21 import xml
.etree
.ElementTree
23 from ..compat
import functools
# isort: split
24 from ..compat
import compat_etree_fromstring
, compat_expanduser
, compat_os_name
25 from ..cookies
import LenientSimpleCookie
26 from ..downloader
import FileDownloader
27 from ..downloader
.f4m
import get_base_url
, remove_encrypted_media
64 parse_m3u8_attributes
,
91 """Information Extractor class.
93 Information extractors are the classes that, given a URL, extract
94 information about the video (or videos) the URL refers to. This
95 information includes the real video URL, the video title, author and
96 others. The information is stored in a dictionary which is then
97 passed to the YoutubeDL. The YoutubeDL processes this
98 information possibly downloading the video to the file system, among
99 other possible outcomes.
101 The type field determines the type of the result.
102 By far the most common value (and the default if _type is missing) is
103 "video", which indicates a single video.
105 For a video, the dictionaries must include the following fields:
107 id: Video identifier.
108 title: Video title, unescaped. Set to an empty string if video has
109 no title as opposed to "None" which signifies that the
110 extractor failed to obtain a title
112 Additionally, it must contain either a formats entry or a url one:
114 formats: A list of dictionaries for each format available, ordered
115 from worst to best quality.
118 * url The mandatory URL representing the media:
119 for plain file media - HTTP URL of this file,
121 for HLS - URL of the M3U8 media playlist,
122 for HDS - URL of the F4M manifest,
124 - HTTP URL to plain file media (in case of
126 - URL of the MPD manifest or base URL
127 representing the media if MPD manifest
128 is parsed from a string (in case of
130 for MSS - URL of the ISM manifest.
132 The URL of the manifest file in case of
134 for HLS - URL of the M3U8 master playlist,
135 for HDS - URL of the F4M manifest,
136 for DASH - URL of the MPD manifest,
137 for MSS - URL of the ISM manifest.
138 * manifest_stream_number (For internal use only)
139 The index of the stream in the manifest file
140 * ext Will be calculated from URL if missing
141 * format A human-readable description of the format
142 ("mp4 container with h264/opus").
143 Calculated from the format_id, width, height.
144 and format_note fields if missing.
145 * format_id A short description of the format
146 ("mp4_h264_opus" or "19").
147 Technically optional, but strongly recommended.
148 * format_note Additional info about the format
149 ("3D" or "DASH video")
150 * width Width of the video, if known
151 * height Height of the video, if known
152 * resolution Textual description of width and height
153 * dynamic_range The dynamic range of the video. One of:
154 "SDR" (None), "HDR10", "HDR10+, "HDR12", "HLG, "DV"
155 * tbr Average bitrate of audio and video in KBit/s
156 * abr Average audio bitrate in KBit/s
157 * acodec Name of the audio codec in use
158 * asr Audio sampling rate in Hertz
159 * audio_channels Number of audio channels
160 * vbr Average video bitrate in KBit/s
162 * vcodec Name of the video codec in use
163 * container Name of the container format
164 * filesize The number of bytes, if known in advance
165 * filesize_approx An estimate for the number of bytes
166 * player_url SWF Player URL (used for rtmpdump).
167 * protocol The protocol that will be used for the actual
168 download, lower-case. One of "http", "https" or
169 one of the protocols defined in downloader.PROTOCOL_MAP
171 Base URL for fragments. Each fragment's path
172 value (if present) will be relative to
174 * fragments A list of fragments of a fragmented media.
175 Each fragment entry must contain either an url
176 or a path. If an url is present it should be
177 considered by a client. Otherwise both path and
178 fragment_base_url must be present. Here is
179 the list of all potential fields:
180 * "url" - fragment's URL
181 * "path" - fragment's path relative to
183 * "duration" (optional, int or float)
184 * "filesize" (optional, int)
185 * is_from_start Is a live format that can be downloaded
186 from the start. Boolean
187 * preference Order number of this format. If this field is
188 present and not None, the formats get sorted
189 by this field, regardless of all other values.
190 -1 for default (order by other properties),
191 -2 or smaller for less than default.
192 < -1000 to hide the format (if there is
193 another one which is strictly better)
194 * language Language code, e.g. "de" or "en-US".
195 * language_preference Is this in the language mentioned in
197 10 if it's what the URL is about,
198 -1 for default (don't know),
199 -10 otherwise, other values reserved for now.
200 * quality Order number of the video quality of this
201 format, irrespective of the file format.
202 -1 for default (order by other properties),
203 -2 or smaller for less than default.
204 * source_preference Order number for this video source
205 (quality takes higher priority)
206 -1 for default (order by other properties),
207 -2 or smaller for less than default.
208 * http_headers A dictionary of additional HTTP headers
209 to add to the request.
210 * stretched_ratio If given and not 1, indicates that the
211 video's pixels are not square.
212 width : height ratio as float.
213 * no_resume The server does not support resuming the
214 (HTTP or RTMP) download. Boolean.
215 * has_drm The format has DRM and cannot be downloaded. Boolean
216 * downloader_options A dictionary of downloader options
217 (For internal use only)
218 * http_chunk_size Chunk size for HTTP downloads
219 * ffmpeg_args Extra arguments for ffmpeg downloader
220 RTMP formats can also have the additional fields: page_url,
221 app, play_path, tc_url, flash_version, rtmp_live, rtmp_conn,
222 rtmp_protocol, rtmp_real_time
224 url: Final video URL.
225 ext: Video filename extension.
226 format: The video format, defaults to ext (used for --get-format)
227 player_url: SWF Player URL (used for rtmpdump).
229 The following fields are optional:
231 direct: True if a direct video file was given (must only be set by GenericIE)
232 alt_title: A secondary title of the video.
233 display_id An alternative identifier for the video, not necessarily
234 unique, but available before title. Typically, id is
235 something like "4234987", title "Dancing naked mole rats",
236 and display_id "dancing-naked-mole-rats"
237 thumbnails: A list of dictionaries, with the following entries:
238 * "id" (optional, string) - Thumbnail format ID
240 * "preference" (optional, int) - quality of the image
241 * "width" (optional, int)
242 * "height" (optional, int)
243 * "resolution" (optional, string "{width}x{height}",
245 * "filesize" (optional, int)
246 * "http_headers" (dict) - HTTP headers for the request
247 thumbnail: Full URL to a video thumbnail image.
248 description: Full video description.
249 uploader: Full name of the video uploader.
250 license: License name the video is licensed under.
251 creator: The creator of the video.
252 timestamp: UNIX timestamp of the moment the video was uploaded
253 upload_date: Video upload date in UTC (YYYYMMDD).
254 If not explicitly set, calculated from timestamp
255 release_timestamp: UNIX timestamp of the moment the video was released.
256 If it is not clear whether to use timestamp or this, use the former
257 release_date: The date (YYYYMMDD) when the video was released in UTC.
258 If not explicitly set, calculated from release_timestamp
259 modified_timestamp: UNIX timestamp of the moment the video was last modified.
260 modified_date: The date (YYYYMMDD) when the video was last modified in UTC.
261 If not explicitly set, calculated from modified_timestamp
262 uploader_id: Nickname or id of the video uploader.
263 uploader_url: Full URL to a personal webpage of the video uploader.
264 channel: Full name of the channel the video is uploaded on.
265 Note that channel fields may or may not repeat uploader
266 fields. This depends on a particular extractor.
267 channel_id: Id of the channel.
268 channel_url: Full URL to a channel webpage.
269 channel_follower_count: Number of followers of the channel.
270 location: Physical location where the video was filmed.
271 subtitles: The available subtitles as a dictionary in the format
272 {tag: subformats}. "tag" is usually a language code, and
273 "subformats" is a list sorted from lower to higher
274 preference, each element is a dictionary with the "ext"
276 * "data": The subtitles file contents
277 * "url": A URL pointing to the subtitles file
278 It can optionally also have:
279 * "name": Name or description of the subtitles
280 * "http_headers": A dictionary of additional HTTP headers
281 to add to the request.
282 "ext" will be calculated from URL if missing
283 automatic_captions: Like 'subtitles'; contains automatically generated
284 captions instead of normal subtitles
285 duration: Length of the video in seconds, as an integer or float.
286 view_count: How many users have watched the video on the platform.
287 concurrent_view_count: How many users are currently watching the video on the platform.
288 like_count: Number of positive ratings of the video
289 dislike_count: Number of negative ratings of the video
290 repost_count: Number of reposts of the video
291 average_rating: Average rating give by users, the scale used depends on the webpage
292 comment_count: Number of comments on the video
293 comments: A list of comments, each with one or more of the following
294 properties (all but one of text or html optional):
295 * "author" - human-readable name of the comment author
296 * "author_id" - user ID of the comment author
297 * "author_thumbnail" - The thumbnail of the comment author
299 * "html" - Comment as HTML
300 * "text" - Plain text of the comment
301 * "timestamp" - UNIX timestamp of comment
302 * "parent" - ID of the comment this one is replying to.
303 Set to "root" to indicate that this is a
304 comment to the original video.
305 * "like_count" - Number of positive ratings of the comment
306 * "dislike_count" - Number of negative ratings of the comment
307 * "is_favorited" - Whether the comment is marked as
308 favorite by the video uploader
309 * "author_is_uploader" - Whether the comment is made by
311 age_limit: Age restriction for the video, as an integer (years)
312 webpage_url: The URL to the video webpage, if given to yt-dlp it
313 should allow to get the same result again. (It will be set
314 by YoutubeDL if it's missing)
315 categories: A list of categories that the video falls in, for example
317 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
318 cast: A list of the video cast
319 is_live: True, False, or None (=unknown). Whether this video is a
320 live stream that goes on instead of a fixed-length video.
321 was_live: True, False, or None (=unknown). Whether this video was
322 originally a live stream.
323 live_status: None (=unknown), 'is_live', 'is_upcoming', 'was_live', 'not_live',
324 or 'post_live' (was live, but VOD is not yet processed)
325 If absent, automatically set from is_live, was_live
326 start_time: Time in seconds where the reproduction should start, as
327 specified in the URL.
328 end_time: Time in seconds where the reproduction should end, as
329 specified in the URL.
330 chapters: A list of dictionaries, with the following entries:
331 * "start_time" - The start time of the chapter in seconds
332 * "end_time" - The end time of the chapter in seconds
333 * "title" (optional, string)
334 playable_in_embed: Whether this video is allowed to play in embedded
335 players on other sites. Can be True (=always allowed),
336 False (=never allowed), None (=unknown), or a string
337 specifying the criteria for embedability; e.g. 'whitelist'
338 availability: Under what condition the video is available. One of
339 'private', 'premium_only', 'subscriber_only', 'needs_auth',
340 'unlisted' or 'public'. Use 'InfoExtractor._availability'
342 _old_archive_ids: A list of old archive ids needed for backward compatibility
343 __post_extractor: A function to be called just before the metadata is
344 written to either disk, logger or console. The function
345 must return a dict which will be added to the info_dict.
346 This is usefull for additional information that is
347 time-consuming to extract. Note that the fields thus
348 extracted will not be available to output template and
349 match_filter. So, only "comments" and "comment_count" are
350 currently allowed to be extracted via this method.
352 The following fields should only be used when the video belongs to some logical
355 chapter: Name or title of the chapter the video belongs to.
356 chapter_number: Number of the chapter the video belongs to, as an integer.
357 chapter_id: Id of the chapter the video belongs to, as a unicode string.
359 The following fields should only be used when the video is an episode of some
360 series, programme or podcast:
362 series: Title of the series or programme the video episode belongs to.
363 series_id: Id of the series or programme the video episode belongs to, as a unicode string.
364 season: Title of the season the video episode belongs to.
365 season_number: Number of the season the video episode belongs to, as an integer.
366 season_id: Id of the season the video episode belongs to, as a unicode string.
367 episode: Title of the video episode. Unlike mandatory video title field,
368 this field should denote the exact title of the video episode
369 without any kind of decoration.
370 episode_number: Number of the video episode within a season, as an integer.
371 episode_id: Id of the video episode, as a unicode string.
373 The following fields should only be used when the media is a track or a part of
376 track: Title of the track.
377 track_number: Number of the track within an album or a disc, as an integer.
378 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
380 artist: Artist(s) of the track.
381 genre: Genre(s) of the track.
382 album: Title of the album the track belongs to.
383 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
384 album_artist: List of all artists appeared on the album (e.g.
385 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
387 disc_number: Number of the disc or other physical medium the track belongs to,
389 release_year: Year (YYYY) when the album was released.
390 composer: Composer of the piece
392 The following fields should only be set for clips that should be cut from the original video:
394 section_start: Start time of the section in seconds
395 section_end: End time of the section in seconds
397 The following fields should only be set for storyboards:
398 rows: Number of rows in each storyboard fragment, as an integer
399 columns: Number of columns in each storyboard fragment, as an integer
401 Unless mentioned otherwise, the fields should be Unicode strings.
403 Unless mentioned otherwise, None is equivalent to absence of information.
406 _type "playlist" indicates multiple videos.
407 There must be a key "entries", which is a list, an iterable, or a PagedList
408 object, each element of which is a valid dictionary by this specification.
410 Additionally, playlists can have "id", "title", and any other relevant
411 attributes with the same semantics as videos (see above).
413 It can also have the following optional fields:
415 playlist_count: The total number of videos in a playlist. If not given,
416 YoutubeDL tries to calculate it from "entries"
419 _type "multi_video" indicates that there are multiple videos that
420 form a single show, for examples multiple acts of an opera or TV episode.
421 It must have an entries key like a playlist and contain all the keys
422 required for a video at the same time.
425 _type "url" indicates that the video must be extracted from another
426 location, possibly by a different extractor. Its only required key is:
427 "url" - the next URL to extract.
428 The key "ie_key" can be set to the class name (minus the trailing "IE",
429 e.g. "Youtube") if the extractor class is known in advance.
430 Additionally, the dictionary may have any properties of the resolved entity
431 known in advance, for example "title" if the title of the referred video is
435 _type "url_transparent" entities have the same specification as "url", but
436 indicate that the given additional information is more precise than the one
437 associated with the resolved URL.
438 This is useful when a site employs a video service that hosts the video and
439 its technical metadata, but that video service does not embed a useful
440 title, description etc.
443 Subclasses of this should also be added to the list of extractors and
444 should define a _VALID_URL regexp and, re-define the _real_extract() and
445 (optionally) _real_initialize() methods.
447 Subclasses may also override suitable() if necessary, but ensure the function
448 signature is preserved and that this function imports everything it needs
449 (except other extractors), so that lazy_extractors works correctly.
451 Subclasses can define a list of _EMBED_REGEX, which will be searched for in
452 the HTML of Generic webpages. It may also override _extract_embed_urls
453 or _extract_from_webpage as necessary. While these are normally classmethods,
454 _extract_from_webpage is allowed to be an instance method.
456 _extract_from_webpage may raise self.StopExtraction() to stop further
457 processing of the webpage and obtain exclusive rights to it. This is useful
458 when the extractor cannot reliably be matched using just the URL,
459 e.g. invidious/peertube instances
461 Embed-only extractors can be defined by setting _VALID_URL = False.
463 To support username + password (or netrc) login, the extractor must define a
464 _NETRC_MACHINE and re-define _perform_login(username, password) and
465 (optionally) _initialize_pre_login() methods. The _perform_login method will
466 be called between _initialize_pre_login and _real_initialize if credentials
467 are passed by the user. In cases where it is necessary to have the login
468 process as part of the extraction rather than initialization, _perform_login
469 can be left undefined.
471 _GEO_BYPASS attribute may be set to False in order to disable
472 geo restriction bypass mechanisms for a particular extractor.
473 Though it won't disable explicit geo restriction bypass based on
474 country code provided with geo_bypass_country.
476 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
477 countries for this extractor. One of these countries will be used by
478 geo restriction bypass mechanism right away in order to bypass
479 geo restriction, of course, if the mechanism is not disabled.
481 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
482 IP blocks in CIDR notation for this extractor. One of these IP blocks
483 will be used by geo restriction bypass mechanism similarly
486 The _ENABLED attribute should be set to False for IEs that
487 are disabled by default and must be explicitly enabled.
489 The _WORKING attribute should be set to False for broken IEs
490 in order to warn the users and skip the tests.
495 _x_forwarded_for_ip
= None
497 _GEO_COUNTRIES
= None
498 _GEO_IP_BLOCKS
= None
501 _NETRC_MACHINE
= None
507 def _login_hint(self
, method
=NO_DEFAULT
, netrc
=None):
508 password_hint
= f
'--username and --password, or --netrc ({netrc or self._NETRC_MACHINE}) to provide account credentials'
511 'any': f
'Use --cookies, --cookies-from-browser, {password_hint}',
512 'password': f
'Use {password_hint}',
514 'Use --cookies-from-browser or --cookies for the authentication. '
515 'See https://github.com/yt-dlp/yt-dlp/wiki/FAQ#how-do-i-pass-cookies-to-yt-dlp for how to manually pass cookies'),
516 }[method
if method
is not NO_DEFAULT
else 'any' if self
.supports_login() else 'cookies']
518 def __init__(self
, downloader
=None):
519 """Constructor. Receives an optional downloader (a YoutubeDL instance).
520 If a downloader is not passed during initialization,
521 it must be set using "set_downloader()" before "extract()" is called"""
523 self
._x
_forwarded
_for
_ip
= None
524 self
._printed
_messages
= set()
525 self
.set_downloader(downloader
)
528 def _match_valid_url(cls
, url
):
529 if cls
._VALID
_URL
is False:
531 # This does not use has/getattr intentionally - we want to know whether
532 # we have cached the regexp for *this* class, whereas getattr would also
533 # match the superclass
534 if '_VALID_URL_RE' not in cls
.__dict
__:
535 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
536 return cls
._VALID
_URL
_RE
.match(url
)
539 def suitable(cls
, url
):
540 """Receives a URL and returns True if suitable for this IE."""
541 # This function must import everything it needs (except other extractors),
542 # so that lazy_extractors works correctly
543 return cls
._match
_valid
_url
(url
) is not None
546 def _match_id(cls
, url
):
547 return cls
._match
_valid
_url
(url
).group('id')
550 def get_temp_id(cls
, url
):
552 return cls
._match
_id
(url
)
553 except (IndexError, AttributeError):
558 """Getter method for _WORKING."""
562 def supports_login(cls
):
563 return bool(cls
._NETRC
_MACHINE
)
565 def initialize(self
):
566 """Initializes an instance (authentication, etc)."""
