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
,
95 class InfoExtractor(object):
96 """Information Extractor class.
98 Information extractors are the classes that, given a URL, extract
99 information about the video (or videos) the URL refers to. This
100 information includes the real video URL, the video title, author and
101 others. The information is stored in a dictionary which is then
102 passed to the YoutubeDL. The YoutubeDL processes this
103 information possibly downloading the video to the file system, among
104 other possible outcomes.
106 The type field determines the type of the result.
107 By far the most common value (and the default if _type is missing) is
108 "video", which indicates a single video.
110 For a video, the dictionaries must include the following fields:
112 id: Video identifier.
113 title: Video title, unescaped.
115 Additionally, it must contain either a formats entry or a url one:
117 formats: A list of dictionaries for each format available, ordered
118 from worst to best quality.
121 * url The mandatory URL representing the media:
122 for plain file media - HTTP URL of this file,
124 for HLS - URL of the M3U8 media playlist,
125 for HDS - URL of the F4M manifest,
127 - HTTP URL to plain file media (in case of
129 - URL of the MPD manifest or base URL
130 representing the media if MPD manifest
131 is parsed from a string (in case of
133 for MSS - URL of the ISM manifest.
135 The URL of the manifest file in case of
137 for HLS - URL of the M3U8 master playlist,
138 for HDS - URL of the F4M manifest,
139 for DASH - URL of the MPD manifest,
140 for MSS - URL of the ISM manifest.
141 * ext Will be calculated from URL if missing
142 * format A human-readable description of the format
143 ("mp4 container with h264/opus").
144 Calculated from the format_id, width, height.
145 and format_note fields if missing.
146 * format_id A short description of the format
147 ("mp4_h264_opus" or "19").
148 Technically optional, but strongly recommended.
149 * format_note Additional info about the format
150 ("3D" or "DASH video")
151 * width Width of the video, if known
152 * height Height of the video, if known
153 * resolution Textual description of width and height
154 * dynamic_range The dynamic range of the video. One of:
155 "SDR" (None), "HDR10", "HDR10+, "HDR12", "HLG, "DV"
156 * tbr Average bitrate of audio and video in KBit/s
157 * abr Average audio bitrate in KBit/s
158 * acodec Name of the audio codec in use
159 * asr Audio sampling rate in Hertz
160 * vbr Average video bitrate in KBit/s
162 * vcodec Name of the video codec in use
163 * container Name of the container format
164 * filesize The number of bytes, if known in advance
165 * filesize_approx An estimate for the number of bytes
166 * player_url SWF Player URL (used for rtmpdump).
167 * protocol The protocol that will be used for the actual
168 download, lower-case. One of "http", "https" or
169 one of the protocols defined in downloader.PROTOCOL_MAP
171 Base URL for fragments. Each fragment's path
172 value (if present) will be relative to
174 * fragments A list of fragments of a fragmented media.
175 Each fragment entry must contain either an url
176 or a path. If an url is present it should be
177 considered by a client. Otherwise both path and
178 fragment_base_url must be present. Here is
179 the list of all potential fields:
180 * "url" - fragment's URL
181 * "path" - fragment's path relative to
183 * "duration" (optional, int or float)
184 * "filesize" (optional, int)
185 * is_from_start Is a live format that can be downloaded
186 from the start. Boolean
187 * preference Order number of this format. If this field is
188 present and not None, the formats get sorted
189 by this field, regardless of all other values.
190 -1 for default (order by other properties),
191 -2 or smaller for less than default.
192 < -1000 to hide the format (if there is
193 another one which is strictly better)
194 * language Language code, e.g. "de" or "en-US".
195 * language_preference Is this in the language mentioned in
197 10 if it's what the URL is about,
198 -1 for default (don't know),
199 -10 otherwise, other values reserved for now.
200 * quality Order number of the video quality of this
201 format, irrespective of the file format.
202 -1 for default (order by other properties),
203 -2 or smaller for less than default.
204 * source_preference Order number for this video source
205 (quality takes higher priority)
206 -1 for default (order by other properties),
207 -2 or smaller for less than default.
208 * http_headers A dictionary of additional HTTP headers
209 to add to the request.
210 * stretched_ratio If given and not 1, indicates that the
211 video's pixels are not square.
212 width : height ratio as float.
213 * no_resume The server does not support resuming the
214 (HTTP or RTMP) download. Boolean.
215 * has_drm The format has DRM and cannot be downloaded. Boolean
216 * downloader_options A dictionary of downloader options as
217 described in FileDownloader
218 RTMP formats can also have the additional fields: page_url,
219 app, play_path, tc_url, flash_version, rtmp_live, rtmp_conn,
220 rtmp_protocol, rtmp_real_time
222 url: Final video URL.
223 ext: Video filename extension.
224 format: The video format, defaults to ext (used for --get-format)
225 player_url: SWF Player URL (used for rtmpdump).
227 The following fields are optional:
229 alt_title: A secondary title of the video.
230 display_id An alternative identifier for the video, not necessarily
231 unique, but available before title. Typically, id is
232 something like "4234987", title "Dancing naked mole rats",
233 and display_id "dancing-naked-mole-rats"
234 thumbnails: A list of dictionaries, with the following entries:
235 * "id" (optional, string) - Thumbnail format ID
237 * "preference" (optional, int) - quality of the image
238 * "width" (optional, int)
239 * "height" (optional, int)
240 * "resolution" (optional, string "{width}x{height}",
242 * "filesize" (optional, int)
243 * "http_headers" (dict) - HTTP headers for the request
244 thumbnail: Full URL to a video thumbnail image.
245 description: Full video description.
246 uploader: Full name of the video uploader.
247 license: License name the video is licensed under.
248 creator: The creator of the video.
249 timestamp: UNIX timestamp of the moment the video was uploaded
250 upload_date: Video upload date (YYYYMMDD).
251 If not explicitly set, calculated from timestamp
252 release_timestamp: UNIX timestamp of the moment the video was released.
253 If it is not clear whether to use timestamp or this, use the former
254 release_date: The date (YYYYMMDD) when the video was released.
255 If not explicitly set, calculated from release_timestamp
256 modified_timestamp: UNIX timestamp of the moment the video was last modified.
257 modified_date: The date (YYYYMMDD) when the video was last modified.
258 If not explicitly set, calculated from modified_timestamp
259 uploader_id: Nickname or id of the video uploader.
260 uploader_url: Full URL to a personal webpage of the video uploader.
261 channel: Full name of the channel the video is uploaded on.
262 Note that channel fields may or may not repeat uploader
263 fields. This depends on a particular extractor.
264 channel_id: Id of the channel.
265 channel_url: Full URL to a channel webpage.
266 channel_follower_count: Number of followers of the channel.
267 location: Physical location where the video was filmed.
268 subtitles: The available subtitles as a dictionary in the format
269 {tag: subformats}. "tag" is usually a language code, and
270 "subformats" is a list sorted from lower to higher
271 preference, each element is a dictionary with the "ext"
273 * "data": The subtitles file contents
274 * "url": A URL pointing to the subtitles file
275 It can optionally also have:
276 * "name": Name or description of the subtitles
277 * http_headers: A dictionary of additional HTTP headers
278 to add to the request.
279 "ext" will be calculated from URL if missing
280 automatic_captions: Like 'subtitles'; contains automatically generated
281 captions instead of normal subtitles
282 duration: Length of the video in seconds, as an integer or float.
283 view_count: How many users have watched the video on the platform.
284 like_count: Number of positive ratings of the video
285 dislike_count: Number of negative ratings of the video
286 repost_count: Number of reposts of the video
287 average_rating: Average rating give by users, the scale used depends on the webpage
288 comment_count: Number of comments on the video
289 comments: A list of comments, each with one or more of the following
290 properties (all but one of text or html optional):
291 * "author" - human-readable name of the comment author
292 * "author_id" - user ID of the comment author
293 * "author_thumbnail" - The thumbnail of the comment author
295 * "html" - Comment as HTML
296 * "text" - Plain text of the comment
297 * "timestamp" - UNIX timestamp of comment
298 * "parent" - ID of the comment this one is replying to.
299 Set to "root" to indicate that this is a
300 comment to the original video.
301 * "like_count" - Number of positive ratings of the comment
302 * "dislike_count" - Number of negative ratings of the comment
303 * "is_favorited" - Whether the comment is marked as
304 favorite by the video uploader
305 * "author_is_uploader" - Whether the comment is made by
307 age_limit: Age restriction for the video, as an integer (years)
308 webpage_url: The URL to the video webpage, if given to yt-dlp it
309 should allow to get the same result again. (It will be set
310 by YoutubeDL if it's missing)
311 categories: A list of categories that the video falls in, for example
313 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
314 cast: A list of the video cast
315 is_live: True, False, or None (=unknown). Whether this video is a
316 live stream that goes on instead of a fixed-length video.
317 was_live: True, False, or None (=unknown). Whether this video was
318 originally a live stream.
319 live_status: 'is_live', 'is_upcoming', 'was_live', 'not_live' or None (=unknown)
320 If absent, automatically set from is_live, was_live
321 start_time: Time in seconds where the reproduction should start, as
322 specified in the URL.
323 end_time: Time in seconds where the reproduction should end, as
324 specified in the URL.
325 chapters: A list of dictionaries, with the following entries:
326 * "start_time" - The start time of the chapter in seconds
327 * "end_time" - The end time of the chapter in seconds
328 * "title" (optional, string)
329 playable_in_embed: Whether this video is allowed to play in embedded
330 players on other sites. Can be True (=always allowed),
331 False (=never allowed), None (=unknown), or a string
332 specifying the criteria for embedability (Eg: 'whitelist')
333 availability: Under what condition the video is available. One of
334 'private', 'premium_only', 'subscriber_only', 'needs_auth',
335 'unlisted' or 'public'. Use 'InfoExtractor._availability'
337 __post_extractor: A function to be called just before the metadata is
338 written to either disk, logger or console. The function
339 must return a dict which will be added to the info_dict.
340 This is usefull for additional information that is
341 time-consuming to extract. Note that the fields thus
342 extracted will not be available to output template and
343 match_filter. So, only "comments" and "comment_count" are
344 currently allowed to be extracted via this method.
346 The following fields should only be used when the video belongs to some logical
349 chapter: Name or title of the chapter the video belongs to.
350 chapter_number: Number of the chapter the video belongs to, as an integer.
351 chapter_id: Id of the chapter the video belongs to, as a unicode string.
353 The following fields should only be used when the video is an episode of some
354 series, programme or podcast:
356 series: Title of the series or programme the video episode belongs to.
357 series_id: Id of the series or programme the video episode belongs to, as a unicode string.
358 season: Title of the season the video episode belongs to.
359 season_number: Number of the season the video episode belongs to, as an integer.
360 season_id: Id of the season the video episode belongs to, as a unicode string.
361 episode: Title of the video episode. Unlike mandatory video title field,
362 this field should denote the exact title of the video episode
363 without any kind of decoration.
364 episode_number: Number of the video episode within a season, as an integer.
365 episode_id: Id of the video episode, as a unicode string.
367 The following fields should only be used when the media is a track or a part of
370 track: Title of the track.
371 track_number: Number of the track within an album or a disc, as an integer.
372 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
374 artist: Artist(s) of the track.
375 genre: Genre(s) of the track.
376 album: Title of the album the track belongs to.
377 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
378 album_artist: List of all artists appeared on the album (e.g.
379 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
381 disc_number: Number of the disc or other physical medium the track belongs to,
383 release_year: Year (YYYY) when the album was released.
384 composer: Composer of the piece
386 Unless mentioned otherwise, the fields should be Unicode strings.
388 Unless mentioned otherwise, None is equivalent to absence of information.
391 _type "playlist" indicates multiple videos.
392 There must be a key "entries", which is a list, an iterable, or a PagedList
393 object, each element of which is a valid dictionary by this specification.
395 Additionally, playlists can have "id", "title", and any other relevent
396 attributes with the same semantics as videos (see above).
398 It can also have the following optional fields:
400 playlist_count: The total number of videos in a playlist. If not given,
401 YoutubeDL tries to calculate it from "entries"
404 _type "multi_video" indicates that there are multiple videos that
405 form a single show, for examples multiple acts of an opera or TV episode.
406 It must have an entries key like a playlist and contain all the keys
407 required for a video at the same time.
410 _type "url" indicates that the video must be extracted from another
411 location, possibly by a different extractor. Its only required key is:
412 "url" - the next URL to extract.
413 The key "ie_key" can be set to the class name (minus the trailing "IE",
414 e.g. "Youtube") if the extractor class is known in advance.
415 Additionally, the dictionary may have any properties of the resolved entity
416 known in advance, for example "title" if the title of the referred video is
420 _type "url_transparent" entities have the same specification as "url", but
421 indicate that the given additional information is more precise than the one
422 associated with the resolved URL.
423 This is useful when a site employs a video service that hosts the video and
424 its technical metadata, but that video service does not embed a useful
425 title, description etc.
428 Subclasses of this one should re-define the _real_initialize() and
429 _real_extract() methods and define a _VALID_URL regexp.
430 Probably, they should also be added to the list of extractors.
432 Subclasses may also override suitable() if necessary, but ensure the function
433 signature is preserved and that this function imports everything it needs
434 (except other extractors), so that lazy_extractors works correctly
436 _GEO_BYPASS attribute may be set to False in order to disable
437 geo restriction bypass mechanisms for a particular extractor.
438 Though it won't disable explicit geo restriction bypass based on
439 country code provided with geo_bypass_country.
441 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
442 countries for this extractor. One of these countries will be used by
443 geo restriction bypass mechanism right away in order to bypass
444 geo restriction, of course, if the mechanism is not disabled.
446 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
447 IP blocks in CIDR notation for this extractor. One of these IP blocks
448 will be used by geo restriction bypass mechanism similarly
451 The _WORKING attribute should be set to False for broken IEs
452 in order to warn the users and skip the tests.
457 _x_forwarded_for_ip
= None
459 _GEO_COUNTRIES
= None
460 _GEO_IP_BLOCKS
= None
464 'any': 'Use --cookies, --username and --password, or --netrc to provide account credentials',
466 'Use --cookies-from-browser or --cookies for the authentication. '
467 'See https://github.com/ytdl-org/youtube-dl#how-do-i-pass-cookies-to-youtube-dl for how to manually pass cookies'),
468 'password': 'Use --username and --password, or --netrc to provide account credentials',
471 def __init__(self
, downloader
=None):
472 """Constructor. Receives an optional downloader (a YoutubeDL instance).
473 If a downloader is not passed during initialization,
474 it must be set using "set_downloader()" before "extract()" is called"""
476 self
._x
_forwarded
_for
_ip
= None
477 self
._printed
_messages
= set()
478 self
.set_downloader(downloader
)
481 def _match_valid_url(cls
, url
):
482 # This does not use has/getattr intentionally - we want to know whether
483 # we have cached the regexp for *this* class, whereas getattr would also
484 # match the superclass
485 if '_VALID_URL_RE' not in cls
.__dict
__:
486 if '_VALID_URL' not in cls
.__dict
__:
487 cls
._VALID
_URL
= cls
._make
_valid
_url
()
488 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
489 return cls
._VALID
_URL
_RE
.match(url
)
492 def suitable(cls
, url
):
493 """Receives a URL and returns True if suitable for this IE."""
494 # This function must import everything it needs (except other extractors),
495 # so that lazy_extractors works correctly
496 return cls
._match
_valid
_url
(url
) is not None
499 def _match_id(cls
, url
):
500 return cls
._match
_valid
_url
(url
).group('id')
503 def get_temp_id(cls
, url
):
505 return cls
._match
_id
(url
)
506 except (IndexError, AttributeError):
511 """Getter method for _WORKING."""
514 def initialize(self
):
515 """Initializes an instance (authentication, etc)."""
