2 from __future__
import unicode_literals
17 from ..compat
import (
18 compat_cookiejar_Cookie
,
19 compat_cookies_SimpleCookie
,
21 compat_etree_fromstring
,
28 compat_urllib_parse_unquote
,
29 compat_urllib_parse_urlencode
,
30 compat_urllib_request
,
32 compat_xml_parse_error
,
34 from ..downloader
import FileDownloader
35 from ..downloader
.f4m
import (
37 remove_encrypted_media
,
69 parse_m3u8_attributes
,
95 class InfoExtractor(object):
96 """Information Extractor class.
98 Information extractors are the classes that, given a URL, extract
99 information about the video (or videos) the URL refers to. This
100 information includes the real video URL, the video title, author and
101 others. The information is stored in a dictionary which is then
102 passed to the YoutubeDL. The YoutubeDL processes this
103 information possibly downloading the video to the file system, among
104 other possible outcomes.
106 The type field determines the type of the result.
107 By far the most common value (and the default if _type is missing) is
108 "video", which indicates a single video.
110 For a video, the dictionaries must include the following fields:
112 id: Video identifier.
113 title: Video title, unescaped.
115 Additionally, it must contain either a formats entry or a url one:
117 formats: A list of dictionaries for each format available, ordered
118 from worst to best quality.
121 * url The mandatory URL representing the media:
122 for plain file media - HTTP URL of this file,
124 for HLS - URL of the M3U8 media playlist,
125 for HDS - URL of the F4M manifest,
127 - HTTP URL to plain file media (in case of
129 - URL of the MPD manifest or base URL
130 representing the media if MPD manifest
131 is parsed from a string (in case of
133 for MSS - URL of the ISM manifest.
135 The URL of the manifest file in case of
137 for HLS - URL of the M3U8 master playlist,
138 for HDS - URL of the F4M manifest,
139 for DASH - URL of the MPD manifest,
140 for MSS - URL of the ISM manifest.
141 * ext Will be calculated from URL if missing
142 * format A human-readable description of the format
143 ("mp4 container with h264/opus").
144 Calculated from the format_id, width, height.
145 and format_note fields if missing.
146 * format_id A short description of the format
147 ("mp4_h264_opus" or "19").
148 Technically optional, but strongly recommended.
149 * format_note Additional info about the format
150 ("3D" or "DASH video")
151 * width Width of the video, if known
152 * height Height of the video, if known
153 * resolution Textual description of width and height
154 * dynamic_range The dynamic range of the video. One of:
155 "SDR" (None), "HDR10", "HDR10+, "HDR12", "HLG, "DV"
156 * tbr Average bitrate of audio and video in KBit/s
157 * abr Average audio bitrate in KBit/s
158 * acodec Name of the audio codec in use
159 * asr Audio sampling rate in Hertz
160 * vbr Average video bitrate in KBit/s
162 * vcodec Name of the video codec in use
163 * container Name of the container format
164 * filesize The number of bytes, if known in advance
165 * filesize_approx An estimate for the number of bytes
166 * player_url SWF Player URL (used for rtmpdump).
167 * protocol The protocol that will be used for the actual
168 download, lower-case. One of "http", "https" or
169 one of the protocols defined in downloader.PROTOCOL_MAP
171 Base URL for fragments. Each fragment's path
172 value (if present) will be relative to
174 * fragments A list of fragments of a fragmented media.
175 Each fragment entry must contain either an url
176 or a path. If an url is present it should be
177 considered by a client. Otherwise both path and
178 fragment_base_url must be present. Here is
179 the list of all potential fields:
180 * "url" - fragment's URL
181 * "path" - fragment's path relative to
183 * "duration" (optional, int or float)
184 * "filesize" (optional, int)
185 * is_from_start Is a live format that can be downloaded
186 from the start. Boolean
187 * preference Order number of this format. If this field is
188 present and not None, the formats get sorted
189 by this field, regardless of all other values.
190 -1 for default (order by other properties),
191 -2 or smaller for less than default.
192 < -1000 to hide the format (if there is
193 another one which is strictly better)
194 * language Language code, e.g. "de" or "en-US".
195 * language_preference Is this in the language mentioned in
197 10 if it's what the URL is about,
198 -1 for default (don't know),
199 -10 otherwise, other values reserved for now.
200 * quality Order number of the video quality of this
201 format, irrespective of the file format.
202 -1 for default (order by other properties),
203 -2 or smaller for less than default.
204 * source_preference Order number for this video source
205 (quality takes higher priority)
206 -1 for default (order by other properties),
207 -2 or smaller for less than default.
208 * http_headers A dictionary of additional HTTP headers
209 to add to the request.
210 * stretched_ratio If given and not 1, indicates that the
211 video's pixels are not square.
212 width : height ratio as float.
213 * no_resume The server does not support resuming the
214 (HTTP or RTMP) download. Boolean.
215 * has_drm The format has DRM and cannot be downloaded. Boolean
216 * downloader_options A dictionary of downloader options as
217 described in FileDownloader
218 RTMP formats can also have the additional fields: page_url,
219 app, play_path, tc_url, flash_version, rtmp_live, rtmp_conn,
220 rtmp_protocol, rtmp_real_time
222 url: Final video URL.
223 ext: Video filename extension.
224 format: The video format, defaults to ext (used for --get-format)
225 player_url: SWF Player URL (used for rtmpdump).
227 The following fields are optional:
229 direct: True if a direct video file was given (must only be set by GenericIE)
230 alt_title: A secondary title of the video.
231 display_id An alternative identifier for the video, not necessarily
232 unique, but available before title. Typically, id is
233 something like "4234987", title "Dancing naked mole rats",
234 and display_id "dancing-naked-mole-rats"
235 thumbnails: A list of dictionaries, with the following entries:
236 * "id" (optional, string) - Thumbnail format ID
238 * "preference" (optional, int) - quality of the image
239 * "width" (optional, int)
240 * "height" (optional, int)
241 * "resolution" (optional, string "{width}x{height}",
243 * "filesize" (optional, int)
244 * "http_headers" (dict) - HTTP headers for the request
245 thumbnail: Full URL to a video thumbnail image.
246 description: Full video description.
247 uploader: Full name of the video uploader.
248 license: License name the video is licensed under.
249 creator: The creator of the video.
250 timestamp: UNIX timestamp of the moment the video was uploaded
251 upload_date: Video upload date (YYYYMMDD).
252 If not explicitly set, calculated from timestamp
253 release_timestamp: UNIX timestamp of the moment the video was released.
254 If it is not clear whether to use timestamp or this, use the former
255 release_date: The date (YYYYMMDD) when the video was released.
256 If not explicitly set, calculated from release_timestamp
257 modified_timestamp: UNIX timestamp of the moment the video was last modified.
258 modified_date: The date (YYYYMMDD) when the video was last modified.
259 If not explicitly set, calculated from modified_timestamp
260 uploader_id: Nickname or id of the video uploader.
261 uploader_url: Full URL to a personal webpage of the video uploader.
262 channel: Full name of the channel the video is uploaded on.
263 Note that channel fields may or may not repeat uploader
264 fields. This depends on a particular extractor.
265 channel_id: Id of the channel.
266 channel_url: Full URL to a channel webpage.
267 channel_follower_count: Number of followers of the channel.
268 location: Physical location where the video was filmed.
269 subtitles: The available subtitles as a dictionary in the format
270 {tag: subformats}. "tag" is usually a language code, and
271 "subformats" is a list sorted from lower to higher
272 preference, each element is a dictionary with the "ext"
274 * "data": The subtitles file contents
275 * "url": A URL pointing to the subtitles file
276 It can optionally also have:
277 * "name": Name or description of the subtitles
278 * "http_headers": A dictionary of additional HTTP headers
279 to add to the request.
280 "ext" will be calculated from URL if missing
281 automatic_captions: Like 'subtitles'; contains automatically generated
282 captions instead of normal subtitles
283 duration: Length of the video in seconds, as an integer or float.
284 view_count: How many users have watched the video on the platform.
285 like_count: Number of positive ratings of the video
286 dislike_count: Number of negative ratings of the video
287 repost_count: Number of reposts of the video
288 average_rating: Average rating give by users, the scale used depends on the webpage
289 comment_count: Number of comments on the video
290 comments: A list of comments, each with one or more of the following
291 properties (all but one of text or html optional):
292 * "author" - human-readable name of the comment author
293 * "author_id" - user ID of the comment author
294 * "author_thumbnail" - The thumbnail of the comment author
296 * "html" - Comment as HTML
297 * "text" - Plain text of the comment
298 * "timestamp" - UNIX timestamp of comment
299 * "parent" - ID of the comment this one is replying to.
300 Set to "root" to indicate that this is a
301 comment to the original video.
302 * "like_count" - Number of positive ratings of the comment
303 * "dislike_count" - Number of negative ratings of the comment
304 * "is_favorited" - Whether the comment is marked as
305 favorite by the video uploader
306 * "author_is_uploader" - Whether the comment is made by
308 age_limit: Age restriction for the video, as an integer (years)
309 webpage_url: The URL to the video webpage, if given to yt-dlp it
310 should allow to get the same result again. (It will be set
311 by YoutubeDL if it's missing)
312 categories: A list of categories that the video falls in, for example
314 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
315 cast: A list of the video cast
316 is_live: True, False, or None (=unknown). Whether this video is a
317 live stream that goes on instead of a fixed-length video.
318 was_live: True, False, or None (=unknown). Whether this video was
319 originally a live stream.
320 live_status: 'is_live', 'is_upcoming', 'was_live', 'not_live' or None (=unknown)
321 If absent, automatically set from is_live, was_live
322 start_time: Time in seconds where the reproduction should start, as
323 specified in the URL.
324 end_time: Time in seconds where the reproduction should end, as
325 specified in the URL.
326 chapters: A list of dictionaries, with the following entries:
327 * "start_time" - The start time of the chapter in seconds
328 * "end_time" - The end time of the chapter in seconds
329 * "title" (optional, string)
330 playable_in_embed: Whether this video is allowed to play in embedded
331 players on other sites. Can be True (=always allowed),
332 False (=never allowed), None (=unknown), or a string
333 specifying the criteria for embedability (Eg: 'whitelist')
334 availability: Under what condition the video is available. One of
335 'private', 'premium_only', 'subscriber_only', 'needs_auth',
336 'unlisted' or 'public'. Use 'InfoExtractor._availability'
338 __post_extractor: A function to be called just before the metadata is
339 written to either disk, logger or console. The function
340 must return a dict which will be added to the info_dict.
341 This is usefull for additional information that is
342 time-consuming to extract. Note that the fields thus
343 extracted will not be available to output template and
344 match_filter. So, only "comments" and "comment_count" are
345 currently allowed to be extracted via this method.
347 The following fields should only be used when the video belongs to some logical
350 chapter: Name or title of the chapter the video belongs to.
351 chapter_number: Number of the chapter the video belongs to, as an integer.
352 chapter_id: Id of the chapter the video belongs to, as a unicode string.
354 The following fields should only be used when the video is an episode of some
355 series, programme or podcast:
357 series: Title of the series or programme the video episode belongs to.
358 series_id: Id of the series or programme the video episode belongs to, as a unicode string.
359 season: Title of the season the video episode belongs to.
360 season_number: Number of the season the video episode belongs to, as an integer.
361 season_id: Id of the season the video episode belongs to, as a unicode string.
362 episode: Title of the video episode. Unlike mandatory video title field,
363 this field should denote the exact title of the video episode
364 without any kind of decoration.
365 episode_number: Number of the video episode within a season, as an integer.
366 episode_id: Id of the video episode, as a unicode string.
368 The following fields should only be used when the media is a track or a part of
371 track: Title of the track.
372 track_number: Number of the track within an album or a disc, as an integer.
373 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
375 artist: Artist(s) of the track.
376 genre: Genre(s) of the track.
377 album: Title of the album the track belongs to.
378 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
379 album_artist: List of all artists appeared on the album (e.g.
380 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
382 disc_number: Number of the disc or other physical medium the track belongs to,
384 release_year: Year (YYYY) when the album was released.
385 composer: Composer of the piece
387 Unless mentioned otherwise, the fields should be Unicode strings.
389 Unless mentioned otherwise, None is equivalent to absence of information.
392 _type "playlist" indicates multiple videos.
393 There must be a key "entries", which is a list, an iterable, or a PagedList
394 object, each element of which is a valid dictionary by this specification.
396 Additionally, playlists can have "id", "title", and any other relevent
397 attributes with the same semantics as videos (see above).
399 It can also have the following optional fields:
401 playlist_count: The total number of videos in a playlist. If not given,
402 YoutubeDL tries to calculate it from "entries"
405 _type "multi_video" indicates that there are multiple videos that
406 form a single show, for examples multiple acts of an opera or TV episode.
407 It must have an entries key like a playlist and contain all the keys
408 required for a video at the same time.
411 _type "url" indicates that the video must be extracted from another
412 location, possibly by a different extractor. Its only required key is:
413 "url" - the next URL to extract.
414 The key "ie_key" can be set to the class name (minus the trailing "IE",
415 e.g. "Youtube") if the extractor class is known in advance.
416 Additionally, the dictionary may have any properties of the resolved entity
417 known in advance, for example "title" if the title of the referred video is
421 _type "url_transparent" entities have the same specification as "url", but
422 indicate that the given additional information is more precise than the one
423 associated with the resolved URL.
424 This is useful when a site employs a video service that hosts the video and
425 its technical metadata, but that video service does not embed a useful
426 title, description etc.
429 Subclasses of this should define a _VALID_URL regexp and, re-define the
430 _real_extract() and (optionally) _real_initialize() methods.
431 Probably, they should also be added to the list of extractors.
433 Subclasses may also override suitable() if necessary, but ensure the function
434 signature is preserved and that this function imports everything it needs
435 (except other extractors), so that lazy_extractors works correctly
437 _GEO_BYPASS attribute may be set to False in order to disable
438 geo restriction bypass mechanisms for a particular extractor.
439 Though it won't disable explicit geo restriction bypass based on
440 country code provided with geo_bypass_country.
442 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
443 countries for this extractor. One of these countries will be used by
444 geo restriction bypass mechanism right away in order to bypass
445 geo restriction, of course, if the mechanism is not disabled.
447 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
448 IP blocks in CIDR notation for this extractor. One of these IP blocks
449 will be used by geo restriction bypass mechanism similarly
452 The _WORKING attribute should be set to False for broken IEs
453 in order to warn the users and skip the tests.
458 _x_forwarded_for_ip
= None
460 _GEO_COUNTRIES
= None
461 _GEO_IP_BLOCKS
= None
465 'any': 'Use --cookies, --username and --password, or --netrc to provide account credentials',
467 'Use --cookies-from-browser or --cookies for the authentication. '
468 'See https://github.com/ytdl-org/youtube-dl#how-do-i-pass-cookies-to-youtube-dl for how to manually pass cookies'),
469 'password': 'Use --username and --password, or --netrc to provide account credentials',
472 def __init__(self
, downloader
=None):
473 """Constructor. Receives an optional downloader (a YoutubeDL instance).
474 If a downloader is not passed during initialization,
475 it must be set using "set_downloader()" before "extract()" is called"""
477 self
._x
_forwarded
_for
_ip
= None
478 self
._printed
_messages
= set()
479 self
.set_downloader(downloader
)
482 def _match_valid_url(cls
, url
):
483 # This does not use has/getattr intentionally - we want to know whether
484 # we have cached the regexp for *this* class, whereas getattr would also
485 # match the superclass
486 if '_VALID_URL_RE' not in cls
.__dict
__:
487 if '_VALID_URL' not in cls
.__dict
__:
488 cls
._VALID
_URL
= cls
._make
_valid
_url
()
489 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
490 return cls
._VALID
_URL
_RE
.match(url
)
493 def suitable(cls
, url
):
494 """Receives a URL and returns True if suitable for this IE."""
495 # This function must import everything it needs (except other extractors),
496 # so that lazy_extractors works correctly
497 return cls
._match
_valid
_url
(url
) is not None
500 def _match_id(cls
, url
):
501 return cls
._match
_valid
_url
(url
).group('id')
504 def get_temp_id(cls
, url
):
506 return cls
._match
_id
(url
)
507 except (IndexError, AttributeError):
512 """Getter method for _WORKING."""
515 def initialize(self
):
516 """Initializes an instance (authentication, etc)."""
