20 import xml
.etree
.ElementTree
22 from ..compat
import functools
# isort: split
23 from ..compat
import compat_etree_fromstring
, compat_expanduser
, compat_os_name
24 from ..downloader
import FileDownloader
25 from ..downloader
.f4m
import get_base_url
, remove_encrypted_media
61 parse_m3u8_attributes
,
88 """Information Extractor class.
90 Information extractors are the classes that, given a URL, extract
91 information about the video (or videos) the URL refers to. This
92 information includes the real video URL, the video title, author and
93 others. The information is stored in a dictionary which is then
94 passed to the YoutubeDL. The YoutubeDL processes this
95 information possibly downloading the video to the file system, among
96 other possible outcomes.
98 The type field determines the type of the result.
99 By far the most common value (and the default if _type is missing) is
100 "video", which indicates a single video.
102 For a video, the dictionaries must include the following fields:
104 id: Video identifier.
105 title: Video title, unescaped. Set to an empty string if video has
106 no title as opposed to "None" which signifies that the
107 extractor failed to obtain a title
109 Additionally, it must contain either a formats entry or a url one:
111 formats: A list of dictionaries for each format available, ordered
112 from worst to best quality.
115 * url The mandatory URL representing the media:
116 for plain file media - HTTP URL of this file,
118 for HLS - URL of the M3U8 media playlist,
119 for HDS - URL of the F4M manifest,
121 - HTTP URL to plain file media (in case of
123 - URL of the MPD manifest or base URL
124 representing the media if MPD manifest
125 is parsed from a string (in case of
127 for MSS - URL of the ISM manifest.
129 The URL of the manifest file in case of
131 for HLS - URL of the M3U8 master playlist,
132 for HDS - URL of the F4M manifest,
133 for DASH - URL of the MPD manifest,
134 for MSS - URL of the ISM manifest.
135 * manifest_stream_number (For internal use only)
136 The index of the stream in the manifest file
137 * ext Will be calculated from URL if missing
138 * format A human-readable description of the format
139 ("mp4 container with h264/opus").
140 Calculated from the format_id, width, height.
141 and format_note fields if missing.
142 * format_id A short description of the format
143 ("mp4_h264_opus" or "19").
144 Technically optional, but strongly recommended.
145 * format_note Additional info about the format
146 ("3D" or "DASH video")
147 * width Width of the video, if known
148 * height Height of the video, if known
149 * resolution Textual description of width and height
150 * dynamic_range The dynamic range of the video. One of:
151 "SDR" (None), "HDR10", "HDR10+, "HDR12", "HLG, "DV"
152 * tbr Average bitrate of audio and video in KBit/s
153 * abr Average audio bitrate in KBit/s
154 * acodec Name of the audio codec in use
155 * asr Audio sampling rate in Hertz
156 * vbr Average video bitrate in KBit/s
158 * vcodec Name of the video codec in use
159 * container Name of the container format
160 * filesize The number of bytes, if known in advance
161 * filesize_approx An estimate for the number of bytes
162 * player_url SWF Player URL (used for rtmpdump).
163 * protocol The protocol that will be used for the actual
164 download, lower-case. One of "http", "https" or
165 one of the protocols defined in downloader.PROTOCOL_MAP
167 Base URL for fragments. Each fragment's path
168 value (if present) will be relative to
170 * fragments A list of fragments of a fragmented media.
171 Each fragment entry must contain either an url
172 or a path. If an url is present it should be
173 considered by a client. Otherwise both path and
174 fragment_base_url must be present. Here is
175 the list of all potential fields:
176 * "url" - fragment's URL
177 * "path" - fragment's path relative to
179 * "duration" (optional, int or float)
180 * "filesize" (optional, int)
181 * is_from_start Is a live format that can be downloaded
182 from the start. Boolean
183 * preference Order number of this format. If this field is
184 present and not None, the formats get sorted
185 by this field, regardless of all other values.
186 -1 for default (order by other properties),
187 -2 or smaller for less than default.
188 < -1000 to hide the format (if there is
189 another one which is strictly better)
190 * language Language code, e.g. "de" or "en-US".
191 * language_preference Is this in the language mentioned in
193 10 if it's what the URL is about,
194 -1 for default (don't know),
195 -10 otherwise, other values reserved for now.
196 * quality Order number of the video quality of this
197 format, irrespective of the file format.
198 -1 for default (order by other properties),
199 -2 or smaller for less than default.
200 * source_preference Order number for this video source
201 (quality takes higher priority)
202 -1 for default (order by other properties),
203 -2 or smaller for less than default.
204 * http_headers A dictionary of additional HTTP headers
205 to add to the request.
206 * stretched_ratio If given and not 1, indicates that the
207 video's pixels are not square.
208 width : height ratio as float.
209 * no_resume The server does not support resuming the
210 (HTTP or RTMP) download. Boolean.
211 * has_drm The format has DRM and cannot be downloaded. Boolean
212 * downloader_options A dictionary of downloader options
213 (For internal use only)
214 * http_chunk_size Chunk size for HTTP downloads
215 * ffmpeg_args Extra arguments for ffmpeg downloader
216 RTMP formats can also have the additional fields: page_url,
217 app, play_path, tc_url, flash_version, rtmp_live, rtmp_conn,
218 rtmp_protocol, rtmp_real_time
220 url: Final video URL.
221 ext: Video filename extension.
222 format: The video format, defaults to ext (used for --get-format)
223 player_url: SWF Player URL (used for rtmpdump).
225 The following fields are optional:
227 direct: True if a direct video file was given (must only be set by GenericIE)
228 alt_title: A secondary title of the video.
229 display_id An alternative identifier for the video, not necessarily
230 unique, but available before title. Typically, id is
231 something like "4234987", title "Dancing naked mole rats",
232 and display_id "dancing-naked-mole-rats"
233 thumbnails: A list of dictionaries, with the following entries:
234 * "id" (optional, string) - Thumbnail format ID
236 * "preference" (optional, int) - quality of the image
237 * "width" (optional, int)
238 * "height" (optional, int)
239 * "resolution" (optional, string "{width}x{height}",
241 * "filesize" (optional, int)
242 * "http_headers" (dict) - HTTP headers for the request
243 thumbnail: Full URL to a video thumbnail image.
244 description: Full video description.
245 uploader: Full name of the video uploader.
246 license: License name the video is licensed under.
247 creator: The creator of the video.
248 timestamp: UNIX timestamp of the moment the video was uploaded
249 upload_date: Video upload date in UTC (YYYYMMDD).
250 If not explicitly set, calculated from timestamp
251 release_timestamp: UNIX timestamp of the moment the video was released.
252 If it is not clear whether to use timestamp or this, use the former
253 release_date: The date (YYYYMMDD) when the video was released in UTC.
254 If not explicitly set, calculated from release_timestamp
255 modified_timestamp: UNIX timestamp of the moment the video was last modified.
256 modified_date: The date (YYYYMMDD) when the video was last modified in UTC.
257 If not explicitly set, calculated from modified_timestamp
258 uploader_id: Nickname or id of the video uploader.
259 uploader_url: Full URL to a personal webpage of the video uploader.
260 channel: Full name of the channel the video is uploaded on.
261 Note that channel fields may or may not repeat uploader
262 fields. This depends on a particular extractor.
263 channel_id: Id of the channel.
264 channel_url: Full URL to a channel webpage.
265 channel_follower_count: Number of followers of the channel.
266 location: Physical location where the video was filmed.
267 subtitles: The available subtitles as a dictionary in the format
268 {tag: subformats}. "tag" is usually a language code, and
269 "subformats" is a list sorted from lower to higher
270 preference, each element is a dictionary with the "ext"
272 * "data": The subtitles file contents
273 * "url": A URL pointing to the subtitles file
274 It can optionally also have:
275 * "name": Name or description of the subtitles
276 * "http_headers": A dictionary of additional HTTP headers
277 to add to the request.
278 "ext" will be calculated from URL if missing
279 automatic_captions: Like 'subtitles'; contains automatically generated
280 captions instead of normal subtitles
281 duration: Length of the video in seconds, as an integer or float.
282 view_count: How many users have watched the video on the platform.
283 like_count: Number of positive ratings of the video
284 dislike_count: Number of negative ratings of the video
285 repost_count: Number of reposts of the video
286 average_rating: Average rating give by users, the scale used depends on the webpage
287 comment_count: Number of comments on the video
288 comments: A list of comments, each with one or more of the following
289 properties (all but one of text or html optional):
290 * "author" - human-readable name of the comment author
291 * "author_id" - user ID of the comment author
292 * "author_thumbnail" - The thumbnail of the comment author
294 * "html" - Comment as HTML
295 * "text" - Plain text of the comment
296 * "timestamp" - UNIX timestamp of comment
297 * "parent" - ID of the comment this one is replying to.
