_ism
_formats
_and
_subtitles
(*args
, **kwargs
)
2980 self
._report
_ignoring
_subs
('ISM')
2983 def _extract_ism_formats_and_subtitles(self
, ism_url
, video_id
, ism_id
=None, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
):
2984 res
= self
._download
_xml
_handle
(
2986 note
='Downloading ISM manifest' if note
is None else note
,
2987 errnote
='Failed to download ISM manifest' if errnote
is None else errnote
,
2988 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
2995 return self
._parse
_ism
_formats
_and
_subtitles
(ism_doc
, urlh
.geturl(), ism_id
)
2997 def _parse_ism_formats_and_subtitles(self
, ism_doc
, ism_url
, ism_id
=None):
2999 Parse formats from ISM manifest.
3001 1. [MS-SSTR]: Smooth Streaming Protocol,
3002 https://msdn.microsoft.com/en-us/library/ff469518.aspx
3004 if ism_doc
.get('IsLive') == 'TRUE':
3007 duration
= int(ism_doc
.attrib
['Duration'])
3008 timescale
= int_or_none(ism_doc
.get('TimeScale')) or 10000000
3012 for stream
in ism_doc
.findall('StreamIndex'):
3013 stream_type
= stream
.get('Type')
3014 if stream_type
not in ('video', 'audio', 'text'):
3016 url_pattern
= stream
.attrib
['Url']
3017 stream_timescale
= int_or_none(stream
.get('TimeScale')) or timescale
3018 stream_name
= stream
.get('Name')
3019 stream_language
= stream
.get('Language', 'und')
3020 for track
in stream
.findall('QualityLevel'):
3021 fourcc
= track
.get('FourCC') or ('AACL' if track
.get('AudioTag') == '255' else None)
3022 # TODO: add support for WVC1 and WMAP
3023 if fourcc
not in ('H264', 'AVC1', 'AACL', 'TTML'):
3024 self
.report_warning('%s is not a supported codec' % fourcc
)
3026 tbr
= int(track
.attrib
['Bitrate']) // 1000
3027 # [1] does not mention Width and Height attributes. However,
3028 # they're often present while MaxWidth and MaxHeight are
3029 # missing, so should be used as fallbacks
3030 width
= int_or_none(track
.get('MaxWidth') or track
.get('Width'))
3031 height
= int_or_none(track
.get('MaxHeight') or track
.get('Height'))
3032 sampling_rate
= int_or_none(track
.get('SamplingRate'))
3034 track_url_pattern
= re
.sub(r
'{[Bb]itrate}', track
.attrib
['Bitrate'], url_pattern
)
3035 track_url_pattern
= compat_urlparse
.urljoin(ism_url
, track_url_pattern
)
3041 stream_fragments
= stream
.findall('c')
3042 for stream_fragment_index
, stream_fragment
in enumerate(stream_fragments
):
3043 fragment_ctx
['time'] = int_or_none(stream_fragment
.get('t')) or fragment_ctx
['time']
3044 fragment_repeat
= int_or_none(stream_fragment
.get('r')) or 1
3045 fragment_ctx
['duration'] = int_or_none(stream_fragment
.get('d'))
3046 if not fragment_ctx
['duration']:
3048 next_fragment_time
= int(stream_fragment
[stream_fragment_index
+ 1].attrib
['t'])
3050 next_fragment_time
= duration
3051 fragment_ctx
['duration'] = (next_fragment_time
- fragment_ctx
['time']) / fragment_repeat
3052 for _
in range(fragment_repeat
):
3054 'url': re
.sub(r
'{start[ _]time}', compat_str(fragment_ctx
['time']), track_url_pattern
),
3055 'duration': fragment_ctx
['duration'] / stream_timescale
,
3057 fragment_ctx
['time'] += fragment_ctx
['duration']
3059 if stream_type
== 'text':
3060 subtitles
.setdefault(stream_language
