20 import xml
.etree
.ElementTree
22 from ..compat
import functools
# isort: split
23 from ..compat
import compat_etree_fromstring
, compat_expanduser
, compat_os_name
24 from ..downloader
import FileDownloader
25 from ..downloader
.f4m
import get_base_url
, remove_encrypted_media
62 parse_m3u8_attributes
,
89 """Information Extractor class.
91 Information extractors are the classes that, given a URL, extract
92 information about the video (or videos) the URL refers to. This
93 information includes the real video URL, the video title, author and
94 others. The information is stored in a dictionary which is then
95 passed to the YoutubeDL. The YoutubeDL processes this
96 information possibly downloading the video to the file system, among
97 other possible outcomes.
99 The type field determines the type of the result.
100 By far the most common value (and the default if _type is missing) is
101 "video", which indicates a single video.
103 For a video, the dictionaries must include the following fields:
105 id: Video identifier.
106 title: Video title, unescaped. Set to an empty string if video has
107 no title as opposed to "None" which signifies that the
108 extractor failed to obtain a title
110 Additionally, it must contain either a formats entry or a url one:
112 formats: A list of dictionaries for each format available, ordered
113 from worst to best quality.
116 * url The mandatory URL representing the media:
117 for plain file media - HTTP URL of this file,
119 for HLS - URL of the M3U8 media playlist,
120 for HDS - URL of the F4M manifest,
122 - HTTP URL to plain file media (in case of
124 - URL of the MPD manifest or base URL
125 representing the media if MPD manifest
126 is parsed from a string (in case of
128 for MSS - URL of the ISM manifest.
130 The URL of the manifest file in case of
132 for HLS - URL of the M3U8 master playlist,
133 for HDS - URL of the F4M manifest,
134 for DASH - URL of the MPD manifest,
135 for MSS - URL of the ISM manifest.
136 * manifest_stream_number (For internal use only)
137 The index of the stream in the manifest file
138 * ext Will be calculated from URL if missing
139 * format A human-readable description of the format
140 ("mp4 container with h264/opus").
141 Calculated from the format_id, width, height.
142 and format_note fields if missing.
143 * format_id A short description of the format
144 ("mp4_h264_opus" or "19").
145 Technically optional, but strongly recommended.
146 * format_note Additional info about the format
147 ("3D" or "DASH video")
148 * width Width of the video, if known
149 * height Height of the video, if known
150 * resolution Textual description of width and height
151 * dynamic_range The dynamic range of the video. One of:
152 "SDR" (None), "HDR10", "HDR10+, "HDR12", "HLG, "DV"
153 * tbr Average bitrate of audio and video in KBit/s
154 * abr Average audio bitrate in KBit/s
155 * acodec Name of the audio codec in use
156 * asr Audio sampling rate in Hertz
157 * audio_channels Number of audio channels
158 * vbr Average video bitrate in KBit/s
160 * vcodec Name of the video codec in use
161 * container Name of the container format
162 * filesize The number of bytes, if known in advance
163 * filesize_approx An estimate for the number of bytes
164 * player_url SWF Player URL (used for rtmpdump).
165 * protocol The protocol that will be used for the actual
166 download, lower-case. One of "http", "https" or
167 one of the protocols defined in downloader.PROTOCOL_MAP
169 Base URL for fragments. Each fragment's path
170 value (if present) will be relative to
172 * fragments A list of fragments of a fragmented media.
173 Each fragment entry must contain either an url
174 or a path. If an url is present it should be
175 considered by a client. Otherwise both path and
176 fragment_base_url must be present. Here is
177 the list of all potential fields:
178 * "url" - fragment's URL
179 * "path" - fragment's path relative to
181 * "duration" (optional, int or float)
182 * "filesize" (optional, int)
183 * is_from_start Is a live format that can be downloaded
184 from the start. Boolean
185 * preference Order number of this format. If this field is
186 present and not None, the formats get sorted
187 by this field, regardless of all other values.
188 -1 for default (order by other properties),
189 -2 or smaller for less than default.
190 < -1000 to hide the format (if there is
191 another one which is strictly better)
192 * language Language code, e.g. "de" or "en-US".
193 * language_preference Is this in the language mentioned in
195 10 if it's what the URL is about,
196 -1 for default (don't know),
197 -10 otherwise, other values reserved for now.
198 * quality Order number of the video quality of this
199 format, irrespective of the file format.
200 -1 for default (order by other properties),
201 -2 or smaller for less than default.
202 * source_preference Order number for this video source
203 (quality takes higher priority)
204 -1 for default (order by other properties),
205 -2 or smaller for less than default.
206 * http_headers A dictionary of additional HTTP headers
207 to add to the request.
208 * stretched_ratio If given and not 1, indicates that the
209 video's pixels are not square.
210 width : height ratio as float.
211 * no_resume The server does not support resuming the
212 (HTTP or RTMP) download. Boolean.
213 * has_drm The format has DRM and cannot be downloaded. Boolean
214 * downloader_options A dictionary of downloader options
215 (For internal use only)
216 * http_chunk_size Chunk size for HTTP downloads
217 * ffmpeg_args Extra arguments for ffmpeg downloader
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 in UTC (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 in UTC.
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 in UTC.
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: None (=unknown), 'is_live', 'is_upcoming', 'was_live', 'not_live',
321 or 'post_live' (was live, but VOD is not yet processed)
322 If absent, automatically set from is_live, was_live
323 start_time: Time in seconds where the reproduction should start, as
324 specified in the URL.
325 end_time: Time in seconds where the reproduction should end, as
326 specified in the URL.
327 chapters: A list of dictionaries, with the following entries:
328 * "start_time" - The start time of the chapter in seconds
329 * "end_time" - The end time of the chapter in seconds
330 * "title" (optional, string)
331 playable_in_embed: Whether this video is allowed to play in embedded
332 players on other sites. Can be True (=always allowed),
333 False (=never allowed), None (=unknown), or a string
334 specifying the criteria for embedability; e.g. 'whitelist'
335 availability: Under what condition the video is available. One of
336 'private', 'premium_only', 'subscriber_only', 'needs_auth',
337 'unlisted' or 'public'. Use 'InfoExtractor._availability'
339 _old_archive_ids: A list of old archive ids needed for backward compatibility
340 __post_extractor: A function to be called just before the metadata is
341 written to either disk, logger or console. The function
342 must return a dict which will be added to the info_dict.
343 This is usefull for additional information that is
344 time-consuming to extract. Note that the fields thus
345 extracted will not be available to output template and
346 match_filter. So, only "comments" and "comment_count" are
347 currently allowed to be extracted via this method.
349 The following fields should only be used when the video belongs to some logical
352 chapter: Name or title of the chapter the video belongs to.
353 chapter_number: Number of the chapter the video belongs to, as an integer.
354 chapter_id: Id of the chapter the video belongs to, as a unicode string.
356 The following fields should only be used when the video is an episode of some
357 series, programme or podcast:
359 series: Title of the series or programme the video episode belongs to.
360 series_id: Id of the series or programme the video episode belongs to, as a unicode string.
361 season: Title of the season the video episode belongs to.
362 season_number: Number of the season the video episode belongs to, as an integer.
363 season_id: Id of the season the video episode belongs to, as a unicode string.
364 episode: Title of the video episode. Unlike mandatory video title field,
365 this field should denote the exact title of the video episode
366 without any kind of decoration.
367 episode_number: Number of the video episode within a season, as an integer.
368 episode_id: Id of the video episode, as a unicode string.
370 The following fields should only be used when the media is a track or a part of
373 track: Title of the track.
374 track_number: Number of the track within an album or a disc, as an integer.
375 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
377 artist: Artist(s) of the track.
378 genre: Genre(s) of the track.
379 album: Title of the album the track belongs to.
380 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
381 album_artist: List of all artists appeared on the album (e.g.
382 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
384 disc_number: Number of the disc or other physical medium the track belongs to,
386 release_year: Year (YYYY) when the album was released.
387 composer: Composer of the piece
389 The following fields should only be set for clips that should be cut from the original video:
391 section_start: Start time of the section in seconds
392 section_end: End time of the section in seconds
394 The following fields should only be set for storyboards:
395 rows: Number of rows in each storyboard fragment, as an integer
396 columns: Number of columns in each storyboard fragment, as an integer
398 Unless mentioned otherwise, the fields should be Unicode strings.
400 Unless mentioned otherwise, None is equivalent to absence of information.
403 _type "playlist" indicates multiple videos.
404 There must be a key "entries", which is a list, an iterable, or a PagedList
405 object, each element of which is a valid dictionary by this specification.
407 Additionally, playlists can have "id", "title", and any other relevant
408 attributes with the same semantics as videos (see above).
410 It can also have the following optional fields:
412 playlist_count: The total number of videos in a playlist. If not given,
413 YoutubeDL tries to calculate it from "entries"
416 _type "multi_video" indicates that there are multiple videos that
417 form a single show, for examples multiple acts of an opera or TV episode.
418 It must have an entries key like a playlist and contain all the keys
419 required for a video at the same time.
422 _type "url" indicates that the video must be extracted from another
423 location, possibly by a different extractor. Its only required key is:
424 "url" - the next URL to extract.
425 The key "ie_key" can be set to the class name (minus the trailing "IE",
426 e.g. "Youtube") if the extractor class is known in advance.
427 Additionally, the dictionary may have any properties of the resolved entity
428 known in advance, for example "title" if the title of the referred video is
432 _type "url_transparent" entities have the same specification as "url", but
433 indicate that the given additional information is more precise than the one
434 associated with the resolved URL.
435 This is useful when a site employs a video service that hosts the video and
436 its technical metadata, but that video service does not embed a useful
437 title, description etc.
440 Subclasses of this should also be added to the list of extractors and
441 should define a _VALID_URL regexp and, re-define the _real_extract() and
442 (optionally) _real_initialize() methods.
444 Subclasses may also override suitable() if necessary, but ensure the function
445 signature is preserved and that this function imports everything it needs
446 (except other extractors), so that lazy_extractors works correctly.
448 Subclasses can define a list of _EMBED_REGEX, which will be searched for in
449 the HTML of Generic webpages. It may also override _extract_embed_urls
450 or _extract_from_webpage as necessary. While these are normally classmethods,
451 _extract_from_webpage is allowed to be an instance method.
453 _extract_from_webpage may raise self.StopExtraction() to stop further
454 processing of the webpage and obtain exclusive rights to it. This is useful
455 when the extractor cannot reliably be matched using just the URL,
456 e.g. invidious/peertube instances
458 Embed-only extractors can be defined by setting _VALID_URL = False.
460 To support username + password (or netrc) login, the extractor must define a
461 _NETRC_MACHINE and re-define _perform_login(username, password) and
462 (optionally) _initialize_pre_login() methods. The _perform_login method will
463 be called between _initialize_pre_login and _real_initialize if credentials
464 are passed by the user. In cases where it is necessary to have the login
465 process as part of the extraction rather than initialization, _perform_login
466 can be left undefined.
468 _GEO_BYPASS attribute may be set to False in order to disable
469 geo restriction bypass mechanisms for a particular extractor.
470 Though it won't disable explicit geo restriction bypass based on
471 country code provided with geo_bypass_country.
473 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
474 countries for this extractor. One of these countries will be used by
475 geo restriction bypass mechanism right away in order to bypass
476 geo restriction, of course, if the mechanism is not disabled.
478 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
479 IP blocks in CIDR notation for this extractor. One of these IP blocks
480 will be used by geo restriction bypass mechanism similarly
483 The _WORKING attribute should be set to False for broken IEs
484 in order to warn the users and skip the tests.
489 _x_forwarded_for_ip
= None
491 _GEO_COUNTRIES
= None
492 _GEO_IP_BLOCKS
= None
494 _NETRC_MACHINE
= None
500 def _login_hint(self
, method
=NO_DEFAULT
, netrc
=None):
501 password_hint
= f
'--username and --password, or --netrc ({netrc or self._NETRC_MACHINE}) to provide account credentials'
504 'any': f
'Use --cookies, --cookies-from-browser, {password_hint}',
505 'password': f
'Use {password_hint}',
507 'Use --cookies-from-browser or --cookies for the authentication. '
508 'See https://github.com/ytdl-org/youtube-dl#how-do-i-pass-cookies-to-youtube-dl for how to manually pass cookies'),
509 }[method
if method
is not NO_DEFAULT
else 'any' if self
.supports_login() else 'cookies']
511 def __init__(self
, downloader
=None):
512 """Constructor. Receives an optional downloader (a YoutubeDL instance).
513 If a downloader is not passed during initialization,
514 it must be set using "set_downloader()" before "extract()" is called"""
516 self
._x
_forwarded
_for
_ip
= None
517 self
._printed
_messages
= set()
518 self
.set_downloader(downloader
)
521 def _match_valid_url(cls
, url
):
522 if cls
._VALID
_URL
is False:
524 # This does not use has/getattr intentionally - we want to know whether
525 # we have cached the regexp for *this* class, whereas getattr would also
526 # match the superclass
527 if '_VALID_URL_RE' not in cls
.__dict
__:
528 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
529 return cls
._VALID
_URL
_RE
.match(url
)
532 def suitable(cls
, url
):
533 """Receives a URL and returns True if suitable for this IE."""
534 # This function must import everything it needs (except other extractors),
535 # so that lazy_extractors works correctly
536 return cls
._match
_valid
_url
(url
) is not None
539 def _match_id(cls
, url
):
540 return cls
._match
_valid
_url
(url
).group('id')
543 def get_temp_id(cls
, url
):
545 return cls
._match
_id
(url
)
546 except (IndexError, AttributeError):
551 """Getter method for _WORKING."""
555 def supports_login(cls
):
556 return bool(cls
._NETRC
_MACHINE
)
558 def initialize(self
):
559 """Initializes an instance (authentication, etc)."""
560 self
._printed
_messages
= set()
561 self
._initialize
_geo
_bypass
({
562 'countries': self
._GEO
_COUNTRIES
,
563 'ip_blocks': self
._GEO
_IP
_BLOCKS
,
566 self
._initialize
_pre
_login
()
567 if self
.supports_login():
568 username
, password
= self
._get
_login
_info
()
570 self
._perform
_login
(username
, password
)
571 elif self
.get_param('username') and False not in (self
.IE_DESC
, self
._NETRC
_MACHINE
):
572 self
.report_warning(f
'Login with password is not supported for this website. {self._login_hint("cookies")}')
573 self
._real
_initialize
()
576 def _initialize_geo_bypass(self
, geo_bypass_context
):
578 Initialize geo restriction bypass mechanism.
580 This method is used to initialize geo bypass mechanism based on faking
581 X-Forwarded-For HTTP header. A random country from provided country list
582 is selected and a random IP belonging to this country is generated. This
583 IP will be passed as X-Forwarded-For HTTP header in all subsequent
586 This method will be used for initial geo bypass mechanism initialization
587 during the instance initialization with _GEO_COUNTRIES and
590 You may also manually call it from extractor's code if geo bypass
591 information is not available beforehand (e.g. obtained during
592 extraction) or due to some other reason. In this case you should pass
593 this information in geo bypass context passed as first argument. It may
594 contain following fields:
596 countries: List of geo unrestricted countries (similar
598 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
599 (similar to _GEO_IP_BLOCKS)
602 if not self
._x
_forwarded
_for
_ip
:
604 # Geo bypass mechanism is explicitly disabled by user
605 if not self
.get_param('geo_bypass', True):
608 if not geo_bypass_context
:
609 geo_bypass_context
= {}
611 # Backward compatibility: previously _initialize_geo_bypass
612 # expected a list of countries, some 3rd party code may still use
614 if isinstance(geo_bypass_context
, (list, tuple)):
615 geo_bypass_context
= {
616 'countries': geo_bypass_context
,
619 # The whole point of geo bypass mechanism is to fake IP
620 # as X-Forwarded-For HTTP header based on some IP block or
623 # Path 1: bypassing based on IP block in CIDR notation
625 # Explicit IP block specified by user, use it right away
626 # regardless of whether extractor is geo bypassable or not
627 ip_block
= self
.get_param('geo_bypass_ip_block', None)
629 # Otherwise use random IP block from geo bypass context but only
630 # if extractor is known as geo bypassable
632 ip_blocks
= geo_bypass_context
.get('ip_blocks')
633 if self
._GEO
_BYPASS
and ip_blocks
:
634 ip_block
= random
.choice(ip_blocks
)
637 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(ip_block
)
638 self
.write_debug(f
'Using fake IP {self._x_forwarded_for_ip} as X-Forwarded-For')
641 # Path 2: bypassing based on country code
643 # Explicit country code specified by user, use it right away
644 # regardless of whether extractor is geo bypassable or not
645 country
= self
.get_param('geo_bypass_country', None)
647 # Otherwise use random country code from geo bypass context but
648 # only if extractor is known as geo bypassable
650 countries
= geo_bypass_context
.get('countries')
651 if self
._GEO
_BYPASS
and countries
:
652 country
= random
.choice(countries
)
655 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country
)
656 self
._downloader
.write_debug(
657 f
'Using fake IP {self._x_forwarded_for_ip} ({country.upper()}) as X-Forwarded-For')
659 def extract(self
, url
):
660 """Extracts URL information and returns it in list of dicts."""