298 Set to "root" to indicate that this is a
299 comment to the original video.
300 * "like_count" - Number of positive ratings of the comment
301 * "dislike_count" - Number of negative ratings of the comment
302 * "is_favorited" - Whether the comment is marked as
303 favorite by the video uploader
304 * "author_is_uploader" - Whether the comment is made by
306 age_limit: Age restriction for the video, as an integer (years)
307 webpage_url: The URL to the video webpage, if given to yt-dlp it
308 should allow to get the same result again. (It will be set
309 by YoutubeDL if it's missing)
310 categories: A list of categories that the video falls in, for example
312 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
313 cast: A list of the video cast
314 is_live: True, False, or None (=unknown). Whether this video is a
315 live stream that goes on instead of a fixed-length video.
316 was_live: True, False, or None (=unknown). Whether this video was
317 originally a live stream.
318 live_status: None (=unknown), 'is_live', 'is_upcoming', 'was_live', 'not_live'
319 or 'post_live' (was live, but VOD is not yet processed)
320 If absent, automatically set from is_live, was_live
321 start_time: Time in seconds where the reproduction should start, as
322 specified in the URL.
323 end_time: Time in seconds where the reproduction should end, as
324 specified in the URL.
325 chapters: A list of dictionaries, with the following entries:
326 * "start_time" - The start time of the chapter in seconds
327 * "end_time" - The end time of the chapter in seconds
328 * "title" (optional, string)
329 playable_in_embed: Whether this video is allowed to play in embedded
330 players on other sites. Can be True (=always allowed),
331 False (=never allowed), None (=unknown), or a string
332 specifying the criteria for embedability (Eg: 'whitelist')
333 availability: Under what condition the video is available. One of
334 'private', 'premium_only', 'subscriber_only', 'needs_auth',
335 'unlisted' or 'public'. Use 'InfoExtractor._availability'
337 _old_archive_ids: A list of old archive ids needed for backward compatibility
338 __post_extractor: A function to be called just before the metadata is
339 written to either disk, logger or console. The function
340 must return a dict which will be added to the info_dict.
341 This is usefull for additional information that is
342 time-consuming to extract. Note that the fields thus
343 extracted will not be available to output template and
344 match_filter. So, only "comments" and "comment_count" are
345 currently allowed to be extracted via this method.
347 The following fields should only be used when the video belongs to some logical
350 chapter: Name or title of the chapter the video belongs to.
351 chapter_number: Number of the chapter the video belongs to, as an integer.
352 chapter_id: Id of the chapter the video belongs to, as a unicode string.
354 The following fields should only be used when the video is an episode of some
355 series, programme or podcast:
357 series: Title of the series or programme the video episode belongs to.
358 series_id: Id of the series or programme the video episode belongs to, as a unicode string.
359 season: Title of the season the video episode belongs to.
360 season_number: Number of the season the video episode belongs to, as an integer.
361 season_id: Id of the season the video episode belongs to, as a unicode string.
362 episode: Title of the video episode. Unlike mandatory video title field,
363 this field should denote the exact title of the video episode
364 without any kind of decoration.
365 episode_number: Number of the video episode within a season, as an integer.
366 episode_id: Id of the video episode, as a unicode string.
368 The following fields should only be used when the media is a track or a part of
371 track: Title of the track.
372 track_number: Number of the track within an album or a disc, as an integer.
373 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
375 artist: Artist(s) of the track.
376 genre: Genre(s) of the track.
377 album: Title of the album the track belongs to.
378 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
379 album_artist: List of all artists appeared on the album (e.g.
380 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
382 disc_number: Number of the disc or other physical medium the track belongs to,
384 release_year: Year (YYYY) when the album was released.
385 composer: Composer of the piece
387 The following fields should only be set for clips that should be cut from the original video:
389 section_start: Start time of the section in seconds
390 section_end: End time of the section in seconds
392 The following fields should only be set for storyboards:
393 rows: Number of rows in each storyboard fragment, as an integer
394 columns: Number of columns in each storyboard fragment, as an integer
396 Unless mentioned otherwise, the fields should be Unicode strings.
398 Unless mentioned otherwise, None is equivalent to absence of information.
401 _type "playlist" indicates multiple videos.
402 There must be a key "entries", which is a list, an iterable, or a PagedList
403 object, each element of which is a valid dictionary by this specification.
405 Additionally, playlists can have "id", "title", and any other relevant
406 attributes with the same semantics as videos (see above).
408 It can also have the following optional fields:
410 playlist_count: The total number of videos in a playlist. If not given,
411 YoutubeDL tries to calculate it from "entries"
414 _type "multi_video" indicates that there are multiple videos that
415 form a single show, for examples multiple acts of an opera or TV episode.
416 It must have an entries key like a playlist and contain all the keys
417 required for a video at the same time.
420 _type "url" indicates that the video must be extracted from another
421 location, possibly by a different extractor. Its only required key is:
422 "url" - the next URL to extract.
423 The key "ie_key" can be set to the class name (minus the trailing "IE",
424 e.g. "Youtube") if the extractor class is known in advance.
425 Additionally, the dictionary may have any properties of the resolved entity
426 known in advance, for example "title" if the title of the referred video is
430 _type "url_transparent" entities have the same specification as "url", but
431 indicate that the given additional information is more precise than the one
432 associated with the resolved URL.
433 This is useful when a site employs a video service that hosts the video and
434 its technical metadata, but that video service does not embed a useful
435 title, description etc.
438 Subclasses of this should also be added to the list of extractors and
439 should define a _VALID_URL regexp and, re-define the _real_extract() and
440 (optionally) _real_initialize() methods.
442 Subclasses may also override suitable() if necessary, but ensure the function
443 signature is preserved and that this function imports everything it needs
444 (except other extractors), so that lazy_extractors works correctly.
446 Subclasses can define a list of _EMBED_REGEX, which will be searched for in
447 the HTML of Generic webpages. It may also override _extract_embed_urls
448 or _extract_from_webpage as necessary. While these are normally classmethods,
449 _extract_from_webpage is allowed to be an instance method.
451 _extract_from_webpage may raise self.StopExtraction() to stop further
452 processing of the webpage and obtain exclusive rights to it. This is useful
453 when the extractor cannot reliably be matched using just the URL.
454 Eg: invidious/peertube instances
456 Embed-only extractors can be defined by setting _VALID_URL = False.
458 To support username + password (or netrc) login, the extractor must define a
459 _NETRC_MACHINE and re-define _perform_login(username, password) and
460 (optionally) _initialize_pre_login() methods. The _perform_login method will
461 be called between _initialize_pre_login and _real_initialize if credentials
462 are passed by the user. In cases where it is necessary to have the login
463 process as part of the extraction rather than initialization, _perform_login
464 can be left undefined.
466 _GEO_BYPASS attribute may be set to False in order to disable
467 geo restriction bypass mechanisms for a particular extractor.
468 Though it won't disable explicit geo restriction bypass based on
469 country code provided with geo_bypass_country.
471 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
472 countries for this extractor. One of these countries will be used by
473 geo restriction bypass mechanism right away in order to bypass
474 geo restriction, of course, if the mechanism is not disabled.
476 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
477 IP blocks in CIDR notation for this extractor. One of these IP blocks
478 will be used by geo restriction bypass mechanism similarly
481 The _WORKING attribute should be set to False for broken IEs
482 in order to warn the users and skip the tests.
487 _x_forwarded_for_ip
= None
489 _GEO_COUNTRIES
= None
490 _GEO_IP_BLOCKS
= None
492 _NETRC_MACHINE
= None
498 def _login_hint(self
, method
=NO_DEFAULT
, netrc
=None):
499 password_hint
= f
'--username and --password, or --netrc ({netrc or self._NETRC_MACHINE}) to provide account credentials'
502 'any': f
'Use --cookies, --cookies-from-browser, {password_hint}',
503 'password': f
'Use {password_hint}',
505 'Use --cookies-from-browser or --cookies for the authentication. '
506 'See https://github.com/ytdl-org/youtube-dl#how-do-i-pass-cookies-to-youtube-dl for how to manually pass cookies'),
507 }[method
if method
is not NO_DEFAULT
else 'any' if self
.supports_login() else 'cookies']
509 def __init__(self
, downloader
=None):
510 """Constructor. Receives an optional downloader (a YoutubeDL instance).