, []).append({
3064 'manifest_url': ism_url
,
3065 'fragments': fragments
,
3066 '_download_params': {
3067 'stream_type': stream_type
,
3068 'duration': duration
,
3069 'timescale': stream_timescale
,
3071 'language': stream_language
,
3072 'codec_private_data': track
.get('CodecPrivateData'),
3075 elif stream_type
in ('video', 'audio'):
3077 'format_id': join_nonempty(ism_id
, stream_name
, tbr
),
3079 'manifest_url': ism_url
,
3080 'ext': 'ismv' if stream_type
== 'video' else 'isma',
3084 'asr': sampling_rate
,
3085 'vcodec': 'none' if stream_type
== 'audio' else fourcc
,
3086 'acodec': 'none' if stream_type
== 'video' else fourcc
,
3088 'fragments': fragments
,
3089 'has_drm': ism_doc
.find('Protection') is not None,
3090 '_download_params': {
3091 'stream_type': stream_type
,
3092 'duration': duration
,
3093 'timescale': stream_timescale
,
3094 'width': width
or 0,
3095 'height': height
or 0,
3097 'language': stream_language
,
3098 'codec_private_data': track
.get('CodecPrivateData'),
3099 'sampling_rate': sampling_rate
,
3100 'channels': int_or_none(track
.get('Channels', 2)),
3101 'bits_per_sample': int_or_none(track
.get('BitsPerSample', 16)),
3102 'nal_unit_length_field': int_or_none(track
.get('NALUnitLengthField', 4)),
3105 return formats
, subtitles
3107 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):
3108 def absolute_url(item_url
):
3109 return urljoin(base_url
, item_url
)
3111 def parse_content_type(content_type
):
3112 if not content_type
:
3114 ctr
= re
.search(r
'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type
)
3116 mimetype
, codecs
= ctr
.groups()
3117 f
= parse_codecs(codecs
)
3118 f
['ext'] = mimetype2ext(mimetype
)
3122 def _media_formats(src
, cur_media_type
, type_info
={}):
3123 full_url
= absolute_url(src
)
3124 ext
= type_info
.get('ext') or determine_ext(full_url
)
3126 is_plain_url
= False
3127 formats
= self
._extract
_m
3u8_formats
(
3128 full_url
, video_id
, ext
='mp4',
3129 entry_protocol
=m3u8_entry_protocol
, m3u8_id
=m3u8_id
,
3130 preference
=preference
, quality
=quality
, fatal
=False)
3132 is_plain_url
= False
3133 formats
= self
._extract
_mpd
_formats
(
3134 full_url
, video_id
, mpd_id
=mpd_id
, fatal
=False)
3139 'vcodec': 'none' if cur_media_type
== 'audio' else None,
3141 return is_plain_url
, formats
3144 # amp-video and amp-audio are very similar to their HTML5 counterparts
3145 # so we wll include them right here (see
3146 # https://www.ampproject.org/docs/reference/components/amp-video)
3147 # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/
3148 _MEDIA_TAG_NAME_RE
= r
'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)'
3149 media_tags
= [(media_tag
, media_tag_name
, media_type
, '')
3150 for media_tag
, media_tag_name
, media_type
3151 in re
.findall(r
'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE
, webpage
)]
3152 media_tags
.extend(re
.findall(
3153 # We only allow video|audio followed by a whitespace or '>'.
3154 # Allowing more characters may end up in significant slow down (see
3155 # https://github.com/ytdl-org/youtube-dl/issues/11979, example URL:
3156 # http://www.porntrex.com/maps/videositemap.xml).