665 self
.write_debug('Extracting URL: %s' % url
)
666 ie_result
= self
._real
_extract
(url
)
667 if ie_result
is None:
669 if self
._x
_forwarded
_for
_ip
:
670 ie_result
['__x_forwarded_for_ip'] = self
._x
_forwarded
_for
_ip
671 subtitles
= ie_result
.get('subtitles') or {}
672 if 'no-live-chat' in self
.get_param('compat_opts'):
673 for lang
in ('live_chat', 'comments', 'danmaku'):
674 subtitles
.pop(lang
, None)
676 except GeoRestrictedError
as e
:
677 if self
.__maybe
_fake
_ip
_and
_retry
(e
.countries
):
680 except UnsupportedError
:
682 except ExtractorError
as e
:
684 'video_id': e
.video_id
or self
.get_temp_id(url
),
686 'tb': e
.traceback
or sys
.exc_info()[2],
687 'expected': e
.expected
,
690 if hasattr(e
, 'countries'):
691 kwargs
['countries'] = e
.countries
692 raise type(e
)(e
.orig_msg
, **kwargs
)
693 except http
.client
.IncompleteRead
as e
:
694 raise ExtractorError('A network error has occurred.', cause
=e
, expected
=True, video_id
=self
.get_temp_id(url
))
695 except (KeyError, StopIteration) as e
:
696 raise ExtractorError('An extractor error has occurred.', cause
=e
, video_id
=self
.get_temp_id(url
))
698 def __maybe_fake_ip_and_retry(self
, countries
):
699 if (not self
.get_param('geo_bypass_country', None)
701 and self
.get_param('geo_bypass', True)
702 and not self
._x
_forwarded
_for
_ip
704 country_code
= random
.choice(countries
)
705 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country_code
)
706 if self
._x
_forwarded
_for
_ip
:
708 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
709 % (self
._x
_forwarded
_for
_ip
, country_code
.upper()))
713 def set_downloader(self
, downloader
):
714 """Sets a YoutubeDL instance as the downloader for this IE."""
715 self
._downloader
= downloader
719 return self
._downloader
.cache
723 return self
._downloader
.cookiejar
725 def _initialize_pre_login(self
):
726 """ Initialization before login. Redefine in subclasses."""
729 def _perform_login(self
, username
, password
):
730 """ Login with username and password. Redefine in subclasses."""
733 def _real_initialize(self
):
734 """Real initialization process. Redefine in subclasses."""
737 def _real_extract(self
, url
):
738 """Real extraction process. Redefine in subclasses."""
739 raise NotImplementedError('This method must be implemented by subclasses')
743 """A string for getting the InfoExtractor with get_info_extractor"""
744 return cls
.__name
__[:-2]
748 return cls
.__name
__[:-2]
751 def __can_accept_status_code(err
, expected_status
):
752 assert isinstance(err
, urllib
.error
.HTTPError
)
753 if expected_status
is None:
755 elif callable(expected_status
):
756 return expected_status(err
.code
) is True
758 return err
.code
in variadic(expected_status
)
760 def _create_request(self
, url_or_request
, data
=None, headers
=None, query
=None):
761 if isinstance(url_or_request
, urllib
.request
.Request
):
762 return update_Request(url_or_request
, data
=data
, headers
=headers
, query
=query
)
764 url_or_request
= update_url_query(url_or_request
, query
)
765 return sanitized_Request(url_or_request
, data
, headers
or {})
767 def _request_webpage(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, data
=None, headers
=None, query
=None, expected_status
=None):
769 Return the response handle.
771 See _download_webpage docstring for arguments specification.
773 if not self
._downloader
._first
_webpage
_request
:
774 sleep_interval
= self
.get_param('sleep_interval_requests') or 0
775 if sleep_interval
> 0:
776 self
.to_screen('Sleeping %s seconds ...' % sleep_interval
)
777 time
.sleep(sleep_interval
)
779 self
._downloader
._first
_webpage
_request
= False
782 self
.report_download_webpage(video_id
)
783 elif note
is not False:
785 self
.to_screen(str(note
))
787 self
.to_screen(f
'{video_id}: {note}')
789 # Some sites check X-Forwarded-For HTTP header in order to figure out
790 # the origin of the client behind proxy. This allows bypassing geo
791 # restriction by faking this header's value to IP that belongs to some
792 # geo unrestricted country. We will do so once we encounter any
793 # geo restriction error.
794 if self
._x
_forwarded
_for
_ip
:
795 headers
= (headers
or {}).copy()
796 headers
.setdefault('X-Forwarded-For', self
._x
_forwarded
_for
_ip
)
799 return self
._downloader
.urlopen(self
._create
_request
(url_or_request
, data
, headers
, query
))
800 except network_exceptions
as err
:
801 if isinstance(err
, urllib
.error
.HTTPError
):
802 if self
.__can
_accept
_status
_code
(err
, expected_status
):
803 # Retain reference to error to prevent file object from
804 # being closed before it can be read. Works around the
805 # effects of <https://bugs.python.org/issue15002>
806 # introduced in Python 3.4.1.
813 errnote
= 'Unable to download webpage'
815 errmsg
= f
'{errnote}: {error_to_compat_str(err)}'
817 raise ExtractorError(errmsg
, cause
=err
)
819 self
.report_warning(errmsg
)
822 def _download_webpage_handle(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True,
823 encoding
=None, data
=None, headers
={}, query={}
, expected_status
=None):
825 Return a tuple (page content as string, URL handle).
828 url_or_request -- plain text URL as a string or
829 a urllib.request.Request object
830 video_id -- Video/playlist/item identifier (string)
833 note -- note printed before downloading (string)
834 errnote -- note printed in case of an error (string)
835 fatal -- flag denoting whether error should be considered fatal,
836 i.e. whether it should cause ExtractionError to be raised,
837 otherwise a warning will be reported and extraction continued
838 encoding -- encoding for a page content decoding, guessed automatically
839 when not explicitly specified
840 data -- POST data (bytes)
841 headers -- HTTP headers (dict)
842 query -- URL query (dict)
843 expected_status -- allows to accept failed HTTP requests (non 2xx
844 status code) by explicitly specifying a set of accepted status
845 codes. Can be any of the following entities:
846 - an integer type specifying an exact failed status code to
848 - a list or a tuple of integer types specifying a list of
849 failed status codes to accept
850 - a callable accepting an actual failed status code and
851 returning True if it should be accepted
852 Note that this argument does not affect success status codes (2xx)
853 which are always accepted.
856 # Strip hashes from the URL (#1038)
857 if isinstance(url_or_request
, str):
858 url_or_request
= url_or_request
.partition('#')[0]
860 urlh
= self
._request
_webpage
(url_or_request
, video_id
, note
, errnote
, fatal
, data
=data
, headers
=headers
, query
=query
, expected_status
=expected_status
)
864 content
= self
._webpage
_read
_content
(urlh
, url_or_request
, video_id
, note
, errnote
, fatal
, encoding
=encoding
)
865 return (content
, urlh
)
868 def _guess_encoding_from_content(content_type
, webpage_bytes
):
869 m
= re
.match(r
'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type
)
871 encoding
= m
.group(1)
873 m
= re
.search(br
'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
874 webpage_bytes[:1024])
876 encoding = m.group(1).decode('ascii')
877 elif webpage_bytes.startswith(b'\xff\xfe'):
884 def __check_blocked(self, content):
885 first_block = content[:512]
886 if ('<title>Access to this site is blocked</title>' in content
887 and 'Websense' in first_block):
888 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
889 blocked_iframe = self._html_search_regex(
890 r'<iframe src="([^
"]+)"', content,
891 'Websense information URL
', default=None)
893 msg += ' Visit
%s for more details
' % blocked_iframe
894 raise ExtractorError(msg, expected=True)
895 if '<title
>The URL you requested has been blocked
</title
>' in first_block:
897 'Access to this webpage has been blocked by Indian censorship
. '
898 'Use a VPN
or proxy
server (with --proxy
) to route around it
.')
899 block_msg = self._html_search_regex(
900 r'</h1
><p
>(.*?
)</p
>',
901 content, 'block message
', default=None)
903 msg += ' (Message
: "%s")' % block_msg.replace('\n', ' ')
904 raise ExtractorError(msg, expected=True)
905 if ('<title
>TTK
:: Доступ к ресурсу ограничен
</title
>' in content
906 and 'blocklist
.rkn
.gov
.ru
' in content):
907 raise ExtractorError(
908 'Access to this webpage has been blocked by decision of the Russian government
. '
909 'Visit http
://blocklist
.rkn
.gov
.ru
/ for a block reason
.',
912 def _request_dump_filename(self, url, video_id):
913 basen = f'{video_id}_{url}
'
914 trim_length = self.get_param('trim_file_name
') or 240
915 if len(basen) > trim_length:
916 h = '___
' + hashlib.md5(basen.encode('utf
-8')).hexdigest()
917 basen = basen[:trim_length - len(h)] + h
918 filename = sanitize_filename(f'{basen}
.dump
', restricted=True)
919 # Working around MAX_PATH limitation on Windows (see
920 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
921 if compat_os_name == 'nt
':
922 absfilepath = os.path.abspath(filename)
923 if len(absfilepath) > 259:
924 filename = fR'\\?\{absfilepath}
'
927 def __decode_webpage(self, webpage_bytes, encoding, headers):
929 encoding = self._guess_encoding_from_content(headers.get('Content
-Type
', ''), webpage_bytes)
931 return webpage_bytes.decode(encoding, 'replace
')
933 return webpage_bytes.decode('utf
-8', 'replace
')
935 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
936 webpage_bytes = urlh.read()
937 if prefix is not None:
938 webpage_bytes = prefix + webpage_bytes
939 if self.get_param('dump_intermediate_pages
', False):
940 self.to_screen('Dumping request to
' + urlh.geturl())
941 dump = base64.b64encode(webpage_bytes).decode('ascii
')
942 self._downloader.to_screen(dump)
943 if self.get_param('write_pages
'):
944 filename = self._request_dump_filename(urlh.geturl(), video_id)
945 self.to_screen(f'Saving request to {filename}
')
946 with open(filename, 'wb
') as outf:
947 outf.write(webpage_bytes)
949 content = self.__decode_webpage(webpage_bytes, encoding, urlh.headers)
950 self.__check_blocked(content)
954 def __print_error(self, errnote, fatal, video_id, err):
956 raise ExtractorError(f'{video_id}
: {errnote}
', cause=err)
958 self.report_warning(f'{video_id}
: {errnote}
: {err}
')
960 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True, errnote=None):
962 xml_string = transform_source(xml_string)
964 return compat_etree_fromstring(xml_string.encode('utf
-8'))
965 except xml.etree.ElementTree.ParseError as ve:
966 self.__print_error('Failed to parse XML
' if errnote is None else errnote, fatal, video_id, ve)
968 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True, errnote=None, **parser_kwargs):
971 json_string, cls=LenientJSONDecoder, strict=False, transform_source=transform_source, **parser_kwargs)
972 except ValueError as ve:
973 self.__print_error('Failed to parse JSON
' if errnote is None else errnote, fatal, video_id, ve)
975 def _parse_socket_response_as_json(self, data, *args, **kwargs):
976 return self._parse_json(data[data.find('{'):data.rfind('}
') + 1], *args, **kwargs)
978 def __create_download_methods(name, parser, note, errnote, return_value):
980 def parse(ie, content, *args, errnote=errnote, **kwargs):
984 kwargs['errnote
'] = errnote
985 # parser is fetched by name so subclasses can override it
986 return getattr(ie, parser)(content, *args, **kwargs)
988 def download_handle(self, url_or_request, video_id, note=note, errnote=errnote, transform_source=None,
989 fatal=True, encoding=None, data=None, headers={}, query={}, expected_status=None):
990 res = self._download_webpage_handle(
991 url_or_request, video_id, note=note, errnote=errnote, fatal=fatal, encoding=encoding,
992 data=data, headers=headers, query=query, expected_status=expected_status)
996 return parse(self, content, video_id, transform_source=transform_source, fatal=fatal, errnote=errnote), urlh
998 def download_content(self, url_or_request, video_id, note=note, errnote=errnote, transform_source=None,
999 fatal=True, encoding=None, data=None, headers={}, query={}, expected_status=None):
1000 if self.get_param('load_pages
'):
1001 url_or_request = self._create_request(url_or_request, data, headers, query)
1002 filename = self._request_dump_filename(url_or_request.full_url, video_id)
1003 self.to_screen(f'Loading request
from {filename}
')
1005 with open(filename, 'rb
') as dumpf:
1006 webpage_bytes = dumpf.read()
1007 except OSError as e:
1008 self.report_warning(f'Unable to load request
from disk
: {e}
')
1010 content = self.__decode_webpage(webpage_bytes, encoding, url_or_request.headers)
1011 return parse(self, content, video_id, transform_source=transform_source, fatal=fatal, errnote=errnote)
1015 'transform_source
': transform_source,
1017 'encoding
': encoding,
1021 'expected_status
': expected_status,
1024 kwargs.pop('transform_source
')
1025 # The method is fetched by name so subclasses can override _download_..._handle
1026 res = getattr(self, download_handle.__name__)(url_or_request, video_id, **kwargs)
1027 return res if res is False else res[0]
1029 def impersonate(func, name, return_value):
1030 func.__name__, func.__qualname__ = name, f'InfoExtractor
.{name}
'
1032 @param transform_source Apply this transformation before parsing
1033 @returns {return_value}
1035 See _download_webpage_handle docstring for other arguments specification
1038 impersonate(download_handle, f'_download_{name}_handle
', f'({return_value}
, URL handle
)')
1039 impersonate(download_content, f'_download_{name}
', f'{return_value}
')
1040 return download_handle, download_content
1042 _download_xml_handle, _download_xml = __create_download_methods(
1043 'xml
', '_parse_xml
', 'Downloading XML
', 'Unable to download XML
', 'xml
as an xml
.etree
.ElementTree
.Element
')
1044 _download_json_handle, _download_json = __create_download_methods(
1045 'json
', '_parse_json
', 'Downloading JSON metadata
', 'Unable to download JSON metadata
', 'JSON
object as a
dict')
1046 _download_socket_json_handle, _download_socket_json = __create_download_methods(
1047 'socket_json
', '_parse_socket_response_as_json
', 'Polling socket
', 'Unable to poll socket
', 'JSON
object as a
dict')
1048 __download_webpage = __create_download_methods('webpage
', None, None, None, 'data of the page
as a string
')[1]
1050 def _download_webpage(
1051 self, url_or_request, video_id, note=None, errnote=None,
1052 fatal=True, tries=1, timeout=NO_DEFAULT, *args, **kwargs):
1054 Return the data of the page as a string.
1057 tries -- number of tries
1058 timeout -- sleep interval between tries
1060 See _download_webpage_handle docstring for other arguments specification.
1063 R''' # NB: These are unused; should they be deprecated?
1065 self._downloader.deprecation_warning('tries argument
is deprecated
in InfoExtractor
._download
_webpage
')
1066 if timeout is NO_DEFAULT:
1069 self._downloader.deprecation_warning('timeout argument
is deprecated
in InfoExtractor
._download
_webpage
')
1075 return self.__download_webpage(url_or_request, video_id, note, errnote, None, fatal, *args, **kwargs)
1076 except http.client.IncompleteRead as e:
1078 if try_count >= tries:
1080 self._sleep(timeout, video_id)
1082 def report_warning(self, msg, video_id=None, *args, only_once=False, **kwargs):
1083 idstr = format_field(video_id, None, '%s: ')
1084 msg = f'[{self.IE_NAME}
] {idstr}{msg}
'
1086 if f'WARNING
: {msg}
' in self._printed_messages:
1088 self._printed_messages.add(f'WARNING
: {msg}
')
1089 self._downloader.report_warning(msg, *args, **kwargs)
1091 def to_screen(self, msg, *args, **kwargs):
1092 """Print msg to screen, prefixing it with '[ie_name
]'"""
1093 self._downloader.to_screen(f'[{self.IE_NAME}
] {msg}
', *args, **kwargs)
1095 def write_debug(self, msg, *args, **kwargs):
1096 self._downloader.write_debug(f'[{self.IE_NAME}
] {msg}
', *args, **kwargs)
1098 def get_param(self, name, default=None, *args, **kwargs):
1099 if self._downloader:
1100 return self._downloader.params.get(name, default, *args, **kwargs)
1103 def report_drm(self, video_id, partial=False):
1104 self.raise_no_formats('This video
is DRM protected
', expected=True, video_id=video_id)
1106 def report_extraction(self, id_or_name):
1107 """Report information extraction."""
1108 self.to_screen('%s: Extracting information
' % id_or_name)
1110 def report_download_webpage(self, video_id):
1111 """Report webpage download."""
1112 self.to_screen('%s: Downloading webpage
' % video_id)
1114 def report_age_confirmation(self):
1115 """Report attempt to confirm age."""
1116 self.to_screen('Confirming age
')
1118 def report_login(self):
1119 """Report attempt to log in."""
1120 self.to_screen('Logging
in')
1122 def raise_login_required(
1123 self, msg='This video
is only available
for registered users
',
1124 metadata_available=False, method=NO_DEFAULT):
1125 if metadata_available and (
1126 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1127 self.report_warning(msg)
1129 msg += format_field(self._login_hint(method), None, '. %s')
1130 raise ExtractorError(msg, expected=True)
1132 def raise_geo_restricted(
1133 self, msg='This video
is not available
from your location due to geo restriction
',
1134 countries=None, metadata_available=False):
1135 if metadata_available and (
1136 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1137 self.report_warning(msg)
1139 raise GeoRestrictedError(msg, countries=countries)
1141 def raise_no_formats(self, msg, expected=False, video_id=None):
1143 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1144 self.report_warning(msg, video_id)
1145 elif isinstance(msg, ExtractorError):
1148 raise ExtractorError(msg, expected=expected, video_id=video_id)
1150 # Methods for following #608
1152 def url_result(url, ie=None, video_id=None, video_title=None, *, url_transparent=False, **kwargs):
1153 """Returns a URL that points to a page that should be processed"""
1155 kwargs['ie_key
'] = ie if isinstance(ie, str) else ie.ie_key()
1156 if video_id is not None:
1157 kwargs['id'] = video_id
1158 if video_title is not None:
1159 kwargs['title
'] = video_title
1162 '_type
': 'url_transparent
' if url_transparent else 'url
',
1167 def playlist_from_matches(cls, matches, playlist_id=None, playlist_title=None,
1168 getter=IDENTITY, ie=None, video_kwargs=None, **kwargs):
1169 return cls.playlist_result(
1170 (cls.url_result(m, ie, **(video_kwargs or {})) for m in orderedSet(map(getter, matches), lazy=True)),
1171 playlist_id, playlist_title, **kwargs)
1174 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None, *, multi_video=False, **kwargs):
1175 """Returns a playlist"""
1177 kwargs['id'] = playlist_id
1179 kwargs['title
'] = playlist_title
1180 if playlist_description is not None:
1181 kwargs['description
'] = playlist_description
1184 '_type
': 'multi_video
' if multi_video else 'playlist
',
1188 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1190 Perform a regex search on the given string, using a single or a list of
1191 patterns returning the first matching group.