511 If a downloader is not passed during initialization,
512 it must be set using "set_downloader()" before "extract()" is called"""
514 self
._x
_forwarded
_for
_ip
= None
515 self
._printed
_messages
= set()
516 self
.set_downloader(downloader
)
519 def _match_valid_url(cls
, url
):
520 if cls
._VALID
_URL
is False:
522 # This does not use has/getattr intentionally - we want to know whether
523 # we have cached the regexp for *this* class, whereas getattr would also
524 # match the superclass
525 if '_VALID_URL_RE' not in cls
.__dict
__:
526 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
527 return cls
._VALID
_URL
_RE
.match(url
)
530 def suitable(cls
, url
):
531 """Receives a URL and returns True if suitable for this IE."""
532 # This function must import everything it needs (except other extractors),
533 # so that lazy_extractors works correctly
534 return cls
._match
_valid
_url
(url
) is not None
537 def _match_id(cls
, url
):
538 return cls
._match
_valid
_url
(url
).group('id')
541 def get_temp_id(cls
, url
):
543 return cls
._match
_id
(url
)
544 except (IndexError, AttributeError):
549 """Getter method for _WORKING."""
553 def supports_login(cls
):
554 return bool(cls
._NETRC
_MACHINE
)
556 def initialize(self
):
557 """Initializes an instance (authentication, etc)."""
558 self
._printed
_messages
= set()
559 self
._initialize
_geo
_bypass
({
560 'countries': self
._GEO
_COUNTRIES
,
561 'ip_blocks': self
._GEO
_IP
_BLOCKS
,
564 self
._initialize
_pre
_login
()
565 if self
.supports_login():
566 username
, password
= self
._get
_login
_info
()
568 self
._perform
_login
(username
, password
)
569 elif self
.get_param('username') and False not in (self
.IE_DESC
, self
._NETRC
_MACHINE
):
570 self
.report_warning(f
'Login with password is not supported for this website. {self._login_hint("cookies")}')
571 self
._real
_initialize
()
574 def _initialize_geo_bypass(self
, geo_bypass_context
):
576 Initialize geo restriction bypass mechanism.
578 This method is used to initialize geo bypass mechanism based on faking
579 X-Forwarded-For HTTP header. A random country from provided country list
580 is selected and a random IP belonging to this country is generated. This
581 IP will be passed as X-Forwarded-For HTTP header in all subsequent
584 This method will be used for initial geo bypass mechanism initialization
585 during the instance initialization with _GEO_COUNTRIES and
588 You may also manually call it from extractor's code if geo bypass
589 information is not available beforehand (e.g. obtained during
590 extraction) or due to some other reason. In this case you should pass
591 this information in geo bypass context passed as first argument. It may
592 contain following fields:
594 countries: List of geo unrestricted countries (similar
596 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
597 (similar to _GEO_IP_BLOCKS)
600 if not self
._x
_forwarded
_for
_ip
:
602 # Geo bypass mechanism is explicitly disabled by user
603 if not self
.get_param('geo_bypass', True):
606 if not geo_bypass_context
:
607 geo_bypass_context
= {}
609 # Backward compatibility: previously _initialize_geo_bypass
610 # expected a list of countries, some 3rd party code may still use
612 if isinstance(geo_bypass_context
, (list, tuple)):
613 geo_bypass_context
= {
614 'countries': geo_bypass_context
,
617 # The whole point of geo bypass mechanism is to fake IP
618 # as X-Forwarded-For HTTP header based on some IP block or
621 # Path 1: bypassing based on IP block in CIDR notation
623 # Explicit IP block specified by user, use it right away
624 # regardless of whether extractor is geo bypassable or not
625 ip_block
= self
.get_param('geo_bypass_ip_block', None)
627 # Otherwise use random IP block from geo bypass context but only
628 # if extractor is known as geo bypassable
630 ip_blocks
= geo_bypass_context
.get('ip_blocks')
631 if self
._GEO
_BYPASS
and ip_blocks
:
632 ip_block
= random
.choice(ip_blocks
)
635 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(ip_block
)
636 self
.write_debug(f
'Using fake IP {self._x_forwarded_for_ip} as X-Forwarded-For')
639 # Path 2: bypassing based on country code
641 # Explicit country code specified by user, use it right away
642 # regardless of whether extractor is geo bypassable or not
643 country
= self
.get_param('geo_bypass_country', None)
645 # Otherwise use random country code from geo bypass context but
646 # only if extractor is known as geo bypassable
648 countries
= geo_bypass_context
.get('countries')
649 if self
._GEO
_BYPASS
and countries
:
650 country
= random
.choice(countries
)
653 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country
)
654 self
._downloader
.write_debug(
655 f
'Using fake IP {self._x_forwarded_for_ip} ({country.upper()}) as X-Forwarded-For')
657 def extract(self
, url
):
658 """Extracts URL information and returns it in list of dicts."""
663 self
.write_debug('Extracting URL: %s' % url
)
664 ie_result
= self
._real
_extract
(url
)
665 if ie_result
is None:
667 if self
._x
_forwarded
_for
_ip
:
668 ie_result
['__x_forwarded_for_ip'] = self
._x
_forwarded
_for
_ip
669 subtitles
= ie_result
.get('subtitles') or {}
670 if 'no-live-chat' in self
.get_param('compat_opts'):
671 for lang
in ('live_chat', 'comments', 'danmaku'):
672 subtitles
.pop(lang
, None)
674 except GeoRestrictedError
as e
:
675 if self
.__maybe
_fake
_ip
_and
_retry
(e
.countries
):
678 except UnsupportedError
:
680 except ExtractorError
as e
:
682 'video_id': e
.video_id
or self
.get_temp_id(url
),
684 'tb': e
.traceback
or sys
.exc_info()[2],
685 'expected': e
.expected
,
688 if hasattr(e
, 'countries'):
689 kwargs
['countries'] = e
.countries
690 raise type(e
)(e
.orig_msg
, **kwargs
)
691 except http
.client
.IncompleteRead
as e
:
692 raise ExtractorError('A network error has occurred.', cause
=e
, expected
=True, video_id
=self
.get_temp_id(url
))
693 except (KeyError, StopIteration) as e
:
694 raise ExtractorError('An extractor error has occurred.', cause
=e
, video_id
=self
.get_temp_id(url
))
696 def __maybe_fake_ip_and_retry(self
, countries
):
697 if (not self
.get_param('geo_bypass_country', None)
699 and self
.get_param('geo_bypass', True)
700 and not self
._x
_forwarded
_for
_ip
702 country_code
= random
.choice(countries
)
703 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country_code
)
704 if self
._x
_forwarded
_for
_ip
:
706 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
707 % (self
._x
_forwarded
_for
_ip
, country_code
.upper()))
711 def set_downloader(self
, downloader
):
712 """Sets a YoutubeDL instance as the downloader for this IE."""
713 self
._downloader
= downloader
717 return self
._downloader
.cache
721 return self
._downloader
.cookiejar
723 def _initialize_pre_login(self
):
724 """ Initialization before login. Redefine in subclasses."""
727 def _perform_login(self
, username
, password
):
728 """ Login with username and password. Redefine in subclasses."""
731 def _real_initialize(self
):
732 """Real initialization process. Redefine in subclasses."""
735 def _real_extract(self
, url
):
736 """Real extraction process. Redefine in subclasses."""
737 raise NotImplementedError('This method must be implemented by subclasses')
741 """A string for getting the InfoExtractor with get_info_extractor"""
742 return cls
.__name
__[:-2]
746 return cls
.__name
__[:-2]
749 def __can_accept_status_code(err
, expected_status
):
750 assert isinstance(err
, urllib
.error
.HTTPError
)
751 if expected_status
is None:
753 elif callable(expected_status
):
754 return expected_status(err
.code
) is True
756 return err
.code
in variadic(expected_status
)
758 def _create_request(self
, url_or_request
, data
=None, headers
=None, query
=None):
759 if isinstance(url_or_request
, urllib
.request
.Request
):
760 return update_Request(url_or_request
, data
=data
, headers
=headers
, query
=query
)
762 url_or_request
= update_url_query(url_or_request
, query
)
763 return sanitized_Request(url_or_request
, data
, headers
or {})
765 def _request_webpage(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, data
=None, headers
=None, query
=None, expected_status
=None):
767 Return the response handle.
769 See _download_webpage docstring for arguments specification.
771 if not self
._downloader
._first
_webpage
_request
:
772 sleep_interval
= self
.get_param('sleep_interval_requests') or 0
773 if sleep_interval
> 0:
774 self
.to_screen('Sleeping %s seconds ...' % sleep_interval
)
775 time
.sleep(sleep_interval
)
777 self
._downloader
._first
_webpage
_request
= False
780 self
.report_download_webpage(video_id
)
781 elif note
is not False:
783 self
.to_screen(str(note
))
785 self
.to_screen(f
'{video_id}: {note}')
787 # Some sites check X-Forwarded-For HTTP header in order to figure out
788 # the origin of the client behind proxy. This allows bypassing geo
789 # restriction by faking this header's value to IP that belongs to some
790 # geo unrestricted country. We will do so once we encounter any
791 # geo restriction error.