3157 r
'(?s)(<(?P<tag>%s)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>' % _MEDIA_TAG_NAME_RE
, webpage
))
3158 for media_tag
, _
, media_type
, media_content
in media_tags
:
3163 media_attributes
= extract_attributes(media_tag
)
3164 src
= strip_or_none(media_attributes
.get('src'))
3166 _
, formats
= _media_formats(src
, media_type
)
3167 media_info
['formats'].extend(formats
)
3168 media_info
['thumbnail'] = absolute_url(media_attributes
.get('poster'))
3170 for source_tag
in re
.findall(r
'<source[^>]+>', media_content
):
3171 s_attr
= extract_attributes(source_tag
)
3172 # data-video-src and data-src are non standard but seen
3173 # several times in the wild
3174 src
= strip_or_none(dict_get(s_attr
, ('src', 'data-video-src', 'data-src')))
3177 f
= parse_content_type(s_attr
.get('type'))
3178 is_plain_url
, formats
= _media_formats(src
, media_type
, f
)
3180 # width, height, res, label and title attributes are
3181 # all not standard but seen several times in the wild
3184 for lbl
in ('label', 'title')
3185 if str_or_none(s_attr
.get(lbl
))
3187 width
= int_or_none(s_attr
.get('width'))
3188 height
= (int_or_none(s_attr
.get('height'))
3189 or int_or_none(s_attr
.get('res')))
3190 if not width
or not height
:
3192 resolution
= parse_resolution(lbl
)
3195 width
= width
or resolution
.get('width')
3196 height
= height
or resolution
.get('height')
3198 tbr
= parse_bitrate(lbl
)
3207 'format_id': s_attr
.get('label') or s_attr
.get('title'),
3209 f
.update(formats
[0])
3210 media_info
['formats'].append(f
)
3212 media_info
['formats'].extend(formats
)
3213 for track_tag
in re
.findall(r
'<track[^>]+>', media_content
):
3214 track_attributes
= extract_attributes(track_tag
)
3215 kind
= track_attributes
.get('kind')
3216 if not kind
or kind
in ('subtitles', 'captions'):
3217 src
= strip_or_none(track_attributes
.get('src'))
3220 lang
= track_attributes
.get('srclang') or track_attributes
.get('lang') or track_attributes
.get('label')
3221 media_info
['subtitles'].setdefault(lang
, []).append({
3222 'url': absolute_url(src
),
3224 for f
in media_info
['formats']:
3225 f
.setdefault('http_headers', {})['Referer'] = base_url
3226 if media_info
['formats'] or media_info
['subtitles']:
3227 entries
.append(media_info
)
3230 def _extract_akamai_formats(self
, *args
, **kwargs
):
3231 fmts
, subs
= self
._extract
_akamai
_formats
_and
_subtitles
(*args
, **kwargs
)
3233 self
._report
_ignoring
_subs
('akamai')
3236 def _extract_akamai_formats_and_subtitles(self
, manifest_url
, video_id
, hosts
={}):
3237 signed
= 'hdnea=' in manifest_url
3239 # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
3240 manifest_url
= re
.sub(
3241 r
'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
3242 '', manifest_url
).strip('?')
3247 hdcore_sign
= 'hdcore=3.7.0'
3248 f4m_url
= re
.sub(r
'(https?://[^/]+)/i/', r
'\1/z/', manifest_url
).replace('/master.m3u8', '/manifest.f4m')
3249 hds_host
= hosts
.get('hds')
3251 f4m_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hds_host
, f4m_url
)
3252 if 'hdcore=' not in f4m_url
:
3253 f4m_url
+= ('&' if '?' in f4m_url
else '?') + hdcore_sign
3254 f4m_formats
= self
._extract
_f
4m
_formats
(
3255 f4m_url
, video_id
, f4m_id
='hds', fatal
=False)
3256 for entry
in f4m_formats
:
3257 entry
.update({'extra_param_to_segment_url': hdcore_sign}
)
3258 formats
.extend(f4m_formats
)
3260 m3u8_url
= re
.sub(r
'(https?://[^/]+)/z/', r
'\1/i/', manifest_url
).replace('/manifest.f4m', '/master.m3u8')
3261 hls_host
= hosts
.get('hls')
3263 m3u8_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hls_host
, m3u8_url
)
3264 m3u8_formats
, m3u8_subtitles
= self
._extract
_m
3u8_formats
_and
_subtitles
(
3265 m3u8_url
, video_id
, 'mp4', 'm3u8_native',
3266 m3u8_id
='hls', fatal
=False)
3267 formats
.extend(m3u8_formats
)
3268 subtitles
= self
._merge
_subtitles
(subtitles
, m3u8_subtitles
)
3270 http_host
= hosts
.get('http')
3271 if http_host
and m3u8_formats
and not signed
:
3272 REPL_REGEX
= r
'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
3273 qualities
= re
.match(REPL_REGEX
, m3u8_url
).group(2).split(',')
3274 qualities_length
= len(qualities
)
3275 if len(m3u8_formats
) in (qualities_length
, qualities_length
+ 1):
3277 for f
in m3u8_formats
:
3278 if f
['vcodec'] != 'none':
3279 for protocol
in ('http', 'https'):
3281 del http_f
['manifest_url']
3283 REPL_REGEX
, protocol
+ r
'://%s/\g<1>%s\3' % (http_host
, qualities
[i
]), f
['url'])
3285 'format_id': http_f
['format_id'].replace('hls-', protocol
+ '-'),
3287 'protocol': protocol
,
3289 formats
.append(http_f
)
3292 return formats
, subtitles
3294 def _extract_wowza_formats(self
