2 from __future__
import unicode_literals
17 from ..compat
import (
18 compat_cookiejar_Cookie
,
19 compat_cookies_SimpleCookie
,
21 compat_etree_fromstring
,
28 compat_urllib_parse_unquote
,
29 compat_urllib_parse_urlencode
,
30 compat_urllib_request
,
32 compat_xml_parse_error
,
34 from ..downloader
import FileDownloader
35 from ..downloader
.f4m
import (
37 remove_encrypted_media
,
69 parse_m3u8_attributes
,
94 class InfoExtractor(object):
95 """Information Extractor class.
97 Information extractors are the classes that, given a URL, extract
98 information about the video (or videos) the URL refers to. This
99 information includes the real video URL, the video title, author and
100 others. The information is stored in a dictionary which is then
101 passed to the YoutubeDL. The YoutubeDL processes this
102 information possibly downloading the video to the file system, among
103 other possible outcomes.
105 The type field determines the type of the result.
106 By far the most common value (and the default if _type is missing) is
107 "video", which indicates a single video.
109 For a video, the dictionaries must include the following fields:
111 id: Video identifier.
112 title: Video title, unescaped.
114 Additionally, it must contain either a formats entry or a url one:
116 formats: A list of dictionaries for each format available, ordered
117 from worst to best quality.
120 * url The mandatory URL representing the media:
121 for plain file media - HTTP URL of this file,
123 for HLS - URL of the M3U8 media playlist,
124 for HDS - URL of the F4M manifest,
126 - HTTP URL to plain file media (in case of
128 - URL of the MPD manifest or base URL
129 representing the media if MPD manifest
130 is parsed from a string (in case of
132 for MSS - URL of the ISM manifest.
134 The URL of the manifest file in case of
136 for HLS - URL of the M3U8 master playlist,
137 for HDS - URL of the F4M manifest,
138 for DASH - URL of the MPD manifest,
139 for MSS - URL of the ISM manifest.
140 * ext Will be calculated from URL if missing
141 * format A human-readable description of the format
142 ("mp4 container with h264/opus").
143 Calculated from the format_id, width, height.
144 and format_note fields if missing.
145 * format_id A short description of the format
146 ("mp4_h264_opus" or "19").
147 Technically optional, but strongly recommended.
148 * format_note Additional info about the format
149 ("3D" or "DASH video")
150 * width Width of the video, if known
151 * height Height of the video, if known
152 * resolution Textual description of width and height
153 * dynamic_range The dynamic range of the video. One of:
154 "SDR" (None), "HDR10", "HDR10+, "HDR12", "HLG, "DV"
155 * tbr Average bitrate of audio and video in KBit/s
156 * abr Average audio bitrate in KBit/s
157 * acodec Name of the audio codec in use
158 * asr Audio sampling rate in Hertz
159 * vbr Average video bitrate in KBit/s
161 * vcodec Name of the video codec in use
162 * container Name of the container format
163 * filesize The number of bytes, if known in advance
164 * filesize_approx An estimate for the number of bytes
165 * player_url SWF Player URL (used for rtmpdump).
166 * protocol The protocol that will be used for the actual
167 download, lower-case. One of "http", "https" or
168 one of the protocols defined in downloader.PROTOCOL_MAP
170 Base URL for fragments. Each fragment's path
171 value (if present) will be relative to
173 * fragments A list of fragments of a fragmented media.
174 Each fragment entry must contain either an url
175 or a path. If an url is present it should be
176 considered by a client. Otherwise both path and
177 fragment_base_url must be present. Here is
178 the list of all potential fields:
179 * "url" - fragment's URL
180 * "path" - fragment's path relative to
182 * "duration" (optional, int or float)
183 * "filesize" (optional, int)
184 * is_from_start Is a live format that can be downloaded
185 from the start. Boolean
186 * preference Order number of this format. If this field is
187 present and not None, the formats get sorted
188 by this field, regardless of all other values.
189 -1 for default (order by other properties),
190 -2 or smaller for less than default.
191 < -1000 to hide the format (if there is
192 another one which is strictly better)
193 * language Language code, e.g. "de" or "en-US".
194 * language_preference Is this in the language mentioned in
196 10 if it's what the URL is about,
197 -1 for default (don't know),
198 -10 otherwise, other values reserved for now.
199 * quality Order number of the video quality of this
200 format, irrespective of the file format.
201 -1 for default (order by other properties),
202 -2 or smaller for less than default.
203 * source_preference Order number for this video source
204 (quality takes higher priority)
205 -1 for default (order by other properties),
206 -2 or smaller for less than default.
207 * http_headers A dictionary of additional HTTP headers
208 to add to the request.
209 * stretched_ratio If given and not 1, indicates that the
210 video's pixels are not square.
211 width : height ratio as float.
212 * no_resume The server does not support resuming the
213 (HTTP or RTMP) download. Boolean.
214 * has_drm The format has DRM and cannot be downloaded. Boolean
215 * downloader_options A dictionary of downloader options as
216 described in FileDownloader
217 RTMP formats can also have the additional fields: page_url,
218 app, play_path, tc_url, flash_version, rtmp_live, rtmp_conn,
219 rtmp_protocol, rtmp_real_time
221 url: Final video URL.
222 ext: Video filename extension.
223 format: The video format, defaults to ext (used for --get-format)
224 player_url: SWF Player URL (used for rtmpdump).
226 The following fields are optional:
228 alt_title: A secondary title of the video.
229 display_id An alternative identifier for the video, not necessarily
230 unique, but available before title. Typically, id is
231 something like "4234987", title "Dancing naked mole rats",
232 and display_id "dancing-naked-mole-rats"
233 thumbnails: A list of dictionaries, with the following entries:
234 * "id" (optional, string) - Thumbnail format ID
236 * "preference" (optional, int) - quality of the image
237 * "width" (optional, int)
238 * "height" (optional, int)
239 * "resolution" (optional, string "{width}x{height}",
241 * "filesize" (optional, int)
242 * "http_headers" (dict) - HTTP headers for the request
243 thumbnail: Full URL to a video thumbnail image.
244 description: Full video description.
245 uploader: Full name of the video uploader.
246 license: License name the video is licensed under.
247 creator: The creator of the video.
248 timestamp: UNIX timestamp of the moment the video was uploaded
249 upload_date: Video upload date (YYYYMMDD).
250 If not explicitly set, calculated from timestamp
251 release_timestamp: UNIX timestamp of the moment the video was released.
252 If it is not clear whether to use timestamp or this, use the former
253 release_date: The date (YYYYMMDD) when the video was released.
254 If not explicitly set, calculated from release_timestamp
255 modified_timestamp: UNIX timestamp of the moment the video was last modified.
256 modified_date: The date (YYYYMMDD) when the video was last modified.
257 If not explicitly set, calculated from modified_timestamp
258 uploader_id: Nickname or id of the video uploader.
259 uploader_url: Full URL to a personal webpage of the video uploader.
260 channel: Full name of the channel the video is uploaded on.
261 Note that channel fields may or may not repeat uploader
262 fields. This depends on a particular extractor.
263 channel_id: Id of the channel.
264 channel_url: Full URL to a channel webpage.
265 channel_follower_count: Number of followers of the channel.
266 location: Physical location where the video was filmed.
267 subtitles: The available subtitles as a dictionary in the format
268 {tag: subformats}. "tag" is usually a language code, and
269 "subformats" is a list sorted from lower to higher
270 preference, each element is a dictionary with the "ext"
272 * "data": The subtitles file contents
273 * "url": A URL pointing to the subtitles file
274 It can optionally also have:
275 * "name": Name or description of the subtitles
276 * http_headers: A dictionary of additional HTTP headers
277 to add to the request.
278 "ext" will be calculated from URL if missing
279 automatic_captions: Like 'subtitles'; contains automatically generated
280 captions instead of normal subtitles
281 duration: Length of the video in seconds, as an integer or float.
282 view_count: How many users have watched the video on the platform.
283 like_count: Number of positive ratings of the video
284 dislike_count: Number of negative ratings of the video
285 repost_count: Number of reposts of the video
286 average_rating: Average rating give by users, the scale used depends on the webpage
287 comment_count: Number of comments on the video
288 comments: A list of comments, each with one or more of the following
289 properties (all but one of text or html optional):
290 * "author" - human-readable name of the comment author
291 * "author_id" - user ID of the comment author
292 * "author_thumbnail" - The thumbnail of the comment author
294 * "html" - Comment as HTML
295 * "text" - Plain text of the comment
296 * "timestamp" - UNIX timestamp of comment
297 * "parent" - ID of the comment this one is replying to.
298 Set to "root" to indicate that this is a
299 comment to the original video.
300 * "like_count" - Number of positive ratings of the comment
301 * "dislike_count" - Number of negative ratings of the comment
302 * "is_favorited" - Whether the comment is marked as
303 favorite by the video uploader
304 * "author_is_uploader" - Whether the comment is made by
306 age_limit: Age restriction for the video, as an integer (years)
307 webpage_url: The URL to the video webpage, if given to yt-dlp it
308 should allow to get the same result again. (It will be set
309 by YoutubeDL if it's missing)
310 categories: A list of categories that the video falls in, for example
312 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
313 cast: A list of the video cast
314 is_live: True, False, or None (=unknown). Whether this video is a
315 live stream that goes on instead of a fixed-length video.
316 was_live: True, False, or None (=unknown). Whether this video was
317 originally a live stream.
318 live_status: 'is_live', 'is_upcoming', 'was_live', 'not_live' or None (=unknown)
319 If absent, automatically set from is_live, was_live
320 start_time: Time in seconds where the reproduction should start, as
321 specified in the URL.
322 end_time: Time in seconds where the reproduction should end, as
323 specified in the URL.
324 chapters: A list of dictionaries, with the following entries:
325 * "start_time" - The start time of the chapter in seconds
326 * "end_time" - The end time of the chapter in seconds
327 * "title" (optional, string)
328 playable_in_embed: Whether this video is allowed to play in embedded
329 players on other sites. Can be True (=always allowed),
330 False (=never allowed), None (=unknown), or a string
331 specifying the criteria for embedability (Eg: 'whitelist')
332 availability: Under what condition the video is available. One of
333 'private', 'premium_only', 'subscriber_only', 'needs_auth',
334 'unlisted' or 'public'. Use 'InfoExtractor._availability'
336 __post_extractor: A function to be called just before the metadata is
337 written to either disk, logger or console. The function
338 must return a dict which will be added to the info_dict.
339 This is usefull for additional information that is
340 time-consuming to extract. Note that the fields thus
341 extracted will not be available to output template and
342 match_filter. So, only "comments" and "comment_count" are
343 currently allowed to be extracted via this method.
345 The following fields should only be used when the video belongs to some logical
348 chapter: Name or title of the chapter the video belongs to.
349 chapter_number: Number of the chapter the video belongs to, as an integer.
350 chapter_id: Id of the chapter the video belongs to, as a unicode string.
352 The following fields should only be used when the video is an episode of some
353 series, programme or podcast:
355 series: Title of the series or programme the video episode belongs to.
356 series_id: Id of the series or programme the video episode belongs to, as a unicode string.
357 season: Title of the season the video episode belongs to.
358 season_number: Number of the season the video episode belongs to, as an integer.
359 season_id: Id of the season the video episode belongs to, as a unicode string.
360 episode: Title of the video episode. Unlike mandatory video title field,
361 this field should denote the exact title of the video episode
362 without any kind of decoration.
363 episode_number: Number of the video episode within a season, as an integer.
364 episode_id: Id of the video episode, as a unicode string.
366 The following fields should only be used when the media is a track or a part of
369 track: Title of the track.
370 track_number: Number of the track within an album or a disc, as an integer.
371 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
373 artist: Artist(s) of the track.
374 genre: Genre(s) of the track.
375 album: Title of the album the track belongs to.
376 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
377 album_artist: List of all artists appeared on the album (e.g.
378 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
380 disc_number: Number of the disc or other physical medium the track belongs to,
382 release_year: Year (YYYY) when the album was released.
383 composer: Composer of the piece
385 Unless mentioned otherwise, the fields should be Unicode strings.
387 Unless mentioned otherwise, None is equivalent to absence of information.
390 _type "playlist" indicates multiple videos.
391 There must be a key "entries", which is a list, an iterable, or a PagedList
392 object, each element of which is a valid dictionary by this specification.
394 Additionally, playlists can have "id", "title", and any other relevent
395 attributes with the same semantics as videos (see above).
397 It can also have the following optional fields:
399 playlist_count: The total number of videos in a playlist. If not given,
400 YoutubeDL tries to calculate it from "entries"
403 _type "multi_video" indicates that there are multiple videos that
404 form a single show, for examples multiple acts of an opera or TV episode.
405 It must have an entries key like a playlist and contain all the keys
406 required for a video at the same time.
409 _type "url" indicates that the video must be extracted from another
410 location, possibly by a different extractor. Its only required key is:
411 "url" - the next URL to extract.
412 The key "ie_key" can be set to the class name (minus the trailing "IE",
413 e.g. "Youtube") if the extractor class is known in advance.
414 Additionally, the dictionary may have any properties of the resolved entity
415 known in advance, for example "title" if the title of the referred video is
419 _type "url_transparent" entities have the same specification as "url", but
420 indicate that the given additional information is more precise than the one
421 associated with the resolved URL.
422 This is useful when a site employs a video service that hosts the video and
423 its technical metadata, but that video service does not embed a useful
424 title, description etc.
427 Subclasses of this one should re-define the _real_initialize() and
428 _real_extract() methods and define a _VALID_URL regexp.
429 Probably, they should also be added to the list of extractors.
431 Subclasses may also override suitable() if necessary, but ensure the function
432 signature is preserved and that this function imports everything it needs
433 (except other extractors), so that lazy_extractors works correctly
435 _GEO_BYPASS attribute may be set to False in order to disable
436 geo restriction bypass mechanisms for a particular extractor.
437 Though it won't disable explicit geo restriction bypass based on
438 country code provided with geo_bypass_country.
440 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
441 countries for this extractor. One of these countries will be used by
442 geo restriction bypass mechanism right away in order to bypass
443 geo restriction, of course, if the mechanism is not disabled.
445 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
446 IP blocks in CIDR notation for this extractor. One of these IP blocks
447 will be used by geo restriction bypass mechanism similarly
450 The _WORKING attribute should be set to False for broken IEs
451 in order to warn the users and skip the tests.
456 _x_forwarded_for_ip
= None
458 _GEO_COUNTRIES
= None
459 _GEO_IP_BLOCKS
= None
463 'any': 'Use --cookies, --username and --password, or --netrc to provide account credentials',
465 'Use --cookies-from-browser or --cookies for the authentication. '
466 'See https://github.com/ytdl-org/youtube-dl#how-do-i-pass-cookies-to-youtube-dl for how to manually pass cookies'),
467 'password': 'Use --username and --password, or --netrc to provide account credentials',
470 def __init__(self
, downloader
=None):
471 """Constructor. Receives an optional downloader (a YoutubeDL instance).
472 If a downloader is not passed during initialization,
473 it must be set using "set_downloader()" before "extract()" is called"""
475 self
._x
_forwarded
_for
_ip
= None
476 self
._printed
_messages
= set()
477 self
.set_downloader(downloader
)
480 def _match_valid_url(cls
, url
):
481 # This does not use has/getattr intentionally - we want to know whether
482 # we have cached the regexp for *this* class, whereas getattr would also
483 # match the superclass
484 if '_VALID_URL_RE' not in cls
.__dict
__:
485 if '_VALID_URL' not in cls
.__dict
__:
486 cls
._VALID
_URL
= cls
._make
_valid
_url
()
487 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
488 return cls
._VALID
_URL
_RE
.match(url
)
491 def suitable(cls
, url
):
492 """Receives a URL and returns True if suitable for this IE."""
493 # This function must import everything it needs (except other extractors),
494 # so that lazy_extractors works correctly
495 return cls
._match
_valid
_url
(url
) is not None
498 def _match_id(cls
, url
):
499 return cls
._match
_valid
_url
(url
).group('id')
502 def get_temp_id(cls
, url
):
504 return cls
._match
_id
(url
)
505 except (IndexError, AttributeError):
510 """Getter method for _WORKING."""
513 def initialize(self
):
514 """Initializes an instance (authentication, etc)."""