792 if self
._x
_forwarded
_for
_ip
:
793 headers
= (headers
or {}).copy()
794 headers
.setdefault('X-Forwarded-For', self
._x
_forwarded
_for
_ip
)
797 return self
._downloader
.urlopen(self
._create
_request
(url_or_request
, data
, headers
, query
))
798 except network_exceptions
as err
:
799 if isinstance(err
, urllib
.error
.HTTPError
):
800 if self
.__can
_accept
_status
_code
(err
, expected_status
):
801 # Retain reference to error to prevent file object from
802 # being closed before it can be read. Works around the
803 # effects of <https://bugs.python.org/issue15002>
804 # introduced in Python 3.4.1.
811 errnote
= 'Unable to download webpage'
813 errmsg
= f
'{errnote}: {error_to_compat_str(err)}'
815 raise ExtractorError(errmsg
, cause
=err
)
817 self
.report_warning(errmsg
)
820 def _download_webpage_handle(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True,
821 encoding
=None, data
=None, headers
={}, query={}
, expected_status
=None):
823 Return a tuple (page content as string, URL handle).
826 url_or_request -- plain text URL as a string or
827 a urllib.request.Request object
828 video_id -- Video/playlist/item identifier (string)
831 note -- note printed before downloading (string)
832 errnote -- note printed in case of an error (string)
833 fatal -- flag denoting whether error should be considered fatal,
834 i.e. whether it should cause ExtractionError to be raised,
835 otherwise a warning will be reported and extraction continued
836 encoding -- encoding for a page content decoding, guessed automatically
837 when not explicitly specified
838 data -- POST data (bytes)
839 headers -- HTTP headers (dict)
840 query -- URL query (dict)
841 expected_status -- allows to accept failed HTTP requests (non 2xx
842 status code) by explicitly specifying a set of accepted status
843 codes. Can be any of the following entities:
844 - an integer type specifying an exact failed status code to
846 - a list or a tuple of integer types specifying a list of
847 failed status codes to accept
848 - a callable accepting an actual failed status code and
849 returning True if it should be accepted
850 Note that this argument does not affect success status codes (2xx)
851 which are always accepted.
854 # Strip hashes from the URL (#1038)
855 if isinstance(url_or_request
, str):
856 url_or_request
= url_or_request
.partition('#')[0]
858 urlh
= self
._request
_webpage
(url_or_request
, video_id
, note
, errnote
, fatal
, data
=data
, headers
=headers
, query
=query
, expected_status
=expected_status
)
862 content
= self
._webpage
_read
_content
(urlh
, url_or_request
, video_id
, note
, errnote
, fatal
, encoding
=encoding
)
863 return (content
, urlh
)
866 def _guess_encoding_from_content(content_type
, webpage_bytes
):
867 m
= re
.match(r
'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type
)
869 encoding
= m
.group(1)
871 m
= re
.search(br
'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
872 webpage_bytes[:1024])
874 encoding = m.group(1).decode('ascii')
875 elif webpage_bytes.startswith(b'\xff\xfe'):
882 def __check_blocked(self, content):
883 first_block = content[:512]
884 if ('<title>Access to this site is blocked</title>' in content
885 and 'Websense' in first_block):
886 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
887 blocked_iframe = self._html_search_regex(
888 r'<iframe src="([^
"]+)"', content,
889 'Websense information URL
', default=None)
891 msg += ' Visit
%s for more details
' % blocked_iframe
892 raise ExtractorError(msg, expected=True)
893 if '<title
>The URL you requested has been blocked
</title
>' in first_block:
895 'Access to this webpage has been blocked by Indian censorship
. '
896 'Use a VPN
or proxy
server (with --proxy
) to route around it
.')
897 block_msg = self._html_search_regex(
898 r'</h1
><p
>(.*?
)</p
>',
899 content, 'block message
', default=None)
901 msg += ' (Message
: "%s")' % block_msg.replace('\n', ' ')
902 raise ExtractorError(msg, expected=True)
903 if ('<title
>TTK
:: Доступ к ресурсу ограничен
</title
>' in content
904 and 'blocklist
.rkn
.gov
.ru
' in content):
905 raise ExtractorError(
906 'Access to this webpage has been blocked by decision of the Russian government
. '
907 'Visit http
://blocklist
.rkn
.gov
.ru
/ for a block reason
.',
910 def _request_dump_filename(self, url, video_id):
911 basen = f'{video_id}_{url}
'
912 trim_length = self.get_param('trim_file_name
') or 240
913 if len(basen) > trim_length:
914 h = '___
' + hashlib.md5(basen.encode('utf
-8')).hexdigest()
915 basen = basen[:trim_length - len(h)] + h
916 filename = sanitize_filename(f'{basen}
.dump
', restricted=True)
917 # Working around MAX_PATH limitation on Windows (see
918 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
919 if compat_os_name == 'nt
':
920 absfilepath = os.path.abspath(filename)
921 if len(absfilepath) > 259:
922 filename = fR'\\?\{absfilepath}
'
925 def __decode_webpage(self, webpage_bytes, encoding, headers):
927 encoding = self._guess_encoding_from_content(headers.get('Content
-Type
', ''), webpage_bytes)
929 return webpage_bytes.decode(encoding, 'replace
')
931 return webpage_bytes.decode('utf
-8', 'replace
')
933 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
934 webpage_bytes = urlh.read()
935 if prefix is not None:
936 webpage_bytes = prefix + webpage_bytes
937 if self.get_param('dump_intermediate_pages
', False):
938 self.to_screen('Dumping request to
' + urlh.geturl())
939 dump = base64.b64encode(webpage_bytes).decode('ascii
')
940 self._downloader.to_screen(dump)
941 if self.get_param('write_pages
'):
942 filename = self._request_dump_filename(urlh.geturl(), video_id)
943 self.to_screen(f'Saving request to {filename}
')
944 with open(filename, 'wb
') as outf:
945 outf.write(webpage_bytes)
947 content = self.__decode_webpage(webpage_bytes, encoding, urlh.headers)
948 self.__check_blocked(content)
952 def __print_error(self, errnote, fatal, video_id, err):
954 raise ExtractorError(f'{video_id}
: {errnote}
', cause=err)
956 self.report_warning(f'{video_id}
: {errnote}
: {err}
')
958 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True, errnote=None):
960 xml_string = transform_source(xml_string)
962 return compat_etree_fromstring(xml_string.encode('utf
-8'))
963 except xml.etree.ElementTree.ParseError as ve:
964 self.__print_error('Failed to parse XML
' if errnote is None else errnote, fatal, video_id, ve)
966 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True, errnote=None, **parser_kwargs):
969 json_string, cls=LenientJSONDecoder, strict=False, transform_source=transform_source, **parser_kwargs)
970 except ValueError as ve:
971 self.__print_error('Failed to parse JSON
' if errnote is None else errnote, fatal, video_id, ve)
973 def _parse_socket_response_as_json(self, data, *args, **kwargs):
974 return self._parse_json(data[data.find('{'):data.rfind('}
') + 1], *args, **kwargs)
976 def __create_download_methods(name, parser, note, errnote, return_value):
978 def parse(ie, content, *args, errnote=errnote, **kwargs):
982 kwargs['errnote
'] = errnote
983 # parser is fetched by name so subclasses can override it
984 return getattr(ie, parser)(content, *args, **kwargs)
986 def download_handle(self, url_or_request, video_id, note=note, errnote=errnote, transform_source=None,
987 fatal=True, encoding=None, data=None, headers={}, query={}, expected_status=None):
988 res = self._download_webpage_handle(
989 url_or_request, video_id, note=note, errnote=errnote, fatal=fatal, encoding=encoding,
990 data=data, headers=headers, query=query, expected_status=expected_status)
994 return parse(self, content, video_id, transform_source=transform_source, fatal=fatal, errnote=errnote), urlh
996 def download_content(self, url_or_request, video_id, note=note, errnote=errnote, transform_source=None,
997 fatal=True, encoding=None, data=None, headers={}, query={}, expected_status=None):
998 if self.get_param('load_pages
'):
999 url_or_request = self._create_request(url_or_request, data, headers, query)
1000 filename = self._request_dump_filename(url_or_request.full_url, video_id)
1001 self.to_screen(f'Loading request
from {filename}
')
1003 with open(filename, 'rb
') as dumpf:
1004 webpage_bytes = dumpf.read()
1005 except OSError as e:
1006 self.report_warning(f'Unable to load request
from disk
: {e}
')
1008 content = self.__decode_webpage(webpage_bytes, encoding, url_or_request.headers)
1009 return parse(self, content, video_id, transform_source=transform_source, fatal=fatal, errnote=errnote)
1013 'transform_source
': transform_source,
1015 'encoding
': encoding,
1019 'expected_status
': expected_status,
1022 kwargs.pop('transform_source
')
1023 # The method is fetched by name so subclasses can override _download_..._handle
1024 res = getattr(self, download_handle.__name__)(url_or_request, video_id, **kwargs)
1025 return res if res is False else res[0]
1027 def impersonate(func, name, return_value):
1028 func.__name__, func.__qualname__ = name, f'InfoExtractor
.{name}
'
1030 @param transform_source Apply this transformation before parsing
1031 @returns {return_value}
1033 See _download_webpage_handle docstring for other arguments specification
1036 impersonate(download_handle, f'_download_{name}_handle
', f'({return_value}
, URL handle
)')
1037 impersonate(download_content, f'_download_{name}
', f'{return_value}
')
1038 return download_handle, download_content
1040 _download_xml_handle, _download_xml = __create_download_methods(
1041 'xml
', '_parse_xml
', 'Downloading XML
', 'Unable to download XML
', 'xml
as an xml
.etree
.ElementTree
.Element
')
1042 _download_json_handle, _download_json = __create_download_methods(
1043 'json
', '_parse_json
', 'Downloading JSON metadata
', 'Unable to download JSON metadata
', 'JSON
object as a
dict')
1044 _download_socket_json_handle, _download_socket_json = __create_download_methods(
1045 'socket_json
', '_parse_socket_response_as_json
', 'Polling socket
', 'Unable to poll socket
', 'JSON
object as a
dict')
1046 __download_webpage = __create_download_methods('webpage
', None, None, None, 'data of the page
as a string
')[1]
1048 def _download_webpage(
1049 self, url_or_request, video_id, note=None, errnote=None,
1050 fatal=True, tries=1, timeout=NO_DEFAULT, *args, **kwargs):