, url
, video_id
, m3u8_entry_protocol
='m3u8_native', skip_protocols
=[]):
3295 query
= compat_urlparse
.urlparse(url
).query
3296 url
= re
.sub(r
'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url
)
3298 r
'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url
)
3299 url_base
= mobj
.group('url')
3300 http_base_url
= '%s%s:%s' % ('http', mobj
.group('s') or '', url_base
)
3303 def manifest_url(manifest
):
3304 m_url
= '%s/%s' % (http_base_url
, manifest
)
3306 m_url
+= '?%s' % query
3309 if 'm3u8' not in skip_protocols
:
3310 formats
.extend(self
._extract
_m
3u8_formats
(
3311 manifest_url('playlist.m3u8'), video_id
, 'mp4',
3312 m3u8_entry_protocol
, m3u8_id
='hls', fatal
=False))
3313 if 'f4m' not in skip_protocols
:
3314 formats
.extend(self
._extract
_f
4m
_formats
(
3315 manifest_url('manifest.f4m'),
3316 video_id
, f4m_id
='hds', fatal
=False))
3317 if 'dash' not in skip_protocols
:
3318 formats
.extend(self
._extract
_mpd
_formats
(
3319 manifest_url('manifest.mpd'),
3320 video_id
, mpd_id
='dash', fatal
=False))
3321 if re
.search(r
'(?:/smil:|\.smil)', url_base
):
3322 if 'smil' not in skip_protocols
:
3323 rtmp_formats
= self
._extract
_smil
_formats
(
3324 manifest_url('jwplayer.smil'),
3325 video_id
, fatal
=False)
3326 for rtmp_format
in rtmp_formats
:
3327 rtsp_format
= rtmp_format
.copy()
3328 rtsp_format
['url'] = '%s/%s' % (rtmp_format
['url'], rtmp_format
['play_path'])
3329 del rtsp_format
['play_path']
3330 del rtsp_format
['ext']
3331 rtsp_format
.update({
3332 'url': rtsp_format
['url'].replace('rtmp://', 'rtsp://'),
3333 'format_id': rtmp_format
['format_id'].replace('rtmp', 'rtsp'),
3336 formats
.extend([rtmp_format
, rtsp_format
])
3338 for protocol
in ('rtmp', 'rtsp'):
3339 if protocol
not in skip_protocols
:
3341 'url': '%s:%s' % (protocol
, url_base
),
3342 'format_id': protocol
,
3343 'protocol': protocol
,
3347 def _find_jwplayer_data(self
, webpage
, video_id
=None, transform_source
=js_to_json
):
3349 r
'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P
=quote
)\
)(?
!</script
>).*?\
.setup\s
*\
((?P
<options
>[^
)]+)\
)',
3353 jwplayer_data = self._parse_json(mobj.group('options
'),
3355 transform_source=transform_source)
3356 except ExtractorError:
3359 if isinstance(jwplayer_data, dict):
3360 return jwplayer_data
3362 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
3363 jwplayer_data = self._find_jwplayer_data(
3364 webpage, video_id, transform_source=js_to_json)
3365 return self._parse_jwplayer_data(
3366 jwplayer_data, video_id, *args, **kwargs)
3368 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
3369 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3370 # JWPlayer backward compatibility: flattened playlists
3371 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
3372 if 'playlist
' not in jwplayer_data:
3373 jwplayer_data = {'playlist': [jwplayer_data]}
3377 # JWPlayer backward compatibility: single playlist item
3378 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
3379 if not isinstance(jwplayer_data['playlist
'], list):
3380 jwplayer_data['playlist
'] = [jwplayer_data['playlist
']]
3382 for video_data in jwplayer_data['playlist
']:
3383 # JWPlayer backward compatibility: flattened sources
3384 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
3385 if 'sources
' not in video_data:
3386 video_data['sources
'] = [video_data]
3388 this_video_id = video_id or video_data['mediaid
']
3390 formats = self._parse_jwplayer_formats(
3391 video_data['sources
'], video_id=this_video_id, m3u8_id=m3u8_id,
3392 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
3395 tracks = video_data.get('tracks
')
3396 if tracks and isinstance(tracks, list):
3397 for track in tracks:
3398 if not isinstance(track, dict):
3400 track_kind = track.get('kind
')
3401 if not track_kind or not isinstance(track_kind, compat_str):
3403 if track_kind.lower() not in ('captions
', 'subtitles
'):
3405 track_url = urljoin(base_url, track.get('file'))
3408 subtitles.setdefault(track.get('label
') or 'en
', []).append({
3409 'url
': self._proto_relative_url(track_url)
3413 'id': this_video_id,
3414 'title
': unescapeHTML(video_data['title
'] if require_title else video_data.get('title
')),
3415 'description
': clean_html(video_data.get('description
')),
3416 'thumbnail
': urljoin(base_url, self._proto_relative_url(video_data.get('image
'))),
3417 'timestamp
': int_or_none(video_data.get('pubdate
')),
3418 'duration
': float_or_none(jwplayer_data.get('duration
') or video_data.get('duration
')),
3419 'subtitles
': subtitles,
3421 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
3422 if len(formats) == 1 and re.search(r'^
(?