517 self
._printed
_messages
= set()
518 self
._initialize
_geo
_bypass
({
519 'countries': self
._GEO
_COUNTRIES
,
520 'ip_blocks': self
._GEO
_IP
_BLOCKS
,
523 self
._real
_initialize
()
526 def _initialize_geo_bypass(self
, geo_bypass_context
):
528 Initialize geo restriction bypass mechanism.
530 This method is used to initialize geo bypass mechanism based on faking
531 X-Forwarded-For HTTP header. A random country from provided country list
532 is selected and a random IP belonging to this country is generated. This
533 IP will be passed as X-Forwarded-For HTTP header in all subsequent
536 This method will be used for initial geo bypass mechanism initialization
537 during the instance initialization with _GEO_COUNTRIES and
540 You may also manually call it from extractor's code if geo bypass
541 information is not available beforehand (e.g. obtained during
542 extraction) or due to some other reason. In this case you should pass
543 this information in geo bypass context passed as first argument. It may
544 contain following fields:
546 countries: List of geo unrestricted countries (similar
548 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
549 (similar to _GEO_IP_BLOCKS)
552 if not self
._x
_forwarded
_for
_ip
:
554 # Geo bypass mechanism is explicitly disabled by user
555 if not self
.get_param('geo_bypass', True):
558 if not geo_bypass_context
:
559 geo_bypass_context
= {}
561 # Backward compatibility: previously _initialize_geo_bypass
562 # expected a list of countries, some 3rd party code may still use
564 if isinstance(geo_bypass_context
, (list, tuple)):
565 geo_bypass_context
= {
566 'countries': geo_bypass_context
,
569 # The whole point of geo bypass mechanism is to fake IP
570 # as X-Forwarded-For HTTP header based on some IP block or
573 # Path 1: bypassing based on IP block in CIDR notation
575 # Explicit IP block specified by user, use it right away
576 # regardless of whether extractor is geo bypassable or not
577 ip_block
= self
.get_param('geo_bypass_ip_block', None)
579 # Otherwise use random IP block from geo bypass context but only
580 # if extractor is known as geo bypassable
582 ip_blocks
= geo_bypass_context
.get('ip_blocks')
583 if self
._GEO
_BYPASS
and ip_blocks
:
584 ip_block
= random
.choice(ip_blocks
)
587 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(ip_block
)
588 self
._downloader
.write_debug(
589 '[debug] Using fake IP %s as X-Forwarded-For' % self
._x
_forwarded
_for
_ip
)
592 # Path 2: bypassing based on country code
594 # Explicit country code specified by user, use it right away
595 # regardless of whether extractor is geo bypassable or not
596 country
= self
.get_param('geo_bypass_country', None)
598 # Otherwise use random country code from geo bypass context but
599 # only if extractor is known as geo bypassable
601 countries
= geo_bypass_context
.get('countries')
602 if self
._GEO
_BYPASS
and countries
:
603 country
= random
.choice(countries
)
606 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country
)
607 self
._downloader
.write_debug(
608 'Using fake IP %s (%s) as X-Forwarded-For' % (self
._x
_forwarded
_for
_ip
, country
.upper()))
610 def extract(self
, url
):
611 """Extracts URL information and returns it in list of dicts."""
616 self
.write_debug('Extracting URL: %s' % url
)
617 ie_result
= self
._real
_extract
(url
)
618 if ie_result
is None:
620 if self
._x
_forwarded
_for
_ip
:
621 ie_result
['__x_forwarded_for_ip'] = self
._x
_forwarded
_for
_ip
622 subtitles
= ie_result
.get('subtitles')
623 if (subtitles
and 'live_chat' in subtitles
624 and 'no-live-chat' in self
.get_param('compat_opts', [])):
625 del subtitles
['live_chat']
627 except GeoRestrictedError
as e
:
628 if self
.__maybe
_fake
_ip
_and
_retry
(e
.countries
):
631 except UnsupportedError
:
633 except ExtractorError
as e
:
635 'video_id': e
.video_id
or self
.get_temp_id(url
),
637 'tb': e
.traceback
or sys
.exc_info()[2],
638 'expected': e
.expected
,
641 if hasattr(e
, 'countries'):
642 kwargs
['countries'] = e
.countries
643 raise type(e
)(e
.orig_msg
, **kwargs
)
644 except compat_http_client
.IncompleteRead
as e
:
645 raise ExtractorError('A network error has occurred.', cause
=e
, expected
=True, video_id
=self
.get_temp_id(url
))
646 except (KeyError, StopIteration) as e
:
647 raise ExtractorError('An extractor error has occurred.', cause
=e
, video_id
=self
.get_temp_id(url
))
649 def __maybe_fake_ip_and_retry(self
, countries
):
650 if (not self
.get_param('geo_bypass_country', None)
652 and self
.get_param('geo_bypass', True)
653 and not self
._x
_forwarded
_for
_ip
655 country_code
= random
.choice(countries
)
656 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country_code
)
657 if self
._x
_forwarded
_for
_ip
:
659 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
660 % (self
._x
_forwarded
_for
_ip
, country_code
.upper()))
664 def set_downloader(self
, downloader
):
665 """Sets a YoutubeDL instance as the downloader for this IE."""
666 self
._downloader
= downloader
668 def _real_initialize(self
):
669 """Real initialization process. Redefine in subclasses."""
672 def _real_extract(self
, url
):
673 """Real extraction process. Redefine in subclasses."""
674 raise NotImplementedError('This method must be implemented by subclasses')
678 """A string for getting the InfoExtractor with get_info_extractor"""
679 return cls
.__name
__[:-2]
683 return compat_str(type(self
).__name
__[:-2])
686 def __can_accept_status_code(err
, expected_status
):
687 assert isinstance(err
, compat_urllib_error
.HTTPError
)
688 if expected_status
is None:
690 elif callable(expected_status
):
691 return expected_status(err
.code
) is True
693 return err
.code
in variadic(expected_status
)
695 def _request_webpage(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
, expected_status
=None):
697 Return the response handle.
699 See _download_webpage docstring for arguments specification.
701 if not self
._downloader
._first
_webpage
_request
:
702 sleep_interval
= self
.get_param('sleep_interval_requests') or 0
703 if sleep_interval
> 0:
704 self
.to_screen('Sleeping %s seconds ...' % sleep_interval
)
705 time
.sleep(sleep_interval
)
707 self
._downloader
._first
_webpage
_request
= False
710 self
.report_download_webpage(video_id
)
711 elif note
is not False:
713 self
.to_screen('%s' % (note
,))
715 self
.to_screen('%s: %s' % (video_id
, note
))
717 # Some sites check X-Forwarded-For HTTP header in order to figure out
718 # the origin of the client behind proxy. This allows bypassing geo
719 # restriction by faking this header's value to IP that belongs to some
720 # geo unrestricted country. We will do so once we encounter any
721 # geo restriction error.
722 if self
._x
_forwarded
_for
_ip
:
723 if 'X-Forwarded-For' not in headers
:
724 headers
['X-Forwarded-For'] = self
._x
_forwarded
_for
_ip
726 if isinstance(url_or_request
, compat_urllib_request
.Request
):
727 url_or_request
= update_Request(
728 url_or_request
, data
=data
, headers
=headers
, query
=query
)
731 url_or_request
= update_url_query(url_or_request
, query
)
732 if data
is not None or headers
:
733 url_or_request
= sanitized_Request(url_or_request
, data
, headers
)
735 return self
._downloader
.urlopen(url_or_request
)
736 except network_exceptions
as err
:
737 if isinstance(err
, compat_urllib_error
.HTTPError
):
738 if self
.__can
_accept
_status
_code
(err
, expected_status
):
739 # Retain reference to error to prevent file object from
740 # being closed before it can be read. Works around the
741 # effects of <https://bugs.python.org/issue15002>
742 # introduced in Python 3.4.1.
749 errnote
= 'Unable to download webpage'
751 errmsg
= '%s: %s' % (errnote
, error_to_compat_str(err
))
753 raise ExtractorError(errmsg
, cause
=err
)
755 self
.report_warning(errmsg
)
758 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):
760 Return a tuple (page content as string, URL handle).
762 See _download_webpage docstring for arguments specification.
764 # Strip hashes from the URL (#1038)
765 if isinstance(url_or_request
, (compat_str
, str)):
766 url_or_request
= url_or_request
.partition('#')[0]
768 urlh
= self
._request
_webpage
(url_or_request
, video_id
, note
, errnote
, fatal
, data
=data
, headers
=headers
, query
=query
, expected_status
=expected_status
)
772 content
= self
._webpage
_read
_content
(urlh
, url_or_request
, video_id
, note
, errnote
, fatal
, encoding
=encoding
)
773 return (content
, urlh
)
776 def _guess_encoding_from_content(content_type
, webpage_bytes
):
777 m
= re
.match(r
'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type
)
779 encoding
= m
.group(1)
781 m
= re
.search(br
'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
782 webpage_bytes[:1024])
784 encoding = m.group(1).decode('ascii')
785 elif webpage_bytes.startswith(b'\xff\xfe'):
792 def __check_blocked(self, content):
793 first_block = content[:512]
794 if ('<title>Access to this site is blocked</title>' in content
795 and 'Websense' in first_block):
796 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
797 blocked_iframe = self._html_search_regex(
798 r'<iframe src="([^
"]+)"', content,
799 'Websense information URL
', default=None)
801 msg += ' Visit
%s for more details
' % blocked_iframe
802 raise ExtractorError(msg, expected=True)
803 if '<title
>The URL you requested has been blocked
</title
>' in first_block:
805 'Access to this webpage has been blocked by Indian censorship
. '
806 'Use a VPN
or proxy
server (with --proxy
) to route around it
.')
807 block_msg = self._html_search_regex(
808 r'</h1
><p
>(.*?
)</p
>',
809 content, 'block message
', default=None)
811 msg += ' (Message
: "%s")' % block_msg.replace('\n', ' ')
812 raise ExtractorError(msg, expected=True)
813 if ('<title
>TTK
:: Доступ к ресурсу ограничен
</title
>' in content
814 and 'blocklist
.rkn
.gov
.ru
' in content):
815 raise ExtractorError(
816 'Access to this webpage has been blocked by decision of the Russian government
. '
817 'Visit http
://blocklist
.rkn
.gov
.ru
/ for a block reason
.',
820 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
821 content_type = urlh.headers.get('Content
-Type
', '')
822 webpage_bytes = urlh.read()
823 if prefix is not None:
824 webpage_bytes = prefix + webpage_bytes
826 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
827 if self.get_param('dump_intermediate_pages
', False):
828 self.to_screen('Dumping request to
' + urlh.geturl())
829 dump = base64.b64encode(webpage_bytes).decode('ascii
')
830 self._downloader.to_screen(dump)
831 if self.get_param('write_pages
', False):
832 basen = '%s_%s' % (video_id, urlh.geturl())
833 trim_length = self.get_param('trim_file_name
') or 240
834 if len(basen) > trim_length:
835 h = '___
' + hashlib.md5(basen.encode('utf
-8')).hexdigest()
836 basen = basen[:trim_length - len(h)] + h
837 raw_filename = basen + '.dump
'
838 filename = sanitize_filename(raw_filename, restricted=True)
839 self.to_screen('Saving request to
' + filename)
840 # Working around MAX_PATH limitation on Windows (see
841 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
842 if compat_os_name == 'nt
':
843 absfilepath = os.path.abspath(filename)
844 if len(absfilepath) > 259:
845 filename = '\\\\?
\\' + absfilepath
846 with open(filename, 'wb
') as outf:
847 outf.write(webpage_bytes)
850 content = webpage_bytes.decode(encoding, 'replace
')
852 content = webpage_bytes.decode('utf
-8', 'replace
')
854 self.__check_blocked(content)
858 def _download_webpage(
859 self, url_or_request, video_id, note=None, errnote=None,
860 fatal=True, tries=1, timeout=5, encoding=None, data=None,
861 headers={}, query={}, expected_status=None):
863 Return the data of the page as a string.
866 url_or_request -- plain text URL as a string or
867 a compat_urllib_request.Requestobject
868 video_id -- Video/playlist/item identifier (string)
871 note -- note printed before downloading (string)
872 errnote -- note printed in case of an error (string)
873 fatal -- flag denoting whether error should be considered fatal,
874 i.e. whether it should cause ExtractionError to be raised,
875 otherwise a warning will be reported and extraction continued
876 tries -- number of tries
877 timeout -- sleep interval between tries
878 encoding -- encoding for a page content decoding, guessed automatically
879 when not explicitly specified
880 data -- POST data (bytes)
881 headers -- HTTP headers (dict)
882 query -- URL query (dict)
883 expected_status -- allows to accept failed HTTP requests (non 2xx
884 status code) by explicitly specifying a set of accepted status
885 codes. Can be any of the following entities:
886 - an integer type specifying an exact failed status code to
888 - a list or a tuple of integer types specifying a list of
889 failed status codes to accept
890 - a callable accepting an actual failed status code and
891 returning True if it should be accepted
892 Note that this argument does not affect success status codes (2xx)
893 which are always accepted.
898 while success is False:
900 res = self._download_webpage_handle(
901 url_or_request, video_id, note, errnote, fatal,
902 encoding=encoding, data=data, headers=headers, query=query,
903 expected_status=expected_status)
905 except compat_http_client.IncompleteRead as e:
907 if try_count >= tries:
909 self._sleep(timeout, video_id)
916 def _download_xml_handle(
917 self, url_or_request, video_id, note='Downloading XML
',
918 errnote='Unable to download XML
', transform_source=None,
919 fatal=True, encoding=None, data=None, headers={}, query={},
920 expected_status=None):
922 Return a tuple (xml as an compat_etree_Element, URL handle).
924 See _download_webpage docstring for arguments specification.
926 res = self._download_webpage_handle(
927 url_or_request, video_id, note, errnote, fatal=fatal,
928 encoding=encoding, data=data, headers=headers, query=query,
929 expected_status=expected_status)
932 xml_string, urlh = res
933 return self._parse_xml(
934 xml_string, video_id, transform_source=transform_source,
938 self, url_or_request, video_id,
939 note='Downloading XML
', errnote='Unable to download XML
',
940 transform_source=None, fatal=True, encoding=None,
941 data=None, headers={}, query={}, expected_status=None):
943 Return the xml as an compat_etree_Element.
945 See _download_webpage docstring for arguments specification.
947 res = self._download_xml_handle(
948 url_or_request, video_id, note=note, errnote=errnote,
949 transform_source=transform_source, fatal=fatal, encoding=encoding,
950 data=data, headers=headers, query=query,
951 expected_status=expected_status)
952 return res if res is False else res[0]
954 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
956 xml_string = transform_source(xml_string)
958 return compat_etree_fromstring(xml_string.encode('utf
-8'))
959 except compat_xml_parse_error as ve:
960 errmsg = '%s: Failed to parse XML
' % video_id
962 raise ExtractorError(errmsg, cause=ve)
964 self.report_warning(errmsg + str(ve))
966 def _download_json_handle(
967 self, url_or_request, video_id, note='Downloading JSON metadata
',
968 errnote='Unable to download JSON metadata
', transform_source=None,
969 fatal=True, encoding=None, data=None, headers={}, query={},
970 expected_status=None):
972 Return a tuple (JSON object, URL handle).
974 See _download_webpage docstring for arguments specification.
976 res = self._download_webpage_handle(
977 url_or_request, video_id, note, errnote, fatal=fatal,
978 encoding=encoding, data=data, headers=headers, query=query,
979 expected_status=expected_status)
982 json_string, urlh = res
983 return self._parse_json(
984 json_string, video_id, transform_source=transform_source,
988 self, url_or_request, video_id, note='Downloading JSON metadata
',
989 errnote='Unable to download JSON metadata
', transform_source=None,
990 fatal=True, encoding=None, data=None, headers={}, query={},
991 expected_status=None):
993 Return the JSON object as a dict.
995 See _download_webpage docstring for arguments specification.
997 res = self._download_json_handle(
998 url_or_request, video_id, note=note, errnote=errnote,
999 transform_source=transform_source, fatal=fatal, encoding=encoding,
1000 data=data, headers=headers, query=query,
1001 expected_status=expected_status)
1002 return res if res is False else res[0]
1004 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
1005 if transform_source:
1006 json_string = transform_source(json_string)
1008 return json.loads(json_string)
1009 except ValueError as ve:
1010 errmsg = '%s: Failed to parse JSON
' % video_id
1012 raise ExtractorError(errmsg, cause=ve)
1014 self.report_warning(errmsg + str(ve))
1016 def _parse_socket_response_as_json(self, data, video_id, transform_source=None, fatal=True):
1017 return self._parse_json(
1018 data[data.find('{'):data.rfind('}
') + 1],
1019 video_id, transform_source, fatal)
1021 def _download_socket_json_handle(
1022 self, url_or_request, video_id, note='Polling socket
',
1023 errnote='Unable to poll socket
', transform_source=None,
1024 fatal=True, encoding=None, data=None, headers={}, query={},
1025 expected_status=None):
1027 Return a tuple (JSON object, URL handle).
1029 See _download_webpage docstring for arguments specification.
1031 res = self._download_webpage_handle(
1032 url_or_request, video_id, note, errnote, fatal=fatal,
1033 encoding=encoding, data=data, headers=headers, query=query,
1034 expected_status=expected_status)
1038 return self._parse_socket_response_as_json(
1039 webpage, video_id, transform_source=transform_source,
1042 def _download_socket_json(
1043 self, url_or_request, video_id, note='Polling socket
',
1044 errnote='Unable to poll socket
', transform_source=None,
1045 fatal=True, encoding=None, data=None, headers={}, query={},
1046 expected_status=None):
1048 Return the JSON object as a dict.
1050 See _download_webpage docstring for arguments specification.
1052 res = self._download_socket_json_handle(
1053 url_or_request, video_id, note=note, errnote=errnote,
1054 transform_source=transform_source, fatal=fatal, encoding=encoding,
1055 data=data, headers=headers, query=query,
1056 expected_status=expected_status)
1057 return res if res is False else res[0]
1059 def report_warning(self, msg, video_id=None, *args, only_once=False, **kwargs):
1060 idstr = format_field(video_id, template='%s: ')
1061 msg = f'[{self.IE_NAME}
] {idstr}{msg}
'
1063 if f'WARNING
: {msg}
' in self._printed_messages:
1065 self._printed_messages.add(f'WARNING
: {msg}
')
1066 self._downloader.report_warning(msg, *args, **kwargs)
1068 def to_screen(self, msg, *args, **kwargs):
1069 """Print msg to screen, prefixing it with '[ie_name
]'"""
1070 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg), *args, **kwargs)
1072 def write_debug(self, msg, *args, **kwargs):
1073 self._downloader.write_debug('[%s] %s' % (self.IE_NAME, msg), *args, **kwargs)
1075 def get_param(self, name, default=None, *args, **kwargs):
1076 if self._downloader:
1077 return self._downloader.params.get(name, default, *args, **kwargs)
1080 def report_drm(self, video_id, partial=False):
1081 self.raise_no_formats('This video
is DRM protected
', expected=True, video_id=video_id)
1083 def report_extraction(self, id_or_name):
1084 """Report information extraction."""