1192 In case of failure return a default value or raise a WARNING or a
1193 RegexNotFoundError, depending on fatal, specifying the field name.
1197 elif isinstance(pattern, (str, re.Pattern)):
1198 mobj = re.search(pattern, string, flags)
1201 mobj = re.search(p, string, flags)
1205 _name = self._downloader._format_err(name, self._downloader.Styles.EMPHASIS)
1209 # return the first matching group
1210 return next(g for g in mobj.groups() if g is not None)
1211 elif isinstance(group, (list, tuple)):
1212 return tuple(mobj.group(g) for g in group)
1214 return mobj.group(group)
1215 elif default is not NO_DEFAULT:
1218 raise RegexNotFoundError('Unable to extract
%s' % _name)
1220 self.report_warning('unable to extract
%s' % _name + bug_reports_message())
1223 def _search_json(self, start_pattern, string, name, video_id, *, end_pattern='',
1224 contains_pattern='(?s
:.+)', fatal=True, default=NO_DEFAULT, **kwargs):
1225 """Searches string for the JSON object specified by start_pattern"""
1226 # NB: end_pattern is only used to reduce the size of the initial match
1227 if default is NO_DEFAULT:
1228 default, has_default = {}, False
1230 fatal, has_default = False, True
1232 json_string = self._search_regex(
1233 rf'{start_pattern}\s
*(?P
<json
>{{\s*{contains_pattern}\s
*}})\s
*{end_pattern}
',
1234 string, name, group='json
', fatal=fatal, default=None if has_default else NO_DEFAULT)
1238 _name = self._downloader._format_err(name, self._downloader.Styles.EMPHASIS)
1240 return self._parse_json(json_string, video_id, ignore_extra=True, **kwargs)
1241 except ExtractorError as e:
1243 raise ExtractorError(
1244 f'Unable to extract {_name}
- Failed to parse JSON
', cause=e.cause, video_id=video_id)
1245 elif not has_default:
1246 self.report_warning(
1247 f'Unable to extract {_name}
- Failed to parse JSON
: {e}
', video_id=video_id)
1250 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1252 Like _search_regex, but strips HTML tags and unescapes entities.
1254 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1256 return clean_html(res).strip()
1260 def _get_netrc_login_info(self, netrc_machine=None):
1263 netrc_machine = netrc_machine or self._NETRC_MACHINE
1265 if self.get_param('usenetrc
', False):
1267 netrc_file = compat_expanduser(self.get_param('netrc_location
') or '~
')
1268 if os.path.isdir(netrc_file):
1269 netrc_file = os.path.join(netrc_file, '.netrc
')
1270 info = netrc.netrc(file=netrc_file).authenticators(netrc_machine)
1271 if info is not None:
1275 raise netrc.NetrcParseError(
1276 'No authenticators
for %s' % netrc_machine)
1277 except (OSError, netrc.NetrcParseError) as err:
1278 self.report_warning(
1279 'parsing
.netrc
: %s' % error_to_compat_str(err))
1281 return username, password
1283 def _get_login_info(self, username_option='username
', password_option='password
', netrc_machine=None):
1285 Get the login info as (username, password)
1286 First look for the manually specified credentials using username_option
1287 and password_option as keys in params dictionary. If no such credentials
1288 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1290 If there's no info available
, return (None, None)
1293 # Attempt to use provided username and password or .netrc data
1294 username = self.get_param(username_option)
1295 if username is not None:
1296 password = self.get_param(password_option)
1298 username, password = self._get_netrc_login_info(netrc_machine)
1300 return username, password
1302 def _get_tfa_info(self, note='two-factor verification code'):
1304 Get the two
-factor authentication info
1305 TODO
- asking the user will be required
for sms
/phone verify
1306 currently just uses the command line option
1307 If there
's no info available, return None
1310 tfa = self.get_param('twofactor
')
1314 return getpass.getpass('Type
%s and press
[Return
]: ' % note)
1316 # Helper functions for extracting OpenGraph info
1318 def _og_regexes(prop):
1319 content_re = r'content
=(?
:"([^"]+?
)"|\'([^\']+?)\'|\s*([^\s"\'=<>`
]+?
))'
1320 property_re = (r'(?
:name|
property)=(?
:\'og
%(sep)s%(prop)s\'|
"og%(sep)s%(prop)s"|\s
*og
%(sep)s%(prop)s\b)'
1321 % {'prop': re.escape(prop), 'sep': '(?::|[:-])'})
1322 template = r'<meta
[^
>]+?
%s[^
>]+?
%s'
1324 template % (property_re, content_re),
1325 template % (content_re, property_re),
1329 def _meta_regex(prop):
1330 return r'''(?isx)<meta
1331 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1332 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1334 def _og_search_property(self, prop, html, name=None, **kargs):
1335 prop = variadic(prop)
1337 name = 'OpenGraph
%s' % prop[0]
1340 og_regexes.extend(self._og_regexes(p))
1341 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1344 return unescapeHTML(escaped)
1346 def _og_search_thumbnail(self, html, **kargs):
1347 return self._og_search_property('image
', html, 'thumbnail URL
', fatal=False, **kargs)
1349 def _og_search_description(self, html, **kargs):
1350 return self._og_search_property('description
', html, fatal=False, **kargs)
1352 def _og_search_title(self, html, *, fatal=False, **kargs):
1353 return self._og_search_property('title
', html, fatal=fatal, **kargs)
1355 def _og_search_video_url(self, html, name='video url
', secure=True, **kargs):
1356 regexes = self._og_regexes('video
') + self._og_regexes('video
:url
')
1358 regexes = self._og_regexes('video
:secure_url
') + regexes
1359 return self._html_search_regex(regexes, html, name, **kargs)
1361 def _og_search_url(self, html, **kargs):
1362 return self._og_search_property('url
', html, **kargs)
1364 def _html_extract_title(self, html, name='title
', *, fatal=False, **kwargs):
1365 return self._html_search_regex(r'(?s
)<title
\b[^
>]*>([^
<]+)</title
>', html, name, fatal=fatal, **kwargs)
1367 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1368 name = variadic(name)
1369 if display_name is None:
1370 display_name = name[0]
1371 return self._html_search_regex(
1372 [self._meta_regex(n) for n in name],
1373 html, display_name, fatal=fatal, group='content
', **kwargs)
1375 def _dc_search_uploader(self, html):
1376 return self._html_search_meta('dc
.creator
', html, 'uploader
')
1379 def _rta_search(html):
1380 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1381 if re.search(r'(?ix
)<meta\s
+name
="rating"\s
+'
1382 r' content
="RTA-5042-1996-1400-1577-RTA"',
1386 # And then there are the jokers who advertise that they use RTA, but actually don't
.
1387 AGE_LIMIT_MARKERS
= [
1388 r
'Proudly Labeled <a href="http://www\.rtalabel\.org/" title="Restricted to Adults">RTA</a>',
1390 if any(re
.search(marker
, html
) for marker
in AGE_LIMIT_MARKERS
):
1394 def _media_rating_search(self
, html
):
1395 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1396 rating
= self
._html
_search
_meta
('rating', html
)
1408 return RATING_TABLE
.get(rating
.lower())
1410 def _family_friendly_search(self
, html
):
1411 # See http://schema.org/VideoObject
1412 family_friendly
= self
._html
_search
_meta
(
1413 'isFamilyFriendly', html
, default
=None)
1415 if not family_friendly
:
1424 return RATING_TABLE
.get(family_friendly
.lower())
1426 def _twitter_search_player(self
, html
):
1427 return self
._html
_search
_meta
('twitter:player', html
,
1428 'twitter card player')
1430 def _yield_json_ld(self
, html
, video_id
, *, fatal
=True, default
=NO_DEFAULT
):
1431 """Yield all json ld objects in the html"""
1432 if default
is not NO_DEFAULT
:
1434 for mobj
in re
.finditer(JSON_LD_RE
, html
):
1435 json_ld_item
= self
._parse
_json
(mobj
.group('json_ld'), video_id
, fatal
=fatal
)
1436 for json_ld
in variadic(json_ld_item
):
1437 if isinstance(json_ld
, dict):
1440 def _search_json_ld(self
, html
, video_id
, expected_type
=None, *, fatal
=True, default
=NO_DEFAULT
):
1441 """Search for a video in any json ld in the html"""
1442 if default
is not NO_DEFAULT
:
1444 info
= self
._json
_ld
(
1445 list(self
._yield
_json
_ld
(html
, video_id
, fatal
=fatal
, default
=default
)),
1446 video_id
, fatal
=fatal
, expected_type
=expected_type
)
1449 if default
is not NO_DEFAULT
:
1452 raise RegexNotFoundError('Unable to extract JSON-LD')
1454 self
.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1457 def _json_ld(self
, json_ld
, video_id
, fatal
=True, expected_type
=None):
1458 if isinstance(json_ld
, str):
1459 json_ld
= self
._parse
_json
(json_ld
, video_id
, fatal
=fatal
)
1463 if not isinstance(json_ld
, (list, tuple, dict)):
1465 if isinstance(json_ld
, dict):
1468 INTERACTION_TYPE_MAP
= {
1469 'CommentAction': 'comment',
1470 'AgreeAction': 'like',
1471 'DisagreeAction': 'dislike',
1472 'LikeAction': 'like',
1473 'DislikeAction': 'dislike',
1474 'ListenAction': 'view',
1475 'WatchAction': 'view',
1476 'ViewAction': 'view',
1479 def is_type(e
, *expected_types
):
1480 type = variadic(traverse_obj(e
, '@type'))
1481 return any(x
in type for x
in expected_types
)
1483 def extract_interaction_type(e
):
1484 interaction_type
= e
.get('interactionType')
1485 if isinstance(interaction_type
, dict):
1486 interaction_type
= interaction_type
.get('@type')
1487 return str_or_none(interaction_type
)
1489 def extract_interaction_statistic(e
):
1490 interaction_statistic
= e
.get('interactionStatistic')
1491 if isinstance(interaction_statistic
, dict):
1492 interaction_statistic
= [interaction_statistic
]
1493 if not isinstance(interaction_statistic
, list):
1495 for is_e
in interaction_statistic
:
1496 if not is_type(is_e
, 'InteractionCounter'):
1498 interaction_type
= extract_interaction_type(is_e
)
1499 if not interaction_type
:
1501 # For interaction count some sites provide string instead of
1502 # an integer (as per spec) with non digit characters (e.g. ",")
1503 # so extracting count with more relaxed str_to_int
1504 interaction_count
= str_to_int(is_e
.get('userInteractionCount'))
1505 if interaction_count
is None:
1507 count_kind
= INTERACTION_TYPE_MAP
.get(interaction_type
.split('/')[-1])
1510 count_key
= '%s_count' % count_kind
1511 if info
.get(count_key
) is not None:
1513 info
[count_key
] = interaction_count
1515 def extract_chapter_information(e
):
1517 'title': part
.get('name'),
1518 'start_time': part
.get('startOffset'),
1519 'end_time': part
.get('endOffset'),
1520 } for part
in variadic(e
.get('hasPart') or []) if part
.get('@type') == 'Clip']
1521 for idx
, (last_c
, current_c
, next_c
) in enumerate(zip(
1522 [{'end_time': 0}
] + chapters
, chapters
, chapters
[1:])):
1523 current_c
['end_time'] = current_c
['end_time'] or next_c
['start_time']
1524 current_c
['start_time'] = current_c
['start_time'] or last_c
['end_time']
1525 if None in current_c
.values():
1526 self
.report_warning(f
'Chapter {idx} contains broken data. Not extracting chapters')
1529 chapters
[-1]['end_time'] = chapters
[-1]['end_time'] or info
['duration']
1530 info
['chapters'] = chapters
1532 def extract_video_object(e
):
1533 assert is_type(e
, 'VideoObject')
1534 author
= e
.get('author')
1536 'url': url_or_none(e
.get('contentUrl')),
1537 'title': unescapeHTML(e
.get('name')),
1538 'description': unescapeHTML(e
.get('description')),
1539 'thumbnails': [{'url': unescapeHTML(url)}
1540 for url
in variadic(traverse_obj(e
, 'thumbnailUrl', 'thumbnailURL'))
1541 if url_or_none(url
)],
1542 'duration': parse_duration(e
.get('duration')),
1543 'timestamp': unified_timestamp(e
.get('uploadDate')),
1544 # author can be an instance of 'Organization' or 'Person' types.
1545 # both types can have 'name' property(inherited from 'Thing' type). [1]
1546 # however some websites are using 'Text' type instead.
1547 # 1. https://schema.org/VideoObject
1548 'uploader': author
.get('name') if isinstance(author
, dict) else author
if isinstance(author
, str) else None,
1549 'filesize': int_or_none(float_or_none(e
.get('contentSize'))),
1550 'tbr': int_or_none(e
.get('bitrate')),
1551 'width': int_or_none(e
.get('width')),
1552 'height': int_or_none(e
.get('height')),
1553 'view_count': int_or_none(e
.get('interactionCount')),
1555 extract_interaction_statistic(e
)
1556 extract_chapter_information(e
)
1558 def traverse_json_ld(json_ld
, at_top_level
=True):
1560 if at_top_level
and '@context' not in e
:
1562 if at_top_level
and set(e
.keys()) == {'@context', '@graph'}
:
1563 traverse_json_ld(variadic(e
['@graph'], allowed_types
=(dict,)), at_top_level
=False)
1565 if expected_type
is not None and not is_type(e
, expected_type
):
1567 rating
= traverse_obj(e
, ('aggregateRating', 'ratingValue'), expected_type
=float_or_none
)
1568 if rating
is not None:
1569 info
['average_rating'] = rating
1570 if is_type(e
, 'TVEpisode', 'Episode'):
1571 episode_name
= unescapeHTML(e
.get('name'))
1573 'episode': episode_name
,
1574 'episode_number': int_or_none(e
.get('episodeNumber')),
1575 'description': unescapeHTML(e
.get('description')),
1577 if not info
.get('title') and episode_name
:
1578 info
['title'] = episode_name
1579 part_of_season
= e
.get('partOfSeason')
1580 if is_type(part_of_season
, 'TVSeason', 'Season', 'CreativeWorkSeason'):
1582 'season': unescapeHTML(part_of_season
.get('name')),
1583 'season_number': int_or_none(part_of_season
.get('seasonNumber')),
1585 part_of_series
= e
.get('partOfSeries') or e
.get('partOfTVSeries')
1586 if is_type(part_of_series
, 'TVSeries', 'Series', 'CreativeWorkSeries'):
1587 info
['series'] = unescapeHTML(part_of_series
.get('name'))
1588 elif is_type(e
, 'Movie'):
1590 'title': unescapeHTML(e
.get('name')),
1591 'description': unescapeHTML(e
.get('description')),
1592 'duration': parse_duration(e
.get('duration')),
1593 'timestamp': unified_timestamp(e
.get('dateCreated')),
1595 elif is_type(e
, 'Article', 'NewsArticle'):
1597 'timestamp': parse_iso8601(e
.get('datePublished')),
1598 'title': unescapeHTML(e
.get('headline')),
1599 'description': unescapeHTML(e
.get('articleBody') or e
.get('description')),
1601 if is_type(traverse_obj(e
, ('video', 0)), 'VideoObject'):
1602 extract_video_object(e
['video'][0])
1603 elif is_type(traverse_obj(e
, ('subjectOf', 0)), 'VideoObject'):
1604 extract_video_object(e
['subjectOf'][0])
1605 elif is_type(e
, 'VideoObject'):
1606 extract_video_object(e
)
1607 if expected_type
is None:
1611 video
= e
.get('video')
1612 if is_type(video
, 'VideoObject'):
1613 extract_video_object(video
)
1614 if expected_type
is None:
1618 traverse_json_ld(json_ld
)
1620 return filter_dict(info
)
1622 def _search_nextjs_data(self
, webpage
, video_id
, *, transform_source
=None, fatal
=True, **kw
):
1623 return self
._parse
_json
(
1625 r
'(?s)<script[^>]+id=[\'"]__NEXT_DATA__[\'"][^
>]*>([^
<]+)</script
>',
1626 webpage, 'next
.js data
', fatal=fatal, **kw),
1627 video_id, transform_source=transform_source, fatal=fatal)
1629 def _search_nuxt_data(self, webpage, video_id, context_name='__NUXT__
', *, fatal=True, traverse=('data
', 0)):
1630 """Parses Nuxt.js metadata. This works as long as the function __NUXT__ invokes is a pure function"""
1631 rectx = re.escape(context_name)
1632 FUNCTION_RE = r'\
(function\
((?P
<arg_keys
>.*?
)\
){return\s+(?P<js>{.*?}
)\s
*;?\s
*}\
((?P
<arg_vals
>.*?