515 self
._printed
_messages
= set()
516 self
._initialize
_geo
_bypass
({
517 'countries': self
._GEO
_COUNTRIES
,
518 'ip_blocks': self
._GEO
_IP
_BLOCKS
,
521 self
._real
_initialize
()
524 def _initialize_geo_bypass(self
, geo_bypass_context
):
526 Initialize geo restriction bypass mechanism.
528 This method is used to initialize geo bypass mechanism based on faking
529 X-Forwarded-For HTTP header. A random country from provided country list
530 is selected and a random IP belonging to this country is generated. This
531 IP will be passed as X-Forwarded-For HTTP header in all subsequent
534 This method will be used for initial geo bypass mechanism initialization
535 during the instance initialization with _GEO_COUNTRIES and
538 You may also manually call it from extractor's code if geo bypass
539 information is not available beforehand (e.g. obtained during
540 extraction) or due to some other reason. In this case you should pass
541 this information in geo bypass context passed as first argument. It may
542 contain following fields:
544 countries: List of geo unrestricted countries (similar
546 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
547 (similar to _GEO_IP_BLOCKS)
550 if not self
._x
_forwarded
_for
_ip
:
552 # Geo bypass mechanism is explicitly disabled by user
553 if not self
.get_param('geo_bypass', True):
556 if not geo_bypass_context
:
557 geo_bypass_context
= {}
559 # Backward compatibility: previously _initialize_geo_bypass
560 # expected a list of countries, some 3rd party code may still use
562 if isinstance(geo_bypass_context
, (list, tuple)):
563 geo_bypass_context
= {
564 'countries': geo_bypass_context
,
567 # The whole point of geo bypass mechanism is to fake IP
568 # as X-Forwarded-For HTTP header based on some IP block or
571 # Path 1: bypassing based on IP block in CIDR notation
573 # Explicit IP block specified by user, use it right away
574 # regardless of whether extractor is geo bypassable or not
575 ip_block
= self
.get_param('geo_bypass_ip_block', None)
577 # Otherwise use random IP block from geo bypass context but only
578 # if extractor is known as geo bypassable
580 ip_blocks
= geo_bypass_context
.get('ip_blocks')
581 if self
._GEO
_BYPASS
and ip_blocks
:
582 ip_block
= random
.choice(ip_blocks
)
585 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(ip_block
)
586 self
._downloader
.write_debug(
587 '[debug] Using fake IP %s as X-Forwarded-For' % self
._x
_forwarded
_for
_ip
)
590 # Path 2: bypassing based on country code
592 # Explicit country code specified by user, use it right away
593 # regardless of whether extractor is geo bypassable or not
594 country
= self
.get_param('geo_bypass_country', None)
596 # Otherwise use random country code from geo bypass context but
597 # only if extractor is known as geo bypassable
599 countries
= geo_bypass_context
.get('countries')
600 if self
._GEO
_BYPASS
and countries
:
601 country
= random
.choice(countries
)
604 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country
)
605 self
._downloader
.write_debug(
606 'Using fake IP %s (%s) as X-Forwarded-For' % (self
._x
_forwarded
_for
_ip
, country
.upper()))
608 def extract(self
, url
):
609 """Extracts URL information and returns it in list of dicts."""
614 self
.write_debug('Extracting URL: %s' % url
)
615 ie_result
= self
._real
_extract
(url
)
616 if ie_result
is None:
618 if self
._x
_forwarded
_for
_ip
:
619 ie_result
['__x_forwarded_for_ip'] = self
._x
_forwarded
_for
_ip
620 subtitles
= ie_result
.get('subtitles')
621 if (subtitles
and 'live_chat' in subtitles
622 and 'no-live-chat' in self
.get_param('compat_opts', [])):
623 del subtitles
['live_chat']
625 except GeoRestrictedError
as e
:
626 if self
.__maybe
_fake
_ip
_and
_retry
(e
.countries
):
629 except UnsupportedError
:
631 except ExtractorError
as e
:
633 'video_id': e
.video_id
or self
.get_temp_id(url
),
635 'tb': e
.traceback
or sys
.exc_info()[2],
636 'expected': e
.expected
,
639 if hasattr(e
, 'countries'):
640 kwargs
['countries'] = e
.countries
641 raise type(e
)(e
.msg
, **kwargs
)
642 except compat_http_client
.IncompleteRead
as e
:
643 raise ExtractorError('A network error has occurred.', cause
=e
, expected
=True, video_id
=self
.get_temp_id(url
))
644 except (KeyError, StopIteration) as e
:
645 raise ExtractorError('An extractor error has occurred.', cause
=e
, video_id
=self
.get_temp_id(url
))
647 def __maybe_fake_ip_and_retry(self
, countries
):
648 if (not self
.get_param('geo_bypass_country', None)
650 and self
.get_param('geo_bypass', True)
651 and not self
._x
_forwarded
_for
_ip
653 country_code
= random
.choice(countries
)
654 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country_code
)
655 if self
._x
_forwarded
_for
_ip
:
657 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
658 % (self
._x
_forwarded
_for
_ip
, country_code
.upper()))
662 def set_downloader(self
, downloader
):
663 """Sets the downloader for this IE."""
664 self
._downloader
= downloader
666 def _real_initialize(self
):
667 """Real initialization process. Redefine in subclasses."""
670 def _real_extract(self
, url
):
671 """Real extraction process. Redefine in subclasses."""
676 """A string for getting the InfoExtractor with get_info_extractor"""
677 return cls
.__name
__[:-2]
681 return compat_str(type(self
).__name
__[:-2])
684 def __can_accept_status_code(err
, expected_status
):
685 assert isinstance(err
, compat_urllib_error
.HTTPError
)
686 if expected_status
is None:
688 elif callable(expected_status
):
689 return expected_status(err
.code
) is True
691 return err
.code
in variadic(expected_status
)
693 def _request_webpage(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
, expected_status
=None):
695 Return the response handle.
697 See _download_webpage docstring for arguments specification.
699 if not self
._downloader
._first
_webpage
_request
:
700 sleep_interval
= self
.get_param('sleep_interval_requests') or 0
701 if sleep_interval
> 0:
702 self
.to_screen('Sleeping %s seconds ...' % sleep_interval
)
703 time
.sleep(sleep_interval
)
705 self
._downloader
._first
_webpage
_request
= False
708 self
.report_download_webpage(video_id
)
709 elif note
is not False:
711 self
.to_screen('%s' % (note
,))
713 self
.to_screen('%s: %s' % (video_id
, note
))
715 # Some sites check X-Forwarded-For HTTP header in order to figure out
716 # the origin of the client behind proxy. This allows bypassing geo
717 # restriction by faking this header's value to IP that belongs to some
718 # geo unrestricted country. We will do so once we encounter any
719 # geo restriction error.
720 if self
._x
_forwarded
_for
_ip
:
721 if 'X-Forwarded-For' not in headers
:
722 headers
['X-Forwarded-For'] = self
._x
_forwarded
_for
_ip
724 if isinstance(url_or_request
, compat_urllib_request
.Request
):
725 url_or_request
= update_Request(
726 url_or_request
, data
=data
, headers
=headers
, query
=query
)
729 url_or_request
= update_url_query(url_or_request
, query
)
730 if data
is not None or headers
:
731 url_or_request
= sanitized_Request(url_or_request
, data
, headers
)
733 return self
._downloader
.urlopen(url_or_request
)
734 except network_exceptions
as err
:
735 if isinstance(err
, compat_urllib_error
.HTTPError
):
736 if self
.__can
_accept
_status
_code
(err
, expected_status
):
737 # Retain reference to error to prevent file object from
738 # being closed before it can be read. Works around the
739 # effects of <https://bugs.python.org/issue15002>
740 # introduced in Python 3.4.1.
747 errnote
= 'Unable to download webpage'
749 errmsg
= '%s: %s' % (errnote
, error_to_compat_str(err
))
751 raise ExtractorError(errmsg
, sys
.exc_info()[2], cause
=err
)
753 self
.report_warning(errmsg
)
756 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):
758 Return a tuple (page content as string, URL handle).
760 See _download_webpage docstring for arguments specification.
762 # Strip hashes from the URL (#1038)
763 if isinstance(url_or_request
, (compat_str
, str)):
764 url_or_request
= url_or_request
.partition('#')[0]
766 urlh
= self
._request
_webpage
(url_or_request
, video_id
, note
, errnote
, fatal
, data
=data
, headers
=headers
, query
=query
, expected_status
=expected_status
)
770 content
= self
._webpage
_read
_content
(urlh
, url_or_request
, video_id
, note
, errnote
, fatal
, encoding
=encoding
)
771 return (content
, urlh
)
774 def _guess_encoding_from_content(content_type
, webpage_bytes
):
775 m
= re
.match(r
'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type
)
777 encoding
= m
.group(1)
779 m
= re
.search(br
'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
780 webpage_bytes[:1024])
782 encoding = m.group(1).decode('ascii')
783 elif webpage_bytes.startswith(b'\xff\xfe'):
790 def __check_blocked(self, content):
791 first_block = content[:512]
792 if ('<title>Access to this site is blocked</title>' in content
793 and 'Websense' in first_block):
794 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
795 blocked_iframe = self._html_search_regex(
796 r'<iframe src="([^
"]+)"', content,
797 'Websense information URL
', default=None)
799 msg += ' Visit
%s for more details
' % blocked_iframe
800 raise ExtractorError(msg, expected=True)
801 if '<title
>The URL you requested has been blocked
</title
>' in first_block:
803 'Access to this webpage has been blocked by Indian censorship
. '
804 'Use a VPN
or proxy
server (with --proxy
) to route around it
.')
805 block_msg = self._html_search_regex(
806 r'</h1
><p
>(.*?
)</p
>',
807 content, 'block message
', default=None)
809 msg += ' (Message
: "%s")' % block_msg.replace('\n', ' ')
810 raise ExtractorError(msg, expected=True)
811 if ('<title
>TTK
:: Доступ к ресурсу ограничен
</title
>' in content
812 and 'blocklist
.rkn
.gov
.ru
' in content):
813 raise ExtractorError(
814 'Access to this webpage has been blocked by decision of the Russian government
. '
815 'Visit http
://blocklist
.rkn
.gov
.ru
/ for a block reason
.',
818 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
819 content_type = urlh.headers.get('Content
-Type
', '')
820 webpage_bytes = urlh.read()
821 if prefix is not None:
822 webpage_bytes = prefix + webpage_bytes
824 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
825 if self.get_param('dump_intermediate_pages
', False):
826 self.to_screen('Dumping request to
' + urlh.geturl())
827 dump = base64.b64encode(webpage_bytes).decode('ascii
')
828 self._downloader.to_screen(dump)
829 if self.get_param('write_pages
', False):
830 basen = '%s_%s' % (video_id, urlh.geturl())
831 trim_length = self.get_param('trim_file_name
') or 240
832 if len(basen) > trim_length:
833 h = '___
' + hashlib.md5(basen.encode('utf
-8')).hexdigest()
834 basen = basen[:trim_length - len(h)] + h
835 raw_filename = basen + '.dump
'
836 filename = sanitize_filename(raw_filename, restricted=True)
837 self.to_screen('Saving request to
' + filename)
838 # Working around MAX_PATH limitation on Windows (see
839 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
840 if compat_os_name == 'nt
':
841 absfilepath = os.path.abspath(filename)
842 if len(absfilepath) > 259:
843 filename = '\\\\?
\\' + absfilepath
844 with open(filename, 'wb
') as outf:
845 outf.write(webpage_bytes)
848 content = webpage_bytes.decode(encoding, 'replace
')
850 content = webpage_bytes.decode('utf
-8', 'replace
')
852 self.__check_blocked(content)
856 def _download_webpage(
857 self, url_or_request, video_id, note=None, errnote=None,
858 fatal=True, tries=1, timeout=5, encoding=None, data=None,
859 headers={}, query={}, expected_status=None):
861 Return the data of the page as a string.
864 url_or_request -- plain text URL as a string or
865 a compat_urllib_request.Requestobject
866 video_id -- Video/playlist/item identifier (string)
869 note -- note printed before downloading (string)
870 errnote -- note printed in case of an error (string)
871 fatal -- flag denoting whether error should be considered fatal,
872 i.e. whether it should cause ExtractionError to be raised,
873 otherwise a warning will be reported and extraction continued
874 tries -- number of tries
875 timeout -- sleep interval between tries
876 encoding -- encoding for a page content decoding, guessed automatically
877 when not explicitly specified
878 data -- POST data (bytes)
879 headers -- HTTP headers (dict)
880 query -- URL query (dict)
881 expected_status -- allows to accept failed HTTP requests (non 2xx
882 status code) by explicitly specifying a set of accepted status
883 codes. Can be any of the following entities:
884 - an integer type specifying an exact failed status code to
886 - a list or a tuple of integer types specifying a list of
887 failed status codes to accept
888 - a callable accepting an actual failed status code and
889 returning True if it should be accepted
890 Note that this argument does not affect success status codes (2xx)
891 which are always accepted.
896 while success is False:
898 res = self._download_webpage_handle(
899 url_or_request, video_id, note, errnote, fatal,
900 encoding=encoding, data=data, headers=headers, query=query,
901 expected_status=expected_status)
903 except compat_http_client.IncompleteRead as e:
905 if try_count >= tries:
907 self._sleep(timeout, video_id)
914 def _download_xml_handle(
915 self, url_or_request, video_id, note='Downloading XML
',
916 errnote='Unable to download XML
', transform_source=None,
917 fatal=True, encoding=None, data=None, headers={}, query={},
918 expected_status=None):
920 Return a tuple (xml as an compat_etree_Element, URL handle).
922 See _download_webpage docstring for arguments specification.
924 res = self._download_webpage_handle(
925 url_or_request, video_id, note, errnote, fatal=fatal,
926 encoding=encoding, data=data, headers=headers, query=query,
927 expected_status=expected_status)
930 xml_string, urlh = res
931 return self._parse_xml(
932 xml_string, video_id, transform_source=transform_source,
936 self, url_or_request, video_id,
937 note='Downloading XML
', errnote='Unable to download XML
',
938 transform_source=None, fatal=True, encoding=None,
939 data=None, headers={}, query={}, expected_status=None):
941 Return the xml as an compat_etree_Element.
943 See _download_webpage docstring for arguments specification.
945 res = self._download_xml_handle(
946 url_or_request, video_id, note=note, errnote=errnote,
947 transform_source=transform_source, fatal=fatal, encoding=encoding,
948 data=data, headers=headers, query=query,
949 expected_status=expected_status)
950 return res if res is False else res[0]
952 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
954 xml_string = transform_source(xml_string)
956 return compat_etree_fromstring(xml_string.encode('utf
-8'))
957 except compat_xml_parse_error as ve:
958 errmsg = '%s: Failed to parse XML
' % video_id
960 raise ExtractorError(errmsg, cause=ve)
962 self.report_warning(errmsg + str(ve))
964 def _download_json_handle(
965 self, url_or_request, video_id, note='Downloading JSON metadata
',
966 errnote='Unable to download JSON metadata
', transform_source=None,
967 fatal=True, encoding=None, data=None, headers={}, query={},
968 expected_status=None):
970 Return a tuple (JSON object, URL handle).
972 See _download_webpage docstring for arguments specification.
974 res = self._download_webpage_handle(
975 url_or_request, video_id, note, errnote, fatal=fatal,
976 encoding=encoding, data=data, headers=headers, query=query,
977 expected_status=expected_status)
980 json_string, urlh = res
981 return self._parse_json(
982 json_string, video_id, transform_source=transform_source,
986 self, url_or_request, video_id, note='Downloading JSON metadata
',
987 errnote='Unable to download JSON metadata
', transform_source=None,
988 fatal=True, encoding=None, data=None, headers={}, query={},
989 expected_status=None):
991 Return the JSON object as a dict.
993 See _download_webpage docstring for arguments specification.
995 res = self._download_json_handle(
996 url_or_request, video_id, note=note, errnote=errnote,
997 transform_source=transform_source, fatal=fatal, encoding=encoding,
998 data=data, headers=headers, query=query,
999 expected_status=expected_status)
1000 return res if res is False else res[0]
1002 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
1003 if transform_source:
1004 json_string = transform_source(json_string)
1006 return json.loads(json_string)
1007 except ValueError as ve:
1008 errmsg = '%s: Failed to parse JSON
' % video_id
1010 raise ExtractorError(errmsg, cause=ve)
1012 self.report_warning(errmsg + str(ve))
1014 def _parse_socket_response_as_json(self, data, video_id, transform_source=None, fatal=True):
1015 return self._parse_json(
1016 data[data.find('{'):data.rfind('}
') + 1],
1017 video_id, transform_source, fatal)
1019 def _download_socket_json_handle(
1020 self, url_or_request, video_id, note='Polling socket
',
1021 errnote='Unable to poll socket
', transform_source=None,
1022 fatal=True, encoding=None, data=None, headers={}, query={},
1023 expected_status=None):
1025 Return a tuple (JSON object, URL handle).
1027 See _download_webpage docstring for arguments specification.
1029 res = self._download_webpage_handle(
1030 url_or_request, video_id, note, errnote, fatal=fatal,
1031 encoding=encoding, data=data, headers=headers, query=query,
1032 expected_status=expected_status)
1036 return self._parse_socket_response_as_json(
1037 webpage, video_id, transform_source=transform_source,
1040 def _download_socket_json(
1041 self, url_or_request, video_id, note='Polling socket
',
1042 errnote='Unable to poll socket
', transform_source=None,
1043 fatal=True, encoding=None, data=None, headers={}, query={},
1044 expected_status=None):
1046 Return the JSON object as a dict.
1048 See _download_webpage docstring for arguments specification.
1050 res = self._download_socket_json_handle(
1051 url_or_request, video_id, note=note, errnote=errnote,
1052 transform_source=transform_source, fatal=fatal, encoding=encoding,
1053 data=data, headers=headers, query=query,
1054 expected_status=expected_status)
1055 return res if res is False else res[0]
1057 def report_warning(self, msg, video_id=None, *args, only_once=False, **kwargs):
1058 idstr = format_field(video_id, template='%s: ')
1059 msg = f'[{self.IE_NAME}
] {idstr}{msg}
'
1061 if f'WARNING
: {msg}
' in self._printed_messages:
1063 self._printed_messages.add(f'WARNING
: {msg}
')
1064 self._downloader.report_warning(msg, *args, **kwargs)
1066 def to_screen(self, msg, *args, **kwargs):
1067 """Print msg to screen, prefixing it with '[ie_name
]'"""
1068 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg), *args, **kwargs)
1070 def write_debug(self, msg, *args, **kwargs):
1071 self._downloader.write_debug('[%s] %s' % (self.IE_NAME, msg), *args, **kwargs)
1073 def get_param(self, name, default=None, *args, **kwargs):
1074 if self._downloader:
1075 return self._downloader.params.get(name, default, *args, **kwargs)
1078 def report_drm(self, video_id, partial=False):
1079 self.raise_no_formats('This video
is DRM protected
', expected=True, video_id=video_id)
1081 def report_extraction(self, id_or_name):
1082 """Report information extraction."""