516 self
._printed
_messages
= set()
517 self
._initialize
_geo
_bypass
({
518 'countries': self
._GEO
_COUNTRIES
,
519 'ip_blocks': self
._GEO
_IP
_BLOCKS
,
522 self
._real
_initialize
()
525 def _initialize_geo_bypass(self
, geo_bypass_context
):
527 Initialize geo restriction bypass mechanism.
529 This method is used to initialize geo bypass mechanism based on faking
530 X-Forwarded-For HTTP header. A random country from provided country list
531 is selected and a random IP belonging to this country is generated. This
532 IP will be passed as X-Forwarded-For HTTP header in all subsequent
535 This method will be used for initial geo bypass mechanism initialization
536 during the instance initialization with _GEO_COUNTRIES and
539 You may also manually call it from extractor's code if geo bypass
540 information is not available beforehand (e.g. obtained during
541 extraction) or due to some other reason. In this case you should pass
542 this information in geo bypass context passed as first argument. It may
543 contain following fields:
545 countries: List of geo unrestricted countries (similar
547 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
548 (similar to _GEO_IP_BLOCKS)
551 if not self
._x
_forwarded
_for
_ip
:
553 # Geo bypass mechanism is explicitly disabled by user
554 if not self
.get_param('geo_bypass', True):
557 if not geo_bypass_context
:
558 geo_bypass_context
= {}
560 # Backward compatibility: previously _initialize_geo_bypass
561 # expected a list of countries, some 3rd party code may still use
563 if isinstance(geo_bypass_context
, (list, tuple)):
564 geo_bypass_context
= {
565 'countries': geo_bypass_context
,
568 # The whole point of geo bypass mechanism is to fake IP
569 # as X-Forwarded-For HTTP header based on some IP block or
572 # Path 1: bypassing based on IP block in CIDR notation
574 # Explicit IP block specified by user, use it right away
575 # regardless of whether extractor is geo bypassable or not
576 ip_block
= self
.get_param('geo_bypass_ip_block', None)
578 # Otherwise use random IP block from geo bypass context but only
579 # if extractor is known as geo bypassable
581 ip_blocks
= geo_bypass_context
.get('ip_blocks')
582 if self
._GEO
_BYPASS
and ip_blocks
:
583 ip_block
= random
.choice(ip_blocks
)
586 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(ip_block
)
587 self
._downloader
.write_debug(
588 '[debug] Using fake IP %s as X-Forwarded-For' % self
._x
_forwarded
_for
_ip
)
591 # Path 2: bypassing based on country code
593 # Explicit country code specified by user, use it right away
594 # regardless of whether extractor is geo bypassable or not
595 country
= self
.get_param('geo_bypass_country', None)
597 # Otherwise use random country code from geo bypass context but
598 # only if extractor is known as geo bypassable
600 countries
= geo_bypass_context
.get('countries')
601 if self
._GEO
_BYPASS
and countries
:
602 country
= random
.choice(countries
)
605 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country
)
606 self
._downloader
.write_debug(
607 'Using fake IP %s (%s) as X-Forwarded-For' % (self
._x
_forwarded
_for
_ip
, country
.upper()))
609 def extract(self
, url
):
610 """Extracts URL information and returns it in list of dicts."""
615 self
.write_debug('Extracting URL: %s' % url
)
616 ie_result
= self
._real
_extract
(url
)
617 if ie_result
is None:
619 if self
._x
_forwarded
_for
_ip
:
620 ie_result
['__x_forwarded_for_ip'] = self
._x
_forwarded
_for
_ip
621 subtitles
= ie_result
.get('subtitles')
622 if (subtitles
and 'live_chat' in subtitles
623 and 'no-live-chat' in self
.get_param('compat_opts', [])):
624 del subtitles
['live_chat']
626 except GeoRestrictedError
as e
:
627 if self
.__maybe
_fake
_ip
_and
_retry
(e
.countries
):
630 except UnsupportedError
:
632 except ExtractorError
as e
:
634 'video_id': e
.video_id
or self
.get_temp_id(url
),
636 'tb': e
.traceback
or sys
.exc_info()[2],
637 'expected': e
.expected
,
640 if hasattr(e
, 'countries'):
641 kwargs
['countries'] = e
.countries
642 raise type(e
)(e
.orig_msg
, **kwargs
)
643 except compat_http_client
.IncompleteRead
as e
:
644 raise ExtractorError('A network error has occurred.', cause
=e
, expected
=True, video_id
=self
.get_temp_id(url
))
645 except (KeyError, StopIteration) as e
:
646 raise ExtractorError('An extractor error has occurred.', cause
=e
, video_id
=self
.get_temp_id(url
))
648 def __maybe_fake_ip_and_retry(self
, countries
):
649 if (not self
.get_param('geo_bypass_country', None)
651 and self
.get_param('geo_bypass', True)
652 and not self
._x
_forwarded
_for
_ip
654 country_code
= random
.choice(countries
)
655 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country_code
)
656 if self
._x
_forwarded
_for
_ip
:
658 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
659 % (self
._x
_forwarded
_for
_ip
, country_code
.upper()))
663 def set_downloader(self
, downloader
):
664 """Sets the downloader for this IE."""
665 self
._downloader
= downloader
667 def _real_initialize(self
):
668 """Real initialization process. Redefine in subclasses."""
671 def _real_extract(self
, url
):
672 """Real extraction process. Redefine in subclasses."""
677 """A string for getting the InfoExtractor with get_info_extractor"""
678 return cls
.__name
__[:-2]
682 return compat_str(type(self
).__name
__[:-2])
685 def __can_accept_status_code(err
, expected_status
):
686 assert isinstance(err
, compat_urllib_error
.HTTPError
)
687 if expected_status
is None:
689 elif callable(expected_status
):
690 return expected_status(err
.code
) is True
692 return err
.code
in variadic(expected_status
)
694 def _request_webpage(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
, expected_status
=None):
696 Return the response handle.
698 See _download_webpage docstring for arguments specification.
700 if not self
._downloader
._first
_webpage
_request
:
701 sleep_interval
= self
.get_param('sleep_interval_requests') or 0
702 if sleep_interval
> 0:
703 self
.to_screen('Sleeping %s seconds ...' % sleep_interval
)
704 time
.sleep(sleep_interval
)
706 self
._downloader
._first
_webpage
_request
= False
709 self
.report_download_webpage(video_id
)
710 elif note
is not False:
712 self
.to_screen('%s' % (note
,))
714 self
.to_screen('%s: %s' % (video_id
, note
))
716 # Some sites check X-Forwarded-For HTTP header in order to figure out
717 # the origin of the client behind proxy. This allows bypassing geo
718 # restriction by faking this header's value to IP that belongs to some
719 # geo unrestricted country. We will do so once we encounter any
720 # geo restriction error.
721 if self
._x
_forwarded
_for
_ip
:
722 if 'X-Forwarded-For' not in headers
:
723 headers
['X-Forwarded-For'] = self
._x
_forwarded
_for
_ip
725 if isinstance(url_or_request
, compat_urllib_request
.Request
):
726 url_or_request
= update_Request(
727 url_or_request
, data
=data
, headers
=headers
, query
=query
)
730 url_or_request
= update_url_query(url_or_request
, query
)
731 if data
is not None or headers
:
732 url_or_request
= sanitized_Request(url_or_request
, data
, headers
)
734 return self
._downloader
.urlopen(url_or_request
)
735 except network_exceptions
as err
:
736 if isinstance(err
, compat_urllib_error
.HTTPError
):
737 if self
.__can
_accept
_status
_code
(err
, expected_status
):
738 # Retain reference to error to prevent file object from
739 # being closed before it can be read. Works around the
740 # effects of <https://bugs.python.org/issue15002>
741 # introduced in Python 3.4.1.
748 errnote
= 'Unable to download webpage'
750 errmsg
= '%s: %s' % (errnote
, error_to_compat_str(err
))
752 raise ExtractorError(errmsg
, sys
.exc_info()[2], cause
=err
)
754 self
.report_warning(errmsg
)
757 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):
759 Return a tuple (page content as string, URL handle).
761 See _download_webpage docstring for arguments specification.
763 # Strip hashes from the URL (#1038)
764 if isinstance(url_or_request
, (compat_str
, str)):
765 url_or_request
= url_or_request
.partition('#')[0]
767 urlh
= self
._request
_webpage
(url_or_request
, video_id
, note
, errnote
, fatal
, data
=data
, headers
=headers
, query
=query
, expected_status
=expected_status
)
771 content
= self
._webpage
_read
_content
(urlh
, url_or_request
, video_id
, note
, errnote
, fatal
, encoding
=encoding
)
772 return (content
, urlh
)
775 def _guess_encoding_from_content(content_type
, webpage_bytes
):
776 m
= re
.match(r
'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type
)
778 encoding
= m
.group(1)
780 m
= re
.search(br
'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
781 webpage_bytes[:1024])
783 encoding = m.group(1).decode('ascii')
784 elif webpage_bytes.startswith(b'\xff\xfe'):
791 def __check_blocked(self, content):
792 first_block = content[:512]
793 if ('<title>Access to this site is blocked</title>' in content
794 and 'Websense' in first_block):
795 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
796 blocked_iframe = self._html_search_regex(
797 r'<iframe src="([^
"]+)"', content,
798 'Websense information URL
', default=None)
800 msg += ' Visit
%s for more details
' % blocked_iframe
801 raise ExtractorError(msg, expected=True)
802 if '<title
>The URL you requested has been blocked
</title
>' in first_block:
804 'Access to this webpage has been blocked by Indian censorship
. '
805 'Use a VPN
or proxy
server (with --proxy
) to route around it
.')
806 block_msg = self._html_search_regex(
807 r'</h1
><p
>(.*?
)</p
>',
808 content, 'block message
', default=None)
810 msg += ' (Message
: "%s")' % block_msg.replace('\n', ' ')
811 raise ExtractorError(msg, expected=True)
812 if ('<title
>TTK
:: Доступ к ресурсу ограничен
</title
>' in content
813 and 'blocklist
.rkn
.gov
.ru
' in content):
814 raise ExtractorError(
815 'Access to this webpage has been blocked by decision of the Russian government
. '
816 'Visit http
://blocklist
.rkn
.gov
.ru
/ for a block reason
.',
819 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
820 content_type = urlh.headers.get('Content
-Type
', '')
821 webpage_bytes = urlh.read()
822 if prefix is not None:
823 webpage_bytes = prefix + webpage_bytes
825 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
826 if self.get_param('dump_intermediate_pages
', False):
827 self.to_screen('Dumping request to
' + urlh.geturl())
828 dump = base64.b64encode(webpage_bytes).decode('ascii
')
829 self._downloader.to_screen(dump)
830 if self.get_param('write_pages
', False):
831 basen = '%s_%s' % (video_id, urlh.geturl())
832 trim_length = self.get_param('trim_file_name
') or 240
833 if len(basen) > trim_length:
834 h = '___
' + hashlib.md5(basen.encode('utf
-8')).hexdigest()
835 basen = basen[:trim_length - len(h)] + h
836 raw_filename = basen + '.dump
'
837 filename = sanitize_filename(raw_filename, restricted=True)
838 self.to_screen('Saving request to
' + filename)
839 # Working around MAX_PATH limitation on Windows (see
840 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
841 if compat_os_name == 'nt
':
842 absfilepath = os.path.abspath(filename)
843 if len(absfilepath) > 259:
844 filename = '\\\\?
\\' + absfilepath
845 with open(filename, 'wb
') as outf:
846 outf.write(webpage_bytes)
849 content = webpage_bytes.decode(encoding, 'replace
')
851 content = webpage_bytes.decode('utf
-8', 'replace
')
853 self.__check_blocked(content)
857 def _download_webpage(
858 self, url_or_request, video_id, note=None, errnote=None,
859 fatal=True, tries=1, timeout=5, encoding=None, data=None,
860 headers={}, query={}, expected_status=None):
862 Return the data of the page as a string.
865 url_or_request -- plain text URL as a string or
866 a compat_urllib_request.Requestobject
867 video_id -- Video/playlist/item identifier (string)
870 note -- note printed before downloading (string)
871 errnote -- note printed in case of an error (string)
872 fatal -- flag denoting whether error should be considered fatal,
873 i.e. whether it should cause ExtractionError to be raised,
874 otherwise a warning will be reported and extraction continued
875 tries -- number of tries
876 timeout -- sleep interval between tries
877 encoding -- encoding for a page content decoding, guessed automatically
878 when not explicitly specified
879 data -- POST data (bytes)
880 headers -- HTTP headers (dict)
881 query -- URL query (dict)
882 expected_status -- allows to accept failed HTTP requests (non 2xx
883 status code) by explicitly specifying a set of accepted status
884 codes. Can be any of the following entities:
885 - an integer type specifying an exact failed status code to
887 - a list or a tuple of integer types specifying a list of
888 failed status codes to accept
889 - a callable accepting an actual failed status code and
890 returning True if it should be accepted
891 Note that this argument does not affect success status codes (2xx)
892 which are always accepted.
897 while success is False:
899 res = self._download_webpage_handle(
900 url_or_request, video_id, note, errnote, fatal,
901 encoding=encoding, data=data, headers=headers, query=query,
902 expected_status=expected_status)
904 except compat_http_client.IncompleteRead as e:
906 if try_count >= tries:
908 self._sleep(timeout, video_id)
915 def _download_xml_handle(
916 self, url_or_request, video_id, note='Downloading XML
',
917 errnote='Unable to download XML
', transform_source=None,
918 fatal=True, encoding=None, data=None, headers={}, query={},
919 expected_status=None):
921 Return a tuple (xml as an compat_etree_Element, URL handle).
923 See _download_webpage docstring for arguments specification.
925 res = self._download_webpage_handle(
926 url_or_request, video_id, note, errnote, fatal=fatal,
927 encoding=encoding, data=data, headers=headers, query=query,
928 expected_status=expected_status)
931 xml_string, urlh = res
932 return self._parse_xml(
933 xml_string, video_id, transform_source=transform_source,
937 self, url_or_request, video_id,
938 note='Downloading XML
', errnote='Unable to download XML
',
939 transform_source=None, fatal=True, encoding=None,
940 data=None, headers={}, query={}, expected_status=None):
942 Return the xml as an compat_etree_Element.
944 See _download_webpage docstring for arguments specification.
946 res = self._download_xml_handle(
947 url_or_request, video_id, note=note, errnote=errnote,
948 transform_source=transform_source, fatal=fatal, encoding=encoding,
949 data=data, headers=headers, query=query,
950 expected_status=expected_status)
951 return res if res is False else res[0]
953 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
955 xml_string = transform_source(xml_string)
957 return compat_etree_fromstring(xml_string.encode('utf
-8'))
958 except compat_xml_parse_error as ve:
959 errmsg = '%s: Failed to parse XML
' % video_id
961 raise ExtractorError(errmsg, cause=ve)
963 self.report_warning(errmsg + str(ve))
965 def _download_json_handle(
966 self, url_or_request, video_id, note='Downloading JSON metadata
',
967 errnote='Unable to download JSON metadata
', transform_source=None,
968 fatal=True, encoding=None, data=None, headers={}, query={},
969 expected_status=None):
971 Return a tuple (JSON object, URL handle).
973 See _download_webpage docstring for arguments specification.
975 res = self._download_webpage_handle(
976 url_or_request, video_id, note, errnote, fatal=fatal,
977 encoding=encoding, data=data, headers=headers, query=query,
978 expected_status=expected_status)
981 json_string, urlh = res
982 return self._parse_json(
983 json_string, video_id, transform_source=transform_source,
987 self, url_or_request, video_id, note='Downloading JSON metadata
',
988 errnote='Unable to download JSON metadata
', transform_source=None,
989 fatal=True, encoding=None, data=None, headers={}, query={},
990 expected_status=None):
992 Return the JSON object as a dict.
994 See _download_webpage docstring for arguments specification.
996 res = self._download_json_handle(
997 url_or_request, video_id, note=note, errnote=errnote,
998 transform_source=transform_source, fatal=fatal, encoding=encoding,
999 data=data, headers=headers, query=query,
1000 expected_status=expected_status)
1001 return res if res is False else res[0]
1003 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
1004 if transform_source:
1005 json_string = transform_source(json_string)
1007 return json.loads(json_string)
1008 except ValueError as ve:
1009 errmsg = '%s: Failed to parse JSON
' % video_id
1011 raise ExtractorError(errmsg, cause=ve)
1013 self.report_warning(errmsg + str(ve))
1015 def _parse_socket_response_as_json(self, data, video_id, transform_source=None, fatal=True):
1016 return self._parse_json(
1017 data[data.find('{'):data.rfind('}
') + 1],
1018 video_id, transform_source, fatal)
1020 def _download_socket_json_handle(
1021 self, url_or_request, video_id, note='Polling socket
',
1022 errnote='Unable to poll socket
', transform_source=None,
1023 fatal=True, encoding=None, data=None, headers={}, query={},
1024 expected_status=None):
1026 Return a tuple (JSON object, URL handle).
1028 See _download_webpage docstring for arguments specification.
1030 res = self._download_webpage_handle(
1031 url_or_request, video_id, note, errnote, fatal=fatal,
1032 encoding=encoding, data=data, headers=headers, query=query,
1033 expected_status=expected_status)
1037 return self._parse_socket_response_as_json(
1038 webpage, video_id, transform_source=transform_source,
1041 def _download_socket_json(
1042 self, url_or_request, video_id, note='Polling socket
',
1043 errnote='Unable to poll socket
', transform_source=None,
1044 fatal=True, encoding=None, data=None, headers={}, query={},
1045 expected_status=None):
1047 Return the JSON object as a dict.
1049 See _download_webpage docstring for arguments specification.
1051 res = self._download_socket_json_handle(
1052 url_or_request, video_id, note=note, errnote=errnote,
1053 transform_source=transform_source, fatal=fatal, encoding=encoding,
1054 data=data, headers=headers, query=query,
1055 expected_status=expected_status)
1056 return res if res is False else res[0]
1058 def report_warning(self, msg, video_id=None, *args, only_once=False, **kwargs):
1059 idstr = format_field(video_id, template='%s: ')
1060 msg = f'[{self.IE_NAME}
] {idstr}{msg}
'
1062 if f'WARNING
: {msg}
' in self._printed_messages:
1064 self._printed_messages.add(f'WARNING
: {msg}
')
1065 self._downloader.report_warning(msg, *args, **kwargs)
1067 def to_screen(self, msg, *args, **kwargs):
1068 """Print msg to screen, prefixing it with '[ie_name
]'"""
1069 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg), *args, **kwargs)
1071 def write_debug(self, msg, *args, **kwargs):
1072 self._downloader.write_debug('[%s] %s' % (self.IE_NAME, msg), *args, **kwargs)
1074 def get_param(self, name, default=None, *args, **kwargs):
1075 if self._downloader:
1076 return self._downloader.params.get(name, default, *args, **kwargs)
1079 def report_drm(self, video_id, partial=False):
1080 self.raise_no_formats('This video
is DRM protected
', expected=True, video_id=video_id)
1082 def report_extraction(self, id_or_name):
1083 """Report information extraction."""