)\
)'
1633 js, arg_keys, arg_vals = self._search_regex(
1634 (rf'<script
>\s
*window\
.{rectx}
={FUNCTION_RE}\s
*\
)\s
*;?\s
*</script
>', rf'{rectx}\
(.*?{FUNCTION_RE}
'),
1635 webpage, context_name, group=('js
', 'arg_keys
', 'arg_vals
'), fatal=fatal)
1637 args = dict(zip(arg_keys.split(','), arg_vals.split(',')))
1639 for key, val in args.items():
1640 if val in ('undefined
', 'void
0'):
1643 ret = self._parse_json(js, video_id, transform_source=functools.partial(js_to_json, vars=args), fatal=fatal)
1644 return traverse_obj(ret, traverse) or {}
1647 def _hidden_inputs(html):
1648 html = re.sub(r'<!--(?
:(?
!<!--).)*-->', '', html)
1650 for input in re.findall(r'(?i
)(<input[^
>]+>)', html):
1651 attrs = extract_attributes(input)
1654 if attrs.get('type') not in ('hidden
', 'submit
'):
1656 name = attrs.get('name
') or attrs.get('id')
1657 value = attrs.get('value
')
1658 if name and value is not None:
1659 hidden_inputs[name] = value
1660 return hidden_inputs
1662 def _form_hidden_inputs(self, form_id, html):
1663 form = self._search_regex(
1664 r'(?
is)<form
[^
>]+?
id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1665 html, '%s form' % form_id, group='form')
1666 return self._hidden_inputs(form)
1669 regex = r' *((?P<reverse>\+)?(?P<field>[a-zA-Z0-9_]+)((?P<separator>[~:])(?P<limit>.*?))?)? *$'
1671 default = ('hidden', 'aud_or_vid', 'hasvid', 'ie_pref', 'lang', 'quality',
1672 'res', 'fps', 'hdr:12', 'vcodec:vp9.2', 'channels', 'acodec',
1673 'size', 'br', 'asr', 'proto', 'ext', 'hasaud', 'source', 'id') # These must not be aliases
1674 ytdl_default = ('hasaud', 'lang', 'quality', 'tbr', 'filesize', 'vbr',
1675 'height', 'width', 'proto', 'vext', 'abr', 'aext',
1676 'fps', 'fs_approx', 'source', 'id')
1679 'vcodec': {'type': 'ordered', 'regex': True,
1680 'order': ['av0?1', 'vp0?9.2', 'vp0?9', '[hx]265|he?vc?', '[hx]264|avc', 'vp0?8', 'mp4v|h263', 'theora', '', None, 'none']},
1681 'acodec': {'type': 'ordered', 'regex': True,
1682 'order': ['[af]lac', 'wav|aiff', 'opus', 'vorbis|ogg', 'aac', 'mp?4a?', 'mp3', 'e-?a?c-?3', 'ac-?3', 'dts', '', None, 'none']},
1683 'hdr': {'type': 'ordered', 'regex': True, 'field': 'dynamic_range',
1684 'order': ['dv', '(hdr)?12', r'(hdr)?10\+', '(hdr)?10', 'hlg', '', 'sdr', None]},
1685 'proto': {'type': 'ordered', 'regex': True, 'field': 'protocol',
1686 'order': ['(ht|f)tps', '(ht|f)tp$', 'm3u8.*', '.*dash', 'websocket_frag', 'rtmpe?', '', 'mms|rtsp', 'ws|websocket', 'f4']},
1687 'vext': {'type': 'ordered', 'field': 'video_ext',
1688 'order': ('mp4', 'webm', 'flv', '', 'none'),
1689 'order_free': ('webm', 'mp4', 'flv', '', 'none')},
1690 'aext': {'type': 'ordered', 'field': 'audio_ext',
1691 'order': ('m4a', 'aac', 'mp3', 'ogg', 'opus', 'webm', '', 'none'),
1692 'order_free': ('opus', 'ogg', 'webm', 'm4a', 'mp3', 'aac', '', 'none')},
1693 'hidden': {'visible': False, 'forced': True, 'type': 'extractor', 'max': -1000},
1694 'aud_or_vid': {'visible': False, 'forced': True, 'type': 'multiple',
1695 'field': ('vcodec', 'acodec'),
1696 'function': lambda it: int(any(v != 'none' for v in it))},
1697 'ie_pref': {'priority': True, 'type': 'extractor'},
1698 'hasvid': {'priority': True, 'field': 'vcodec', 'type': 'boolean', 'not_in_list': ('none',)},
1699 'hasaud': {'field': 'acodec', 'type': 'boolean', 'not_in_list': ('none',)},
1700 'lang': {'convert': 'float', 'field': 'language_preference', 'default': -1},
1701 'quality': {'convert': 'float', 'default': -1},
1702 'filesize': {'convert': 'bytes'},
1703 'fs_approx': {'convert': 'bytes', 'field': 'filesize_approx'},
1704 'id': {'convert': 'string', 'field': 'format_id'},
1705 'height': {'convert': 'float_none'},
1706 'width': {'convert': 'float_none'},
1707 'fps': {'convert': 'float_none'},
1708 'channels': {'convert': 'float_none', 'field': 'audio_channels'},
1709 'tbr': {'convert': 'float_none'},
1710 'vbr': {'convert': 'float_none'},
1711 'abr': {'convert': 'float_none'},
1712 'asr': {'convert': 'float_none'},
1713 'source': {'convert': 'float', 'field': 'source_preference', 'default': -1},
1715 'codec': {'type': 'combined', 'field': ('vcodec', 'acodec')},
1716 'br': {'type': 'combined', 'field': ('tbr', 'vbr', 'abr'), 'same_limit': True},
1717 'size': {'type': 'combined', 'same_limit': True, 'field': ('filesize', 'fs_approx')},
1718 'ext': {'type': 'combined', 'field': ('vext', 'aext')},
1719 'res': {'type': 'multiple', 'field': ('height', 'width'),
1720 'function': lambda it: (lambda l: min(l) if l else 0)(tuple(filter(None, it)))},
1722 # Actual field names
1723 'format_id': {'type': 'alias', 'field': 'id'},
1724 'preference': {'type': 'alias', 'field': 'ie_pref'},
1725 'language_preference': {'type': 'alias', 'field': 'lang'},
1726 'source_preference': {'type': 'alias', 'field': 'source'},
1727 'protocol': {'type': 'alias', 'field': 'proto'},
1728 'filesize_approx': {'type': 'alias', 'field': 'fs_approx'},
1729 'audio_channels': {'type': 'alias', 'field': 'channels'},
1732 'dimension': {'type': 'alias', 'field': 'res', 'deprecated': True},
1733 'resolution': {'type': 'alias', 'field': 'res', 'deprecated': True},
1734 'extension': {'type': 'alias', 'field': 'ext', 'deprecated': True},
1735 'bitrate': {'type': 'alias', 'field': 'br', 'deprecated': True},
1736 'total_bitrate': {'type': 'alias', 'field': 'tbr', 'deprecated': True},
1737 'video_bitrate': {'type': 'alias', 'field': 'vbr', 'deprecated': True},
1738 'audio_bitrate': {'type': 'alias', 'field': 'abr', 'deprecated': True},
1739 'framerate': {'type': 'alias', 'field': 'fps', 'deprecated': True},
1740 'filesize_estimate': {'type': 'alias', 'field': 'size', 'deprecated': True},
1741 'samplerate': {'type': 'alias', 'field': 'asr', 'deprecated': True},
1742 'video_ext': {'type': 'alias', 'field': 'vext', 'deprecated': True},
1743 'audio_ext': {'type': 'alias', 'field': 'aext', 'deprecated': True},
1744 'video_codec': {'type': 'alias', 'field': 'vcodec', 'deprecated': True},
1745 'audio_codec': {'type': 'alias', 'field': 'acodec', 'deprecated': True},
1746 'video': {'type': 'alias', 'field': 'hasvid', 'deprecated': True},
1747 'has_video': {'type': 'alias', 'field': 'hasvid', 'deprecated': True},
1748 'audio': {'type': 'alias', 'field': 'hasaud', 'deprecated': True},
1749 'has_audio': {'type': 'alias', 'field': 'hasaud', 'deprecated': True},
1750 'extractor': {'type': 'alias', 'field': 'ie_pref', 'deprecated': True},
1751 'extractor_preference': {'type': 'alias', 'field': 'ie_pref', 'deprecated': True},
1754 def __init__(self, ie, field_preference):
1756 self.ydl = ie._downloader
1757 self.evaluate_params(self.ydl.params, field_preference)
1758 if ie.get_param('verbose'):
1759 self.print_verbose_info(self.ydl.write_debug)
1761 def _get_field_setting(self, field, key):
1762 if field not in self.settings:
1763 if key in ('forced', 'priority'):
1765 self.ydl.deprecation_warning(
1766 f'Using arbitrary fields ({field}) for format sorting is deprecated '
1767 'and may be removed in a future version')
1768 self.settings[field] = {}
1769 propObj = self.settings[field]
1770 if key not in propObj:
1771 type = propObj.get('type')
1773 default = 'preference' if type == 'extractor' else (field,) if type in ('combined', 'multiple') else field
1774 elif key == 'convert':
1775 default = 'order' if type == 'ordered' else 'float_string' if field else 'ignore'
1777 default = {'type': 'field', 'visible': True, 'order': [], 'not_in_list': (None,)}.get(key, None)
1778 propObj[key] = default
1781 def _resolve_field_value(self, field, value, convertNone=False):
1786 value = value.lower()
1787 conversion = self._get_field_setting(field, 'convert')
1788 if conversion == 'ignore':
1790 if conversion == 'string':
1792 elif conversion == 'float_none':
1793 return float_or_none(value)
1794 elif conversion == 'bytes':
1795 return FileDownloader.parse_bytes(value)
1796 elif conversion == 'order':
1797 order_list = (self._use_free_order and self._get_field_setting(field, 'order_free')) or self._get_field_setting(field, 'order')
1798 use_regex = self._get_field_setting(field, 'regex')
1799 list_length = len(order_list)
1800 empty_pos = order_list.index('') if '' in order_list else list_length + 1
1801 if use_regex and value is not None:
1802 for i, regex in enumerate(order_list):
1803 if regex and re.match(regex, value):
1804 return list_length - i
1805 return list_length - empty_pos # not in list
1806 else: # not regex or value = None
1807 return list_length - (order_list.index(value) if value in order_list else empty_pos)
1809 if value.isnumeric():
1812 self.settings[field]['convert'] = 'string'
1815 def evaluate_params(self, params, sort_extractor):
1816 self._use_free_order = params.get('prefer_free_formats', False)
1817 self._sort_user = params.get('format_sort', [])
1818 self._sort_extractor = sort_extractor
1820 def add_item(field, reverse, closest, limit_text):
1821 field = field.lower()
1822 if field in self._order:
1824 self._order.append(field)
1825 limit = self._resolve_field_value(field, limit_text)
1828 'closest': False if limit is None else closest,
1829 'limit_text': limit_text,
1831 if field in self.settings:
1832 self.settings[field].update(data)
1834 self.settings[field] = data
1837 tuple(field for field in self.default if self._get_field_setting(field, 'forced'))
1838 + (tuple() if params.get('format_sort_force', False)
1839 else tuple(field for field in self.default if self._get_field_setting(field, 'priority')))
1840 + tuple(self._sort_user) + tuple(sort_extractor) + self.default)
1842 for item in sort_list:
1843 match = re.match(self.regex, item)
1845 raise ExtractorError('Invalid format sort string "%s" given by extractor' % item)
1846 field = match.group('field')
1849 if self._get_field_setting(field, 'type') == 'alias':
1850 alias, field = field, self._get_field_setting(field, 'field')
1851 if self._get_field_setting(alias, 'deprecated'):
1852 self.ydl.deprecation_warning(
1853 f'Format sorting alias {alias} is deprecated '
1854 f'and may be removed in a future version. Please use {field} instead')
1855 reverse = match.group('reverse') is not None
1856 closest = match.group('separator') == '~'
1857 limit_text = match.group('limit')
1859 has_limit = limit_text is not None
1860 has_multiple_fields = self._get_field_setting(field, 'type') == 'combined'
1861 has_multiple_limits = has_limit and has_multiple_fields and not self._get_field_setting(field, 'same_limit')
1863 fields = self._get_field_setting(field, 'field') if has_multiple_fields else (field,)
1864 limits = limit_text.split(':') if has_multiple_limits else (limit_text,) if has_limit else tuple()
1865 limit_count = len(limits)
1866 for (i, f) in enumerate(fields):
1867 add_item(f, reverse, closest,
1868 limits[i] if i < limit_count
1869 else limits[0] if has_limit and not has_multiple_limits
1872 def print_verbose_info(self, write_debug):
1874 write_debug('Sort order given by user: %s' % ', '.join(self._sort_user))
1875 if self._sort_extractor:
1876 write_debug('Sort order given by extractor: %s' % ', '.join(self._sort_extractor))
1877 write_debug('Formats sorted by: %s' % ', '.join(['%s%s%s' % (
1878 '+' if self._get_field_setting(field, 'reverse') else '', field,
1879 '%s%s(%s)' % ('~' if self._get_field_setting(field, 'closest') else ':',
1880 self._get_field_setting(field, 'limit_text'),
1881 self._get_field_setting(field, 'limit'))
1882 if self._get_field_setting(field, 'limit_text') is not None else '')
1883 for field in self._order if self._get_field_setting(field, 'visible')]))
1885 def _calculate_field_preference_from_value(self, format, field, type, value):
1886 reverse = self._get_field_setting(field, 'reverse')
1887 closest = self._get_field_setting(field, 'closest')
1888 limit = self._get_field_setting(field, 'limit')
1890 if type == 'extractor':
1891 maximum = self._get_field_setting(field, 'max')
1892 if value is None or (maximum is not None and value >= maximum):
1894 elif type == 'boolean':
1895 in_list = self._get_field_setting(field, 'in_list')
1896 not_in_list = self._get_field_setting(field, 'not_in_list')
1897 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
1898 elif type == 'ordered':
1899 value = self._resolve_field_value(field, value, True)
1901 # try to convert to number
1902 val_num = float_or_none(value, default=self._get_field_setting(field, 'default'))
1903 is_num = self._get_field_setting(field, 'convert') != 'string' and val_num is not None
1907 return ((-10, 0) if value is None
1908 else (1, value, 0) if not is_num # if a field has mixed strings and numbers, strings are sorted higher
1909 else (0, -abs(value - limit), value - limit if reverse else limit - value) if closest
1910 else (0, value, 0) if not reverse and (limit is None or value <= limit)
1911 else (0, -value, 0) if limit is None or (reverse and value == limit) or value > limit
1912 else (-1, value, 0))
1914 def _calculate_field_preference(self, format, field):
1915 type = self._get_field_setting(field, 'type') # extractor, boolean, ordered, field, multiple
1916 get_value = lambda f: format.get(self._get_field_setting(f, 'field'))
1917 if type == 'multiple':
1918 type = 'field' # Only 'field' is allowed in multiple for now
1919 actual_fields = self._get_field_setting(field, 'field')
1921 value = self._get_field_setting(field, 'function')(get_value(f) for f in actual_fields)
1923 value = get_value(field)
1924 return self._calculate_field_preference_from_value(format, field, type, value)
1926 def calculate_preference(self, format):
1927 # Determine missing protocol
1928 if not format.get('protocol'):
1929 format['protocol'] = determine_protocol(format)
1931 # Determine missing ext
1932 if not format.get('ext') and 'url' in format:
1933 format['ext'] = determine_ext(format['url'])
1934 if format.get('vcodec') == 'none':
1935 format['audio_ext'] = format['ext'] if format.get('acodec') != 'none' else 'none'
1936 format['video_ext'] = 'none'
1938 format['video_ext'] = format['ext']
1939 format['audio_ext'] = 'none'
1940 # if format.get('preference') is None and format.get('ext') in ('f4f', 'f4m'): # Not supported?