1083 self.to_screen('%s: Extracting information
' % id_or_name)
1085 def report_download_webpage(self, video_id):
1086 """Report webpage download."""
1087 self.to_screen('%s: Downloading webpage
' % video_id)
1089 def report_age_confirmation(self):
1090 """Report attempt to confirm age."""
1091 self.to_screen('Confirming age
')
1093 def report_login(self):
1094 """Report attempt to log in."""
1095 self.to_screen('Logging
in')
1097 def raise_login_required(
1098 self, msg='This video
is only available
for registered users
',
1099 metadata_available=False, method='any
'):
1100 if metadata_available and (
1101 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1102 self.report_warning(msg)
1103 if method is not None:
1104 msg = '%s. %s' % (msg, self._LOGIN_HINTS[method])
1105 raise ExtractorError(msg, expected=True)
1107 def raise_geo_restricted(
1108 self, msg='This video
is not available
from your location due to geo restriction
',
1109 countries=None, metadata_available=False):
1110 if metadata_available and (
1111 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1112 self.report_warning(msg)
1114 raise GeoRestrictedError(msg, countries=countries)
1116 def raise_no_formats(self, msg, expected=False, video_id=None):
1118 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1119 self.report_warning(msg, video_id)
1120 elif isinstance(msg, ExtractorError):
1123 raise ExtractorError(msg, expected=expected, video_id=video_id)
1125 # Methods for following #608
1127 def url_result(url, ie=None, video_id=None, video_title=None, *, url_transparent=False, **kwargs):
1128 """Returns a URL that points to a page that should be processed"""
1130 kwargs['ie_key
'] = ie if isinstance(ie, str) else ie.ie_key()
1131 if video_id is not None:
1132 kwargs['id'] = video_id
1133 if video_title is not None:
1134 kwargs['title
'] = video_title
1137 '_type
': 'url_transparent
' if url_transparent else 'url
',
1141 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None, **kwargs):
1142 urls = (self.url_result(self._proto_relative_url(m), ie)
1143 for m in orderedSet(map(getter, matches) if getter else matches))
1144 return self.playlist_result(urls, playlist_id, playlist_title, **kwargs)
1147 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None, *, multi_video=False, **kwargs):
1148 """Returns a playlist"""
1150 kwargs['id'] = playlist_id
1152 kwargs['title
'] = playlist_title
1153 if playlist_description is not None:
1154 kwargs['description
'] = playlist_description
1157 '_type
': 'multi_video
' if multi_video else 'playlist
',
1161 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1163 Perform a regex search on the given string, using a single or a list of
1164 patterns returning the first matching group.
1165 In case of failure return a default value or raise a WARNING or a
1166 RegexNotFoundError, depending on fatal, specifying the field name.
1168 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
1169 mobj = re.search(pattern, string, flags)
1172 mobj = re.search(p, string, flags)
1176 _name = self._downloader._format_err(name, self._downloader.Styles.EMPHASIS)
1180 # return the first matching group
1181 return next(g for g in mobj.groups() if g is not None)
1182 elif isinstance(group, (list, tuple)):
1183 return tuple(mobj.group(g) for g in group)
1185 return mobj.group(group)
1186 elif default is not NO_DEFAULT:
1189 raise RegexNotFoundError('Unable to extract
%s' % _name)
1191 self.report_warning('unable to extract
%s' % _name + bug_reports_message())
1194 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1196 Like _search_regex, but strips HTML tags and unescapes entities.
1198 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1200 return clean_html(res).strip()
1204 def _get_netrc_login_info(self, netrc_machine=None):
1207 netrc_machine = netrc_machine or self._NETRC_MACHINE
1209 if self.get_param('usenetrc
', False):
1211 netrc_file = compat_expanduser(self.get_param('netrc_location
') or '~
')
1212 if os.path.isdir(netrc_file):
1213 netrc_file = os.path.join(netrc_file, '.netrc
')
1214 info = netrc.netrc(file=netrc_file).authenticators(netrc_machine)
1215 if info is not None:
1219 raise netrc.NetrcParseError(
1220 'No authenticators
for %s' % netrc_machine)
1221 except (IOError, netrc.NetrcParseError) as err:
1222 self.report_warning(
1223 'parsing
.netrc
: %s' % error_to_compat_str(err))
1225 return username, password
1227 def _get_login_info(self, username_option='username
', password_option='password
', netrc_machine=None):
1229 Get the login info as (username, password)
1230 First look for the manually specified credentials using username_option
1231 and password_option as keys in params dictionary. If no such credentials
1232 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1234 If there's no info available
, return (None, None)
1237 # Attempt to use provided username and password or .netrc data
1238 username = self.get_param(username_option)
1239 if username is not None:
1240 password = self.get_param(password_option)
1242 username, password = self._get_netrc_login_info(netrc_machine)
1244 return username, password
1246 def _get_tfa_info(self, note='two-factor verification code'):
1248 Get the two
-factor authentication info
1249 TODO
- asking the user will be required
for sms
/phone verify
1250 currently just uses the command line option
1251 If there
's no info available, return None
1254 tfa = self.get_param('twofactor
')
1258 return compat_getpass('Type
%s and press
[Return
]: ' % note)
1260 # Helper functions for extracting OpenGraph info
1262 def _og_regexes(prop):
1263 content_re = r'content
=(?
:"([^"]+?
)"|\'([^\']+?)\'|\s*([^\s"\'=<>`
]+?
))'
1264 property_re = (r'(?
:name|
property)=(?
:\'og
[:-]%(prop)s\'|
"og[:-]%(prop)s"|\s
*og
[:-]%(prop)s\b)'
1265 % {'prop': re.escape(prop)})
1266 template = r'<meta
[^
>]+?
%s[^
>]+?
%s'
1268 template % (property_re, content_re),
1269 template % (content_re, property_re),
1273 def _meta_regex(prop):
1274 return r'''(?isx)<meta
1275 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1276 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1278 def _og_search_property(self, prop, html, name=None, **kargs):
1279 prop = variadic(prop)
1281 name = 'OpenGraph
%s' % prop[0]
1284 og_regexes.extend(self._og_regexes(p))
1285 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1288 return unescapeHTML(escaped)
1290 def _og_search_thumbnail(self, html, **kargs):
1291 return self._og_search_property('image
', html, 'thumbnail URL
', fatal=False, **kargs)
1293 def _og_search_description(self, html, **kargs):
1294 return self._og_search_property('description
', html, fatal=False, **kargs)
1296 def _og_search_title(self, html, **kargs):
1297 kargs.setdefault('fatal
', False)
1298 return self._og_search_property('title
', html, **kargs)
1300 def _og_search_video_url(self, html, name='video url
', secure=True, **kargs):
1301 regexes = self._og_regexes('video
') + self._og_regexes('video
:url
')
1303 regexes = self._og_regexes('video
:secure_url
') + regexes
1304 return self._html_search_regex(regexes, html, name, **kargs)
1306 def _og_search_url(self, html, **kargs):
1307 return self._og_search_property('url
', html, **kargs)
1309 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1310 name = variadic(name)
1311 if display_name is None:
1312 display_name = name[0]
1313 return self._html_search_regex(
1314 [self._meta_regex(n) for n in name],
1315 html, display_name, fatal=fatal, group='content
', **kwargs)
1317 def _dc_search_uploader(self, html):
1318 return self._html_search_meta('dc
.creator
', html, 'uploader
')
1320 def _rta_search(self, html):
1321 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1322 if re.search(r'(?ix
)<meta\s
+name
="rating"\s
+'
1323 r' content
="RTA-5042-1996-1400-1577-RTA"',
1328 def _media_rating_search(self, html):
1329 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1330 rating = self._html_search_meta('rating
', html)
1342 return RATING_TABLE.get(rating.lower())
1344 def _family_friendly_search(self, html):
1345 # See http://schema.org/VideoObject
1346 family_friendly = self._html_search_meta(
1347 'isFamilyFriendly
', html, default=None)
1349 if not family_friendly:
1358 return RATING_TABLE.get(family_friendly.lower())
1360 def _twitter_search_player(self, html):
1361 return self._html_search_meta('twitter
:player
', html,
1362 'twitter card player
')
1364 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
1365 json_ld_list = list(re.finditer(JSON_LD_RE, html))
1366 default = kwargs.get('default
', NO_DEFAULT)
1367 # JSON-LD may be malformed and thus `fatal` should be respected.
1368 # At the same time `default` may be passed that assumes `fatal=False`
1369 # for _search_regex. Let's simulate the same behavior here
as well
.
1370 fatal
= kwargs
.get('fatal', True) if default
is NO_DEFAULT
else False
1372 for mobj
in json_ld_list
:
1373 json_ld_item
= self
._parse
_json
(
1374 mobj
.group('json_ld'), video_id
, fatal
=fatal
)
1375 if not json_ld_item
:
1377 if isinstance(json_ld_item
, dict):
1378 json_ld
.append(json_ld_item
)
1379 elif isinstance(json_ld_item
, (list, tuple)):
1380 json_ld
.extend(json_ld_item
)
1382 json_ld
= self
._json
_ld
(json_ld
, video_id
, fatal
=fatal
, expected_type
=expected_type
)
1385 if default
is not NO_DEFAULT
:
1388 raise RegexNotFoundError('Unable to extract JSON-LD')
1390 self
.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1393 def _json_ld(self
, json_ld
, video_id
, fatal
=True, expected_type
=None):
1394 if isinstance(json_ld
, compat_str
):
1395 json_ld
= self
._parse
_json
(json_ld
, video_id
, fatal
=fatal
)
1399 if not isinstance(json_ld
, (list, tuple, dict)):
1401 if isinstance(json_ld
, dict):
1404 INTERACTION_TYPE_MAP
= {
1405 'CommentAction': 'comment',
1406 'AgreeAction': 'like',
1407 'DisagreeAction': 'dislike',
1408 'LikeAction': 'like',
1409 'DislikeAction': 'dislike',
1410 'ListenAction': 'view',
1411 'WatchAction': 'view',
1412 'ViewAction': 'view',
1415 def extract_interaction_type(e
):
1416 interaction_type
= e
.get('interactionType')
1417 if isinstance(interaction_type
, dict):
1418 interaction_type
= interaction_type
.get('@type')
1419 return str_or_none(interaction_type
)
1421 def extract_interaction_statistic(e
):
1422 interaction_statistic
= e
.get('interactionStatistic')
1423 if isinstance(interaction_statistic
, dict):
1424 interaction_statistic
= [interaction_statistic
]
1425 if not isinstance(interaction_statistic
, list):
1427 for is_e
in interaction_statistic
:
1428 if not isinstance(is_e
, dict):
1430 if is_e
.get('@type') != 'InteractionCounter':
1432 interaction_type
= extract_interaction_type(is_e
)
1433 if not interaction_type
:
1435 # For interaction count some sites provide string instead of
1436 # an integer (as per spec) with non digit characters (e.g. ",")
1437 # so extracting count with more relaxed str_to_int
1438 interaction_count
= str_to_int(is_e
.get('userInteractionCount'))
1439 if interaction_count
is None:
1441 count_kind
= INTERACTION_TYPE_MAP
.get(interaction_type
.split('/')[-1])
1444 count_key
= '%s_count' % count_kind
1445 if info
.get(count_key
) is not None:
1447 info
[count_key
] = interaction_count
1449 def extract_chapter_information(e
):
1451 'title': part
.get('name'),
1452 'start_time': part
.get('startOffset'),
1453 'end_time': part
.get('endOffset'),
1454 } for part
in variadic(e
.get('hasPart') or []) if part
.get('@type') == 'Clip']
1455 for idx
, (last_c
, current_c
, next_c
) in enumerate(zip(
1456 [{'end_time': 0}
] + chapters
, chapters
, chapters
[1:])):
1457 current_c
['end_time'] = current_c
['end_time'] or next_c
['start_time']
1458 current_c
['start_time'] = current_c
['start_time'] or last_c
['end_time']
1459 if None in current_c
.values():
1460 self
.report_warning(f
'Chapter {idx} contains broken data. Not extracting chapters')
1463 chapters
[-1]['end_time'] = chapters
[-1]['end_time'] or info
['duration']
1464 info
['chapters'] = chapters
1466 def extract_video_object(e
):
1467 assert e
['@type'] == 'VideoObject'
1468 author
= e
.get('author')
1470 'url': url_or_none(e
.get('contentUrl')),
1471 'title': unescapeHTML(e
.get('name')),
1472 'description': unescapeHTML(e
.get('description')),
1473 'thumbnails': [{'url': url_or_none(url)}
1474 for url
in variadic(traverse_obj(e
, 'thumbnailUrl', 'thumbnailURL'))],
1475 'duration': parse_duration(e
.get('duration')),
1476 'timestamp': unified_timestamp(e
.get('uploadDate')),
1477 # author can be an instance of 'Organization' or 'Person' types.
1478 # both types can have 'name' property(inherited from 'Thing' type). [1]
1479 # however some websites are using 'Text' type instead.
1480 # 1. https://schema.org/VideoObject
1481 'uploader': author
.get('name') if isinstance(author
, dict) else author
if isinstance(author
, compat_str
) else None,
1482 'filesize': float_or_none(e
.get('contentSize')),
1483 'tbr': int_or_none(e
.get('bitrate')),
1484 'width': int_or_none(e
.get('width')),
1485 'height': int_or_none(e
.get('height')),
1486 'view_count': int_or_none(e
.get('interactionCount')),
1488 extract_interaction_statistic(e
)
1489 extract_chapter_information(e
)
1491 def traverse_json_ld(json_ld
, at_top_level
=True):
1493 if at_top_level
and '@context' not in e
:
1495 if at_top_level
and set(e
.keys()) == {'@context', '@graph'}
:
1496 traverse_json_ld(variadic(e
['@graph'], allowed_types
=(dict,)), at_top_level
=False)
1498 item_type
= e
.get('@type')
1499 if expected_type
is not None and expected_type
!= item_type
:
1501 rating
= traverse_obj(e
, ('aggregateRating', 'ratingValue'), expected_type
=float_or_none
)
1502 if rating
is not None:
1503 info
['average_rating'] = rating
1504 if item_type
in ('TVEpisode', 'Episode'):
1505 episode_name
= unescapeHTML(e
.get('name'))
1507 'episode': episode_name
,
1508 'episode_number': int_or_none(e
.get('episodeNumber')),
1509 'description': unescapeHTML(e
.get('description')),
1511 if not info
.get('title') and episode_name
:
1512 info
['title'] = episode_name
1513 part_of_season
= e
.get('partOfSeason')
1514 if isinstance(part_of_season
, dict) and part_of_season
.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1516 'season': unescapeHTML(part_of_season
.get('name')),
1517 'season_number': int_or_none(part_of_season
.get('seasonNumber')),
1519 part_of_series
= e
.get('partOfSeries') or e
.get('partOfTVSeries')
1520 if isinstance(part_of_series
, dict) and part_of_series
.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1521 info
['series'] = unescapeHTML(part_of_series
.get('name'))
1522 elif item_type
== 'Movie':
1524 'title': unescapeHTML(e
.get('name')),
1525 'description': unescapeHTML(e
.get('description')),
1526 'duration': parse_duration(e
.get('duration')),
1527 'timestamp': unified_timestamp(e
.get('dateCreated')),
1529 elif item_type
in ('Article', 'NewsArticle'):
1531 'timestamp': parse_iso8601(e
.get('datePublished')),
1532 'title': unescapeHTML(e
.get('headline')),
1533 'description': unescapeHTML(e
.get('articleBody') or e
.get('description')),
1535 if traverse_obj(e
, ('video', 0, '@type')) == 'VideoObject':
1536 extract_video_object(e
['video'][0])
1537 elif item_type
== 'VideoObject':
1538 extract_video_object(e
)
1539 if expected_type
is None:
1543 video
= e
.get('video')
1544 if isinstance(video
, dict) and video
.get('@type') == 'VideoObject':
1545 extract_video_object(video
)
1546 if expected_type
is None:
1550 traverse_json_ld(json_ld
)
1552 return dict((k
, v
) for k
, v
in info
.items() if v
is not None)
1554 def _search_nextjs_data(self
, webpage
, video_id
, *, transform_source
=None, fatal
=True, **kw
):
1555 return self
._parse
_json
(
1557 r
'(?s)<script[^>]+id=[\'"]__NEXT_DATA__[\'"][^
>]*>([^
<]+)</script
>',
1558 webpage, 'next
.js data
', fatal=fatal, **kw),
1559 video_id, transform_source=transform_source, fatal=fatal)
1561 def _search_nuxt_data(self, webpage, video_id, context_name='__NUXT__
'):
1562 ''' Parses Nuxt.js metadata. This works as long as the function __NUXT__ invokes is a pure function. '''
1563 # not all website do this, but it can be changed
1564 # https://stackoverflow.com/questions/67463109/how-to-change-or-hide-nuxt-and-nuxt-keyword-in-page-source
1565 rectx = re.escape(context_name)
1566 js, arg_keys, arg_vals = self._search_regex(
1567 (r'<script
>window\
.%s=\
(function\
((?P
<arg_keys
>.*?