1052 Return the data of the page as a string.
1055 tries -- number of tries
1056 timeout -- sleep interval between tries
1058 See _download_webpage_handle docstring for other arguments specification.
1061 R''' # NB: These are unused; should they be deprecated?
1063 self._downloader.deprecation_warning('tries argument
is deprecated
in InfoExtractor
._download
_webpage
')
1064 if timeout is NO_DEFAULT:
1067 self._downloader.deprecation_warning('timeout argument
is deprecated
in InfoExtractor
._download
_webpage
')
1073 return self.__download_webpage(url_or_request, video_id, note, errnote, None, fatal, *args, **kwargs)
1074 except http.client.IncompleteRead as e:
1076 if try_count >= tries:
1078 self._sleep(timeout, video_id)
1080 def report_warning(self, msg, video_id=None, *args, only_once=False, **kwargs):
1081 idstr = format_field(video_id, None, '%s: ')
1082 msg = f'[{self.IE_NAME}
] {idstr}{msg}
'
1084 if f'WARNING
: {msg}
' in self._printed_messages:
1086 self._printed_messages.add(f'WARNING
: {msg}
')
1087 self._downloader.report_warning(msg, *args, **kwargs)
1089 def to_screen(self, msg, *args, **kwargs):
1090 """Print msg to screen, prefixing it with '[ie_name
]'"""
1091 self._downloader.to_screen(f'[{self.IE_NAME}
] {msg}
', *args, **kwargs)
1093 def write_debug(self, msg, *args, **kwargs):
1094 self._downloader.write_debug(f'[{self.IE_NAME}
] {msg}
', *args, **kwargs)
1096 def get_param(self, name, default=None, *args, **kwargs):
1097 if self._downloader:
1098 return self._downloader.params.get(name, default, *args, **kwargs)
1101 def report_drm(self, video_id, partial=False):
1102 self.raise_no_formats('This video
is DRM protected
', expected=True, video_id=video_id)
1104 def report_extraction(self, id_or_name):
1105 """Report information extraction."""
1106 self.to_screen('%s: Extracting information
' % id_or_name)
1108 def report_download_webpage(self, video_id):
1109 """Report webpage download."""
1110 self.to_screen('%s: Downloading webpage
' % video_id)
1112 def report_age_confirmation(self):
1113 """Report attempt to confirm age."""
1114 self.to_screen('Confirming age
')
1116 def report_login(self):
1117 """Report attempt to log in."""
1118 self.to_screen('Logging
in')
1120 def raise_login_required(
1121 self, msg='This video
is only available
for registered users
',
1122 metadata_available=False, method=NO_DEFAULT):
1123 if metadata_available and (
1124 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1125 self.report_warning(msg)
1127 msg += format_field(self._login_hint(method), None, '. %s')
1128 raise ExtractorError(msg, expected=True)
1130 def raise_geo_restricted(
1131 self, msg='This video
is not available
from your location due to geo restriction
',
1132 countries=None, metadata_available=False):
1133 if metadata_available and (
1134 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1135 self.report_warning(msg)
1137 raise GeoRestrictedError(msg, countries=countries)
1139 def raise_no_formats(self, msg, expected=False, video_id=None):
1141 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1142 self.report_warning(msg, video_id)
1143 elif isinstance(msg, ExtractorError):
1146 raise ExtractorError(msg, expected=expected, video_id=video_id)
1148 # Methods for following #608
1150 def url_result(url, ie=None, video_id=None, video_title=None, *, url_transparent=False, **kwargs):
1151 """Returns a URL that points to a page that should be processed"""
1153 kwargs['ie_key
'] = ie if isinstance(ie, str) else ie.ie_key()
1154 if video_id is not None:
1155 kwargs['id'] = video_id
1156 if video_title is not None:
1157 kwargs['title
'] = video_title
1160 '_type
': 'url_transparent
' if url_transparent else 'url
',
1165 def playlist_from_matches(cls, matches, playlist_id=None, playlist_title=None,
1166 getter=IDENTITY, ie=None, video_kwargs=None, **kwargs):
1167 return cls.playlist_result(
1168 (cls.url_result(m, ie, **(video_kwargs or {})) for m in orderedSet(map(getter, matches), lazy=True)),
1169 playlist_id, playlist_title, **kwargs)
1172 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None, *, multi_video=False, **kwargs):
1173 """Returns a playlist"""
1175 kwargs['id'] = playlist_id
1177 kwargs['title
'] = playlist_title
1178 if playlist_description is not None:
1179 kwargs['description
'] = playlist_description
1182 '_type
': 'multi_video
' if multi_video else 'playlist
',
1186 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1188 Perform a regex search on the given string, using a single or a list of
1189 patterns returning the first matching group.
1190 In case of failure return a default value or raise a WARNING or a
1191 RegexNotFoundError, depending on fatal, specifying the field name.
1195 elif isinstance(pattern, (str, re.Pattern)):
1196 mobj = re.search(pattern, string, flags)
1199 mobj = re.search(p, string, flags)
1203 _name = self._downloader._format_err(name, self._downloader.Styles.EMPHASIS)
1207 # return the first matching group
1208 return next(g for g in mobj.groups() if g is not None)
1209 elif isinstance(group, (list, tuple)):
1210 return tuple(mobj.group(g) for g in group)
1212 return mobj.group(group)
1213 elif default is not NO_DEFAULT:
1216 raise RegexNotFoundError('Unable to extract
%s' % _name)
1218 self.report_warning('unable to extract
%s' % _name + bug_reports_message())
1221 def _search_json(self, start_pattern, string, name, video_id, *, end_pattern='',
1222 contains_pattern='(?s
:.+)', fatal=True, default=NO_DEFAULT, **kwargs):
1223 """Searches string for the JSON object specified by start_pattern"""
1224 # NB: end_pattern is only used to reduce the size of the initial match
1225 if default is NO_DEFAULT:
1226 default, has_default = {}, False
1228 fatal, has_default = False, True
1230 json_string = self._search_regex(
1231 rf'{start_pattern}\s
*(?P
<json
>{{\s*{contains_pattern}\s
*}})\s
*{end_pattern}
',
1232 string, name, group='json
', fatal=fatal, default=None if has_default else NO_DEFAULT)
1236 _name = self._downloader._format_err(name, self._downloader.Styles.EMPHASIS)
1238 return self._parse_json(json_string, video_id, ignore_extra=True, **kwargs)
1239 except ExtractorError as e:
1241 raise ExtractorError(
1242 f'Unable to extract {_name}
- Failed to parse JSON
', cause=e.cause, video_id=video_id)
1243 elif not has_default:
1244 self.report_warning(
1245 f'Unable to extract {_name}
- Failed to parse JSON
: {e}
', video_id=video_id)
1248 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1250 Like _search_regex, but strips HTML tags and unescapes entities.