567 self
._printed
_messages
= set()
568 self
._initialize
_geo
_bypass
({
569 'countries': self
._GEO
_COUNTRIES
,
570 'ip_blocks': self
._GEO
_IP
_BLOCKS
,
573 self
._initialize
_pre
_login
()
574 if self
.supports_login():
575 username
, password
= self
._get
_login
_info
()
577 self
._perform
_login
(username
, password
)
578 elif self
.get_param('username') and False not in (self
.IE_DESC
, self
._NETRC
_MACHINE
):
579 self
.report_warning(f
'Login with password is not supported for this website. {self._login_hint("cookies")}')
580 self
._real
_initialize
()
583 def _initialize_geo_bypass(self
, geo_bypass_context
):
585 Initialize geo restriction bypass mechanism.
587 This method is used to initialize geo bypass mechanism based on faking
588 X-Forwarded-For HTTP header. A random country from provided country list
589 is selected and a random IP belonging to this country is generated. This
590 IP will be passed as X-Forwarded-For HTTP header in all subsequent
593 This method will be used for initial geo bypass mechanism initialization
594 during the instance initialization with _GEO_COUNTRIES and
597 You may also manually call it from extractor's code if geo bypass
598 information is not available beforehand (e.g. obtained during
599 extraction) or due to some other reason. In this case you should pass
600 this information in geo bypass context passed as first argument. It may
601 contain following fields:
603 countries: List of geo unrestricted countries (similar
605 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
606 (similar to _GEO_IP_BLOCKS)
609 if not self
._x
_forwarded
_for
_ip
:
611 # Geo bypass mechanism is explicitly disabled by user
612 if not self
.get_param('geo_bypass', True):
615 if not geo_bypass_context
:
616 geo_bypass_context
= {}
618 # Backward compatibility: previously _initialize_geo_bypass
619 # expected a list of countries, some 3rd party code may still use
621 if isinstance(geo_bypass_context
, (list, tuple)):
622 geo_bypass_context
= {
623 'countries': geo_bypass_context
,
626 # The whole point of geo bypass mechanism is to fake IP
627 # as X-Forwarded-For HTTP header based on some IP block or
630 # Path 1: bypassing based on IP block in CIDR notation
632 # Explicit IP block specified by user, use it right away
633 # regardless of whether extractor is geo bypassable or not
634 ip_block
= self
.get_param('geo_bypass_ip_block', None)
636 # Otherwise use random IP block from geo bypass context but only
637 # if extractor is known as geo bypassable
639 ip_blocks
= geo_bypass_context
.get('ip_blocks')
640 if self
._GEO
_BYPASS
and ip_blocks
:
641 ip_block
= random
.choice(ip_blocks
)
644 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(ip_block
)
645 self
.write_debug(f
'Using fake IP {self._x_forwarded_for_ip} as X-Forwarded-For')
648 # Path 2: bypassing based on country code
650 # Explicit country code specified by user, use it right away
651 # regardless of whether extractor is geo bypassable or not
652 country
= self
.get_param('geo_bypass_country', None)
654 # Otherwise use random country code from geo bypass context but
655 # only if extractor is known as geo bypassable
657 countries
= geo_bypass_context
.get('countries')
658 if self
._GEO
_BYPASS
and countries
:
659 country
= random
.choice(countries
)
662 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country
)
663 self
._downloader
.write_debug(
664 f
'Using fake IP {self._x_forwarded_for_ip} ({country.upper()}) as X-Forwarded-For')
666 def extract(self
, url
):
667 """Extracts URL information and returns it in list of dicts."""
672 self
.write_debug('Extracting URL: %s' % url
)
673 ie_result
= self
._real
_extract
(url
)
674 if ie_result
is None:
676 if self
._x
_forwarded
_for
_ip
:
677 ie_result
['__x_forwarded_for_ip'] = self
._x
_forwarded
_for
_ip
678 subtitles
= ie_result
.get('subtitles') or {}
679 if 'no-live-chat' in self
.get_param('compat_opts'):
680 for lang
in ('live_chat', 'comments', 'danmaku'):
681 subtitles
.pop(lang
, None)
683 except GeoRestrictedError
as e
:
684 if self
.__maybe
_fake
_ip
_and
_retry
(e
.countries
):
687 except UnsupportedError
:
689 except ExtractorError
as e
:
691 'video_id': e
.video_id
or self
.get_temp_id(url
),
693 'tb': e
.traceback
or sys
.exc_info()[2],
694 'expected': e
.expected
,
697 if hasattr(e
, 'countries'):
698 kwargs
['countries'] = e
.countries
699 raise type(e
)(e
.orig_msg
, **kwargs
)
700 except http
.client
.IncompleteRead
as e
:
701 raise ExtractorError('A network error has occurred.', cause
=e
, expected
=True, video_id
=self
.get_temp_id(url
))
702 except (KeyError, StopIteration) as e
:
703 raise ExtractorError('An extractor error has occurred.', cause
=e
, video_id
=self
.get_temp_id(url
))
705 def __maybe_fake_ip_and_retry(self
, countries
):
706 if (not self
.get_param('geo_bypass_country', None)
708 and self
.get_param('geo_bypass', True)
709 and not self
._x
_forwarded
_for
_ip
711 country_code
= random
.choice(countries
)
712 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country_code
)
713 if self
._x
_forwarded
_for
_ip
:
715 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
716 % (self
._x
_forwarded
_for
_ip
, country_code
.upper()))
720 def set_downloader(self
, downloader
):
721 """Sets a YoutubeDL instance as the downloader for this IE."""
722 self
._downloader
= downloader
726 return self
._downloader
.cache
730 return self
._downloader
.cookiejar
732 def _initialize_pre_login(self
):
733 """ Initialization before login. Redefine in subclasses."""
736 def _perform_login(self
, username
, password
):
737 """ Login with username and password. Redefine in subclasses."""
740 def _real_initialize(self
):
741 """Real initialization process. Redefine in subclasses."""
744 def _real_extract(self
, url
):
745 """Real extraction process. Redefine in subclasses."""
746 raise NotImplementedError('This method must be implemented by subclasses')
750 """A string for getting the InfoExtractor with get_info_extractor"""
751 return cls
.__name
__[:-2]
755 return cls
.__name
__[:-2]
758 def __can_accept_status_code(err
, expected_status
):
759 assert isinstance(err
, urllib
.error
.HTTPError
)
760 if expected_status
is None:
762 elif callable(expected_status
):
763 return expected_status(err
.code
) is True
765 return err
.code
in variadic(expected_status
)
767 def _create_request(self
, url_or_request
, data
=None, headers
=None, query
=None):
768 if isinstance(url_or_request
, urllib
.request
.Request
):
769 return update_Request(url_or_request
, data
=data
, headers
=headers
, query
=query
)
771 url_or_request
= update_url_query(url_or_request
, query
)
772 return sanitized_Request(url_or_request
, data
, headers
or {})
774 def _request_webpage(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, data
=None, headers
=None, query
=None, expected_status
=None):
776 Return the response handle.
778 See _download_webpage docstring for arguments specification.
780 if not self
._downloader
._first
_webpage
_request
:
781 sleep_interval
= self
.get_param('sleep_interval_requests') or 0
782 if sleep_interval
> 0:
783 self
.to_screen('Sleeping %s seconds ...' % sleep_interval
)
784 time
.sleep(sleep_interval
)
786 self
._downloader
._first
_webpage
_request
= False
789 self
.report_download_webpage(video_id
)
790 elif note
is not False:
792 self
.to_screen(str(note
))
794 self
.to_screen(f
'{video_id}: {note}')
796 # Some sites check X-Forwarded-For HTTP header in order to figure out
797 # the origin of the client behind proxy. This allows bypassing geo
798 # restriction by faking this header's value to IP that belongs to some
799 # geo unrestricted country. We will do so once we encounter any
800 # geo restriction error.
801 if self
._x
_forwarded
_for
_ip
:
802 headers
= (headers
or {}).copy()
803 headers
.setdefault('X-Forwarded-For', self
._x
_forwarded
_for
_ip
)
806 return self
._downloader
.urlopen(self
._create
_request
(url_or_request
, data
, headers
, query
))
807 except network_exceptions
as err
:
808 if isinstance(err
, urllib
.error
.HTTPError
):
809 if self
.__can
_accept
_status
_code
(err
, expected_status
):
810 # Retain reference to error to prevent file object from
811 # being closed before it can be read. Works around the
812 # effects of <https://bugs.python.org/issue15002>
813 # introduced in Python 3.4.1.
820 errnote
= 'Unable to download webpage'
822 errmsg
= f
'{errnote}: {error_to_compat_str(err)}'
824 raise ExtractorError(errmsg
, cause
=err
)
826 self
.report_warning(errmsg
)
829 def _download_webpage_handle(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True,
830 encoding
=None, data
=None, headers
={}, query={}
, expected_status
=None):
832 Return a tuple (page content as string, URL handle).
835 url_or_request -- plain text URL as a string or
836 a urllib.request.Request object
837 video_id -- Video/playlist/item identifier (string)
840 note -- note printed before downloading (string)
841 errnote -- note printed in case of an error (string)
842 fatal -- flag denoting whether error should be considered fatal,
843 i.e. whether it should cause ExtractionError to be raised,
844 otherwise a warning will be reported and extraction continued
845 encoding -- encoding for a page content decoding, guessed automatically
846 when not explicitly specified
847 data -- POST data (bytes)
848 headers -- HTTP headers (dict)
849 query -- URL query (dict)
850 expected_status -- allows to accept failed HTTP requests (non 2xx
851 status code) by explicitly specifying a set of accepted status
852 codes. Can be any of the following entities:
853 - an integer type specifying an exact failed status code to
855 - a list or a tuple of integer types specifying a list of
856 failed status codes to accept
857 - a callable accepting an actual failed status code and
858 returning True if it should be accepted
859 Note that this argument does not affect success status codes (2xx)
860 which are always accepted.
863 # Strip hashes from the URL (#1038)
864 if isinstance(url_or_request
, str):
865 url_or_request
= url_or_request
.partition('#')[0]
867 urlh
= self
._request
_webpage
(url_or_request
, video_id
, note
, errnote
, fatal
, data
=data
, headers
=headers
, query
=query
, expected_status
=expected_status
)
871 content
= self
._webpage
_read
_content
(urlh
, url_or_request
, video_id
, note
, errnote
, fatal
, encoding
=encoding
)
872 return (content
, urlh
)
875 def _guess_encoding_from_content(content_type
, webpage_bytes
):
876 m
= re
.match(r
'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type
)
878 encoding
= m
.group(1)
880 m
= re
.search(br
'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
881 webpage_bytes[:1024])
883 encoding = m.group(1).decode('ascii')
884 elif webpage_bytes.startswith(b'\xff\xfe'):
891 def __check_blocked(self, content):
892 first_block = content[:512]
893 if ('<title>Access to this site is blocked</title>' in content
894 and 'Websense' in first_block):
895 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
896 blocked_iframe = self._html_search_regex(
897 r'<iframe src="([^
"]+)"', content,
898 'Websense information URL
', default=None)
900 msg += ' Visit
%s for more details
' % blocked_iframe
901 raise ExtractorError(msg, expected=True)
902 if '<title
>The URL you requested has been blocked
</title
>' in first_block:
904 'Access to this webpage has been blocked by Indian censorship
. '
905 'Use a VPN
or proxy
server (with --proxy
) to route around it
.')
906 block_msg = self._html_search_regex(
907 r'</h1
><p
>(.*?
)</p
>',
908 content, 'block message
', default=None)
910 msg += ' (Message
: "%s")' % block_msg.replace('\n', ' ')
911 raise ExtractorError(msg, expected=True)
912 if ('<title
>TTK
:: Доступ к ресурсу ограничен
</title
>' in content
913 and 'blocklist
.rkn
.gov
.ru
' in content):
914 raise ExtractorError(
915 'Access to this webpage has been blocked by decision of the Russian government
. '
916 'Visit http
://blocklist
.rkn
.gov
.ru
/ for a block reason
.',
919 def _request_dump_filename(self, url, video_id):
920 basen = f'{video_id}_{url}
'
921 trim_length = self.get_param('trim_file_name
') or 240
922 if len(basen) > trim_length:
923 h = '___
' + hashlib.md5(basen.encode('utf
-8')).hexdigest()
924 basen = basen[:trim_length - len(h)] + h
925 filename = sanitize_filename(f'{basen}
.dump
', restricted=True)
926 # Working around MAX_PATH limitation on Windows (see
927 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
928 if compat_os_name == 'nt
':
929 absfilepath = os.path.abspath(filename)
930 if len(absfilepath) > 259:
931 filename = fR'\\?\{absfilepath}
'
934 def __decode_webpage(self, webpage_bytes, encoding, headers):
936 encoding = self._guess_encoding_from_content(headers.get('Content
-Type
', ''), webpage_bytes)
938 return webpage_bytes.decode(encoding, 'replace
')
940 return webpage_bytes.decode('utf
-8', 'replace
')
942 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
943 webpage_bytes = urlh.read()
944 if prefix is not None:
945 webpage_bytes = prefix + webpage_bytes
946 if self.get_param('dump_intermediate_pages
', False):
947 self.to_screen('Dumping request to
' + urlh.geturl())
948 dump = base64.b64encode(webpage_bytes).decode('ascii
')
949 self._downloader.to_screen(dump)
950 if self.get_param('write_pages
'):
951 filename = self._request_dump_filename(urlh.geturl(), video_id)
952 self.to_screen(f'Saving request to {filename}
')
953 with open(filename, 'wb
') as outf:
954 outf.write(webpage_bytes)
956 content = self.__decode_webpage(webpage_bytes, encoding, urlh.headers)
957 self.__check_blocked(content)
961 def __print_error(self, errnote, fatal, video_id, err):
963 raise ExtractorError(f'{video_id}
: {errnote}
', cause=err)
965 self.report_warning(f'{video_id}
: {errnote}
: {err}
')
967 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True, errnote=None):
969 xml_string = transform_source(xml_string)
971 return compat_etree_fromstring(xml_string.encode('utf
-8'))
972 except xml.etree.ElementTree.ParseError as ve:
973 self.__print_error('Failed to parse XML
' if errnote is None else errnote, fatal, video_id, ve)
975 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True, errnote=None, **parser_kwargs):
978 json_string, cls=LenientJSONDecoder, strict=False, transform_source=transform_source, **parser_kwargs)
979 except ValueError as ve:
980 self.__print_error('Failed to parse JSON
' if errnote is None else errnote, fatal, video_id, ve)
982 def _parse_socket_response_as_json(self, data, *args, **kwargs):
983 return self._parse_json(data[data.find('{'):data.rfind('}
') + 1], *args, **kwargs)
985 def __create_download_methods(name, parser, note, errnote, return_value):
987 def parse(ie, content, *args, errnote=errnote, **kwargs):
991 kwargs['errnote
'] = errnote
992 # parser is fetched by name so subclasses can override it
993 return getattr(ie, parser)(content, *args, **kwargs)
995 def download_handle(self, url_or_request, video_id, note=note, errnote=errnote, transform_source=None,
996 fatal=True, encoding=None, data=None, headers={}, query={}, expected_status=None):
997 res = self._download_webpage_handle(
998 url_or_request, video_id, note=note, errnote=errnote, fatal=fatal, encoding=encoding,
999 data=data, headers=headers, query=query, expected_status=expected_status)
1003 return parse(self, content, video_id, transform_source=transform_source, fatal=fatal, errnote=errnote), urlh
1005 def download_content(self, url_or_request, video_id, note=note, errnote=errnote, transform_source=None,
1006 fatal=True, encoding=None, data=None, headers={}, query={}, expected_status=None):
1007 if self.get_param('load_pages
'):
1008 url_or_request = self._create_request(url_or_request, data, headers, query)
1009 filename = self._request_dump_filename(url_or_request.full_url, video_id)
1010 self.to_screen(f'Loading request
from {filename}
')
1012 with open(filename, 'rb
') as dumpf:
1013 webpage_bytes = dumpf.read()
1014 except OSError as e:
1015 self.report_warning(f'Unable to load request
from disk
: {e}
')
1017 content = self.__decode_webpage(webpage_bytes, encoding, url_or_request.headers)
1018 return parse(self, content, video_id, transform_source=transform_source, fatal=fatal, errnote=errnote)
1022 'transform_source
': transform_source,
1024 'encoding
': encoding,
1028 'expected_status
': expected_status,
1031 kwargs.pop('transform_source
')
1032 # The method is fetched by name so subclasses can override _download_..._handle
1033 res = getattr(self, download_handle.__name__)(url_or_request, video_id, **kwargs)
1034 return res if res is False else res[0]
1036 def impersonate(func, name, return_value):
1037 func.__name__, func.__qualname__ = name, f'InfoExtractor
.{name}
'
1039 @param transform_source Apply this transformation before parsing
1040 @returns {return_value}
1042 See _download_webpage_handle docstring for other arguments specification
1045 impersonate(download_handle, f'_download_{name}_handle
', f'({return_value}
, URL handle
)')
1046 impersonate(download_content, f'_download_{name}
', f'{return_value}
')
1047 return download_handle, download_content
1049 _download_xml_handle, _download_xml = __create_download_methods(
1050 'xml
', '_parse_xml
', 'Downloading XML
', 'Unable to download XML
', 'xml
as an xml
.etree
.ElementTree
.Element
')
1051 _download_json_handle, _download_json = __create_download_methods(
1052 'json
', '_parse_json
', 'Downloading JSON metadata
', 'Unable to download JSON metadata
', 'JSON
object as a
dict')
1053 _download_socket_json_handle, _download_socket_json = __create_download_methods(
1054 'socket_json
', '_parse_socket_response_as_json
', 'Polling socket
', 'Unable to poll socket
', 'JSON
object as a
dict')
1055 __download_webpage = __create_download_methods('webpage
', None, None, None, 'data of the page
as a string
')[1]
1057 def _download_webpage(
1058 self, url_or_request, video_id, note=None, errnote=None,
1059 fatal=True, tries=1, timeout=NO_DEFAULT, *args, **kwargs):
1061 Return the data of the page as a string.
1064 tries -- number of tries
1065 timeout -- sleep interval between tries
1067 See _download_webpage_handle docstring for other arguments specification.
1070 R''' # NB: These are unused; should they be deprecated?
1072 self._downloader.deprecation_warning('tries argument
is deprecated
in InfoExtractor
._download
_webpage
')
1073 if timeout is NO_DEFAULT:
1076 self._downloader.deprecation_warning('timeout argument
is deprecated
in InfoExtractor
._download
_webpage
')
1082 return self.__download_webpage(url_or_request, video_id, note, errnote, None, fatal, *args, **kwargs)
1083 except http.client.IncompleteRead as e:
1085 if try_count >= tries:
1087 self._sleep(timeout, video_id)
1089 def report_warning(self, msg, video_id=None, *args, only_once=False, **kwargs):
1090 idstr = format_field(video_id, None, '%s: ')
1091 msg = f'[{self.IE_NAME}
] {idstr}{msg}
'
1093 if f'WARNING
: {msg}
' in self._printed_messages:
1095 self._printed_messages.add(f'WARNING
: {msg}
')
1096 self._downloader.report_warning(msg, *args, **kwargs)
1098 def to_screen(self, msg, *args, **kwargs):
1099 """Print msg to screen, prefixing it with '[ie_name
]'"""
1100 self._downloader.to_screen(f'[{self.IE_NAME}
] {msg}
', *args, **kwargs)
1102 def write_debug(self, msg, *args, **kwargs):
1103 self._downloader.write_debug(f'[{self.IE_NAME}
] {msg}
', *args, **kwargs)
1105 def get_param(self, name, default=None, *args, **kwargs):
1106 if self._downloader:
1107 return self._downloader.params.get(name, default, *args, **kwargs)
1110 def report_drm(self, video_id, partial=False):
1111 self.raise_no_formats('This video
is DRM protected
', expected=True, video_id=video_id)
1113 def report_extraction(self, id_or_name):
1114 """Report information extraction."""