:http|
//).*(?
:youtube\
.com|youtu\
.be
)/.+', formats[0]['url
']):
3424 '_type
': 'url_transparent
',
3425 'url
': formats[0]['url
'],
3428 self._sort_formats(formats)
3429 entry['formats
'] = formats
3430 entries.append(entry)
3431 if len(entries) == 1:
3434 return self.playlist_result(entries)
3436 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
3437 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3440 for source in jwplayer_sources_data:
3441 if not isinstance(source, dict):
3443 source_url = urljoin(
3444 base_url, self._proto_relative_url(source.get('file')))
3445 if not source_url or source_url in urls:
3447 urls.append(source_url)
3448 source_type = source.get('type') or ''
3449 ext = mimetype2ext(source_type) or determine_ext(source_url)
3450 if source_type == 'hls
' or ext == 'm3u8
':
3451 formats.extend(self._extract_m3u8_formats(
3452 source_url, video_id, 'mp4
', entry_protocol='m3u8_native
',
3453 m3u8_id=m3u8_id, fatal=False))
3454 elif source_type == 'dash
' or ext == 'mpd
':
3455 formats.extend(self._extract_mpd_formats(
3456 source_url, video_id, mpd_id=mpd_id, fatal=False))
3458 formats.extend(self._extract_smil_formats(
3459 source_url, video_id, fatal=False))
3460 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3461 elif source_type.startswith('audio
') or ext in (
3462 'oga
', 'aac
', 'mp3
', 'mpeg
', 'vorbis
'):
3469 height = int_or_none(source.get('height
'))
3471 # Often no height is provided but there is a label in
3472 # format like "1080p", "720p SD", or 1080.
3473 height = int_or_none(self._search_regex(
3474 r'^
(\d{3,4}
)[pP
]?
(?
:\b|$
)', compat_str(source.get('label
') or ''),
3475 'height
', default=None))
3478 'width
': int_or_none(source.get('width
')),
3480 'tbr
': int_or_none(source.get('bitrate
')),
3483 if source_url.startswith('rtmp
'):
3484 a_format['ext
'] = 'flv
'
3485 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3486 # of jwplayer.flash.swf
3487 rtmp_url_parts = re.split(
3488 r'((?