)\
)\{return\s(?P<js>\{.*?\}
)\
}\
((?P
<arg_vals
>.+?
)\
)\
);?
</script
>' % rectx,
1568 r'%s\
(.*?\
(function\
((?P
<arg_keys
>.*?
)\
)\{return\s(?P<js>\{.*?\}
)\
}\
((?P
<arg_vals
>.*?
)\
)' % rectx),
1569 webpage, context_name, group=['js
', 'arg_keys
', 'arg_vals
'])
1571 args = dict(zip(arg_keys.split(','), arg_vals.split(',')))
1573 for key, val in args.items():
1574 if val in ('undefined
', 'void
0'):
1577 return self._parse_json(js_to_json(js, args), video_id)['data
'][0]
1580 def _hidden_inputs(html):
1581 html = re.sub(r'<!--(?
:(?
!<!--).)*-->', '', html)
1583 for input in re.findall(r'(?i
)(<input[^
>]+>)', html):
1584 attrs = extract_attributes(input)
1587 if attrs.get('type') not in ('hidden
', 'submit
'):
1589 name = attrs.get('name
') or attrs.get('id')
1590 value = attrs.get('value
')
1591 if name and value is not None:
1592 hidden_inputs[name] = value
1593 return hidden_inputs
1595 def _form_hidden_inputs(self, form_id, html):
1596 form = self._search_regex(
1597 r'(?
is)<form
[^
>]+?
id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1598 html, '%s form' % form_id, group='form')
1599 return self._hidden_inputs(form)
1602 regex = r' *((?P<reverse>\+)?(?P<field>[a-zA-Z0-9_]+)((?P<separator>[~:])(?P<limit>.*?))?)? *$'
1604 default = ('hidden', 'aud_or_vid', 'hasvid', 'ie_pref', 'lang', 'quality',
1605 'res', 'fps', 'hdr:12', 'codec:vp9.2', 'size', 'br', 'asr',
1606 'proto', 'ext', 'hasaud', 'source', 'id') # These must not be aliases
1607 ytdl_default = ('hasaud', 'lang', 'quality', 'tbr', 'filesize', 'vbr',
1608 'height', 'width', 'proto', 'vext', 'abr', 'aext',
1609 'fps', 'fs_approx', 'source', 'id')
1612 'vcodec': {'type': 'ordered', 'regex': True,
1613 'order': ['av0?1', 'vp0?9.2', 'vp0?9', '[hx]265|he?vc?', '[hx]264|avc', 'vp0?8', 'mp4v|h263', 'theora', '', None, 'none']},
1614 'acodec': {'type': 'ordered', 'regex': True,
1615 'order': ['[af]lac', 'wav|aiff', 'opus', 'vorbis', 'aac', 'mp?4a?', 'mp3', 'e-?a?c-?3', 'ac-?3', 'dts', '', None, 'none']},
1616 'hdr': {'type': 'ordered', 'regex': True, 'field': 'dynamic_range',
1617 'order': ['dv', '(hdr)?12', r'(hdr)?10\+', '(hdr)?10', 'hlg', '', 'sdr', None]},
1618 'proto': {'type': 'ordered', 'regex': True, 'field': 'protocol',
1619 'order': ['(ht|f)tps', '(ht|f)tp$', 'm3u8.*', '.*dash', 'websocket_frag', 'rtmpe?', '', 'mms|rtsp', 'ws|websocket', 'f4']},
1620 'vext': {'type': 'ordered', 'field': 'video_ext',
1621 'order': ('mp4', 'webm', 'flv', '', 'none'),
1622 'order_free': ('webm', 'mp4', 'flv', '', 'none')},
1623 'aext': {'type': 'ordered', 'field': 'audio_ext',
1624 'order': ('m4a', 'aac', 'mp3', 'ogg', 'opus', 'webm', '', 'none'),
1625 'order_free': ('opus', 'ogg', 'webm', 'm4a', 'mp3', 'aac', '', 'none')},
1626 'hidden': {'visible': False, 'forced': True, 'type': 'extractor', 'max': -1000},
1627 'aud_or_vid': {'visible': False, 'forced': True, 'type': 'multiple',
1628 'field': ('vcodec', 'acodec'),
1629 'function': lambda it: int(any(v != 'none' for v in it))},
1630 'ie_pref': {'priority': True, 'type': 'extractor'},
1631 'hasvid': {'priority': True, 'field': 'vcodec', 'type': 'boolean', 'not_in_list': ('none',)},
1632 'hasaud': {'field': 'acodec', 'type': 'boolean', 'not_in_list': ('none',)},
1633 'lang': {'convert': 'float', 'field': 'language_preference', 'default': -1},
1634 'quality': {'convert': 'float', 'default': -1},
1635 'filesize': {'convert': 'bytes'},
1636 'fs_approx': {'convert': 'bytes', 'field': 'filesize_approx'},
1637 'id': {'convert': 'string', 'field': 'format_id'},
1638 'height': {'convert': 'float_none'},
1639 'width': {'convert': 'float_none'},
1640 'fps': {'convert': 'float_none'},
1641 'tbr': {'convert': 'float_none'},
1642 'vbr': {'convert': 'float_none'},
1643 'abr': {'convert': 'float_none'},
1644 'asr': {'convert': 'float_none'},
1645 'source': {'convert': 'float', 'field': 'source_preference', 'default': -1},
1647 'codec': {'type': 'combined', 'field': ('vcodec', 'acodec')},
1648 'br': {'type': 'combined', 'field': ('tbr', 'vbr', 'abr'), 'same_limit': True},
1649 'size': {'type': 'combined', 'same_limit': True, 'field': ('filesize', 'fs_approx')},
1650 'ext': {'type': 'combined', 'field': ('vext', 'aext')},
1651 'res': {'type': 'multiple', 'field': ('height', 'width'),
1652 'function': lambda it: (lambda l: min(l) if l else 0)(tuple(filter(None, it)))},
1654 # For compatibility with youtube-dl
1655 'format_id': {'type': 'alias', 'field': 'id'},
1656 'preference': {'type': 'alias', 'field': 'ie_pref'},
1657 'language_preference': {'type': 'alias', 'field': 'lang'},
1660 'dimension': {'type': 'alias', 'field': 'res'},
1661 'resolution': {'type': 'alias', 'field': 'res'},
1662 'extension': {'type': 'alias', 'field': 'ext'},
1663 'bitrate': {'type': 'alias', 'field': 'br'},
1664 'total_bitrate': {'type': 'alias', 'field': 'tbr'},
1665 'video_bitrate': {'type': 'alias', 'field': 'vbr'},
1666 'audio_bitrate': {'type': 'alias', 'field': 'abr'},
1667 'framerate': {'type': 'alias', 'field': 'fps'},
1668 'protocol': {'type': 'alias', 'field': 'proto'},
1669 'source_preference': {'type': 'alias', 'field': 'source'},
1670 'filesize_approx': {'type': 'alias', 'field': 'fs_approx'},
1671 'filesize_estimate': {'type': 'alias', 'field': 'size'},
1672 'samplerate': {'type': 'alias', 'field': 'asr'},
1673 'video_ext': {'type': 'alias', 'field': 'vext'},
1674 'audio_ext': {'type': 'alias', 'field': 'aext'},
1675 'video_codec': {'type': 'alias', 'field': 'vcodec'},
1676 'audio_codec': {'type': 'alias', 'field': 'acodec'},
1677 'video': {'type': 'alias', 'field': 'hasvid'},
1678 'has_video': {'type': 'alias', 'field': 'hasvid'},
1679 'audio': {'type': 'alias', 'field': 'hasaud'},
1680 'has_audio': {'type': 'alias', 'field': 'hasaud'},
1681 'extractor': {'type': 'alias', 'field': 'ie_pref'},
1682 'extractor_preference': {'type': 'alias', 'field': 'ie_pref'},
1685 def __init__(self, ie, field_preference):
1687 self.ydl = ie._downloader
1688 self.evaluate_params(self.ydl.params, field_preference)
1689 if ie.get_param('verbose'):
1690 self.print_verbose_info(self.ydl.write_debug)
1692 def _get_field_setting(self, field, key):
1693 if field not in self.settings:
1694 if key in ('forced', 'priority'):
1696 self.ydl.deprecation_warning(
1697 f'Using arbitrary fields ({field}) for format sorting is deprecated '
1698 'and may be removed in a future version')
1699 self.settings[field] = {}
1700 propObj = self.settings[field]
1701 if key not in propObj:
1702 type = propObj.get('type')
1704 default = 'preference' if type == 'extractor' else (field,) if type in ('combined', 'multiple') else field
1705 elif key == 'convert':
1706 default = 'order' if type == 'ordered' else 'float_string' if field else 'ignore'
1708 default = {'type': 'field', 'visible': True, 'order': [], 'not_in_list': (None,)}.get(key, None)
1709 propObj[key] = default
1712 def _resolve_field_value(self, field, value, convertNone=False):
1717 value = value.lower()
1718 conversion = self._get_field_setting(field, 'convert')
1719 if conversion == 'ignore':
1721 if conversion == 'string':
1723 elif conversion == 'float_none':
1724 return float_or_none(value)
1725 elif conversion == 'bytes':
1726 return FileDownloader.parse_bytes(value)
1727 elif conversion == 'order':
1728 order_list = (self._use_free_order and self._get_field_setting(field, 'order_free')) or self._get_field_setting(field, 'order')
1729 use_regex = self._get_field_setting(field, 'regex')
1730 list_length = len(order_list)
1731 empty_pos = order_list.index('') if '' in order_list else list_length + 1
1732 if use_regex and value is not None:
1733 for i, regex in enumerate(order_list):
1734 if regex and re.match(regex, value):
1735 return list_length - i
1736 return list_length - empty_pos # not in list
1737 else: # not regex or value = None
1738 return list_length - (order_list.index(value) if value in order_list else empty_pos)
1740 if value.isnumeric():
1743 self.settings[field]['convert'] = 'string'
1746 def evaluate_params(self, params, sort_extractor):
1747 self._use_free_order = params.get('prefer_free_formats', False)
1748 self._sort_user = params.get('format_sort', [])
1749 self._sort_extractor = sort_extractor
1751 def add_item(field, reverse, closest, limit_text):
1752 field = field.lower()
1753 if field in self._order:
1755 self._order.append(field)
1756 limit = self._resolve_field_value(field, limit_text)
1759 'closest': False if limit is None else closest,
1760 'limit_text': limit_text,
1762 if field in self.settings:
1763 self.settings[field].update(data)
1765 self.settings[field] = data
1768 tuple(field for field in self.default if self._get_field_setting(field, 'forced'))
1769 + (tuple() if params.get('format_sort_force', False)
1770 else tuple(field for field in self.default if self._get_field_setting(field, 'priority')))
1771 + tuple(self._sort_user) + tuple(sort_extractor) + self.default)
1773 for item in sort_list:
1774 match = re.match(self.regex, item)
1776 raise ExtractorError('Invalid format sort string "%s" given by extractor' % item)
1777 field = match.group('field')
1780 if self._get_field_setting(field, 'type') == 'alias':
1781 alias, field = field, self._get_field_setting(field, 'field')
1782 if alias not in ('format_id', 'preference', 'language_preference'):
1783 self.ydl.deprecation_warning(
1784 f'Format sorting alias {alias} is deprecated '
1785 f'and may be removed in a future version. Please use {field} instead')
1786 reverse = match.group('reverse') is not None
1787 closest = match.group('separator') == '~'
1788 limit_text = match.group('limit')
1790 has_limit = limit_text is not None
1791 has_multiple_fields = self._get_field_setting(field, 'type') == 'combined'
1792 has_multiple_limits = has_limit and has_multiple_fields and not self._get_field_setting(field, 'same_limit')
1794 fields = self._get_field_setting(field, 'field') if has_multiple_fields else (field,)
1795 limits = limit_text.split(':') if has_multiple_limits else (limit_text,) if has_limit else tuple()
1796 limit_count = len(limits)
1797 for (i, f) in enumerate(fields):
1798 add_item(f, reverse, closest,
1799 limits[i] if i < limit_count
1800 else limits[0] if has_limit and not has_multiple_limits
1803 def print_verbose_info(self, write_debug):
1805 write_debug('Sort order given by user: %s' % ', '.join(self._sort_user))
1806 if self._sort_extractor:
1807 write_debug('Sort order given by extractor: %s' % ', '.join(self._sort_extractor))
1808 write_debug('Formats sorted by: %s' % ', '.join(['%s%s%s' % (
1809 '+' if self._get_field_setting(field, 'reverse') else '', field,
1810 '%s%s(%s)' % ('~' if self._get_field_setting(field, 'closest') else ':',
1811 self._get_field_setting(field, 'limit_text'),
1812 self._get_field_setting(field, 'limit'))
1813 if self._get_field_setting(field, 'limit_text') is not None else '')
1814 for field in self._order if self._get_field_setting(field, 'visible')]))
1816 def _calculate_field_preference_from_value(self, format, field, type, value):
1817 reverse = self._get_field_setting(field, 'reverse')
1818 closest = self._get_field_setting(field, 'closest')
1819 limit = self._get_field_setting(field, 'limit')
1821 if type == 'extractor':
1822 maximum = self._get_field_setting(field, 'max')
1823 if value is None or (maximum is not None and value >= maximum):
1825 elif type == 'boolean':
1826 in_list = self._get_field_setting(field, 'in_list')
1827 not_in_list = self._get_field_setting(field, 'not_in_list')
1828 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
1829 elif type == 'ordered':
1830 value = self._resolve_field_value(field, value, True)
1832 # try to convert to number
1833 val_num = float_or_none(value, default=self._get_field_setting(field, 'default'))
1834 is_num = self._get_field_setting(field, 'convert') != 'string' and val_num is not None
1838 return ((-10, 0) if value is None
1839 else (1, value, 0) if not is_num # if a field has mixed strings and numbers, strings are sorted higher
1840 else (0, -abs(value - limit), value - limit if reverse else limit - value) if closest
1841 else (0, value, 0) if not reverse and (limit is None or value <= limit)
1842 else (0, -value, 0) if limit is None or (reverse and value == limit) or value > limit
1843 else (-1, value, 0))
1845 def _calculate_field_preference(self, format, field):
1846 type = self._get_field_setting(field, 'type') # extractor, boolean, ordered, field, multiple
1847 get_value = lambda f: format.get(self._get_field_setting(f, 'field'))
1848 if type == 'multiple':
1849 type = 'field' # Only 'field' is allowed in multiple for now
1850 actual_fields = self._get_field_setting(field, 'field')
1852 value = self._get_field_setting(field, 'function')(get_value(f) for f in actual_fields)
1854 value = get_value(field)
1855 return self._calculate_field_preference_from_value(format, field, type, value)
1857 def calculate_preference(self, format):
1858 # Determine missing protocol
1859 if not format.get('protocol'):
1860 format['protocol'] = determine_protocol(format)
1862 # Determine missing ext
1863 if not format.get('ext') and 'url' in format:
1864 format['ext'] = determine_ext(format['url'])
1865 if format.get('vcodec') == 'none':
1866 format['audio_ext'] = format['ext'] if format.get('acodec') != 'none' else 'none'
1867 format['video_ext'] = 'none'
1869 format['video_ext'] = format['ext']
1870 format['audio_ext'] = 'none'
1871 # if format.get('preference') is None and format.get('ext') in ('f4f', 'f4m'): # Not supported?