1085 self.to_screen('%s: Extracting information
' % id_or_name)
1087 def report_download_webpage(self, video_id):
1088 """Report webpage download."""
1089 self.to_screen('%s: Downloading webpage
' % video_id)
1091 def report_age_confirmation(self):
1092 """Report attempt to confirm age."""
1093 self.to_screen('Confirming age
')
1095 def report_login(self):
1096 """Report attempt to log in."""
1097 self.to_screen('Logging
in')
1099 def raise_login_required(
1100 self, msg='This video
is only available
for registered users
',
1101 metadata_available=False, method='any
'):
1102 if metadata_available and (
1103 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1104 self.report_warning(msg)
1106 if method is not None:
1107 msg = '%s. %s' % (msg, self._LOGIN_HINTS[method])
1108 raise ExtractorError(msg, expected=True)
1110 def raise_geo_restricted(
1111 self, msg='This video
is not available
from your location due to geo restriction
',
1112 countries=None, metadata_available=False):
1113 if metadata_available and (
1114 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1115 self.report_warning(msg)
1117 raise GeoRestrictedError(msg, countries=countries)
1119 def raise_no_formats(self, msg, expected=False, video_id=None):
1121 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1122 self.report_warning(msg, video_id)
1123 elif isinstance(msg, ExtractorError):
1126 raise ExtractorError(msg, expected=expected, video_id=video_id)
1128 # Methods for following #608
1130 def url_result(url, ie=None, video_id=None, video_title=None, *, url_transparent=False, **kwargs):
1131 """Returns a URL that points to a page that should be processed"""
1133 kwargs['ie_key
'] = ie if isinstance(ie, str) else ie.ie_key()
1134 if video_id is not None:
1135 kwargs['id'] = video_id
1136 if video_title is not None:
1137 kwargs['title
'] = video_title
1140 '_type
': 'url_transparent
' if url_transparent else 'url
',
1144 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None, video_kwargs=None, **kwargs):
1145 urls = (self.url_result(self._proto_relative_url(m), ie, **(video_kwargs or {}))
1146 for m in orderedSet(map(getter, matches) if getter else matches))
1147 return self.playlist_result(urls, playlist_id, playlist_title, **kwargs)
1150 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None, *, multi_video=False, **kwargs):
1151 """Returns a playlist"""
1153 kwargs['id'] = playlist_id
1155 kwargs['title
'] = playlist_title
1156 if playlist_description is not None:
1157 kwargs['description
'] = playlist_description
1160 '_type
': 'multi_video
' if multi_video else 'playlist
',
1164 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1166 Perform a regex search on the given string, using a single or a list of
1167 patterns returning the first matching group.
1168 In case of failure return a default value or raise a WARNING or a
1169 RegexNotFoundError, depending on fatal, specifying the field name.
1171 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
1172 mobj = re.search(pattern, string, flags)
1175 mobj = re.search(p, string, flags)
1179 _name = self._downloader._format_err(name, self._downloader.Styles.EMPHASIS)
1183 # return the first matching group
1184 return next(g for g in mobj.groups() if g is not None)
1185 elif isinstance(group, (list, tuple)):
1186 return tuple(mobj.group(g) for g in group)
1188 return mobj.group(group)
1189 elif default is not NO_DEFAULT:
1192 raise RegexNotFoundError('Unable to extract
%s' % _name)
1194 self.report_warning('unable to extract
%s' % _name + bug_reports_message())
1197 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1199 Like _search_regex, but strips HTML tags and unescapes entities.
1201 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1203 return clean_html(res).strip()
1207 def _get_netrc_login_info(self, netrc_machine=None):
1210 netrc_machine = netrc_machine or self._NETRC_MACHINE
1212 if self.get_param('usenetrc
', False):
1214 netrc_file = compat_expanduser(self.get_param('netrc_location
') or '~
')
1215 if os.path.isdir(netrc_file):
1216 netrc_file = os.path.join(netrc_file, '.netrc
')
1217 info = netrc.netrc(file=netrc_file).authenticators(netrc_machine)
1218 if info is not None:
1222 raise netrc.NetrcParseError(
1223 'No authenticators
for %s' % netrc_machine)
1224 except (IOError, netrc.NetrcParseError) as err:
1225 self.report_warning(
1226 'parsing
.netrc
: %s' % error_to_compat_str(err))
1228 return username, password
1230 def _get_login_info(self, username_option='username
', password_option='password
', netrc_machine=None):
1232 Get the login info as (username, password)
1233 First look for the manually specified credentials using username_option
1234 and password_option as keys in params dictionary. If no such credentials
1235 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1237 If there's no info available
, return (None, None)
1240 # Attempt to use provided username and password or .netrc data
1241 username = self.get_param(username_option)
1242 if username is not None:
1243 password = self.get_param(password_option)
1245 username, password = self._get_netrc_login_info(netrc_machine)
1247 return username, password
1249 def _get_tfa_info(self, note='two-factor verification code'):
1251 Get the two
-factor authentication info
1252 TODO
- asking the user will be required
for sms
/phone verify
1253 currently just uses the command line option
1254 If there
's no info available, return None
1257 tfa = self.get_param('twofactor
')
1261 return compat_getpass('Type
%s and press
[Return
]: ' % note)
1263 # Helper functions for extracting OpenGraph info
1265 def _og_regexes(prop):
1266 content_re = r'content
=(?
:"([^"]+?
)"|\'([^\']+?)\'|\s*([^\s"\'=<>`
]+?
))'
1267 property_re = (r'(?
:name|
property)=(?
:\'og
[:-]%(prop)s\'|
"og[:-]%(prop)s"|\s
*og
[:-]%(prop)s\b)'
1268 % {'prop': re.escape(prop)})
1269 template = r'<meta
[^
>]+?
%s[^
>]+?
%s'
1271 template % (property_re, content_re),
1272 template % (content_re, property_re),
1276 def _meta_regex(prop):
1277 return r'''(?isx)<meta
1278 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1279 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1281 def _og_search_property(self, prop, html, name=None, **kargs):
1282 prop = variadic(prop)
1284 name = 'OpenGraph
%s' % prop[0]
1287 og_regexes.extend(self._og_regexes(p))
1288 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1291 return unescapeHTML(escaped)
1293 def _og_search_thumbnail(self, html, **kargs):
1294 return self._og_search_property('image
', html, 'thumbnail URL
', fatal=False, **kargs)
1296 def _og_search_description(self, html, **kargs):
1297 return self._og_search_property('description
', html, fatal=False, **kargs)
1299 def _og_search_title(self, html, **kargs):
1300 kargs.setdefault('fatal
', False)
1301 return self._og_search_property('title
', html, **kargs)
1303 def _og_search_video_url(self, html, name='video url
', secure=True, **kargs):
1304 regexes = self._og_regexes('video
') + self._og_regexes('video
:url
')
1306 regexes = self._og_regexes('video
:secure_url
') + regexes
1307 return self._html_search_regex(regexes, html, name, **kargs)
1309 def _og_search_url(self, html, **kargs):
1310 return self._og_search_property('url
', html, **kargs)
1312 def _html_extract_title(self, html, name, **kwargs):
1313 return self._html_search_regex(
1314 r'(?s
)<title
>(.*?
)</title
>', html, name, **kwargs)
1316 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1317 name = variadic(name)
1318 if display_name is None:
1319 display_name = name[0]
1320 return self._html_search_regex(
1321 [self._meta_regex(n) for n in name],
1322 html, display_name, fatal=fatal, group='content
', **kwargs)
1324 def _dc_search_uploader(self, html):
1325 return self._html_search_meta('dc
.creator
', html, 'uploader
')
1327 def _rta_search(self, html):
1328 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1329 if re.search(r'(?ix
)<meta\s
+name
="rating"\s
+'
1330 r' content
="RTA-5042-1996-1400-1577-RTA"',
1335 def _media_rating_search(self, html):
1336 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1337 rating = self._html_search_meta('rating
', html)
1349 return RATING_TABLE.get(rating.lower())
1351 def _family_friendly_search(self, html):
1352 # See http://schema.org/VideoObject
1353 family_friendly = self._html_search_meta(
1354 'isFamilyFriendly
', html, default=None)
1356 if not family_friendly:
1365 return RATING_TABLE.get(family_friendly.lower())
1367 def _twitter_search_player(self, html):
1368 return self._html_search_meta('twitter
:player
', html,
1369 'twitter card player
')
1371 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
1372 json_ld_list = list(re.finditer(JSON_LD_RE, html))
1373 default = kwargs.get('default
', NO_DEFAULT)
1374 # JSON-LD may be malformed and thus `fatal` should be respected.
1375 # At the same time `default` may be passed that assumes `fatal=False`
1376 # for _search_regex. Let's simulate the same behavior here
as well
.
1377 fatal
= kwargs
.get('fatal', True) if default
is NO_DEFAULT
else False
1379 for mobj
in json_ld_list
:
1380 json_ld_item
= self
._parse
_json
(
1381 mobj
.group('json_ld'), video_id
, fatal
=fatal
)
1382 if not json_ld_item
:
1384 if isinstance(json_ld_item
, dict):
1385 json_ld
.append(json_ld_item
)
1386 elif isinstance(json_ld_item
, (list, tuple)):
1387 json_ld
.extend(json_ld_item
)
1389 json_ld
= self
._json
_ld
(json_ld
, video_id
, fatal
=fatal
, expected_type
=expected_type
)
1392 if default
is not NO_DEFAULT
:
1395 raise RegexNotFoundError('Unable to extract JSON-LD')
1397 self
.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1400 def _json_ld(self
, json_ld
, video_id
, fatal
=True, expected_type
=None):
1401 if isinstance(json_ld
, compat_str
):
1402 json_ld
= self
._parse
_json
(json_ld
, video_id
, fatal
=fatal
)
1406 if not isinstance(json_ld
, (list, tuple, dict)):
1408 if isinstance(json_ld
, dict):
1411 INTERACTION_TYPE_MAP
= {
1412 'CommentAction': 'comment',
1413 'AgreeAction': 'like',
1414 'DisagreeAction': 'dislike',
1415 'LikeAction': 'like',
1416 'DislikeAction': 'dislike',
1417 'ListenAction': 'view',
1418 'WatchAction': 'view',
1419 'ViewAction': 'view',
1422 def extract_interaction_type(e
):
1423 interaction_type
= e
.get('interactionType')
1424 if isinstance(interaction_type
, dict):
1425 interaction_type
= interaction_type
.get('@type')
1426 return str_or_none(interaction_type
)
1428 def extract_interaction_statistic(e
):
1429 interaction_statistic
= e
.get('interactionStatistic')
1430 if isinstance(interaction_statistic
, dict):
1431 interaction_statistic
= [interaction_statistic
]
1432 if not isinstance(interaction_statistic
, list):
1434 for is_e
in interaction_statistic
:
1435 if not isinstance(is_e
, dict):
1437 if is_e
.get('@type') != 'InteractionCounter':
1439 interaction_type
= extract_interaction_type(is_e
)
1440 if not interaction_type
:
1442 # For interaction count some sites provide string instead of
1443 # an integer (as per spec) with non digit characters (e.g. ",")
1444 # so extracting count with more relaxed str_to_int
1445 interaction_count
= str_to_int(is_e
.get('userInteractionCount'))
1446 if interaction_count
is None:
1448 count_kind
= INTERACTION_TYPE_MAP
.get(interaction_type
.split('/')[-1])
1451 count_key
= '%s_count' % count_kind
1452 if info
.get(count_key
) is not None:
1454 info
[count_key
] = interaction_count
1456 def extract_chapter_information(e
):
1458 'title': part
.get('name'),
1459 'start_time': part
.get('startOffset'),
1460 'end_time': part
.get('endOffset'),
1461 } for part
in variadic(e
.get('hasPart') or []) if part
.get('@type') == 'Clip']
1462 for idx
, (last_c
, current_c
, next_c
) in enumerate(zip(
1463 [{'end_time': 0}
] + chapters
, chapters
, chapters
[1:])):
1464 current_c
['end_time'] = current_c
['end_time'] or next_c
['start_time']
1465 current_c
['start_time'] = current_c
['start_time'] or last_c
['end_time']
1466 if None in current_c
.values():
1467 self
.report_warning(f
'Chapter {idx} contains broken data. Not extracting chapters')
1470 chapters
[-1]['end_time'] = chapters
[-1]['end_time'] or info
['duration']
1471 info
['chapters'] = chapters
1473 def extract_video_object(e
):
1474 assert e
['@type'] == 'VideoObject'
1475 author
= e
.get('author')
1477 'url': url_or_none(e
.get('contentUrl')),
1478 'title': unescapeHTML(e
.get('name')),
1479 'description': unescapeHTML(e
.get('description')),
1480 'thumbnails': [{'url': url_or_none(url)}
1481 for url
in variadic(traverse_obj(e
, 'thumbnailUrl', 'thumbnailURL'))],
1482 'duration': parse_duration(e
.get('duration')),
1483 'timestamp': unified_timestamp(e
.get('uploadDate')),
1484 # author can be an instance of 'Organization' or 'Person' types.
1485 # both types can have 'name' property(inherited from 'Thing' type). [1]
1486 # however some websites are using 'Text' type instead.
1487 # 1. https://schema.org/VideoObject
1488 'uploader': author
.get('name') if isinstance(author
, dict) else author
if isinstance(author
, compat_str
) else None,
1489 'filesize': float_or_none(e
.get('contentSize')),
1490 'tbr': int_or_none(e
.get('bitrate')),
1491 'width': int_or_none(e
.get('width')),
1492 'height': int_or_none(e
.get('height')),
1493 'view_count': int_or_none(e
.get('interactionCount')),
1495 extract_interaction_statistic(e
)
1496 extract_chapter_information(e
)
1498 def traverse_json_ld(json_ld
, at_top_level
=True):
1500 if at_top_level
and '@context' not in e
:
1502 if at_top_level
and set(e
.keys()) == {'@context', '@graph'}
:
1503 traverse_json_ld(variadic(e
['@graph'], allowed_types
=(dict,)), at_top_level
=False)
1505 item_type
= e
.get('@type')
1506 if expected_type
is not None and expected_type
!= item_type
:
1508 rating
= traverse_obj(e
, ('aggregateRating', 'ratingValue'), expected_type
=float_or_none
)
1509 if rating
is not None:
1510 info
['average_rating'] = rating
1511 if item_type
in ('TVEpisode', 'Episode'):
1512 episode_name
= unescapeHTML(e
.get('name'))
1514 'episode': episode_name
,
1515 'episode_number': int_or_none(e
.get('episodeNumber')),
1516 'description': unescapeHTML(e
.get('description')),
1518 if not info
.get('title') and episode_name
:
1519 info
['title'] = episode_name
1520 part_of_season
= e
.get('partOfSeason')
1521 if isinstance(part_of_season
, dict) and part_of_season
.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1523 'season': unescapeHTML(part_of_season
.get('name')),
1524 'season_number': int_or_none(part_of_season
.get('seasonNumber')),
1526 part_of_series
= e
.get('partOfSeries') or e
.get('partOfTVSeries')
1527 if isinstance(part_of_series
, dict) and part_of_series
.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1528 info
['series'] = unescapeHTML(part_of_series
.get('name'))
1529 elif item_type
== 'Movie':
1531 'title': unescapeHTML(e
.get('name')),
1532 'description': unescapeHTML(e
.get('description')),
1533 'duration': parse_duration(e
.get('duration')),
1534 'timestamp': unified_timestamp(e
.get('dateCreated')),
1536 elif item_type
in ('Article', 'NewsArticle'):
1538 'timestamp': parse_iso8601(e
.get('datePublished')),
1539 'title': unescapeHTML(e
.get('headline')),
1540 'description': unescapeHTML(e
.get('articleBody') or e
.get('description')),
1542 if traverse_obj(e
, ('video', 0, '@type')) == 'VideoObject':
1543 extract_video_object(e
['video'][0])
1544 elif item_type
== 'VideoObject':
1545 extract_video_object(e
)
1546 if expected_type
is None:
1550 video
= e
.get('video')
1551 if isinstance(video
, dict) and video
.get('@type') == 'VideoObject':
1552 extract_video_object(video
)
1553 if expected_type
is None:
1557 traverse_json_ld(json_ld
)
1559 return dict((k
, v
) for k
, v
in info
.items() if v
is not None)
1561 def _search_nextjs_data(self
, webpage
, video_id
, *, transform_source
=None, fatal
=True, **kw
):
1562 return self
._parse
_json
(
1564 r
'(?s)<script[^>]+id=[\'"]__NEXT_DATA__[\'"][^
>]*>([^
<]+)</script
>',
1565 webpage, 'next
.js data
', fatal=fatal, **kw),
1566 video_id, transform_source=transform_source, fatal=fatal)
1568 def _search_nuxt_data(self, webpage, video_id, context_name='__NUXT__
'):
1569 ''' Parses Nuxt.js metadata. This works as long as the function __NUXT__ invokes is a pure function. '''
1570 # not all website do this, but it can be changed
1571 # https://stackoverflow.com/questions/67463109/how-to-change-or-hide-nuxt-and-nuxt-keyword-in-page-source
1572 rectx = re.escape(context_name)
1573 js, arg_keys, arg_vals = self._search_regex(
1574 (r'<script
>window\
.%s=\
(function\
((?P
<arg_keys
>.*?