1084 self.to_screen('%s: Extracting information
' % id_or_name)
1086 def report_download_webpage(self, video_id):
1087 """Report webpage download."""
1088 self.to_screen('%s: Downloading webpage
' % video_id)
1090 def report_age_confirmation(self):
1091 """Report attempt to confirm age."""
1092 self.to_screen('Confirming age
')
1094 def report_login(self):
1095 """Report attempt to log in."""
1096 self.to_screen('Logging
in')
1098 def raise_login_required(
1099 self, msg='This video
is only available
for registered users
',
1100 metadata_available=False, method='any
'):
1101 if metadata_available and (
1102 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1103 self.report_warning(msg)
1105 if method is not None:
1106 msg = '%s. %s' % (msg, self._LOGIN_HINTS[method])
1107 raise ExtractorError(msg, expected=True)
1109 def raise_geo_restricted(
1110 self, msg='This video
is not available
from your location due to geo restriction
',
1111 countries=None, metadata_available=False):
1112 if metadata_available and (
1113 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1114 self.report_warning(msg)
1116 raise GeoRestrictedError(msg, countries=countries)
1118 def raise_no_formats(self, msg, expected=False, video_id=None):
1120 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1121 self.report_warning(msg, video_id)
1122 elif isinstance(msg, ExtractorError):
1125 raise ExtractorError(msg, expected=expected, video_id=video_id)
1127 # Methods for following #608
1129 def url_result(url, ie=None, video_id=None, video_title=None, *, url_transparent=False, **kwargs):
1130 """Returns a URL that points to a page that should be processed"""
1132 kwargs['ie_key
'] = ie if isinstance(ie, str) else ie.ie_key()
1133 if video_id is not None:
1134 kwargs['id'] = video_id
1135 if video_title is not None:
1136 kwargs['title
'] = video_title
1139 '_type
': 'url_transparent
' if url_transparent else 'url
',
1143 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None, video_kwargs=None, **kwargs):
1144 urls = (self.url_result(self._proto_relative_url(m), ie, **(video_kwargs or {}))
1145 for m in orderedSet(map(getter, matches) if getter else matches))
1146 return self.playlist_result(urls, playlist_id, playlist_title, **kwargs)
1149 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None, *, multi_video=False, **kwargs):
1150 """Returns a playlist"""
1152 kwargs['id'] = playlist_id
1154 kwargs['title
'] = playlist_title
1155 if playlist_description is not None:
1156 kwargs['description
'] = playlist_description
1159 '_type
': 'multi_video
' if multi_video else 'playlist
',
1163 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1165 Perform a regex search on the given string, using a single or a list of
1166 patterns returning the first matching group.
1167 In case of failure return a default value or raise a WARNING or a
1168 RegexNotFoundError, depending on fatal, specifying the field name.
1170 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
1171 mobj = re.search(pattern, string, flags)
1174 mobj = re.search(p, string, flags)
1178 _name = self._downloader._format_err(name, self._downloader.Styles.EMPHASIS)
1182 # return the first matching group
1183 return next(g for g in mobj.groups() if g is not None)
1184 elif isinstance(group, (list, tuple)):
1185 return tuple(mobj.group(g) for g in group)
1187 return mobj.group(group)
1188 elif default is not NO_DEFAULT:
1191 raise RegexNotFoundError('Unable to extract
%s' % _name)
1193 self.report_warning('unable to extract
%s' % _name + bug_reports_message())
1196 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1198 Like _search_regex, but strips HTML tags and unescapes entities.
1200 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1202 return clean_html(res).strip()
1206 def _get_netrc_login_info(self, netrc_machine=None):
1209 netrc_machine = netrc_machine or self._NETRC_MACHINE
1211 if self.get_param('usenetrc
', False):
1213 netrc_file = compat_expanduser(self.get_param('netrc_location
') or '~
')
1214 if os.path.isdir(netrc_file):
1215 netrc_file = os.path.join(netrc_file, '.netrc
')
1216 info = netrc.netrc(file=netrc_file).authenticators(netrc_machine)
1217 if info is not None:
1221 raise netrc.NetrcParseError(
1222 'No authenticators
for %s' % netrc_machine)
1223 except (IOError, netrc.NetrcParseError) as err:
1224 self.report_warning(
1225 'parsing
.netrc
: %s' % error_to_compat_str(err))
1227 return username, password
1229 def _get_login_info(self, username_option='username
', password_option='password
', netrc_machine=None):
1231 Get the login info as (username, password)
1232 First look for the manually specified credentials using username_option
1233 and password_option as keys in params dictionary. If no such credentials
1234 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1236 If there's no info available
, return (None, None)
1239 # Attempt to use provided username and password or .netrc data
1240 username = self.get_param(username_option)
1241 if username is not None:
1242 password = self.get_param(password_option)
1244 username, password = self._get_netrc_login_info(netrc_machine)
1246 return username, password
1248 def _get_tfa_info(self, note='two-factor verification code'):
1250 Get the two
-factor authentication info
1251 TODO
- asking the user will be required
for sms
/phone verify
1252 currently just uses the command line option
1253 If there
's no info available, return None
1256 tfa = self.get_param('twofactor
')
1260 return compat_getpass('Type
%s and press
[Return
]: ' % note)
1262 # Helper functions for extracting OpenGraph info
1264 def _og_regexes(prop):
1265 content_re = r'content
=(?
:"([^"]+?
)"|\'([^\']+?)\'|\s*([^\s"\'=<>`
]+?
))'
1266 property_re = (r'(?
:name|
property)=(?
:\'og
[:-]%(prop)s\'|
"og[:-]%(prop)s"|\s
*og
[:-]%(prop)s\b)'
1267 % {'prop': re.escape(prop)})
1268 template = r'<meta
[^
>]+?
%s[^
>]+?
%s'
1270 template % (property_re, content_re),
1271 template % (content_re, property_re),
1275 def _meta_regex(prop):
1276 return r'''(?isx)<meta
1277 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1278 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1280 def _og_search_property(self, prop, html, name=None, **kargs):
1281 prop = variadic(prop)
1283 name = 'OpenGraph
%s' % prop[0]
1286 og_regexes.extend(self._og_regexes(p))
1287 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1290 return unescapeHTML(escaped)
1292 def _og_search_thumbnail(self, html, **kargs):
1293 return self._og_search_property('image
', html, 'thumbnail URL
', fatal=False, **kargs)
1295 def _og_search_description(self, html, **kargs):
1296 return self._og_search_property('description
', html, fatal=False, **kargs)
1298 def _og_search_title(self, html, **kargs):
1299 kargs.setdefault('fatal
', False)
1300 return self._og_search_property('title
', html, **kargs)
1302 def _og_search_video_url(self, html, name='video url
', secure=True, **kargs):
1303 regexes = self._og_regexes('video
') + self._og_regexes('video
:url
')
1305 regexes = self._og_regexes('video
:secure_url
') + regexes
1306 return self._html_search_regex(regexes, html, name, **kargs)
1308 def _og_search_url(self, html, **kargs):
1309 return self._og_search_property('url
', html, **kargs)
1311 def _html_extract_title(self, html, name, **kwargs):
1312 return self._html_search_regex(
1313 r'(?s
)<title
>(.*?
)</title
>', html, name, **kwargs)
1315 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1316 name = variadic(name)
1317 if display_name is None:
1318 display_name = name[0]
1319 return self._html_search_regex(
1320 [self._meta_regex(n) for n in name],
1321 html, display_name, fatal=fatal, group='content
', **kwargs)
1323 def _dc_search_uploader(self, html):
1324 return self._html_search_meta('dc
.creator
', html, 'uploader
')
1326 def _rta_search(self, html):
1327 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1328 if re.search(r'(?ix
)<meta\s
+name
="rating"\s
+'
1329 r' content
="RTA-5042-1996-1400-1577-RTA"',
1334 def _media_rating_search(self, html):
1335 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1336 rating = self._html_search_meta('rating
', html)
1348 return RATING_TABLE.get(rating.lower())
1350 def _family_friendly_search(self, html):
1351 # See http://schema.org/VideoObject
1352 family_friendly = self._html_search_meta(
1353 'isFamilyFriendly
', html, default=None)
1355 if not family_friendly:
1364 return RATING_TABLE.get(family_friendly.lower())
1366 def _twitter_search_player(self, html):
1367 return self._html_search_meta('twitter
:player
', html,
1368 'twitter card player
')
1370 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
1371 json_ld_list = list(re.finditer(JSON_LD_RE, html))
1372 default = kwargs.get('default
', NO_DEFAULT)
1373 # JSON-LD may be malformed and thus `fatal` should be respected.
1374 # At the same time `default` may be passed that assumes `fatal=False`
1375 # for _search_regex. Let's simulate the same behavior here
as well
.
1376 fatal
= kwargs
.get('fatal', True) if default
is NO_DEFAULT
else False
1378 for mobj
in json_ld_list
:
1379 json_ld_item
= self
._parse
_json
(
1380 mobj
.group('json_ld'), video_id
, fatal
=fatal
)
1381 if not json_ld_item
:
1383 if isinstance(json_ld_item
, dict):
1384 json_ld
.append(json_ld_item
)
1385 elif isinstance(json_ld_item
, (list, tuple)):
1386 json_ld
.extend(json_ld_item
)
1388 json_ld
= self
._json
_ld
(json_ld
, video_id
, fatal
=fatal
, expected_type
=expected_type
)
1391 if default
is not NO_DEFAULT
:
1394 raise RegexNotFoundError('Unable to extract JSON-LD')
1396 self
.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1399 def _json_ld(self
, json_ld
, video_id
, fatal
=True, expected_type
=None):
1400 if isinstance(json_ld
, compat_str
):
1401 json_ld
= self
._parse
_json
(json_ld
, video_id
, fatal
=fatal
)
1405 if not isinstance(json_ld
, (list, tuple, dict)):
1407 if isinstance(json_ld
, dict):
1410 INTERACTION_TYPE_MAP
= {
1411 'CommentAction': 'comment',
1412 'AgreeAction': 'like',
1413 'DisagreeAction': 'dislike',
1414 'LikeAction': 'like',
1415 'DislikeAction': 'dislike',
1416 'ListenAction': 'view',
1417 'WatchAction': 'view',
1418 'ViewAction': 'view',
1421 def extract_interaction_type(e
):
1422 interaction_type
= e
.get('interactionType')
1423 if isinstance(interaction_type
, dict):
1424 interaction_type
= interaction_type
.get('@type')
1425 return str_or_none(interaction_type
)
1427 def extract_interaction_statistic(e
):
1428 interaction_statistic
= e
.get('interactionStatistic')
1429 if isinstance(interaction_statistic
, dict):
1430 interaction_statistic
= [interaction_statistic
]
1431 if not isinstance(interaction_statistic
, list):
1433 for is_e
in interaction_statistic
:
1434 if not isinstance(is_e
, dict):
1436 if is_e
.get('@type') != 'InteractionCounter':
1438 interaction_type
= extract_interaction_type(is_e
)
1439 if not interaction_type
:
1441 # For interaction count some sites provide string instead of
1442 # an integer (as per spec) with non digit characters (e.g. ",")
1443 # so extracting count with more relaxed str_to_int
1444 interaction_count
= str_to_int(is_e
.get('userInteractionCount'))
1445 if interaction_count
is None:
1447 count_kind
= INTERACTION_TYPE_MAP
.get(interaction_type
.split('/')[-1])
1450 count_key
= '%s_count' % count_kind
1451 if info
.get(count_key
) is not None:
1453 info
[count_key
] = interaction_count
1455 def extract_chapter_information(e
):
1457 'title': part
.get('name'),
1458 'start_time': part
.get('startOffset'),
1459 'end_time': part
.get('endOffset'),
1460 } for part
in variadic(e
.get('hasPart') or []) if part
.get('@type') == 'Clip']
1461 for idx
, (last_c
, current_c
, next_c
) in enumerate(zip(
1462 [{'end_time': 0}
] + chapters
, chapters
, chapters
[1:])):
1463 current_c
['end_time'] = current_c
['end_time'] or next_c
['start_time']
1464 current_c
['start_time'] = current_c
['start_time'] or last_c
['end_time']
1465 if None in current_c
.values():
1466 self
.report_warning(f
'Chapter {idx} contains broken data. Not extracting chapters')
1469 chapters
[-1]['end_time'] = chapters
[-1]['end_time'] or info
['duration']
1470 info
['chapters'] = chapters
1472 def extract_video_object(e
):
1473 assert e
['@type'] == 'VideoObject'
1474 author
= e
.get('author')
1476 'url': url_or_none(e
.get('contentUrl')),
1477 'title': unescapeHTML(e
.get('name')),
1478 'description': unescapeHTML(e
.get('description')),
1479 'thumbnails': [{'url': url_or_none(url)}
1480 for url
in variadic(traverse_obj(e
, 'thumbnailUrl', 'thumbnailURL'))],
1481 'duration': parse_duration(e
.get('duration')),
1482 'timestamp': unified_timestamp(e
.get('uploadDate')),
1483 # author can be an instance of 'Organization' or 'Person' types.
1484 # both types can have 'name' property(inherited from 'Thing' type). [1]
1485 # however some websites are using 'Text' type instead.
1486 # 1. https://schema.org/VideoObject
1487 'uploader': author
.get('name') if isinstance(author
, dict) else author
if isinstance(author
, compat_str
) else None,
1488 'filesize': float_or_none(e
.get('contentSize')),
1489 'tbr': int_or_none(e
.get('bitrate')),
1490 'width': int_or_none(e
.get('width')),
1491 'height': int_or_none(e
.get('height')),
1492 'view_count': int_or_none(e
.get('interactionCount')),
1494 extract_interaction_statistic(e
)
1495 extract_chapter_information(e
)
1497 def traverse_json_ld(json_ld
, at_top_level
=True):
1499 if at_top_level
and '@context' not in e
:
1501 if at_top_level
and set(e
.keys()) == {'@context', '@graph'}
:
1502 traverse_json_ld(variadic(e
['@graph'], allowed_types
=(dict,)), at_top_level
=False)
1504 item_type
= e
.get('@type')
1505 if expected_type
is not None and expected_type
!= item_type
:
1507 rating
= traverse_obj(e
, ('aggregateRating', 'ratingValue'), expected_type
=float_or_none
)
1508 if rating
is not None:
1509 info
['average_rating'] = rating
1510 if item_type
in ('TVEpisode', 'Episode'):
1511 episode_name
= unescapeHTML(e
.get('name'))
1513 'episode': episode_name
,
1514 'episode_number': int_or_none(e
.get('episodeNumber')),
1515 'description': unescapeHTML(e
.get('description')),
1517 if not info
.get('title') and episode_name
:
1518 info
['title'] = episode_name
1519 part_of_season
= e
.get('partOfSeason')
1520 if isinstance(part_of_season
, dict) and part_of_season
.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1522 'season': unescapeHTML(part_of_season
.get('name')),
1523 'season_number': int_or_none(part_of_season
.get('seasonNumber')),
1525 part_of_series
= e
.get('partOfSeries') or e
.get('partOfTVSeries')
1526 if isinstance(part_of_series
, dict) and part_of_series
.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1527 info
['series'] = unescapeHTML(part_of_series
.get('name'))
1528 elif item_type
== 'Movie':
1530 'title': unescapeHTML(e
.get('name')),
1531 'description': unescapeHTML(e
.get('description')),
1532 'duration': parse_duration(e
.get('duration')),
1533 'timestamp': unified_timestamp(e
.get('dateCreated')),
1535 elif item_type
in ('Article', 'NewsArticle'):
1537 'timestamp': parse_iso8601(e
.get('datePublished')),
1538 'title': unescapeHTML(e
.get('headline')),
1539 'description': unescapeHTML(e
.get('articleBody') or e
.get('description')),
1541 if traverse_obj(e
, ('video', 0, '@type')) == 'VideoObject':
1542 extract_video_object(e
['video'][0])
1543 elif item_type
== 'VideoObject':
1544 extract_video_object(e
)
1545 if expected_type
is None:
1549 video
= e
.get('video')
1550 if isinstance(video
, dict) and video
.get('@type') == 'VideoObject':
1551 extract_video_object(video
)
1552 if expected_type
is None:
1556 traverse_json_ld(json_ld
)
1558 return dict((k
, v
) for k
, v
in info
.items() if v
is not None)
1560 def _search_nextjs_data(self
, webpage
, video_id
, *, transform_source
=None, fatal
=True, **kw
):
1561 return self
._parse
_json
(
1563 r
'(?s)<script[^>]+id=[\'"]__NEXT_DATA__[\'"][^
>]*>([^
<]+)</script
>',
1564 webpage, 'next
.js data
', fatal=fatal, **kw),
1565 video_id, transform_source=transform_source, fatal=fatal)
1567 def _search_nuxt_data(self, webpage, video_id, context_name='__NUXT__
'):
1568 ''' Parses Nuxt.js metadata. This works as long as the function __NUXT__ invokes is a pure function. '''
1569 # not all website do this, but it can be changed
1570 # https://stackoverflow.com/questions/67463109/how-to-change-or-hide-nuxt-and-nuxt-keyword-in-page-source
1571 rectx = re.escape(context_name)
1572 js, arg_keys, arg_vals = self._search_regex(
1573 (r'<script
>window\
.%s=\
(function\
((?P
<arg_keys
>.*?