1872 # format['preference'] = -1000
1874 # Determine missing bitrates
1875 if format.get('tbr') is None:
1876 if format.get('vbr') is not None and format.get('abr') is not None:
1877 format['tbr'] = format.get('vbr', 0) + format.get('abr', 0)
1879 if format.get('vcodec') != 'none' and format.get('vbr') is None:
1880 format['vbr'] = format.get('tbr') - format.get('abr', 0)
1881 if format.get('acodec') != 'none' and format.get('abr') is None:
1882 format['abr'] = format.get('tbr') - format.get('vbr', 0)
1884 return tuple(self._calculate_field_preference(format, field) for field in self._order)
1886 def _sort_formats(self, formats, field_preference=[]):
1889 format_sort = self.FormatSort(self, field_preference)
1890 formats.sort(key=lambda f: format_sort.calculate_preference(f))
1892 def _check_formats(self, formats, video_id):
1894 formats[:] = filter(
1895 lambda f: self._is_valid_url(
1897 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1901 def _remove_duplicate_formats(formats):
1905 if f['url'] not in format_urls:
1906 format_urls.add(f['url'])
1907 unique_formats.append(f)
1908 formats[:] = unique_formats
1910 def _is_valid_url(self, url, video_id, item='video', headers={}):
1911 url = self._proto_relative_url(url, scheme='http:')
1912 # For now assume non HTTP(S) URLs always valid
1913 if not (url.startswith('http://') or url.startswith('https://')):
1916 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1918 except ExtractorError as e:
1920 '%s: %s URL is invalid, skipping: %s'
1921 % (video_id, item, error_to_compat_str(e.cause)))
1924 def http_scheme(self):
1925 """ Either "http
:" or "https
:", depending on the user's preferences """
1928 if self.get_param('prefer_insecure', False)
1931 def _proto_relative_url(self, url, scheme=None):
1934 if url.startswith('//'):
1936 scheme = self.http_scheme()
1941 def _sleep(self, timeout, video_id, msg_template=None):
1942 if msg_template is None:
1943 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1944 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1948 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1949 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1950 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
1951 manifest = self._download_xml(
1952 manifest_url, video_id, 'Downloading f4m manifest',
1953 'Unable to download f4m manifest',
1954 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1955 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
1956 transform_source=transform_source,
1957 fatal=fatal, data=data, headers=headers, query=query)
1959 if manifest is False:
1962 return self._parse_f4m_formats(
1963 manifest, manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1964 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1966 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1967 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1968 fatal=True, m3u8_id=None):
1969 if not isinstance(manifest, compat_etree_Element) and not fatal:
1972 # currently yt-dlp cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1973 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1974 if akamai_pv is not None and ';' in akamai_pv.text:
1975 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1976 if playerVerificationChallenge.strip() != '':
1980 manifest_version = '1.0'
1981 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1983 manifest_version = '2.0'
1984 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1985 # Remove unsupported DRM protected media from final formats
1986 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
1987 media_nodes = remove_encrypted_media(media_nodes)
1991 manifest_base_url = get_base_url(manifest)
1993 bootstrap_info = xpath_element(
1994 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1995 'bootstrap info', default=None)
1998 mime_type = xpath_text(
1999 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
2000 'base URL', default=None)
2001 if mime_type and mime_type.startswith('audio/'):
2004 for i, media_el in enumerate(media_nodes):
2005 tbr = int_or_none(media_el.attrib.get('bitrate'))
2006 width = int_or_none(media_el.attrib.get('width'))
2007 height = int_or_none(media_el.attrib.get('height'))
2008 format_id = join_nonempty(f4m_id, tbr or i)
2009 # If <bootstrapInfo> is present, the specified f4m is a
2010 # stream-level manifest, and only set-level manifests may refer to
2011 # external resources. See section 11.4 and section 4 of F4M spec
2012 if bootstrap_info is None:
2014 # @href is introduced in 2.0, see section 11.6 of F4M spec
2015 if manifest_version == '2.0':
2016 media_url = media_el.attrib.get('href')
2017 if media_url is None:
2018 media_url = media_el.attrib.get('url')
2022 media_url if media_url.startswith('http://') or media_url.startswith('https://')
2023 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
2024 # If media_url is itself a f4m manifest do the recursive extraction
2025 # since bitrates in parent manifest (this one) and media_url manifest
2026 # may differ leading to inability to resolve the format by requested
2027 # bitrate in f4m downloader
2028 ext = determine_ext(manifest_url)
2030 f4m_formats = self._extract_f4m_formats(
2031 manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
2032 transform_source=transform_source, fatal=fatal)
2033 # Sometimes stream-level manifest contains single media entry that
2034 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
2035 # At the same time parent's media entry in set-level manifest may
2036 # contain it. We will copy it from parent in such cases.
2037 if len(f4m_formats) == 1:
2040 'tbr': f.get('tbr') or tbr,
2041 'width': f.get('width') or width,
2042 'height': f.get('height') or height,
2043 'format_id': f.get('format_id') if not tbr else format_id,
2046 formats.extend(f4m_formats)
2049 formats.extend(self._extract_m3u8_formats(
2050 manifest_url, video_id, 'mp4', preference=preference,
2051 quality=quality, m3u8_id=m3u8_id, fatal=fatal))
2054 'format_id': format_id,
2055 'url': manifest_url,
2056 'manifest_url': manifest_url,
2057 'ext': 'flv' if bootstrap_info is not None else None,
2063 'preference': preference,
2068 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, quality=None, m3u8_id=None):
2070 'format_id': join_nonempty(m3u8_id, 'meta'),
2074 'preference': preference - 100 if preference else -100,
2076 'resolution': 'multiple',
2077 'format_note': 'Quality selection URL',
2080 def _report_ignoring_subs(self, name):
2081 self.report_warning(bug_reports_message(
2082 f'Ignoring subtitle tracks found in the {name} manifest; '
2083 'if any subtitle tracks are missing,'
2086 def _extract_m3u8_formats(self, *args, **kwargs):
2087 fmts, subs = self._extract_m3u8_formats_and_subtitles(*args, **kwargs)
2089 self._report_ignoring_subs('HLS')
2092 def _extract_m3u8_formats_and_subtitles(
2093 self, m3u8_url, video_id, ext=None, entry_protocol='m3u8_native',
2094 preference=None, quality=None, m3u8_id=None, note=None,
2095 errnote=None, fatal=True, live=False, data=None, headers={},
2098 res = self._download_webpage_handle(
2100 note='Downloading m3u8 information' if note is None else note,
2101 errnote='Failed to download m3u8 information' if errnote is None else errnote,
2102 fatal=fatal, data=data, headers=headers, query=query)
2107 m3u8_doc, urlh = res
2108 m3u8_url = urlh.geturl()
2110 return self._parse_m3u8_formats_and_subtitles(
2111 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
2112 preference=preference, quality=quality, m3u8_id=m3u8_id,
2113 note=note, errnote=errnote, fatal=fatal, live=live, data=data,
2114 headers=headers, query=query, video_id=video_id)
2116 def _parse_m3u8_formats_and_subtitles(
2117 self, m3u8_doc, m3u8_url=None, ext=None, entry_protocol='m3u8_native',
2118 preference=None, quality=None, m3u8_id=None, live=False, note=None,
2119 errnote=None, fatal=True, data=None, headers={}, query={},
2121 formats, subtitles = [], {}
2123 has_drm = re.search('|'.join([
2124 r'#EXT-X-FAXS-CM:', # Adobe Flash Access
2125 r'#EXT-X-(?:SESSION-)?KEY:.*?URI="skd
://', # Apple FairPlay
2128 def format_url(url):
2129 return url if re.match(r'^https?
://', url) else compat_urlparse.urljoin(m3u8_url, url)
2131 if self.get_param('hls_split_discontinuity
', False):
2132 def _extract_m3u8_playlist_indices(manifest_url=None, m3u8_doc=None):
2134 if not manifest_url:
2136 m3u8_doc = self._download_webpage(
2137 manifest_url, video_id, fatal=fatal, data=data, headers=headers,
2138 note=False, errnote='Failed to download m3u8 playlist information
')
2139 if m3u8_doc is False:
2141 return range(1 + sum(line.startswith('#EXT-X-DISCONTINUITY') for line in m3u8_doc.splitlines()))
2144 def _extract_m3u8_playlist_indices(*args
, **kwargs
):
2148 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
2149 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
2150 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
2152 # We should try extracting formats only from master playlists [1, 4.3.4],
2153 # i.e. playlists that describe available qualities. On the other hand
2154 # media playlists [1, 4.3.3] should be returned as is since they contain
2155 # just the media without qualities renditions.
2156 # Fortunately, master playlist can be easily distinguished from media
2157 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
2158 # master playlist tags MUST NOT appear in a media playlist and vice versa.
2159 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
2160 # media playlist and MUST NOT appear in master playlist thus we can
2161 # clearly detect media playlist with this criterion.
2163 if '#EXT-X-TARGETDURATION' in m3u8_doc
: # media playlist, return as is
2165 'format_id': join_nonempty(m3u8_id
, idx
),
2166 'format_index': idx
,
2167 'url': m3u8_url
or encode_data_uri(m3u8_doc
.encode('utf-8'), 'application/x-mpegurl'),
2169 'protocol': entry_protocol
,
2170 'preference': preference
,
2173 } for idx
in _extract_m3u8_playlist_indices(m3u8_doc
=m3u8_doc
)]
2175 return formats
, subtitles
2178 last_stream_inf
= {}
2180 def extract_media(x_media_line
):
2181 media
= parse_m3u8_attributes(x_media_line
)
2182 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
2183 media_type
, group_id
, name
= media
.get('TYPE'), media
.get('GROUP-ID'), media
.get('NAME')
2184 if not (media_type
and group_id
and name
):
2186 groups
.setdefault(group_id
, []).append(media
)
2187 # <https://tools.ietf.org/html/rfc8216#section-4.3.4.1>
2188 if media_type
== 'SUBTITLES':
2189 # According to RFC 8216 §4.3.4.2.1, URI is REQUIRED in the
2190 # EXT-X-MEDIA tag if the media type is SUBTITLES.
2191 # However, lack of URI has been spotted in the wild.
2192 # e.g. NebulaIE; see https://github.com/yt-dlp/yt-dlp/issues/339
2193 if not media
.get('URI'):
2195 url
= format_url(media
['URI'])
2198 'ext': determine_ext(url
),
2200 if sub_info
['ext'] == 'm3u8':
2201 # Per RFC 8216 §3.1, the only possible subtitle format m3u8
2202 # files may contain is WebVTT:
2203 # <https://tools.ietf.org/html/rfc8216#section-3.1>
2204 sub_info
['ext'] = 'vtt'
2205 sub_info
['protocol'] = 'm3u8_native'
2206 lang
= media
.get('LANGUAGE') or 'und'
2207 subtitles
.setdefault(lang
, []).append(sub_info
)
2208 if media_type
not in ('VIDEO', 'AUDIO'):
2210 media_url
= media
.get('URI')
2212 manifest_url
= format_url(media_url
)
2214 'format_id': join_nonempty(m3u8_id
, group_id
, name
, idx
),
2215 'format_note': name
,
2216 'format_index': idx
,
2217 'url': manifest_url
,
2218 'manifest_url': m3u8_url
,
2219 'language': media
.get('LANGUAGE'),
2221 'protocol': entry_protocol
,
2222 'preference': preference
,
2224 'vcodec': 'none' if media_type
== 'AUDIO' else None,
2225 } for idx
in _extract_m3u8_playlist_indices(manifest_url
))
2227 def build_stream_name():
2228 # Despite specification does not mention NAME attribute for
2229 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
2230 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
2231 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
2232 stream_name
= last_stream_inf
.get('NAME')
2235 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
2236 # from corresponding rendition group
2237 stream_group_id
= last_stream_inf
.get('VIDEO')
2238 if not stream_group_id
:
2240 stream_group
= groups
.get(stream_group_id
)
2241 if not stream_group
:
2242 return stream_group_id
2243 rendition
= stream_group
[0]
2244 return rendition
.get('NAME') or stream_group_id
2246 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
2247 # chance to detect video only formats when EXT-X-STREAM-INF tags
2248 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
2249 for line
in m3u8_doc
.splitlines():
2250 if line
.startswith('#EXT-X-MEDIA:'):
2253 for line
in m3u8_doc
.splitlines():
2254 if line
.startswith('#EXT-X-STREAM-INF:'):
2255 last_stream_inf
= parse_m3u8_attributes(line
)
2256 elif line
.startswith('#') or not line
.strip():
2259 tbr
= float_or_none(
2260 last_stream_inf
.get('AVERAGE-BANDWIDTH')
2261 or last_stream_inf
.get('BANDWIDTH'), scale
=1000)
2262 manifest_url
= format_url(line
.strip())
2264 for idx
in _extract_m3u8_playlist_indices(manifest_url
):
2265 format_id
= [m3u8_id
, None, idx
]
2266 # Bandwidth of live streams may differ over time thus making
2267 # format_id unpredictable. So it's better to keep provided
2270 stream_name
= build_stream_name()
2271 format_id
[1] = stream_name
or '%d' % (tbr
or len(formats
))
2273 'format_id': join_nonempty(*format_id
),
2274 'format_index': idx
,
2275 'url': manifest_url
,
2276 'manifest_url': m3u8_url
,
2279 'fps': float_or_none(last_stream_inf
.get('FRAME-RATE')),
2280 'protocol': entry_protocol
,
2281 'preference': preference
,
2284 resolution
= last_stream_inf
.get('RESOLUTION')
2286 mobj
= re
.search(r
'(?P<width>\d+)[xX](?P<height>\d+)', resolution
)
2288 f
['width'] = int(mobj
.group('width'))
2289 f
['height'] = int(mobj
.group('height'))
2290 # Unified Streaming Platform
2292 r
'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f
['url'])
2294 abr
, vbr
= mobj
.groups()
2295 abr
, vbr
= float_or_none(abr
, 1000), float_or_none(vbr
, 1000)
2300 codecs
= parse_codecs(last_stream_inf
.get('CODECS'))
2302 audio_group_id
= last_stream_inf
.get('AUDIO')
2303 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
2304 # references a rendition group MUST have a CODECS attribute.
2305 # However, this is not always respected, for example, [2]
2306 # contains EXT-X-STREAM-INF tag which references AUDIO
2307 # rendition group but does not have CODECS and despite
2308 # referencing an audio group it represents a complete
2309 # (with audio and video) format. So, for such cases we will
2310 # ignore references to rendition groups and treat them
2311 # as complete formats.