1252 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1254 return clean_html(res).strip()
1258 def _get_netrc_login_info(self, netrc_machine=None):
1261 netrc_machine = netrc_machine or self._NETRC_MACHINE
1263 if self.get_param('usenetrc
', False):
1265 netrc_file = compat_expanduser(self.get_param('netrc_location
') or '~
')
1266 if os.path.isdir(netrc_file):
1267 netrc_file = os.path.join(netrc_file, '.netrc
')
1268 info = netrc.netrc(file=netrc_file).authenticators(netrc_machine)
1269 if info is not None:
1273 raise netrc.NetrcParseError(
1274 'No authenticators
for %s' % netrc_machine)
1275 except (OSError, netrc.NetrcParseError) as err:
1276 self.report_warning(
1277 'parsing
.netrc
: %s' % error_to_compat_str(err))
1279 return username, password
1281 def _get_login_info(self, username_option='username
', password_option='password
', netrc_machine=None):
1283 Get the login info as (username, password)
1284 First look for the manually specified credentials using username_option
1285 and password_option as keys in params dictionary. If no such credentials
1286 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1288 If there's no info available
, return (None, None)
1291 # Attempt to use provided username and password or .netrc data
1292 username = self.get_param(username_option)
1293 if username is not None:
1294 password = self.get_param(password_option)
1296 username, password = self._get_netrc_login_info(netrc_machine)
1298 return username, password
1300 def _get_tfa_info(self, note='two-factor verification code'):
1302 Get the two
-factor authentication info
1303 TODO
- asking the user will be required
for sms
/phone verify
1304 currently just uses the command line option
1305 If there
's no info available, return None
1308 tfa = self.get_param('twofactor
')
1312 return getpass.getpass('Type
%s and press
[Return
]: ' % note)
1314 # Helper functions for extracting OpenGraph info
1316 def _og_regexes(prop):
1317 content_re = r'content
=(?
:"([^"]+?
)"|\'([^\']+?)\'|\s*([^\s"\'=<>`
]+?
))'
1318 property_re = (r'(?
:name|
property)=(?
:\'og
%(sep)s%(prop)s\'|
"og%(sep)s%(prop)s"|\s
*og
%(sep)s%(prop)s\b)'
1319 % {'prop': re.escape(prop), 'sep': '(?::|[:-])'})
1320 template = r'<meta
[^
>]+?
%s[^
>]+?
%s'
1322 template % (property_re, content_re),
1323 template % (content_re, property_re),
1327 def _meta_regex(prop):
1328 return r'''(?isx)<meta
1329 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1330 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1332 def _og_search_property(self, prop, html, name=None, **kargs):
1333 prop = variadic(prop)
1335 name = 'OpenGraph
%s' % prop[0]
1338 og_regexes.extend(self._og_regexes(p))
1339 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1342 return unescapeHTML(escaped)
1344 def _og_search_thumbnail(self, html, **kargs):
1345 return self._og_search_property('image
', html, 'thumbnail URL
', fatal=False, **kargs)
1347 def _og_search_description(self, html, **kargs):
1348 return self._og_search_property('description
', html, fatal=False, **kargs)
1350 def _og_search_title(self, html, *, fatal=False, **kargs):
1351 return self._og_search_property('title
', html, fatal=fatal, **kargs)
1353 def _og_search_video_url(self, html, name='video url
', secure=True, **kargs):
1354 regexes = self._og_regexes('video
') + self._og_regexes('video
:url
')
1356 regexes = self._og_regexes('video
:secure_url
') + regexes
1357 return self._html_search_regex(regexes, html, name, **kargs)
1359 def _og_search_url(self, html, **kargs):
1360 return self._og_search_property('url
', html, **kargs)
1362 def _html_extract_title(self, html, name='title
', *, fatal=False, **kwargs):
1363 return self._html_search_regex(r'(?s
)<title
\b[^
>]*>([^
<]+)</title
>', html, name, fatal=fatal, **kwargs)
1365 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1366 name = variadic(name)
1367 if display_name is None:
1368 display_name = name[0]
1369 return self._html_search_regex(
1370 [self._meta_regex(n) for n in name],
1371 html, display_name, fatal=fatal, group='content
', **kwargs)
1373 def _dc_search_uploader(self, html):
1374 return self._html_search_meta('dc
.creator
', html, 'uploader
')
1377 def _rta_search(html):
1378 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1379 if re.search(r'(?ix
)<meta\s
+name
="rating"\s
+'
1380 r' content
="RTA-5042-1996-1400-1577-RTA"',
1384 # And then there are the jokers who advertise that they use RTA, but actually don't
.
1385 AGE_LIMIT_MARKERS
= [
1386 r
'Proudly Labeled <a href="http://www\.rtalabel\.org/" title="Restricted to Adults">RTA</a>',
1388 if any(re
.search(marker
, html
) for marker
in AGE_LIMIT_MARKERS
):
1392 def _media_rating_search(self
, html
):
1393 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1394 rating
= self
._html
_search
_meta
('rating', html
)
1406 return RATING_TABLE
.get(rating
.lower())
1408 def _family_friendly_search(self
, html
):
1409 # See http://schema.org/VideoObject
1410 family_friendly
= self
._html
_search
_meta
(
1411 'isFamilyFriendly', html
, default
=None)
1413 if not family_friendly
:
1422 return RATING_TABLE
.get(family_friendly
.lower())
1424 def _twitter_search_player(self
, html
):
1425 return self
._html
_search
_meta
('twitter:player', html
,
1426 'twitter card player')
1428 def _yield_json_ld(self
, html
, video_id
, *, fatal
=True, default
=NO_DEFAULT
):
1429 """Yield all json ld objects in the html"""
1430 if default
is not NO_DEFAULT
:
1432 for mobj
in re
.finditer(JSON_LD_RE
, html
):
1433 json_ld_item
= self
._parse
_json
(mobj
.group('json_ld'), video_id
, fatal
=fatal
)
1434 for json_ld
in variadic(json_ld_item
):
1435 if isinstance(json_ld
, dict):
1438 def _search_json_ld(self
, html
, video_id
, expected_type
=None, *, fatal
=True, default
=NO_DEFAULT
):
1439 """Search for a video in any json ld in the html"""
1440 if default
is not NO_DEFAULT
:
1442 info
= self
._json
_ld
(
1443 list(self
._yield
_json
_ld
(html
, video_id
, fatal
=fatal
, default
=default
)),
1444 video_id
, fatal
=fatal
, expected_type
=expected_type
)
1447 if default
is not NO_DEFAULT
:
1450 raise RegexNotFoundError('Unable to extract JSON-LD')
1452 self
.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1455 def _json_ld(self
, json_ld
, video_id
, fatal
=True, expected_type
=None):
1456 if isinstance(json_ld
, str):
1457 json_ld
= self
._parse
_json
(json_ld
, video_id
, fatal
=fatal
)
1461 if not isinstance(json_ld
, (list, tuple, dict)):
1463 if isinstance(json_ld
, dict):
1466 INTERACTION_TYPE_MAP
= {
1467 'CommentAction': 'comment',
1468 'AgreeAction': 'like',
1469 'DisagreeAction': 'dislike',
1470 'LikeAction': 'like',
1471 'DislikeAction': 'dislike',
1472 'ListenAction': 'view',
1473 'WatchAction': 'view',
1474 'ViewAction': 'view',
1477 def is_type(e
, *expected_types
):
1478 type = variadic(traverse_obj(e
, '@type'))
1479 return any(x
in type for x
in expected_types
)
1481 def extract_interaction_type(e
):
1482 interaction_type
= e
.get('interactionType')
1483 if isinstance(interaction_type
, dict):
1484 interaction_type
= interaction_type
.get('@type')
1485 return str_or_none(interaction_type
)
1487 def extract_interaction_statistic(e
):
1488 interaction_statistic
= e
.get('interactionStatistic')
1489 if isinstance(interaction_statistic
, dict):
1490 interaction_statistic
= [interaction_statistic
]
1491 if not isinstance(interaction_statistic
, list):
1493 for is_e
in interaction_statistic
:
1494 if not is_type(is_e
, 'InteractionCounter'):
1496 interaction_type
= extract_interaction_type(is_e
)
1497 if not interaction_type
:
1499 # For interaction count some sites provide string instead of
1500 # an integer (as per spec) with non digit characters (e.g. ",")
1501 # so extracting count with more relaxed str_to_int
1502 interaction_count
= str_to_int(is_e
.get('userInteractionCount'))
1503 if interaction_count
is None:
1505 count_kind
= INTERACTION_TYPE_MAP
.get(interaction_type
.split('/')[-1])
1508 count_key
= '%s_count' % count_kind
1509 if info
.get(count_key
) is not None:
1511 info
[count_key
] = interaction_count
1513 def extract_chapter_information(e
):
1515 'title': part
.get('name'),
1516 'start_time': part
.get('startOffset'),
1517 'end_time': part
.get('endOffset'),
1518 } for part
in variadic(e
.get('hasPart') or []) if part
.get('@type') == 'Clip']
1519 for idx
, (last_c
, current_c
, next_c
) in enumerate(zip(
1520 [{'end_time': 0}
] + chapters
, chapters
, chapters
[1:])):
1521 current_c
['end_time'] = current_c
['end_time'] or next_c
['start_time']
1522 current_c
['start_time'] = current_c
['start_time'] or last_c
['end_time']
1523 if None in current_c
.values():
1524 self
.report_warning(f
'Chapter {idx} contains broken data. Not extracting chapters')
1527 chapters
[-1]['end_time'] = chapters
[-1]['end_time'] or info
['duration']
1528 info
['chapters'] = chapters
1530 def extract_video_object(e
):
1531 assert is_type(e
, 'VideoObject')
1532 author
= e
.get('author')
1534 'url': url_or_none(e
.get('contentUrl')),
1535 'title': unescapeHTML(e
.get('name')),
1536 'description': unescapeHTML(e
.get('description')),
1537 'thumbnails': [{'url': unescapeHTML(url)}
1538 for url
in variadic(traverse_obj(e
, 'thumbnailUrl', 'thumbnailURL'))
1539 if url_or_none(url
)],
1540 'duration': parse_duration(e
.get('duration')),
1541 'timestamp': unified_timestamp(e
.get('uploadDate')),
1542 # author can be an instance of 'Organization' or 'Person' types.