)\
)\{return\s(?P<js>\{.*?\}
)\
}\
((?P
<arg_vals
>.+?
)\
)\
);?
</script
>' % rectx,
1575 r'%s\
(.*?\
(function\
((?P
<arg_keys
>.*?
)\
)\{return\s(?P<js>\{.*?\}
)\
}\
((?P
<arg_vals
>.*?
)\
)' % rectx),
1576 webpage, context_name, group=['js
', 'arg_keys
', 'arg_vals
'])
1578 args = dict(zip(arg_keys.split(','), arg_vals.split(',')))
1580 for key, val in args.items():
1581 if val in ('undefined
', 'void
0'):
1584 return self._parse_json(js_to_json(js, args), video_id)['data
'][0]
1587 def _hidden_inputs(html):
1588 html = re.sub(r'<!--(?
:(?
!<!--).)*-->', '', html)
1590 for input in re.findall(r'(?i
)(<input[^
>]+>)', html):
1591 attrs = extract_attributes(input)
1594 if attrs.get('type') not in ('hidden
', 'submit
'):
1596 name = attrs.get('name
') or attrs.get('id')
1597 value = attrs.get('value
')
1598 if name and value is not None:
1599 hidden_inputs[name] = value
1600 return hidden_inputs
1602 def _form_hidden_inputs(self, form_id, html):
1603 form = self._search_regex(
1604 r'(?
is)<form
[^
>]+?
id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1605 html, '%s form' % form_id, group='form')
1606 return self._hidden_inputs(form)
1609 regex = r' *((?P<reverse>\+)?(?P<field>[a-zA-Z0-9_]+)((?P<separator>[~:])(?P<limit>.*?))?)? *$'
1611 default = ('hidden', 'aud_or_vid', 'hasvid', 'ie_pref', 'lang', 'quality',
1612 'res', 'fps', 'hdr:12', 'codec:vp9.2', 'size', 'br', 'asr',
1613 'proto', 'ext', 'hasaud', 'source', 'id') # These must not be aliases
1614 ytdl_default = ('hasaud', 'lang', 'quality', 'tbr', 'filesize', 'vbr',
1615 'height', 'width', 'proto', 'vext', 'abr', 'aext',
1616 'fps', 'fs_approx', 'source', 'id')
1619 'vcodec': {'type': 'ordered', 'regex': True,
1620 'order': ['av0?1', 'vp0?9.2', 'vp0?9', '[hx]265|he?vc?', '[hx]264|avc', 'vp0?8', 'mp4v|h263', 'theora', '', None, 'none']},
1621 'acodec': {'type': 'ordered', 'regex': True,
1622 'order': ['[af]lac', 'wav|aiff', 'opus', 'vorbis|ogg', 'aac', 'mp?4a?', 'mp3', 'e-?a?c-?3', 'ac-?3', 'dts', '', None, 'none']},
1623 'hdr': {'type': 'ordered', 'regex': True, 'field': 'dynamic_range',
1624 'order': ['dv', '(hdr)?12', r'(hdr)?10\+', '(hdr)?10', 'hlg', '', 'sdr', None]},
1625 'proto': {'type': 'ordered', 'regex': True, 'field': 'protocol',
1626 'order': ['(ht|f)tps', '(ht|f)tp$', 'm3u8.*', '.*dash', 'websocket_frag', 'rtmpe?', '', 'mms|rtsp', 'ws|websocket', 'f4']},
1627 'vext': {'type': 'ordered', 'field': 'video_ext',
1628 'order': ('mp4', 'webm', 'flv', '', 'none'),
1629 'order_free': ('webm', 'mp4', 'flv', '', 'none')},
1630 'aext': {'type': 'ordered', 'field': 'audio_ext',
1631 'order': ('m4a', 'aac', 'mp3', 'ogg', 'opus', 'webm', '', 'none'),
1632 'order_free': ('opus', 'ogg', 'webm', 'm4a', 'mp3', 'aac', '', 'none')},
1633 'hidden': {'visible': False, 'forced': True, 'type': 'extractor', 'max': -1000},
1634 'aud_or_vid': {'visible': False, 'forced': True, 'type': 'multiple',
1635 'field': ('vcodec', 'acodec'),
1636 'function': lambda it: int(any(v != 'none' for v in it))},
1637 'ie_pref': {'priority': True, 'type': 'extractor'},
1638 'hasvid': {'priority': True, 'field': 'vcodec', 'type': 'boolean', 'not_in_list': ('none',)},
1639 'hasaud': {'field': 'acodec', 'type': 'boolean', 'not_in_list': ('none',)},
1640 'lang': {'convert': 'float', 'field': 'language_preference', 'default': -1},
1641 'quality': {'convert': 'float', 'default': -1},
1642 'filesize': {'convert': 'bytes'},
1643 'fs_approx': {'convert': 'bytes', 'field': 'filesize_approx'},
1644 'id': {'convert': 'string', 'field': 'format_id'},
1645 'height': {'convert': 'float_none'},
1646 'width': {'convert': 'float_none'},
1647 'fps': {'convert': 'float_none'},
1648 'tbr': {'convert': 'float_none'},
1649 'vbr': {'convert': 'float_none'},
1650 'abr': {'convert': 'float_none'},
1651 'asr': {'convert': 'float_none'},
1652 'source': {'convert': 'float', 'field': 'source_preference', 'default': -1},
1654 'codec': {'type': 'combined', 'field': ('vcodec', 'acodec')},
1655 'br': {'type': 'combined', 'field': ('tbr', 'vbr', 'abr'), 'same_limit': True},
1656 'size': {'type': 'combined', 'same_limit': True, 'field': ('filesize', 'fs_approx')},
1657 'ext': {'type': 'combined', 'field': ('vext', 'aext')},
1658 'res': {'type': 'multiple', 'field': ('height', 'width'),
1659 'function': lambda it: (lambda l: min(l) if l else 0)(tuple(filter(None, it)))},
1661 # For compatibility with youtube-dl
1662 'format_id': {'type': 'alias', 'field': 'id'},
1663 'preference': {'type': 'alias', 'field': 'ie_pref'},
1664 'language_preference': {'type': 'alias', 'field': 'lang'},
1665 'source_preference': {'type': 'alias', 'field': 'source'},
1666 'protocol': {'type': 'alias', 'field': 'proto'},
1667 'filesize_approx': {'type': 'alias', 'field': 'fs_approx'},
1670 'dimension': {'type': 'alias', 'field': 'res', 'deprecated': True},
1671 'resolution': {'type': 'alias', 'field': 'res', 'deprecated': True},
1672 'extension': {'type': 'alias', 'field': 'ext', 'deprecated': True},
1673 'bitrate': {'type': 'alias', 'field': 'br', 'deprecated': True},
1674 'total_bitrate': {'type': 'alias', 'field': 'tbr', 'deprecated': True},
1675 'video_bitrate': {'type': 'alias', 'field': 'vbr', 'deprecated': True},
1676 'audio_bitrate': {'type': 'alias', 'field': 'abr', 'deprecated': True},
1677 'framerate': {'type': 'alias', 'field': 'fps', 'deprecated': True},
1678 'filesize_estimate': {'type': 'alias', 'field': 'size', 'deprecated': True},
1679 'samplerate': {'type': 'alias', 'field': 'asr', 'deprecated': True},
1680 'video_ext': {'type': 'alias', 'field': 'vext', 'deprecated': True},
1681 'audio_ext': {'type': 'alias', 'field': 'aext', 'deprecated': True},
1682 'video_codec': {'type': 'alias', 'field': 'vcodec', 'deprecated': True},
1683 'audio_codec': {'type': 'alias', 'field': 'acodec', 'deprecated': True},
1684 'video': {'type': 'alias', 'field': 'hasvid', 'deprecated': True},
1685 'has_video': {'type': 'alias', 'field': 'hasvid', 'deprecated': True},
1686 'audio': {'type': 'alias', 'field': 'hasaud', 'deprecated': True},
1687 'has_audio': {'type': 'alias', 'field': 'hasaud', 'deprecated': True},
1688 'extractor': {'type': 'alias', 'field': 'ie_pref', 'deprecated': True},
1689 'extractor_preference': {'type': 'alias', 'field': 'ie_pref', 'deprecated': True},
1692 def __init__(self, ie, field_preference):
1694 self.ydl = ie._downloader
1695 self.evaluate_params(self.ydl.params, field_preference)
1696 if ie.get_param('verbose'):
1697 self.print_verbose_info(self.ydl.write_debug)
1699 def _get_field_setting(self, field, key):
1700 if field not in self.settings:
1701 if key in ('forced', 'priority'):
1703 self.ydl.deprecation_warning(
1704 f'Using arbitrary fields ({field}) for format sorting is deprecated '
1705 'and may be removed in a future version')
1706 self.settings[field] = {}
1707 propObj = self.settings[field]
1708 if key not in propObj:
1709 type = propObj.get('type')
1711 default = 'preference' if type == 'extractor' else (field,) if type in ('combined', 'multiple') else field
1712 elif key == 'convert':
1713 default = 'order' if type == 'ordered' else 'float_string' if field else 'ignore'
1715 default = {'type': 'field', 'visible': True, 'order': [], 'not_in_list': (None,)}.get(key, None)
1716 propObj[key] = default
1719 def _resolve_field_value(self, field, value, convertNone=False):
1724 value = value.lower()
1725 conversion = self._get_field_setting(field, 'convert')
1726 if conversion == 'ignore':
1728 if conversion == 'string':
1730 elif conversion == 'float_none':
1731 return float_or_none(value)
1732 elif conversion == 'bytes':
1733 return FileDownloader.parse_bytes(value)
1734 elif conversion == 'order':
1735 order_list = (self._use_free_order and self._get_field_setting(field, 'order_free')) or self._get_field_setting(field, 'order')
1736 use_regex = self._get_field_setting(field, 'regex')
1737 list_length = len(order_list)
1738 empty_pos = order_list.index('') if '' in order_list else list_length + 1
1739 if use_regex and value is not None:
1740 for i, regex in enumerate(order_list):
1741 if regex and re.match(regex, value):
1742 return list_length - i
1743 return list_length - empty_pos # not in list
1744 else: # not regex or value = None
1745 return list_length - (order_list.index(value) if value in order_list else empty_pos)
1747 if value.isnumeric():
1750 self.settings[field]['convert'] = 'string'
1753 def evaluate_params(self, params, sort_extractor):
1754 self._use_free_order = params.get('prefer_free_formats', False)
1755 self._sort_user = params.get('format_sort', [])
1756 self._sort_extractor = sort_extractor
1758 def add_item(field, reverse, closest, limit_text):
1759 field = field.lower()
1760 if field in self._order:
1762 self._order.append(field)
1763 limit = self._resolve_field_value(field, limit_text)
1766 'closest': False if limit is None else closest,
1767 'limit_text': limit_text,
1769 if field in self.settings:
1770 self.settings[field].update(data)
1772 self.settings[field] = data
1775 tuple(field for field in self.default if self._get_field_setting(field, 'forced'))
1776 + (tuple() if params.get('format_sort_force', False)
1777 else tuple(field for field in self.default if self._get_field_setting(field, 'priority')))
1778 + tuple(self._sort_user) + tuple(sort_extractor) + self.default)
1780 for item in sort_list:
1781 match = re.match(self.regex, item)
1783 raise ExtractorError('Invalid format sort string "%s" given by extractor' % item)
1784 field = match.group('field')
1787 if self._get_field_setting(field, 'type') == 'alias':
1788 alias, field = field, self._get_field_setting(field, 'field')
1789 if self._get_field_setting(alias, 'deprecated'):
1790 self.ydl.deprecation_warning(
1791 f'Format sorting alias {alias} is deprecated '
1792 f'and may be removed in a future version. Please use {field} instead')
1793 reverse = match.group('reverse') is not None
1794 closest = match.group('separator') == '~'
1795 limit_text = match.group('limit')
1797 has_limit = limit_text is not None
1798 has_multiple_fields = self._get_field_setting(field, 'type') == 'combined'
1799 has_multiple_limits = has_limit and has_multiple_fields and not self._get_field_setting(field, 'same_limit')
1801 fields = self._get_field_setting(field, 'field') if has_multiple_fields else (field,)
1802 limits = limit_text.split(':') if has_multiple_limits else (limit_text,) if has_limit else tuple()
1803 limit_count = len(limits)
1804 for (i, f) in enumerate(fields):
1805 add_item(f, reverse, closest,
1806 limits[i] if i < limit_count
1807 else limits[0] if has_limit and not has_multiple_limits
1810 def print_verbose_info(self, write_debug):
1812 write_debug('Sort order given by user: %s' % ', '.join(self._sort_user))
1813 if self._sort_extractor:
1814 write_debug('Sort order given by extractor: %s' % ', '.join(self._sort_extractor))
1815 write_debug('Formats sorted by: %s' % ', '.join(['%s%s%s' % (
1816 '+' if self._get_field_setting(field, 'reverse') else '', field,
1817 '%s%s(%s)' % ('~' if self._get_field_setting(field, 'closest') else ':',
1818 self._get_field_setting(field, 'limit_text'),
1819 self._get_field_setting(field, 'limit'))
1820 if self._get_field_setting(field, 'limit_text') is not None else '')
1821 for field in self._order if self._get_field_setting(field, 'visible')]))
1823 def _calculate_field_preference_from_value(self, format, field, type, value):
1824 reverse = self._get_field_setting(field, 'reverse')
1825 closest = self._get_field_setting(field, 'closest')
1826 limit = self._get_field_setting(field, 'limit')
1828 if type == 'extractor':
1829 maximum = self._get_field_setting(field, 'max')
1830 if value is None or (maximum is not None and value >= maximum):
1832 elif type == 'boolean':
1833 in_list = self._get_field_setting(field, 'in_list')
1834 not_in_list = self._get_field_setting(field, 'not_in_list')
1835 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
1836 elif type == 'ordered':
1837 value = self._resolve_field_value(field, value, True)
1839 # try to convert to number
1840 val_num = float_or_none(value, default=self._get_field_setting(field, 'default'))
1841 is_num = self._get_field_setting(field, 'convert') != 'string' and val_num is not None
1845 return ((-10, 0) if value is None
1846 else (1, value, 0) if not is_num # if a field has mixed strings and numbers, strings are sorted higher
1847 else (0, -abs(value - limit), value - limit if reverse else limit - value) if closest
1848 else (0, value, 0) if not reverse and (limit is None or value <= limit)
1849 else (0, -value, 0) if limit is None or (reverse and value == limit) or value > limit
1850 else (-1, value, 0))
1852 def _calculate_field_preference(self, format, field):
1853 type = self._get_field_setting(field, 'type') # extractor, boolean, ordered, field, multiple
1854 get_value = lambda f: format.get(self._get_field_setting(f, 'field'))
1855 if type == 'multiple':
1856 type = 'field' # Only 'field' is allowed in multiple for now
1857 actual_fields = self._get_field_setting(field, 'field')
1859 value = self._get_field_setting(field, 'function')(get_value(f) for f in actual_fields)
1861 value = get_value(field)
1862 return self._calculate_field_preference_from_value(format, field, type, value)
1864 def calculate_preference(self, format):
1865 # Determine missing protocol
1866 if not format.get('protocol'):
1867 format['protocol'] = determine_protocol(format)
1869 # Determine missing ext
1870 if not format.get('ext') and 'url' in format:
1871 format['ext'] = determine_ext(format['url'])
1872 if format.get('vcodec') == 'none':
1873 format['audio_ext'] = format['ext'] if format.get('acodec') != 'none' else 'none'
1874 format['video_ext'] = 'none'
1876 format['video_ext'] = format['ext']
1877 format['audio_ext'] = 'none'
1878 # if format.get('preference') is None and format.get('ext') in ('f4f', 'f4m'): # Not supported?