2312 if audio_group_id
and codecs
and f
.get('vcodec') != 'none':
2313 audio_group
= groups
.get(audio_group_id
)
2314 if audio_group
and audio_group
[0].get('URI'):
2315 # TODO: update acodec for audio only formats with
2317 f
['acodec'] = 'none'
2318 if not f
.get('ext'):
2319 f
['ext'] = 'm4a' if f
.get('vcodec') == 'none' else 'mp4'
2323 progressive_uri
= last_stream_inf
.get('PROGRESSIVE-URI')
2326 del http_f
['manifest_url']
2328 'format_id': f
['format_id'].replace('hls-', 'http-'),
2330 'url': progressive_uri
,
2332 formats
.append(http_f
)
2334 last_stream_inf
= {}
2335 return formats
, subtitles
2337 def _extract_m3u8_vod_duration(
2338 self
, m3u8_vod_url
, video_id
, note
=None, errnote
=None, data
=None, headers
={}, query={}
):
2340 m3u8_vod
= self
._download
_webpage
(
2341 m3u8_vod_url
, video_id
,
2342 note
='Downloading m3u8 VOD manifest' if note
is None else note
,
2343 errnote
='Failed to download VOD manifest' if errnote
is None else errnote
,
2344 fatal
=False, data
=data
, headers
=headers
, query
=query
)
2346 return self
._parse
_m
3u8_vod
_duration
(m3u8_vod
or '', video_id
)
2348 def _parse_m3u8_vod_duration(self
, m3u8_vod
, video_id
):
2349 if '#EXT-X-PLAYLIST-TYPE:VOD' not in m3u8_vod
:
2353 float(line
[len('#EXTINF:'):].split(',')[0])
2354 for line
in m3u8_vod
.splitlines() if line
.startswith('#EXTINF:'))) or None
2357 def _xpath_ns(path
, namespace
=None):
2361 for c
in path
.split('/'):
2362 if not c
or c
== '.':
2365 out
.append('{%s}%s' % (namespace
, c
))
2366 return '/'.join(out
)
2368 def _extract_smil_formats_and_subtitles(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None, transform_source
=None):
2369 smil
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
, transform_source
=transform_source
)
2375 namespace
= self
._parse
_smil
_namespace
(smil
)
2377 fmts
= self
._parse
_smil
_formats
(
2378 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2379 subs
= self
._parse
_smil
_subtitles
(
2380 smil
, namespace
=namespace
)
2384 def _extract_smil_formats(self
, *args
, **kwargs
):
2385 fmts
, subs
= self
._extract
_smil
_formats
_and
_subtitles
(*args
, **kwargs
)
2387 self
._report
_ignoring
_subs
('SMIL')
2390 def _extract_smil_info(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None):
2391 smil
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
)
2394 return self
._parse
_smil
(smil
, smil_url
, video_id
, f4m_params
=f4m_params
)
2396 def _download_smil(self
, smil_url
, video_id
, fatal
=True, transform_source
=None):
2397 return self
._download
_xml
(
2398 smil_url
, video_id
, 'Downloading SMIL file',
2399 'Unable to download SMIL file', fatal
=fatal
, transform_source
=transform_source
)
2401 def _parse_smil(self
, smil
, smil_url
, video_id
, f4m_params
=None):
2402 namespace
= self
._parse
_smil
_namespace
(smil
)
2404 formats
= self
._parse
_smil
_formats
(
2405 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2406 subtitles
= self
._parse
_smil
_subtitles
(smil
, namespace
=namespace
)
2408 video_id
= os
.path
.splitext(url_basename(smil_url
))[0]
2412 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2413 name
= meta
.attrib
.get('name')
2414 content
= meta
.attrib
.get('content')
2415 if not name
or not content
:
2417 if not title
and name
== 'title':
2419 elif not description
and name
in ('description', 'abstract'):
2420 description
= content
2421 elif not upload_date
and name
== 'date':
2422 upload_date
= unified_strdate(content
)
2425 'id': image
.get('type'),
2426 'url': image
.get('src'),
2427 'width': int_or_none(image
.get('width')),
2428 'height': int_or_none(image
.get('height')),
2429 } for image
in smil
.findall(self
._xpath
_ns
('.//image', namespace
)) if image
.get('src')]
2433 'title': title
or video_id
,
2434 'description': description
,
2435 'upload_date': upload_date
,
2436 'thumbnails': thumbnails
,
2438 'subtitles': subtitles
,
2441 def _parse_smil_namespace(self
, smil
):
2442 return self
._search
_regex
(
2443 r
'(?i)^{([^}]+)?}smil$', smil
.tag
, 'namespace', default
=None)
2445 def _parse_smil_formats(self
, smil
, smil_url
, video_id
, namespace
=None, f4m_params
=None, transform_rtmp_url
=None):
2447 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2448 b
= meta
.get('base') or meta
.get('httpBase')
2460 media
= smil
.findall(self
._xpath
_ns
('.//video', namespace
)) + smil
.findall(self
._xpath
_ns
('.//audio', namespace
))
2461 for medium
in media
:
2462 src
= medium
.get('src')
2463 if not src
or src
in srcs
:
2467 bitrate
= float_or_none(medium
.get('system-bitrate') or medium
.get('systemBitrate'), 1000)
2468 filesize
= int_or_none(medium
.get('size') or medium
.get('fileSize'))
2469 width
= int_or_none(medium
.get('width'))
2470 height
= int_or_none(medium
.get('height'))
2471 proto
= medium
.get('proto')
2472 ext
= medium
.get('ext')
2473 src_ext
= determine_ext(src
)
2474 streamer
= medium
.get('streamer') or base
2476 if proto
== 'rtmp' or streamer
.startswith('rtmp'):
2482 'format_id': 'rtmp-%d' % (rtmp_count
if bitrate
is None else bitrate
),
2484 'filesize': filesize
,
2488 if transform_rtmp_url
:
2489 streamer
, src
= transform_rtmp_url(streamer
, src
)
2490 formats
[-1].update({
2496 src_url
= src
if src
.startswith('http') else compat_urlparse
.urljoin(base
, src
)
2497 src_url
= src_url
.strip()
2499 if proto
== 'm3u8' or src_ext
== 'm3u8':
2500 m3u8_formats
= self
._extract
_m
3u8_formats
(
2501 src_url
, video_id
, ext
or 'mp4', m3u8_id
='hls', fatal
=False)
2502 if len(m3u8_formats
) == 1:
2504 m3u8_formats
[0].update({
2505 'format_id': 'hls-%d' % (m3u8_count
if bitrate
is None else bitrate
),
2510 formats
.extend(m3u8_formats
)
2511 elif src_ext
== 'f4m':
2516 'plugin': 'flowplayer-3.2.0.1',
2518 f4m_url
+= '&' if '?' in f4m_url
else '?'
2519 f4m_url
+= compat_urllib_parse_urlencode(f4m_params
)
2520 formats
.extend(self
._extract
_f
4m
_formats
(f4m_url
, video_id
, f4m_id
='hds', fatal
=False))
2521 elif src_ext
== 'mpd':
2522 formats
.extend(self
._extract
_mpd
_formats
(
2523 src_url
, video_id
, mpd_id
='dash', fatal
=False))
2524 elif re
.search(r
'\.ism/[Mm]anifest', src_url
):
2525 formats
.extend(self
._extract
_ism
_formats
(
2526 src_url
, video_id
, ism_id
='mss', fatal
=False))
2527 elif src_url
.startswith('http') and self
._is
_valid
_url
(src
, video_id
):
2531 'ext': ext
or src_ext
or 'flv',
2532 'format_id': 'http-%d' % (bitrate
or http_count
),
2534 'filesize': filesize
,
2539 for medium
in smil
.findall(self
._xpath
_ns
('.//imagestream', namespace
)):
2540 src
= medium
.get('src')
2541 if not src
or src
in srcs
:
2547 'format_id': 'imagestream-%d' % (imgs_count
),
2549 'ext': mimetype2ext(medium
.get('type')),
2552 'width': int_or_none(medium
.get('width')),
2553 'height': int_or_none(medium
.get('height')),
2554 'format_note': 'SMIL storyboards',
2559 def _parse_smil_subtitles(self
, smil
, namespace
=None, subtitles_lang
='en'):
2562 for num
, textstream
in enumerate(smil
.findall(self
._xpath
_ns
('.//textstream', namespace
))):
2563 src
= textstream
.get('src')
2564 if not src
or src
in urls
:
2567 ext
= textstream
.get('ext') or mimetype2ext(textstream
.get('type')) or determine_ext(src
)
2568 lang
= textstream
.get('systemLanguage') or textstream
.get('systemLanguageName') or textstream
.get('lang') or subtitles_lang
2569 subtitles
.setdefault(lang
, []).append({
2575 def _extract_xspf_playlist(self
, xspf_url
, playlist_id
, fatal
=True):
2576 xspf
= self
._download
_xml
(
2577 xspf_url
, playlist_id
, 'Downloading xpsf playlist',
2578 'Unable to download xspf manifest', fatal
=fatal
)
2581 return self
._parse
_xspf
(
2582 xspf
, playlist_id
, xspf_url
=xspf_url
,
2583 xspf_base_url
=base_url(xspf_url
))
2585 def _parse_xspf(self
, xspf_doc
, playlist_id
, xspf_url
=None, xspf_base_url
=None):
2587 'xspf': 'http://xspf.org/ns/0/',
2588 's1': 'http://static.streamone.nl/player/ns/0',
2592 for track
in xspf_doc
.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP
)):
2594 track
, xpath_with_ns('./xspf:title', NS_MAP
), 'title', default
=playlist_id
)
2595 description
= xpath_text(
2596 track
, xpath_with_ns('./xspf:annotation', NS_MAP
), 'description')
2597 thumbnail
= xpath_text(
2598 track
, xpath_with_ns('./xspf:image', NS_MAP
), 'thumbnail')
2599 duration
= float_or_none(
2600 xpath_text(track
, xpath_with_ns('./xspf:duration', NS_MAP
), 'duration'), 1000)
2603 for location
in track
.findall(xpath_with_ns('./xspf:location', NS_MAP
)):
2604 format_url
= urljoin(xspf_base_url
, location
.text
)
2609 'manifest_url': xspf_url
,
2610 'format_id': location
.get(xpath_with_ns('s1:label', NS_MAP
)),
2611 'width': int_or_none(location
.get(xpath_with_ns('s1:width', NS_MAP
))),
2612 'height': int_or_none(location
.get(xpath_with_ns('s1:height', NS_MAP
))),
2614 self
._sort
_formats
(formats
)
2619 'description': description
,
2620 'thumbnail': thumbnail
,
2621 'duration': duration
,
2626 def _extract_mpd_formats(self
, *args
, **kwargs
):
2627 fmts
, subs
= self
._extract
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2629 self
._report
_ignoring
_subs
('DASH')
2632 def _extract_mpd_formats_and_subtitles(
2633 self
, mpd_url
, video_id
, mpd_id
=None, note
=None, errnote
=None,
2634 fatal
=True, data
=None, headers
={}, query={}
):
2635 res
= self
._download
_xml
_handle
(
2637 note
='Downloading MPD manifest' if note
is None else note
,
2638 errnote
='Failed to download MPD manifest' if errnote
is None else errnote
,
2639 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
2645 mpd_base_url
= base_url(urlh
.geturl())
2647 return self
._parse
_mpd
_formats
_and
_subtitles
(
2648 mpd_doc
, mpd_id
, mpd_base_url
, mpd_url
)
2650 def _parse_mpd_formats(self
, *args
, **kwargs
):
2651 fmts
, subs
= self
._parse
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2653 self
._report
_ignoring
_subs
('DASH')
2656 def _parse_mpd_formats_and_subtitles(
2657 self
, mpd_doc
, mpd_id
=None, mpd_base_url
='', mpd_url
=None):
2659 Parse formats from MPD manifest.
2661 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
2662 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
2663 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
2665 if not self
.get_param('dynamic_mpd', True):
2666 if mpd_doc
.get('type') == 'dynamic':
2669 namespace
= self
._search
_regex
(r
'(?i)^{([^}]+)?}MPD$', mpd_doc
.tag
, 'namespace', default
=None)
2672 return self
._xpath
_ns
(path
, namespace
)
2674 def is_drm_protected(element
):
2675 return element
.find(_add_ns('ContentProtection')) is not None
2677 def extract_multisegment_info(element
, ms_parent_info
):
2678 ms_info
= ms_parent_info
.copy()
2680 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2681 # common attributes and elements. We will only extract relevant
2683 def extract_common(source
):
2684 segment_timeline
= source
.find(_add_ns('SegmentTimeline'))
2685 if segment_timeline
is not None:
2686 s_e
= segment_timeline
.findall(_add_ns('S'))
2688 ms_info
['total_number'] = 0
2691 r
= int(s
.get('r', 0))
2692 ms_info
['total_number'] += 1 + r
2693 ms_info
['s'].append({
2694 't': int(s
.get('t', 0)),
2695 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2696 'd': int(s
.attrib
['d']),
2699 start_number
= source
.get('startNumber')
2701 ms_info
['start_number'] = int(start_number
)
2702 timescale
= source
.get('timescale')
2704 ms_info
['timescale'] = int(timescale
)
2705 segment_duration
= source
.get('duration')
2706 if segment_duration
:
2707 ms_info
['segment_duration'] = float(segment_duration
)
2709 def extract_Initialization(source
):
2710 initialization
= source
.find(_add_ns('Initialization'))
2711 if initialization
is not None:
2712 ms_info
['initialization_url'] = initialization
.attrib
['sourceURL']
2714 segment_list
= element
.find(_add_ns('SegmentList'))
2715 if segment_list
is not None:
2716 extract_common(segment_list
)
2717 extract_Initialization(segment_list
)
2718 segment_urls_e
= segment_list
.findall(_add_ns('SegmentURL'))
2720 ms_info
['segment_urls'] = [segment
.attrib
['media'] for segment
in segment_urls_e
]
2722 segment_template
= element
.find(_add_ns('SegmentTemplate'))
2723 if segment_template
is not None:
2724 extract_common(segment_template
)
2725 media
= segment_template
.get('media')
2727 ms_info
['media'] = media
2728 initialization
= segment_template
.get('initialization')
2730 ms_info
['initialization'] = initialization
2732 extract_Initialization(segment_template
)
2735 mpd_duration
= parse_duration(mpd_doc
.get('mediaPresentationDuration'))
2736 formats
, subtitles
= [], {}
2737 stream_numbers
= collections
.defaultdict(int)
2738 for period
in mpd_doc
.findall(_add_ns('Period')):
2739 period_duration
= parse_duration(period
.get('duration')) or mpd_duration
2740 period_ms_info
= extract_multisegment_info(period
, {
2744 for adaptation_set
in period
.findall(_add_ns('AdaptationSet')):
2745 adaption_set_ms_info
= extract_multisegment_info(adaptation_set
, period_ms_info
)
2746 for representation
in adaptation_set
.findall(_add_ns('Representation')):
2747 representation_attrib
= adaptation_set
.attrib
.copy()
2748 representation_attrib
.update(representation
.attrib
)
2749 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2750 mime_type
= representation_attrib
['mimeType']
2751 content_type
= representation_attrib
.get('contentType', mime_type
.split('/')[0])
2753 codecs
= parse_codecs(representation_attrib
.get('codecs', ''))
2754 if content_type
not in ('video', 'audio', 'text'):
2755 if mime_type
== 'image/jpeg':
2756 content_type
= mime_type
2757 elif codecs
['vcodec'] != 'none':
2758 content_type
= 'video'
2759 elif codecs
['acodec'] != 'none':
2760 content_type
= 'audio'
2761 elif codecs
.get('tcodec', 'none') != 'none':
2762 content_type
= 'text'
2763 elif mimetype2ext(mime_type
) in ('tt', 'dfxp', 'ttml', 'xml', 'json'):
2764 content_type
= 'text'
2766 self
.report_warning('Unknown MIME type %s in DASH manifest' % mime_type
)
2770 for element
in (representation
, adaptation_set
, period
, mpd_doc
):
2771 base_url_e
= element
.find(_add_ns('BaseURL'))
2772 if base_url_e
is not None:
2773 base_url
= base_url_e
.text
+ base_url
2774 if re
.match(r
'^https?://', base_url
):
2776 if mpd_base_url
and base_url
.startswith('/'):
2777 base_url
= compat_urlparse
.urljoin(mpd_base_url
, base_url
)
2778 elif mpd_base_url
and not re
.match(r
'^https?://', base_url
):
2779 if not mpd_base_url
.endswith('/'):
2781 base_url
= mpd_base_url
+ base_url
2782 representation_id
= representation_attrib
.get('id')
2783 lang
= representation_attrib
.get('lang')
2784 url_el
= representation
.find(_add_ns('BaseURL'))
2785 filesize
= int_or_none(url_el
.attrib
.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el
is not None else None)
2786 bandwidth
= int_or_none(representation_attrib
.get('bandwidth'))
2787 if representation_id
is not None:
2788 format_id
= representation_id
2790 format_id
= content_type
2792 format_id
= mpd_id
+ '-' + format_id
2793 if content_type
in ('video', 'audio'):
2795 'format_id': format_id
,
2796 'manifest_url': mpd_url
,
2797 'ext': mimetype2ext(mime_type
),
2798 'width': int_or_none(representation_attrib
.get('width')),
2799 'height': int_or_none(representation_attrib
.get('height')),
2800 'tbr': float_or_none(bandwidth
, 1000),
2801 'asr': int_or_none(representation_attrib
.get('audioSamplingRate')),
2802 'fps': int_or_none(representation_attrib
.get('frameRate')),
2803 'language': lang
if lang
not in ('mul', 'und', 'zxx', 'mis') else None,
2804 'format_note': 'DASH %s' % content_type
,
2805 'filesize': filesize
,
2806 'container': mimetype2ext(mime_type
) + '_dash',
2809 elif content_type
== 'text':
2811 'ext': mimetype2ext(mime_type
),
2812 'manifest_url': mpd_url
,
2813 'filesize': filesize
,
2815 elif content_type
== 'image/jpeg':
2816 # See test case in VikiIE
2817 # https://www.viki.com/videos/1175236v-choosing-spouse-by-lottery-episode-1
2819 'format_id': format_id
,
2821 'manifest_url': mpd_url
,
2822 'format_note': 'DASH storyboards (jpeg)',
2826 if is_drm_protected(adaptation_set
) or is_drm_protected(representation
):
2828 representation_ms_info
= extract_multisegment_info(representation
, adaption_set_ms_info
)
2830 def prepare_template(template_name
, identifiers
):
2831 tmpl
= representation_ms_info
[template_name
]
2832 # First of, % characters outside $...$ templates
2833 # must be escaped by doubling for proper processing
2834 # by % operator string formatting used further (see
2835 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2841 in_template
= not in_template
2842 elif c
== '%' and not in_template
:
2844 # Next, $...$ templates are translated to their
2845 # %(...) counterparts to be used with % operator
2846 if representation_id
is not None:
2847 t
= t
.replace('$RepresentationID$', representation_id
)
2848 t
= re
.sub(r
'\$(%s)\$' % '|'.join(identifiers
), r
'%(\1)d', t
)
2849 t
= re
.sub(r
'\$(%s)%%([^$]+)\$' % '|'.join(identifiers
), r
'%(\1)\2', t
)
2850 t
.replace('$$', '$')
2853 # @initialization is a regular template like @media one
2854 # so it should be handled just the same way (see
2855 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2856 if 'initialization' in representation_ms_info
:
2857 initialization_template
= prepare_template(
2859 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2860 # $Time$ shall not be included for @initialization thus
2861 # only $Bandwidth$ remains
2863 representation_ms_info
['initialization_url'] = initialization_template
% {
2864 'Bandwidth': bandwidth
,
2867 def location_key(location
):
2868 return 'url' if re
.match(r
'^https?://', location
) else 'path'
2870 if 'segment_urls' not in representation_ms_info
and 'media' in representation_ms_info
:
2872 media_template
= prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2873 media_location_key
= location_key(media_template
)
2875 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2876 # can't be used at the same time
2877 if '%(Number' in media_template
and 's' not in representation_ms_info
:
2878 segment_duration
= None
2879 if 'total_number' not in representation_ms_info
and 'segment_duration' in representation_ms_info
:
2880 segment_duration
= float_or_none(representation_ms_info
['segment_duration'], representation_ms_info
['timescale'])
2881 representation_ms_info
['total_number'] = int(math
.ceil(float(period_duration
) / segment_duration
))
2882 representation_ms_info
['fragments'] = [{
2883 media_location_key
: media_template
% {
2884 'Number': segment_number
,
2885 'Bandwidth': bandwidth
,
2887 'duration': segment_duration
,
2888 } for segment_number
in range(
2889 representation_ms_info
['start_number'],
2890 representation_ms_info
['total_number'] + representation_ms_info
['start_number'])]
2892 # $Number*$ or $Time$ in media template with S list available
2893 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2894 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2895 representation_ms_info
['fragments'] = []
2898 segment_number
= representation_ms_info
['start_number']
2900 def add_segment_url():
2901 segment_url
= media_template
% {
2902 'Time': segment_time
,
2903 'Bandwidth': bandwidth
,
2904 'Number': segment_number
,
2906 representation_ms_info
['fragments'].append({
2907 media_location_key
: segment_url
,
2908 'duration': float_or_none(segment_d
, representation_ms_info
['timescale']),
2911 for num
, s
in enumerate(representation_ms_info
['s']):
2912 segment_time
= s
.get('t') or segment_time
2916 for r
in range(s
.get('r', 0)):
2917 segment_time
+= segment_d
2920 segment_time
+= segment_d
2921 elif 'segment_urls' in representation_ms_info
and 's' in representation_ms_info
:
2923 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2924 # or any YouTube dashsegments video
2927 timescale
= representation_ms_info
['timescale']
2928 for s
in representation_ms_info
['s']:
2929 duration
= float_or_none(s
['d'], timescale
)
2930 for r
in range(s
.get('r', 0) + 1):
2931 segment_uri
= representation_ms_info
['segment_urls'][segment_index
]
2933 location_key(segment_uri
): segment_uri
,
2934 'duration': duration
,
2937 representation_ms_info
['fragments'] = fragments
2938 elif 'segment_urls' in representation_ms_info
:
2939 # Segment URLs with no SegmentTimeline
2940 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2941 # https://github.com/ytdl-org/youtube-dl/pull/14844
2943 segment_duration
= float_or_none(
2944 representation_ms_info
['segment_duration'],
2945 representation_ms_info
['timescale']) if 'segment_duration' in representation_ms_info
else None
2946 for segment_url
in representation_ms_info
['segment_urls']:
2948 location_key(segment_url
): segment_url
,
2950 if segment_duration
:
2951 fragment
['duration'] = segment_duration
2952 fragments
.append(fragment
)
2953 representation_ms_info
['fragments'] = fragments
2954 # If there is a fragments key available then we correctly recognized fragmented media.