:mp4|mp3|flv
):)', source_url, 1)
3489 if len(rtmp_url_parts) == 3:
3490 rtmp_url, prefix, play_path = rtmp_url_parts
3493 'play_path
': prefix + play_path,
3496 a_format.update(rtmp_params)
3497 formats.append(a_format)
3500 def _live_title(self, name):
3501 self._downloader.deprecation_warning('yt_dlp
.InfoExtractor
._live
_title
is deprecated
and does
not work
as expected
')
3504 def _int(self, v, name, fatal=False, **kwargs):
3505 res = int_or_none(v, **kwargs)
3506 if 'get_attr
' in kwargs:
3507 print(getattr(v, kwargs['get_attr
']))
3509 msg = 'Failed to extract
%s: Could
not parse value
%r' % (name, v)
3511 raise ExtractorError(msg)
3513 self.report_warning(msg)
3516 def _float(self, v, name, fatal=False, **kwargs):
3517 res = float_or_none(v, **kwargs)
3519 msg = 'Failed to extract
%s: Could
not parse value
%r' % (name, v)
3521 raise ExtractorError(msg)
3523 self.report_warning(msg)
3526 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
3527 path='/', secure=False, discard=False, rest={}, **kwargs):
3528 cookie = compat_cookiejar_Cookie(
3529 0, name, value, port, port is not None, domain, True,
3530 domain.startswith('.'), path, True, secure, expire_time,
3531 discard, None, None, rest)
3532 self._downloader.cookiejar.set_cookie(cookie)
3534 def _get_cookies(self, url):
3535 """ Return a compat_cookies_SimpleCookie with the cookies for the url """
3536 req = sanitized_Request(url)
3537 self._downloader.cookiejar.add_cookie_header(req)
3538 return compat_cookies_SimpleCookie(req.get_header('Cookie
'))
3540 def _apply_first_set_cookie_header(self, url_handle, cookie):
3542 Apply first Set-Cookie header instead of the last. Experimental.
3544 Some sites (e.g. [1-3]) may serve two cookies under the same name
3545 in Set-Cookie header and expect the first (old) one to be set rather
3546 than second (new). However, as of RFC6265 the newer one cookie
3547 should be set into cookie store what actually happens.
3548 We will workaround this issue by resetting the cookie to
3549 the first one manually.
3550 1. https://new.vk.com/
3551 2. https://github.com/ytdl-org/youtube-dl/issues/9841#issuecomment-227871201
3552 3. https://learning.oreilly.com/
3554 for header, cookies in url_handle.headers.items():
3555 if header.lower() != 'set-cookie
':
3557 if sys.version_info[0] >= 3:
3558 cookies = cookies.encode('iso
-8859-1')
3559 cookies = cookies.decode('utf
-8')
3560 cookie_value = re.search(
3561 r'%s=(.+?
);.*?
\b[Dd
]omain
=(.+?
)(?
:[,;]|$
)' % cookie, cookies)
3563 value, domain = cookie_value.groups()
3564 self._set_cookie(domain, cookie, value)
3567 def get_testcases(self, include_onlymatching=False):
3568 t = getattr(self, '_TEST
', None)
3570 assert not hasattr(self, '_TESTS
'), \
3571 '%s has _TEST
and _TESTS
' % type(self).__name__
3574 tests = getattr(self, '_TESTS
', [])
3576 if not include_onlymatching and t.get('only_matching
', False):
3578 t['name
'] = type(self).__name__[:-len('IE
')]
3581 def is_suitable(self, age_limit):
3582 """ Test whether the extractor is generally suitable for the given
3583 age limit (i.e. pornographic sites are not, all others usually are) """
3585 any_restricted = False
3586 for tc in self.get_testcases(include_onlymatching=False):
3587 if tc.get('playlist
', []):
3588 tc = tc['playlist
'][0]
3589 is_restricted = age_restricted(
3590 tc.get('info_dict
', {}).get('age_limit
'), age_limit)
3591 if not is_restricted:
3593 any_restricted = any_restricted or is_restricted
3594 return not any_restricted
3596 def extract_subtitles(self, *args, **kwargs):
3597 if (self.get_param('writesubtitles
', False)
3598 or self.get_param('listsubtitles
')):
3599 return self._get_subtitles(*args, **kwargs)
3602 def _get_subtitles(self, *args, **kwargs):
3603 raise NotImplementedError('This method must be implemented by subclasses
')
3605 def extract_comments(self, *args, **kwargs):
3606 if not self.get_param('getcomments
'):
3608 generator = self._get_comments(*args, **kwargs)
3615 comments.append(next(generator))
3616 except StopIteration:
3618 except KeyboardInterrupt:
3619 self.to_screen('Interrupted by user
')
3620 except Exception as e:
3621 if self.get_param('ignoreerrors
') is not True:
3623 self._downloader.report_error(e)
3624 comment_count = len(comments)
3625 self.to_screen(f'Extracted {comment_count} comments
')
3627 'comments
': comments,
3628 'comment_count