1115 self.to_screen('%s: Extracting information
' % id_or_name)
1117 def report_download_webpage(self, video_id):
1118 """Report webpage download."""
1119 self.to_screen('%s: Downloading webpage
' % video_id)
1121 def report_age_confirmation(self):
1122 """Report attempt to confirm age."""
1123 self.to_screen('Confirming age
')
1125 def report_login(self):
1126 """Report attempt to log in."""
1127 self.to_screen('Logging
in')
1129 def raise_login_required(
1130 self, msg='This video
is only available
for registered users
',
1131 metadata_available=False, method=NO_DEFAULT):
1132 if metadata_available and (
1133 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1134 self.report_warning(msg)
1136 msg += format_field(self._login_hint(method), None, '. %s')
1137 raise ExtractorError(msg, expected=True)
1139 def raise_geo_restricted(
1140 self, msg='This video
is not available
from your location due to geo restriction
',
1141 countries=None, metadata_available=False):
1142 if metadata_available and (
1143 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1144 self.report_warning(msg)
1146 raise GeoRestrictedError(msg, countries=countries)
1148 def raise_no_formats(self, msg, expected=False, video_id=None):
1150 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1151 self.report_warning(msg, video_id)
1152 elif isinstance(msg, ExtractorError):
1155 raise ExtractorError(msg, expected=expected, video_id=video_id)
1157 # Methods for following #608
1159 def url_result(url, ie=None, video_id=None, video_title=None, *, url_transparent=False, **kwargs):
1160 """Returns a URL that points to a page that should be processed"""
1162 kwargs['ie_key
'] = ie if isinstance(ie, str) else ie.ie_key()
1163 if video_id is not None:
1164 kwargs['id'] = video_id
1165 if video_title is not None:
1166 kwargs['title
'] = video_title
1169 '_type
': 'url_transparent
' if url_transparent else 'url
',
1174 def playlist_from_matches(cls, matches, playlist_id=None, playlist_title=None,
1175 getter=IDENTITY, ie=None, video_kwargs=None, **kwargs):
1176 return cls.playlist_result(
1177 (cls.url_result(m, ie, **(video_kwargs or {})) for m in orderedSet(map(getter, matches), lazy=True)),
1178 playlist_id, playlist_title, **kwargs)
1181 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None, *, multi_video=False, **kwargs):
1182 """Returns a playlist"""
1184 kwargs['id'] = playlist_id
1186 kwargs['title
'] = playlist_title
1187 if playlist_description is not None:
1188 kwargs['description
'] = playlist_description
1191 '_type
': 'multi_video
' if multi_video else 'playlist
',
1195 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1197 Perform a regex search on the given string, using a single or a list of
1198 patterns returning the first matching group.
1199 In case of failure return a default value or raise a WARNING or a
1200 RegexNotFoundError, depending on fatal, specifying the field name.
1204 elif isinstance(pattern, (str, re.Pattern)):
1205 mobj = re.search(pattern, string, flags)
1208 mobj = re.search(p, string, flags)
1212 _name = self._downloader._format_err(name, self._downloader.Styles.EMPHASIS)
1216 # return the first matching group
1217 return next(g for g in mobj.groups() if g is not None)
1218 elif isinstance(group, (list, tuple)):
1219 return tuple(mobj.group(g) for g in group)
1221 return mobj.group(group)
1222 elif default is not NO_DEFAULT:
1225 raise RegexNotFoundError('Unable to extract
%s' % _name)
1227 self.report_warning('unable to extract
%s' % _name + bug_reports_message())
1230 def _search_json(self, start_pattern, string, name, video_id, *, end_pattern='',
1231 contains_pattern=r'{(?s:.+)}
', fatal=True, default=NO_DEFAULT, **kwargs):
1232 """Searches string for the JSON object specified by start_pattern"""
1233 # NB: end_pattern is only used to reduce the size of the initial match
1234 if default is NO_DEFAULT:
1235 default, has_default = {}, False
1237 fatal, has_default = False, True
1239 json_string = self._search_regex(
1240 rf'(?
:{start_pattern}
)\s
*(?P
<json
>{contains_pattern}
)\s
*(?
:{end_pattern}
)',
1241 string, name, group='json
', fatal=fatal, default=None if has_default else NO_DEFAULT)
1245 _name = self._downloader._format_err(name, self._downloader.Styles.EMPHASIS)
1247 return self._parse_json(json_string, video_id, ignore_extra=True, **kwargs)
1248 except ExtractorError as e:
1250 raise ExtractorError(
1251 f'Unable to extract {_name}
- Failed to parse JSON
', cause=e.cause, video_id=video_id)
1252 elif not has_default:
1253 self.report_warning(
1254 f'Unable to extract {_name}
- Failed to parse JSON
: {e}
', video_id=video_id)
1257 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1259 Like _search_regex, but strips HTML tags and unescapes entities.
1261 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1263 return clean_html(res).strip()
1267 def _get_netrc_login_info(self, netrc_machine=None):
1270 netrc_machine = netrc_machine or self._NETRC_MACHINE
1272 if self.get_param('usenetrc
', False):
1274 netrc_file = compat_expanduser(self.get_param('netrc_location
') or '~
')
1275 if os.path.isdir(netrc_file):
1276 netrc_file = os.path.join(netrc_file, '.netrc
')
1277 info = netrc.netrc(file=netrc_file).authenticators(netrc_machine)
1278 if info is not None:
1282 raise netrc.NetrcParseError(
1283 'No authenticators
for %s' % netrc_machine)
1284 except (OSError, netrc.NetrcParseError) as err:
1285 self.report_warning(
1286 'parsing
.netrc
: %s' % error_to_compat_str(err))
1288 return username, password
1290 def _get_login_info(self, username_option='username
', password_option='password
', netrc_machine=None):
1292 Get the login info as (username, password)
1293 First look for the manually specified credentials using username_option
1294 and password_option as keys in params dictionary. If no such credentials
1295 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1297 If there's no info available
, return (None, None)
1300 # Attempt to use provided username and password or .netrc data
1301 username = self.get_param(username_option)
1302 if username is not None:
1303 password = self.get_param(password_option)
1305 username, password = self._get_netrc_login_info(netrc_machine)
1307 return username, password
1309 def _get_tfa_info(self, note='two-factor verification code'):
1311 Get the two
-factor authentication info
1312 TODO
- asking the user will be required
for sms
/phone verify
1313 currently just uses the command line option
1314 If there
's no info available, return None
1317 tfa = self.get_param('twofactor
')
1321 return getpass.getpass('Type
%s and press
[Return
]: ' % note)
1323 # Helper functions for extracting OpenGraph info
1325 def _og_regexes(prop):
1326 content_re = r'content
=(?
:"([^"]+?
)"|\'([^\']+?)\'|\s*([^\s"\'=<>`
]+?
))'
1327 property_re = (r'(?
:name|
property)=(?
:\'og
%(sep)s%(prop)s\'|
"og%(sep)s%(prop)s"|\s
*og
%(sep)s%(prop)s\b)'
1328 % {'prop': re.escape(prop), 'sep': '(?::|[:-])'})
1329 template = r'<meta
[^
>]+?
%s[^
>]+?
%s'
1331 template % (property_re, content_re),
1332 template % (content_re, property_re),
1336 def _meta_regex(prop):
1337 return r'''(?isx)<meta
1338 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1339 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1341 def _og_search_property(self, prop, html, name=None, **kargs):
1342 prop = variadic(prop)
1344 name = 'OpenGraph
%s' % prop[0]
1347 og_regexes.extend(self._og_regexes(p))
1348 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1351 return unescapeHTML(escaped)
1353 def _og_search_thumbnail(self, html, **kargs):
1354 return self._og_search_property('image
', html, 'thumbnail URL
', fatal=False, **kargs)
1356 def _og_search_description(self, html, **kargs):
1357 return self._og_search_property('description
', html, fatal=False, **kargs)
1359 def _og_search_title(self, html, *, fatal=False, **kargs):
1360 return self._og_search_property('title
', html, fatal=fatal, **kargs)
1362 def _og_search_video_url(self, html, name='video url
', secure=True, **kargs):
1363 regexes = self._og_regexes('video
') + self._og_regexes('video
:url
')
1365 regexes = self._og_regexes('video
:secure_url
') + regexes
1366 return self._html_search_regex(regexes, html, name, **kargs)
1368 def _og_search_url(self, html, **kargs):
1369 return self._og_search_property('url
', html, **kargs)
1371 def _html_extract_title(self, html, name='title
', *, fatal=False, **kwargs):
1372 return self._html_search_regex(r'(?s
)<title
\b[^
>]*>([^
<]+)</title
>', html, name, fatal=fatal, **kwargs)
1374 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1375 name = variadic(name)
1376 if display_name is None:
1377 display_name = name[0]
1378 return self._html_search_regex(
1379 [self._meta_regex(n) for n in name],
1380 html, display_name, fatal=fatal, group='content
', **kwargs)
1382 def _dc_search_uploader(self, html):
1383 return self._html_search_meta('dc
.creator
', html, 'uploader
')
1386 def _rta_search(html):
1387 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1388 if re.search(r'(?ix
)<meta\s
+name
="rating"\s
+'
1389 r' content
="RTA-5042-1996-1400-1577-RTA"',
1393 # And then there are the jokers who advertise that they use RTA, but actually don't
.
1394 AGE_LIMIT_MARKERS
= [
1395 r
'Proudly Labeled <a href="http://www\.rtalabel\.org/" title="Restricted to Adults">RTA</a>',
1397 if any(re
.search(marker
, html
) for marker
in AGE_LIMIT_MARKERS
):
1401 def _media_rating_search(self
, html
):
1402 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1403 rating
= self
._html
_search
_meta
('rating', html
)
1415 return RATING_TABLE
.get(rating
.lower())
1417 def _family_friendly_search(self
, html
):
1418 # See http://schema.org/VideoObject
1419 family_friendly
= self
._html
_search
_meta
(
1420 'isFamilyFriendly', html
, default
=None)
1422 if not family_friendly
:
1431 return RATING_TABLE
.get(family_friendly
.lower())
1433 def _twitter_search_player(self
, html
):
1434 return self
._html
_search
_meta
('twitter:player', html
,
1435 'twitter card player')
1437 def _yield_json_ld(self
, html
, video_id
, *, fatal
=True, default
=NO_DEFAULT
):
1438 """Yield all json ld objects in the html"""
1439 if default
is not NO_DEFAULT
:
1441 for mobj
in re
.finditer(JSON_LD_RE
, html
):
1442 json_ld_item
= self
._parse
_json
(mobj
.group('json_ld'), video_id
, fatal
=fatal
)
1443 for json_ld
in variadic(json_ld_item
):
1444 if isinstance(json_ld
, dict):
1447 def _search_json_ld(self
, html
, video_id
, expected_type
=None, *, fatal
=True, default
=NO_DEFAULT
):
1448 """Search for a video in any json ld in the html"""
1449 if default
is not NO_DEFAULT
:
1451 info
= self
._json
_ld
(
1452 list(self
._yield
_json
_ld
(html
, video_id
, fatal
=fatal
, default
=default
)),
1453 video_id
, fatal
=fatal
, expected_type
=expected_type
)
1456 if default
is not NO_DEFAULT
:
1459 raise RegexNotFoundError('Unable to extract JSON-LD')
1461 self
.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1464 def _json_ld(self
, json_ld
, video_id
, fatal
=True, expected_type
=None):
1465 if isinstance(json_ld
, str):
1466 json_ld
= self
._parse
_json
(json_ld
, video_id
, fatal
=fatal
)
1470 if not isinstance(json_ld
, (list, tuple, dict)):
1472 if isinstance(json_ld
, dict):
1475 INTERACTION_TYPE_MAP
= {
1476 'CommentAction': 'comment',
1477 'AgreeAction': 'like',
1478 'DisagreeAction': 'dislike',
1479 'LikeAction': 'like',
1480 'DislikeAction': 'dislike',
1481 'ListenAction': 'view',
1482 'WatchAction': 'view',
1483 'ViewAction': 'view',
1486 def is_type(e
, *expected_types
):
1487 type = variadic(traverse_obj(e
, '@type'))
1488 return any(x
in type for x
in expected_types
)
1490 def extract_interaction_type(e
):
1491 interaction_type
= e
.get('interactionType')
1492 if isinstance(interaction_type
, dict):
1493 interaction_type
= interaction_type
.get('@type')
1494 return str_or_none(interaction_type
)
1496 def extract_interaction_statistic(e
):
1497 interaction_statistic
= e
.get('interactionStatistic')
1498 if isinstance(interaction_statistic
, dict):
1499 interaction_statistic
= [interaction_statistic
]
1500 if not isinstance(interaction_statistic
, list):
1502 for is_e
in interaction_statistic
:
1503 if not is_type(is_e
, 'InteractionCounter'):
1505 interaction_type
= extract_interaction_type(is_e
)
1506 if not interaction_type
:
1508 # For interaction count some sites provide string instead of
1509 # an integer (as per spec) with non digit characters (e.g. ",")
1510 # so extracting count with more relaxed str_to_int
1511 interaction_count
= str_to_int(is_e
.get('userInteractionCount'))
1512 if interaction_count
is None:
1514 count_kind
= INTERACTION_TYPE_MAP
.get(interaction_type
.split('/')[-1])
1517 count_key
= '%s_count' % count_kind
1518 if info
.get(count_key
) is not None:
1520 info
[count_key
] = interaction_count
1522 def extract_chapter_information(e
):
1524 'title': part
.get('name'),
1525 'start_time': part
.get('startOffset'),
1526 'end_time': part
.get('endOffset'),
1527 } for part
in variadic(e
.get('hasPart') or []) if part
.get('@type') == 'Clip']
1528 for idx
, (last_c
, current_c
, next_c
) in enumerate(zip(
1529 [{'end_time': 0}
] + chapters
, chapters
, chapters
[1:])):
1530 current_c
['end_time'] = current_c
['end_time'] or next_c
['start_time']
1531 current_c
['start_time'] = current_c
['start_time'] or last_c
['end_time']
1532 if None in current_c
.values():
1533 self
.report_warning(f
'Chapter {idx} contains broken data. Not extracting chapters')
1536 chapters
[-1]['end_time'] = chapters
[-1]['end_time'] or info
['duration']
1537 info
['chapters'] = chapters
1539 def extract_video_object(e
):
1540 author
= e
.get('author')
1542 'url': url_or_none(e
.get('contentUrl')),
1543 'ext': mimetype2ext(e
.get('encodingFormat')),
1544 'title': unescapeHTML(e
.get('name')),
1545 'description': unescapeHTML(e
.get('description')),
1546 'thumbnails': [{'url': unescapeHTML(url)}
1547 for url
in variadic(traverse_obj(e
, 'thumbnailUrl', 'thumbnailURL'))
1548 if url_or_none(url
)],
1549 'duration': parse_duration(e
.get('duration')),
1550 'timestamp': unified_timestamp(e
.get('uploadDate')),
1551 # author can be an instance of 'Organization' or 'Person' types.
1552 # both types can have 'name' property(inherited from 'Thing' type). [1]
1553 # however some websites are using 'Text' type instead.