2955 # Otherwise we will assume unfragmented media with direct access. Technically, such
2956 # assumption is not necessarily correct since we may simply have no support for
2957 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2958 if 'fragments' in representation_ms_info
:
2960 # NB: mpd_url may be empty when MPD manifest is parsed from a string
2961 'url': mpd_url
or base_url
,
2962 'fragment_base_url': base_url
,
2964 'protocol': 'http_dash_segments' if mime_type
!= 'image/jpeg' else 'mhtml',
2966 if 'initialization_url' in representation_ms_info
:
2967 initialization_url
= representation_ms_info
['initialization_url']
2968 if not f
.get('url'):
2969 f
['url'] = initialization_url
2970 f
['fragments'].append({location_key(initialization_url): initialization_url}
)
2971 f
['fragments'].extend(representation_ms_info
['fragments'])
2973 # Assuming direct URL to unfragmented media.
2975 if content_type
in ('video', 'audio', 'image/jpeg'):
2976 f
['manifest_stream_number'] = stream_numbers
[f
['url']]
2977 stream_numbers
[f
['url']] += 1
2979 elif content_type
== 'text':
2980 subtitles
.setdefault(lang
or 'und', []).append(f
)
2982 return formats
, subtitles
2984 def _extract_ism_formats(self
, *args
, **kwargs
):
2985 fmts
, subs
= self
._extract
_ism
_formats
_and
_subtitles
(*args
, **kwargs
)
2987 self
._report
_ignoring
_subs
('ISM')
2990 def _extract_ism_formats_and_subtitles(self
, ism_url
, video_id
, ism_id
=None, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
):
2991 res
= self
._download
_xml
_handle
(
2993 note
='Downloading ISM manifest' if note
is None else note
,
2994 errnote
='Failed to download ISM manifest' if errnote
is None else errnote
,
2995 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
3002 return self
._parse
_ism
_formats
_and
_subtitles
(ism_doc
, urlh
.geturl(), ism_id
)
3004 def _parse_ism_formats_and_subtitles(self
, ism_doc
, ism_url
, ism_id
=None):
3006 Parse formats from ISM manifest.
3008 1. [MS-SSTR]: Smooth Streaming Protocol,
3009 https://msdn.microsoft.com/en-us/library/ff469518.aspx
3011 if ism_doc
.get('IsLive') == 'TRUE':
3014 duration
= int(ism_doc
.attrib
['Duration'])
3015 timescale
= int_or_none(ism_doc
.get('TimeScale')) or 10000000
3019 for stream
in ism_doc
.findall('StreamIndex'):
3020 stream_type
= stream
.get('Type')
3021 if stream_type
not in ('video', 'audio', 'text'):
3023 url_pattern
= stream
.attrib
['Url']
3024 stream_timescale
= int_or_none(stream
.get('TimeScale')) or timescale
3025 stream_name
= stream
.get('Name')
3026 stream_language
= stream
.get('Language', 'und')
3027 for track
in stream
.findall('QualityLevel'):
3028 fourcc
= track
.get('FourCC') or ('AACL' if track
.get('AudioTag') == '255' else None)
3029 # TODO: add support for WVC1 and WMAP
3030 if fourcc
not in ('H264', 'AVC1', 'AACL', 'TTML'):
3031 self
.report_warning('%s is not a supported codec' % fourcc
)
3033 tbr
= int(track
.attrib
['Bitrate']) // 1000
3034 # [1] does not mention Width and Height attributes. However,
3035 # they're often present while MaxWidth and MaxHeight are
3036 # missing, so should be used as fallbacks
3037 width
= int_or_none(track
.get('MaxWidth') or track
.get('Width'))
3038 height
= int_or_none(track
.get('MaxHeight') or track
.get('Height'))
3039 sampling_rate
= int_or_none(track
.get('SamplingRate'))
3041 track_url_pattern
= re
.sub(r
'{[Bb]itrate}', track
.attrib
['Bitrate'], url_pattern
)
3042 track_url_pattern
= compat_urlparse
.urljoin(ism_url
, track_url_pattern
)
3048 stream_fragments
= stream
.findall('c')
3049 for stream_fragment_index
, stream_fragment
in enumerate(stream_fragments
):
3050 fragment_ctx
['time'] = int_or_none(stream_fragment
.get('t')) or fragment_ctx
['time']
3051 fragment_repeat
= int_or_none(stream_fragment
.get('r')) or 1
3052 fragment_ctx
['duration'] = int_or_none(stream_fragment
.get('d'))
3053 if not fragment_ctx
['duration']:
3055 next_fragment_time
= int(stream_fragment
[stream_fragment_index
+ 1].attrib
['t'])
3057 next_fragment_time
= duration
3058 fragment_ctx
['duration'] = (next_fragment_time
- fragment_ctx
['time']) / fragment_repeat
3059 for _
in range(fragment_repeat
):
3061 'url': re
.sub(r
'{start[ _]time}', compat_str(fragment_ctx
['time']), track_url_pattern
),
3062 'duration': fragment_ctx
['duration'] / stream_timescale
,
3064 fragment_ctx
['time'] += fragment_ctx
['duration']
3066 if stream_type
== 'text':
3067 subtitles
.setdefault(stream_language
, []).append({
3071 'manifest_url': ism_url
,
3072 'fragments': fragments
,
3073 '_download_params': {
3074 'stream_type': stream_type
,
3075 'duration': duration
,
3076 'timescale': stream_timescale
,
3078 'language': stream_language
,
3079 'codec_private_data': track
.get('CodecPrivateData'),
3082 elif stream_type
in ('video', 'audio'):
3084 'format_id': join_nonempty(ism_id
, stream_name
, tbr
),
3086 'manifest_url': ism_url
,
3087 'ext': 'ismv' if stream_type
== 'video' else 'isma',
3091 'asr': sampling_rate
,
3092 'vcodec': 'none' if stream_type
== 'audio' else fourcc
,
3093 'acodec': 'none' if stream_type
== 'video' else fourcc
,
3095 'fragments': fragments
,
3096 'has_drm': ism_doc
.find('Protection') is not None,
3097 '_download_params': {
3098 'stream_type': stream_type
,
3099 'duration': duration
,
3100 'timescale': stream_timescale
,
3101 'width': width
or 0,
3102 'height': height
or 0,
3104 'language': stream_language
,
3105 'codec_private_data': track
.get('CodecPrivateData'),
3106 'sampling_rate': sampling_rate
,
3107 'channels': int_or_none(track
.get('Channels', 2)),
3108 'bits_per_sample': int_or_none(track
.get('BitsPerSample', 16)),
3109 'nal_unit_length_field': int_or_none(track
.get('NALUnitLengthField', 4)),
3112 return formats
, subtitles
3114 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):
3115 def absolute_url(item_url
):
3116 return urljoin(base_url
, item_url
)
3118 def parse_content_type(content_type
):
3119 if not content_type
:
3121 ctr
= re
.search(r
'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type
)
3123 mimetype
, codecs
= ctr
.groups()
3124 f
= parse_codecs(codecs
)
3125 f
['ext'] = mimetype2ext(mimetype
)
3129 def _media_formats(src
, cur_media_type
, type_info
={}):
3130 full_url
= absolute_url(src
)
3131 ext
= type_info
.get('ext') or determine_ext(full_url
)
3133 is_plain_url
= False
3134 formats
= self
._extract
_m
3u8_formats
(
3135 full_url
, video_id
, ext
='mp4',
3136 entry_protocol
=m3u8_entry_protocol
, m3u8_id
=m3u8_id
,
3137 preference
=preference
, quality
=quality
, fatal
=False)
3139 is_plain_url
= False
3140 formats
= self
._extract
_mpd
_formats
(
3141 full_url
, video_id
, mpd_id
=mpd_id
, fatal
=False)
3146 'vcodec': 'none' if cur_media_type
== 'audio' else None,
3148 return is_plain_url
, formats
3151 # amp-video and amp-audio are very similar to their HTML5 counterparts
3152 # so we wll include them right here (see
3153 # https://www.ampproject.org/docs/reference/components/amp-video)
3154 # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/
3155 _MEDIA_TAG_NAME_RE
= r
'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)'
3156 media_tags
= [(media_tag
, media_tag_name
, media_type
, '')
3157 for media_tag
, media_tag_name
, media_type
3158 in re
.findall(r
'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE
, webpage
)]
3159 media_tags
.extend(re
.findall(
3160 # We only allow video|audio followed by a whitespace or '>'.
3161 # Allowing more characters may end up in significant slow down (see
3162 # https://github.com/ytdl-org/youtube-dl/issues/11979, example URL:
3163 # http://www.porntrex.com/maps/videositemap.xml).
3164 r
'(?s)(<(?P<tag>%s)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>' % _MEDIA_TAG_NAME_RE
, webpage
))
3165 for media_tag
, _
, media_type
, media_content
in media_tags
:
3170 media_attributes
= extract_attributes(media_tag
)
3171 src
= strip_or_none(media_attributes
.get('src'))
3173 _
, formats
= _media_formats(src
, media_type
)
3174 media_info
['formats'].extend(formats
)
3175 media_info
['thumbnail'] = absolute_url(media_attributes
.get('poster'))
3177 for source_tag
in re
.findall(r
'<source[^>]+>', media_content
):
3178 s_attr
= extract_attributes(source_tag
)
3179 # data-video-src and data-src are non standard but seen
3180 # several times in the wild
3181 src
= strip_or_none(dict_get(s_attr
, ('src', 'data-video-src', 'data-src')))
3184 f
= parse_content_type(s_attr
.get('type'))
3185 is_plain_url
, formats
= _media_formats(src
, media_type
, f
)
3187 # width, height, res, label and title attributes are
3188 # all not standard but seen several times in the wild
3191 for lbl
in ('label', 'title')
3192 if str_or_none(s_attr
.get(lbl
))
3194 width
= int_or_none(s_attr
.get('width'))
3195 height
= (int_or_none(s_attr
.get('height'))
3196 or int_or_none(s_attr
.get('res')))
3197 if not width
or not height
:
3199 resolution
= parse_resolution(lbl
)
3202 width
= width
or resolution
.get('width')
3203 height
= height
or resolution
.get('height')
3205 tbr
= parse_bitrate(lbl
)
3214 'format_id': s_attr
.get('label') or s_attr
.get('title'),
3216 f
.update(formats
[0])
3217 media_info
['formats'].append(f
)
3219 media_info
['formats'].extend(formats
)
3220 for track_tag
in re
.findall(r
'<track[^>]+>', media_content
):
3221 track_attributes
= extract_attributes(track_tag
)
3222 kind
= track_attributes
.get('kind')
3223 if not kind
or kind
in ('subtitles', 'captions'):
3224 src
= strip_or_none(track_attributes
.get('src'))
3227 lang
= track_attributes
.get('srclang') or track_attributes
.get('lang') or track_attributes
.get('label')
3228 media_info
['subtitles'].setdefault(lang
, []).append({
3229 'url': absolute_url(src
),
3231 for f
in media_info
['formats']:
3232 f
.setdefault('http_headers', {})['Referer'] = base_url
3233 if media_info
['formats'] or media_info
['subtitles']:
3234 entries
.append(media_info
)
3237 def _extract_akamai_formats(self
, *args
, **kwargs
):
3238 fmts
, subs
= self
._extract
_akamai
_formats
_and
_subtitles
(*args
, **kwargs
)
3240 self
._report
_ignoring
_subs
('akamai')
3243 def _extract_akamai_formats_and_subtitles(self
, manifest_url
, video_id
, hosts
={}):
3244 signed
= 'hdnea=' in manifest_url
3246 # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
3247 manifest_url
= re
.sub(
3248 r
'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
3249 '', manifest_url
).strip('?')