1879 # format['preference'] = -1000
1881 # Determine missing bitrates
1882 if format.get('tbr') is None:
1883 if format.get('vbr') is not None and format.get('abr') is not None:
1884 format['tbr'] = format.get('vbr', 0) + format.get('abr', 0)
1886 if format.get('vcodec') != 'none' and format.get('vbr') is None:
1887 format['vbr'] = format.get('tbr') - format.get('abr', 0)
1888 if format.get('acodec') != 'none' and format.get('abr') is None:
1889 format['abr'] = format.get('tbr') - format.get('vbr', 0)
1891 return tuple(self._calculate_field_preference(format, field) for field in self._order)
1893 def _sort_formats(self, formats, field_preference=[]):
1896 format_sort = self.FormatSort(self, field_preference)
1897 formats.sort(key=lambda f: format_sort.calculate_preference(f))
1899 def _check_formats(self, formats, video_id):
1901 formats[:] = filter(
1902 lambda f: self._is_valid_url(
1904 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1908 def _remove_duplicate_formats(formats):
1912 if f['url'] not in format_urls:
1913 format_urls.add(f['url'])
1914 unique_formats.append(f)
1915 formats[:] = unique_formats
1917 def _is_valid_url(self, url, video_id, item='video', headers={}):
1918 url = self._proto_relative_url(url, scheme='http:')
1919 # For now assume non HTTP(S) URLs always valid
1920 if not (url.startswith('http://') or url.startswith('https://')):
1923 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1925 except ExtractorError as e:
1927 '%s: %s URL is invalid, skipping: %s'
1928 % (video_id, item, error_to_compat_str(e.cause)))
1931 def http_scheme(self):
1932 """ Either "http
:" or "https
:", depending on the user's preferences """
1935 if self.get_param('prefer_insecure', False)
1938 def _proto_relative_url(self, url, scheme=None):
1941 if url.startswith('//'):
1943 scheme = self.http_scheme()
1948 def _sleep(self, timeout, video_id, msg_template=None):
1949 if msg_template is None:
1950 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1951 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1955 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1956 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1957 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
1958 manifest = self._download_xml(
1959 manifest_url, video_id, 'Downloading f4m manifest',
1960 'Unable to download f4m manifest',
1961 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1962 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
1963 transform_source=transform_source,
1964 fatal=fatal, data=data, headers=headers, query=query)
1966 if manifest is False:
1969 return self._parse_f4m_formats(
1970 manifest, manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1971 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1973 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1974 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1975 fatal=True, m3u8_id=None):
1976 if not isinstance(manifest, compat_etree_Element) and not fatal:
1979 # currently yt-dlp cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1980 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1981 if akamai_pv is not None and ';' in akamai_pv.text:
1982 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1983 if playerVerificationChallenge.strip() != '':
1987 manifest_version = '1.0'
1988 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1990 manifest_version = '2.0'
1991 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1992 # Remove unsupported DRM protected media from final formats
1993 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
1994 media_nodes = remove_encrypted_media(media_nodes)
1998 manifest_base_url = get_base_url(manifest)
2000 bootstrap_info = xpath_element(
2001 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
2002 'bootstrap info', default=None)
2005 mime_type = xpath_text(
2006 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
2007 'base URL', default=None)
2008 if mime_type and mime_type.startswith('audio/'):
2011 for i, media_el in enumerate(media_nodes):
2012 tbr = int_or_none(media_el.attrib.get('bitrate'))
2013 width = int_or_none(media_el.attrib.get('width'))
2014 height = int_or_none(media_el.attrib.get('height'))
2015 format_id = join_nonempty(f4m_id, tbr or i)
2016 # If <bootstrapInfo> is present, the specified f4m is a
2017 # stream-level manifest, and only set-level manifests may refer to
2018 # external resources. See section 11.4 and section 4 of F4M spec
2019 if bootstrap_info is None:
2021 # @href is introduced in 2.0, see section 11.6 of F4M spec
2022 if manifest_version == '2.0':
2023 media_url = media_el.attrib.get('href')
2024 if media_url is None:
2025 media_url = media_el.attrib.get('url')
2029 media_url if media_url.startswith('http://') or media_url.startswith('https://')
2030 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
2031 # If media_url is itself a f4m manifest do the recursive extraction
2032 # since bitrates in parent manifest (this one) and media_url manifest
2033 # may differ leading to inability to resolve the format by requested
2034 # bitrate in f4m downloader
2035 ext = determine_ext(manifest_url)
2037 f4m_formats = self._extract_f4m_formats(
2038 manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
2039 transform_source=transform_source, fatal=fatal)
2040 # Sometimes stream-level manifest contains single media entry that
2041 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
2042 # At the same time parent's media entry in set-level manifest may
2043 # contain it. We will copy it from parent in such cases.
2044 if len(f4m_formats) == 1:
2047 'tbr': f.get('tbr') or tbr,
2048 'width': f.get('width') or width,
2049 'height': f.get('height') or height,
2050 'format_id': f.get('format_id') if not tbr else format_id,
2053 formats.extend(f4m_formats)
2056 formats.extend(self._extract_m3u8_formats(
2057 manifest_url, video_id, 'mp4', preference=preference,
2058 quality=quality, m3u8_id=m3u8_id, fatal=fatal))
2061 'format_id': format_id,
2062 'url': manifest_url,
2063 'manifest_url': manifest_url,
2064 'ext': 'flv' if bootstrap_info is not None else None,
2070 'preference': preference,
2075 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, quality=None, m3u8_id=None):
2077 'format_id': join_nonempty(m3u8_id, 'meta'),
2081 'preference': preference - 100 if preference else -100,
2083 'resolution': 'multiple',
2084 'format_note': 'Quality selection URL',
2087 def _report_ignoring_subs(self, name):
2088 self.report_warning(bug_reports_message(
2089 f'Ignoring subtitle tracks found in the {name} manifest; '
2090 'if any subtitle tracks are missing,'
2093 def _extract_m3u8_formats(self, *args, **kwargs):
2094 fmts, subs = self._extract_m3u8_formats_and_subtitles(*args, **kwargs)
2096 self._report_ignoring_subs('HLS')
2099 def _extract_m3u8_formats_and_subtitles(
2100 self, m3u8_url, video_id, ext=None, entry_protocol='m3u8_native',
2101 preference=None, quality=None, m3u8_id=None, note=None,
2102 errnote=None, fatal=True, live=False, data=None, headers={},
2105 res = self._download_webpage_handle(
2107 note='Downloading m3u8 information' if note is None else note,
2108 errnote='Failed to download m3u8 information' if errnote is None else errnote,
2109 fatal=fatal, data=data, headers=headers, query=query)
2114 m3u8_doc, urlh = res
2115 m3u8_url = urlh.geturl()
2117 return self._parse_m3u8_formats_and_subtitles(
2118 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
2119 preference=preference, quality=quality, m3u8_id=m3u8_id,
2120 note=note, errnote=errnote, fatal=fatal, live=live, data=data,
2121 headers=headers, query=query, video_id=video_id)
2123 def _parse_m3u8_formats_and_subtitles(
2124 self, m3u8_doc, m3u8_url=None, ext=None, entry_protocol='m3u8_native',
2125 preference=None, quality=None, m3u8_id=None, live=False, note=None,
2126 errnote=None, fatal=True, data=None, headers={}, query={},
2128 formats, subtitles = [], {}
2130 has_drm = re.search('|'.join([
2131 r'#EXT-X-FAXS-CM:', # Adobe Flash Access
2132 r'#EXT-X-(?:SESSION-)?KEY:.*?URI="skd
://', # Apple FairPlay
2135 def format_url(url):
2136 return url if re.match(r'^https?
://', url) else compat_urlparse.urljoin(m3u8_url, url)
2138 if self.get_param('hls_split_discontinuity
', False):
2139 def _extract_m3u8_playlist_indices(manifest_url=None, m3u8_doc=None):
2141 if not manifest_url:
2143 m3u8_doc = self._download_webpage(
2144 manifest_url, video_id, fatal=fatal, data=data, headers=headers,
2145 note=False, errnote='Failed to download m3u8 playlist information
')
2146 if m3u8_doc is False:
2148 return range(1 + sum(line.startswith('#EXT-X-DISCONTINUITY') for line in m3u8_doc.splitlines()))
2151 def _extract_m3u8_playlist_indices(*args
, **kwargs
):
2155 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
2156 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
2157 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
2159 # We should try extracting formats only from master playlists [1, 4.3.4],
2160 # i.e. playlists that describe available qualities. On the other hand
2161 # media playlists [1, 4.3.3] should be returned as is since they contain
2162 # just the media without qualities renditions.
2163 # Fortunately, master playlist can be easily distinguished from media
2164 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
2165 # master playlist tags MUST NOT appear in a media playlist and vice versa.
2166 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
2167 # media playlist and MUST NOT appear in master playlist thus we can
2168 # clearly detect media playlist with this criterion.
2170 if '#EXT-X-TARGETDURATION' in m3u8_doc
: # media playlist, return as is
2172 'format_id': join_nonempty(m3u8_id
, idx
),
2173 'format_index': idx
,
2174 'url': m3u8_url
or encode_data_uri(m3u8_doc
.encode('utf-8'), 'application/x-mpegurl'),
2176 'protocol': entry_protocol
,
2177 'preference': preference
,
2180 } for idx
in _extract_m3u8_playlist_indices(m3u8_doc
=m3u8_doc
)]
2182 return formats
, subtitles
2185 last_stream_inf
= {}
2187 def extract_media(x_media_line
):
2188 media
= parse_m3u8_attributes(x_media_line
)
2189 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
2190 media_type
, group_id
, name
= media
.get('TYPE'), media
.get('GROUP-ID'), media
.get('NAME')
2191 if not (media_type
and group_id
and name
):
2193 groups
.setdefault(group_id
, []).append(media
)
2194 # <https://tools.ietf.org/html/rfc8216#section-4.3.4.1>
2195 if media_type
== 'SUBTITLES':
2196 # According to RFC 8216 §4.3.4.2.1, URI is REQUIRED in the
2197 # EXT-X-MEDIA tag if the media type is SUBTITLES.
2198 # However, lack of URI has been spotted in the wild.
2199 # e.g. NebulaIE; see https://github.com/yt-dlp/yt-dlp/issues/339
2200 if not media
.get('URI'):
2202 url
= format_url(media
['URI'])
2205 'ext': determine_ext(url
),
2207 if sub_info
['ext'] == 'm3u8':
2208 # Per RFC 8216 §3.1, the only possible subtitle format m3u8
2209 # files may contain is WebVTT:
2210 # <https://tools.ietf.org/html/rfc8216#section-3.1>
2211 sub_info
['ext'] = 'vtt'
2212 sub_info
['protocol'] = 'm3u8_native'
2213 lang
= media
.get('LANGUAGE') or 'und'
2214 subtitles
.setdefault(lang
, []).append(sub_info
)
2215 if media_type
not in ('VIDEO', 'AUDIO'):
2217 media_url
= media
.get('URI')
2219 manifest_url
= format_url(media_url
)
2221 'format_id': join_nonempty(m3u8_id
, group_id
, name
, idx
),
2222 'format_note': name
,
2223 'format_index': idx
,
2224 'url': manifest_url
,
2225 'manifest_url': m3u8_url
,
2226 'language': media
.get('LANGUAGE'),
2228 'protocol': entry_protocol
,
2229 'preference': preference
,
2231 'vcodec': 'none' if media_type
== 'AUDIO' else None,
2232 } for idx
in _extract_m3u8_playlist_indices(manifest_url
))
2234 def build_stream_name():
2235 # Despite specification does not mention NAME attribute for
2236 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
2237 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
2238 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
2239 stream_name
= last_stream_inf
.get('NAME')
2242 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
2243 # from corresponding rendition group
2244 stream_group_id
= last_stream_inf
.get('VIDEO')
2245 if not stream_group_id
:
2247 stream_group
= groups
.get(stream_group_id
)
2248 if not stream_group
:
2249 return stream_group_id
2250 rendition
= stream_group
[0]
2251 return rendition
.get('NAME') or stream_group_id
2253 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
2254 # chance to detect video only formats when EXT-X-STREAM-INF tags
2255 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
2256 for line
in m3u8_doc
.splitlines():
2257 if line
.startswith('#EXT-X-MEDIA:'):
2260 for line
in m3u8_doc
.splitlines():
2261 if line
.startswith('#EXT-X-STREAM-INF:'):
2262 last_stream_inf
= parse_m3u8_attributes(line
)
2263 elif line
.startswith('#') or not line
.strip():
2266 tbr
= float_or_none(
2267 last_stream_inf
.get('AVERAGE-BANDWIDTH')
2268 or last_stream_inf
.get('BANDWIDTH'), scale
=1000)
2269 manifest_url
= format_url(line
.strip())
2271 for idx
in _extract_m3u8_playlist_indices(manifest_url
):
2272 format_id
= [m3u8_id
, None, idx
]
2273 # Bandwidth of live streams may differ over time thus making
2274 # format_id unpredictable. So it's better to keep provided
2277 stream_name
= build_stream_name()
2278 format_id
[1] = stream_name
or '%d' % (tbr
or len(formats
))
2280 'format_id': join_nonempty(*format_id
),
2281 'format_index': idx
,
2282 'url': manifest_url
,
2283 'manifest_url': m3u8_url
,
2286 'fps': float_or_none(last_stream_inf
.get('FRAME-RATE')),
2287 'protocol': entry_protocol
,
2288 'preference': preference
,
2291 resolution
= last_stream_inf
.get('RESOLUTION')
2293 mobj
= re
.search(r
'(?P<width>\d+)[xX](?P<height>\d+)', resolution
)
2295 f
['width'] = int(mobj
.group('width'))
2296 f
['height'] = int(mobj
.group('height'))
2297 # Unified Streaming Platform
2299 r
'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f
['url'])
2301 abr
, vbr
= mobj
.groups()
2302 abr
, vbr
= float_or_none(abr
, 1000), float_or_none(vbr
, 1000)
2307 codecs
= parse_codecs(last_stream_inf
.get('CODECS'))
2309 audio_group_id
= last_stream_inf
.get('AUDIO')
2310 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
2311 # references a rendition group MUST have a CODECS attribute.
2312 # However, this is not always respected, for example, [2]
2313 # contains EXT-X-STREAM-INF tag which references AUDIO
2314 # rendition group but does not have CODECS and despite
2315 # referencing an audio group it represents a complete
2316 # (with audio and video) format. So, for such cases we will
2317 # ignore references to rendition groups and treat them
2318 # as complete formats.