1941 # format['preference'] = -1000
1943 # Determine missing bitrates
1944 if format.get('tbr') is None:
1945 if format.get('vbr') is not None and format.get('abr') is not None:
1946 format['tbr'] = format.get('vbr', 0) + format.get('abr', 0)
1948 if format.get('vcodec') != 'none' and format.get('vbr') is None:
1949 format['vbr'] = format.get('tbr') - format.get('abr', 0)
1950 if format.get('acodec') != 'none' and format.get('abr') is None:
1951 format['abr'] = format.get('tbr') - format.get('vbr', 0)
1953 return tuple(self._calculate_field_preference(format, field) for field in self._order)
1955 def _sort_formats(self, formats, field_preference=[]):
1958 formats.sort(key=self.FormatSort(self, field_preference).calculate_preference)
1960 def _check_formats(self, formats, video_id):
1962 formats[:] = filter(
1963 lambda f: self._is_valid_url(
1965 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1969 def _remove_duplicate_formats(formats):
1973 if f['url'] not in format_urls:
1974 format_urls.add(f['url'])
1975 unique_formats.append(f)
1976 formats[:] = unique_formats
1978 def _is_valid_url(self, url, video_id, item='video', headers={}):
1979 url = self._proto_relative_url(url, scheme='http:')
1980 # For now assume non HTTP(S) URLs always valid
1981 if not (url.startswith('http://') or url.startswith('https://')):
1984 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1986 except ExtractorError as e:
1988 '%s: %s URL is invalid, skipping: %s'
1989 % (video_id, item, error_to_compat_str(e.cause)))
1992 def http_scheme(self):
1993 """ Either "http
:" or "https
:", depending on the user's preferences """
1996 if self.get_param('prefer_insecure', False)
1999 def _proto_relative_url(self, url, scheme=None):
2000 scheme = scheme or self.http_scheme()
2001 assert scheme.endswith(':')
2002 return sanitize_url(url, scheme=scheme[:-1])
2004 def _sleep(self, timeout, video_id, msg_template=None):
2005 if msg_template is None:
2006 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
2007 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
2011 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
2012 transform_source=lambda s: fix_xml_ampersands(s).strip(),
2013 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
2014 res = self._download_xml_handle(
2015 manifest_url, video_id, 'Downloading f4m manifest',
2016 'Unable to download f4m manifest',
2017 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
2018 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
2019 transform_source=transform_source,
2020 fatal=fatal, data=data, headers=headers, query=query)
2024 manifest, urlh = res
2025 manifest_url = urlh.geturl()
2027 return self._parse_f4m_formats(
2028 manifest, manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
2029 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
2031 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
2032 transform_source=lambda s: fix_xml_ampersands(s).strip(),
2033 fatal=True, m3u8_id=None):
2034 if not isinstance(manifest, xml.etree.ElementTree.Element) and not fatal:
2037 # currently yt-dlp cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
2038 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
2039 if akamai_pv is not None and ';' in akamai_pv.text:
2040 playerVerificationChallenge = akamai_pv.text.split(';')[0]
2041 if playerVerificationChallenge.strip() != '':
2045 manifest_version = '1.0'
2046 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
2048 manifest_version = '2.0'
2049 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
2050 # Remove unsupported DRM protected media from final formats
2051 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
2052 media_nodes = remove_encrypted_media(media_nodes)
2056 manifest_base_url = get_base_url(manifest)
2058 bootstrap_info = xpath_element(
2059 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
2060 'bootstrap info', default=None)
2063 mime_type = xpath_text(
2064 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
2065 'base URL', default=None)
2066 if mime_type and mime_type.startswith('audio/'):
2069 for i, media_el in enumerate(media_nodes):
2070 tbr = int_or_none(media_el.attrib.get('bitrate'))
2071 width = int_or_none(media_el.attrib.get('width'))
2072 height = int_or_none(media_el.attrib.get('height'))
2073 format_id = join_nonempty(f4m_id, tbr or i)
2074 # If <bootstrapInfo> is present, the specified f4m is a
2075 # stream-level manifest, and only set-level manifests may refer to
2076 # external resources. See section 11.4 and section 4 of F4M spec
2077 if bootstrap_info is None:
2079 # @href is introduced in 2.0, see section 11.6 of F4M spec
2080 if manifest_version == '2.0':
2081 media_url = media_el.attrib.get('href')
2082 if media_url is None:
2083 media_url = media_el.attrib.get('url')
2087 media_url if media_url.startswith('http://') or media_url.startswith('https://')
2088 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
2089 # If media_url is itself a f4m manifest do the recursive extraction
2090 # since bitrates in parent manifest (this one) and media_url manifest
2091 # may differ leading to inability to resolve the format by requested
2092 # bitrate in f4m downloader
2093 ext = determine_ext(manifest_url)
2095 f4m_formats = self._extract_f4m_formats(
2096 manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
2097 transform_source=transform_source, fatal=fatal)
2098 # Sometimes stream-level manifest contains single media entry that
2099 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
2100 # At the same time parent's media entry in set-level manifest may
2101 # contain it. We will copy it from parent in such cases.
2102 if len(f4m_formats) == 1:
2105 'tbr': f.get('tbr') or tbr,
2106 'width': f.get('width') or width,
2107 'height': f.get('height') or height,
2108 'format_id': f.get('format_id') if not tbr else format_id,
2111 formats.extend(f4m_formats)
2114 formats.extend(self._extract_m3u8_formats(
2115 manifest_url, video_id, 'mp4', preference=preference,
2116 quality=quality, m3u8_id=m3u8_id, fatal=fatal))
2119 'format_id': format_id,
2120 'url': manifest_url,
2121 'manifest_url': manifest_url,
2122 'ext': 'flv' if bootstrap_info is not None else None,
2128 'preference': preference,
2133 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, quality=None, m3u8_id=None):
2135 'format_id': join_nonempty(m3u8_id, 'meta'),
2139 'preference': preference - 100 if preference else -100,
2141 'resolution': 'multiple',
2142 'format_note': 'Quality selection URL',
2145 def _report_ignoring_subs(self, name):
2146 self.report_warning(bug_reports_message(
2147 f'Ignoring subtitle tracks found in the {name} manifest; '
2148 'if any subtitle tracks are missing,'
2151 def _extract_m3u8_formats(self, *args, **kwargs):
2152 fmts, subs = self._extract_m3u8_formats_and_subtitles(*args, **kwargs)
2154 self._report_ignoring_subs('HLS')
2157 def _extract_m3u8_formats_and_subtitles(
2158 self, m3u8_url, video_id, ext=None, entry_protocol='m3u8_native',
2159 preference=None, quality=None, m3u8_id=None, note=None,
2160 errnote=None, fatal=True, live=False, data=None, headers={},
2163 res = self._download_webpage_handle(
2165 note='Downloading m3u8 information' if note is None else note,
2166 errnote='Failed to download m3u8 information' if errnote is None else errnote,
2167 fatal=fatal, data=data, headers=headers, query=query)
2172 m3u8_doc, urlh = res
2173 m3u8_url = urlh.geturl()
2175 return self._parse_m3u8_formats_and_subtitles(
2176 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
2177 preference=preference, quality=quality, m3u8_id=m3u8_id,
2178 note=note, errnote=errnote, fatal=fatal, live=live, data=data,
2179 headers=headers, query=query, video_id=video_id)
2181 def _parse_m3u8_formats_and_subtitles(
2182 self, m3u8_doc, m3u8_url=None, ext=None, entry_protocol='m3u8_native',
2183 preference=None, quality=None, m3u8_id=None, live=False, note=None,
2184 errnote=None, fatal=True, data=None, headers={}, query={},
2186 formats, subtitles = [], {}
2188 has_drm = re.search('|'.join([
2189 r'#EXT-X-FAXS-CM:', # Adobe Flash Access
2190 r'#EXT-X-(?:SESSION-)?KEY:.*?URI="skd
://', # Apple FairPlay
2193 def format_url(url):
2194 return url if re.match(r'^https?
://', url) else urllib.parse.urljoin(m3u8_url, url)
2196 if self.get_param('hls_split_discontinuity
', False):
2197 def _extract_m3u8_playlist_indices(manifest_url=None, m3u8_doc=None):
2199 if not manifest_url:
2201 m3u8_doc = self._download_webpage(
2202 manifest_url, video_id, fatal=fatal, data=data, headers=headers,
2203 note=False, errnote='Failed to download m3u8 playlist information
')
2204 if m3u8_doc is False:
2206 return range(1 + sum(line.startswith('#EXT-X-DISCONTINUITY') for line in m3u8_doc.splitlines()))
2209 def _extract_m3u8_playlist_indices(*args
, **kwargs
):
2213 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
2214 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
2215 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
2217 # We should try extracting formats only from master playlists [1, 4.3.4],
2218 # i.e. playlists that describe available qualities. On the other hand
2219 # media playlists [1, 4.3.3] should be returned as is since they contain
2220 # just the media without qualities renditions.
2221 # Fortunately, master playlist can be easily distinguished from media
2222 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
2223 # master playlist tags MUST NOT appear in a media playlist and vice versa.
2224 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
2225 # media playlist and MUST NOT appear in master playlist thus we can
2226 # clearly detect media playlist with this criterion.
2228 if '#EXT-X-TARGETDURATION' in m3u8_doc
: # media playlist, return as is
2230 'format_id': join_nonempty(m3u8_id
, idx
),
2231 'format_index': idx
,
2232 'url': m3u8_url
or encode_data_uri(m3u8_doc
.encode('utf-8'), 'application/x-mpegurl'),
2234 'protocol': entry_protocol
,
2235 'preference': preference
,
2238 } for idx
in _extract_m3u8_playlist_indices(m3u8_doc
=m3u8_doc
)]
2240 return formats
, subtitles
2243 last_stream_inf
= {}
2245 def extract_media(x_media_line
):
2246 media
= parse_m3u8_attributes(x_media_line
)
2247 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
2248 media_type
, group_id
, name
= media
.get('TYPE'), media
.get('GROUP-ID'), media
.get('NAME')
2249 if not (media_type
and group_id
and name
):
2251 groups
.setdefault(group_id
, []).append(media
)
2252 # <https://tools.ietf.org/html/rfc8216#section-4.3.4.1>
2253 if media_type
== 'SUBTITLES':
2254 # According to RFC 8216 §4.3.4.2.1, URI is REQUIRED in the
2255 # EXT-X-MEDIA tag if the media type is SUBTITLES.
2256 # However, lack of URI has been spotted in the wild.
2257 # e.g. NebulaIE; see https://github.com/yt-dlp/yt-dlp/issues/339
2258 if not media
.get('URI'):
2260 url
= format_url(media
['URI'])
2263 'ext': determine_ext(url
),
2265 if sub_info
['ext'] == 'm3u8':
2266 # Per RFC 8216 §3.1, the only possible subtitle format m3u8
2267 # files may contain is WebVTT:
2268 # <https://tools.ietf.org/html/rfc8216#section-3.1>
2269 sub_info
['ext'] = 'vtt'
2270 sub_info
['protocol'] = 'm3u8_native'
2271 lang
= media
.get('LANGUAGE') or 'und'
2272 subtitles
.setdefault(lang
, []).append(sub_info
)
2273 if media_type
not in ('VIDEO', 'AUDIO'):
2275 media_url
= media
.get('URI')
2277 manifest_url
= format_url(media_url
)
2279 'format_id': join_nonempty(m3u8_id
, group_id
, name
, idx
),
2280 'format_note': name
,
2281 'format_index': idx
,
2282 'url': manifest_url
,
2283 'manifest_url': m3u8_url
,
2284 'language': media
.get('LANGUAGE'),
2286 'protocol': entry_protocol
,
2287 'preference': preference
,
2289 'vcodec': 'none' if media_type
== 'AUDIO' else None,
2290 } for idx
in _extract_m3u8_playlist_indices(manifest_url
))
2292 def build_stream_name():
2293 # Despite specification does not mention NAME attribute for
2294 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
2295 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
2296 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
2297 stream_name
= last_stream_inf
.get('NAME')
2300 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
2301 # from corresponding rendition group
2302 stream_group_id
= last_stream_inf
.get('VIDEO')
2303 if not stream_group_id
:
2305 stream_group
= groups
.get(stream_group_id
)
2306 if not stream_group
:
2307 return stream_group_id
2308 rendition
= stream_group
[0]
2309 return rendition
.get('NAME') or stream_group_id
2311 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
2312 # chance to detect video only formats when EXT-X-STREAM-INF tags
2313 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
2314 for line
in m3u8_doc
.splitlines():
2315 if line
.startswith('#EXT-X-MEDIA:'):
2318 for line
in m3u8_doc
.splitlines():
2319 if line
.startswith('#EXT-X-STREAM-INF:'):
2320 last_stream_inf
= parse_m3u8_attributes(line
)
2321 elif line
.startswith('#') or not line
.strip():
2324 tbr
= float_or_none(
2325 last_stream_inf
.get('AVERAGE-BANDWIDTH')
2326 or last_stream_inf
.get('BANDWIDTH'), scale
=1000)
2327 manifest_url
= format_url(line
.strip())
2329 for idx
in _extract_m3u8_playlist_indices(manifest_url
):
2330 format_id
= [m3u8_id
, None, idx
]
2331 # Bandwidth of live streams may differ over time thus making
2332 # format_id unpredictable. So it's better to keep provided
2335 stream_name
= build_stream_name()
2336 format_id
[1] = stream_name
or '%d' % (tbr
or len(formats
))
2338 'format_id': join_nonempty(*format_id
),
2339 'format_index': idx
,
2340 'url': manifest_url
,
2341 'manifest_url': m3u8_url
,
2344 'fps': float_or_none(last_stream_inf
.get('FRAME-RATE')),
2345 'protocol': entry_protocol
,
2346 'preference': preference
,
2349 resolution
= last_stream_inf
.get('RESOLUTION')
2351 mobj
= re
.search(r
'(?P<width>\d+)[xX](?P<height>\d+)', resolution
)
2353 f
['width'] = int(mobj
.group('width'))
2354 f
['height'] = int(mobj
.group('height'))
2355 # Unified Streaming Platform
2357 r
'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f
['url'])
2359 abr
, vbr
= mobj
.groups()
2360 abr
, vbr
= float_or_none(abr
, 1000), float_or_none(vbr
, 1000)
2365 codecs
= parse_codecs(last_stream_inf
.get('CODECS'))
2367 audio_group_id
= last_stream_inf
.get('AUDIO')
2368 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
2369 # references a rendition group MUST have a CODECS attribute.
2370 # However, this is not always respected. E.g. [2]
2371 # contains EXT-X-STREAM-INF tag which references AUDIO
2372 # rendition group but does not have CODECS and despite
2373 # referencing an audio group it represents a complete
2374 # (with audio and video) format. So, for such cases we will
2375 # ignore references to rendition groups and treat them
2376 # as complete formats.
2377 if audio_group_id
and codecs
and f
.get('vcodec') != 'none':
2378 audio_group
= groups
.get(audio_group_id
)
2379 if audio_group
and audio_group
[0].get('URI'):
2380 # TODO: update acodec for audio only formats with
2382 f
['acodec'] = 'none'
2383 if not f
.get('ext'):
2384 f
['ext'] = 'm4a' if f
.get('vcodec') == 'none' else 'mp4'
2388 progressive_uri
= last_stream_inf
.get('PROGRESSIVE-URI')
2391 del http_f
['manifest_url']
2393 'format_id': f
['format_id'].replace('hls-', 'http-'),
2395 'url': progressive_uri
,
2397 formats
.append(http_f
)
2399 last_stream_inf
= {}
2400 return formats
, subtitles
2402 def _extract_m3u8_vod_duration(
2403 self
, m3u8_vod_url
, video_id
, note
=None, errnote
=None, data
=None, headers
={}, query={}
):
2405 m3u8_vod
= self
._download
_webpage
(
2406 m3u8_vod_url
, video_id
,
2407 note
='Downloading m3u8 VOD manifest' if note
is None else note
,
2408 errnote
='Failed to download VOD manifest' if errnote
is None else errnote
,
2409 fatal
=False, data
=data
, headers
=headers
, query
=query
)
2411 return self
._parse
_m
3u8_vod
_duration
(m3u8_vod
or '', video_id
)
2413 def _parse_m3u8_vod_duration(self
, m3u8_vod
, video_id
):
2414 if '#EXT-X-PLAYLIST-TYPE:VOD' not in m3u8_vod
:
2418 float(line
[len('#EXTINF:'):].split(',')[0])
2419 for line
in m3u8_vod
.splitlines() if line
.startswith('#EXTINF:'))) or None
2422 def _xpath_ns(path
, namespace
=None):
2426 for c
in path
.split('/'):
2427 if not c
or c
== '.':
2430 out
.append('{%s}%s' % (namespace
, c
))
2431 return '/'.join(out
)
2433 def _extract_smil_formats_and_subtitles(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None, transform_source
=None):
2434 res
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
, transform_source
=transform_source
)
2440 smil_url
= urlh
.geturl()
2442 namespace
= self
._parse
_smil
_namespace
(smil
)
2444 fmts
= self
._parse
_smil
_formats
(
2445 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2446 subs
= self
._parse
_smil
_subtitles
(
2447 smil
, namespace
=namespace
)
2451 def _extract_smil_formats(self
, *args
, **kwargs
):
2452 fmts
, subs
= self
._extract
_smil
_formats
_and
_subtitles
(*args
, **kwargs
)
2454 self
._report
_ignoring
_subs
('SMIL')
2457 def _extract_smil_info(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None):
2458 res
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
)
2463 smil_url
= urlh
.geturl()
2465 return self
._parse
_smil
(smil
, smil_url
, video_id
, f4m_params
=f4m_params
)
2467 def _download_smil(self
, smil_url
, video_id
, fatal
=True, transform_source
=None):
2468 return self
._download
_xml
_handle
(
2469 smil_url
, video_id
, 'Downloading SMIL file',
2470 'Unable to download SMIL file', fatal
=fatal
, transform_source
=transform_source
)
2472 def _parse_smil(self
, smil
, smil_url
, video_id
, f4m_params
=None):
2473 namespace
= self
._parse
_smil
_namespace
(smil
)
2475 formats
= self
._parse
_smil
_formats
(
2476 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2477 subtitles
= self
._parse
_smil
_subtitles
(smil
, namespace
=namespace
)
2479 video_id
= os
.path
.splitext(url_basename(smil_url
))[0]
2483 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2484 name
= meta
.attrib
.get('name')
2485 content
= meta
.attrib
.get('content')
2486 if not name
or not content
:
2488 if not title
and name
== 'title':
2490 elif not description
and name
in ('description', 'abstract'):
2491 description
= content
2492 elif not upload_date
and name
== 'date':
2493 upload_date
= unified_strdate(content
)
2496 'id': image
.get('type'),
2497 'url': image
.get('src'),
2498 'width': int_or_none(image
.get('width')),
2499 'height': int_or_none(image
.get('height')),
2500 } for image
in smil
.findall(self
._xpath
_ns
('.//image', namespace
)) if image
.get('src')]
2504 'title': title
or video_id
,
2505 'description': description
,
2506 'upload_date': upload_date
,
2507 'thumbnails': thumbnails
,
2509 'subtitles': subtitles
,
2512 def _parse_smil_namespace(self
, smil
):
2513 return self
._search
_regex
(
2514 r
'(?i)^{([^}]+)?}smil$', smil
.tag
, 'namespace', default
=None)
2516 def _parse_smil_formats(self
, smil
, smil_url
, video_id
, namespace
=None, f4m_params
=None, transform_rtmp_url
=None):
2518 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2519 b
= meta
.get('base') or meta
.get('httpBase')
2531 media
= smil
.findall(self
._xpath
_ns
('.//video', namespace
)) + smil
.findall(self
._xpath
_ns
('.//audio', namespace
))
2532 for medium
in media
:
2533 src
= medium
.get('src')
2534 if not src
or src
in srcs
:
2538 bitrate
= float_or_none(medium
.get('system-bitrate') or medium
.get('systemBitrate'), 1000)
2539 filesize
= int_or_none(medium
.get('size') or medium
.get('fileSize'))
2540 width
= int_or_none(medium
.get('width'))
2541 height
= int_or_none(medium
.get('height'))
2542 proto
= medium
.get('proto')
2543 ext
= medium
.get('ext')
2544 src_ext
= determine_ext(src
)
2545 streamer
= medium
.get('streamer') or base
2547 if proto
== 'rtmp' or streamer
.startswith('rtmp'):
2553 'format_id': 'rtmp-%d' % (rtmp_count
if bitrate
is None else bitrate
),
2555 'filesize': filesize
,
2559 if transform_rtmp_url
:
2560 streamer
, src
= transform_rtmp_url(streamer
, src
)
2561 formats
[-1].update({
2567 src_url
= src
if src
.startswith('http') else urllib
.parse
.urljoin(base
, src
)
2568 src_url
= src_url
.strip()
2570 if proto
== 'm3u8' or src_ext
== 'm3u8':
2571 m3u8_formats
= self
._extract
_m
3u8_formats
(
2572 src_url
, video_id
, ext
or 'mp4', m3u8_id
='hls', fatal
=False)
2573 if len(m3u8_formats
) == 1:
2575 m3u8_formats
[0].update({
2576 'format_id': 'hls-%d' % (m3u8_count
if bitrate
is None else bitrate
),
2581 formats
.extend(m3u8_formats
)
2582 elif src_ext
== 'f4m':
2587 'plugin': 'flowplayer-3.2.0.1',
2589 f4m_url
+= '&' if '?' in f4m_url
else '?'