)\
)\{return\s(?P<js>\{.*?\}
)\
}\
((?P
<arg_vals
>.+?
)\
)\
);?
</script
>' % rectx,
1574 r'%s\
(.*?\
(function\
((?P
<arg_keys
>.*?
)\
)\{return\s(?P<js>\{.*?\}
)\
}\
((?P
<arg_vals
>.*?
)\
)' % rectx),
1575 webpage, context_name, group=['js
', 'arg_keys
', 'arg_vals
'])
1577 args = dict(zip(arg_keys.split(','), arg_vals.split(',')))
1579 for key, val in args.items():
1580 if val in ('undefined
', 'void
0'):
1583 return self._parse_json(js_to_json(js, args), video_id)['data
'][0]
1586 def _hidden_inputs(html):
1587 html = re.sub(r'<!--(?
:(?
!<!--).)*-->', '', html)
1589 for input in re.findall(r'(?i
)(<input[^
>]+>)', html):
1590 attrs = extract_attributes(input)
1593 if attrs.get('type') not in ('hidden
', 'submit
'):
1595 name = attrs.get('name
') or attrs.get('id')
1596 value = attrs.get('value
')
1597 if name and value is not None:
1598 hidden_inputs[name] = value
1599 return hidden_inputs
1601 def _form_hidden_inputs(self, form_id, html):
1602 form = self._search_regex(
1603 r'(?
is)<form
[^
>]+?
id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1604 html, '%s form' % form_id, group='form')
1605 return self._hidden_inputs(form)
1608 regex = r' *((?P<reverse>\+)?(?P<field>[a-zA-Z0-9_]+)((?P<separator>[~:])(?P<limit>.*?))?)? *$'
1610 default = ('hidden', 'aud_or_vid', 'hasvid', 'ie_pref', 'lang', 'quality',
1611 'res', 'fps', 'hdr:12', 'codec:vp9.2', 'size', 'br', 'asr',
1612 'proto', 'ext', 'hasaud', 'source', 'id') # These must not be aliases
1613 ytdl_default = ('hasaud', 'lang', 'quality', 'tbr', 'filesize', 'vbr',
1614 'height', 'width', 'proto', 'vext', 'abr', 'aext',
1615 'fps', 'fs_approx', 'source', 'id')
1618 'vcodec': {'type': 'ordered', 'regex': True,
1619 'order': ['av0?1', 'vp0?9.2', 'vp0?9', '[hx]265|he?vc?', '[hx]264|avc', 'vp0?8', 'mp4v|h263', 'theora', '', None, 'none']},
1620 'acodec': {'type': 'ordered', 'regex': True,
1621 'order': ['[af]lac', 'wav|aiff', 'opus', 'vorbis|ogg', 'aac', 'mp?4a?', 'mp3', 'e-?a?c-?3', 'ac-?3', 'dts', '', None, 'none']},
1622 'hdr': {'type': 'ordered', 'regex': True, 'field': 'dynamic_range',
1623 'order': ['dv', '(hdr)?12', r'(hdr)?10\+', '(hdr)?10', 'hlg', '', 'sdr', None]},
1624 'proto': {'type': 'ordered', 'regex': True, 'field': 'protocol',
1625 'order': ['(ht|f)tps', '(ht|f)tp$', 'm3u8.*', '.*dash', 'websocket_frag', 'rtmpe?', '', 'mms|rtsp', 'ws|websocket', 'f4']},
1626 'vext': {'type': 'ordered', 'field': 'video_ext',
1627 'order': ('mp4', 'webm', 'flv', '', 'none'),
1628 'order_free': ('webm', 'mp4', 'flv', '', 'none')},
1629 'aext': {'type': 'ordered', 'field': 'audio_ext',
1630 'order': ('m4a', 'aac', 'mp3', 'ogg', 'opus', 'webm', '', 'none'),
1631 'order_free': ('opus', 'ogg', 'webm', 'm4a', 'mp3', 'aac', '', 'none')},
1632 'hidden': {'visible': False, 'forced': True, 'type': 'extractor', 'max': -1000},
1633 'aud_or_vid': {'visible': False, 'forced': True, 'type': 'multiple',
1634 'field': ('vcodec', 'acodec'),
1635 'function': lambda it: int(any(v != 'none' for v in it))},
1636 'ie_pref': {'priority': True, 'type': 'extractor'},
1637 'hasvid': {'priority': True, 'field': 'vcodec', 'type': 'boolean', 'not_in_list': ('none',)},
1638 'hasaud': {'field': 'acodec', 'type': 'boolean', 'not_in_list': ('none',)},
1639 'lang': {'convert': 'float', 'field': 'language_preference', 'default': -1},
1640 'quality': {'convert': 'float', 'default': -1},
1641 'filesize': {'convert': 'bytes'},
1642 'fs_approx': {'convert': 'bytes', 'field': 'filesize_approx'},
1643 'id': {'convert': 'string', 'field': 'format_id'},
1644 'height': {'convert': 'float_none'},
1645 'width': {'convert': 'float_none'},
1646 'fps': {'convert': 'float_none'},
1647 'tbr': {'convert': 'float_none'},
1648 'vbr': {'convert': 'float_none'},
1649 'abr': {'convert': 'float_none'},
1650 'asr': {'convert': 'float_none'},
1651 'source': {'convert': 'float', 'field': 'source_preference', 'default': -1},
1653 'codec': {'type': 'combined', 'field': ('vcodec', 'acodec')},
1654 'br': {'type': 'combined', 'field': ('tbr', 'vbr', 'abr'), 'same_limit': True},
1655 'size': {'type': 'combined', 'same_limit': True, 'field': ('filesize', 'fs_approx')},
1656 'ext': {'type': 'combined', 'field': ('vext', 'aext')},
1657 'res': {'type': 'multiple', 'field': ('height', 'width'),
1658 'function': lambda it: (lambda l: min(l) if l else 0)(tuple(filter(None, it)))},
1660 # For compatibility with youtube-dl
1661 'format_id': {'type': 'alias', 'field': 'id'},
1662 'preference': {'type': 'alias', 'field': 'ie_pref'},
1663 'language_preference': {'type': 'alias', 'field': 'lang'},
1666 'dimension': {'type': 'alias', 'field': 'res'},
1667 'resolution': {'type': 'alias', 'field': 'res'},
1668 'extension': {'type': 'alias', 'field': 'ext'},
1669 'bitrate': {'type': 'alias', 'field': 'br'},
1670 'total_bitrate': {'type': 'alias', 'field': 'tbr'},
1671 'video_bitrate': {'type': 'alias', 'field': 'vbr'},
1672 'audio_bitrate': {'type': 'alias', 'field': 'abr'},
1673 'framerate': {'type': 'alias', 'field': 'fps'},
1674 'protocol': {'type': 'alias', 'field': 'proto'},
1675 'source_preference': {'type': 'alias', 'field': 'source'},
1676 'filesize_approx': {'type': 'alias', 'field': 'fs_approx'},
1677 'filesize_estimate': {'type': 'alias', 'field': 'size'},
1678 'samplerate': {'type': 'alias', 'field': 'asr'},
1679 'video_ext': {'type': 'alias', 'field': 'vext'},
1680 'audio_ext': {'type': 'alias', 'field': 'aext'},
1681 'video_codec': {'type': 'alias', 'field': 'vcodec'},
1682 'audio_codec': {'type': 'alias', 'field': 'acodec'},
1683 'video': {'type': 'alias', 'field': 'hasvid'},
1684 'has_video': {'type': 'alias', 'field': 'hasvid'},
1685 'audio': {'type': 'alias', 'field': 'hasaud'},
1686 'has_audio': {'type': 'alias', 'field': 'hasaud'},
1687 'extractor': {'type': 'alias', 'field': 'ie_pref'},
1688 'extractor_preference': {'type': 'alias', 'field': 'ie_pref'},
1691 def __init__(self, ie, field_preference):
1693 self.ydl = ie._downloader
1694 self.evaluate_params(self.ydl.params, field_preference)
1695 if ie.get_param('verbose'):
1696 self.print_verbose_info(self.ydl.write_debug)
1698 def _get_field_setting(self, field, key):
1699 if field not in self.settings:
1700 if key in ('forced', 'priority'):
1702 self.ydl.deprecation_warning(
1703 f'Using arbitrary fields ({field}) for format sorting is deprecated '
1704 'and may be removed in a future version')
1705 self.settings[field] = {}
1706 propObj = self.settings[field]
1707 if key not in propObj:
1708 type = propObj.get('type')
1710 default = 'preference' if type == 'extractor' else (field,) if type in ('combined', 'multiple') else field
1711 elif key == 'convert':
1712 default = 'order' if type == 'ordered' else 'float_string' if field else 'ignore'
1714 default = {'type': 'field', 'visible': True, 'order': [], 'not_in_list': (None,)}.get(key, None)
1715 propObj[key] = default
1718 def _resolve_field_value(self, field, value, convertNone=False):
1723 value = value.lower()
1724 conversion = self._get_field_setting(field, 'convert')
1725 if conversion == 'ignore':
1727 if conversion == 'string':
1729 elif conversion == 'float_none':
1730 return float_or_none(value)
1731 elif conversion == 'bytes':
1732 return FileDownloader.parse_bytes(value)
1733 elif conversion == 'order':
1734 order_list = (self._use_free_order and self._get_field_setting(field, 'order_free')) or self._get_field_setting(field, 'order')
1735 use_regex = self._get_field_setting(field, 'regex')
1736 list_length = len(order_list)
1737 empty_pos = order_list.index('') if '' in order_list else list_length + 1
1738 if use_regex and value is not None:
1739 for i, regex in enumerate(order_list):
1740 if regex and re.match(regex, value):
1741 return list_length - i
1742 return list_length - empty_pos # not in list
1743 else: # not regex or value = None
1744 return list_length - (order_list.index(value) if value in order_list else empty_pos)
1746 if value.isnumeric():
1749 self.settings[field]['convert'] = 'string'
1752 def evaluate_params(self, params, sort_extractor):
1753 self._use_free_order = params.get('prefer_free_formats', False)
1754 self._sort_user = params.get('format_sort', [])
1755 self._sort_extractor = sort_extractor
1757 def add_item(field, reverse, closest, limit_text):
1758 field = field.lower()
1759 if field in self._order:
1761 self._order.append(field)
1762 limit = self._resolve_field_value(field, limit_text)
1765 'closest': False if limit is None else closest,
1766 'limit_text': limit_text,
1768 if field in self.settings:
1769 self.settings[field].update(data)
1771 self.settings[field] = data
1774 tuple(field for field in self.default if self._get_field_setting(field, 'forced'))
1775 + (tuple() if params.get('format_sort_force', False)
1776 else tuple(field for field in self.default if self._get_field_setting(field, 'priority')))
1777 + tuple(self._sort_user) + tuple(sort_extractor) + self.default)
1779 for item in sort_list:
1780 match = re.match(self.regex, item)
1782 raise ExtractorError('Invalid format sort string "%s" given by extractor' % item)
1783 field = match.group('field')
1786 if self._get_field_setting(field, 'type') == 'alias':
1787 alias, field = field, self._get_field_setting(field, 'field')
1788 if alias not in ('format_id', 'preference', 'language_preference'):
1789 self.ydl.deprecation_warning(
1790 f'Format sorting alias {alias} is deprecated '
1791 f'and may be removed in a future version. Please use {field} instead')
1792 reverse = match.group('reverse') is not None
1793 closest = match.group('separator') == '~'
1794 limit_text = match.group('limit')
1796 has_limit = limit_text is not None
1797 has_multiple_fields = self._get_field_setting(field, 'type') == 'combined'
1798 has_multiple_limits = has_limit and has_multiple_fields and not self._get_field_setting(field, 'same_limit')
1800 fields = self._get_field_setting(field, 'field') if has_multiple_fields else (field,)
1801 limits = limit_text.split(':') if has_multiple_limits else (limit_text,) if has_limit else tuple()
1802 limit_count = len(limits)
1803 for (i, f) in enumerate(fields):
1804 add_item(f, reverse, closest,
1805 limits[i] if i < limit_count
1806 else limits[0] if has_limit and not has_multiple_limits
1809 def print_verbose_info(self, write_debug):
1811 write_debug('Sort order given by user: %s' % ', '.join(self._sort_user))
1812 if self._sort_extractor:
1813 write_debug('Sort order given by extractor: %s' % ', '.join(self._sort_extractor))
1814 write_debug('Formats sorted by: %s' % ', '.join(['%s%s%s' % (
1815 '+' if self._get_field_setting(field, 'reverse') else '', field,
1816 '%s%s(%s)' % ('~' if self._get_field_setting(field, 'closest') else ':',
1817 self._get_field_setting(field, 'limit_text'),
1818 self._get_field_setting(field, 'limit'))
1819 if self._get_field_setting(field, 'limit_text') is not None else '')
1820 for field in self._order if self._get_field_setting(field, 'visible')]))
1822 def _calculate_field_preference_from_value(self, format, field, type, value):
1823 reverse = self._get_field_setting(field, 'reverse')
1824 closest = self._get_field_setting(field, 'closest')
1825 limit = self._get_field_setting(field, 'limit')
1827 if type == 'extractor':
1828 maximum = self._get_field_setting(field, 'max')
1829 if value is None or (maximum is not None and value >= maximum):
1831 elif type == 'boolean':
1832 in_list = self._get_field_setting(field, 'in_list')
1833 not_in_list = self._get_field_setting(field, 'not_in_list')
1834 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
1835 elif type == 'ordered':
1836 value = self._resolve_field_value(field, value, True)
1838 # try to convert to number
1839 val_num = float_or_none(value, default=self._get_field_setting(field, 'default'))
1840 is_num = self._get_field_setting(field, 'convert') != 'string' and val_num is not None
1844 return ((-10, 0) if value is None
1845 else (1, value, 0) if not is_num # if a field has mixed strings and numbers, strings are sorted higher
1846 else (0, -abs(value - limit), value - limit if reverse else limit - value) if closest
1847 else (0, value, 0) if not reverse and (limit is None or value <= limit)
1848 else (0, -value, 0) if limit is None or (reverse and value == limit) or value > limit
1849 else (-1, value, 0))
1851 def _calculate_field_preference(self, format, field):
1852 type = self._get_field_setting(field, 'type') # extractor, boolean, ordered, field, multiple
1853 get_value = lambda f: format.get(self._get_field_setting(f, 'field'))
1854 if type == 'multiple':
1855 type = 'field' # Only 'field' is allowed in multiple for now
1856 actual_fields = self._get_field_setting(field, 'field')
1858 value = self._get_field_setting(field, 'function')(get_value(f) for f in actual_fields)
1860 value = get_value(field)
1861 return self._calculate_field_preference_from_value(format, field, type, value)
1863 def calculate_preference(self, format):
1864 # Determine missing protocol
1865 if not format.get('protocol'):
1866 format['protocol'] = determine_protocol(format)
1868 # Determine missing ext
1869 if not format.get('ext') and 'url' in format:
1870 format['ext'] = determine_ext(format['url'])
1871 if format.get('vcodec') == 'none':
1872 format['audio_ext'] = format['ext'] if format.get('acodec') != 'none' else 'none'
1873 format['video_ext'] = 'none'
1875 format['video_ext'] = format['ext']
1876 format['audio_ext'] = 'none'
1877 # if format.get('preference') is None and format.get('ext') in ('f4f', 'f4m'): # Not supported?