2319 if audio_group_id
and codecs
and f
.get('vcodec') != 'none':
2320 audio_group
= groups
.get(audio_group_id
)
2321 if audio_group
and audio_group
[0].get('URI'):
2322 # TODO: update acodec for audio only formats with
2324 f
['acodec'] = 'none'
2325 if not f
.get('ext'):
2326 f
['ext'] = 'm4a' if f
.get('vcodec') == 'none' else 'mp4'
2330 progressive_uri
= last_stream_inf
.get('PROGRESSIVE-URI')
2333 del http_f
['manifest_url']
2335 'format_id': f
['format_id'].replace('hls-', 'http-'),
2337 'url': progressive_uri
,
2339 formats
.append(http_f
)
2341 last_stream_inf
= {}
2342 return formats
, subtitles
2344 def _extract_m3u8_vod_duration(
2345 self
, m3u8_vod_url
, video_id
, note
=None, errnote
=None, data
=None, headers
={}, query={}
):
2347 m3u8_vod
= self
._download
_webpage
(
2348 m3u8_vod_url
, video_id
,
2349 note
='Downloading m3u8 VOD manifest' if note
is None else note
,
2350 errnote
='Failed to download VOD manifest' if errnote
is None else errnote
,
2351 fatal
=False, data
=data
, headers
=headers
, query
=query
)
2353 return self
._parse
_m
3u8_vod
_duration
(m3u8_vod
or '', video_id
)
2355 def _parse_m3u8_vod_duration(self
, m3u8_vod
, video_id
):
2356 if '#EXT-X-PLAYLIST-TYPE:VOD' not in m3u8_vod
:
2360 float(line
[len('#EXTINF:'):].split(',')[0])
2361 for line
in m3u8_vod
.splitlines() if line
.startswith('#EXTINF:'))) or None
2364 def _xpath_ns(path
, namespace
=None):
2368 for c
in path
.split('/'):
2369 if not c
or c
== '.':
2372 out
.append('{%s}%s' % (namespace
, c
))
2373 return '/'.join(out
)
2375 def _extract_smil_formats_and_subtitles(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None, transform_source
=None):
2376 smil
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
, transform_source
=transform_source
)
2382 namespace
= self
._parse
_smil
_namespace
(smil
)
2384 fmts
= self
._parse
_smil
_formats
(
2385 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2386 subs
= self
._parse
_smil
_subtitles
(
2387 smil
, namespace
=namespace
)
2391 def _extract_smil_formats(self
, *args
, **kwargs
):
2392 fmts
, subs
= self
._extract
_smil
_formats
_and
_subtitles
(*args
, **kwargs
)
2394 self
._report
_ignoring
_subs
('SMIL')
2397 def _extract_smil_info(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None):
2398 smil
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
)
2401 return self
._parse
_smil
(smil
, smil_url
, video_id
, f4m_params
=f4m_params
)
2403 def _download_smil(self
, smil_url
, video_id
, fatal
=True, transform_source
=None):
2404 return self
._download
_xml
(
2405 smil_url
, video_id
, 'Downloading SMIL file',
2406 'Unable to download SMIL file', fatal
=fatal
, transform_source
=transform_source
)
2408 def _parse_smil(self
, smil
, smil_url
, video_id
, f4m_params
=None):
2409 namespace
= self
._parse
_smil
_namespace
(smil
)
2411 formats
= self
._parse
_smil
_formats
(
2412 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2413 subtitles
= self
._parse
_smil
_subtitles
(smil
, namespace
=namespace
)
2415 video_id
= os
.path
.splitext(url_basename(smil_url
))[0]
2419 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2420 name
= meta
.attrib
.get('name')
2421 content
= meta
.attrib
.get('content')
2422 if not name
or not content
:
2424 if not title
and name
== 'title':
2426 elif not description
and name
in ('description', 'abstract'):
2427 description
= content
2428 elif not upload_date
and name
== 'date':
2429 upload_date
= unified_strdate(content
)
2432 'id': image
.get('type'),
2433 'url': image
.get('src'),
2434 'width': int_or_none(image
.get('width')),
2435 'height': int_or_none(image
.get('height')),
2436 } for image
in smil
.findall(self
._xpath
_ns
('.//image', namespace
)) if image
.get('src')]
2440 'title': title
or video_id
,
2441 'description': description
,
2442 'upload_date': upload_date
,
2443 'thumbnails': thumbnails
,
2445 'subtitles': subtitles
,
2448 def _parse_smil_namespace(self
, smil
):
2449 return self
._search
_regex
(
2450 r
'(?i)^{([^}]+)?}smil$', smil
.tag
, 'namespace', default
=None)
2452 def _parse_smil_formats(self
, smil
, smil_url
, video_id
, namespace
=None, f4m_params
=None, transform_rtmp_url
=None):
2454 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2455 b
= meta
.get('base') or meta
.get('httpBase')
2467 media
= smil
.findall(self
._xpath
_ns
('.//video', namespace
)) + smil
.findall(self
._xpath
_ns
('.//audio', namespace
))
2468 for medium
in media
:
2469 src
= medium
.get('src')
2470 if not src
or src
in srcs
:
2474 bitrate
= float_or_none(medium
.get('system-bitrate') or medium
.get('systemBitrate'), 1000)
2475 filesize
= int_or_none(medium
.get('size') or medium
.get('fileSize'))
2476 width
= int_or_none(medium
.get('width'))
2477 height
= int_or_none(medium
.get('height'))
2478 proto
= medium
.get('proto')
2479 ext
= medium
.get('ext')
2480 src_ext
= determine_ext(src
)
2481 streamer
= medium
.get('streamer') or base
2483 if proto
== 'rtmp' or streamer
.startswith('rtmp'):
2489 'format_id': 'rtmp-%d' % (rtmp_count
if bitrate
is None else bitrate
),
2491 'filesize': filesize
,
2495 if transform_rtmp_url
:
2496 streamer
, src
= transform_rtmp_url(streamer
, src
)
2497 formats
[-1].update({
2503 src_url
= src
if src
.startswith('http') else compat_urlparse
.urljoin(base
, src
)
2504 src_url
= src_url
.strip()
2506 if proto
== 'm3u8' or src_ext
== 'm3u8':
2507 m3u8_formats
= self
._extract
_m
3u8_formats
(
2508 src_url
, video_id
, ext
or 'mp4', m3u8_id
='hls', fatal
=False)
2509 if len(m3u8_formats
) == 1:
2511 m3u8_formats
[0].update({
2512 'format_id': 'hls-%d' % (m3u8_count
if bitrate
is None else bitrate
),
2517 formats
.extend(m3u8_formats
)
2518 elif src_ext
== 'f4m':
2523 'plugin': 'flowplayer-3.2.0.1',
2525 f4m_url
+= '&' if '?' in f4m_url
else '?'
2526 f4m_url
+= compat_urllib_parse_urlencode(f4m_params
)
2527 formats
.extend(self
._extract
_f
4m
_formats
(f4m_url
, video_id
, f4m_id
='hds', fatal
=False))
2528 elif src_ext
== 'mpd':
2529 formats
.extend(self
._extract
_mpd
_formats
(
2530 src_url
, video_id
, mpd_id
='dash', fatal
=False))
2531 elif re
.search(r
'\.ism/[Mm]anifest', src_url
):
2532 formats
.extend(self
._extract
_ism
_formats
(
2533 src_url
, video_id
, ism_id
='mss', fatal
=False))
2534 elif src_url
.startswith('http') and self
._is
_valid
_url
(src
, video_id
):
2538 'ext': ext
or src_ext
or 'flv',
2539 'format_id': 'http-%d' % (bitrate
or http_count
),
2541 'filesize': filesize
,
2546 for medium
in smil
.findall(self
._xpath
_ns
('.//imagestream', namespace
)):
2547 src
= medium
.get('src')
2548 if not src
or src
in srcs
:
2554 'format_id': 'imagestream-%d' % (imgs_count
),
2556 'ext': mimetype2ext(medium
.get('type')),
2559 'width': int_or_none(medium
.get('width')),
2560 'height': int_or_none(medium
.get('height')),
2561 'format_note': 'SMIL storyboards',
2566 def _parse_smil_subtitles(self
, smil
, namespace
=None, subtitles_lang
='en'):
2569 for num
, textstream
in enumerate(smil
.findall(self
._xpath
_ns
('.//textstream', namespace
))):
2570 src
= textstream
.get('src')
2571 if not src
or src
in urls
:
2574 ext
= textstream
.get('ext') or mimetype2ext(textstream
.get('type')) or determine_ext(src
)
2575 lang
= textstream
.get('systemLanguage') or textstream
.get('systemLanguageName') or textstream
.get('lang') or subtitles_lang
2576 subtitles
.setdefault(lang
, []).append({
2582 def _extract_xspf_playlist(self
, xspf_url
, playlist_id
, fatal
=True):
2583 xspf
= self
._download
_xml
(
2584 xspf_url
, playlist_id
, 'Downloading xpsf playlist',
2585 'Unable to download xspf manifest', fatal
=fatal
)
2588 return self
._parse
_xspf
(
2589 xspf
, playlist_id
, xspf_url
=xspf_url
,
2590 xspf_base_url
=base_url(xspf_url
))
2592 def _parse_xspf(self
, xspf_doc
, playlist_id
, xspf_url
=None, xspf_base_url
=None):
2594 'xspf': 'http://xspf.org/ns/0/',
2595 's1': 'http://static.streamone.nl/player/ns/0',
2599 for track
in xspf_doc
.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP
)):
2601 track
, xpath_with_ns('./xspf:title', NS_MAP
), 'title', default
=playlist_id
)
2602 description
= xpath_text(
2603 track
, xpath_with_ns('./xspf:annotation', NS_MAP
), 'description')
2604 thumbnail
= xpath_text(
2605 track
, xpath_with_ns('./xspf:image', NS_MAP
), 'thumbnail')
2606 duration
= float_or_none(
2607 xpath_text(track
, xpath_with_ns('./xspf:duration', NS_MAP
), 'duration'), 1000)
2610 for location
in track
.findall(xpath_with_ns('./xspf:location', NS_MAP
)):
2611 format_url
= urljoin(xspf_base_url
, location
.text
)
2616 'manifest_url': xspf_url
,
2617 'format_id': location
.get(xpath_with_ns('s1:label', NS_MAP
)),
2618 'width': int_or_none(location
.get(xpath_with_ns('s1:width', NS_MAP
))),
2619 'height': int_or_none(location
.get(xpath_with_ns('s1:height', NS_MAP
))),
2621 self
._sort
_formats
(formats
)
2626 'description': description
,
2627 'thumbnail': thumbnail
,
2628 'duration': duration
,
2633 def _extract_mpd_formats(self
, *args
, **kwargs
):
2634 fmts
, subs
= self
._extract
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2636 self
._report
_ignoring
_subs
('DASH')
2639 def _extract_mpd_formats_and_subtitles(
2640 self
, mpd_url
, video_id
, mpd_id
=None, note
=None, errnote
=None,
2641 fatal
=True, data
=None, headers
={}, query={}
):
2642 res
= self
._download
_xml
_handle
(
2644 note
='Downloading MPD manifest' if note
is None else note
,
2645 errnote
='Failed to download MPD manifest' if errnote
is None else errnote
,
2646 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
2652 mpd_base_url
= base_url(urlh
.geturl())
2654 return self
._parse
_mpd
_formats
_and
_subtitles
(
2655 mpd_doc
, mpd_id
, mpd_base_url
, mpd_url
)
2657 def _parse_mpd_formats(self
, *args
, **kwargs
):
2658 fmts
, subs
= self
._parse
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2660 self
._report
_ignoring
_subs
('DASH')
2663 def _parse_mpd_formats_and_subtitles(
2664 self
, mpd_doc
, mpd_id
=None, mpd_base_url
='', mpd_url
=None):
2666 Parse formats from MPD manifest.
2668 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
2669 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
2670 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
2672 if not self
.get_param('dynamic_mpd', True):
2673 if mpd_doc
.get('type') == 'dynamic':
2676 namespace
= self
._search
_regex
(r
'(?i)^{([^}]+)?}MPD$', mpd_doc
.tag
, 'namespace', default
=None)
2679 return self
._xpath
_ns
(path
, namespace
)
2681 def is_drm_protected(element
):
2682 return element
.find(_add_ns('ContentProtection')) is not None
2684 def extract_multisegment_info(element
, ms_parent_info
):
2685 ms_info
= ms_parent_info
.copy()
2687 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2688 # common attributes and elements. We will only extract relevant
2690 def extract_common(source
):
2691 segment_timeline
= source
.find(_add_ns('SegmentTimeline'))
2692 if segment_timeline
is not None:
2693 s_e
= segment_timeline
.findall(_add_ns('S'))
2695 ms_info
['total_number'] = 0
2698 r
= int(s
.get('r', 0))
2699 ms_info
['total_number'] += 1 + r
2700 ms_info
['s'].append({
2701 't': int(s
.get('t', 0)),
2702 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2703 'd': int(s
.attrib
['d']),
2706 start_number
= source
.get('startNumber')
2708 ms_info
['start_number'] = int(start_number
)
2709 timescale
= source
.get('timescale')
2711 ms_info
['timescale'] = int(timescale
)
2712 segment_duration
= source
.get('duration')
2713 if segment_duration
:
2714 ms_info
['segment_duration'] = float(segment_duration
)
2716 def extract_Initialization(source
):
2717 initialization
= source
.find(_add_ns('Initialization'))
2718 if initialization
is not None:
2719 ms_info
['initialization_url'] = initialization
.attrib
['sourceURL']
2721 segment_list
= element
.find(_add_ns('SegmentList'))
2722 if segment_list
is not None:
2723 extract_common(segment_list
)
2724 extract_Initialization(segment_list
)
2725 segment_urls_e
= segment_list
.findall(_add_ns('SegmentURL'))
2727 ms_info
['segment_urls'] = [segment
.attrib
['media'] for segment
in segment_urls_e
]
2729 segment_template
= element
.find(_add_ns('SegmentTemplate'))
2730 if segment_template
is not None:
2731 extract_common(segment_template
)
2732 media
= segment_template
.get('media')
2734 ms_info
['media'] = media
2735 initialization
= segment_template
.get('initialization')
2737 ms_info
['initialization'] = initialization
2739 extract_Initialization(segment_template
)
2742 mpd_duration
= parse_duration(mpd_doc
.get('mediaPresentationDuration'))
2743 formats
, subtitles
= [], {}
2744 stream_numbers
= collections
.defaultdict(int)
2745 for period
in mpd_doc
.findall(_add_ns('Period')):
2746 period_duration
= parse_duration(period
.get('duration')) or mpd_duration
2747 period_ms_info
= extract_multisegment_info(period
, {
2751 for adaptation_set
in period
.findall(_add_ns('AdaptationSet')):
2752 adaption_set_ms_info
= extract_multisegment_info(adaptation_set
, period_ms_info
)
2753 for representation
in adaptation_set
.findall(_add_ns('Representation')):
2754 representation_attrib
= adaptation_set
.attrib
.copy()
2755 representation_attrib
.update(representation
.attrib
)
2756 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2757 mime_type
= representation_attrib
['mimeType']
2758 content_type
= representation_attrib
.get('contentType', mime_type
.split('/')[0])
2760 codecs
= parse_codecs(representation_attrib
.get('codecs', ''))
2761 if content_type
not in ('video', 'audio', 'text'):
2762 if mime_type
== 'image/jpeg':
2763 content_type
= mime_type
2764 elif codecs
['vcodec'] != 'none':
2765 content_type
= 'video'
2766 elif codecs
['acodec'] != 'none':
2767 content_type
= 'audio'
2768 elif codecs
.get('tcodec', 'none') != 'none':
2769 content_type
= 'text'
2770 elif mimetype2ext(mime_type
) in ('tt', 'dfxp', 'ttml', 'xml', 'json'):
2771 content_type
= 'text'
2773 self
.report_warning('Unknown MIME type %s in DASH manifest' % mime_type
)
2777 for element
in (representation
, adaptation_set
, period
, mpd_doc
):
2778 base_url_e
= element
.find(_add_ns('BaseURL'))
2779 if base_url_e
is not None:
2780 base_url
= base_url_e
.text
+ base_url
2781 if re
.match(r
'^https?://', base_url
):
2783 if mpd_base_url
and base_url
.startswith('/'):
2784 base_url
= compat_urlparse
.urljoin(mpd_base_url
, base_url
)
2785 elif mpd_base_url
and not re
.match(r
'^https?://', base_url
):
2786 if not mpd_base_url
.endswith('/'):
2788 base_url
= mpd_base_url
+ base_url
2789 representation_id
= representation_attrib
.get('id')
2790 lang
= representation_attrib
.get('lang')
2791 url_el
= representation
.find(_add_ns('BaseURL'))
2792 filesize
= int_or_none(url_el
.attrib
.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el
is not None else None)
2793 bandwidth
= int_or_none(representation_attrib
.get('bandwidth'))
2794 if representation_id
is not None:
2795 format_id
= representation_id
2797 format_id
= content_type
2799 format_id
= mpd_id
+ '-' + format_id
2800 if content_type
in ('video', 'audio'):
2802 'format_id': format_id
,
2803 'manifest_url': mpd_url
,
2804 'ext': mimetype2ext(mime_type
),
2805 'width': int_or_none(representation_attrib
.get('width')),
2806 'height': int_or_none(representation_attrib
.get('height')),
2807 'tbr': float_or_none(bandwidth
, 1000),
2808 'asr': int_or_none(representation_attrib
.get('audioSamplingRate')),
2809 'fps': int_or_none(representation_attrib
.get('frameRate')),
2810 'language': lang
if lang
not in ('mul', 'und', 'zxx', 'mis') else None,
2811 'format_note': 'DASH %s' % content_type
,
2812 'filesize': filesize
,
2813 'container': mimetype2ext(mime_type
) + '_dash',
2816 elif content_type
== 'text':
2818 'ext': mimetype2ext(mime_type
),
2819 'manifest_url': mpd_url
,
2820 'filesize': filesize
,
2822 elif content_type
== 'image/jpeg':
2823 # See test case in VikiIE
2824 # https://www.viki.com/videos/1175236v-choosing-spouse-by-lottery-episode-1
2826 'format_id': format_id
,
2828 'manifest_url': mpd_url
,
2829 'format_note': 'DASH storyboards (jpeg)',
2833 if is_drm_protected(adaptation_set
) or is_drm_protected(representation
):
2835 representation_ms_info
= extract_multisegment_info(representation
, adaption_set_ms_info
)
2837 def prepare_template(template_name
, identifiers
):
2838 tmpl
= representation_ms_info
[template_name
]
2839 # First of, % characters outside $...$ templates
2840 # must be escaped by doubling for proper processing
2841 # by % operator string formatting used further (see
2842 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2848 in_template
= not in_template
2849 elif c
== '%' and not in_template
:
2851 # Next, $...$ templates are translated to their
2852 # %(...) counterparts to be used with % operator
2853 if representation_id
is not None:
2854 t
= t
.replace('$RepresentationID$', representation_id
)
2855 t
= re
.sub(r
'\$(%s)\$' % '|'.join(identifiers
), r
'%(\1)d', t
)
2856 t
= re
.sub(r
'\$(%s)%%([^$]+)\$' % '|'.join(identifiers
), r
'%(\1)\2', t
)
2857 t
.replace('$$', '$')
2860 # @initialization is a regular template like @media one
2861 # so it should be handled just the same way (see
2862 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2863 if 'initialization' in representation_ms_info
:
2864 initialization_template
= prepare_template(
2866 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2867 # $Time$ shall not be included for @initialization thus
2868 # only $Bandwidth$ remains
2870 representation_ms_info
['initialization_url'] = initialization_template
% {
2871 'Bandwidth': bandwidth
,
2874 def location_key(location
):
2875 return 'url' if re
.match(r
'^https?://', location
) else 'path'
2877 if 'segment_urls' not in representation_ms_info
and 'media' in representation_ms_info
:
2879 media_template
= prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2880 media_location_key
= location_key(media_template
)
2882 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2883 # can't be used at the same time
2884 if '%(Number' in media_template
and 's' not in representation_ms_info
:
2885 segment_duration
= None
2886 if 'total_number' not in representation_ms_info
and 'segment_duration' in representation_ms_info
:
2887 segment_duration
= float_or_none(representation_ms_info
['segment_duration'], representation_ms_info
['timescale'])
2888 representation_ms_info
['total_number'] = int(math
.ceil(
2889 float_or_none(period_duration
, segment_duration
, default
=0)))
2890 representation_ms_info
['fragments'] = [{
2891 media_location_key
: media_template
% {
2892 'Number': segment_number
,
2893 'Bandwidth': bandwidth
,
2895 'duration': segment_duration
,
2896 } for segment_number
in range(
2897 representation_ms_info
['start_number'],
2898 representation_ms_info
['total_number'] + representation_ms_info
['start_number'])]
2900 # $Number*$ or $Time$ in media template with S list available
2901 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2902 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2903 representation_ms_info
['fragments'] = []
2906 segment_number
= representation_ms_info
['start_number']
2908 def add_segment_url():
2909 segment_url
= media_template
% {
2910 'Time': segment_time
,
2911 'Bandwidth': bandwidth
,
2912 'Number': segment_number
,
2914 representation_ms_info
['fragments'].append({
2915 media_location_key
: segment_url
,
2916 'duration': float_or_none(segment_d
, representation_ms_info
['timescale']),
2919 for num
, s
in enumerate(representation_ms_info
['s']):
2920 segment_time
= s
.get('t') or segment_time
2924 for r
in range(s
.get('r', 0)):
2925 segment_time
+= segment_d
2928 segment_time
+= segment_d
2929 elif 'segment_urls' in representation_ms_info
and 's' in representation_ms_info
:
2931 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2932 # or any YouTube dashsegments video
2935 timescale
= representation_ms_info
['timescale']
2936 for s
in representation_ms_info
['s']:
2937 duration
= float_or_none(s
['d'], timescale
)
2938 for r
in range(s
.get('r', 0) + 1):
2939 segment_uri
= representation_ms_info
['segment_urls'][segment_index
]
2941 location_key(segment_uri
): segment_uri
,
2942 'duration': duration
,
2945 representation_ms_info
['fragments'] = fragments
2946 elif 'segment_urls' in representation_ms_info
:
2947 # Segment URLs with no SegmentTimeline
2948 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2949 # https://github.com/ytdl-org/youtube-dl/pull/14844
2951 segment_duration
= float_or_none(
2952 representation_ms_info
['segment_duration'],
2953 representation_ms_info
['timescale']) if 'segment_duration' in representation_ms_info
else None
2954 for segment_url
in representation_ms_info
['segment_urls']:
2956 location_key(segment_url
): segment_url
,
2958 if segment_duration
:
2959 fragment
['duration'] = segment_duration
2960 fragments
.append(fragment
)
2961 representation_ms_info
['fragments'] = fragments
2962 # If there is a fragments key available then we correctly recognized fragmented media.