1543 # both types can have 'name' property(inherited from 'Thing' type). [1]
1544 # however some websites are using 'Text' type instead.
1545 # 1. https://schema.org/VideoObject
1546 'uploader': author
.get('name') if isinstance(author
, dict) else author
if isinstance(author
, str) else None,
1547 'filesize': int_or_none(float_or_none(e
.get('contentSize'))),
1548 'tbr': int_or_none(e
.get('bitrate')),
1549 'width': int_or_none(e
.get('width')),
1550 'height': int_or_none(e
.get('height')),
1551 'view_count': int_or_none(e
.get('interactionCount')),
1553 extract_interaction_statistic(e
)
1554 extract_chapter_information(e
)
1556 def traverse_json_ld(json_ld
, at_top_level
=True):
1558 if at_top_level
and '@context' not in e
:
1560 if at_top_level
and set(e
.keys()) == {'@context', '@graph'}
:
1561 traverse_json_ld(variadic(e
['@graph'], allowed_types
=(dict,)), at_top_level
=False)
1563 if expected_type
is not None and not is_type(e
, expected_type
):
1565 rating
= traverse_obj(e
, ('aggregateRating', 'ratingValue'), expected_type
=float_or_none
)
1566 if rating
is not None:
1567 info
['average_rating'] = rating
1568 if is_type(e
, 'TVEpisode', 'Episode'):
1569 episode_name
= unescapeHTML(e
.get('name'))
1571 'episode': episode_name
,
1572 'episode_number': int_or_none(e
.get('episodeNumber')),
1573 'description': unescapeHTML(e
.get('description')),
1575 if not info
.get('title') and episode_name
:
1576 info
['title'] = episode_name
1577 part_of_season
= e
.get('partOfSeason')
1578 if is_type(part_of_season
, 'TVSeason', 'Season', 'CreativeWorkSeason'):
1580 'season': unescapeHTML(part_of_season
.get('name')),
1581 'season_number': int_or_none(part_of_season
.get('seasonNumber')),
1583 part_of_series
= e
.get('partOfSeries') or e
.get('partOfTVSeries')
1584 if is_type(part_of_series
, 'TVSeries', 'Series', 'CreativeWorkSeries'):
1585 info
['series'] = unescapeHTML(part_of_series
.get('name'))
1586 elif is_type(e
, 'Movie'):
1588 'title': unescapeHTML(e
.get('name')),
1589 'description': unescapeHTML(e
.get('description')),
1590 'duration': parse_duration(e
.get('duration')),
1591 'timestamp': unified_timestamp(e
.get('dateCreated')),
1593 elif is_type(e
, 'Article', 'NewsArticle'):
1595 'timestamp': parse_iso8601(e
.get('datePublished')),
1596 'title': unescapeHTML(e
.get('headline')),
1597 'description': unescapeHTML(e
.get('articleBody') or e
.get('description')),
1599 if is_type(traverse_obj(e
, ('video', 0)), 'VideoObject'):
1600 extract_video_object(e
['video'][0])
1601 elif is_type(traverse_obj(e
, ('subjectOf', 0)), 'VideoObject'):
1602 extract_video_object(e
['subjectOf'][0])
1603 elif is_type(e
, 'VideoObject'):
1604 extract_video_object(e
)
1605 if expected_type
is None:
1609 video
= e
.get('video')
1610 if is_type(video
, 'VideoObject'):
1611 extract_video_object(video
)
1612 if expected_type
is None:
1616 traverse_json_ld(json_ld
)
1618 return filter_dict(info
)
1620 def _search_nextjs_data(self
, webpage
, video_id
, *, transform_source
=None, fatal
=True, **kw
):
1621 return self
._parse
_json
(
1623 r
'(?s)<script[^>]+id=[\'"]__NEXT_DATA__[\'"][^
>]*>([^
<]+)</script
>',
1624 webpage, 'next
.js data
', fatal=fatal, **kw),
1625 video_id, transform_source=transform_source, fatal=fatal)
1627 def _search_nuxt_data(self, webpage, video_id, context_name='__NUXT__
', *, fatal=True, traverse=('data
', 0)):
1628 """Parses Nuxt.js metadata. This works as long as the function __NUXT__ invokes is a pure function"""
1629 rectx = re.escape(context_name)
1630 FUNCTION_RE = r'\
(function\
((?P
<arg_keys
>.*?
)\
){return\s+(?P<js>{.*?}
)\s
*;?\s
*}\
((?P
<arg_vals
>.*?
)\
)'
1631 js, arg_keys, arg_vals = self._search_regex(
1632 (rf'<script
>\s
*window\
.{rectx}
={FUNCTION_RE}\s
*\
)\s
*;?\s
*</script
>', rf'{rectx}\
(.*?{FUNCTION_RE}
'),
1633 webpage, context_name, group=('js
', 'arg_keys
', 'arg_vals
'), fatal=fatal)
1635 args = dict(zip(arg_keys.split(','), arg_vals.split(',')))
1637 for key, val in args.items():
1638 if val in ('undefined
', 'void
0'):
1641 ret = self._parse_json(js, video_id, transform_source=functools.partial(js_to_json, vars=args), fatal=fatal)
1642 return traverse_obj(ret, traverse) or {}
1645 def _hidden_inputs(html):
1646 html = re.sub(r'<!--(?
:(?