1878 # format['preference'] = -1000
1880 # Determine missing bitrates
1881 if format.get('tbr') is None:
1882 if format.get('vbr') is not None and format.get('abr') is not None:
1883 format['tbr'] = format.get('vbr', 0) + format.get('abr', 0)
1885 if format.get('vcodec') != 'none' and format.get('vbr') is None:
1886 format['vbr'] = format.get('tbr') - format.get('abr', 0)
1887 if format.get('acodec') != 'none' and format.get('abr') is None:
1888 format['abr'] = format.get('tbr') - format.get('vbr', 0)
1890 return tuple(self._calculate_field_preference(format, field) for field in self._order)
1892 def _sort_formats(self, formats, field_preference=[]):
1895 format_sort = self.FormatSort(self, field_preference)
1896 formats.sort(key=lambda f: format_sort.calculate_preference(f))
1898 def _check_formats(self, formats, video_id):
1900 formats[:] = filter(
1901 lambda f: self._is_valid_url(
1903 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1907 def _remove_duplicate_formats(formats):
1911 if f['url'] not in format_urls:
1912 format_urls.add(f['url'])
1913 unique_formats.append(f)
1914 formats[:] = unique_formats
1916 def _is_valid_url(self, url, video_id, item='video', headers={}):
1917 url = self._proto_relative_url(url, scheme='http:')
1918 # For now assume non HTTP(S) URLs always valid
1919 if not (url.startswith('http://') or url.startswith('https://')):
1922 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1924 except ExtractorError as e:
1926 '%s: %s URL is invalid, skipping: %s'
1927 % (video_id, item, error_to_compat_str(e.cause)))
1930 def http_scheme(self):
1931 """ Either "http
:" or "https
:", depending on the user's preferences """
1934 if self.get_param('prefer_insecure', False)
1937 def _proto_relative_url(self, url, scheme=None):
1940 if url.startswith('//'):
1942 scheme = self.http_scheme()
1947 def _sleep(self, timeout, video_id, msg_template=None):
1948 if msg_template is None:
1949 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1950 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1954 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1955 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1956 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
1957 manifest = self._download_xml(
1958 manifest_url, video_id, 'Downloading f4m manifest',
1959 'Unable to download f4m manifest',
1960 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1961 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
1962 transform_source=transform_source,
1963 fatal=fatal, data=data, headers=headers, query=query)
1965 if manifest is False:
1968 return self._parse_f4m_formats(
1969 manifest, manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1970 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1972 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1973 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1974 fatal=True, m3u8_id=None):
1975 if not isinstance(manifest, compat_etree_Element) and not fatal:
1978 # currently yt-dlp cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1979 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1980 if akamai_pv is not None and ';' in akamai_pv.text:
1981 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1982 if playerVerificationChallenge.strip() != '':
1986 manifest_version = '1.0'
1987 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1989 manifest_version = '2.0'
1990 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1991 # Remove unsupported DRM protected media from final formats
1992 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
1993 media_nodes = remove_encrypted_media(media_nodes)
1997 manifest_base_url = get_base_url(manifest)
1999 bootstrap_info = xpath_element(
2000 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
2001 'bootstrap info', default=None)
2004 mime_type = xpath_text(
2005 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
2006 'base URL', default=None)
2007 if mime_type and mime_type.startswith('audio/'):
2010 for i, media_el in enumerate(media_nodes):
2011 tbr = int_or_none(media_el.attrib.get('bitrate'))
2012 width = int_or_none(media_el.attrib.get('width'))
2013 height = int_or_none(media_el.attrib.get('height'))
2014 format_id = join_nonempty(f4m_id, tbr or i)
2015 # If <bootstrapInfo> is present, the specified f4m is a
2016 # stream-level manifest, and only set-level manifests may refer to
2017 # external resources. See section 11.4 and section 4 of F4M spec
2018 if bootstrap_info is None:
2020 # @href is introduced in 2.0, see section 11.6 of F4M spec
2021 if manifest_version == '2.0':
2022 media_url = media_el.attrib.get('href')
2023 if media_url is None:
2024 media_url = media_el.attrib.get('url')
2028 media_url if media_url.startswith('http://') or media_url.startswith('https://')
2029 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
2030 # If media_url is itself a f4m manifest do the recursive extraction
2031 # since bitrates in parent manifest (this one) and media_url manifest
2032 # may differ leading to inability to resolve the format by requested
2033 # bitrate in f4m downloader
2034 ext = determine_ext(manifest_url)
2036 f4m_formats = self._extract_f4m_formats(
2037 manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
2038 transform_source=transform_source, fatal=fatal)
2039 # Sometimes stream-level manifest contains single media entry that
2040 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
2041 # At the same time parent's media entry in set-level manifest may
2042 # contain it. We will copy it from parent in such cases.
2043 if len(f4m_formats) == 1:
2046 'tbr': f.get('tbr') or tbr,
2047 'width': f.get('width') or width,
2048 'height': f.get('height') or height,
2049 'format_id': f.get('format_id') if not tbr else format_id,
2052 formats.extend(f4m_formats)
2055 formats.extend(self._extract_m3u8_formats(
2056 manifest_url, video_id, 'mp4', preference=preference,
2057 quality=quality, m3u8_id=m3u8_id, fatal=fatal))
2060 'format_id': format_id,
2061 'url': manifest_url,
2062 'manifest_url': manifest_url,
2063 'ext': 'flv' if bootstrap_info is not None else None,
2069 'preference': preference,
2074 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, quality=None, m3u8_id=None):
2076 'format_id': join_nonempty(m3u8_id, 'meta'),
2080 'preference': preference - 100 if preference else -100,
2082 'resolution': 'multiple',
2083 'format_note': 'Quality selection URL',
2086 def _report_ignoring_subs(self, name):
2087 self.report_warning(bug_reports_message(
2088 f'Ignoring subtitle tracks found in the {name} manifest; '
2089 'if any subtitle tracks are missing,'
2092 def _extract_m3u8_formats(self, *args, **kwargs):
2093 fmts, subs = self._extract_m3u8_formats_and_subtitles(*args, **kwargs)
2095 self._report_ignoring_subs('HLS')
2098 def _extract_m3u8_formats_and_subtitles(
2099 self, m3u8_url, video_id, ext=None, entry_protocol='m3u8_native',
2100 preference=None, quality=None, m3u8_id=None, note=None,
2101 errnote=None, fatal=True, live=False, data=None, headers={},
2104 res = self._download_webpage_handle(
2106 note='Downloading m3u8 information' if note is None else note,
2107 errnote='Failed to download m3u8 information' if errnote is None else errnote,
2108 fatal=fatal, data=data, headers=headers, query=query)
2113 m3u8_doc, urlh = res
2114 m3u8_url = urlh.geturl()
2116 return self._parse_m3u8_formats_and_subtitles(
2117 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
2118 preference=preference, quality=quality, m3u8_id=m3u8_id,
2119 note=note, errnote=errnote, fatal=fatal, live=live, data=data,
2120 headers=headers, query=query, video_id=video_id)
2122 def _parse_m3u8_formats_and_subtitles(
2123 self, m3u8_doc, m3u8_url=None, ext=None, entry_protocol='m3u8_native',
2124 preference=None, quality=None, m3u8_id=None, live=False, note=None,
2125 errnote=None, fatal=True, data=None, headers={}, query={},
2127 formats, subtitles = [], {}
2129 has_drm = re.search('|'.join([
2130 r'#EXT-X-FAXS-CM:', # Adobe Flash Access
2131 r'#EXT-X-(?:SESSION-)?KEY:.*?URI="skd
://', # Apple FairPlay
2134 def format_url(url):
2135 return url if re.match(r'^https?
://', url) else compat_urlparse.urljoin(m3u8_url, url)
2137 if self.get_param('hls_split_discontinuity
', False):
2138 def _extract_m3u8_playlist_indices(manifest_url=None, m3u8_doc=None):
2140 if not manifest_url:
2142 m3u8_doc = self._download_webpage(
2143 manifest_url, video_id, fatal=fatal, data=data, headers=headers,
2144 note=False, errnote='Failed to download m3u8 playlist information
')
2145 if m3u8_doc is False:
2147 return range(1 + sum(line.startswith('#EXT-X-DISCONTINUITY') for line in m3u8_doc.splitlines()))
2150 def _extract_m3u8_playlist_indices(*args
, **kwargs
):
2154 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
2155 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
2156 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
2158 # We should try extracting formats only from master playlists [1, 4.3.4],
2159 # i.e. playlists that describe available qualities. On the other hand
2160 # media playlists [1, 4.3.3] should be returned as is since they contain
2161 # just the media without qualities renditions.
2162 # Fortunately, master playlist can be easily distinguished from media
2163 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
2164 # master playlist tags MUST NOT appear in a media playlist and vice versa.
2165 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
2166 # media playlist and MUST NOT appear in master playlist thus we can
2167 # clearly detect media playlist with this criterion.
2169 if '#EXT-X-TARGETDURATION' in m3u8_doc
: # media playlist, return as is
2171 'format_id': join_nonempty(m3u8_id
, idx
),
2172 'format_index': idx
,
2173 'url': m3u8_url
or encode_data_uri(m3u8_doc
.encode('utf-8'), 'application/x-mpegurl'),
2175 'protocol': entry_protocol
,
2176 'preference': preference
,
2179 } for idx
in _extract_m3u8_playlist_indices(m3u8_doc
=m3u8_doc
)]
2181 return formats
, subtitles
2184 last_stream_inf
= {}
2186 def extract_media(x_media_line
):
2187 media
= parse_m3u8_attributes(x_media_line
)
2188 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
2189 media_type
, group_id
, name
= media
.get('TYPE'), media
.get('GROUP-ID'), media
.get('NAME')
2190 if not (media_type
and group_id
and name
):
2192 groups
.setdefault(group_id
, []).append(media
)
2193 # <https://tools.ietf.org/html/rfc8216#section-4.3.4.1>
2194 if media_type
== 'SUBTITLES':
2195 # According to RFC 8216 §4.3.4.2.1, URI is REQUIRED in the
2196 # EXT-X-MEDIA tag if the media type is SUBTITLES.
2197 # However, lack of URI has been spotted in the wild.
2198 # e.g. NebulaIE; see https://github.com/yt-dlp/yt-dlp/issues/339
2199 if not media
.get('URI'):
2201 url
= format_url(media
['URI'])
2204 'ext': determine_ext(url
),
2206 if sub_info
['ext'] == 'm3u8':
2207 # Per RFC 8216 §3.1, the only possible subtitle format m3u8
2208 # files may contain is WebVTT:
2209 # <https://tools.ietf.org/html/rfc8216#section-3.1>
2210 sub_info
['ext'] = 'vtt'
2211 sub_info
['protocol'] = 'm3u8_native'
2212 lang
= media
.get('LANGUAGE') or 'und'
2213 subtitles
.setdefault(lang
, []).append(sub_info
)
2214 if media_type
not in ('VIDEO', 'AUDIO'):
2216 media_url
= media
.get('URI')
2218 manifest_url
= format_url(media_url
)
2220 'format_id': join_nonempty(m3u8_id
, group_id
, name
, idx
),
2221 'format_note': name
,
2222 'format_index': idx
,
2223 'url': manifest_url
,
2224 'manifest_url': m3u8_url
,
2225 'language': media
.get('LANGUAGE'),
2227 'protocol': entry_protocol
,
2228 'preference': preference
,
2230 'vcodec': 'none' if media_type
== 'AUDIO' else None,
2231 } for idx
in _extract_m3u8_playlist_indices(manifest_url
))
2233 def build_stream_name():
2234 # Despite specification does not mention NAME attribute for
2235 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
2236 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
2237 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
2238 stream_name
= last_stream_inf
.get('NAME')
2241 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
2242 # from corresponding rendition group
2243 stream_group_id
= last_stream_inf
.get('VIDEO')
2244 if not stream_group_id
:
2246 stream_group
= groups
.get(stream_group_id
)
2247 if not stream_group
:
2248 return stream_group_id
2249 rendition
= stream_group
[0]
2250 return rendition
.get('NAME') or stream_group_id
2252 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
2253 # chance to detect video only formats when EXT-X-STREAM-INF tags
2254 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
2255 for line
in m3u8_doc
.splitlines():
2256 if line
.startswith('#EXT-X-MEDIA:'):
2259 for line
in m3u8_doc
.splitlines():
2260 if line
.startswith('#EXT-X-STREAM-INF:'):
2261 last_stream_inf
= parse_m3u8_attributes(line
)
2262 elif line
.startswith('#') or not line
.strip():
2265 tbr
= float_or_none(
2266 last_stream_inf
.get('AVERAGE-BANDWIDTH')
2267 or last_stream_inf
.get('BANDWIDTH'), scale
=1000)
2268 manifest_url
= format_url(line
.strip())
2270 for idx
in _extract_m3u8_playlist_indices(manifest_url
):
2271 format_id
= [m3u8_id
, None, idx
]
2272 # Bandwidth of live streams may differ over time thus making
2273 # format_id unpredictable. So it's better to keep provided
2276 stream_name
= build_stream_name()
2277 format_id
[1] = stream_name
or '%d' % (tbr
or len(formats
))
2279 'format_id': join_nonempty(*format_id
),
2280 'format_index': idx
,
2281 'url': manifest_url
,
2282 'manifest_url': m3u8_url
,
2285 'fps': float_or_none(last_stream_inf
.get('FRAME-RATE')),
2286 'protocol': entry_protocol
,
2287 'preference': preference
,
2290 resolution
= last_stream_inf
.get('RESOLUTION')
2292 mobj
= re
.search(r
'(?P<width>\d+)[xX](?P<height>\d+)', resolution
)
2294 f
['width'] = int(mobj
.group('width'))
2295 f
['height'] = int(mobj
.group('height'))
2296 # Unified Streaming Platform
2298 r
'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f
['url'])
2300 abr
, vbr
= mobj
.groups()
2301 abr
, vbr
= float_or_none(abr
, 1000), float_or_none(vbr
, 1000)
2306 codecs
= parse_codecs(last_stream_inf
.get('CODECS'))
2308 audio_group_id
= last_stream_inf
.get('AUDIO')
2309 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
2310 # references a rendition group MUST have a CODECS attribute.
2311 # However, this is not always respected, for example, [2]
2312 # contains EXT-X-STREAM-INF tag which references AUDIO
2313 # rendition group but does not have CODECS and despite
2314 # referencing an audio group it represents a complete
2315 # (with audio and video) format. So, for such cases we will
2316 # ignore references to rendition groups and treat them
2317 # as complete formats.