1554 # 1. https://schema.org/VideoObject
1555 'uploader': author
.get('name') if isinstance(author
, dict) else author
if isinstance(author
, str) else None,
1556 'artist': traverse_obj(e
, ('byArtist', 'name'), expected_type
=str),
1557 'filesize': int_or_none(float_or_none(e
.get('contentSize'))),
1558 'tbr': int_or_none(e
.get('bitrate')),
1559 'width': int_or_none(e
.get('width')),
1560 'height': int_or_none(e
.get('height')),
1561 'view_count': int_or_none(e
.get('interactionCount')),
1562 'tags': try_call(lambda: e
.get('keywords').split(',')),
1564 if is_type(e
, 'AudioObject'):
1567 'abr': int_or_none(e
.get('bitrate')),
1569 extract_interaction_statistic(e
)
1570 extract_chapter_information(e
)
1572 def traverse_json_ld(json_ld
, at_top_level
=True):
1574 if at_top_level
and '@context' not in e
:
1576 if at_top_level
and set(e
.keys()) == {'@context', '@graph'}
:
1577 traverse_json_ld(variadic(e
['@graph'], allowed_types
=(dict,)), at_top_level
=False)
1579 if expected_type
is not None and not is_type(e
, expected_type
):
1581 rating
= traverse_obj(e
, ('aggregateRating', 'ratingValue'), expected_type
=float_or_none
)
1582 if rating
is not None:
1583 info
['average_rating'] = rating
1584 if is_type(e
, 'TVEpisode', 'Episode'):
1585 episode_name
= unescapeHTML(e
.get('name'))
1587 'episode': episode_name
,
1588 'episode_number': int_or_none(e
.get('episodeNumber')),
1589 'description': unescapeHTML(e
.get('description')),
1591 if not info
.get('title') and episode_name
:
1592 info
['title'] = episode_name
1593 part_of_season
= e
.get('partOfSeason')
1594 if is_type(part_of_season
, 'TVSeason', 'Season', 'CreativeWorkSeason'):
1596 'season': unescapeHTML(part_of_season
.get('name')),
1597 'season_number': int_or_none(part_of_season
.get('seasonNumber')),
1599 part_of_series
= e
.get('partOfSeries') or e
.get('partOfTVSeries')
1600 if is_type(part_of_series
, 'TVSeries', 'Series', 'CreativeWorkSeries'):
1601 info
['series'] = unescapeHTML(part_of_series
.get('name'))
1602 elif is_type(e
, 'Movie'):
1604 'title': unescapeHTML(e
.get('name')),
1605 'description': unescapeHTML(e
.get('description')),
1606 'duration': parse_duration(e
.get('duration')),
1607 'timestamp': unified_timestamp(e
.get('dateCreated')),
1609 elif is_type(e
, 'Article', 'NewsArticle'):
1611 'timestamp': parse_iso8601(e
.get('datePublished')),
1612 'title': unescapeHTML(e
.get('headline')),
1613 'description': unescapeHTML(e
.get('articleBody') or e
.get('description')),
1615 if is_type(traverse_obj(e
, ('video', 0)), 'VideoObject'):
1616 extract_video_object(e
['video'][0])
1617 elif is_type(traverse_obj(e
, ('subjectOf', 0)), 'VideoObject'):
1618 extract_video_object(e
['subjectOf'][0])
1619 elif is_type(e
, 'VideoObject', 'AudioObject'):
1620 extract_video_object(e
)
1621 if expected_type
is None:
1625 video
= e
.get('video')
1626 if is_type(video
, 'VideoObject'):
1627 extract_video_object(video
)
1628 if expected_type
is None:
1632 traverse_json_ld(json_ld
)
1634 return filter_dict(info
)
1636 def _search_nextjs_data(self
, webpage
, video_id
, *, transform_source
=None, fatal
=True, **kw
):
1637 return self
._parse
_json
(
1639 r
'(?s)<script[^>]+id=[\'"]__NEXT_DATA__[\'"][^
>]*>([^
<]+)</script
>',
1640 webpage, 'next
.js data
', fatal=fatal, **kw),
1641 video_id, transform_source=transform_source, fatal=fatal)
1643 def _search_nuxt_data(self, webpage, video_id, context_name='__NUXT__
', *, fatal=True, traverse=('data
', 0)):
1644 """Parses Nuxt.js metadata. This works as long as the function __NUXT__ invokes is a pure function"""
1645 rectx = re.escape(context_name)
1646 FUNCTION_RE = r'\
(function\
((?P
<arg_keys
>.*?
)\
){return\s+(?P<js>{.*?}
)\s
*;?\s
*}\
((?P
<arg_vals
>.*?
)\
)'
1647 js, arg_keys, arg_vals = self._search_regex(
1648 (rf'<script
>\s
*window\
.{rectx}
={FUNCTION_RE}\s
*\
)\s
*;?\s
*</script
>', rf'{rectx}\
(.*?{FUNCTION_RE}
'),
1649 webpage, context_name, group=('js
', 'arg_keys
', 'arg_vals
'), fatal=fatal)
1651 args = dict(zip(arg_keys.split(','), arg_vals.split(',')))
1653 for key, val in args.items():
1654 if val in ('undefined
', 'void
0'):
1657 ret = self._parse_json(js, video_id, transform_source=functools.partial(js_to_json, vars=args), fatal=fatal)
1658 return traverse_obj(ret, traverse) or {}
1661 def _hidden_inputs(html):
1662 html = re.sub(r'<!--(?
:(?
!<!--).)*-->', '', html)
1664 for input in re.findall(r'(?i
)(<input[^
>]+>)', html):
1665 attrs = extract_attributes(input)
1668 if attrs.get('type') not in ('hidden
', 'submit
'):
1670 name = attrs.get('name
') or attrs.get('id')
1671 value = attrs.get('value
')
1672 if name and value is not None:
1673 hidden_inputs[name] = value
1674 return hidden_inputs
1676 def _form_hidden_inputs(self, form_id, html):
1677 form = self._search_regex(
1678 r'(?
is)<form
[^
>]+?
id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1679 html, '%s form' % form_id, group='form')
1680 return self._hidden_inputs(form)
1683 regex = r' *((?P<reverse>\+)?(?P<field>[a-zA-Z0-9_]+)((?P<separator>[~:])(?P<limit>.*?))?)? *$'
1685 default = ('hidden', 'aud_or_vid', 'hasvid', 'ie_pref', 'lang', 'quality',
1686 'res', 'fps', 'hdr:12', 'vcodec:vp9.2', 'channels', 'acodec',
1687 'size', 'br', 'asr', 'proto', 'ext', 'hasaud', 'source', 'id') # These must not be aliases
1688 ytdl_default = ('hasaud', 'lang', 'quality', 'tbr', 'filesize', 'vbr',
1689 'height', 'width', 'proto', 'vext', 'abr', 'aext',
1690 'fps', 'fs_approx', 'source', 'id')
1693 'vcodec': {'type': 'ordered', 'regex': True,
1694 'order': ['av0?1', 'vp0?9.2', 'vp0?9', '[hx]265|he?vc?', '[hx]264|avc', 'vp0?8', 'mp4v|h263', 'theora', '', None, 'none']},
1695 'acodec': {'type': 'ordered', 'regex': True,
1696 'order': ['[af]lac', 'wav|aiff', 'opus', 'vorbis|ogg', 'aac', 'mp?4a?', 'mp3', 'e-?a?c-?3', 'ac-?3', 'dts', '', None, 'none']},
1697 'hdr': {'type': 'ordered', 'regex': True, 'field': 'dynamic_range',
1698 'order': ['dv', '(hdr)?12', r'(hdr)?10\+', '(hdr)?10', 'hlg', '', 'sdr', None]},
1699 'proto': {'type': 'ordered', 'regex': True, 'field': 'protocol',
1700 'order': ['(ht|f)tps', '(ht|f)tp$', 'm3u8.*', '.*dash', 'websocket_frag', 'rtmpe?', '', 'mms|rtsp', 'ws|websocket', 'f4']},
1701 'vext': {'type': 'ordered', 'field': 'video_ext',
1702 'order': ('mp4', 'webm', 'flv', '', 'none'),
1703 'order_free': ('webm', 'mp4', 'flv', '', 'none')},
1704 'aext': {'type': 'ordered', 'field': 'audio_ext',
1705 'order': ('m4a', 'aac', 'mp3', 'ogg', 'opus', 'webm', '', 'none'),
1706 'order_free': ('ogg', 'opus', 'webm', 'mp3', 'm4a', 'aac', '', 'none')},
1707 'hidden': {'visible': False, 'forced': True, 'type': 'extractor', 'max': -1000},
1708 'aud_or_vid': {'visible': False, 'forced': True, 'type': 'multiple',
1709 'field': ('vcodec', 'acodec'),
1710 'function': lambda it: int(any(v != 'none' for v in it))},
1711 'ie_pref': {'priority': True, 'type': 'extractor'},
1712 'hasvid': {'priority': True, 'field': 'vcodec', 'type': 'boolean', 'not_in_list': ('none',)},
1713 'hasaud': {'field': 'acodec', 'type': 'boolean', 'not_in_list': ('none',)},
1714 'lang': {'convert': 'float', 'field': 'language_preference', 'default': -1},
1715 'quality': {'convert': 'float', 'default': -1},
1716 'filesize': {'convert': 'bytes'},
1717 'fs_approx': {'convert': 'bytes', 'field': 'filesize_approx'},
1718 'id': {'convert': 'string', 'field': 'format_id'},
1719 'height': {'convert': 'float_none'},
1720 'width': {'convert': 'float_none'},
1721 'fps': {'convert': 'float_none'},
1722 'channels': {'convert': 'float_none', 'field': 'audio_channels'},
1723 'tbr': {'convert': 'float_none'},
1724 'vbr': {'convert': 'float_none'},
1725 'abr': {'convert': 'float_none'},
1726 'asr': {'convert': 'float_none'},
1727 'source': {'convert': 'float', 'field': 'source_preference', 'default': -1},
1729 'codec': {'type': 'combined', 'field': ('vcodec', 'acodec')},
1730 'br': {'type': 'combined', 'field': ('tbr', 'vbr', 'abr'), 'same_limit': True},
1731 'size': {'type': 'combined', 'same_limit': True, 'field': ('filesize', 'fs_approx')},
1732 'ext': {'type': 'combined', 'field': ('vext', 'aext')},
1733 'res': {'type': 'multiple', 'field': ('height', 'width'),
1734 'function': lambda it: (lambda l: min(l) if l else 0)(tuple(filter(None, it)))},
1736 # Actual field names
1737 'format_id': {'type': 'alias', 'field': 'id'},
1738 'preference': {'type': 'alias', 'field': 'ie_pref'},
1739 'language_preference': {'type': 'alias', 'field': 'lang'},
1740 'source_preference': {'type': 'alias', 'field': 'source'},
1741 'protocol': {'type': 'alias', 'field': 'proto'},
1742 'filesize_approx': {'type': 'alias', 'field': 'fs_approx'},
1743 'audio_channels': {'type': 'alias', 'field': 'channels'},
1746 'dimension': {'type': 'alias', 'field': 'res', 'deprecated': True},
1747 'resolution': {'type': 'alias', 'field': 'res', 'deprecated': True},
1748 'extension': {'type': 'alias', 'field': 'ext', 'deprecated': True},
1749 'bitrate': {'type': 'alias', 'field': 'br', 'deprecated': True},
1750 'total_bitrate': {'type': 'alias', 'field': 'tbr', 'deprecated': True},
1751 'video_bitrate': {'type': 'alias', 'field': 'vbr', 'deprecated': True},
1752 'audio_bitrate': {'type': 'alias', 'field': 'abr', 'deprecated': True},
1753 'framerate': {'type': 'alias', 'field': 'fps', 'deprecated': True},
1754 'filesize_estimate': {'type': 'alias', 'field': 'size', 'deprecated': True},
1755 'samplerate': {'type': 'alias', 'field': 'asr', 'deprecated': True},
1756 'video_ext': {'type': 'alias', 'field': 'vext', 'deprecated': True},
1757 'audio_ext': {'type': 'alias', 'field': 'aext', 'deprecated': True},
1758 'video_codec': {'type': 'alias', 'field': 'vcodec', 'deprecated': True},
1759 'audio_codec': {'type': 'alias', 'field': 'acodec', 'deprecated': True},
1760 'video': {'type': 'alias', 'field': 'hasvid', 'deprecated': True},
1761 'has_video': {'type': 'alias', 'field': 'hasvid', 'deprecated': True},
1762 'audio': {'type': 'alias', 'field': 'hasaud', 'deprecated': True},
1763 'has_audio': {'type': 'alias', 'field': 'hasaud', 'deprecated': True},
1764 'extractor': {'type': 'alias', 'field': 'ie_pref', 'deprecated': True},
1765 'extractor_preference': {'type': 'alias', 'field': 'ie_pref', 'deprecated': True},
1768 def __init__(self, ie, field_preference):
1770 self.ydl = ie._downloader
1771 self.evaluate_params(self.ydl.params, field_preference)
1772 if ie.get_param('verbose'):
1773 self.print_verbose_info(self.ydl.write_debug)
1775 def _get_field_setting(self, field, key):
1776 if field not in self.settings:
1777 if key in ('forced', 'priority'):
1779 self.ydl.deprecated_feature(f'Using arbitrary fields ({field}) for format sorting is '
1780 'deprecated and may be removed in a future version')
1781 self.settings[field] = {}
1782 propObj = self.settings[field]
1783 if key not in propObj:
1784 type = propObj.get('type')
1786 default = 'preference' if type == 'extractor' else (field,) if type in ('combined', 'multiple') else field
1787 elif key == 'convert':
1788 default = 'order' if type == 'ordered' else 'float_string' if field else 'ignore'
1790 default = {'type': 'field', 'visible': True, 'order': [], 'not_in_list': (None,)}.get(key, None)
1791 propObj[key] = default
1794 def _resolve_field_value(self, field, value, convertNone=False):
1799 value = value.lower()
1800 conversion = self._get_field_setting(field, 'convert')
1801 if conversion == 'ignore':
1803 if conversion == 'string':
1805 elif conversion == 'float_none':
1806 return float_or_none(value)
1807 elif conversion == 'bytes':
1808 return FileDownloader.parse_bytes(value)
1809 elif conversion == 'order':
1810 order_list = (self._use_free_order and self._get_field_setting(field, 'order_free')) or self._get_field_setting(field, 'order')
1811 use_regex = self._get_field_setting(field, 'regex')
1812 list_length = len(order_list)
1813 empty_pos = order_list.index('') if '' in order_list else list_length + 1
1814 if use_regex and value is not None:
1815 for i, regex in enumerate(order_list):
1816 if regex and re.match(regex, value):
1817 return list_length - i
1818 return list_length - empty_pos # not in list
1819 else: # not regex or value = None
1820 return list_length - (order_list.index(value) if value in order_list else empty_pos)
1822 if value.isnumeric():
1825 self.settings[field]['convert'] = 'string'
1828 def evaluate_params(self, params, sort_extractor):
1829 self._use_free_order = params.get('prefer_free_formats', False)
1830 self._sort_user = params.get('format_sort', [])
1831 self._sort_extractor = sort_extractor
1833 def add_item(field, reverse, closest, limit_text):
1834 field = field.lower()
1835 if field in self._order:
1837 self._order.append(field)
1838 limit = self._resolve_field_value(field, limit_text)
1841 'closest': False if limit is None else closest,
1842 'limit_text': limit_text,
1844 if field in self.settings:
1845 self.settings[field].update(data)
1847 self.settings[field] = data
1850 tuple(field for field in self.default if self._get_field_setting(field, 'forced'))
1851 + (tuple() if params.get('format_sort_force', False)
1852 else tuple(field for field in self.default if self._get_field_setting(field, 'priority')))
1853 + tuple(self._sort_user) + tuple(sort_extractor) + self.default)
1855 for item in sort_list:
1856 match = re.match(self.regex, item)
1858 raise ExtractorError('Invalid format sort string "%s" given by extractor' % item)
1859 field = match.group('field')
1862 if self._get_field_setting(field, 'type') == 'alias':
1863 alias, field = field, self._get_field_setting(field, 'field')
1864 if self._get_field_setting(alias, 'deprecated'):
1865 self.ydl.deprecated_feature(f'Format sorting alias {alias} is deprecated and may '
1866 f'be removed in a future version. Please use {field} instead')
1867 reverse = match.group('reverse') is not None
1868 closest = match.group('separator') == '~'
1869 limit_text = match.group('limit')
1871 has_limit = limit_text is not None
1872 has_multiple_fields = self._get_field_setting(field, 'type') == 'combined'
1873 has_multiple_limits = has_limit and has_multiple_fields and not self._get_field_setting(field, 'same_limit')
1875 fields = self._get_field_setting(field, 'field') if has_multiple_fields else (field,)
1876 limits = limit_text.split(':') if has_multiple_limits else (limit_text,) if has_limit else tuple()
1877 limit_count = len(limits)
1878 for (i, f) in enumerate(fields):
1879 add_item(f, reverse, closest,
1880 limits[i] if i < limit_count
1881 else limits[0] if has_limit and not has_multiple_limits
1884 def print_verbose_info(self, write_debug):
1886 write_debug('Sort order given by user: %s' % ', '.join(self._sort_user))
1887 if self._sort_extractor:
1888 write_debug('Sort order given by extractor: %s' % ', '.join(self._sort_extractor))
1889 write_debug('Formats sorted by: %s' % ', '.join(['%s%s%s' % (
1890 '+' if self._get_field_setting(field, 'reverse') else '', field,
1891 '%s%s(%s)' % ('~' if self._get_field_setting(field, 'closest') else ':',
1892 self._get_field_setting(field, 'limit_text'),
1893 self._get_field_setting(field, 'limit'))
1894 if self._get_field_setting(field, 'limit_text') is not None else '')
1895 for field in self._order if self._get_field_setting(field, 'visible')]))
1897 def _calculate_field_preference_from_value(self, format, field, type, value):
1898 reverse = self._get_field_setting(field, 'reverse')
1899 closest = self._get_field_setting(field, 'closest')
1900 limit = self._get_field_setting(field, 'limit')
1902 if type == 'extractor':
1903 maximum = self._get_field_setting(field, 'max')
1904 if value is None or (maximum is not None and value >= maximum):
1906 elif type == 'boolean':
1907 in_list = self._get_field_setting(field, 'in_list')
1908 not_in_list = self._get_field_setting(field, 'not_in_list')
1909 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
1910 elif type == 'ordered':
1911 value = self._resolve_field_value(field, value, True)
1913 # try to convert to number
1914 val_num = float_or_none(value, default=self._get_field_setting(field, 'default'))
1915 is_num = self._get_field_setting(field, 'convert') != 'string' and val_num is not None
1919 return ((-10, 0) if value is None
1920 else (1, value, 0) if not is_num # if a field has mixed strings and numbers, strings are sorted higher
1921 else (0, -abs(value - limit), value - limit if reverse else limit - value) if closest
1922 else (0, value, 0) if not reverse and (limit is None or value <= limit)
1923 else (0, -value, 0) if limit is None or (reverse and value == limit) or value > limit
1924 else (-1, value, 0))
1926 def _calculate_field_preference(self, format, field):
1927 type = self._get_field_setting(field, 'type') # extractor, boolean, ordered, field, multiple
1928 get_value = lambda f: format.get(self._get_field_setting(f, 'field'))
1929 if type == 'multiple':
1930 type = 'field' # Only 'field' is allowed in multiple for now
1931 actual_fields = self._get_field_setting(field, 'field')
1933 value = self._get_field_setting(field, 'function')(get_value(f) for f in actual_fields)
1935 value = get_value(field)
1936 return self._calculate_field_preference_from_value(format, field, type, value)
1938 def calculate_preference(self, format):
1939 # Determine missing protocol
1940 if not format.get('protocol'):
1941 format['protocol'] = determine_protocol(format)
1943 # Determine missing ext
1944 if not format.get('ext') and 'url' in format:
1945 format['ext'] = determine_ext(format['url'])
1946 if format.get('vcodec') == 'none':
1947 format['audio_ext'] = format['ext'] if format.get('acodec') != 'none' else 'none'
1948 format['video_ext'] = 'none'
1950 format['video_ext'] = format['ext']
1951 format['audio_ext'] = 'none'
1952 # if format.get('preference') is None and format.get('ext') in ('f4f', 'f4m'): # Not supported?