2963 # Otherwise we will assume unfragmented media with direct access. Technically, such
2964 # assumption is not necessarily correct since we may simply have no support for
2965 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2966 if 'fragments' in representation_ms_info
:
2968 # NB: mpd_url may be empty when MPD manifest is parsed from a string
2969 'url': mpd_url
or base_url
,
2970 'fragment_base_url': base_url
,
2972 'protocol': 'http_dash_segments' if mime_type
!= 'image/jpeg' else 'mhtml',
2974 if 'initialization_url' in representation_ms_info
:
2975 initialization_url
= representation_ms_info
['initialization_url']
2976 if not f
.get('url'):
2977 f
['url'] = initialization_url
2978 f
['fragments'].append({location_key(initialization_url): initialization_url}
)
2979 f
['fragments'].extend(representation_ms_info
['fragments'])
2980 if not period_duration
:
2981 period_duration
= try_get(
2982 representation_ms_info
,
2983 lambda r
: sum(frag
['duration'] for frag
in r
['fragments']), float)
2985 # Assuming direct URL to unfragmented media.
2987 if content_type
in ('video', 'audio', 'image/jpeg'):
2988 f
['manifest_stream_number'] = stream_numbers
[f
['url']]
2989 stream_numbers
[f
['url']] += 1
2991 elif content_type
== 'text':
2992 subtitles
.setdefault(lang
or 'und', []).append(f
)
2994 return formats
, subtitles
2996 def _extract_ism_formats(self
, *args
, **kwargs
):
2997 fmts
, subs
= self
._extract
_ism
_formats
_and
_subtitles
(*args
, **kwargs
)
2999 self
._report
_ignoring
_subs
('ISM')
3002 def _extract_ism_formats_and_subtitles(self
, ism_url
, video_id
, ism_id
=None, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
):
3003 res
= self
._download
_xml
_handle
(
3005 note
='Downloading ISM manifest' if note
is None else note
,
3006 errnote
='Failed to download ISM manifest' if errnote
is None else errnote
,
3007 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
3014 return self
._parse
_ism
_formats
_and
_subtitles
(ism_doc
, urlh
.geturl(), ism_id
)
3016 def _parse_ism_formats_and_subtitles(self
, ism_doc
, ism_url
, ism_id
=None):
3018 Parse formats from ISM manifest.
3020 1. [MS-SSTR]: Smooth Streaming Protocol,
3021 https://msdn.microsoft.com/en-us/library/ff469518.aspx
3023 if ism_doc
.get('IsLive') == 'TRUE':
3026 duration
= int(ism_doc
.attrib
['Duration'])
3027 timescale
= int_or_none(ism_doc
.get('TimeScale')) or 10000000
3031 for stream
in ism_doc
.findall('StreamIndex'):
3032 stream_type
= stream
.get('Type')
3033 if stream_type
not in ('video', 'audio', 'text'):
3035 url_pattern
= stream
.attrib
['Url']
3036 stream_timescale
= int_or_none(stream
.get('TimeScale')) or timescale
3037 stream_name
= stream
.get('Name')
3038 stream_language
= stream
.get('Language', 'und')
3039 for track
in stream
.findall('QualityLevel'):
3040 fourcc
= track
.get('FourCC') or ('AACL' if track
.get('AudioTag') == '255' else None)
3041 # TODO: add support for WVC1 and WMAP
3042 if fourcc
not in ('H264', 'AVC1', 'AACL', 'TTML'):
3043 self
.report_warning('%s is not a supported codec' % fourcc
)
3045 tbr
= int(track
.attrib
['Bitrate']) // 1000
3046 # [1] does not mention Width and Height attributes. However,
3047 # they're often present while MaxWidth and MaxHeight are
3048 # missing, so should be used as fallbacks
3049 width
= int_or_none(track
.get('MaxWidth') or track
.get('Width'))
3050 height
= int_or_none(track
.get('MaxHeight') or track
.get('Height'))
3051 sampling_rate
= int_or_none(track
.get('SamplingRate'))
3053 track_url_pattern
= re
.sub(r
'{[Bb]itrate}', track
.attrib
['Bitrate'], url_pattern
)
3054 track_url_pattern
= compat_urlparse
.urljoin(ism_url
, track_url_pattern
)
3060 stream_fragments
= stream
.findall('c')
3061 for stream_fragment_index
, stream_fragment
in enumerate(stream_fragments
):
3062 fragment_ctx
['time'] = int_or_none(stream_fragment
.get('t')) or fragment_ctx
['time']
3063 fragment_repeat
= int_or_none(stream_fragment
.get('r')) or 1
3064 fragment_ctx
['duration'] = int_or_none(stream_fragment
.get('d'))
3065 if not fragment_ctx
['duration']:
3067 next_fragment_time
= int(stream_fragment
[stream_fragment_index
+ 1].attrib
['t'])
3069 next_fragment_time
= duration
3070 fragment_ctx
['duration'] = (next_fragment_time
- fragment_ctx
['time']) / fragment_repeat
3071 for _
in range(fragment_repeat
):
3073 'url': re
.sub(r
'{start[ _]time}', compat_str(fragment_ctx
['time']), track_url_pattern
),
3074 'duration': fragment_ctx
['duration'] / stream_timescale
,
3076 fragment_ctx
['time'] += fragment_ctx
['duration']
3078 if stream_type
== 'text':
3079 subtitles
.setdefault(stream_language
, []).append({
3083 'manifest_url': ism_url
,
3084 'fragments': fragments
,
3085 '_download_params': {
3086 'stream_type': stream_type
,
3087 'duration': duration
,
3088 'timescale': stream_timescale
,
3090 'language': stream_language
,
3091 'codec_private_data': track
.get('CodecPrivateData'),
3094 elif stream_type
in ('video', 'audio'):
3096 'format_id': join_nonempty(ism_id
, stream_name
, tbr
),
3098 'manifest_url': ism_url
,
3099 'ext': 'ismv' if stream_type
== 'video' else 'isma',
3103 'asr': sampling_rate
,
3104 'vcodec': 'none' if stream_type
== 'audio' else fourcc
,
3105 'acodec': 'none' if stream_type
== 'video' else fourcc
,
3107 'fragments': fragments
,
3108 'has_drm': ism_doc
.find('Protection') is not None,
3109 '_download_params': {
3110 'stream_type': stream_type
,
3111 'duration': duration
,
3112 'timescale': stream_timescale
,
3113 'width': width
or 0,
3114 'height': height
or 0,
3116 'language': stream_language
,
3117 'codec_private_data': track
.get('CodecPrivateData'),
3118 'sampling_rate': sampling_rate
,
3119 'channels': int_or_none(track
.get('Channels', 2)),
3120 'bits_per_sample': int_or_none(track
.get('BitsPerSample', 16)),
3121 'nal_unit_length_field': int_or_none(track
.get('NALUnitLengthField', 4)),
3124 return formats
, subtitles
3126 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):
3127 def absolute_url(item_url
):
3128 return urljoin(base_url
, item_url
)
3130 def parse_content_type(content_type
):
3131 if not content_type
:
3133 ctr
= re
.search(r
'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type
)
3135 mimetype
, codecs
= ctr
.groups()
3136 f
= parse_codecs(codecs
)
3137 f
['ext'] = mimetype2ext(mimetype
)
3141 def _media_formats(src
, cur_media_type
, type_info
={}):
3142 full_url
= absolute_url(src
)
3143 ext
= type_info
.get('ext') or determine_ext(full_url
)
3145 is_plain_url
= False
3146 formats
= self
._extract
_m
3u8_formats
(
3147 full_url
, video_id
, ext
='mp4',
3148 entry_protocol
=m3u8_entry_protocol
, m3u8_id
=m3u8_id
,
3149 preference
=preference
, quality
=quality
, fatal
=False)
3151 is_plain_url
= False
3152 formats
= self
._extract
_mpd
_formats
(
3153 full_url
, video_id
, mpd_id
=mpd_id
, fatal
=False)
3158 'vcodec': 'none' if cur_media_type
== 'audio' else None,
3160 return is_plain_url
, formats
3163 # amp-video and amp-audio are very similar to their HTML5 counterparts
3164 # so we wll include them right here (see
3165 # https://www.ampproject.org/docs/reference/components/amp-video)
3166 # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/
3167 _MEDIA_TAG_NAME_RE
= r
'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)'
3168 media_tags
= [(media_tag
, media_tag_name
, media_type
, '')
3169 for media_tag
, media_tag_name
, media_type
3170 in re
.findall(r
'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE
, webpage
)]
3171 media_tags
.extend(re
.findall(
3172 # We only allow video|audio followed by a whitespace or '>'.
3173 # Allowing more characters may end up in significant slow down (see
3174 # https://github.com/ytdl-org/youtube-dl/issues/11979, example URL:
3175 # http://www.porntrex.com/maps/videositemap.xml).
3176 r
'(?s)(<(?P<tag>%s)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>' % _MEDIA_TAG_NAME_RE
, webpage
))
3177 for media_tag
, _
, media_type
, media_content
in media_tags
:
3182 media_attributes
= extract_attributes(media_tag
)
3183 src
= strip_or_none(media_attributes
.get('src'))
3185 _
, formats
= _media_formats(src
, media_type
)
3186 media_info
['formats'].extend(formats
)
3187 media_info
['thumbnail'] = absolute_url(media_attributes
.get('poster'))
3189 for source_tag
in re
.findall(r
'<source[^>]+>', media_content
):
3190 s_attr
= extract_attributes(source_tag
)
3191 # data-video-src and data-src are non standard but seen
3192 # several times in the wild
3193 src
= strip_or_none(dict_get(s_attr
, ('src', 'data-video-src', 'data-src')))
3196 f
= parse_content_type(s_attr
.get('type'))
3197 is_plain_url
, formats
= _media_formats(src
, media_type
, f
)
3199 # width, height, res, label and title attributes are
3200 # all not standard but seen several times in the wild
3203 for lbl
in ('label', 'title')
3204 if str_or_none(s_attr
.get(lbl
))
3206 width
= int_or_none(s_attr
.get('width'))
3207 height
= (int_or_none(s_attr
.get('height'))
3208 or int_or_none(s_attr
.get('res')))
3209 if not width
or not height
:
3211 resolution
= parse_resolution(lbl
)
3214 width
= width
or resolution
.get('width')
3215 height
= height
or resolution
.get('height')
3217 tbr
= parse_bitrate(lbl
)
3226 'format_id': s_attr
.get('label') or s_attr
.get('title'),
3228 f
.update(formats
[0])
3229 media_info
['formats'].append(f
)
3231 media_info
['formats'].extend(formats
)
3232 for track_tag
in re
.findall(r
'<track[^>]+>', media_content
):
3233 track_attributes
= extract_attributes(track_tag
)
3234 kind
= track_attributes
.get('kind')
3235 if not kind
or kind
in ('subtitles', 'captions'):
3236 src
= strip_or_none(track_attributes
.get('src'))
3239 lang
= track_attributes
.get('srclang') or track_attributes
.get('lang') or track_attributes
.get('label')
3240 media_info
['subtitles'].setdefault(lang
, []).append({
3241 'url': absolute_url(src
),
3243 for f
in media_info
['formats']:
3244 f
.setdefault('http_headers', {})['Referer'] = base_url
3245 if media_info
['formats'] or media_info
['subtitles']:
3246 entries
.append(media_info
)
3249 def _extract_akamai_formats(self
, *args
, **kwargs
):
3250 fmts
, subs
= self
._extract
_akamai
_formats
_and
_subtitles
(*args
, **kwargs
)
3252 self
._report
_ignoring
_subs
('akamai')
3255 def _extract_akamai_formats_and_subtitles(self
, manifest_url
, video_id
, hosts
={}):
3256 signed
= 'hdnea=' in manifest_url
3258 # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
3259 manifest_url
= re
.sub(
3260 r
'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
3261 '', manifest_url
).strip('?')