2590 f4m_url
+= urllib
.parse
.urlencode(f4m_params
)
2591 formats
.extend(self
._extract
_f
4m
_formats
(f4m_url
, video_id
, f4m_id
='hds', fatal
=False))
2592 elif src_ext
== 'mpd':
2593 formats
.extend(self
._extract
_mpd
_formats
(
2594 src_url
, video_id
, mpd_id
='dash', fatal
=False))
2595 elif re
.search(r
'\.ism/[Mm]anifest', src_url
):
2596 formats
.extend(self
._extract
_ism
_formats
(
2597 src_url
, video_id
, ism_id
='mss', fatal
=False))
2598 elif src_url
.startswith('http') and self
._is
_valid
_url
(src
, video_id
):
2602 'ext': ext
or src_ext
or 'flv',
2603 'format_id': 'http-%d' % (bitrate
or http_count
),
2605 'filesize': filesize
,
2610 for medium
in smil
.findall(self
._xpath
_ns
('.//imagestream', namespace
)):
2611 src
= medium
.get('src')
2612 if not src
or src
in srcs
:
2618 'format_id': 'imagestream-%d' % (imgs_count
),
2620 'ext': mimetype2ext(medium
.get('type')),
2623 'width': int_or_none(medium
.get('width')),
2624 'height': int_or_none(medium
.get('height')),
2625 'format_note': 'SMIL storyboards',
2630 def _parse_smil_subtitles(self
, smil
, namespace
=None, subtitles_lang
='en'):
2633 for num
, textstream
in enumerate(smil
.findall(self
._xpath
_ns
('.//textstream', namespace
))):
2634 src
= textstream
.get('src')
2635 if not src
or src
in urls
:
2638 ext
= textstream
.get('ext') or mimetype2ext(textstream
.get('type')) or determine_ext(src
)
2639 lang
= textstream
.get('systemLanguage') or textstream
.get('systemLanguageName') or textstream
.get('lang') or subtitles_lang
2640 subtitles
.setdefault(lang
, []).append({
2646 def _extract_xspf_playlist(self
, xspf_url
, playlist_id
, fatal
=True):
2647 res
= self
._download
_xml
_handle
(
2648 xspf_url
, playlist_id
, 'Downloading xpsf playlist',
2649 'Unable to download xspf manifest', fatal
=fatal
)
2654 xspf_url
= urlh
.geturl()
2656 return self
._parse
_xspf
(
2657 xspf
, playlist_id
, xspf_url
=xspf_url
,
2658 xspf_base_url
=base_url(xspf_url
))
2660 def _parse_xspf(self
, xspf_doc
, playlist_id
, xspf_url
=None, xspf_base_url
=None):
2662 'xspf': 'http://xspf.org/ns/0/',
2663 's1': 'http://static.streamone.nl/player/ns/0',
2667 for track
in xspf_doc
.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP
)):
2669 track
, xpath_with_ns('./xspf:title', NS_MAP
), 'title', default
=playlist_id
)
2670 description
= xpath_text(
2671 track
, xpath_with_ns('./xspf:annotation', NS_MAP
), 'description')
2672 thumbnail
= xpath_text(
2673 track
, xpath_with_ns('./xspf:image', NS_MAP
), 'thumbnail')
2674 duration
= float_or_none(
2675 xpath_text(track
, xpath_with_ns('./xspf:duration', NS_MAP
), 'duration'), 1000)
2678 for location
in track
.findall(xpath_with_ns('./xspf:location', NS_MAP
)):
2679 format_url
= urljoin(xspf_base_url
, location
.text
)
2684 'manifest_url': xspf_url
,
2685 'format_id': location
.get(xpath_with_ns('s1:label', NS_MAP
)),
2686 'width': int_or_none(location
.get(xpath_with_ns('s1:width', NS_MAP
))),
2687 'height': int_or_none(location
.get(xpath_with_ns('s1:height', NS_MAP
))),
2689 self
._sort
_formats
(formats
)
2694 'description': description
,
2695 'thumbnail': thumbnail
,
2696 'duration': duration
,
2701 def _extract_mpd_formats(self
, *args
, **kwargs
):
2702 fmts
, subs
= self
._extract
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2704 self
._report
_ignoring
_subs
('DASH')
2707 def _extract_mpd_formats_and_subtitles(
2708 self
, mpd_url
, video_id
, mpd_id
=None, note
=None, errnote
=None,
2709 fatal
=True, data
=None, headers
={}, query={}
):
2710 res
= self
._download
_xml
_handle
(
2712 note
='Downloading MPD manifest' if note
is None else note
,
2713 errnote
='Failed to download MPD manifest' if errnote
is None else errnote
,
2714 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
2721 # We could have been redirected to a new url when we retrieved our mpd file.
2722 mpd_url
= urlh
.geturl()
2723 mpd_base_url
= base_url(mpd_url
)
2725 return self
._parse
_mpd
_formats
_and
_subtitles
(
2726 mpd_doc
, mpd_id
, mpd_base_url
, mpd_url
)
2728 def _parse_mpd_formats(self
, *args
, **kwargs
):
2729 fmts
, subs
= self
._parse
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2731 self
._report
_ignoring
_subs
('DASH')
2734 def _parse_mpd_formats_and_subtitles(
2735 self
, mpd_doc
, mpd_id
=None, mpd_base_url
='', mpd_url
=None):
2737 Parse formats from MPD manifest.
2739 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
2740 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
2741 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
2743 if not self
.get_param('dynamic_mpd', True):
2744 if mpd_doc
.get('type') == 'dynamic':
2747 namespace
= self
._search
_regex
(r
'(?i)^{([^}]+)?}MPD$', mpd_doc
.tag
, 'namespace', default
=None)
2750 return self
._xpath
_ns
(path
, namespace
)
2752 def is_drm_protected(element
):
2753 return element
.find(_add_ns('ContentProtection')) is not None
2755 def extract_multisegment_info(element
, ms_parent_info
):
2756 ms_info
= ms_parent_info
.copy()
2758 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2759 # common attributes and elements. We will only extract relevant
2761 def extract_common(source
):
2762 segment_timeline
= source
.find(_add_ns('SegmentTimeline'))
2763 if segment_timeline
is not None:
2764 s_e
= segment_timeline
.findall(_add_ns('S'))
2766 ms_info
['total_number'] = 0
2769 r
= int(s
.get('r', 0))
2770 ms_info
['total_number'] += 1 + r
2771 ms_info
['s'].append({
2772 't': int(s
.get('t', 0)),
2773 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2774 'd': int(s
.attrib
['d']),
2777 start_number
= source
.get('startNumber')
2779 ms_info
['start_number'] = int(start_number
)
2780 timescale
= source
.get('timescale')
2782 ms_info
['timescale'] = int(timescale
)
2783 segment_duration
= source
.get('duration')
2784 if segment_duration
:
2785 ms_info
['segment_duration'] = float(segment_duration
)
2787 def extract_Initialization(source
):
2788 initialization
= source
.find(_add_ns('Initialization'))
2789 if initialization
is not None:
2790 ms_info
['initialization_url'] = initialization
.attrib
['sourceURL']
2792 segment_list
= element
.find(_add_ns('SegmentList'))
2793 if segment_list
is not None:
2794 extract_common(segment_list
)
2795 extract_Initialization(segment_list
)
2796 segment_urls_e
= segment_list
.findall(_add_ns('SegmentURL'))
2798 ms_info
['segment_urls'] = [segment
.attrib
['media'] for segment
in segment_urls_e
]
2800 segment_template
= element
.find(_add_ns('SegmentTemplate'))
2801 if segment_template
is not None:
2802 extract_common(segment_template
)
2803 media
= segment_template
.get('media')
2805 ms_info
['media'] = media
2806 initialization
= segment_template
.get('initialization')
2808 ms_info
['initialization'] = initialization
2810 extract_Initialization(segment_template
)
2813 mpd_duration
= parse_duration(mpd_doc
.get('mediaPresentationDuration'))
2814 formats
, subtitles
= [], {}
2815 stream_numbers
= collections
.defaultdict(int)
2816 for period
in mpd_doc
.findall(_add_ns('Period')):
2817 period_duration
= parse_duration(period
.get('duration')) or mpd_duration
2818 period_ms_info
= extract_multisegment_info(period
, {
2822 for adaptation_set
in period
.findall(_add_ns('AdaptationSet')):
2823 adaption_set_ms_info
= extract_multisegment_info(adaptation_set
, period_ms_info
)
2824 for representation
in adaptation_set
.findall(_add_ns('Representation')):
2825 representation_attrib
= adaptation_set
.attrib
.copy()
2826 representation_attrib
.update(representation
.attrib
)
2827 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2828 mime_type
= representation_attrib
['mimeType']
2829 content_type
= representation_attrib
.get('contentType', mime_type
.split('/')[0])
2831 codec_str
= representation_attrib
.get('codecs', '')
2832 # Some kind of binary subtitle found in some youtube livestreams
2833 if mime_type
== 'application/x-rawcc':
2834 codecs
= {'scodec': codec_str}
2836 codecs
= parse_codecs(codec_str
)
2837 if content_type
not in ('video', 'audio', 'text'):
2838 if mime_type
== 'image/jpeg':
2839 content_type
= mime_type
2840 elif codecs
.get('vcodec', 'none') != 'none':
2841 content_type
= 'video'
2842 elif codecs
.get('acodec', 'none') != 'none':
2843 content_type
= 'audio'
2844 elif codecs
.get('scodec', 'none') != 'none':
2845 content_type
= 'text'
2846 elif mimetype2ext(mime_type
) in ('tt', 'dfxp', 'ttml', 'xml', 'json'):
2847 content_type
= 'text'
2849 self
.report_warning('Unknown MIME type %s in DASH manifest' % mime_type
)
2853 for element
in (representation
, adaptation_set
, period
, mpd_doc
):
2854 base_url_e
= element
.find(_add_ns('BaseURL'))
2855 if try_call(lambda: base_url_e
.text
) is not None:
2856 base_url
= base_url_e
.text
+ base_url
2857 if re
.match(r
'^https?://', base_url
):
2859 if mpd_base_url
and base_url
.startswith('/'):
2860 base_url
= urllib
.parse
.urljoin(mpd_base_url
, base_url
)
2861 elif mpd_base_url
and not re
.match(r
'^https?://', base_url
):
2862 if not mpd_base_url
.endswith('/'):
2864 base_url
= mpd_base_url
+ base_url
2865 representation_id
= representation_attrib
.get('id')
2866 lang
= representation_attrib
.get('lang')
2867 url_el
= representation
.find(_add_ns('BaseURL'))
2868 filesize
= int_or_none(url_el
.attrib
.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el
is not None else None)
2869 bandwidth
= int_or_none(representation_attrib
.get('bandwidth'))
2870 if representation_id
is not None:
2871 format_id
= representation_id
2873 format_id
= content_type
2875 format_id
= mpd_id
+ '-' + format_id
2876 if content_type
in ('video', 'audio'):
2878 'format_id': format_id
,
2879 'manifest_url': mpd_url
,
2880 'ext': mimetype2ext(mime_type
),
2881 'width': int_or_none(representation_attrib
.get('width')),
2882 'height': int_or_none(representation_attrib
.get('height')),
2883 'tbr': float_or_none(bandwidth
, 1000),
2884 'asr': int_or_none(representation_attrib
.get('audioSamplingRate')),
2885 'fps': int_or_none(representation_attrib
.get('frameRate')),
2886 'language': lang
if lang
not in ('mul', 'und', 'zxx', 'mis') else None,
2887 'format_note': 'DASH %s' % content_type
,
2888 'filesize': filesize
,
2889 'container': mimetype2ext(mime_type
) + '_dash',
2892 elif content_type
== 'text':
2894 'ext': mimetype2ext(mime_type
),
2895 'manifest_url': mpd_url
,
2896 'filesize': filesize
,
2898 elif content_type
== 'image/jpeg':
2899 # See test case in VikiIE
2900 # https://www.viki.com/videos/1175236v-choosing-spouse-by-lottery-episode-1
2902 'format_id': format_id
,
2904 'manifest_url': mpd_url
,
2905 'format_note': 'DASH storyboards (jpeg)',
2909 if is_drm_protected(adaptation_set
) or is_drm_protected(representation
):
2911 representation_ms_info
= extract_multisegment_info(representation
, adaption_set_ms_info
)
2913 def prepare_template(template_name
, identifiers
):
2914 tmpl
= representation_ms_info
[template_name
]
2915 # First of, % characters outside $...$ templates
2916 # must be escaped by doubling for proper processing
2917 # by % operator string formatting used further (see
2918 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2924 in_template
= not in_template
2925 elif c
== '%' and not in_template
:
2927 # Next, $...$ templates are translated to their
2928 # %(...) counterparts to be used with % operator
2929 if representation_id
is not None:
2930 t
= t
.replace('$RepresentationID$', representation_id
)
2931 t
= re
.sub(r
'\$(%s)\$' % '|'.join(identifiers
), r
'%(\1)d', t
)
2932 t
= re
.sub(r
'\$(%s)%%([^$]+)\$' % '|'.join(identifiers
), r
'%(\1)\2', t
)
2933 t
.replace('$$', '$')
2936 # @initialization is a regular template like @media one
2937 # so it should be handled just the same way (see
2938 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2939 if 'initialization' in representation_ms_info
:
2940 initialization_template
= prepare_template(
2942 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2943 # $Time$ shall not be included for @initialization thus
2944 # only $Bandwidth$ remains
2946 representation_ms_info
['initialization_url'] = initialization_template
% {
2947 'Bandwidth': bandwidth
,
2950 def location_key(location
):
2951 return 'url' if re
.match(r
'^https?://', location
) else 'path'
2953 if 'segment_urls' not in representation_ms_info
and 'media' in representation_ms_info
:
2955 media_template
= prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2956 media_location_key
= location_key(media_template
)
2958 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2959 # can't be used at the same time
2960 if '%(Number' in media_template
and 's' not in representation_ms_info
:
2961 segment_duration
= None
2962 if 'total_number' not in representation_ms_info
and 'segment_duration' in representation_ms_info
:
2963 segment_duration
= float_or_none(representation_ms_info
['segment_duration'], representation_ms_info
['timescale'])
2964 representation_ms_info
['total_number'] = int(math
.ceil(
2965 float_or_none(period_duration
, segment_duration
, default
=0)))
2966 representation_ms_info
['fragments'] = [{
2967 media_location_key
: media_template
% {
2968 'Number': segment_number
,
2969 'Bandwidth': bandwidth
,
2971 'duration': segment_duration
,
2972 } for segment_number
in range(
2973 representation_ms_info
['start_number'],
2974 representation_ms_info
['total_number'] + representation_ms_info
['start_number'])]
2976 # $Number*$ or $Time$ in media template with S list available
2977 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2978 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2979 representation_ms_info
['fragments'] = []
2982 segment_number
= representation_ms_info
['start_number']
2984 def add_segment_url():
2985 segment_url
= media_template
% {
2986 'Time': segment_time
,
2987 'Bandwidth': bandwidth
,
2988 'Number': segment_number
,
2990 representation_ms_info
['fragments'].append({
2991 media_location_key
: segment_url
,
2992 'duration': float_or_none(segment_d
, representation_ms_info
['timescale']),
2995 for num
, s
in enumerate(representation_ms_info
['s']):
2996 segment_time
= s
.get('t') or segment_time
3000 for r
in range(s
.get('r', 0)):
3001 segment_time
+= segment_d
3004 segment_time
+= segment_d
3005 elif 'segment_urls' in representation_ms_info
and 's' in representation_ms_info
:
3006 # No media template,
3007 # e.g. https://www.youtube.com/watch?v=iXZV5uAYMJI
3008 # or any YouTube dashsegments video
3011 timescale
= representation_ms_info
['timescale']
3012 for s
in representation_ms_info
['s']:
3013 duration
= float_or_none(s
['d'], timescale
)
3014 for r
in range(s
.get('r', 0) + 1):
3015 segment_uri
= representation_ms_info
['segment_urls'][segment_index
]
3017 location_key(segment_uri
): segment_uri
,
3018 'duration': duration
,
3021 representation_ms_info
['fragments'] = fragments
3022 elif 'segment_urls' in representation_ms_info
:
3023 # Segment URLs with no SegmentTimeline
3024 # E.g. https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
3025 # https://github.com/ytdl-org/youtube-dl/pull/14844
3027 segment_duration
= float_or_none(
3028 representation_ms_info
['segment_duration'],
3029 representation_ms_info
['timescale']) if 'segment_duration' in representation_ms_info
else None
3030 for segment_url
in representation_ms_info
['segment_urls']:
3032 location_key(segment_url
): segment_url
,
3034 if segment_duration
:
3035 fragment
['duration'] = segment_duration
3036 fragments
.append(fragment
)
3037 representation_ms_info
['fragments'] = fragments
3038 # If there is a fragments key available then we correctly recognized fragmented media.