1953 # format['preference'] = -1000
1955 # Determine missing bitrates
1956 if format.get('tbr') is None:
1957 if format.get('vbr') is not None and format.get('abr') is not None:
1958 format['tbr'] = format.get('vbr', 0) + format.get('abr', 0)
1960 if format.get('vcodec') != 'none' and format.get('vbr') is None:
1961 format['vbr'] = format.get('tbr') - format.get('abr', 0)
1962 if format.get('acodec') != 'none' and format.get('abr') is None:
1963 format['abr'] = format.get('tbr') - format.get('vbr', 0)
1965 return tuple(self._calculate_field_preference(format, field) for field in self._order)
1967 def _sort_formats(self, formats, field_preference=[]):
1970 formats.sort(key=self.FormatSort(self, field_preference).calculate_preference)
1972 def _check_formats(self, formats, video_id):
1974 formats[:] = filter(
1975 lambda f: self._is_valid_url(
1977 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1981 def _remove_duplicate_formats(formats):
1985 if f['url'] not in format_urls:
1986 format_urls.add(f['url'])
1987 unique_formats.append(f)
1988 formats[:] = unique_formats
1990 def _is_valid_url(self, url, video_id, item='video', headers={}):
1991 url = self._proto_relative_url(url, scheme='http:')
1992 # For now assume non HTTP(S) URLs always valid
1993 if not (url.startswith('http://') or url.startswith('https://')):
1996 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1998 except ExtractorError as e:
2000 '%s: %s URL is invalid, skipping: %s'
2001 % (video_id, item, error_to_compat_str(e.cause)))
2004 def http_scheme(self):
2005 """ Either "http
:" or "https
:", depending on the user's preferences """
2008 if self.get_param('prefer_insecure', False)
2011 def _proto_relative_url(self, url, scheme=None):
2012 scheme = scheme or self.http_scheme()
2013 assert scheme.endswith(':')
2014 return sanitize_url(url, scheme=scheme[:-1])
2016 def _sleep(self, timeout, video_id, msg_template=None):
2017 if msg_template is None:
2018 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
2019 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
2023 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
2024 transform_source=lambda s: fix_xml_ampersands(s).strip(),
2025 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
2026 res = self._download_xml_handle(
2027 manifest_url, video_id, 'Downloading f4m manifest',
2028 'Unable to download f4m manifest',
2029 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
2030 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
2031 transform_source=transform_source,
2032 fatal=fatal, data=data, headers=headers, query=query)
2036 manifest, urlh = res
2037 manifest_url = urlh.geturl()
2039 return self._parse_f4m_formats(
2040 manifest, manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
2041 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
2043 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
2044 transform_source=lambda s: fix_xml_ampersands(s).strip(),
2045 fatal=True, m3u8_id=None):
2046 if not isinstance(manifest, xml.etree.ElementTree.Element) and not fatal:
2049 # currently yt-dlp cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
2050 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
2051 if akamai_pv is not None and ';' in akamai_pv.text:
2052 playerVerificationChallenge = akamai_pv.text.split(';')[0]
2053 if playerVerificationChallenge.strip() != '':
2057 manifest_version = '1.0'
2058 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
2060 manifest_version = '2.0'
2061 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
2062 # Remove unsupported DRM protected media from final formats
2063 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
2064 media_nodes = remove_encrypted_media(media_nodes)
2068 manifest_base_url = get_base_url(manifest)
2070 bootstrap_info = xpath_element(
2071 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
2072 'bootstrap info', default=None)
2075 mime_type = xpath_text(
2076 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
2077 'base URL', default=None)
2078 if mime_type and mime_type.startswith('audio/'):
2081 for i, media_el in enumerate(media_nodes):
2082 tbr = int_or_none(media_el.attrib.get('bitrate'))
2083 width = int_or_none(media_el.attrib.get('width'))
2084 height = int_or_none(media_el.attrib.get('height'))
2085 format_id = join_nonempty(f4m_id, tbr or i)
2086 # If <bootstrapInfo> is present, the specified f4m is a
2087 # stream-level manifest, and only set-level manifests may refer to
2088 # external resources. See section 11.4 and section 4 of F4M spec
2089 if bootstrap_info is None:
2091 # @href is introduced in 2.0, see section 11.6 of F4M spec
2092 if manifest_version == '2.0':
2093 media_url = media_el.attrib.get('href')
2094 if media_url is None:
2095 media_url = media_el.attrib.get('url')
2099 media_url if media_url.startswith('http://') or media_url.startswith('https://')
2100 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
2101 # If media_url is itself a f4m manifest do the recursive extraction
2102 # since bitrates in parent manifest (this one) and media_url manifest
2103 # may differ leading to inability to resolve the format by requested
2104 # bitrate in f4m downloader
2105 ext = determine_ext(manifest_url)
2107 f4m_formats = self._extract_f4m_formats(
2108 manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
2109 transform_source=transform_source, fatal=fatal)
2110 # Sometimes stream-level manifest contains single media entry that
2111 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
2112 # At the same time parent's media entry in set-level manifest may
2113 # contain it. We will copy it from parent in such cases.
2114 if len(f4m_formats) == 1:
2117 'tbr': f.get('tbr') or tbr,
2118 'width': f.get('width') or width,
2119 'height': f.get('height') or height,
2120 'format_id': f.get('format_id') if not tbr else format_id,
2123 formats.extend(f4m_formats)
2126 formats.extend(self._extract_m3u8_formats(
2127 manifest_url, video_id, 'mp4', preference=preference,
2128 quality=quality, m3u8_id=m3u8_id, fatal=fatal))
2131 'format_id': format_id,
2132 'url': manifest_url,
2133 'manifest_url': manifest_url,
2134 'ext': 'flv' if bootstrap_info is not None else None,
2140 'preference': preference,
2145 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, quality=None, m3u8_id=None):
2147 'format_id': join_nonempty(m3u8_id, 'meta'),
2151 'preference': preference - 100 if preference else -100,
2153 'resolution': 'multiple',
2154 'format_note': 'Quality selection URL',
2157 def _report_ignoring_subs(self, name):
2158 self.report_warning(bug_reports_message(
2159 f'Ignoring subtitle tracks found in the {name} manifest; '
2160 'if any subtitle tracks are missing,'
2163 def _extract_m3u8_formats(self, *args, **kwargs):
2164 fmts, subs = self._extract_m3u8_formats_and_subtitles(*args, **kwargs)
2166 self._report_ignoring_subs('HLS')
2169 def _extract_m3u8_formats_and_subtitles(
2170 self, m3u8_url, video_id, ext=None, entry_protocol='m3u8_native',
2171 preference=None, quality=None, m3u8_id=None, note=None,
2172 errnote=None, fatal=True, live=False, data=None, headers={},
2175 res = self._download_webpage_handle(
2177 note='Downloading m3u8 information' if note is None else note,
2178 errnote='Failed to download m3u8 information' if errnote is None else errnote,
2179 fatal=fatal, data=data, headers=headers, query=query)
2184 m3u8_doc, urlh = res
2185 m3u8_url = urlh.geturl()
2187 return self._parse_m3u8_formats_and_subtitles(
2188 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
2189 preference=preference, quality=quality, m3u8_id=m3u8_id,
2190 note=note, errnote=errnote, fatal=fatal, live=live, data=data,
2191 headers=headers, query=query, video_id=video_id)
2193 def _parse_m3u8_formats_and_subtitles(
2194 self, m3u8_doc, m3u8_url=None, ext=None, entry_protocol='m3u8_native',
2195 preference=None, quality=None, m3u8_id=None, live=False, note=None,
2196 errnote=None, fatal=True, data=None, headers={}, query={},
2198 formats, subtitles = [], {}
2200 has_drm = re.search('|'.join([
2201 r'#EXT-X-FAXS-CM:', # Adobe Flash Access
2202 r'#EXT-X-(?:SESSION-)?KEY:.*?URI="skd
://', # Apple FairPlay
2205 def format_url(url):
2206 return url if re.match(r'^https?
://', url) else urllib.parse.urljoin(m3u8_url, url)
2208 if self.get_param('hls_split_discontinuity
', False):
2209 def _extract_m3u8_playlist_indices(manifest_url=None, m3u8_doc=None):
2211 if not manifest_url:
2213 m3u8_doc = self._download_webpage(
2214 manifest_url, video_id, fatal=fatal, data=data, headers=headers,
2215 note=False, errnote='Failed to download m3u8 playlist information
')
2216 if m3u8_doc is False:
2218 return range(1 + sum(line.startswith('#EXT-X-DISCONTINUITY') for line in m3u8_doc.splitlines()))
2221 def _extract_m3u8_playlist_indices(*args
, **kwargs
):
2225 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
2226 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
2227 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
2229 # We should try extracting formats only from master playlists [1, 4.3.4],
2230 # i.e. playlists that describe available qualities. On the other hand
2231 # media playlists [1, 4.3.3] should be returned as is since they contain
2232 # just the media without qualities renditions.
2233 # Fortunately, master playlist can be easily distinguished from media
2234 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
2235 # master playlist tags MUST NOT appear in a media playlist and vice versa.
2236 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
2237 # media playlist and MUST NOT appear in master playlist thus we can
2238 # clearly detect media playlist with this criterion.
2240 if '#EXT-X-TARGETDURATION' in m3u8_doc
: # media playlist, return as is
2242 'format_id': join_nonempty(m3u8_id
, idx
),
2243 'format_index': idx
,
2244 'url': m3u8_url
or encode_data_uri(m3u8_doc
.encode('utf-8'), 'application/x-mpegurl'),
2246 'protocol': entry_protocol
,
2247 'preference': preference
,
2250 } for idx
in _extract_m3u8_playlist_indices(m3u8_doc
=m3u8_doc
)]
2252 return formats
, subtitles
2255 last_stream_inf
= {}
2257 def extract_media(x_media_line
):
2258 media
= parse_m3u8_attributes(x_media_line
)
2259 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
2260 media_type
, group_id
, name
= media
.get('TYPE'), media
.get('GROUP-ID'), media
.get('NAME')
2261 if not (media_type
and group_id
and name
):
2263 groups
.setdefault(group_id
, []).append(media
)
2264 # <https://tools.ietf.org/html/rfc8216#section-4.3.4.1>
2265 if media_type
== 'SUBTITLES':
2266 # According to RFC 8216 §4.3.4.2.1, URI is REQUIRED in the
2267 # EXT-X-MEDIA tag if the media type is SUBTITLES.
2268 # However, lack of URI has been spotted in the wild.
2269 # e.g. NebulaIE; see https://github.com/yt-dlp/yt-dlp/issues/339
2270 if not media
.get('URI'):
2272 url
= format_url(media
['URI'])
2275 'ext': determine_ext(url
),
2277 if sub_info
['ext'] == 'm3u8':
2278 # Per RFC 8216 §3.1, the only possible subtitle format m3u8
2279 # files may contain is WebVTT:
2280 # <https://tools.ietf.org/html/rfc8216#section-3.1>
2281 sub_info
['ext'] = 'vtt'
2282 sub_info
['protocol'] = 'm3u8_native'
2283 lang
= media
.get('LANGUAGE') or 'und'
2284 subtitles
.setdefault(lang
, []).append(sub_info
)
2285 if media_type
not in ('VIDEO', 'AUDIO'):
2287 media_url
= media
.get('URI')
2289 manifest_url
= format_url(media_url
)
2291 'format_id': join_nonempty(m3u8_id
, group_id
, name
, idx
),
2292 'format_note': name
,
2293 'format_index': idx
,
2294 'url': manifest_url
,
2295 'manifest_url': m3u8_url
,
2296 'language': media
.get('LANGUAGE'),
2298 'protocol': entry_protocol
,
2299 'preference': preference
,
2301 'vcodec': 'none' if media_type
== 'AUDIO' else None,
2302 } for idx
in _extract_m3u8_playlist_indices(manifest_url
))
2304 def build_stream_name():
2305 # Despite specification does not mention NAME attribute for
2306 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
2307 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
2308 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
2309 stream_name
= last_stream_inf
.get('NAME')
2312 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
2313 # from corresponding rendition group
2314 stream_group_id
= last_stream_inf
.get('VIDEO')
2315 if not stream_group_id
:
2317 stream_group
= groups
.get(stream_group_id
)
2318 if not stream_group
:
2319 return stream_group_id
2320 rendition
= stream_group
[0]
2321 return rendition
.get('NAME') or stream_group_id
2323 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
2324 # chance to detect video only formats when EXT-X-STREAM-INF tags
2325 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
2326 for line
in m3u8_doc
.splitlines():
2327 if line
.startswith('#EXT-X-MEDIA:'):
2330 for line
in m3u8_doc
.splitlines():
2331 if line
.startswith('#EXT-X-STREAM-INF:'):
2332 last_stream_inf
= parse_m3u8_attributes(line
)
2333 elif line
.startswith('#') or not line
.strip():
2336 tbr
= float_or_none(
2337 last_stream_inf
.get('AVERAGE-BANDWIDTH')
2338 or last_stream_inf
.get('BANDWIDTH'), scale
=1000)
2339 manifest_url
= format_url(line
.strip())
2341 for idx
in _extract_m3u8_playlist_indices(manifest_url
):
2342 format_id
= [m3u8_id
, None, idx
]
2343 # Bandwidth of live streams may differ over time thus making
2344 # format_id unpredictable. So it's better to keep provided
2347 stream_name
= build_stream_name()
2348 format_id
[1] = stream_name
or '%d' % (tbr
or len(formats
))
2350 'format_id': join_nonempty(*format_id
),
2351 'format_index': idx
,
2352 'url': manifest_url
,
2353 'manifest_url': m3u8_url
,
2356 'fps': float_or_none(last_stream_inf
.get('FRAME-RATE')),
2357 'protocol': entry_protocol
,
2358 'preference': preference
,
2361 resolution
= last_stream_inf
.get('RESOLUTION')
2363 mobj
= re
.search(r
'(?P<width>\d+)[xX](?P<height>\d+)', resolution
)
2365 f
['width'] = int(mobj
.group('width'))
2366 f
['height'] = int(mobj
.group('height'))
2367 # Unified Streaming Platform
2369 r
'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f
['url'])
2371 abr
, vbr
= mobj
.groups()
2372 abr
, vbr
= float_or_none(abr
, 1000), float_or_none(vbr
, 1000)
2377 codecs
= parse_codecs(last_stream_inf
.get('CODECS'))
2379 audio_group_id
= last_stream_inf
.get('AUDIO')
2380 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
2381 # references a rendition group MUST have a CODECS attribute.
2382 # However, this is not always respected. E.g. [2]
2383 # contains EXT-X-STREAM-INF tag which references AUDIO
2384 # rendition group but does not have CODECS and despite
2385 # referencing an audio group it represents a complete
2386 # (with audio and video) format. So, for such cases we will
2387 # ignore references to rendition groups and treat them
2388 # as complete formats.
2389 if audio_group_id
and codecs
and f
.get('vcodec') != 'none':
2390 audio_group
= groups
.get(audio_group_id
)
2391 if audio_group
and audio_group
[0].get('URI'):
2392 # TODO: update acodec for audio only formats with
2394 f
['acodec'] = 'none'
2395 if not f
.get('ext'):
2396 f
['ext'] = 'm4a' if f
.get('vcodec') == 'none' else 'mp4'
2400 progressive_uri
= last_stream_inf
.get('PROGRESSIVE-URI')
2403 del http_f
['manifest_url']
2405 'format_id': f
['format_id'].replace('hls-', 'http-'),
2407 'url': progressive_uri
,
2409 formats
.append(http_f
)
2411 last_stream_inf
= {}
2412 return formats
, subtitles
2414 def _extract_m3u8_vod_duration(
2415 self
, m3u8_vod_url
, video_id
, note
=None, errnote
=None, data
=None, headers
={}, query={}
):
2417 m3u8_vod
= self
._download
_webpage
(
2418 m3u8_vod_url
, video_id
,
2419 note
='Downloading m3u8 VOD manifest' if note
is None else note
,
2420 errnote
='Failed to download VOD manifest' if errnote
is None else errnote
,
2421 fatal
=False, data
=data
, headers
=headers
, query
=query
)
2423 return self
._parse
_m
3u8_vod
_duration
(m3u8_vod
or '', video_id
)
2425 def _parse_m3u8_vod_duration(self
, m3u8_vod
, video_id
):
2426 if '#EXT-X-PLAYLIST-TYPE:VOD' not in m3u8_vod
:
2430 float(line
[len('#EXTINF:'):].split(',')[0])
2431 for line
in m3u8_vod
.splitlines() if line
.startswith('#EXTINF:'))) or None
2434 def _xpath_ns(path
, namespace
=None):
2438 for c
in path
.split('/'):
2439 if not c
or c
== '.':
2442 out
.append('{%s}%s' % (namespace
, c
))
2443 return '/'.join(out
)
2445 def _extract_smil_formats_and_subtitles(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None, transform_source
=None):
2446 res
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
, transform_source
=transform_source
)
2452 smil_url
= urlh
.geturl()
2454 namespace
= self
._parse
_smil
_namespace
(smil
)
2456 fmts
= self
._parse
_smil
_formats
(
2457 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2458 subs
= self
._parse
_smil
_subtitles
(
2459 smil
, namespace
=namespace
)
2463 def _extract_smil_formats(self
, *args
, **kwargs
):
2464 fmts
, subs
= self
._extract
_smil
_formats
_and
_subtitles
(*args
, **kwargs
)
2466 self
._report
_ignoring
_subs
('SMIL')
2469 def _extract_smil_info(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None):
2470 res
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
)
2475 smil_url
= urlh
.geturl()
2477 return self
._parse
_smil
(smil
, smil_url
, video_id
, f4m_params
=f4m_params
)
2479 def _download_smil(self
, smil_url
, video_id
, fatal
=True, transform_source
=None):
2480 return self
._download
_xml
_handle
(
2481 smil_url
, video_id
, 'Downloading SMIL file',
2482 'Unable to download SMIL file', fatal
=fatal
, transform_source
=transform_source
)
2484 def _parse_smil(self
, smil
, smil_url
, video_id
, f4m_params
=None):
2485 namespace
= self
._parse
_smil
_namespace
(smil
)
2487 formats
= self
._parse
_smil
_formats
(
2488 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2489 subtitles
= self
._parse
_smil
_subtitles
(smil
, namespace
=namespace
)
2491 video_id
= os
.path
.splitext(url_basename(smil_url
))[0]
2495 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2496 name
= meta
.attrib
.get('name')
2497 content
= meta
.attrib
.get('content')
2498 if not name
or not content
:
2500 if not title
and name
== 'title':
2502 elif not description
and name
in ('description', 'abstract'):
2503 description
= content
2504 elif not upload_date
and name
== 'date':
2505 upload_date
= unified_strdate(content
)
2508 'id': image
.get('type'),
2509 'url': image
.get('src'),
2510 'width': int_or_none(image
.get('width')),
2511 'height': int_or_none(image
.get('height')),
2512 } for image
in smil
.findall(self
._xpath
_ns
('.//image', namespace
)) if image
.get('src')]
2516 'title': title
or video_id
,
2517 'description': description
,
2518 'upload_date': upload_date
,
2519 'thumbnails': thumbnails
,
2521 'subtitles': subtitles
,
2524 def _parse_smil_namespace(self
, smil
):
2525 return self
._search
_regex
(
2526 r
'(?i)^{([^}]+)?}smil$', smil
.tag
, 'namespace', default
=None)
2528 def _parse_smil_formats(self
, smil
, smil_url
, video_id
, namespace
=None, f4m_params
=None, transform_rtmp_url
=None):
2530 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2531 b
= meta
.get('base') or meta
.get('httpBase')
2543 media
= smil
.findall(self
._xpath
_ns
('.//video', namespace
)) + smil
.findall(self
._xpath
_ns
('.//audio', namespace
))
2544 for medium
in media
:
2545 src
= medium
.get('src')
2546 if not src
or src
in srcs
:
2550 bitrate
= float_or_none(medium
.get('system-bitrate') or medium
.get('systemBitrate'), 1000)
2551 filesize
= int_or_none(medium
.get('size') or medium
.get('fileSize'))
2552 width
= int_or_none(medium
.get('width'))
2553 height
= int_or_none(medium
.get('height'))
2554 proto
= medium
.get('proto')
2555 ext
= medium
.get('ext')
2556 src_ext
= determine_ext(src
)
2557 streamer
= medium
.get('streamer') or base
2559 if proto
== 'rtmp' or streamer
.startswith('rtmp'):
2565 'format_id': 'rtmp-%d' % (rtmp_count
if bitrate
is None else bitrate
),
2567 'filesize': filesize
,
2571 if transform_rtmp_url
:
2572 streamer
, src
= transform_rtmp_url(streamer
, src
)
2573 formats
[-1].update({
2579 src_url
= src
if src
.startswith('http') else urllib
.parse
.urljoin(base
, src
)
2580 src_url
= src_url
.strip()
2582 if proto
== 'm3u8' or src_ext
== 'm3u8':
2583 m3u8_formats
= self
._extract
_m
3u8_formats
(
2584 src_url
, video_id
, ext
or 'mp4', m3u8_id
='hls', fatal
=False)
2585 if len(m3u8_formats
) == 1:
2587 m3u8_formats
[0].update({
2588 'format_id': 'hls-%d' % (m3u8_count
if bitrate
is None else bitrate
),
2593 formats
.extend(m3u8_formats
)
2594 elif src_ext
== 'f4m':
2599 'plugin': 'flowplayer-3.2.0.1',
2601 f4m_url
+= '&' if '?' in f4m_url
else '?'