3266 hdcore_sign
= 'hdcore=3.7.0'
3267 f4m_url
= re
.sub(r
'(https?://[^/]+)/i/', r
'\1/z/', manifest_url
).replace('/master.m3u8', '/manifest.f4m')
3268 hds_host
= hosts
.get('hds')
3270 f4m_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hds_host
, f4m_url
)
3271 if 'hdcore=' not in f4m_url
:
3272 f4m_url
+= ('&' if '?' in f4m_url
else '?') + hdcore_sign
3273 f4m_formats
= self
._extract
_f
4m
_formats
(
3274 f4m_url
, video_id
, f4m_id
='hds', fatal
=False)
3275 for entry
in f4m_formats
:
3276 entry
.update({'extra_param_to_segment_url': hdcore_sign}
)
3277 formats
.extend(f4m_formats
)
3279 m3u8_url
= re
.sub(r
'(https?://[^/]+)/z/', r
'\1/i/', manifest_url
).replace('/manifest.f4m', '/master.m3u8')
3280 hls_host
= hosts
.get('hls')
3282 m3u8_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hls_host
, m3u8_url
)
3283 m3u8_formats
, m3u8_subtitles
= self
._extract
_m
3u8_formats
_and
_subtitles
(
3284 m3u8_url
, video_id
, 'mp4', 'm3u8_native',
3285 m3u8_id
='hls', fatal
=False)
3286 formats
.extend(m3u8_formats
)
3287 subtitles
= self
._merge
_subtitles
(subtitles
, m3u8_subtitles
)
3289 http_host
= hosts
.get('http')
3290 if http_host
and m3u8_formats
and not signed
:
3291 REPL_REGEX
= r
'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
3292 qualities
= re
.match(REPL_REGEX
, m3u8_url
).group(2).split(',')
3293 qualities_length
= len(qualities
)
3294 if len(m3u8_formats
) in (qualities_length
, qualities_length
+ 1):
3296 for f
in m3u8_formats
:
3297 if f
['vcodec'] != 'none':
3298 for protocol
in ('http', 'https'):
3300 del http_f
['manifest_url']
3302 REPL_REGEX
, protocol
+ r
'://%s/\g<1>%s\3' % (http_host
, qualities
[i
]), f
['url'])
3304 'format_id': http_f
['format_id'].replace('hls-', protocol
+ '-'),
3306 'protocol': protocol
,
3308 formats
.append(http_f
)
3311 return formats
, subtitles
3313 def _extract_wowza_formats(self
, url
, video_id
, m3u8_entry_protocol
='m3u8_native', skip_protocols
=[]):
3314 query
= compat_urlparse
.urlparse(url
).query
3315 url
= re
.sub(r
'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url
)
3317 r
'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url
)
3318 url_base
= mobj
.group('url')
3319 http_base_url
= '%s%s:%s' % ('http', mobj
.group('s') or '', url_base
)
3322 def manifest_url(manifest
):
3323 m_url
= '%s/%s' % (http_base_url
, manifest
)
3325 m_url
+= '?%s' % query
3328 if 'm3u8' not in skip_protocols
:
3329 formats
.extend(self
._extract
_m
3u8_formats
(
3330 manifest_url('playlist.m3u8'), video_id
, 'mp4',
3331 m3u8_entry_protocol
, m3u8_id
='hls', fatal
=False))
3332 if 'f4m' not in skip_protocols
:
3333 formats
.extend(self
._extract
_f
4m
_formats
(
3334 manifest_url('manifest.f4m'),
3335 video_id
, f4m_id
='hds', fatal
=False))
3336 if 'dash' not in skip_protocols
:
3337 formats
.extend(self
._extract
_mpd
_formats
(
3338 manifest_url('manifest.mpd'),
3339 video_id
, mpd_id
='dash', fatal
=False))
3340 if re
.search(r
'(?:/smil:|\.smil)', url_base
):
3341 if 'smil' not in skip_protocols
:
3342 rtmp_formats
= self
._extract
_smil
_formats
(
3343 manifest_url('jwplayer.smil'),
3344 video_id
, fatal
=False)
3345 for rtmp_format
in rtmp_formats
:
3346 rtsp_format
= rtmp_format
.copy()
3347 rtsp_format
['url'] = '%s/%s' % (rtmp_format
['url'], rtmp_format
['play_path'])
3348 del rtsp_format
['play_path']
3349 del rtsp_format
['ext']
3350 rtsp_format
.update({
3351 'url': rtsp_format
['url'].replace('rtmp://', 'rtsp://'),
3352 'format_id': rtmp_format
['format_id'].replace('rtmp', 'rtsp'),
3355 formats
.extend([rtmp_format
, rtsp_format
])
3357 for protocol
in ('rtmp', 'rtsp'):
3358 if protocol
not in skip_protocols
:
3360 'url': '%s:%s' % (protocol
, url_base
),
3361 'format_id': protocol
,
3362 'protocol': protocol
,
3366 def _find_jwplayer_data(self
, webpage
, video_id
=None, transform_source
=js_to_json
):
3368 r
'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P
=quote
)\
)(?
!</script
>).*?\
.setup\s
*\
((?P
<options
>[^
)]+)\
)',
3372 jwplayer_data = self._parse_json(mobj.group('options
'),
3374 transform_source=transform_source)
3375 except ExtractorError:
3378 if isinstance(jwplayer_data, dict):
3379 return jwplayer_data
3381 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
3382 jwplayer_data = self._find_jwplayer_data(
3383 webpage, video_id, transform_source=js_to_json)
3384 return self._parse_jwplayer_data(
3385 jwplayer_data, video_id, *args, **kwargs)
3387 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
3388 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3389 # JWPlayer backward compatibility: flattened playlists
3390 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
3391 if 'playlist
' not in jwplayer_data:
3392 jwplayer_data = {'playlist': [jwplayer_data]}
3396 # JWPlayer backward compatibility: single playlist item
3397 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
3398 if not isinstance(jwplayer_data['playlist
'], list):
3399 jwplayer_data['playlist
'] = [jwplayer_data['playlist
']]
3401 for video_data in jwplayer_data['playlist
']:
3402 # JWPlayer backward compatibility: flattened sources
3403 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
3404 if 'sources
' not in video_data:
3405 video_data['sources
'] = [video_data]
3407 this_video_id = video_id or video_data['mediaid
']
3409 formats = self._parse_jwplayer_formats(
3410 video_data['sources
'], video_id=this_video_id, m3u8_id=m3u8_id,
3411 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
3414 tracks = video_data.get('tracks
')
3415 if tracks and isinstance(tracks, list):
3416 for track in tracks:
3417 if not isinstance(track, dict):
3419 track_kind = track.get('kind
')
3420 if not track_kind or not isinstance(track_kind, compat_str):
3422 if track_kind.lower() not in ('captions
', 'subtitles
'):
3424 track_url = urljoin(base_url, track.get('file'))
3427 subtitles.setdefault(track.get('label
') or 'en
', []).append({
3428 'url
': self._proto_relative_url(track_url)
3432 'id': this_video_id,
3433 'title
': unescapeHTML(video_data['title
'] if require_title else video_data.get('title
')),
3434 'description
': clean_html(video_data.get('description
')),
3435 'thumbnail
': urljoin(base_url, self._proto_relative_url(video_data.get('image
'))),
3436 'timestamp
': int_or_none(video_data.get('pubdate
')),
3437 'duration
': float_or_none(jwplayer_data.get('duration
') or video_data.get('duration
')),
3438 'subtitles
': subtitles,
3440 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
3441 if len(formats) == 1 and re.search(r'^
(?
:http|
//).*(?
:youtube\
.com|youtu\
.be
)/.+', formats[0]['url
']):
3443 '_type
': 'url_transparent
',
3444 'url
': formats[0]['url
'],
3447 self._sort_formats(formats)
3448 entry['formats
'] = formats
3449 entries.append(entry)
3450 if len(entries) == 1:
3453 return self.playlist_result(entries)
3455 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
3456 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3459 for source in jwplayer_sources_data:
3460 if not isinstance(source, dict):
3462 source_url = urljoin(
3463 base_url, self._proto_relative_url(source.get('file')))
3464 if not source_url or source_url in urls:
3466 urls.append(source_url)
3467 source_type = source.get('type') or ''
3468 ext = mimetype2ext(source_type) or determine_ext(source_url)
3469 if source_type == 'hls
' or ext == 'm3u8
':
3470 formats.extend(self._extract_m3u8_formats(
3471 source_url, video_id, 'mp4
', entry_protocol='m3u8_native
',
3472 m3u8_id=m3u8_id, fatal=False))
3473 elif source_type == 'dash
' or ext == 'mpd
':
3474 formats.extend(self._extract_mpd_formats(
3475 source_url, video_id, mpd_id=mpd_id, fatal=False))
3477 formats.extend(self._extract_smil_formats(
3478 source_url, video_id, fatal=False))
3479 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3480 elif source_type.startswith('audio
') or ext in (
3481 'oga
', 'aac
', 'mp3
', 'mpeg
', 'vorbis
'):
3488 height = int_or_none(source.get('height
'))
3490 # Often no height is provided but there is a label in
3491 # format like "1080p", "720p SD", or 1080.
3492 height = int_or_none(self._search_regex(
3493 r'^
(\d{3,4}
)[pP
]?
(?
:\b|$
)', compat_str(source.get('label
') or ''),
3494 'height
', default=None))
3497 'width
': int_or_none(source.get('width
')),
3499 'tbr
': int_or_none(source.get('bitrate
')),
3502 if source_url.startswith('rtmp
'):
3503 a_format['ext
'] = 'flv
'
3504 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3505 # of jwplayer.flash.swf
3506 rtmp_url_parts = re.split(
3507 r'((?
:mp4|mp3|flv
):)', source_url, 1)
3508 if len(rtmp_url_parts) == 3:
3509 rtmp_url, prefix, play_path = rtmp_url_parts
3512 'play_path
': prefix + play_path,
3515 a_format.update(rtmp_params)
3516 formats.append(a_format)
3519 def _live_title(self, name):
3520 self._downloader.deprecation_warning('yt_dlp
.InfoExtractor
._live
_title
is deprecated
and does
not work
as expected
')
3523 def _int(self, v, name, fatal=False, **kwargs):
3524 res = int_or_none(v, **kwargs)
3526 msg = 'Failed to extract
%s: Could
not parse value
%r' % (name, v)
3528 raise ExtractorError(msg)
3530 self.report_warning(msg)
3533 def _float(self, v, name, fatal=False, **kwargs):
3534 res = float_or_none(v, **kwargs)
3536 msg = 'Failed to extract
%s: Could
not parse value
%r' % (name, v)
3538 raise ExtractorError(msg)
3540 self.report_warning(msg)
3543 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
3544 path='/', secure=False, discard=False, rest={}, **kwargs):
3545 cookie = compat_cookiejar_Cookie(
3546 0, name, value, port, port is not None, domain, True,
3547 domain.startswith('.'), path, True, secure, expire_time,
3548 discard, None, None, rest)
3549 self._downloader.cookiejar.set_cookie(cookie)
3551 def _get_cookies(self, url):
3552 """ Return a compat_cookies_SimpleCookie with the cookies for the url """
3553 req = sanitized_Request(url)
3554 self._downloader.cookiejar.add_cookie_header(req)
3555 return compat_cookies_SimpleCookie(req.get_header('Cookie
'))
3557 def _apply_first_set_cookie_header(self, url_handle, cookie):
3559 Apply first Set-Cookie header instead of the last. Experimental.
3561 Some sites (e.g. [1-3]) may serve two cookies under the same name
3562 in Set-Cookie header and expect the first (old) one to be set rather
3563 than second (new). However, as of RFC6265 the newer one cookie
3564 should be set into cookie store what actually happens.
3565 We will workaround this issue by resetting the cookie to
3566 the first one manually.
3567 1. https://new.vk.com/
3568 2. https://github.com/ytdl-org/youtube-dl/issues/9841#issuecomment-227871201
3569 3. https://learning.oreilly.com/
3571 for header, cookies in url_handle.headers.items():
3572 if header.lower() != 'set-cookie
':
3574 if sys.version_info[0] >= 3:
3575 cookies = cookies.encode('iso
-8859-1')
3576 cookies = cookies.decode('utf
-8')
3577 cookie_value = re.search(
3578 r'%s=(.+?
);.*?
\b[Dd
]omain
=(.+?
)(?
:[,;]|$
)' % cookie, cookies)
3580 value, domain = cookie_value.groups()
3581 self._set_cookie(domain, cookie, value)
3584 def get_testcases(self, include_onlymatching=False):
3585 t = getattr(self, '_TEST
', None)
3587 assert not hasattr(self, '_TESTS
'), \
3588 '%s has _TEST
and _TESTS
' % type(self).__name__
3591 tests = getattr(self, '_TESTS
', [])
3593 if not include_onlymatching and t.get('only_matching
', False):
3595 t['name
'] = type(self).__name__[:-len('IE
')]
3598 def is_suitable(self, age_limit):
3599 """ Test whether the extractor is generally suitable for the given
3600 age limit (i.e. pornographic sites are not, all others usually are) """
3602 any_restricted = False
3603 for tc in self.get_testcases(include_onlymatching=False):
3604 if tc.get('playlist
', []):
3605 tc = tc['playlist
'][0]
3606 is_restricted = age_restricted(
3607 tc.get('info_dict
', {}).get('age_limit
'), age_limit)
3608 if not is_restricted:
3610 any_restricted = any_restricted or is_restricted
3611 return not any_restricted
3613 def extract_subtitles(self, *args, **kwargs):
3614 if (self.get_param('writesubtitles
', False)
3615 or self.get_param('listsubtitles
')):
3616 return self._get_subtitles(*args, **kwargs)
3619 def _get_subtitles(self, *args, **kwargs):
3620 raise NotImplementedError('This method must be implemented by subclasses
')
3622 def extract_comments(self, *args, **kwargs):
3623 if not self.get_param('getcomments
'):
3625 generator = self._get_comments(*args, **kwargs)
3632 comments.append(next(generator))
3633 except StopIteration:
3635 except KeyboardInterrupt:
3636 self.to_screen('Interrupted by user
')
3637 except Exception as e:
3638 if self.get_param('ignoreerrors
') is not True:
3640 self._downloader.report_error(e)
3641 comment_count = len(comments)
3642 self.to_screen(f'Extracted {comment_count} comments
')
3644 'comments
': comments,
3645 'comment_count
': None if interrupted else comment_count
3649 def _get_comments(self, *args, **kwargs):
3650 raise NotImplementedError('This method must be implemented by subclasses
')
3653 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
3654 """ Merge subtitle items for one language. Items with duplicated URLs
3655 will be dropped. """
3656 list1_urls = set([item['url
'] for item in subtitle_list1])
3657 ret = list(subtitle_list1)
3658 ret.extend([item for item in subtitle_list2 if item['url
'] not in list1_urls])
3662 def _merge_subtitles(cls, *dicts, target=None):
3663 """ Merge subtitle dictionaries, language by language. """
3667 for lang, subs in d.items():
3668 target[lang] = cls._merge_subtitle_items(target.get(lang, []), subs)
3671 def extract_automatic_captions(self, *args, **kwargs):
3672 if (self.get_param('writeautomaticsub
', False)
3673 or self.get_param('listsubtitles
')):
3674 return self._get_automatic_captions(*args, **kwargs)
3677 def _get_automatic_captions(self, *args, **kwargs):
3678 raise NotImplementedError('This method must be implemented by subclasses
')
3680 def mark_watched(self, *args, **kwargs):
3681 if not self.get_param('mark_watched
', False):
3683 if (hasattr(self, '_NETRC_MACHINE
') and self._get_login_info()[0] is not None
3684 or self.get_param('cookiefile
')
3685 or self.get_param('cookiesfrombrowser
')):
3686 self._mark_watched(*args, **kwargs)
3688 def _mark_watched(self, *args, **kwargs):
3689 raise NotImplementedError('This method must be implemented by subclasses
')
3691 def geo_verification_headers(self):
3693 geo_verification_proxy = self.get_param('geo_verification_proxy
')
3694 if geo_verification_proxy:
3695 headers['Ytdl
-request
-proxy
'] = geo_verification_proxy
3698 def _generic_id(self, url):
3699 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
3701 def _generic_title(self, url):
3702 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
3705 def _availability(is_private=None, needs_premium=None, needs_subscription=None, needs_auth=None, is_unlisted=None):
3706 all_known = all(map(
3707 lambda x: x is not None,
3708 (is_private, needs_premium, needs_subscription, needs_auth, is_unlisted)))
3710 'private
' if is_private
3711 else 'premium_only
' if needs_premium
3712 else 'subscriber_only
' if needs_subscription
3713 else 'needs_auth
' if needs_auth
3714 else 'unlisted
' if is_unlisted
3715 else 'public
' if all_known
3718 def _configuration_arg(self, key, default=NO_DEFAULT, *, ie_key=None, casesense=False):
3720 @returns A list of values for the extractor argument given by "key"
3721 or "default" if no such key is present
3722 @param default The default value to return when the key is not present (default: [])
3723 @param casesense When false, the values are converted to lower case
3726 self._downloader.params, ('extractor_args
', (ie_key or self.ie_key()).lower(), key))
3728 return [] if default is NO_DEFAULT else default
3729 return list(val) if casesense else [x.lower() for x in val]
3731 def _yes_playlist(self, playlist_id, video_id, smuggled_data=None, *, playlist_label='playlist
', video_label='video
'):
3732 if not playlist_id or not video_id:
3735 no_playlist = (smuggled_data or {}).get('force_noplaylist
')
3736 if no_playlist is not None:
3737 return not no_playlist
3739 video_id = '' if video_id is True else f' {video_id}
'
3740 playlist_id = '' if playlist_id is True else f' {playlist_id}
'
3741 if self.get_param('noplaylist
'):
3742 self.to_screen(f'Downloading just the {video_label}{video_id} because of
--no
-playlist
')
3744 self.to_screen(f'Downloading {playlist_label}{playlist_id}
- add
--no
-playlist to download just the {video_label}{video_id}
')
3748 class SearchInfoExtractor(InfoExtractor):
3750 Base class for paged search queries extractors.
3751 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
3752 Instances should define _SEARCH_KEY and optionally _MAX_RESULTS
3755 _MAX_RESULTS = float('inf
')
3758 def _make_valid_url(cls):
3759 return r'%s(?P
<prefix
>|
[1-9][0-9]*|all
):(?P
<query
>[\s\S
]+)' % cls._SEARCH_KEY
3761 def _real_extract(self, query):
3762 prefix, query = self._match_valid_url(query).group('prefix
', 'query
')
3764 return self._get_n_results(query, 1)
3765 elif prefix == 'all
':
3766 return self._get_n_results(query, self._MAX_RESULTS)
3770 raise ExtractorError('invalid download number
%s for query
"%s"' % (n, query))
3771 elif n > self._MAX_RESULTS:
3772 self.report_warning('%s returns
max %i results (you requested
%i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3773 n = self._MAX_RESULTS
3774 return self._get_n_results(query, n)
3776 def _get_n_results(self, query, n):
3777 """Get a specified number of results for a query.
3778 Either this function or _search_results must be overridden by subclasses """
3779 return self.playlist_result(
3780 itertools.islice(self._search_results(query), 0, None if n == float('inf
') else n),
3783 def _search_results(self, query):
3784 """Returns an iterator of search results"""
3785 raise NotImplementedError('This method must be implemented by subclasses
')
3788 def SEARCH_KEY(self):
3789 return self._SEARCH_KEY