3039 # Otherwise we will assume unfragmented media with direct access. Technically, such
3040 # assumption is not necessarily correct since we may simply have no support for
3041 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
3042 if 'fragments' in representation_ms_info
:
3044 # NB: mpd_url may be empty when MPD manifest is parsed from a string
3045 'url': mpd_url
or base_url
,
3046 'fragment_base_url': base_url
,
3048 'protocol': 'http_dash_segments' if mime_type
!= 'image/jpeg' else 'mhtml',
3050 if 'initialization_url' in representation_ms_info
:
3051 initialization_url
= representation_ms_info
['initialization_url']
3052 if not f
.get('url'):
3053 f
['url'] = initialization_url
3054 f
['fragments'].append({location_key(initialization_url): initialization_url}
)
3055 f
['fragments'].extend(representation_ms_info
['fragments'])
3056 if not period_duration
:
3057 period_duration
= try_get(
3058 representation_ms_info
,
3059 lambda r
: sum(frag
['duration'] for frag
in r
['fragments']), float)
3061 # Assuming direct URL to unfragmented media.
3063 if content_type
in ('video', 'audio', 'image/jpeg'):
3064 f
['manifest_stream_number'] = stream_numbers
[f
['url']]
3065 stream_numbers
[f
['url']] += 1
3067 elif content_type
== 'text':
3068 subtitles
.setdefault(lang
or 'und', []).append(f
)
3070 return formats
, subtitles
3072 def _extract_ism_formats(self
, *args
, **kwargs
):
3073 fmts
, subs
= self
._extract
_ism
_formats
_and
_subtitles
(*args
, **kwargs
)
3075 self
._report
_ignoring
_subs
('ISM')
3078 def _extract_ism_formats_and_subtitles(self
, ism_url
, video_id
, ism_id
=None, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
):
3079 res
= self
._download
_xml
_handle
(
3081 note
='Downloading ISM manifest' if note
is None else note
,
3082 errnote
='Failed to download ISM manifest' if errnote
is None else errnote
,
3083 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
3090 return self
._parse
_ism
_formats
_and
_subtitles
(ism_doc
, urlh
.geturl(), ism_id
)
3092 def _parse_ism_formats_and_subtitles(self
, ism_doc
, ism_url
, ism_id
=None):
3094 Parse formats from ISM manifest.
3096 1. [MS-SSTR]: Smooth Streaming Protocol,
3097 https://msdn.microsoft.com/en-us/library/ff469518.aspx
3099 if ism_doc
.get('IsLive') == 'TRUE':
3102 duration
= int(ism_doc
.attrib
['Duration'])
3103 timescale
= int_or_none(ism_doc
.get('TimeScale')) or 10000000
3107 for stream
in ism_doc
.findall('StreamIndex'):
3108 stream_type
= stream
.get('Type')
3109 if stream_type
not in ('video', 'audio', 'text'):
3111 url_pattern
= stream
.attrib
['Url']
3112 stream_timescale
= int_or_none(stream
.get('TimeScale')) or timescale
3113 stream_name
= stream
.get('Name')
3114 stream_language
= stream
.get('Language', 'und')
3115 for track
in stream
.findall('QualityLevel'):
3116 fourcc
= track
.get('FourCC') or ('AACL' if track
.get('AudioTag') == '255' else None)
3117 # TODO: add support for WVC1 and WMAP
3118 if fourcc
not in ('H264', 'AVC1', 'AACL', 'TTML'):
3119 self
.report_warning('%s is not a supported codec' % fourcc
)
3121 tbr
= int(track
.attrib
['Bitrate']) // 1000
3122 # [1] does not mention Width and Height attributes. However,
3123 # they're often present while MaxWidth and MaxHeight are
3124 # missing, so should be used as fallbacks
3125 width
= int_or_none(track
.get('MaxWidth') or track
.get('Width'))
3126 height
= int_or_none(track
.get('MaxHeight') or track
.get('Height'))
3127 sampling_rate
= int_or_none(track
.get('SamplingRate'))
3129 track_url_pattern
= re
.sub(r
'{[Bb]itrate}', track
.attrib
['Bitrate'], url_pattern
)
3130 track_url_pattern
= urllib
.parse
.urljoin(ism_url
, track_url_pattern
)
3136 stream_fragments
= stream
.findall('c')
3137 for stream_fragment_index
, stream_fragment
in enumerate(stream_fragments
):
3138 fragment_ctx
['time'] = int_or_none(stream_fragment
.get('t')) or fragment_ctx
['time']
3139 fragment_repeat
= int_or_none(stream_fragment
.get('r')) or 1
3140 fragment_ctx
['duration'] = int_or_none(stream_fragment
.get('d'))
3141 if not fragment_ctx
['duration']:
3143 next_fragment_time
= int(stream_fragment
[stream_fragment_index
+ 1].attrib
['t'])
3145 next_fragment_time
= duration
3146 fragment_ctx
['duration'] = (next_fragment_time
- fragment_ctx
['time']) / fragment_repeat
3147 for _
in range(fragment_repeat
):
3149 'url': re
.sub(r
'{start[ _]time}', str(fragment_ctx
['time']), track_url_pattern
),
3150 'duration': fragment_ctx
['duration'] / stream_timescale
,
3152 fragment_ctx
['time'] += fragment_ctx
['duration']
3154 if stream_type
== 'text':
3155 subtitles
.setdefault(stream_language
, []).append({
3159 'manifest_url': ism_url
,
3160 'fragments': fragments
,
3161 '_download_params': {
3162 'stream_type': stream_type
,
3163 'duration': duration
,
3164 'timescale': stream_timescale
,
3166 'language': stream_language
,
3167 'codec_private_data': track
.get('CodecPrivateData'),
3170 elif stream_type
in ('video', 'audio'):
3172 'format_id': join_nonempty(ism_id
, stream_name
, tbr
),
3174 'manifest_url': ism_url
,
3175 'ext': 'ismv' if stream_type
== 'video' else 'isma',
3179 'asr': sampling_rate
,
3180 'vcodec': 'none' if stream_type
== 'audio' else fourcc
,
3181 'acodec': 'none' if stream_type
== 'video' else fourcc
,
3183 'fragments': fragments
,
3184 'has_drm': ism_doc
.find('Protection') is not None,
3185 '_download_params': {
3186 'stream_type': stream_type
,
3187 'duration': duration
,
3188 'timescale': stream_timescale
,
3189 'width': width
or 0,
3190 'height': height
or 0,
3192 'language': stream_language
,
3193 'codec_private_data': track
.get('CodecPrivateData'),
3194 'sampling_rate': sampling_rate
,
3195 'channels': int_or_none(track
.get('Channels', 2)),
3196 'bits_per_sample': int_or_none(track
.get('BitsPerSample', 16)),
3197 'nal_unit_length_field': int_or_none(track
.get('NALUnitLengthField', 4)),
3200 return formats
, subtitles
3202 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):
3203 def absolute_url(item_url
):
3204 return urljoin(base_url
, item_url
)
3206 def parse_content_type(content_type
):
3207 if not content_type
:
3209 ctr
= re
.search(r
'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type
)
3211 mimetype
, codecs
= ctr
.groups()
3212 f
= parse_codecs(codecs
)
3213 f
['ext'] = mimetype2ext(mimetype
)
3217 def _media_formats(src
, cur_media_type
, type_info
=None):
3218 type_info
= type_info
or {}
3219 full_url
= absolute_url(src
)
3220 ext
= type_info
.get('ext') or determine_ext(full_url
)
3222 is_plain_url
= False
3223 formats
= self
._extract
_m
3u8_formats
(
3224 full_url
, video_id
, ext
='mp4',
3225 entry_protocol
=m3u8_entry_protocol
, m3u8_id
=m3u8_id
,
3226 preference
=preference
, quality
=quality
, fatal
=False)
3228 is_plain_url
= False
3229 formats
= self
._extract
_mpd
_formats
(
3230 full_url
, video_id
, mpd_id
=mpd_id
, fatal
=False)
3235 'vcodec': 'none' if cur_media_type
== 'audio' else None,
3238 return is_plain_url
, formats
3241 # amp-video and amp-audio are very similar to their HTML5 counterparts
3242 # so we will include them right here (see
3243 # https://www.ampproject.org/docs/reference/components/amp-video)
3244 # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/
3245 _MEDIA_TAG_NAME_RE
= r
'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)'
3246 media_tags
= [(media_tag
, media_tag_name
, media_type
, '')
3247 for media_tag
, media_tag_name
, media_type
3248 in re
.findall(r
'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE
, webpage
)]
3249 media_tags
.extend(re
.findall(
3250 # We only allow video|audio followed by a whitespace or '>'.
3251 # Allowing more characters may end up in significant slow down (see
3252 # https://github.com/ytdl-org/youtube-dl/issues/11979,
3253 # e.g. http://www.porntrex.com/maps/videositemap.xml).
3254 r
'(?s)(<(?P<tag>%s)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>' % _MEDIA_TAG_NAME_RE
, webpage
))
3255 for media_tag
, _
, media_type
, media_content
in media_tags
:
3260 media_attributes
= extract_attributes(media_tag
)
3261 src
= strip_or_none(media_attributes
.get('src'))
3263 f
= parse_content_type(media_attributes
.get('type'))
3264 _
, formats
= _media_formats(src
, media_type
, f
)
3265 media_info
['formats'].extend(formats
)
3266 media_info
['thumbnail'] = absolute_url(media_attributes
.get('poster'))
3268 for source_tag
in re
.findall(r
'<source[^>]+>', media_content
):
3269 s_attr
= extract_attributes(source_tag
)
3270 # data-video-src and data-src are non standard but seen
3271 # several times in the wild
3272 src
= strip_or_none(dict_get(s_attr
, ('src', 'data-video-src', 'data-src')))
3275 f
= parse_content_type(s_attr
.get('type'))
3276 is_plain_url
, formats
= _media_formats(src
, media_type
, f
)
3278 # width, height, res, label and title attributes are
3279 # all not standard but seen several times in the wild
3282 for lbl
in ('label', 'title')
3283 if str_or_none(s_attr
.get(lbl
))
3285 width
= int_or_none(s_attr
.get('width'))
3286 height
= (int_or_none(s_attr
.get('height'))
3287 or int_or_none(s_attr
.get('res')))
3288 if not width
or not height
:
3290 resolution
= parse_resolution(lbl
)
3293 width
= width
or resolution
.get('width')
3294 height
= height
or resolution
.get('height')
3296 tbr
= parse_bitrate(lbl
)
3305 'format_id': s_attr
.get('label') or s_attr
.get('title'),
3307 f
.update(formats
[0])
3308 media_info
['formats'].append(f
)
3310 media_info
['formats'].extend(formats
)
3311 for track_tag
in re
.findall(r
'<track[^>]+>', media_content
):
3312 track_attributes
= extract_attributes(track_tag
)
3313 kind
= track_attributes
.get('kind')
3314 if not kind
or kind
in ('subtitles', 'captions'):
3315 src
= strip_or_none(track_attributes
.get('src'))
3318 lang
= track_attributes
.get('srclang') or track_attributes
.get('lang') or track_attributes
.get('label')
3319 media_info
['subtitles'].setdefault(lang
, []).append({
3320 'url': absolute_url(src
),
3322 for f
in media_info
['formats']:
3323 f
.setdefault('http_headers', {})['Referer'] = base_url
3324 if media_info
['formats'] or media_info
['subtitles']:
3325 entries
.append(media_info
)
3328 def _extract_akamai_formats(self
, *args
, **kwargs
):
3329 fmts
, subs
= self
._extract
_akamai
_formats
_and
_subtitles
(*args
, **kwargs
)
3331 self
._report
_ignoring
_subs
('akamai')
3334 def _extract_akamai_formats_and_subtitles(self
, manifest_url
, video_id
, hosts
={}):
3335 signed
= 'hdnea=' in manifest_url
3337 # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
3338 manifest_url
= re
.sub(
3339 r
'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
3340 '', manifest_url
).strip('?')
3345 hdcore_sign
= 'hdcore=3.7.0'
3346 f4m_url
= re
.sub(r
'(https?://[^/]+)/i/', r
'\1/z/', manifest_url
).replace('/master.m3u8', '/manifest.f4m')
3347 hds_host
= hosts
.get('hds')
3349 f4m_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hds_host
, f4m_url
)
3350 if 'hdcore=' not in f4m_url
:
3351 f4m_url
+= ('&' if '?' in f4m_url
else '?') + hdcore_sign
3352 f4m_formats
= self
._extract
_f
4m
_formats
(
3353 f4m_url
, video_id
, f4m_id
='hds', fatal
=False)
3354 for entry
in f4m_formats
:
3355 entry
.update({'extra_param_to_segment_url': hdcore_sign}
)
3356 formats
.extend(f4m_formats
)
3358 m3u8_url
= re
.sub(r
'(https?://[^/]+)/z/', r
'\1/i/', manifest_url
).replace('/manifest.f4m', '/master.m3u8')
3359 hls_host
= hosts
.get('hls')
3361 m3u8_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hls_host
, m3u8_url
)
3362 m3u8_formats
, m3u8_subtitles
= self
._extract
_m
3u8_formats
_and
_subtitles
(
3363 m3u8_url
, video_id
, 'mp4', 'm3u8_native',
3364 m3u8_id
='hls', fatal
=False)
3365 formats
.extend(m3u8_formats
)
3366 subtitles
= self
._merge
_subtitles
(subtitles
, m3u8_subtitles
)
3368 http_host
= hosts
.get('http')
3369 if http_host
and m3u8_formats
and not signed
:
3370 REPL_REGEX
= r
'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
3371 qualities
= re
.match(REPL_REGEX
, m3u8_url
).group(2).split(',')
3372 qualities_length
= len(qualities
)
3373 if len(m3u8_formats
) in (qualities_length
, qualities_length
+ 1):
3375 for f
in m3u8_formats
:
3376 if f
['vcodec'] != 'none':
3377 for protocol
in ('http', 'https'):
3379 del http_f
['manifest_url']
3381 REPL_REGEX
, protocol
+ fr
'://{http_host}/\g<1>{qualities[i]}\3', f
['url'])
3383 'format_id': http_f
['format_id'].replace('hls-', protocol
+ '-'),
3385 'protocol': protocol
,
3387 formats
.append(http_f
)
3390 return formats
, subtitles
3392 def _extract_wowza_formats(self
, url
, video_id
, m3u8_entry_protocol
='m3u8_native', skip_protocols
=[]):
3393 query
= urllib
.parse
.urlparse(url
).query
3394 url
= re
.sub(r
'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url
)
3396 r
'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url
)
3397 url_base
= mobj
.group('url')
3398 http_base_url
= '%s%s:%s' % ('http', mobj
.group('s') or '', url_base
)
3401 def manifest_url(manifest
):
3402 m_url
= f
'{http_base_url}/{manifest}'
3404 m_url
+= '?%s' % query
3407 if 'm3u8' not in skip_protocols
:
3408 formats
.extend(self
._extract
_m
3u8_formats
(
3409 manifest_url('playlist.m3u8'), video_id
, 'mp4',
3410 m3u8_entry_protocol
, m3u8_id
='hls', fatal
=False))
3411 if 'f4m' not in skip_protocols
:
3412 formats
.extend(self
._extract
_f
4m
_formats
(
3413 manifest_url('manifest.f4m'),
3414 video_id
, f4m_id
='hds', fatal
=False))
3415 if 'dash' not in skip_protocols
:
3416 formats
.extend(self
._extract
_mpd
_formats
(
3417 manifest_url('manifest.mpd'),
3418 video_id
, mpd_id
='dash', fatal
=False))
3419 if re
.search(r
'(?:/smil:|\.smil)', url_base
):
3420 if 'smil' not in skip_protocols
:
3421 rtmp_formats
= self
._extract
_smil
_formats
(
3422 manifest_url('jwplayer.smil'),
3423 video_id
, fatal
=False)
3424 for rtmp_format
in rtmp_formats
:
3425 rtsp_format
= rtmp_format
.copy()
3426 rtsp_format
['url'] = '%s/%s' % (rtmp_format
['url'], rtmp_format
['play_path'])
3427 del rtsp_format
['play_path']
3428 del rtsp_format
['ext']
3429 rtsp_format
.update({
3430 'url': rtsp_format
['url'].replace('rtmp://', 'rtsp://'),
3431 'format_id': rtmp_format
['format_id'].replace('rtmp', 'rtsp'),
3434 formats
.extend([rtmp_format
, rtsp_format
])
3436 for protocol
in ('rtmp', 'rtsp'):
3437 if protocol
not in skip_protocols
:
3439 'url': f
'{protocol}:{url_base}',
3440 'format_id': protocol
,
3441 'protocol': protocol
,
3445 def _find_jwplayer_data(self
, webpage
, video_id
=None, transform_source
=js_to_json
):
3447 r
'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P
=quote
)\
)(?