2602 f4m_url
+= urllib
.parse
.urlencode(f4m_params
)
2603 formats
.extend(self
._extract
_f
4m
_formats
(f4m_url
, video_id
, f4m_id
='hds', fatal
=False))
2604 elif src_ext
== 'mpd':
2605 formats
.extend(self
._extract
_mpd
_formats
(
2606 src_url
, video_id
, mpd_id
='dash', fatal
=False))
2607 elif re
.search(r
'\.ism/[Mm]anifest', src_url
):
2608 formats
.extend(self
._extract
_ism
_formats
(
2609 src_url
, video_id
, ism_id
='mss', fatal
=False))
2610 elif src_url
.startswith('http') and self
._is
_valid
_url
(src
, video_id
):
2614 'ext': ext
or src_ext
or 'flv',
2615 'format_id': 'http-%d' % (bitrate
or http_count
),
2617 'filesize': filesize
,
2622 for medium
in smil
.findall(self
._xpath
_ns
('.//imagestream', namespace
)):
2623 src
= medium
.get('src')
2624 if not src
or src
in srcs
:
2630 'format_id': 'imagestream-%d' % (imgs_count
),
2632 'ext': mimetype2ext(medium
.get('type')),
2635 'width': int_or_none(medium
.get('width')),
2636 'height': int_or_none(medium
.get('height')),
2637 'format_note': 'SMIL storyboards',
2642 def _parse_smil_subtitles(self
, smil
, namespace
=None, subtitles_lang
='en'):
2645 for num
, textstream
in enumerate(smil
.findall(self
._xpath
_ns
('.//textstream', namespace
))):
2646 src
= textstream
.get('src')
2647 if not src
or src
in urls
:
2650 ext
= textstream
.get('ext') or mimetype2ext(textstream
.get('type')) or determine_ext(src
)
2651 lang
= textstream
.get('systemLanguage') or textstream
.get('systemLanguageName') or textstream
.get('lang') or subtitles_lang
2652 subtitles
.setdefault(lang
, []).append({
2658 def _extract_xspf_playlist(self
, xspf_url
, playlist_id
, fatal
=True):
2659 res
= self
._download
_xml
_handle
(
2660 xspf_url
, playlist_id
, 'Downloading xpsf playlist',
2661 'Unable to download xspf manifest', fatal
=fatal
)
2666 xspf_url
= urlh
.geturl()
2668 return self
._parse
_xspf
(
2669 xspf
, playlist_id
, xspf_url
=xspf_url
,
2670 xspf_base_url
=base_url(xspf_url
))
2672 def _parse_xspf(self
, xspf_doc
, playlist_id
, xspf_url
=None, xspf_base_url
=None):
2674 'xspf': 'http://xspf.org/ns/0/',
2675 's1': 'http://static.streamone.nl/player/ns/0',
2679 for track
in xspf_doc
.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP
)):
2681 track
, xpath_with_ns('./xspf:title', NS_MAP
), 'title', default
=playlist_id
)
2682 description
= xpath_text(
2683 track
, xpath_with_ns('./xspf:annotation', NS_MAP
), 'description')
2684 thumbnail
= xpath_text(
2685 track
, xpath_with_ns('./xspf:image', NS_MAP
), 'thumbnail')
2686 duration
= float_or_none(
2687 xpath_text(track
, xpath_with_ns('./xspf:duration', NS_MAP
), 'duration'), 1000)
2690 for location
in track
.findall(xpath_with_ns('./xspf:location', NS_MAP
)):
2691 format_url
= urljoin(xspf_base_url
, location
.text
)
2696 'manifest_url': xspf_url
,
2697 'format_id': location
.get(xpath_with_ns('s1:label', NS_MAP
)),
2698 'width': int_or_none(location
.get(xpath_with_ns('s1:width', NS_MAP
))),
2699 'height': int_or_none(location
.get(xpath_with_ns('s1:height', NS_MAP
))),
2701 self
._sort
_formats
(formats
)
2706 'description': description
,
2707 'thumbnail': thumbnail
,
2708 'duration': duration
,
2713 def _extract_mpd_formats(self
, *args
, **kwargs
):
2714 fmts
, subs
= self
._extract
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2716 self
._report
_ignoring
_subs
('DASH')
2719 def _extract_mpd_formats_and_subtitles(
2720 self
, mpd_url
, video_id
, mpd_id
=None, note
=None, errnote
=None,
2721 fatal
=True, data
=None, headers
={}, query={}
):
2722 res
= self
._download
_xml
_handle
(
2724 note
='Downloading MPD manifest' if note
is None else note
,
2725 errnote
='Failed to download MPD manifest' if errnote
is None else errnote
,
2726 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
2733 # We could have been redirected to a new url when we retrieved our mpd file.
2734 mpd_url
= urlh
.geturl()
2735 mpd_base_url
= base_url(mpd_url
)
2737 return self
._parse
_mpd
_formats
_and
_subtitles
(
2738 mpd_doc
, mpd_id
, mpd_base_url
, mpd_url
)
2740 def _parse_mpd_formats(self
, *args
, **kwargs
):
2741 fmts
, subs
= self
._parse
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2743 self
._report
_ignoring
_subs
('DASH')
2746 def _parse_mpd_formats_and_subtitles(
2747 self
, mpd_doc
, mpd_id
=None, mpd_base_url
='', mpd_url
=None):
2749 Parse formats from MPD manifest.
2751 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
2752 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
2753 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
2755 if not self
.get_param('dynamic_mpd', True):
2756 if mpd_doc
.get('type') == 'dynamic':
2759 namespace
= self
._search
_regex
(r
'(?i)^{([^}]+)?}MPD$', mpd_doc
.tag
, 'namespace', default
=None)
2762 return self
._xpath
_ns
(path
, namespace
)
2764 def is_drm_protected(element
):
2765 return element
.find(_add_ns('ContentProtection')) is not None
2767 def extract_multisegment_info(element
, ms_parent_info
):
2768 ms_info
= ms_parent_info
.copy()
2770 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2771 # common attributes and elements. We will only extract relevant
2773 def extract_common(source
):
2774 segment_timeline
= source
.find(_add_ns('SegmentTimeline'))
2775 if segment_timeline
is not None:
2776 s_e
= segment_timeline
.findall(_add_ns('S'))
2778 ms_info
['total_number'] = 0
2781 r
= int(s
.get('r', 0))
2782 ms_info
['total_number'] += 1 + r
2783 ms_info
['s'].append({
2784 't': int(s
.get('t', 0)),
2785 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2786 'd': int(s
.attrib
['d']),
2789 start_number
= source
.get('startNumber')
2791 ms_info
['start_number'] = int(start_number
)
2792 timescale
= source
.get('timescale')
2794 ms_info
['timescale'] = int(timescale
)
2795 segment_duration
= source
.get('duration')
2796 if segment_duration
:
2797 ms_info
['segment_duration'] = float(segment_duration
)
2799 def extract_Initialization(source
):
2800 initialization
= source
.find(_add_ns('Initialization'))
2801 if initialization
is not None:
2802 ms_info
['initialization_url'] = initialization
.attrib
['sourceURL']
2804 segment_list
= element
.find(_add_ns('SegmentList'))
2805 if segment_list
is not None:
2806 extract_common(segment_list
)
2807 extract_Initialization(segment_list
)
2808 segment_urls_e
= segment_list
.findall(_add_ns('SegmentURL'))
2810 ms_info
['segment_urls'] = [segment
.attrib
['media'] for segment
in segment_urls_e
]
2812 segment_template
= element
.find(_add_ns('SegmentTemplate'))
2813 if segment_template
is not None:
2814 extract_common(segment_template
)
2815 media
= segment_template
.get('media')
2817 ms_info
['media'] = media
2818 initialization
= segment_template
.get('initialization')
2820 ms_info
['initialization'] = initialization
2822 extract_Initialization(segment_template
)
2825 mpd_duration
= parse_duration(mpd_doc
.get('mediaPresentationDuration'))
2826 formats
, subtitles
= [], {}
2827 stream_numbers
= collections
.defaultdict(int)
2828 for period
in mpd_doc
.findall(_add_ns('Period')):
2829 period_duration
= parse_duration(period
.get('duration')) or mpd_duration
2830 period_ms_info
= extract_multisegment_info(period
, {
2834 for adaptation_set
in period
.findall(_add_ns('AdaptationSet')):
2835 adaption_set_ms_info
= extract_multisegment_info(adaptation_set
, period_ms_info
)
2836 for representation
in adaptation_set
.findall(_add_ns('Representation')):
2837 representation_attrib
= adaptation_set
.attrib
.copy()
2838 representation_attrib
.update(representation
.attrib
)
2839 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2840 mime_type
= representation_attrib
['mimeType']
2841 content_type
= representation_attrib
.get('contentType', mime_type
.split('/')[0])
2843 codec_str
= representation_attrib
.get('codecs', '')
2844 # Some kind of binary subtitle found in some youtube livestreams
2845 if mime_type
== 'application/x-rawcc':
2846 codecs
= {'scodec': codec_str}
2848 codecs
= parse_codecs(codec_str
)
2849 if content_type
not in ('video', 'audio', 'text'):
2850 if mime_type
== 'image/jpeg':
2851 content_type
= mime_type
2852 elif codecs
.get('vcodec', 'none') != 'none':
2853 content_type
= 'video'
2854 elif codecs
.get('acodec', 'none') != 'none':
2855 content_type
= 'audio'
2856 elif codecs
.get('scodec', 'none') != 'none':
2857 content_type
= 'text'
2858 elif mimetype2ext(mime_type
) in ('tt', 'dfxp', 'ttml', 'xml', 'json'):
2859 content_type
= 'text'
2861 self
.report_warning('Unknown MIME type %s in DASH manifest' % mime_type
)
2865 for element
in (representation
, adaptation_set
, period
, mpd_doc
):
2866 base_url_e
= element
.find(_add_ns('BaseURL'))
2867 if try_call(lambda: base_url_e
.text
) is not None:
2868 base_url
= base_url_e
.text
+ base_url
2869 if re
.match(r
'^https?://', base_url
):
2871 if mpd_base_url
and base_url
.startswith('/'):
2872 base_url
= urllib
.parse
.urljoin(mpd_base_url
, base_url
)
2873 elif mpd_base_url
and not re
.match(r
'^https?://', base_url
):
2874 if not mpd_base_url
.endswith('/'):
2876 base_url
= mpd_base_url
+ base_url
2877 representation_id
= representation_attrib
.get('id')
2878 lang
= representation_attrib
.get('lang')
2879 url_el
= representation
.find(_add_ns('BaseURL'))
2880 filesize
= int_or_none(url_el
.attrib
.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el
is not None else None)
2881 bandwidth
= int_or_none(representation_attrib
.get('bandwidth'))
2882 if representation_id
is not None:
2883 format_id
= representation_id
2885 format_id
= content_type
2887 format_id
= mpd_id
+ '-' + format_id
2888 if content_type
in ('video', 'audio'):
2890 'format_id': format_id
,
2891 'manifest_url': mpd_url
,
2892 'ext': mimetype2ext(mime_type
),
2893 'width': int_or_none(representation_attrib
.get('width')),
2894 'height': int_or_none(representation_attrib
.get('height')),
2895 'tbr': float_or_none(bandwidth
, 1000),
2896 'asr': int_or_none(representation_attrib
.get('audioSamplingRate')),
2897 'fps': int_or_none(representation_attrib
.get('frameRate')),
2898 'language': lang
if lang
not in ('mul', 'und', 'zxx', 'mis') else None,
2899 'format_note': 'DASH %s' % content_type
,
2900 'filesize': filesize
,
2901 'container': mimetype2ext(mime_type
) + '_dash',
2904 elif content_type
== 'text':
2906 'ext': mimetype2ext(mime_type
),
2907 'manifest_url': mpd_url
,
2908 'filesize': filesize
,
2910 elif content_type
== 'image/jpeg':
2911 # See test case in VikiIE
2912 # https://www.viki.com/videos/1175236v-choosing-spouse-by-lottery-episode-1
2914 'format_id': format_id
,
2916 'manifest_url': mpd_url
,
2917 'format_note': 'DASH storyboards (jpeg)',
2921 if is_drm_protected(adaptation_set
) or is_drm_protected(representation
):
2923 representation_ms_info
= extract_multisegment_info(representation
, adaption_set_ms_info
)
2925 def prepare_template(template_name
, identifiers
):
2926 tmpl
= representation_ms_info
[template_name
]
2927 if representation_id
is not None:
2928 tmpl
= tmpl
.replace('$RepresentationID$', representation_id
)
2929 # First of, % characters outside $...$ templates
2930 # must be escaped by doubling for proper processing
2931 # by % operator string formatting used further (see
2932 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2938 in_template
= not in_template
2939 elif c
== '%' and not in_template
:
2941 # Next, $...$ templates are translated to their
2942 # %(...) counterparts to be used with % operator
2943 t
= re
.sub(r
'\$(%s)\$' % '|'.join(identifiers
), r
'%(\1)d', t
)
2944 t
= re
.sub(r
'\$(%s)%%([^$]+)\$' % '|'.join(identifiers
), r
'%(\1)\2', t
)
2945 t
.replace('$$', '$')
2948 # @initialization is a regular template like @media one
2949 # so it should be handled just the same way (see
2950 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2951 if 'initialization' in representation_ms_info
:
2952 initialization_template
= prepare_template(
2954 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2955 # $Time$ shall not be included for @initialization thus
2956 # only $Bandwidth$ remains
2958 representation_ms_info
['initialization_url'] = initialization_template
% {
2959 'Bandwidth': bandwidth
,
2962 def location_key(location
):
2963 return 'url' if re
.match(r
'^https?://', location
) else 'path'
2965 if 'segment_urls' not in representation_ms_info
and 'media' in representation_ms_info
:
2967 media_template
= prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2968 media_location_key
= location_key(media_template
)
2970 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2971 # can't be used at the same time
2972 if '%(Number' in media_template
and 's' not in representation_ms_info
:
2973 segment_duration
= None
2974 if 'total_number' not in representation_ms_info
and 'segment_duration' in representation_ms_info
:
2975 segment_duration
= float_or_none(representation_ms_info
['segment_duration'], representation_ms_info
['timescale'])
2976 representation_ms_info
['total_number'] = int(math
.ceil(
2977 float_or_none(period_duration
, segment_duration
, default
=0)))
2978 representation_ms_info
['fragments'] = [{
2979 media_location_key
: media_template
% {
2980 'Number': segment_number
,
2981 'Bandwidth': bandwidth
,
2983 'duration': segment_duration
,
2984 } for segment_number
in range(
2985 representation_ms_info
['start_number'],
2986 representation_ms_info
['total_number'] + representation_ms_info
['start_number'])]
2988 # $Number*$ or $Time$ in media template with S list available
2989 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2990 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2991 representation_ms_info
['fragments'] = []
2994 segment_number
= representation_ms_info
['start_number']
2996 def add_segment_url():
2997 segment_url
= media_template
% {
2998 'Time': segment_time
,
2999 'Bandwidth': bandwidth
,
3000 'Number': segment_number
,
3002 representation_ms_info
['fragments'].append({
3003 media_location_key
: segment_url
,
3004 'duration': float_or_none(segment_d
, representation_ms_info
['timescale']),
3007 for num
, s
in enumerate(representation_ms_info
['s']):
3008 segment_time
= s
.get('t') or segment_time
3012 for r
in range(s
.get('r', 0)):
3013 segment_time
+= segment_d
3016 segment_time
+= segment_d
3017 elif 'segment_urls' in representation_ms_info
and 's' in representation_ms_info
:
3018 # No media template,
3019 # e.g. https://www.youtube.com/watch?v=iXZV5uAYMJI
3020 # or any YouTube dashsegments video
3023 timescale
= representation_ms_info
['timescale']
3024 for s
in representation_ms_info
['s']:
3025 duration
= float_or_none(s
['d'], timescale
)
3026 for r
in range(s
.get('r', 0) + 1):
3027 segment_uri
= representation_ms_info
['segment_urls'][segment_index
]
3029 location_key(segment_uri
): segment_uri
,
3030 'duration': duration
,
3033 representation_ms_info
['fragments'] = fragments
3034 elif 'segment_urls' in representation_ms_info
:
3035 # Segment URLs with no SegmentTimeline
3036 # E.g. https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
3037 # https://github.com/ytdl-org/youtube-dl/pull/14844
3039 segment_duration
= float_or_none(
3040 representation_ms_info
['segment_duration'],
3041 representation_ms_info
['timescale']) if 'segment_duration' in representation_ms_info
else None
3042 for segment_url
in representation_ms_info
['segment_urls']:
3044 location_key(segment_url
): segment_url
,
3046 if segment_duration
:
3047 fragment
['duration'] = segment_duration
3048 fragments
.append(fragment
)
3049 representation_ms_info
['fragments'] = fragments
3050 # If there is a fragments key available then we correctly recognized fragmented media.