': None if interrupted else comment_count
3632 def _get_comments(self, *args, **kwargs):
3633 raise NotImplementedError('This method must be implemented by subclasses
')
3636 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
3637 """ Merge subtitle items for one language. Items with duplicated URLs
3638 will be dropped. """
3639 list1_urls = set([item['url
'] for item in subtitle_list1])
3640 ret = list(subtitle_list1)
3641 ret.extend([item for item in subtitle_list2 if item['url
'] not in list1_urls])
3645 def _merge_subtitles(cls, *dicts, target=None):
3646 """ Merge subtitle dictionaries, language by language. """
3650 for lang, subs in d.items():
3651 target[lang] = cls._merge_subtitle_items(target.get(lang, []), subs)
3654 def extract_automatic_captions(self, *args, **kwargs):
3655 if (self.get_param('writeautomaticsub
', False)
3656 or self.get_param('listsubtitles
')):
3657 return self._get_automatic_captions(*args, **kwargs)
3660 def _get_automatic_captions(self, *args, **kwargs):
3661 raise NotImplementedError('This method must be implemented by subclasses
')
3663 def mark_watched(self, *args, **kwargs):
3664 if not self.get_param('mark_watched
', False):
3666 if (self._get_login_info()[0] is not None
3667 or self.get_param('cookiefile
')
3668 or self.get_param('cookiesfrombrowser
')):
3669 self._mark_watched(*args, **kwargs)
3671 def _mark_watched(self, *args, **kwargs):
3672 raise NotImplementedError('This method must be implemented by subclasses
')
3674 def geo_verification_headers(self):
3676 geo_verification_proxy = self.get_param('geo_verification_proxy
')
3677 if geo_verification_proxy:
3678 headers['Ytdl
-request
-proxy
'] = geo_verification_proxy
3681 def _generic_id(self, url):
3682 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
3684 def _generic_title(self, url):
3685 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
3688 def _availability(is_private=None, needs_premium=None, needs_subscription=None, needs_auth=None, is_unlisted=None):
3689 all_known = all(map(
3690 lambda x: x is not None,
3691 (is_private, needs_premium, needs_subscription, needs_auth, is_unlisted)))
3693 'private
' if is_private
3694 else 'premium_only
' if needs_premium
3695 else 'subscriber_only
' if needs_subscription
3696 else 'needs_auth
' if needs_auth
3697 else 'unlisted
' if is_unlisted
3698 else 'public
' if all_known
3701 def _configuration_arg(self, key, default=NO_DEFAULT, *, ie_key=None, casesense=False):
3703 @returns A list of values for the extractor argument given by "key"
3704 or "default" if no such key is present
3705 @param default The default value to return when the key is not present (default: [])
3706 @param casesense When false, the values are converted to lower case
3709 self._downloader.params, ('extractor_args
', (ie_key or self.ie_key()).lower(), key))
3711 return [] if default is NO_DEFAULT else default
3712 return list(val) if casesense else [x.lower() for x in val]
3715 class SearchInfoExtractor(InfoExtractor):
3717 Base class for paged search queries extractors.
3718 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
3719 Instances should define _SEARCH_KEY and optionally _MAX_RESULTS
3722 _MAX_RESULTS = float('inf
')
3725 def _make_valid_url(cls):
3726 return r'%s(?P
<prefix
>|
[1-9][0-9]*|all
):(?P
<query
>[\s\S
]+)' % cls._SEARCH_KEY
3728 def _real_extract(self, query):
3729 prefix, query = self._match_valid_url(query).group('prefix
', 'query
')
3731 return self._get_n_results(query, 1)
3732 elif prefix == 'all
':
3733 return self._get_n_results(query, self._MAX_RESULTS)
3737 raise ExtractorError('invalid download number
%s for query
"%s"' % (n, query))
3738 elif n > self._MAX_RESULTS:
3739 self.report_warning('%s returns
max %i results (you requested
%i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3740 n = self._MAX_RESULTS
3741 return self._get_n_results(query, n)
3743 def _get_n_results(self, query, n):
3744 """Get a specified number of results for a query.
3745 Either this function or _search_results must be overridden by subclasses """
3746 return self.playlist_result(
3747 itertools.islice(self._search_results(query), 0, None if n == float('inf
') else n),
3750 def _search_results(self, query):
3751 """Returns an iterator of search results"""
3752 raise NotImplementedError('This method must be implemented by subclasses
')
3755 def SEARCH_KEY(self):
3756 return self._SEARCH_KEY