2318 if audio_group_id
and codecs
and f
.get('vcodec') != 'none':
2319 audio_group
= groups
.get(audio_group_id
)
2320 if audio_group
and audio_group
[0].get('URI'):
2321 # TODO: update acodec for audio only formats with
2323 f
['acodec'] = 'none'
2324 if not f
.get('ext'):
2325 f
['ext'] = 'm4a' if f
.get('vcodec') == 'none' else 'mp4'
2329 progressive_uri
= last_stream_inf
.get('PROGRESSIVE-URI')
2332 del http_f
['manifest_url']
2334 'format_id': f
['format_id'].replace('hls-', 'http-'),
2336 'url': progressive_uri
,
2338 formats
.append(http_f
)
2340 last_stream_inf
= {}
2341 return formats
, subtitles
2343 def _extract_m3u8_vod_duration(
2344 self
, m3u8_vod_url
, video_id
, note
=None, errnote
=None, data
=None, headers
={}, query={}
):
2346 m3u8_vod
= self
._download
_webpage
(
2347 m3u8_vod_url
, video_id
,
2348 note
='Downloading m3u8 VOD manifest' if note
is None else note
,
2349 errnote
='Failed to download VOD manifest' if errnote
is None else errnote
,
2350 fatal
=False, data
=data
, headers
=headers
, query
=query
)
2352 return self
._parse
_m
3u8_vod
_duration
(m3u8_vod
or '', video_id
)
2354 def _parse_m3u8_vod_duration(self
, m3u8_vod
, video_id
):
2355 if '#EXT-X-PLAYLIST-TYPE:VOD' not in m3u8_vod
:
2359 float(line
[len('#EXTINF:'):].split(',')[0])
2360 for line
in m3u8_vod
.splitlines() if line
.startswith('#EXTINF:'))) or None
2363 def _xpath_ns(path
, namespace
=None):
2367 for c
in path
.split('/'):
2368 if not c
or c
== '.':
2371 out
.append('{%s}%s' % (namespace
, c
))
2372 return '/'.join(out
)
2374 def _extract_smil_formats_and_subtitles(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None, transform_source
=None):
2375 smil
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
, transform_source
=transform_source
)
2381 namespace
= self
._parse
_smil
_namespace
(smil
)
2383 fmts
= self
._parse
_smil
_formats
(
2384 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2385 subs
= self
._parse
_smil
_subtitles
(
2386 smil
, namespace
=namespace
)
2390 def _extract_smil_formats(self
, *args
, **kwargs
):
2391 fmts
, subs
= self
._extract
_smil
_formats
_and
_subtitles
(*args
, **kwargs
)
2393 self
._report
_ignoring
_subs
('SMIL')
2396 def _extract_smil_info(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None):
2397 smil
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
)
2400 return self
._parse
_smil
(smil
, smil_url
, video_id
, f4m_params
=f4m_params
)
2402 def _download_smil(self
, smil_url
, video_id
, fatal
=True, transform_source
=None):
2403 return self
._download
_xml
(
2404 smil_url
, video_id
, 'Downloading SMIL file',
2405 'Unable to download SMIL file', fatal
=fatal
, transform_source
=transform_source
)
2407 def _parse_smil(self
, smil
, smil_url
, video_id
, f4m_params
=None):
2408 namespace
= self
._parse
_smil
_namespace
(smil
)
2410 formats
= self
._parse
_smil
_formats
(
2411 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2412 subtitles
= self
._parse
_smil
_subtitles
(smil
, namespace
=namespace
)
2414 video_id
= os
.path
.splitext(url_basename(smil_url
))[0]
2418 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2419 name
= meta
.attrib
.get('name')
2420 content
= meta
.attrib
.get('content')
2421 if not name
or not content
:
2423 if not title
and name
== 'title':
2425 elif not description
and name
in ('description', 'abstract'):
2426 description
= content
2427 elif not upload_date
and name
== 'date':
2428 upload_date
= unified_strdate(content
)
2431 'id': image
.get('type'),
2432 'url': image
.get('src'),
2433 'width': int_or_none(image
.get('width')),
2434 'height': int_or_none(image
.get('height')),
2435 } for image
in smil
.findall(self
._xpath
_ns
('.//image', namespace
)) if image
.get('src')]
2439 'title': title
or video_id
,
2440 'description': description
,
2441 'upload_date': upload_date
,
2442 'thumbnails': thumbnails
,
2444 'subtitles': subtitles
,
2447 def _parse_smil_namespace(self
, smil
):
2448 return self
._search
_regex
(
2449 r
'(?i)^{([^}]+)?}smil$', smil
.tag
, 'namespace', default
=None)
2451 def _parse_smil_formats(self
, smil
, smil_url
, video_id
, namespace
=None, f4m_params
=None, transform_rtmp_url
=None):
2453 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2454 b
= meta
.get('base') or meta
.get('httpBase')
2466 media
= smil
.findall(self
._xpath
_ns
('.//video', namespace
)) + smil
.findall(self
._xpath
_ns
('.//audio', namespace
))
2467 for medium
in media
:
2468 src
= medium
.get('src')
2469 if not src
or src
in srcs
:
2473 bitrate
= float_or_none(medium
.get('system-bitrate') or medium
.get('systemBitrate'), 1000)
2474 filesize
= int_or_none(medium
.get('size') or medium
.get('fileSize'))
2475 width
= int_or_none(medium
.get('width'))
2476 height
= int_or_none(medium
.get('height'))
2477 proto
= medium
.get('proto')
2478 ext
= medium
.get('ext')
2479 src_ext
= determine_ext(src
)
2480 streamer
= medium
.get('streamer') or base
2482 if proto
== 'rtmp' or streamer
.startswith('rtmp'):
2488 'format_id': 'rtmp-%d' % (rtmp_count
if bitrate
is None else bitrate
),
2490 'filesize': filesize
,
2494 if transform_rtmp_url
:
2495 streamer
, src
= transform_rtmp_url(streamer
, src
)
2496 formats
[-1].update({
2502 src_url
= src
if src
.startswith('http') else compat_urlparse
.urljoin(base
, src
)
2503 src_url
= src_url
.strip()
2505 if proto
== 'm3u8' or src_ext
== 'm3u8':
2506 m3u8_formats
= self
._extract
_m
3u8_formats
(
2507 src_url
, video_id
, ext
or 'mp4', m3u8_id
='hls', fatal
=False)
2508 if len(m3u8_formats
) == 1:
2510 m3u8_formats
[0].update({
2511 'format_id': 'hls-%d' % (m3u8_count
if bitrate
is None else bitrate
),
2516 formats
.extend(m3u8_formats
)
2517 elif src_ext
== 'f4m':
2522 'plugin': 'flowplayer-3.2.0.1',
2524 f4m_url
+= '&' if '?' in f4m_url
else '?'
2525 f4m_url
+= compat_urllib_parse_urlencode(f4m_params
)
2526 formats
.extend(self
._extract
_f
4m
_formats
(f4m_url
, video_id
, f4m_id
='hds', fatal
=False))
2527 elif src_ext
== 'mpd':
2528 formats
.extend(self
._extract
_mpd
_formats
(
2529 src_url
, video_id
, mpd_id
='dash', fatal
=False))
2530 elif re
.search(r
'\.ism/[Mm]anifest', src_url
):
2531 formats
.extend(self
._extract
_ism
_formats
(
2532 src_url
, video_id
, ism_id
='mss', fatal
=False))
2533 elif src_url
.startswith('http') and self
._is
_valid
_url
(src
, video_id
):
2537 'ext': ext
or src_ext
or 'flv',
2538 'format_id': 'http-%d' % (bitrate
or http_count
),
2540 'filesize': filesize
,
2545 for medium
in smil
.findall(self
._xpath
_ns
('.//imagestream', namespace
)):
2546 src
= medium
.get('src')
2547 if not src
or src
in srcs
:
2553 'format_id': 'imagestream-%d' % (imgs_count
),
2555 'ext': mimetype2ext(medium
.get('type')),
2558 'width': int_or_none(medium
.get('width')),
2559 'height': int_or_none(medium
.get('height')),
2560 'format_note': 'SMIL storyboards',
2565 def _parse_smil_subtitles(self
, smil
, namespace
=None, subtitles_lang
='en'):
2568 for num
, textstream
in enumerate(smil
.findall(self
._xpath
_ns
('.//textstream', namespace
))):
2569 src
= textstream
.get('src')
2570 if not src
or src
in urls
:
2573 ext
= textstream
.get('ext') or mimetype2ext(textstream
.get('type')) or determine_ext(src
)
2574 lang
= textstream
.get('systemLanguage') or textstream
.get('systemLanguageName') or textstream
.get('lang') or subtitles_lang
2575 subtitles
.setdefault(lang
, []).append({
2581 def _extract_xspf_playlist(self
, xspf_url
, playlist_id
, fatal
=True):
2582 xspf
= self
._download
_xml
(
2583 xspf_url
, playlist_id
, 'Downloading xpsf playlist',
2584 'Unable to download xspf manifest', fatal
=fatal
)
2587 return self
._parse
_xspf
(
2588 xspf
, playlist_id
, xspf_url
=xspf_url
,
2589 xspf_base_url
=base_url(xspf_url
))
2591 def _parse_xspf(self
, xspf_doc
, playlist_id
, xspf_url
=None, xspf_base_url
=None):
2593 'xspf': 'http://xspf.org/ns/0/',
2594 's1': 'http://static.streamone.nl/player/ns/0',
2598 for track
in xspf_doc
.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP
)):
2600 track
, xpath_with_ns('./xspf:title', NS_MAP
), 'title', default
=playlist_id
)
2601 description
= xpath_text(
2602 track
, xpath_with_ns('./xspf:annotation', NS_MAP
), 'description')
2603 thumbnail
= xpath_text(
2604 track
, xpath_with_ns('./xspf:image', NS_MAP
), 'thumbnail')
2605 duration
= float_or_none(
2606 xpath_text(track
, xpath_with_ns('./xspf:duration', NS_MAP
), 'duration'), 1000)
2609 for location
in track
.findall(xpath_with_ns('./xspf:location', NS_MAP
)):
2610 format_url
= urljoin(xspf_base_url
, location
.text
)
2615 'manifest_url': xspf_url
,
2616 'format_id': location
.get(xpath_with_ns('s1:label', NS_MAP
)),
2617 'width': int_or_none(location
.get(xpath_with_ns('s1:width', NS_MAP
))),
2618 'height': int_or_none(location
.get(xpath_with_ns('s1:height', NS_MAP
))),
2620 self
._sort
_formats
(formats
)
2625 'description': description
,
2626 'thumbnail': thumbnail
,
2627 'duration': duration
,
2632 def _extract_mpd_formats(self
, *args
, **kwargs
):
2633 fmts
, subs
= self
._extract
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2635 self
._report
_ignoring
_subs
('DASH')
2638 def _extract_mpd_formats_and_subtitles(
2639 self
, mpd_url
, video_id
, mpd_id
=None, note
=None, errnote
=None,
2640 fatal
=True, data
=None, headers
={}, query={}
):
2641 res
= self
._download
_xml
_handle
(
2643 note
='Downloading MPD manifest' if note
is None else note
,
2644 errnote
='Failed to download MPD manifest' if errnote
is None else errnote
,
2645 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
2651 mpd_base_url
= base_url(urlh
.geturl())
2653 return self
._parse
_mpd
_formats
_and
_subtitles
(
2654 mpd_doc
, mpd_id
, mpd_base_url
, mpd_url
)
2656 def _parse_mpd_formats(self
, *args
, **kwargs
):
2657 fmts
, subs
= self
._parse
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2659 self
._report
_ignoring
_subs
('DASH')
2662 def _parse_mpd_formats_and_subtitles(
2663 self
, mpd_doc
, mpd_id
=None, mpd_base_url
='', mpd_url
=None):
2665 Parse formats from MPD manifest.
2667 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
2668 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
2669 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
2671 if not self
.get_param('dynamic_mpd', True):
2672 if mpd_doc
.get('type') == 'dynamic':
2675 namespace
= self
._search
_regex
(r
'(?i)^{([^}]+)?}MPD$', mpd_doc
.tag
, 'namespace', default
=None)
2678 return self
._xpath
_ns
(path
, namespace
)
2680 def is_drm_protected(element
):
2681 return element
.find(_add_ns('ContentProtection')) is not None
2683 def extract_multisegment_info(element
, ms_parent_info
):
2684 ms_info
= ms_parent_info
.copy()
2686 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2687 # common attributes and elements. We will only extract relevant
2689 def extract_common(source
):
2690 segment_timeline
= source
.find(_add_ns('SegmentTimeline'))
2691 if segment_timeline
is not None:
2692 s_e
= segment_timeline
.findall(_add_ns('S'))
2694 ms_info
['total_number'] = 0
2697 r
= int(s
.get('r', 0))
2698 ms_info
['total_number'] += 1 + r
2699 ms_info
['s'].append({
2700 't': int(s
.get('t', 0)),
2701 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2702 'd': int(s
.attrib
['d']),
2705 start_number
= source
.get('startNumber')
2707 ms_info
['start_number'] = int(start_number
)
2708 timescale
= source
.get('timescale')
2710 ms_info
['timescale'] = int(timescale
)
2711 segment_duration
= source
.get('duration')
2712 if segment_duration
:
2713 ms_info
['segment_duration'] = float(segment_duration
)
2715 def extract_Initialization(source
):
2716 initialization
= source
.find(_add_ns('Initialization'))
2717 if initialization
is not None:
2718 ms_info
['initialization_url'] = initialization
.attrib
['sourceURL']
2720 segment_list
= element
.find(_add_ns('SegmentList'))
2721 if segment_list
is not None:
2722 extract_common(segment_list
)
2723 extract_Initialization(segment_list
)
2724 segment_urls_e
= segment_list
.findall(_add_ns('SegmentURL'))
2726 ms_info
['segment_urls'] = [segment
.attrib
['media'] for segment
in segment_urls_e
]
2728 segment_template
= element
.find(_add_ns('SegmentTemplate'))
2729 if segment_template
is not None:
2730 extract_common(segment_template
)
2731 media
= segment_template
.get('media')
2733 ms_info
['media'] = media
2734 initialization
= segment_template
.get('initialization')
2736 ms_info
['initialization'] = initialization
2738 extract_Initialization(segment_template
)
2741 mpd_duration
= parse_duration(mpd_doc
.get('mediaPresentationDuration'))
2742 formats
, subtitles
= [], {}
2743 stream_numbers
= collections
.defaultdict(int)
2744 for period
in mpd_doc
.findall(_add_ns('Period')):
2745 period_duration
= parse_duration(period
.get('duration')) or mpd_duration
2746 period_ms_info
= extract_multisegment_info(period
, {
2750 for adaptation_set
in period
.findall(_add_ns('AdaptationSet')):
2751 adaption_set_ms_info
= extract_multisegment_info(adaptation_set
, period_ms_info
)
2752 for representation
in adaptation_set
.findall(_add_ns('Representation')):
2753 representation_attrib
= adaptation_set
.attrib
.copy()
2754 representation_attrib
.update(representation
.attrib
)
2755 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2756 mime_type
= representation_attrib
['mimeType']
2757 content_type
= representation_attrib
.get('contentType', mime_type
.split('/')[0])
2759 codecs
= parse_codecs(representation_attrib
.get('codecs', ''))
2760 if content_type
not in ('video', 'audio', 'text'):
2761 if mime_type
== 'image/jpeg':
2762 content_type
= mime_type
2763 elif codecs
['vcodec'] != 'none':
2764 content_type
= 'video'
2765 elif codecs
['acodec'] != 'none':
2766 content_type
= 'audio'
2767 elif codecs
.get('tcodec', 'none') != 'none':
2768 content_type
= 'text'
2769 elif mimetype2ext(mime_type
) in ('tt', 'dfxp', 'ttml', 'xml', 'json'):
2770 content_type
= 'text'
2772 self
.report_warning('Unknown MIME type %s in DASH manifest' % mime_type
)
2776 for element
in (representation
, adaptation_set
, period
, mpd_doc
):
2777 base_url_e
= element
.find(_add_ns('BaseURL'))
2778 if base_url_e
is not None:
2779 base_url
= base_url_e
.text
+ base_url
2780 if re
.match(r
'^https?://', base_url
):
2782 if mpd_base_url
and base_url
.startswith('/'):
2783 base_url
= compat_urlparse
.urljoin(mpd_base_url
, base_url
)
2784 elif mpd_base_url
and not re
.match(r
'^https?://', base_url
):
2785 if not mpd_base_url
.endswith('/'):
2787 base_url
= mpd_base_url
+ base_url
2788 representation_id
= representation_attrib
.get('id')
2789 lang
= representation_attrib
.get('lang')
2790 url_el
= representation
.find(_add_ns('BaseURL'))
2791 filesize
= int_or_none(url_el
.attrib
.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el
is not None else None)
2792 bandwidth
= int_or_none(representation_attrib
.get('bandwidth'))
2793 if representation_id
is not None:
2794 format_id
= representation_id
2796 format_id
= content_type
2798 format_id
= mpd_id
+ '-' + format_id
2799 if content_type
in ('video', 'audio'):
2801 'format_id': format_id
,
2802 'manifest_url': mpd_url
,
2803 'ext': mimetype2ext(mime_type
),
2804 'width': int_or_none(representation_attrib
.get('width')),
2805 'height': int_or_none(representation_attrib
.get('height')),
2806 'tbr': float_or_none(bandwidth
, 1000),
2807 'asr': int_or_none(representation_attrib
.get('audioSamplingRate')),
2808 'fps': int_or_none(representation_attrib
.get('frameRate')),
2809 'language': lang
if lang
not in ('mul', 'und', 'zxx', 'mis') else None,
2810 'format_note': 'DASH %s' % content_type
,
2811 'filesize': filesize
,
2812 'container': mimetype2ext(mime_type
) + '_dash',
2815 elif content_type
== 'text':
2817 'ext': mimetype2ext(mime_type
),
2818 'manifest_url': mpd_url
,
2819 'filesize': filesize
,
2821 elif content_type
== 'image/jpeg':
2822 # See test case in VikiIE
2823 # https://www.viki.com/videos/1175236v-choosing-spouse-by-lottery-episode-1
2825 'format_id': format_id
,
2827 'manifest_url': mpd_url
,
2828 'format_note': 'DASH storyboards (jpeg)',
2832 if is_drm_protected(adaptation_set
) or is_drm_protected(representation
):
2834 representation_ms_info
= extract_multisegment_info(representation
, adaption_set_ms_info
)
2836 def prepare_template(template_name
, identifiers
):
2837 tmpl
= representation_ms_info
[template_name
]
2838 # First of, % characters outside $...$ templates
2839 # must be escaped by doubling for proper processing
2840 # by % operator string formatting used further (see
2841 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2847 in_template
= not in_template
2848 elif c
== '%' and not in_template
:
2850 # Next, $...$ templates are translated to their
2851 # %(...) counterparts to be used with % operator
2852 if representation_id
is not None:
2853 t
= t
.replace('$RepresentationID$', representation_id
)
2854 t
= re
.sub(r
'\$(%s)\$' % '|'.join(identifiers
), r
'%(\1)d', t
)
2855 t
= re
.sub(r
'\$(%s)%%([^$]+)\$' % '|'.join(identifiers
), r
'%(\1)\2', t
)
2856 t
.replace('$$', '$')
2859 # @initialization is a regular template like @media one
2860 # so it should be handled just the same way (see
2861 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2862 if 'initialization' in representation_ms_info
:
2863 initialization_template
= prepare_template(
2865 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2866 # $Time$ shall not be included for @initialization thus
2867 # only $Bandwidth$ remains
2869 representation_ms_info
['initialization_url'] = initialization_template
% {
2870 'Bandwidth': bandwidth
,
2873 def location_key(location
):
2874 return 'url' if re
.match(r
'^https?://', location
) else 'path'
2876 if 'segment_urls' not in representation_ms_info
and 'media' in representation_ms_info
:
2878 media_template
= prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2879 media_location_key
= location_key(media_template
)
2881 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2882 # can't be used at the same time
2883 if '%(Number' in media_template
and 's' not in representation_ms_info
:
2884 segment_duration
= None
2885 if 'total_number' not in representation_ms_info
and 'segment_duration' in representation_ms_info
:
2886 segment_duration
= float_or_none(representation_ms_info
['segment_duration'], representation_ms_info
['timescale'])
2887 representation_ms_info
['total_number'] = int(math
.ceil(
2888 float_or_none(period_duration
, segment_duration
, default
=0)))
2889 representation_ms_info
['fragments'] = [{
2890 media_location_key
: media_template
% {
2891 'Number': segment_number
,
2892 'Bandwidth': bandwidth
,
2894 'duration': segment_duration
,
2895 } for segment_number
in range(
2896 representation_ms_info
['start_number'],
2897 representation_ms_info
['total_number'] + representation_ms_info
['start_number'])]
2899 # $Number*$ or $Time$ in media template with S list available
2900 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2901 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2902 representation_ms_info
['fragments'] = []
2905 segment_number
= representation_ms_info
['start_number']
2907 def add_segment_url():
2908 segment_url
= media_template
% {
2909 'Time': segment_time
,
2910 'Bandwidth': bandwidth
,
2911 'Number': segment_number
,
2913 representation_ms_info
['fragments'].append({
2914 media_location_key
: segment_url
,
2915 'duration': float_or_none(segment_d
, representation_ms_info
['timescale']),
2918 for num
, s
in enumerate(representation_ms_info
['s']):
2919 segment_time
= s
.get('t') or segment_time
2923 for r
in range(s
.get('r', 0)):
2924 segment_time
+= segment_d
2927 segment_time
+= segment_d
2928 elif 'segment_urls' in representation_ms_info
and 's' in representation_ms_info
:
2930 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2931 # or any YouTube dashsegments video
2934 timescale
= representation_ms_info
['timescale']
2935 for s
in representation_ms_info
['s']:
2936 duration
= float_or_none(s
['d'], timescale
)
2937 for r
in range(s
.get('r', 0) + 1):
2938 segment_uri
= representation_ms_info
['segment_urls'][segment_index
]
2940 location_key(segment_uri
): segment_uri
,
2941 'duration': duration
,
2944 representation_ms_info
['fragments'] = fragments
2945 elif 'segment_urls' in representation_ms_info
:
2946 # Segment URLs with no SegmentTimeline
2947 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2948 # https://github.com/ytdl-org/youtube-dl/pull/14844
2950 segment_duration
= float_or_none(
2951 representation_ms_info
['segment_duration'],
2952 representation_ms_info
['timescale']) if 'segment_duration' in representation_ms_info
else None
2953 for segment_url
in representation_ms_info
['segment_urls']:
2955 location_key(segment_url
): segment_url
,
2957 if segment_duration
:
2958 fragment
['duration'] = segment_duration
2959 fragments
.append(fragment
)
2960 representation_ms_info
['fragments'] = fragments
2961 # If there is a fragments key available then we correctly recognized fragmented media.