3051 # Otherwise we will assume unfragmented media with direct access. Technically, such
3052 # assumption is not necessarily correct since we may simply have no support for
3053 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
3054 if 'fragments' in representation_ms_info
:
3056 # NB: mpd_url may be empty when MPD manifest is parsed from a string
3057 'url': mpd_url
or base_url
,
3058 'fragment_base_url': base_url
,
3060 'protocol': 'http_dash_segments' if mime_type
!= 'image/jpeg' else 'mhtml',
3062 if 'initialization_url' in representation_ms_info
:
3063 initialization_url
= representation_ms_info
['initialization_url']
3064 if not f
.get('url'):
3065 f
['url'] = initialization_url
3066 f
['fragments'].append({location_key(initialization_url): initialization_url}
)
3067 f
['fragments'].extend(representation_ms_info
['fragments'])
3068 if not period_duration
:
3069 period_duration
= try_get(
3070 representation_ms_info
,
3071 lambda r
: sum(frag
['duration'] for frag
in r
['fragments']), float)
3073 # Assuming direct URL to unfragmented media.
3075 if content_type
in ('video', 'audio', 'image/jpeg'):
3076 f
['manifest_stream_number'] = stream_numbers
[f
['url']]
3077 stream_numbers
[f
['url']] += 1
3079 elif content_type
== 'text':
3080 subtitles
.setdefault(lang
or 'und', []).append(f
)
3082 return formats
, subtitles
3084 def _extract_ism_formats(self
, *args
, **kwargs
):
3085 fmts
, subs
= self
._extract
_ism
_formats
_and
_subtitles
(*args
, **kwargs
)
3087 self
._report
_ignoring
_subs
('ISM')
3090 def _extract_ism_formats_and_subtitles(self
, ism_url
, video_id
, ism_id
=None, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
):
3091 res
= self
._download
_xml
_handle
(
3093 note
='Downloading ISM manifest' if note
is None else note
,
3094 errnote
='Failed to download ISM manifest' if errnote
is None else errnote
,
3095 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
3102 return self
._parse
_ism
_formats
_and
_subtitles
(ism_doc
, urlh
.geturl(), ism_id
)
3104 def _parse_ism_formats_and_subtitles(self
, ism_doc
, ism_url
, ism_id
=None):
3106 Parse formats from ISM manifest.
3108 1. [MS-SSTR]: Smooth Streaming Protocol,
3109 https://msdn.microsoft.com/en-us/library/ff469518.aspx
3111 if ism_doc
.get('IsLive') == 'TRUE':
3114 duration
= int(ism_doc
.attrib
['Duration'])
3115 timescale
= int_or_none(ism_doc
.get('TimeScale')) or 10000000
3119 for stream
in ism_doc
.findall('StreamIndex'):
3120 stream_type
= stream
.get('Type')
3121 if stream_type
not in ('video', 'audio', 'text'):
3123 url_pattern
= stream
.attrib
['Url']
3124 stream_timescale
= int_or_none(stream
.get('TimeScale')) or timescale
3125 stream_name
= stream
.get('Name')
3126 stream_language
= stream
.get('Language', 'und')
3127 for track
in stream
.findall('QualityLevel'):
3128 KNOWN_TAGS
= {'255': 'AACL', '65534': 'EC-3'}
3129 fourcc
= track
.get('FourCC') or KNOWN_TAGS
.get(track
.get('AudioTag'))
3130 # TODO: add support for WVC1 and WMAP
3131 if fourcc
not in ('H264', 'AVC1', 'AACL', 'TTML', 'EC-3'):
3132 self
.report_warning('%s is not a supported codec' % fourcc
)
3134 tbr
= int(track
.attrib
['Bitrate']) // 1000
3135 # [1] does not mention Width and Height attributes. However,
3136 # they're often present while MaxWidth and MaxHeight are
3137 # missing, so should be used as fallbacks
3138 width
= int_or_none(track
.get('MaxWidth') or track
.get('Width'))
3139 height
= int_or_none(track
.get('MaxHeight') or track
.get('Height'))
3140 sampling_rate
= int_or_none(track
.get('SamplingRate'))
3142 track_url_pattern
= re
.sub(r
'{[Bb]itrate}', track
.attrib
['Bitrate'], url_pattern
)
3143 track_url_pattern
= urllib
.parse
.urljoin(ism_url
, track_url_pattern
)
3149 stream_fragments
= stream
.findall('c')
3150 for stream_fragment_index
, stream_fragment
in enumerate(stream_fragments
):
3151 fragment_ctx
['time'] = int_or_none(stream_fragment
.get('t')) or fragment_ctx
['time']
3152 fragment_repeat
= int_or_none(stream_fragment
.get('r')) or 1
3153 fragment_ctx
['duration'] = int_or_none(stream_fragment
.get('d'))
3154 if not fragment_ctx
['duration']:
3156 next_fragment_time
= int(stream_fragment
[stream_fragment_index
+ 1].attrib
['t'])
3158 next_fragment_time
= duration
3159 fragment_ctx
['duration'] = (next_fragment_time
- fragment_ctx
['time']) / fragment_repeat
3160 for _
in range(fragment_repeat
):
3162 'url': re
.sub(r
'{start[ _]time}', str(fragment_ctx
['time']), track_url_pattern
),
3163 'duration': fragment_ctx
['duration'] / stream_timescale
,
3165 fragment_ctx
['time'] += fragment_ctx
['duration']
3167 if stream_type
== 'text':
3168 subtitles
.setdefault(stream_language
, []).append({
3172 'manifest_url': ism_url
,
3173 'fragments': fragments
,
3174 '_download_params': {
3175 'stream_type': stream_type
,
3176 'duration': duration
,
3177 'timescale': stream_timescale
,
3179 'language': stream_language
,
3180 'codec_private_data': track
.get('CodecPrivateData'),
3183 elif stream_type
in ('video', 'audio'):
3185 'format_id': join_nonempty(ism_id
, stream_name
, tbr
),
3187 'manifest_url': ism_url
,
3188 'ext': 'ismv' if stream_type
== 'video' else 'isma',
3192 'asr': sampling_rate
,
3193 'vcodec': 'none' if stream_type
== 'audio' else fourcc
,
3194 'acodec': 'none' if stream_type
== 'video' else fourcc
,
3196 'fragments': fragments
,
3197 'has_drm': ism_doc
.find('Protection') is not None,
3198 '_download_params': {
3199 'stream_type': stream_type
,
3200 'duration': duration
,
3201 'timescale': stream_timescale
,
3202 'width': width
or 0,
3203 'height': height
or 0,
3205 'language': stream_language
,
3206 'codec_private_data': track
.get('CodecPrivateData'),
3207 'sampling_rate': sampling_rate
,
3208 'channels': int_or_none(track
.get('Channels', 2)),
3209 'bits_per_sample': int_or_none(track
.get('BitsPerSample', 16)),
3210 'nal_unit_length_field': int_or_none(track
.get('NALUnitLengthField', 4)),
3213 return formats
, subtitles
3215 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):
3216 def absolute_url(item_url
):
3217 return urljoin(base_url
, item_url
)
3219 def parse_content_type(content_type
):
3220 if not content_type
:
3222 ctr
= re
.search(r
'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type
)
3224 mimetype
, codecs
= ctr
.groups()
3225 f
= parse_codecs(codecs
)
3226 f
['ext'] = mimetype2ext(mimetype
)
3230 def _media_formats(src
, cur_media_type
, type_info
=None):
3231 type_info
= type_info
or {}
3232 full_url
= absolute_url(src
)
3233 ext
= type_info
.get('ext') or determine_ext(full_url
)
3235 is_plain_url
= False
3236 formats
= self
._extract
_m
3u8_formats
(
3237 full_url
, video_id
, ext
='mp4',
3238 entry_protocol
=m3u8_entry_protocol
, m3u8_id
=m3u8_id
,
3239 preference
=preference
, quality
=quality
, fatal
=False)
3241 is_plain_url
= False
3242 formats
= self
._extract
_mpd
_formats
(
3243 full_url
, video_id
, mpd_id
=mpd_id
, fatal
=False)
3248 'vcodec': 'none' if cur_media_type
== 'audio' else None,
3251 return is_plain_url
, formats
3254 # amp-video and amp-audio are very similar to their HTML5 counterparts
3255 # so we will include them right here (see
3256 # https://www.ampproject.org/docs/reference/components/amp-video)
3257 # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/
3258 _MEDIA_TAG_NAME_RE
= r
'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)'
3259 media_tags
= [(media_tag
, media_tag_name
, media_type
, '')
3260 for media_tag
, media_tag_name
, media_type
3261 in re
.findall(r
'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE
, webpage
)]
3262 media_tags
.extend(re
.findall(
3263 # We only allow video|audio followed by a whitespace or '>'.
3264 # Allowing more characters may end up in significant slow down (see
3265 # https://github.com/ytdl-org/youtube-dl/issues/11979,
3266 # e.g. http://www.porntrex.com/maps/videositemap.xml).
3267 r
'(?s)(<(?P<tag>%s)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>' % _MEDIA_TAG_NAME_RE
, webpage
))
3268 for media_tag
, _
, media_type
, media_content
in media_tags
:
3273 media_attributes
= extract_attributes(media_tag
)
3274 src
= strip_or_none(dict_get(media_attributes
, ('src', 'data-video-src', 'data-src', 'data-source')))
3276 f
= parse_content_type(media_attributes
.get('type'))
3277 _
, formats
= _media_formats(src
, media_type
, f
)
3278 media_info
['formats'].extend(formats
)
3279 media_info
['thumbnail'] = absolute_url(media_attributes
.get('poster'))
3281 for source_tag
in re
.findall(r
'<source[^>]+>', media_content
):
3282 s_attr
= extract_attributes(source_tag
)
3283 # data-video-src and data-src are non standard but seen
3284 # several times in the wild
3285 src
= strip_or_none(dict_get(s_attr
, ('src', 'data-video-src', 'data-src', 'data-source')))
3288 f
= parse_content_type(s_attr
.get('type'))
3289 is_plain_url
, formats
= _media_formats(src
, media_type
, f
)
3291 # width, height, res, label and title attributes are
3292 # all not standard but seen several times in the wild
3295 for lbl
in ('label', 'title')
3296 if str_or_none(s_attr
.get(lbl
))
3298 width
= int_or_none(s_attr
.get('width'))
3299 height
= (int_or_none(s_attr
.get('height'))
3300 or int_or_none(s_attr
.get('res')))
3301 if not width
or not height
:
3303 resolution
= parse_resolution(lbl
)
3306 width
= width
or resolution
.get('width')
3307 height
= height
or resolution
.get('height')
3309 tbr
= parse_bitrate(lbl
)
3318 'format_id': s_attr
.get('label') or s_attr
.get('title'),
3320 f
.update(formats
[0])
3321 media_info
['formats'].append(f
)
3323 media_info
['formats'].extend(formats
)
3324 for track_tag
in re
.findall(r
'<track[^>]+>', media_content
):
3325 track_attributes
= extract_attributes(track_tag
)
3326 kind
= track_attributes
.get('kind')
3327 if not kind
or kind
in ('subtitles', 'captions'):
3328 src
= strip_or_none(track_attributes
.get('src'))
3331 lang
= track_attributes
.get('srclang') or track_attributes
.get('lang') or track_attributes
.get('label')
3332 media_info
['subtitles'].setdefault(lang
, []).append({
3333 'url': absolute_url(src
),
3335 for f
in media_info
['formats']:
3336 f
.setdefault('http_headers', {})['Referer'] = base_url
3337 if media_info
['formats'] or media_info
['subtitles']:
3338 entries
.append(media_info
)
3341 def _extract_akamai_formats(self
, *args
, **kwargs
):
3342 fmts
, subs
= self
._extract
_akamai
_formats
_and
_subtitles
(*args
, **kwargs
)
3344 self
._report
_ignoring
_subs
('akamai')
3347 def _extract_akamai_formats_and_subtitles(self
, manifest_url
, video_id
, hosts
={}):
3348 signed
= 'hdnea=' in manifest_url
3350 # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
3351 manifest_url
= re
.sub(
3352 r
'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
3353 '', manifest_url
).strip('?')
3358 hdcore_sign
= 'hdcore=3.7.0'
3359 f4m_url
= re
.sub(r
'(https?://[^/]+)/i/', r
'\1/z/', manifest_url
).replace('/master.m3u8', '/manifest.f4m')
3360 hds_host
= hosts
.get('hds')
3362 f4m_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hds_host
, f4m_url
)
3363 if 'hdcore=' not in f4m_url
:
3364 f4m_url
+= ('&' if '?' in f4m_url
else '?') + hdcore_sign
3365 f4m_formats
= self
._extract
_f
4m
_formats
(
3366 f4m_url
, video_id
, f4m_id
='hds', fatal
=False)
3367 for entry
in f4m_formats
:
3368 entry
.update({'extra_param_to_segment_url': hdcore_sign}
)
3369 formats
.extend(f4m_formats
)
3371 m3u8_url
= re
.sub(r
'(https?://[^/]+)/z/', r
'\1/i/', manifest_url
).replace('/manifest.f4m', '/master.m3u8')
3372 hls_host
= hosts
.get('hls')
3374 m3u8_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hls_host
, m3u8_url
)
3375 m3u8_formats
, m3u8_subtitles
= self
._extract
_m
3u8_formats
_and
_subtitles
(
3376 m3u8_url
, video_id
, 'mp4', 'm3u8_native',
3377 m3u8_id
='hls', fatal
=False)
3378 formats
.extend(m3u8_formats
)
3379 subtitles
= self
._merge
_subtitles
(subtitles
, m3u8_subtitles
)
3381 http_host
= hosts
.get('http')
3382 if http_host
and m3u8_formats
and not signed
:
3383 REPL_REGEX
= r
'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
3384 qualities
= re
.match(REPL_REGEX
, m3u8_url
).group(2).split(',')
3385 qualities_length
= len(qualities
)
3386 if len(m3u8_formats
) in (qualities_length
, qualities_length
+ 1):
3388 for f
in m3u8_formats
:
3389 if f
['vcodec'] != 'none':
3390 for protocol
in ('http', 'https'):
3392 del http_f
['manifest_url']
3394 REPL_REGEX
, protocol
+ fr
'://{http_host}/\g<1>{qualities[i]}\3', f
['url'])
3396 'format_id': http_f
['format_id'].replace('hls-', protocol
+ '-'),
3398 'protocol': protocol
,
3400 formats
.append(http_f
)
3403 return formats
, subtitles
3405 def _extract_wowza_formats(self
, url
, video_id
, m3u8_entry_protocol
='m3u8_native', skip_protocols
=[]):
3406 query
= urllib
.parse
.urlparse(url
).query
3407 url
= re
.sub(r
'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url
)
3409 r
'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url
)
3410 url_base
= mobj
.group('url')
3411 http_base_url
= '%s%s:%s' % ('http', mobj
.group('s') or '', url_base
)
3414 def manifest_url(manifest
):
3415 m_url
= f
'{http_base_url}/{manifest}'
3417 m_url
+= '?%s' % query
3420 if 'm3u8' not in skip_protocols
:
3421 formats
.extend(self
._extract
_m
3u8_formats
(
3422 manifest_url('playlist.m3u8'), video_id
, 'mp4',
3423 m3u8_entry_protocol
, m3u8_id
='hls', fatal
=False))
3424 if 'f4m' not in skip_protocols
:
3425 formats
.extend(self
._extract
_f
4m
_formats
(
3426 manifest_url('manifest.f4m'),
3427 video_id
, f4m_id
='hds', fatal
=False))
3428 if 'dash' not in skip_protocols
:
3429 formats
.extend(self
._extract
_mpd
_formats
(
3430 manifest_url('manifest.mpd'),
3431 video_id
, mpd_id
='dash', fatal
=False))
3432 if re
.search(r
'(?:/smil:|\.smil)', url_base
):
3433 if 'smil' not in skip_protocols
:
3434 rtmp_formats
= self
._extract
_smil
_formats
(
3435 manifest_url('jwplayer.smil'),
3436 video_id
, fatal
=False)
3437 for rtmp_format
in rtmp_formats
:
3438 rtsp_format
= rtmp_format
.copy()
3439 rtsp_format
['url'] = '%s/%s' % (rtmp_format
['url'], rtmp_format
['play_path'])
3440 del rtsp_format
['play_path']
3441 del rtsp_format
['ext']
3442 rtsp_format
.update({
3443 'url': rtsp_format
['url'].replace('rtmp://', 'rtsp://'),
3444 'format_id': rtmp_format
['format_id'].replace('rtmp', 'rtsp'),
3447 formats
.extend([rtmp_format
, rtsp_format
])
3449 for protocol
in ('rtmp', 'rtsp'):
3450 if protocol
not in skip_protocols
:
3452 'url': f
'{protocol}:{url_base}',
3453 'format_id': protocol
,
3454 'protocol': protocol
,
3458 def _find_jwplayer_data(self
, webpage
, video_id
=None, transform_source
=js_to_json
):
3460 r
'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P
=quote
)\
)(?