!</script
>).*?\
.setup\s
*\
((?P
<options
>[^
)]+)\
)',
3451 jwplayer_data = self._parse_json(mobj.group('options
'),
3453 transform_source=transform_source)
3454 except ExtractorError:
3457 if isinstance(jwplayer_data, dict):
3458 return jwplayer_data
3460 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
3461 jwplayer_data = self._find_jwplayer_data(
3462 webpage, video_id, transform_source=js_to_json)
3463 return self._parse_jwplayer_data(
3464 jwplayer_data, video_id, *args, **kwargs)
3466 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
3467 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3468 # JWPlayer backward compatibility: flattened playlists
3469 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
3470 if 'playlist
' not in jwplayer_data:
3471 jwplayer_data = {'playlist': [jwplayer_data]}
3475 # JWPlayer backward compatibility: single playlist item
3476 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
3477 if not isinstance(jwplayer_data['playlist
'], list):
3478 jwplayer_data['playlist
'] = [jwplayer_data['playlist
']]
3480 for video_data in jwplayer_data['playlist
']:
3481 # JWPlayer backward compatibility: flattened sources
3482 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
3483 if 'sources
' not in video_data:
3484 video_data['sources
'] = [video_data]
3486 this_video_id = video_id or video_data['mediaid
']
3488 formats = self._parse_jwplayer_formats(
3489 video_data['sources
'], video_id=this_video_id, m3u8_id=m3u8_id,
3490 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
3493 tracks = video_data.get('tracks
')
3494 if tracks and isinstance(tracks, list):
3495 for track in tracks:
3496 if not isinstance(track, dict):
3498 track_kind = track.get('kind
')
3499 if not track_kind or not isinstance(track_kind, str):
3501 if track_kind.lower() not in ('captions
', 'subtitles
'):
3503 track_url = urljoin(base_url, track.get('file'))
3506 subtitles.setdefault(track.get('label
') or 'en
', []).append({
3507 'url
': self._proto_relative_url(track_url)
3511 'id': this_video_id,
3512 'title
': unescapeHTML(video_data['title
'] if require_title else video_data.get('title
')),
3513 'description
': clean_html(video_data.get('description
')),
3514 'thumbnail
': urljoin(base_url, self._proto_relative_url(video_data.get('image
'))),
3515 'timestamp
': int_or_none(video_data.get('pubdate
')),
3516 'duration
': float_or_none(jwplayer_data.get('duration
') or video_data.get('duration
')),
3517 'subtitles
': subtitles,
3519 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
3520 if len(formats) == 1 and re.search(r'^
(?
:http|
//).*(?
:youtube\
.com|youtu\
.be
)/.+', formats[0]['url
']):
3522 '_type
': 'url_transparent
',
3523 'url
': formats[0]['url
'],
3526 self._sort_formats(formats)
3527 entry['formats
'] = formats
3528 entries.append(entry)
3529 if len(entries) == 1:
3532 return self.playlist_result(entries)
3534 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
3535 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3538 for source in jwplayer_sources_data:
3539 if not isinstance(source, dict):
3541 source_url = urljoin(
3542 base_url, self._proto_relative_url(source.get('file')))
3543 if not source_url or source_url in urls:
3545 urls.append(source_url)
3546 source_type = source.get('type') or ''
3547 ext = mimetype2ext(source_type) or determine_ext(source_url)
3548 if source_type == 'hls
' or ext == 'm3u8
':
3549 formats.extend(self._extract_m3u8_formats(
3550 source_url, video_id, 'mp4
', entry_protocol='m3u8_native
',
3551 m3u8_id=m3u8_id, fatal=False))
3552 elif source_type == 'dash
' or ext == 'mpd
':
3553 formats.extend(self._extract_mpd_formats(
3554 source_url, video_id, mpd_id=mpd_id, fatal=False))
3556 formats.extend(self._extract_smil_formats(
3557 source_url, video_id, fatal=False))
3558 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3559 elif source_type.startswith('audio
') or ext in (
3560 'oga
', 'aac
', 'mp3
', 'mpeg
', 'vorbis
'):
3567 height = int_or_none(source.get('height
'))
3569 # Often no height is provided but there is a label in
3570 # format like "1080p", "720p SD", or 1080.
3571 height = int_or_none(self._search_regex(
3572 r'^
(\d{3,4}
)[pP
]?
(?
:\b|$
)', str(source.get('label
') or ''),
3573 'height
', default=None))
3576 'width
': int_or_none(source.get('width
')),
3578 'tbr
': int_or_none(source.get('bitrate
')),
3581 if source_url.startswith('rtmp
'):
3582 a_format['ext
'] = 'flv
'
3583 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3584 # of jwplayer.flash.swf
3585 rtmp_url_parts = re.split(
3586 r'((?
:mp4|mp3|flv
):)', source_url, 1)
3587 if len(rtmp_url_parts) == 3:
3588 rtmp_url, prefix, play_path = rtmp_url_parts
3591 'play_path
': prefix + play_path,
3594 a_format.update(rtmp_params)
3595 formats.append(a_format)
3598 def _live_title(self, name):
3599 self._downloader.deprecation_warning('yt_dlp
.InfoExtractor
._live
_title
is deprecated
and does
not work
as expected
')
3602 def _int(self, v, name, fatal=False, **kwargs):
3603 res = int_or_none(v, **kwargs)
3605 msg = f'Failed to extract {name}
: Could
not parse value {v!r}
'
3607 raise ExtractorError(msg)
3609 self.report_warning(msg)
3612 def _float(self, v, name, fatal=False, **kwargs):
3613 res = float_or_none(v, **kwargs)
3615 msg = f'Failed to extract {name}
: Could
not parse value {v!r}
'
3617 raise ExtractorError(msg)
3619 self.report_warning(msg)
3622 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
3623 path='/', secure=False, discard=False, rest={}, **kwargs):
3624 cookie = http.cookiejar.Cookie(
3625 0, name, value, port, port is not None, domain, True,
3626 domain.startswith('.'), path, True, secure, expire_time,
3627 discard, None, None, rest)
3628 self.cookiejar.set_cookie(cookie)
3630 def _get_cookies(self, url):
3631 """ Return a http.cookies.SimpleCookie with the cookies for the url """
3632 return http.cookies.SimpleCookie(self._downloader._calc_cookies(url))
3634 def _apply_first_set_cookie_header(self, url_handle, cookie):
3636 Apply first Set-Cookie header instead of the last. Experimental.
3638 Some sites (e.g. [1-3]) may serve two cookies under the same name
3639 in Set-Cookie header and expect the first (old) one to be set rather
3640 than second (new). However, as of RFC6265 the newer one cookie
3641 should be set into cookie store what actually happens.
3642 We will workaround this issue by resetting the cookie to
3643 the first one manually.
3644 1. https://new.vk.com/
3645 2. https://github.com/ytdl-org/youtube-dl/issues/9841#issuecomment-227871201
3646 3. https://learning.oreilly.com/
3648 for header, cookies in url_handle.headers.items():
3649 if header.lower() != 'set-cookie
':
3651 cookies = cookies.encode('iso
-8859-1').decode('utf
-8')
3652 cookie_value = re.search(
3653 r'%s=(.+?
);.*?
\b[Dd
]omain
=(.+?
)(?
:[,;]|$
)' % cookie, cookies)
3655 value, domain = cookie_value.groups()
3656 self._set_cookie(domain, cookie, value)
3660 def get_testcases(cls, include_onlymatching=False):
3661 t = getattr(cls, '_TEST
', None)
3663 assert not hasattr(cls, '_TESTS
'), f'{cls.ie_key()}IE has _TEST
and _TESTS
'
3666 tests = getattr(cls, '_TESTS
', [])
3668 if not include_onlymatching and t.get('only_matching
', False):
3670 t['name
'] = cls.ie_key()
3674 def get_webpage_testcases(cls):
3675 tests = getattr(cls, '_WEBPAGE_TESTS
', [])
3677 t['name
'] = cls.ie_key()
3682 """Get age limit from the testcases"""
3683 return max(traverse_obj(
3684 (*cls.get_testcases(include_onlymatching=False), *cls.get_webpage_testcases()),
3685 (..., (('playlist
', 0), None), 'info_dict
', 'age_limit
')) or [0])
3688 def is_suitable(cls, age_limit):
3689 """Test whether the extractor is generally suitable for the given age limit"""
3690 return not age_restricted(cls.age_limit, age_limit)
3693 def description(cls, *, markdown=True, search_examples=None):
3694 """Description of the extractor"""
3696 if cls._NETRC_MACHINE:
3698 desc += f' [<abbr title
="netrc machine"><em
>{cls._NETRC_MACHINE}
</em
></abbr
>]'
3700 desc += f' [{cls._NETRC_MACHINE}
]'
3701 if cls.IE_DESC is False:
3704 desc += f' {cls.IE_DESC}
'
3706 desc += f'; "{cls.SEARCH_KEY}:" prefix
'
3708 _COUNTS = ('', '5', '10', 'all
')
3709 desc += f' (e
.g
. "{cls.SEARCH_KEY}{random.choice(_COUNTS)}:{random.choice(search_examples)}")'
3710 if not cls.working():
3711 desc += ' (**Currently broken
**)' if markdown else ' (Currently broken
)'
3713 name = f' - **{cls.IE_NAME}
**' if markdown else cls.IE_NAME
3714 return f'{name}
:{desc}
' if desc else name
3716 def extract_subtitles(self, *args, **kwargs):
3717 if (self.get_param('writesubtitles
', False)
3718 or self.get_param('listsubtitles
')):
3719 return self._get_subtitles(*args, **kwargs)
3722 def _get_subtitles(self, *args, **kwargs):
3723 raise NotImplementedError('This method must be implemented by subclasses
')
3725 def extract_comments(self, *args, **kwargs):
3726 if not self.get_param('getcomments
'):
3728 generator = self._get_comments(*args, **kwargs)
3735 comments.append(next(generator))
3736 except StopIteration:
3738 except KeyboardInterrupt:
3739 self.to_screen('Interrupted by user
')
3740 except Exception as e:
3741 if self.get_param('ignoreerrors
') is not True:
3743 self._downloader.report_error(e)
3744 comment_count = len(comments)
3745 self.to_screen(f'Extracted {comment_count} comments
')
3747 'comments
': comments,
3748 'comment_count
': None if interrupted else comment_count
3752 def _get_comments(self, *args, **kwargs):
3753 raise NotImplementedError('This method must be implemented by subclasses
')
3756 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
3757 """ Merge subtitle items for one language. Items with duplicated URLs/data
3758 will be dropped. """
3759 list1_data = {(item.get('url'), item.get('data')) for item in subtitle_list1}
3760 ret = list(subtitle_list1)
3761 ret.extend(item for item in subtitle_list2 if (item.get('url
'), item.get('data
')) not in list1_data)
3765 def _merge_subtitles(cls, *dicts, target=None):
3766 """ Merge subtitle dictionaries, language by language. """
3770 for lang, subs in d.items():
3771 target[lang] = cls._merge_subtitle_items(target.get(lang, []), subs)
3774 def extract_automatic_captions(self, *args, **kwargs):
3775 if (self.get_param('writeautomaticsub
', False)
3776 or self.get_param('listsubtitles
')):
3777 return self._get_automatic_captions(*args, **kwargs)
3780 def _get_automatic_captions(self, *args, **kwargs):
3781 raise NotImplementedError('This method must be implemented by subclasses
')
3783 @functools.cached_property
3784 def _cookies_passed(self):
3785 """Whether cookies have been passed to YoutubeDL"""
3786 return self.get_param('cookiefile
') is not None or self.get_param('cookiesfrombrowser
') is not None
3788 def mark_watched(self, *args, **kwargs):
3789 if not self.get_param('mark_watched
', False):
3791 if self.supports_login() and self._get_login_info()[0] is not None or self._cookies_passed:
3792 self._mark_watched(*args, **kwargs)
3794 def _mark_watched(self, *args, **kwargs):
3795 raise NotImplementedError('This method must be implemented by subclasses
')
3797 def geo_verification_headers(self):
3799 geo_verification_proxy = self.get_param('geo_verification_proxy
')
3800 if geo_verification_proxy:
3801 headers['Ytdl
-request
-proxy
'] = geo_verification_proxy
3805 def _generic_id(url):
3806 return urllib.parse.unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
3809 def _generic_title(url):
3810 return urllib.parse.unquote(os.path.splitext(url_basename(url))[0])
3813 def _availability(is_private=None, needs_premium=None, needs_subscription=None, needs_auth=None, is_unlisted=None):
3814 all_known = all(map(
3815 lambda x: x is not None,
3816 (is_private, needs_premium, needs_subscription, needs_auth, is_unlisted)))
3818 'private
' if is_private
3819 else 'premium_only
' if needs_premium
3820 else 'subscriber_only
' if needs_subscription
3821 else 'needs_auth
' if needs_auth
3822 else 'unlisted
' if is_unlisted
3823 else 'public
' if all_known
3826 def _configuration_arg(self, key, default=NO_DEFAULT, *, ie_key=None, casesense=False):
3828 @returns A list of values for the extractor argument given by "key"
3829 or "default" if no such key is present
3830 @param default The default value to return when the key is not present (default: [])
3831 @param casesense When false, the values are converted to lower case
3834 self._downloader.params, ('extractor_args
', (ie_key or self.ie_key()).lower(), key))
3836 return [] if default is NO_DEFAULT else default
3837 return list(val) if casesense else [x.lower() for x in val]
3839 def _yes_playlist(self, playlist_id, video_id, smuggled_data=None, *, playlist_label='playlist
', video_label='video
'):
3840 if not playlist_id or not video_id:
3843 no_playlist = (smuggled_data or {}).get('force_noplaylist
')
3844 if no_playlist is not None:
3845 return not no_playlist
3847 video_id = '' if video_id is True else f' {video_id}
'
3848 playlist_id = '' if playlist_id is True else f' {playlist_id}
'
3849 if self.get_param('noplaylist
'):
3850 self.to_screen(f'Downloading just the {video_label}{video_id} because of
--no
-playlist
')
3852 self.to_screen(f'Downloading {playlist_label}{playlist_id}
- add
--no
-playlist to download just the {video_label}{video_id}
')
3855 def _error_or_warning(self, err, _count=None, _retries=0, *, fatal=True):
3856 RetryManager.report_retry(err, _count or int(fatal), _retries, info=self.to_screen, warn=self.report_warning,
3857 sleep_func=self.get_param('retry_sleep_functions
', {}).get('extractor
'))
3859 def RetryManager(self, **kwargs):
3860 return RetryManager(self.get_param('extractor_retries
', 3), self._error_or_warning, **kwargs)
3863 def extract_from_webpage(cls, ydl, url, webpage):
3864 ie = (cls if isinstance(cls._extract_from_webpage, types.MethodType)
3865 else ydl.get_info_extractor(cls.ie_key()))
3866 for info in ie._extract_from_webpage(url, webpage) or []:
3867 # url = None since we do not want to set (webpage/original)_url
3868 ydl.add_default_extra_info(info, ie, None)
3872 def _extract_from_webpage(cls, url, webpage):
3873 for embed_url in orderedSet(
3874 cls._extract_embed_urls(url, webpage) or [], lazy=True):
3875 yield cls.url_result(embed_url, cls)
3878 def _extract_embed_urls(cls, url, webpage):
3879 """@returns all the embed urls on the webpage"""
3880 if '_EMBED_URL_RE
' not in cls.__dict__:
3881 assert isinstance(cls._EMBED_REGEX, (list, tuple))
3882 for idx, regex in enumerate(cls._EMBED_REGEX):
3883 assert regex.count('(?P
<url
>') == 1, \
3884 f'{cls.__name__}
._EMBED
_REGEX
[{idx}
] must have exactly
1 url group
\n\t{regex}
'
3885 cls._EMBED_URL_RE = tuple(map(re.compile, cls._EMBED_REGEX))
3887 for regex in cls._EMBED_URL_RE:
3888 for mobj in regex.finditer(webpage):
3889 embed_url = urllib.parse.urljoin(url, unescapeHTML(mobj.group('url
')))
3890 if cls._VALID_URL is False or cls.suitable(embed_url):
3893 class StopExtraction(Exception):
3897 def _extract_url(cls, webpage): # TODO: Remove
3898 """Only for compatibility with some older extractors"""
3899 return next(iter(cls._extract_embed_urls(None, webpage) or []), None)
3902 class SearchInfoExtractor(InfoExtractor):
3904 Base class for paged search queries extractors.
3905 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
3906 Instances should define _SEARCH_KEY and optionally _MAX_RESULTS
3909 _MAX_RESULTS = float('inf
')
3912 def _VALID_URL(cls):
3913 return r'%s(?P
<prefix
>|
[1-9][0-9]*|all
):(?P
<query
>[\s\S
]+)' % cls._SEARCH_KEY
3915 def _real_extract(self, query):
3916 prefix, query = self._match_valid_url(query).group('prefix
', 'query
')
3918 return self._get_n_results(query, 1)
3919 elif prefix == 'all
':
3920 return self._get_n_results(query, self._MAX_RESULTS)
3924 raise ExtractorError(f'invalid download number {n}
for query
"{query}"')
3925 elif n > self._MAX_RESULTS:
3926 self.report_warning('%s returns
max %i results (you requested
%i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3927 n = self._MAX_RESULTS
3928 return self._get_n_results(query, n)
3930 def _get_n_results(self, query, n):
3931 """Get a specified number of results for a query.
3932 Either this function or _search_results must be overridden by subclasses """
3933 return self.playlist_result(
3934 itertools.islice(self._search_results(query), 0, None if n == float('inf
') else n),
3937 def _search_results(self, query):
3938 """Returns an iterator of search results"""
3939 raise NotImplementedError('This method must be implemented by subclasses
')
3942 def SEARCH_KEY(cls):
3943 return cls._SEARCH_KEY