3254 hdcore_sign
= 'hdcore=3.7.0'
3255 f4m_url
= re
.sub(r
'(https?://[^/]+)/i/', r
'\1/z/', manifest_url
).replace('/master.m3u8', '/manifest.f4m')
3256 hds_host
= hosts
.get('hds')
3258 f4m_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hds_host
, f4m_url
)
3259 if 'hdcore=' not in f4m_url
:
3260 f4m_url
+= ('&' if '?' in f4m_url
else '?') + hdcore_sign
3261 f4m_formats
= self
._extract
_f
4m
_formats
(
3262 f4m_url
, video_id
, f4m_id
='hds', fatal
=False)
3263 for entry
in f4m_formats
:
3264 entry
.update({'extra_param_to_segment_url': hdcore_sign}
)
3265 formats
.extend(f4m_formats
)
3267 m3u8_url
= re
.sub(r
'(https?://[^/]+)/z/', r
'\1/i/', manifest_url
).replace('/manifest.f4m', '/master.m3u8')
3268 hls_host
= hosts
.get('hls')
3270 m3u8_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hls_host
, m3u8_url
)
3271 m3u8_formats
, m3u8_subtitles
= self
._extract
_m
3u8_formats
_and
_subtitles
(
3272 m3u8_url
, video_id
, 'mp4', 'm3u8_native',
3273 m3u8_id
='hls', fatal
=False)
3274 formats
.extend(m3u8_formats
)
3275 subtitles
= self
._merge
_subtitles
(subtitles
, m3u8_subtitles
)
3277 http_host
= hosts
.get('http')
3278 if http_host
and m3u8_formats
and not signed
:
3279 REPL_REGEX
= r
'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
3280 qualities
= re
.match(REPL_REGEX
, m3u8_url
).group(2).split(',')
3281 qualities_length
= len(qualities
)
3282 if len(m3u8_formats
) in (qualities_length
, qualities_length
+ 1):
3284 for f
in m3u8_formats
:
3285 if f
['vcodec'] != 'none':
3286 for protocol
in ('http', 'https'):
3288 del http_f
['manifest_url']
3290 REPL_REGEX
, protocol
+ r
'://%s/\g<1>%s\3' % (http_host
, qualities
[i
]), f
['url'])
3292 'format_id': http_f
['format_id'].replace('hls-', protocol
+ '-'),
3294 'protocol': protocol
,
3296 formats
.append(http_f
)
3299 return formats
, subtitles
3301 def _extract_wowza_formats(self
, url
, video_id
, m3u8_entry_protocol
='m3u8_native', skip_protocols
=[]):
3302 query
= compat_urlparse
.urlparse(url
).query
3303 url
= re
.sub(r
'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url
)
3305 r
'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url
)
3306 url_base
= mobj
.group('url')
3307 http_base_url
= '%s%s:%s' % ('http', mobj
.group('s') or '', url_base
)
3310 def manifest_url(manifest
):
3311 m_url
= '%s/%s' % (http_base_url
, manifest
)
3313 m_url
+= '?%s' % query
3316 if 'm3u8' not in skip_protocols
:
3317 formats
.extend(self
._extract
_m
3u8_formats
(
3318 manifest_url('playlist.m3u8'), video_id
, 'mp4',
3319 m3u8_entry_protocol
, m3u8_id
='hls', fatal
=False))
3320 if 'f4m' not in skip_protocols
:
3321 formats
.extend(self
._extract
_f
4m
_formats
(
3322 manifest_url('manifest.f4m'),
3323 video_id
, f4m_id
='hds', fatal
=False))
3324 if 'dash' not in skip_protocols
:
3325 formats
.extend(self
._extract
_mpd
_formats
(
3326 manifest_url('manifest.mpd'),
3327 video_id
, mpd_id
='dash', fatal
=False))
3328 if re
.search(r
'(?:/smil:|\.smil)', url_base
):
3329 if 'smil' not in skip_protocols
:
3330 rtmp_formats
= self
._extract
_smil
_formats
(
3331 manifest_url('jwplayer.smil'),
3332 video_id
, fatal
=False)
3333 for rtmp_format
in rtmp_formats
:
3334 rtsp_format
= rtmp_format
.copy()
3335 rtsp_format
['url'] = '%s/%s' % (rtmp_format
['url'], rtmp_format
['play_path'])
3336 del rtsp_format
['play_path']
3337 del rtsp_format
['ext']
3338 rtsp_format
.update({
3339 'url': rtsp_format
['url'].replace('rtmp://', 'rtsp://'),
3340 'format_id': rtmp_format
['format_id'].replace('rtmp', 'rtsp'),
3343 formats
.extend([rtmp_format
, rtsp_format
])
3345 for protocol
in ('rtmp', 'rtsp'):
3346 if protocol
not in skip_protocols
:
3348 'url': '%s:%s' % (protocol
, url_base
),
3349 'format_id': protocol
,
3350 'protocol': protocol
,
3354 def _find_jwplayer_data(self
, webpage
, video_id
=None, transform_source
=js_to_json
):
3356 r
'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P
=quote
)\
)(?
!</script
>).*?\
.setup\s
*\
((?P
<options
>[^
)]+)\
)',
3360 jwplayer_data = self._parse_json(mobj.group('options
'),
3362 transform_source=transform_source)
3363 except ExtractorError:
3366 if isinstance(jwplayer_data, dict):
3367 return jwplayer_data
3369 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
3370 jwplayer_data = self._find_jwplayer_data(
3371 webpage, video_id, transform_source=js_to_json)
3372 return self._parse_jwplayer_data(
3373 jwplayer_data, video_id, *args, **kwargs)
3375 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
3376 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3377 # JWPlayer backward compatibility: flattened playlists
3378 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
3379 if 'playlist
' not in jwplayer_data:
3380 jwplayer_data = {'playlist': [jwplayer_data]}
3384 # JWPlayer backward compatibility: single playlist item
3385 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
3386 if not isinstance(jwplayer_data['playlist
'], list):
3387 jwplayer_data['playlist
'] = [jwplayer_data['playlist
']]
3389 for video_data in jwplayer_data['playlist
']:
3390 # JWPlayer backward compatibility: flattened sources
3391 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
3392 if 'sources
' not in video_data:
3393 video_data['sources
'] = [video_data]
3395 this_video_id = video_id or video_data['mediaid
']
3397 formats = self._parse_jwplayer_formats(
3398 video_data['sources
'], video_id=this_video_id, m3u8_id=m3u8_id,
3399 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
3402 tracks = video_data.get('tracks
')
3403 if tracks and isinstance(tracks, list):
3404 for track in tracks:
3405 if not isinstance(track, dict):
3407 track_kind = track.get('kind
')
3408 if not track_kind or not isinstance(track_kind, compat_str):
3410 if track_kind.lower() not in ('captions
', 'subtitles
'):
3412 track_url = urljoin(base_url, track.get('file'))
3415 subtitles.setdefault(track.get('label
') or 'en
', []).append({
3416 'url
': self._proto_relative_url(track_url)
3420 'id': this_video_id,
3421 'title
': unescapeHTML(video_data['title
'] if require_title else video_data.get('title
')),
3422 'description
': clean_html(video_data.get('description
')),
3423 'thumbnail
': urljoin(base_url, self._proto_relative_url(video_data.get('image
'))),
3424 'timestamp
': int_or_none(video_data.get('pubdate
')),
3425 'duration
': float_or_none(jwplayer_data.get('duration
') or video_data.get('duration
')),
3426 'subtitles
': subtitles,
3428 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
3429 if len(formats) == 1 and re.search(r'^
(?
:http|
//).*(?
:youtube\
.com|youtu\
.be
)/.+', formats[0]['url
']):
3431 '_type
': 'url_transparent
',
3432 'url
': formats[0]['url
'],
3435 self._sort_formats(formats)
3436 entry['formats
'] = formats
3437 entries.append(entry)
3438 if len(entries) == 1:
3441 return self.playlist_result(entries)
3443 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
3444 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3447 for source in jwplayer_sources_data:
3448 if not isinstance(source, dict):
3450 source_url = urljoin(
3451 base_url, self._proto_relative_url(source.get('file')))
3452 if not source_url or source_url in urls:
3454 urls.append(source_url)
3455 source_type = source.get('type') or ''
3456 ext = mimetype2ext(source_type) or determine_ext(source_url)
3457 if source_type == 'hls
' or ext == 'm3u8
':
3458 formats.extend(self._extract_m3u8_formats(
3459 source_url, video_id, 'mp4
', entry_protocol='m3u8_native
',
3460 m3u8_id=m3u8_id, fatal=False))
3461 elif source_type == 'dash
' or ext == 'mpd
':
3462 formats.extend(self._extract_mpd_formats(
3463 source_url, video_id, mpd_id=mpd_id, fatal=False))
3465 formats.extend(self._extract_smil_formats(
3466 source_url, video_id, fatal=False))
3467 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3468 elif source_type.startswith('audio
') or ext in (
3469 'oga
', 'aac
', 'mp3
', 'mpeg
', 'vorbis
'):
3476 height = int_or_none(source.get('height
'))
3478 # Often no height is provided but there is a label in
3479 # format like "1080p", "720p SD", or 1080.
3480 height = int_or_none(self._search_regex(
3481 r'^
(\d{3,4}
)[pP
]?
(?
:\b|$
)', compat_str(source.get('label
') or ''),
3482 'height
', default=None))
3485 'width
': int_or_none(source.get('width
')),
3487 'tbr
': int_or_none(source.get('bitrate
')),
3490 if source_url.startswith('rtmp
'):
3491 a_format['ext
'] = 'flv
'
3492 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3493 # of jwplayer.flash.swf
3494 rtmp_url_parts = re.split(
3495 r'((?
:mp4|mp3|flv
):)', source_url, 1)
3496 if len(rtmp_url_parts) == 3:
3497 rtmp_url, prefix, play_path = rtmp_url_parts
3500 'play_path
': prefix + play_path,
3503 a_format.update(rtmp_params)
3504 formats.append(a_format)
3507 def _live_title(self, name):
3508 self._downloader.deprecation_warning('yt_dlp
.InfoExtractor
._live
_title
is deprecated
and does
not work
as expected
')
3511 def _int(self, v, name, fatal=False, **kwargs):
3512 res = int_or_none(v, **kwargs)
3514 msg = 'Failed to extract
%s: Could
not parse value
%r' % (name, v)
3516 raise ExtractorError(msg)
3518 self.report_warning(msg)
3521 def _float(self, v, name, fatal=False, **kwargs):
3522 res = float_or_none(v, **kwargs)
3524 msg = 'Failed to extract
%s: Could
not parse value
%r' % (name, v)
3526 raise ExtractorError(msg)
3528 self.report_warning(msg)
3531 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
3532 path='/', secure=False, discard=False, rest={}, **kwargs):
3533 cookie = compat_cookiejar_Cookie(
3534 0, name, value, port, port is not None, domain, True,
3535 domain.startswith('.'), path, True, secure, expire_time,
3536 discard, None, None, rest)
3537 self._downloader.cookiejar.set_cookie(cookie)
3539 def _get_cookies(self, url):
3540 """ Return a compat_cookies_SimpleCookie with the cookies for the url """
3541 req = sanitized_Request(url)
3542 self._downloader.cookiejar.add_cookie_header(req)
3543 return compat_cookies_SimpleCookie(req.get_header('Cookie
'))
3545 def _apply_first_set_cookie_header(self, url_handle, cookie):
3547 Apply first Set-Cookie header instead of the last. Experimental.
3549 Some sites (e.g. [1-3]) may serve two cookies under the same name
3550 in Set-Cookie header and expect the first (old) one to be set rather
3551 than second (new). However, as of RFC6265 the newer one cookie
3552 should be set into cookie store what actually happens.
3553 We will workaround this issue by resetting the cookie to
3554 the first one manually.
3555 1. https://new.vk.com/
3556 2. https://github.com/ytdl-org/youtube-dl/issues/9841#issuecomment-227871201
3557 3. https://learning.oreilly.com/
3559 for header, cookies in url_handle.headers.items():
3560 if header.lower() != 'set-cookie
':
3562 if sys.version_info[0] >= 3:
3563 cookies = cookies.encode('iso
-8859-1')
3564 cookies = cookies.decode('utf
-8')
3565 cookie_value = re.search(
3566 r'%s=(.+?
);.*?
\b[Dd
]omain
=(.+?
)(?
:[,;]|$
)' % cookie, cookies)
3568 value, domain = cookie_value.groups()
3569 self._set_cookie(domain, cookie, value)
3572 def get_testcases(self, include_onlymatching=False):
3573 t = getattr(self, '_TEST
', None)
3575 assert not hasattr(self, '_TESTS
'), \
3576 '%s has _TEST
and _TESTS
' % type(self).__name__
3579 tests = getattr(self, '_TESTS
', [])
3581 if not include_onlymatching and t.get('only_matching
', False):
3583 t['name
'] = type(self).__name__[:-len('IE
')]
3586 def is_suitable(self, age_limit):
3587 """ Test whether the extractor is generally suitable for the given
3588 age limit (i.e. pornographic sites are not, all others usually are) """
3590 any_restricted = False
3591 for tc in self.get_testcases(include_onlymatching=False):
3592 if tc.get('playlist
', []):
3593 tc = tc['playlist
'][0]
3594 is_restricted = age_restricted(
3595 tc.get('info_dict
', {}).get('age_limit
'), age_limit)
3596 if not is_restricted:
3598 any_restricted = any_restricted or is_restricted
3599 return not any_restricted
3601 def extract_subtitles(self, *args, **kwargs):
3602 if (self.get_param('writesubtitles
', False)
3603 or self.get_param('listsubtitles
')):
3604 return self._get_subtitles(*args, **kwargs)
3607 def _get_subtitles(self, *args, **kwargs):
3608 raise NotImplementedError('This method must be implemented by subclasses
')
3610 def extract_comments(self, *args, **kwargs):
3611 if not self.get_param('getcomments
'):
3613 generator = self._get_comments(*args, **kwargs)
3620 comments.append(next(generator))
3621 except StopIteration:
3623 except KeyboardInterrupt:
3624 self.to_screen('Interrupted by user
')
3625 except Exception as e:
3626 if self.get_param('ignoreerrors
') is not True:
3628 self._downloader.report_error(e)
3629 comment_count = len(comments)
3630 self.to_screen(f'Extracted {comment_count} comments
')
3632 'comments
': comments,
3633 'comment_count
': None if interrupted else comment_count
3637 def _get_comments(self, *args, **kwargs):
3638 raise NotImplementedError('This method must be implemented by subclasses
')
3641 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
3642 """ Merge subtitle items for one language. Items with duplicated URLs
3643 will be dropped. """
3644 list1_urls = set([item['url
'] for item in subtitle_list1])
3645 ret = list(subtitle_list1)
3646 ret.extend([item for item in subtitle_list2 if item['url
'] not in list1_urls])
3650 def _merge_subtitles(cls, *dicts, target=None):
3651 """ Merge subtitle dictionaries, language by language. """
3655 for lang, subs in d.items():
3656 target[lang] = cls._merge_subtitle_items(target.get(lang, []), subs)
3659 def extract_automatic_captions(self, *args, **kwargs):
3660 if (self.get_param('writeautomaticsub
', False)
3661 or self.get_param('listsubtitles
')):
3662 return self._get_automatic_captions(*args, **kwargs)
3665 def _get_automatic_captions(self, *args, **kwargs):
3666 raise NotImplementedError('This method must be implemented by subclasses
')
3668 def mark_watched(self, *args, **kwargs):
3669 if not self.get_param('mark_watched
', False):
3671 if (self._get_login_info()[0] is not None
3672 or self.get_param('cookiefile
')
3673 or self.get_param('cookiesfrombrowser
')):
3674 self._mark_watched(*args, **kwargs)
3676 def _mark_watched(self, *args, **kwargs):
3677 raise NotImplementedError('This method must be implemented by subclasses
')
3679 def geo_verification_headers(self):
3681 geo_verification_proxy = self.get_param('geo_verification_proxy
')
3682 if geo_verification_proxy:
3683 headers['Ytdl
-request
-proxy
'] = geo_verification_proxy
3686 def _generic_id(self, url):
3687 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
3689 def _generic_title(self, url):
3690 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
3693 def _availability(is_private=None, needs_premium=None, needs_subscription=None, needs_auth=None, is_unlisted=None):
3694 all_known = all(map(
3695 lambda x: x is not None,
3696 (is_private, needs_premium, needs_subscription, needs_auth, is_unlisted)))
3698 'private
' if is_private
3699 else 'premium_only
' if needs_premium
3700 else 'subscriber_only
' if needs_subscription
3701 else 'needs_auth
' if needs_auth
3702 else 'unlisted
' if is_unlisted
3703 else 'public
' if all_known
3706 def _configuration_arg(self, key, default=NO_DEFAULT, *, ie_key=None, casesense=False):
3708 @returns A list of values for the extractor argument given by "key"
3709 or "default" if no such key is present
3710 @param default The default value to return when the key is not present (default: [])
3711 @param casesense When false, the values are converted to lower case
3714 self._downloader.params, ('extractor_args
', (ie_key or self.ie_key()).lower(), key))
3716 return [] if default is NO_DEFAULT else default
3717 return list(val) if casesense else [x.lower() for x in val]
3719 def _yes_playlist(self, playlist_id, video_id, smuggled_data=None, *, playlist_label='playlist
', video_label='video
'):
3720 if not playlist_id or not video_id:
3723 no_playlist = (smuggled_data or {}).get('force_noplaylist
')
3724 if no_playlist is not None:
3725 return not no_playlist
3727 video_id = '' if video_id is True else f' {video_id}
'
3728 playlist_id = '' if playlist_id is True else f' {playlist_id}
'
3729 if self.get_param('noplaylist
'):
3730 self.to_screen(f'Downloading just the {video_label}{video_id} because of
--no
-playlist
')
3732 self.to_screen(f'Downloading {playlist_label}{playlist_id}
- add
--no
-playlist to download just the {video_label}{video_id}
')
3736 class SearchInfoExtractor(InfoExtractor):
3738 Base class for paged search queries extractors.
3739 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
3740 Instances should define _SEARCH_KEY and optionally _MAX_RESULTS
3743 _MAX_RESULTS = float('inf
')
3746 def _make_valid_url(cls):
3747 return r'%s(?P
<prefix
>|
[1-9][0-9]*|all
):(?P
<query
>[\s\S
]+)' % cls._SEARCH_KEY
3749 def _real_extract(self, query):
3750 prefix, query = self._match_valid_url(query).group('prefix
', 'query
')
3752 return self._get_n_results(query, 1)
3753 elif prefix == 'all
':
3754 return self._get_n_results(query, self._MAX_RESULTS)
3758 raise ExtractorError('invalid download number
%s for query
"%s"' % (n, query))
3759 elif n > self._MAX_RESULTS:
3760 self.report_warning('%s returns
max %i results (you requested
%i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3761 n = self._MAX_RESULTS
3762 return self._get_n_results(query, n)
3764 def _get_n_results(self, query, n):
3765 """Get a specified number of results for a query.
3766 Either this function or _search_results must be overridden by subclasses """
3767 return self.playlist_result(
3768 itertools.islice(self._search_results(query), 0, None if n == float('inf
') else n),
3771 def _search_results(self, query):
3772 """Returns an iterator of search results"""
3773 raise NotImplementedError('This method must be implemented by subclasses
')
3776 def SEARCH_KEY(self):
3777 return self._SEARCH_KEY