!</script
>).*?\
.setup\s
*\
((?P
<options
>[^
)]+)\
)',
3464 jwplayer_data = self._parse_json(mobj.group('options
'),
3466 transform_source=transform_source)
3467 except ExtractorError:
3470 if isinstance(jwplayer_data, dict):
3471 return jwplayer_data
3473 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
3474 jwplayer_data = self._find_jwplayer_data(
3475 webpage, video_id, transform_source=js_to_json)
3476 return self._parse_jwplayer_data(
3477 jwplayer_data, video_id, *args, **kwargs)
3479 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
3480 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3481 # JWPlayer backward compatibility: flattened playlists
3482 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
3483 if 'playlist
' not in jwplayer_data:
3484 jwplayer_data = {'playlist': [jwplayer_data]}
3488 # JWPlayer backward compatibility: single playlist item
3489 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
3490 if not isinstance(jwplayer_data['playlist
'], list):
3491 jwplayer_data['playlist
'] = [jwplayer_data['playlist
']]
3493 for video_data in jwplayer_data['playlist
']:
3494 # JWPlayer backward compatibility: flattened sources
3495 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
3496 if 'sources
' not in video_data:
3497 video_data['sources
'] = [video_data]
3499 this_video_id = video_id or video_data['mediaid
']
3501 formats = self._parse_jwplayer_formats(
3502 video_data['sources
'], video_id=this_video_id, m3u8_id=m3u8_id,
3503 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
3506 tracks = video_data.get('tracks
')
3507 if tracks and isinstance(tracks, list):
3508 for track in tracks:
3509 if not isinstance(track, dict):
3511 track_kind = track.get('kind
')
3512 if not track_kind or not isinstance(track_kind, str):
3514 if track_kind.lower() not in ('captions
', 'subtitles
'):
3516 track_url = urljoin(base_url, track.get('file'))
3519 subtitles.setdefault(track.get('label
') or 'en
', []).append({
3520 'url
': self._proto_relative_url(track_url)
3524 'id': this_video_id,
3525 'title
': unescapeHTML(video_data['title
'] if require_title else video_data.get('title
')),
3526 'description
': clean_html(video_data.get('description
')),
3527 'thumbnail
': urljoin(base_url, self._proto_relative_url(video_data.get('image
'))),
3528 'timestamp
': int_or_none(video_data.get('pubdate
')),
3529 'duration
': float_or_none(jwplayer_data.get('duration
') or video_data.get('duration
')),
3530 'subtitles
': subtitles,
3532 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
3533 if len(formats) == 1 and re.search(r'^
(?
:http|
//).*(?
:youtube\
.com|youtu\
.be
)/.+', formats[0]['url
']):
3535 '_type
': 'url_transparent
',
3536 'url
': formats[0]['url
'],
3539 self._sort_formats(formats)
3540 entry['formats
'] = formats
3541 entries.append(entry)
3542 if len(entries) == 1:
3545 return self.playlist_result(entries)
3547 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
3548 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3551 for source in jwplayer_sources_data:
3552 if not isinstance(source, dict):
3554 source_url = urljoin(
3555 base_url, self._proto_relative_url(source.get('file')))
3556 if not source_url or source_url in urls:
3558 urls.append(source_url)
3559 source_type = source.get('type') or ''
3560 ext = mimetype2ext(source_type) or determine_ext(source_url)
3561 if source_type == 'hls
' or ext == 'm3u8
':
3562 formats.extend(self._extract_m3u8_formats(
3563 source_url, video_id, 'mp4
', entry_protocol='m3u8_native
',
3564 m3u8_id=m3u8_id, fatal=False))
3565 elif source_type == 'dash
' or ext == 'mpd
':
3566 formats.extend(self._extract_mpd_formats(
3567 source_url, video_id, mpd_id=mpd_id, fatal=False))
3569 formats.extend(self._extract_smil_formats(
3570 source_url, video_id, fatal=False))
3571 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3572 elif source_type.startswith('audio
') or ext in (
3573 'oga
', 'aac
', 'mp3
', 'mpeg
', 'vorbis
'):
3580 height = int_or_none(source.get('height
'))
3582 # Often no height is provided but there is a label in
3583 # format like "1080p", "720p SD", or 1080.
3584 height = int_or_none(self._search_regex(
3585 r'^
(\d{3,4}
)[pP
]?
(?
:\b|$
)', str(source.get('label
') or ''),
3586 'height
', default=None))
3589 'width
': int_or_none(source.get('width
')),
3591 'tbr
': int_or_none(source.get('bitrate
'), scale=1000),
3592 'filesize
': int_or_none(source.get('filesize
')),
3595 if source_url.startswith('rtmp
'):
3596 a_format['ext
'] = 'flv
'
3597 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3598 # of jwplayer.flash.swf
3599 rtmp_url_parts = re.split(
3600 r'((?
:mp4|mp3|flv
):)', source_url, 1)
3601 if len(rtmp_url_parts) == 3:
3602 rtmp_url, prefix, play_path = rtmp_url_parts
3605 'play_path
': prefix + play_path,
3608 a_format.update(rtmp_params)
3609 formats.append(a_format)
3612 def _live_title(self, name):
3613 self._downloader.deprecation_warning('yt_dlp
.InfoExtractor
._live
_title
is deprecated
and does
not work
as expected
')
3616 def _int(self, v, name, fatal=False, **kwargs):
3617 res = int_or_none(v, **kwargs)
3619 msg = f'Failed to extract {name}
: Could
not parse value {v!r}
'
3621 raise ExtractorError(msg)
3623 self.report_warning(msg)
3626 def _float(self, v, name, fatal=False, **kwargs):
3627 res = float_or_none(v, **kwargs)
3629 msg = f'Failed to extract {name}
: Could
not parse value {v!r}
'
3631 raise ExtractorError(msg)
3633 self.report_warning(msg)
3636 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
3637 path='/', secure=False, discard=False, rest={}, **kwargs):
3638 cookie = http.cookiejar.Cookie(
3639 0, name, value, port, port is not None, domain, True,
3640 domain.startswith('.'), path, True, secure, expire_time,
3641 discard, None, None, rest)
3642 self.cookiejar.set_cookie(cookie)
3644 def _get_cookies(self, url):
3645 """ Return a http.cookies.SimpleCookie with the cookies for the url """
3646 return LenientSimpleCookie(self._downloader._calc_cookies(url))
3648 def _apply_first_set_cookie_header(self, url_handle, cookie):
3650 Apply first Set-Cookie header instead of the last. Experimental.
3652 Some sites (e.g. [1-3]) may serve two cookies under the same name
3653 in Set-Cookie header and expect the first (old) one to be set rather
3654 than second (new). However, as of RFC6265 the newer one cookie
3655 should be set into cookie store what actually happens.
3656 We will workaround this issue by resetting the cookie to
3657 the first one manually.
3658 1. https://new.vk.com/
3659 2. https://github.com/ytdl-org/youtube-dl/issues/9841#issuecomment-227871201
3660 3. https://learning.oreilly.com/
3662 for header, cookies in url_handle.headers.items():
3663 if header.lower() != 'set-cookie
':
3665 cookies = cookies.encode('iso
-8859-1').decode('utf
-8')
3666 cookie_value = re.search(
3667 r'%s=(.+?
);.*?
\b[Dd
]omain
=(.+?
)(?
:[,;]|$
)' % cookie, cookies)
3669 value, domain = cookie_value.groups()
3670 self._set_cookie(domain, cookie, value)
3674 def get_testcases(cls, include_onlymatching=False):
3675 t = getattr(cls, '_TEST
', None)
3677 assert not hasattr(cls, '_TESTS
'), f'{cls.ie_key()}IE has _TEST
and _TESTS
'
3680 tests = getattr(cls, '_TESTS
', [])
3682 if not include_onlymatching and t.get('only_matching
', False):
3684 t['name
'] = cls.ie_key()
3688 def get_webpage_testcases(cls):
3689 tests = getattr(cls, '_WEBPAGE_TESTS
', [])
3691 t['name
'] = cls.ie_key()
3696 """Get age limit from the testcases"""
3697 return max(traverse_obj(
3698 (*cls.get_testcases(include_onlymatching=False), *cls.get_webpage_testcases()),
3699 (..., (('playlist
', 0), None), 'info_dict
', 'age_limit
')) or [0])
3702 def is_suitable(cls, age_limit):
3703 """Test whether the extractor is generally suitable for the given age limit"""
3704 return not age_restricted(cls.age_limit, age_limit)
3707 def description(cls, *, markdown=True, search_examples=None):
3708 """Description of the extractor"""
3710 if cls._NETRC_MACHINE:
3712 desc += f' [<abbr title
="netrc machine"><em
>{cls._NETRC_MACHINE}
</em
></abbr
>]'
3714 desc += f' [{cls._NETRC_MACHINE}
]'
3715 if cls.IE_DESC is False:
3718 desc += f' {cls.IE_DESC}
'
3720 desc += f'; "{cls.SEARCH_KEY}:" prefix
'
3722 _COUNTS = ('', '5', '10', 'all
')
3723 desc += f' (e
.g
. "{cls.SEARCH_KEY}{random.choice(_COUNTS)}:{random.choice(search_examples)}")'
3724 if not cls.working():
3725 desc += ' (**Currently broken
**)' if markdown else ' (Currently broken
)'
3727 name = f' - **{cls.IE_NAME}
**' if markdown else cls.IE_NAME
3728 return f'{name}
:{desc}
' if desc else name
3730 def extract_subtitles(self, *args, **kwargs):
3731 if (self.get_param('writesubtitles
', False)
3732 or self.get_param('listsubtitles
')):
3733 return self._get_subtitles(*args, **kwargs)
3736 def _get_subtitles(self, *args, **kwargs):
3737 raise NotImplementedError('This method must be implemented by subclasses
')
3739 def extract_comments(self, *args, **kwargs):
3740 if not self.get_param('getcomments
'):
3742 generator = self._get_comments(*args, **kwargs)
3749 comments.append(next(generator))
3750 except StopIteration:
3752 except KeyboardInterrupt:
3753 self.to_screen('Interrupted by user
')
3754 except Exception as e:
3755 if self.get_param('ignoreerrors
') is not True:
3757 self._downloader.report_error(e)
3758 comment_count = len(comments)
3759 self.to_screen(f'Extracted {comment_count} comments
')
3761 'comments
': comments,
3762 'comment_count
': None if interrupted else comment_count
3766 def _get_comments(self, *args, **kwargs):
3767 raise NotImplementedError('This method must be implemented by subclasses
')
3770 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
3771 """ Merge subtitle items for one language. Items with duplicated URLs/data
3772 will be dropped. """
3773 list1_data = {(item.get('url'), item.get('data')) for item in subtitle_list1}
3774 ret = list(subtitle_list1)
3775 ret.extend(item for item in subtitle_list2 if (item.get('url
'), item.get('data
')) not in list1_data)
3779 def _merge_subtitles(cls, *dicts, target=None):
3780 """ Merge subtitle dictionaries, language by language. """
3784 for lang, subs in d.items():
3785 target[lang] = cls._merge_subtitle_items(target.get(lang, []), subs)
3788 def extract_automatic_captions(self, *args, **kwargs):
3789 if (self.get_param('writeautomaticsub
', False)
3790 or self.get_param('listsubtitles
')):
3791 return self._get_automatic_captions(*args, **kwargs)
3794 def _get_automatic_captions(self, *args, **kwargs):
3795 raise NotImplementedError('This method must be implemented by subclasses
')
3797 @functools.cached_property
3798 def _cookies_passed(self):
3799 """Whether cookies have been passed to YoutubeDL"""
3800 return self.get_param('cookiefile
') is not None or self.get_param('cookiesfrombrowser
') is not None
3802 def mark_watched(self, *args, **kwargs):
3803 if not self.get_param('mark_watched
', False):
3805 if self.supports_login() and self._get_login_info()[0] is not None or self._cookies_passed:
3806 self._mark_watched(*args, **kwargs)
3808 def _mark_watched(self, *args, **kwargs):
3809 raise NotImplementedError('This method must be implemented by subclasses
')
3811 def geo_verification_headers(self):
3813 geo_verification_proxy = self.get_param('geo_verification_proxy
')
3814 if geo_verification_proxy:
3815 headers['Ytdl
-request
-proxy
'] = geo_verification_proxy
3819 def _generic_id(url):
3820 return urllib.parse.unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
3823 def _generic_title(url):
3824 return urllib.parse.unquote(os.path.splitext(url_basename(url))[0])
3827 def _availability(is_private=None, needs_premium=None, needs_subscription=None, needs_auth=None, is_unlisted=None):
3828 all_known = all(map(
3829 lambda x: x is not None,
3830 (is_private, needs_premium, needs_subscription, needs_auth, is_unlisted)))
3832 'private
' if is_private
3833 else 'premium_only
' if needs_premium
3834 else 'subscriber_only
' if needs_subscription
3835 else 'needs_auth
' if needs_auth
3836 else 'unlisted
' if is_unlisted
3837 else 'public
' if all_known
3840 def _configuration_arg(self, key, default=NO_DEFAULT, *, ie_key=None, casesense=False):
3842 @returns A list of values for the extractor argument given by "key"
3843 or "default" if no such key is present
3844 @param default The default value to return when the key is not present (default: [])
3845 @param casesense When false, the values are converted to lower case
3848 self._downloader.params, ('extractor_args
', (ie_key or self.ie_key()).lower(), key))
3850 return [] if default is NO_DEFAULT else default
3851 return list(val) if casesense else [x.lower() for x in val]
3853 def _yes_playlist(self, playlist_id, video_id, smuggled_data=None, *, playlist_label='playlist
', video_label='video
'):
3854 if not playlist_id or not video_id:
3857 no_playlist = (smuggled_data or {}).get('force_noplaylist
')
3858 if no_playlist is not None:
3859 return not no_playlist
3861 video_id = '' if video_id is True else f' {video_id}
'
3862 playlist_id = '' if playlist_id is True else f' {playlist_id}
'
3863 if self.get_param('noplaylist
'):
3864 self.to_screen(f'Downloading just the {video_label}{video_id} because of
--no
-playlist
')
3866 self.to_screen(f'Downloading {playlist_label}{playlist_id}
- add
--no
-playlist to download just the {video_label}{video_id}
')
3869 def _error_or_warning(self, err, _count=None, _retries=0, *, fatal=True):
3870 RetryManager.report_retry(
3871 err, _count or int(fatal), _retries,
3872 info=self.to_screen, warn=self.report_warning, error=None if fatal else self.report_warning,
3873 sleep_func=self.get_param('retry_sleep_functions
', {}).get('extractor
'))
3875 def RetryManager(self, **kwargs):
3876 return RetryManager(self.get_param('extractor_retries
', 3), self._error_or_warning, **kwargs)
3879 def extract_from_webpage(cls, ydl, url, webpage):
3880 ie = (cls if isinstance(cls._extract_from_webpage, types.MethodType)
3881 else ydl.get_info_extractor(cls.ie_key()))
3882 for info in ie._extract_from_webpage(url, webpage) or []:
3883 # url = None since we do not want to set (webpage/original)_url
3884 ydl.add_default_extra_info(info, ie, None)
3888 def _extract_from_webpage(cls, url, webpage):
3889 for embed_url in orderedSet(
3890 cls._extract_embed_urls(url, webpage) or [], lazy=True):
3891 yield cls.url_result(embed_url, None if cls._VALID_URL is False else cls)
3894 def _extract_embed_urls(cls, url, webpage):
3895 """@returns all the embed urls on the webpage"""
3896 if '_EMBED_URL_RE
' not in cls.__dict__:
3897 assert isinstance(cls._EMBED_REGEX, (list, tuple))
3898 for idx, regex in enumerate(cls._EMBED_REGEX):
3899 assert regex.count('(?P
<url
>') == 1, \
3900 f'{cls.__name__}
._EMBED
_REGEX
[{idx}
] must have exactly
1 url group
\n\t{regex}
'
3901 cls._EMBED_URL_RE = tuple(map(re.compile, cls._EMBED_REGEX))
3903 for regex in cls._EMBED_URL_RE:
3904 for mobj in regex.finditer(webpage):
3905 embed_url = urllib.parse.urljoin(url, unescapeHTML(mobj.group('url
')))
3906 if cls._VALID_URL is False or cls.suitable(embed_url):
3909 class StopExtraction(Exception):
3913 def _extract_url(cls, webpage): # TODO: Remove
3914 """Only for compatibility with some older extractors"""
3915 return next(iter(cls._extract_embed_urls(None, webpage) or []), None)
3918 def __init_subclass__(cls, *, plugin_name=None, **kwargs):
3920 mro = inspect.getmro(cls)
3921 super_class = cls.__wrapped__ = mro[mro.index(cls) + 1]
3922 cls.IE_NAME, cls.ie_key = f'{super_class.IE_NAME}
+{plugin_name}
', super_class.ie_key
3923 while getattr(super_class, '__wrapped__
', None):
3924 super_class = super_class.__wrapped__
3925 setattr(sys.modules[super_class.__module__], super_class.__name__, cls)
3927 return super().__init_subclass__(**kwargs)
3930 class SearchInfoExtractor(InfoExtractor):
3932 Base class for paged search queries extractors.
3933 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
3934 Instances should define _SEARCH_KEY and optionally _MAX_RESULTS
3937 _MAX_RESULTS = float('inf
')
3940 def _VALID_URL(cls):
3941 return r'%s(?P
<prefix
>|
[1-9][0-9]*|all
):(?P
<query
>[\s\S
]+)' % cls._SEARCH_KEY
3943 def _real_extract(self, query):
3944 prefix, query = self._match_valid_url(query).group('prefix
', 'query
')
3946 return self._get_n_results(query, 1)
3947 elif prefix == 'all
':
3948 return self._get_n_results(query, self._MAX_RESULTS)
3952 raise ExtractorError(f'invalid download number {n}
for query
"{query}"')
3953 elif n > self._MAX_RESULTS:
3954 self.report_warning('%s returns
max %i results (you requested
%i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3955 n = self._MAX_RESULTS
3956 return self._get_n_results(query, n)
3958 def _get_n_results(self, query, n):
3959 """Get a specified number of results for a query.
3960 Either this function or _search_results must be overridden by subclasses """
3961 return self.playlist_result(
3962 itertools.islice(self._search_results(query), 0, None if n == float('inf
') else n),
3965 def _search_results(self, query):
3966 """Returns an iterator of search results"""
3967 raise NotImplementedError('This method must be implemented by subclasses
')
3970 def SEARCH_KEY(cls):
3971 return cls._SEARCH_KEY
3974 class UnsupportedURLIE(InfoExtractor):
3979 def _real_extract(self, url):
3980 raise UnsupportedError(url)