2962 # Otherwise we will assume unfragmented media with direct access. Technically, such
2963 # assumption is not necessarily correct since we may simply have no support for
2964 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2965 if 'fragments' in representation_ms_info
:
2967 # NB: mpd_url may be empty when MPD manifest is parsed from a string
2968 'url': mpd_url
or base_url
,
2969 'fragment_base_url': base_url
,
2971 'protocol': 'http_dash_segments' if mime_type
!= 'image/jpeg' else 'mhtml',
2973 if 'initialization_url' in representation_ms_info
:
2974 initialization_url
= representation_ms_info
['initialization_url']
2975 if not f
.get('url'):
2976 f
['url'] = initialization_url
2977 f
['fragments'].append({location_key(initialization_url): initialization_url}
)
2978 f
['fragments'].extend(representation_ms_info
['fragments'])
2979 if not period_duration
:
2980 period_duration
= try_get(
2981 representation_ms_info
,
2982 lambda r
: sum(frag
['duration'] for frag
in r
['fragments']), float)
2984 # Assuming direct URL to unfragmented media.
2986 if content_type
in ('video', 'audio', 'image/jpeg'):
2987 f
['manifest_stream_number'] = stream_numbers
[f
['url']]
2988 stream_numbers
[f
['url']] += 1
2990 elif content_type
== 'text':
2991 subtitles
.setdefault(lang
or 'und', []).append(f
)
2993 return formats
, subtitles
2995 def _extract_ism_formats(self
, *args
, **kwargs
):
2996 fmts
, subs
= self
._extract
_ism
_formats
_and
_subtitles
(*args
, **kwargs
)
2998 self
._report
_ignoring
_subs
('ISM')
3001 def _extract_ism_formats_and_subtitles(self
, ism_url
, video_id
, ism_id
=None, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
):
3002 res
= self
._download
_xml
_handle
(
3004 note
='Downloading ISM manifest' if note
is None else note
,
3005 errnote
='Failed to download ISM manifest' if errnote
is None else errnote
,
3006 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
3013 return self
._parse
_ism
_formats
_and
_subtitles
(ism_doc
, urlh
.geturl(), ism_id
)
3015 def _parse_ism_formats_and_subtitles(self
, ism_doc
, ism_url
, ism_id
=None):
3017 Parse formats from ISM manifest.
3019 1. [MS-SSTR]: Smooth Streaming Protocol,
3020 https://msdn.microsoft.com/en-us/library/ff469518.aspx
3022 if ism_doc
.get('IsLive') == 'TRUE':
3025 duration
= int(ism_doc
.attrib
['Duration'])
3026 timescale
= int_or_none(ism_doc
.get('TimeScale')) or 10000000
3030 for stream
in ism_doc
.findall('StreamIndex'):
3031 stream_type
= stream
.get('Type')
3032 if stream_type
not in ('video', 'audio', 'text'):
3034 url_pattern
= stream
.attrib
['Url']
3035 stream_timescale
= int_or_none(stream
.get('TimeScale')) or timescale
3036 stream_name
= stream
.get('Name')
3037 stream_language
= stream
.get('Language', 'und')
3038 for track
in stream
.findall('QualityLevel'):
3039 fourcc
= track
.get('FourCC') or ('AACL' if track
.get('AudioTag') == '255' else None)
3040 # TODO: add support for WVC1 and WMAP
3041 if fourcc
not in ('H264', 'AVC1', 'AACL', 'TTML'):
3042 self
.report_warning('%s is not a supported codec' % fourcc
)
3044 tbr
= int(track
.attrib
['Bitrate']) // 1000
3045 # [1] does not mention Width and Height attributes. However,
3046 # they're often present while MaxWidth and MaxHeight are
3047 # missing, so should be used as fallbacks
3048 width
= int_or_none(track
.get('MaxWidth') or track
.get('Width'))
3049 height
= int_or_none(track
.get('MaxHeight') or track
.get('Height'))
3050 sampling_rate
= int_or_none(track
.get('SamplingRate'))
3052 track_url_pattern
= re
.sub(r
'{[Bb]itrate}', track
.attrib
['Bitrate'], url_pattern
)
3053 track_url_pattern
= compat_urlparse
.urljoin(ism_url
, track_url_pattern
)
3059 stream_fragments
= stream
.findall('c')
3060 for stream_fragment_index
, stream_fragment
in enumerate(stream_fragments
):
3061 fragment_ctx
['time'] = int_or_none(stream_fragment
.get('t')) or fragment_ctx
['time']
3062 fragment_repeat
= int_or_none(stream_fragment
.get('r')) or 1
3063 fragment_ctx
['duration'] = int_or_none(stream_fragment
.get('d'))
3064 if not fragment_ctx
['duration']:
3066 next_fragment_time
= int(stream_fragment
[stream_fragment_index
+ 1].attrib
['t'])
3068 next_fragment_time
= duration
3069 fragment_ctx
['duration'] = (next_fragment_time
- fragment_ctx
['time']) / fragment_repeat
3070 for _
in range(fragment_repeat
):
3072 'url': re
.sub(r
'{start[ _]time}', compat_str(fragment_ctx
['time']), track_url_pattern
),
3073 'duration': fragment_ctx
['duration'] / stream_timescale
,
3075 fragment_ctx
['time'] += fragment_ctx
['duration']
3077 if stream_type
== 'text':
3078 subtitles
.setdefault(stream_language
, []).append({
3082 'manifest_url': ism_url
,
3083 'fragments': fragments
,
3084 '_download_params': {
3085 'stream_type': stream_type
,
3086 'duration': duration
,
3087 'timescale': stream_timescale
,
3089 'language': stream_language
,
3090 'codec_private_data': track
.get('CodecPrivateData'),
3093 elif stream_type
in ('video', 'audio'):
3095 'format_id': join_nonempty(ism_id
, stream_name
, tbr
),
3097 'manifest_url': ism_url
,
3098 'ext': 'ismv' if stream_type
== 'video' else 'isma',
3102 'asr': sampling_rate
,
3103 'vcodec': 'none' if stream_type
== 'audio' else fourcc
,
3104 'acodec': 'none' if stream_type
== 'video' else fourcc
,
3106 'fragments': fragments
,
3107 'has_drm': ism_doc
.find('Protection') is not None,
3108 '_download_params': {
3109 'stream_type': stream_type
,
3110 'duration': duration
,
3111 'timescale': stream_timescale
,
3112 'width': width
or 0,
3113 'height': height
or 0,
3115 'language': stream_language
,
3116 'codec_private_data': track
.get('CodecPrivateData'),
3117 'sampling_rate': sampling_rate
,
3118 'channels': int_or_none(track
.get('Channels', 2)),
3119 'bits_per_sample': int_or_none(track
.get('BitsPerSample', 16)),
3120 'nal_unit_length_field': int_or_none(track
.get('NALUnitLengthField', 4)),
3123 return formats
, subtitles
3125 def _parse_html5_media_entries(self
, base_url
, webpage
, video_id
, m3u8_id
=None, m3u8_entry_protocol
='m3u8_native', mpd_id
=None, preference
=None, quality
=None):
3126 def absolute_url(item_url
):
3127 return urljoin(base_url
, item_url
)
3129 def parse_content_type(content_type
):
3130 if not content_type
:
3132 ctr
= re
.search(r
'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type
)
3134 mimetype
, codecs
= ctr
.groups()
3135 f
= parse_codecs(codecs
)
3136 f
['ext'] = mimetype2ext(mimetype
)
3140 def _media_formats(src
, cur_media_type
, type_info
={}):
3141 full_url
= absolute_url(src
)
3142 ext
= type_info
.get('ext') or determine_ext(full_url
)
3144 is_plain_url
= False
3145 formats
= self
._extract
_m
3u8_formats
(
3146 full_url
, video_id
, ext
='mp4',
3147 entry_protocol
=m3u8_entry_protocol
, m3u8_id
=m3u8_id
,
3148 preference
=preference
, quality
=quality
, fatal
=False)
3150 is_plain_url
= False
3151 formats
= self
._extract
_mpd
_formats
(
3152 full_url
, video_id
, mpd_id
=mpd_id
, fatal
=False)
3157 'vcodec': 'none' if cur_media_type
== 'audio' else None,
3159 return is_plain_url
, formats
3162 # amp-video and amp-audio are very similar to their HTML5 counterparts
3163 # so we wll include them right here (see
3164 # https://www.ampproject.org/docs/reference/components/amp-video)
3165 # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/
3166 _MEDIA_TAG_NAME_RE
= r
'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)'
3167 media_tags
= [(media_tag
, media_tag_name
, media_type
, '')
3168 for media_tag
, media_tag_name
, media_type
3169 in re
.findall(r
'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE
, webpage
)]
3170 media_tags
.extend(re
.findall(
3171 # We only allow video|audio followed by a whitespace or '>'.
3172 # Allowing more characters may end up in significant slow down (see
3173 # https://github.com/ytdl-org/youtube-dl/issues/11979, example URL:
3174 # http://www.porntrex.com/maps/videositemap.xml).
3175 r
'(?s)(<(?P<tag>%s)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>' % _MEDIA_TAG_NAME_RE
, webpage
))
3176 for media_tag
, _
, media_type
, media_content
in media_tags
:
3181 media_attributes
= extract_attributes(media_tag
)
3182 src
= strip_or_none(media_attributes
.get('src'))
3184 _
, formats
= _media_formats(src
, media_type
)
3185 media_info
['formats'].extend(formats
)
3186 media_info
['thumbnail'] = absolute_url(media_attributes
.get('poster'))
3188 for source_tag
in re
.findall(r
'<source[^>]+>', media_content
):
3189 s_attr
= extract_attributes(source_tag
)
3190 # data-video-src and data-src are non standard but seen
3191 # several times in the wild
3192 src
= strip_or_none(dict_get(s_attr
, ('src', 'data-video-src', 'data-src')))
3195 f
= parse_content_type(s_attr
.get('type'))
3196 is_plain_url
, formats
= _media_formats(src
, media_type
, f
)
3198 # width, height, res, label and title attributes are
3199 # all not standard but seen several times in the wild
3202 for lbl
in ('label', 'title')
3203 if str_or_none(s_attr
.get(lbl
))
3205 width
= int_or_none(s_attr
.get('width'))
3206 height
= (int_or_none(s_attr
.get('height'))
3207 or int_or_none(s_attr
.get('res')))
3208 if not width
or not height
:
3210 resolution
= parse_resolution(lbl
)
3213 width
= width
or resolution
.get('width')
3214 height
= height
or resolution
.get('height')
3216 tbr
= parse_bitrate(lbl
)
3225 'format_id': s_attr
.get('label') or s_attr
.get('title'),
3227 f
.update(formats
[0])
3228 media_info
['formats'].append(f
)
3230 media_info
['formats'].extend(formats
)
3231 for track_tag
in re
.findall(r
'<track[^>]+>', media_content
):
3232 track_attributes
= extract_attributes(track_tag
)
3233 kind
= track_attributes
.get('kind')
3234 if not kind
or kind
in ('subtitles', 'captions'):
3235 src
= strip_or_none(track_attributes
.get('src'))
3238 lang
= track_attributes
.get('srclang') or track_attributes
.get('lang') or track_attributes
.get('label')
3239 media_info
['subtitles'].setdefault(lang
, []).append({
3240 'url': absolute_url(src
),
3242 for f
in media_info
['formats']:
3243 f
.setdefault('http_headers', {})['Referer'] = base_url
3244 if media_info
['formats'] or media_info
['subtitles']:
3245 entries
.append(media_info
)
3248 def _extract_akamai_formats(self
, *args
, **kwargs
):
3249 fmts
, subs
= self
._extract
_akamai
_formats
_and
_subtitles
(*args
, **kwargs
)
3251 self
._report
_ignoring
_subs
('akamai')
3254 def _extract_akamai_formats_and_subtitles(self
, manifest_url
, video_id
, hosts
={}):
3255 signed
= 'hdnea=' in manifest_url
3257 # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
3258 manifest_url
= re
.sub(
3259 r
'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
3260 '', manifest_url
).strip('?')
3265 hdcore_sign
= 'hdcore=3.7.0'
3266 f4m_url
= re
.sub(r
'(https?://[^/]+)/i/', r
'\1/z/', manifest_url
).replace('/master.m3u8', '/manifest.f4m')
3267 hds_host
= hosts
.get('hds')
3269 f4m_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hds_host
, f4m_url
)
3270 if 'hdcore=' not in f4m_url
:
3271 f4m_url
+= ('&' if '?' in f4m_url
else '?') + hdcore_sign
3272 f4m_formats
= self
._extract
_f
4m
_formats
(
3273 f4m_url
, video_id
, f4m_id
='hds', fatal
=False)
3274 for entry
in f4m_formats
:
3275 entry
.update({'extra_param_to_segment_url': hdcore_sign}
)
3276 formats
.extend(f4m_formats
)
3278 m3u8_url
= re
.sub(r
'(https?://[^/]+)/z/', r
'\1/i/', manifest_url
).replace('/manifest.f4m', '/master.m3u8')
3279 hls_host
= hosts
.get('hls')
3281 m3u8_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hls_host
, m3u8_url
)
3282 m3u8_formats
, m3u8_subtitles
= self
._extract
_m
3u8_formats
_and
_subtitles
(
3283 m3u8_url
, video_id
, 'mp4', 'm3u8_native',
3284 m3u8_id
='hls', fatal
=False)
3285 formats
.extend(m3u8_formats
)
3286 subtitles
= self
._merge
_subtitles
(subtitles
, m3u8_subtitles
)
3288 http_host
= hosts
.get('http')
3289 if http_host
and m3u8_formats
and not signed
:
3290 REPL_REGEX
= r
'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
3291 qualities
= re
.match(REPL_REGEX
, m3u8_url
).group(2).split(',')
3292 qualities_length
= len(qualities
)
3293 if len(m3u8_formats
) in (qualities_length
, qualities_length
+ 1):
3295 for f
in m3u8_formats
:
3296 if f
['vcodec'] != 'none':
3297 for protocol
in ('http', 'https'):
3299 del http_f
['manifest_url']
3301 REPL_REGEX
, protocol
+ r
'://%s/\g<1>%s\3' % (http_host
, qualities
[i
]), f
['url'])
3303 'format_id': http_f
['format_id'].replace('hls-', protocol
+ '-'),
3305 'protocol': protocol
,
3307 formats
.append(http_f
)
3310 return formats
, subtitles
3312 def _extract_wowza_formats(self
, url
, video_id
, m3u8_entry_protocol
='m3u8_native', skip_protocols
=[]):
3313 query
= compat_urlparse
.urlparse(url
).query
3314 url
= re
.sub(r
'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url
)
3316 r
'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url
)
3317 url_base
= mobj
.group('url')
3318 http_base_url
= '%s%s:%s' % ('http', mobj
.group('s') or '', url_base
)
3321 def manifest_url(manifest
):
3322 m_url
= '%s/%s' % (http_base_url
, manifest
)
3324 m_url
+= '?%s' % query
3327 if 'm3u8' not in skip_protocols
:
3328 formats
.extend(self
._extract
_m
3u8_formats
(
3329 manifest_url('playlist.m3u8'), video_id
, 'mp4',
3330 m3u8_entry_protocol
, m3u8_id
='hls', fatal
=False))
3331 if 'f4m' not in skip_protocols
:
3332 formats
.extend(self
._extract
_f
4m
_formats
(
3333 manifest_url('manifest.f4m'),
3334 video_id
, f4m_id
='hds', fatal
=False))
3335 if 'dash' not in skip_protocols
:
3336 formats
.extend(self
._extract
_mpd
_formats
(
3337 manifest_url('manifest.mpd'),
3338 video_id
, mpd_id
='dash', fatal
=False))
3339 if re
.search(r
'(?:/smil:|\.smil)', url_base
):
3340 if 'smil' not in skip_protocols
:
3341 rtmp_formats
= self
._extract
_smil
_formats
(
3342 manifest_url('jwplayer.smil'),
3343 video_id
, fatal
=False)
3344 for rtmp_format
in rtmp_formats
:
3345 rtsp_format
= rtmp_format
.copy()
3346 rtsp_format
['url'] = '%s/%s' % (rtmp_format
['url'], rtmp_format
['play_path'])
3347 del rtsp_format
['play_path']
3348 del rtsp_format
['ext']
3349 rtsp_format
.update({
3350 'url': rtsp_format
['url'].replace('rtmp://', 'rtsp://'),
3351 'format_id': rtmp_format
['format_id'].replace('rtmp', 'rtsp'),
3354 formats
.extend([rtmp_format
, rtsp_format
])
3356 for protocol
in ('rtmp', 'rtsp'):
3357 if protocol
not in skip_protocols
:
3359 'url': '%s:%s' % (protocol
, url_base
),
3360 'format_id': protocol
,
3361 'protocol': protocol
,
3365 def _find_jwplayer_data(self
, webpage
, video_id
=None, transform_source
=js_to_json
):
3367 r
'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P
=quote
)\
)(?
!</script
>).*?\
.setup\s
*\
((?P
<options
>[^
)]+)\
)',
3371 jwplayer_data = self._parse_json(mobj.group('options
'),
3373 transform_source=transform_source)
3374 except ExtractorError:
3377 if isinstance(jwplayer_data, dict):
3378 return jwplayer_data
3380 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
3381 jwplayer_data = self._find_jwplayer_data(
3382 webpage, video_id, transform_source=js_to_json)
3383 return self._parse_jwplayer_data(
3384 jwplayer_data, video_id, *args, **kwargs)
3386 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
3387 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3388 # JWPlayer backward compatibility: flattened playlists
3389 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
3390 if 'playlist
' not in jwplayer_data:
3391 jwplayer_data = {'playlist': [jwplayer_data]}
3395 # JWPlayer backward compatibility: single playlist item
3396 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
3397 if not isinstance(jwplayer_data['playlist
'], list):
3398 jwplayer_data['playlist
'] = [jwplayer_data['playlist
']]
3400 for video_data in jwplayer_data['playlist
']:
3401 # JWPlayer backward compatibility: flattened sources
3402 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
3403 if 'sources
' not in video_data:
3404 video_data['sources
'] = [video_data]
3406 this_video_id = video_id or video_data['mediaid
']
3408 formats = self._parse_jwplayer_formats(
3409 video_data['sources
'], video_id=this_video_id, m3u8_id=m3u8_id,
3410 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
3413 tracks = video_data.get('tracks
')
3414 if tracks and isinstance(tracks, list):
3415 for track in tracks:
3416 if not isinstance(track, dict):
3418 track_kind = track.get('kind
')
3419 if not track_kind or not isinstance(track_kind, compat_str):
3421 if track_kind.lower() not in ('captions
', 'subtitles
'):
3423 track_url = urljoin(base_url, track.get('file'))
3426 subtitles.setdefault(track.get('label
') or 'en
', []).append({
3427 'url
': self._proto_relative_url(track_url)
3431 'id': this_video_id,
3432 'title
': unescapeHTML(video_data['title
'] if require_title else video_data.get('title
')),
3433 'description
': clean_html(video_data.get('description
')),
3434 'thumbnail
': urljoin(base_url, self._proto_relative_url(video_data.get('image
'))),
3435 'timestamp
': int_or_none(video_data.get('pubdate
')),
3436 'duration
': float_or_none(jwplayer_data.get('duration
') or video_data.get('duration
')),
3437 'subtitles
': subtitles,
3439 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
3440 if len(formats) == 1 and re.search(r'^
(?
:http|
//).*(?
:youtube\
.com|youtu\
.be
)/.+', formats[0]['url
']):
3442 '_type
': 'url_transparent
',
3443 'url
': formats[0]['url
'],
3446 self._sort_formats(formats)
3447 entry['formats
'] = formats
3448 entries.append(entry)
3449 if len(entries) == 1:
3452 return self.playlist_result(entries)
3454 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
3455 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3458 for source in jwplayer_sources_data:
3459 if not isinstance(source, dict):
3461 source_url = urljoin(
3462 base_url, self._proto_relative_url(source.get('file')))
3463 if not source_url or source_url in urls:
3465 urls.append(source_url)
3466 source_type = source.get('type') or ''
3467 ext = mimetype2ext(source_type) or determine_ext(source_url)
3468 if source_type == 'hls
' or ext == 'm3u8
':
3469 formats.extend(self._extract_m3u8_formats(
3470 source_url, video_id, 'mp4
', entry_protocol='m3u8_native
',
3471 m3u8_id=m3u8_id, fatal=False))
3472 elif source_type == 'dash
' or ext == 'mpd
':
3473 formats.extend(self._extract_mpd_formats(
3474 source_url, video_id, mpd_id=mpd_id, fatal=False))
3476 formats.extend(self._extract_smil_formats(
3477 source_url, video_id, fatal=False))
3478 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3479 elif source_type.startswith('audio
') or ext in (
3480 'oga
', 'aac
', 'mp3
', 'mpeg
', 'vorbis
'):
3487 height = int_or_none(source.get('height
'))
3489 # Often no height is provided but there is a label in
3490 # format like "1080p", "720p SD", or 1080.
3491 height = int_or_none(self._search_regex(
3492 r'^
(\d{3,4}
)[pP
]?
(?
:\b|$
)', compat_str(source.get('label
') or ''),
3493 'height
', default=None))
3496 'width
': int_or_none(source.get('width
')),
3498 'tbr
': int_or_none(source.get('bitrate
')),
3501 if source_url.startswith('rtmp
'):
3502 a_format['ext
'] = 'flv
'
3503 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3504 # of jwplayer.flash.swf
3505 rtmp_url_parts = re.split(
3506 r'((?
:mp4|mp3|flv
):)', source_url, 1)
3507 if len(rtmp_url_parts) == 3:
3508 rtmp_url, prefix, play_path = rtmp_url_parts
3511 'play_path
': prefix + play_path,
3514 a_format.update(rtmp_params)
3515 formats.append(a_format)
3518 def _live_title(self, name):
3519 self._downloader.deprecation_warning('yt_dlp
.InfoExtractor
._live
_title
is deprecated
and does
not work
as expected
')
3522 def _int(self, v, name, fatal=False, **kwargs):
3523 res = int_or_none(v, **kwargs)
3525 msg = 'Failed to extract
%s: Could
not parse value
%r' % (name, v)
3527 raise ExtractorError(msg)
3529 self.report_warning(msg)
3532 def _float(self, v, name, fatal=False, **kwargs):
3533 res = float_or_none(v, **kwargs)
3535 msg = 'Failed to extract
%s: Could
not parse value
%r' % (name, v)
3537 raise ExtractorError(msg)
3539 self.report_warning(msg)
3542 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
3543 path='/', secure=False, discard=False, rest={}, **kwargs):
3544 cookie = compat_cookiejar_Cookie(
3545 0, name, value, port, port is not None, domain, True,
3546 domain.startswith('.'), path, True, secure, expire_time,
3547 discard, None, None, rest)
3548 self._downloader.cookiejar.set_cookie(cookie)
3550 def _get_cookies(self, url):
3551 """ Return a compat_cookies_SimpleCookie with the cookies for the url """
3552 req = sanitized_Request(url)
3553 self._downloader.cookiejar.add_cookie_header(req)
3554 return compat_cookies_SimpleCookie(req.get_header('Cookie
'))
3556 def _apply_first_set_cookie_header(self, url_handle, cookie):
3558 Apply first Set-Cookie header instead of the last. Experimental.
3560 Some sites (e.g. [1-3]) may serve two cookies under the same name
3561 in Set-Cookie header and expect the first (old) one to be set rather
3562 than second (new). However, as of RFC6265 the newer one cookie
3563 should be set into cookie store what actually happens.
3564 We will workaround this issue by resetting the cookie to
3565 the first one manually.
3566 1. https://new.vk.com/
3567 2. https://github.com/ytdl-org/youtube-dl/issues/9841#issuecomment-227871201
3568 3. https://learning.oreilly.com/
3570 for header, cookies in url_handle.headers.items():
3571 if header.lower() != 'set-cookie
':
3573 if sys.version_info[0] >= 3:
3574 cookies = cookies.encode('iso
-8859-1')
3575 cookies = cookies.decode('utf
-8')
3576 cookie_value = re.search(
3577 r'%s=(.+?
);.*?
\b[Dd
]omain
=(.+?
)(?
:[,;]|$
)' % cookie, cookies)
3579 value, domain = cookie_value.groups()
3580 self._set_cookie(domain, cookie, value)
3583 def get_testcases(self, include_onlymatching=False):
3584 t = getattr(self, '_TEST
', None)
3586 assert not hasattr(self, '_TESTS
'), \
3587 '%s has _TEST
and _TESTS
' % type(self).__name__
3590 tests = getattr(self, '_TESTS
', [])
3592 if not include_onlymatching and t.get('only_matching
', False):
3594 t['name
'] = type(self).__name__[:-len('IE
')]
3597 def is_suitable(self, age_limit):
3598 """ Test whether the extractor is generally suitable for the given
3599 age limit (i.e. pornographic sites are not, all others usually are) """
3601 any_restricted = False
3602 for tc in self.get_testcases(include_onlymatching=False):
3603 if tc.get('playlist
', []):
3604 tc = tc['playlist
'][0]
3605 is_restricted = age_restricted(
3606 tc.get('info_dict
', {}).get('age_limit
'), age_limit)
3607 if not is_restricted:
3609 any_restricted = any_restricted or is_restricted
3610 return not any_restricted
3612 def extract_subtitles(self, *args, **kwargs):
3613 if (self.get_param('writesubtitles
', False)
3614 or self.get_param('listsubtitles
')):
3615 return self._get_subtitles(*args, **kwargs)
3618 def _get_subtitles(self, *args, **kwargs):
3619 raise NotImplementedError('This method must be implemented by subclasses
')
3621 def extract_comments(self, *args, **kwargs):
3622 if not self.get_param('getcomments
'):
3624 generator = self._get_comments(*args, **kwargs)
3631 comments.append(next(generator))
3632 except StopIteration:
3634 except KeyboardInterrupt:
3635 self.to_screen('Interrupted by user
')
3636 except Exception as e:
3637 if self.get_param('ignoreerrors
') is not True:
3639 self._downloader.report_error(e)
3640 comment_count = len(comments)
3641 self.to_screen(f'Extracted {comment_count} comments
')
3643 'comments
': comments,
3644 'comment_count
': None if interrupted else comment_count
3648 def _get_comments(self, *args, **kwargs):
3649 raise NotImplementedError('This method must be implemented by subclasses
')
3652 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
3653 """ Merge subtitle items for one language. Items with duplicated URLs
3654 will be dropped. """
3655 list1_urls = set([item['url
'] for item in subtitle_list1])
3656 ret = list(subtitle_list1)
3657 ret.extend([item for item in subtitle_list2 if item['url
'] not in list1_urls])
3661 def _merge_subtitles(cls, *dicts, target=None):
3662 """ Merge subtitle dictionaries, language by language. """
3666 for lang, subs in d.items():
3667 target[lang] = cls._merge_subtitle_items(target.get(lang, []), subs)
3670 def extract_automatic_captions(self, *args, **kwargs):
3671 if (self.get_param('writeautomaticsub
', False)
3672 or self.get_param('listsubtitles
')):
3673 return self._get_automatic_captions(*args, **kwargs)
3676 def _get_automatic_captions(self, *args, **kwargs):
3677 raise NotImplementedError('This method must be implemented by subclasses
')
3679 def mark_watched(self, *args, **kwargs):
3680 if not self.get_param('mark_watched
', False):
3682 if (hasattr(self, '_NETRC_MACHINE
') and self._get_login_info()[0] is not None
3683 or self.get_param('cookiefile
')
3684 or self.get_param('cookiesfrombrowser
')):
3685 self._mark_watched(*args, **kwargs)
3687 def _mark_watched(self, *args, **kwargs):
3688 raise NotImplementedError('This method must be implemented by subclasses
')
3690 def geo_verification_headers(self):
3692 geo_verification_proxy = self.get_param('geo_verification_proxy
')
3693 if geo_verification_proxy:
3694 headers['Ytdl
-request
-proxy
'] = geo_verification_proxy
3697 def _generic_id(self, url):
3698 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
3700 def _generic_title(self, url):
3701 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
3704 def _availability(is_private=None, needs_premium=None, needs_subscription=None, needs_auth=None, is_unlisted=None):
3705 all_known = all(map(
3706 lambda x: x is not None,
3707 (is_private, needs_premium, needs_subscription, needs_auth, is_unlisted)))
3709 'private
' if is_private
3710 else 'premium_only
' if needs_premium
3711 else 'subscriber_only
' if needs_subscription
3712 else 'needs_auth
' if needs_auth
3713 else 'unlisted
' if is_unlisted
3714 else 'public
' if all_known
3717 def _configuration_arg(self, key, default=NO_DEFAULT, *, ie_key=None, casesense=False):
3719 @returns A list of values for the extractor argument given by "key"
3720 or "default" if no such key is present
3721 @param default The default value to return when the key is not present (default: [])
3722 @param casesense When false, the values are converted to lower case
3725 self._downloader.params, ('extractor_args
', (ie_key or self.ie_key()).lower(), key))
3727 return [] if default is NO_DEFAULT else default
3728 return list(val) if casesense else [x.lower() for x in val]
3730 def _yes_playlist(self, playlist_id, video_id, smuggled_data=None, *, playlist_label='playlist
', video_label='video
'):
3731 if not playlist_id or not video_id:
3734 no_playlist = (smuggled_data or {}).get('force_noplaylist
')
3735 if no_playlist is not None:
3736 return not no_playlist
3738 video_id = '' if video_id is True else f' {video_id}
'
3739 playlist_id = '' if playlist_id is True else f' {playlist_id}
'
3740 if self.get_param('noplaylist
'):
3741 self.to_screen(f'Downloading just the {video_label}{video_id} because of
--no
-playlist
')
3743 self.to_screen(f'Downloading {playlist_label}{playlist_id}
- add
--no
-playlist to download just the {video_label}{video_id}
')
3747 class SearchInfoExtractor(InfoExtractor):
3749 Base class for paged search queries extractors.
3750 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
3751 Instances should define _SEARCH_KEY and optionally _MAX_RESULTS
3754 _MAX_RESULTS = float('inf
')
3757 def _make_valid_url(cls):
3758 return r'%s(?P
<prefix
>|
[1-9][0-9]*|all
):(?P
<query
>[\s\S
]+)' % cls._SEARCH_KEY
3760 def _real_extract(self, query):
3761 prefix, query = self._match_valid_url(query).group('prefix
', 'query
')
3763 return self._get_n_results(query, 1)
3764 elif prefix == 'all
':
3765 return self._get_n_results(query, self._MAX_RESULTS)
3769 raise ExtractorError('invalid download number
%s for query
"%s"' % (n, query))
3770 elif n > self._MAX_RESULTS:
3771 self.report_warning('%s returns
max %i results (you requested
%i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3772 n = self._MAX_RESULTS
3773 return self._get_n_results(query, n)
3775 def _get_n_results(self, query, n):
3776 """Get a specified number of results for a query.
3777 Either this function or _search_results must be overridden by subclasses """
3778 return self.playlist_result(
3779 itertools.islice(self._search_results(query), 0, None if n == float('inf
') else n),
3782 def _search_results(self, query):
3783 """Returns an iterator of search results"""
3784 raise NotImplementedError('This method must be implemented by subclasses
')
3787 def SEARCH_KEY(self):
3788 return self._SEARCH_KEY