!<!--).)*-->', '', html)
1648 for input in re.findall(r'(?i
)(<input[^
>]+>)', html):
1649 attrs = extract_attributes(input)
1652 if attrs.get('type') not in ('hidden
', 'submit
'):
1654 name = attrs.get('name
') or attrs.get('id')
1655 value = attrs.get('value
')
1656 if name and value is not None:
1657 hidden_inputs[name] = value
1658 return hidden_inputs
1660 def _form_hidden_inputs(self, form_id, html):
1661 form = self._search_regex(
1662 r'(?
is)<form
[^
>]+?
id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1663 html, '%s form' % form_id, group='form')
1664 return self._hidden_inputs(form)
1667 regex = r' *((?P<reverse>\+)?(?P<field>[a-zA-Z0-9_]+)((?P<separator>[~:])(?P<limit>.*?))?)? *$'
1669 default = ('hidden', 'aud_or_vid', 'hasvid', 'ie_pref', 'lang', 'quality',
1670 'res', 'fps', 'hdr:12', 'codec:vp9.2', 'size', 'br', 'asr',
1671 'proto', 'ext', 'hasaud', 'source', 'id') # These must not be aliases
1672 ytdl_default = ('hasaud', 'lang', 'quality', 'tbr', 'filesize', 'vbr',
1673 'height', 'width', 'proto', 'vext', 'abr', 'aext',
1674 'fps', 'fs_approx', 'source', 'id')
1677 'vcodec': {'type': 'ordered', 'regex': True,
1678 'order': ['av0?1', 'vp0?9.2', 'vp0?9', '[hx]265|he?vc?', '[hx]264|avc', 'vp0?8', 'mp4v|h263', 'theora', '', None, 'none']},
1679 'acodec': {'type': 'ordered', 'regex': True,
1680 'order': ['[af]lac', 'wav|aiff', 'opus', 'vorbis|ogg', 'aac', 'mp?4a?', 'mp3', 'e-?a?c-?3', 'ac-?3', 'dts', '', None, 'none']},
1681 'hdr': {'type': 'ordered', 'regex': True, 'field': 'dynamic_range',
1682 'order': ['dv', '(hdr)?12', r'(hdr)?10\+', '(hdr)?10', 'hlg', '', 'sdr', None]},
1683 'proto': {'type': 'ordered', 'regex': True, 'field': 'protocol',
1684 'order': ['(ht|f)tps', '(ht|f)tp$', 'm3u8.*', '.*dash', 'websocket_frag', 'rtmpe?', '', 'mms|rtsp', 'ws|websocket', 'f4']},
1685 'vext': {'type': 'ordered', 'field': 'video_ext',
1686 'order': ('mp4', 'webm', 'flv', '', 'none'),
1687 'order_free': ('webm', 'mp4', 'flv', '', 'none')},
1688 'aext': {'type': 'ordered', 'field': 'audio_ext',
1689 'order': ('m4a', 'aac', 'mp3', 'ogg', 'opus', 'webm', '', 'none'),
1690 'order_free': ('opus', 'ogg', 'webm', 'm4a', 'mp3', 'aac', '', 'none')},
1691 'hidden': {'visible': False, 'forced': True, 'type': 'extractor', 'max': -1000},
1692 'aud_or_vid': {'visible': False, 'forced': True, 'type': 'multiple',
1693 'field': ('vcodec', 'acodec'),
1694 'function': lambda it: int(any(v != 'none' for v in it))},
1695 'ie_pref': {'priority': True, 'type': 'extractor'},
1696 'hasvid': {'priority': True, 'field': 'vcodec', 'type': 'boolean', 'not_in_list': ('none',)},
1697 'hasaud': {'field': 'acodec', 'type': 'boolean', 'not_in_list': ('none',)},
1698 'lang': {'convert': 'float', 'field': 'language_preference', 'default': -1},
1699 'quality': {'convert': 'float', 'default': -1},
1700 'filesize': {'convert': 'bytes'},
1701 'fs_approx': {'convert': 'bytes', 'field': 'filesize_approx'},
1702 'id': {'convert': 'string', 'field': 'format_id'},
1703 'height': {'convert': 'float_none'},
1704 'width': {'convert': 'float_none'},
1705 'fps': {'convert': 'float_none'},
1706 'tbr': {'convert': 'float_none'},
1707 'vbr': {'convert': 'float_none'},
1708 'abr': {'convert': 'float_none'},
1709 'asr': {'convert': 'float_none'},
1710 'source': {'convert': 'float', 'field': 'source_preference', 'default': -1},
1712 'codec': {'type': 'combined', 'field': ('vcodec', 'acodec')},
1713 'br': {'type': 'combined', 'field': ('tbr', 'vbr', 'abr'), 'same_limit': True},
1714 'size': {'type': 'combined', 'same_limit': True, 'field': ('filesize', 'fs_approx')},
1715 'ext': {'type': 'combined', 'field': ('vext', 'aext')},
1716 'res': {'type': 'multiple', 'field': ('height', 'width'),
1717 'function': lambda it: (lambda l: min(l) if l else 0)(tuple(filter(None, it)))},
1719 # For compatibility with youtube-dl
1720 'format_id': {'type': 'alias', 'field': 'id'},
1721 'preference': {'type': 'alias', 'field': 'ie_pref'},
1722 'language_preference': {'type': 'alias', 'field': 'lang'},
1723 'source_preference': {'type': 'alias', 'field': 'source'},
1724 'protocol': {'type': 'alias', 'field': 'proto'},
1725 'filesize_approx': {'type': 'alias', 'field': 'fs_approx'},
1728 'dimension': {'type': 'alias', 'field': 'res', 'deprecated': True},
1729 'resolution': {'type': 'alias', 'field': 'res', 'deprecated': True},
1730 'extension': {'type': 'alias', 'field': 'ext', 'deprecated': True},
1731 'bitrate': {'type': 'alias', 'field': 'br', 'deprecated': True},
1732 'total_bitrate': {'type': 'alias', 'field': 'tbr', 'deprecated': True},
1733 'video_bitrate': {'type': 'alias', 'field': 'vbr', 'deprecated': True},
1734 'audio_bitrate': {'type': 'alias', 'field': 'abr', 'deprecated': True},
1735 'framerate': {'type': 'alias', 'field': 'fps', 'deprecated': True},
1736 'filesize_estimate': {'type': 'alias', 'field': 'size', 'deprecated': True},
1737 'samplerate': {'type': 'alias', 'field': 'asr', 'deprecated': True},
1738 'video_ext': {'type': 'alias', 'field': 'vext', 'deprecated': True},
1739 'audio_ext': {'type': 'alias', 'field': 'aext', 'deprecated': True},
1740 'video_codec': {'type': 'alias', 'field': 'vcodec', 'deprecated': True},
1741 'audio_codec': {'type': 'alias', 'field': 'acodec', 'deprecated': True},
1742 'video': {'type': 'alias', 'field': 'hasvid', 'deprecated': True},
1743 'has_video': {'type': 'alias', 'field': 'hasvid', 'deprecated': True},
1744 'audio': {'type': 'alias', 'field': 'hasaud', 'deprecated': True},
1745 'has_audio': {'type': 'alias', 'field': 'hasaud', 'deprecated': True},
1746 'extractor': {'type': 'alias', 'field': 'ie_pref', 'deprecated': True},
1747 'extractor_preference': {'type': 'alias', 'field': 'ie_pref', 'deprecated': True},
1750 def __init__(self, ie, field_preference):
1752 self.ydl = ie._downloader
1753 self.evaluate_params(self.ydl.params, field_preference)
1754 if ie.get_param('verbose'):
1755 self.print_verbose_info(self.ydl.write_debug)
1757 def _get_field_setting(self, field, key):
1758 if field not in self.settings:
1759 if key in ('forced', 'priority'):
1761 self.ydl.deprecation_warning(
1762 f'Using arbitrary fields ({field}) for format sorting is deprecated '
1763 'and may be removed in a future version')
1764 self.settings[field] = {}
1765 propObj = self.settings[field]
1766 if key not in propObj:
1767 type = propObj.get('type')
1769 default = 'preference' if type == 'extractor' else (field,) if type in ('combined', 'multiple') else field
1770 elif key == 'convert':
1771 default = 'order' if type == 'ordered' else 'float_string' if field else 'ignore'
1773 default = {'type': 'field', 'visible': True, 'order': [], 'not_in_list': (None,)}.get(key, None)
1774 propObj[key] = default
1777 def _resolve_field_value(self, field, value, convertNone=False):
1782 value = value.lower()
1783 conversion = self._get_field_setting(field, 'convert')
1784 if conversion == 'ignore':
1786 if conversion == 'string':
1788 elif conversion == 'float_none':
1789 return float_or_none(value)
1790 elif conversion == 'bytes':
1791 return FileDownloader.parse_bytes(value)
1792 elif conversion == 'order':
1793 order_list = (self._use_free_order and self._get_field_setting(field, 'order_free')) or self._get_field_setting(field, 'order')
1794 use_regex = self._get_field_setting(field, 'regex')
1795 list_length = len(order_list)
1796 empty_pos = order_list.index('') if '' in order_list else list_length + 1
1797 if use_regex and value is not None:
1798 for i, regex in enumerate(order_list):
1799 if regex and re.match(regex, value):
1800 return list_length - i
1801 return list_length - empty_pos # not in list
1802 else: # not regex or value = None
1803 return list_length - (order_list.index(value) if value in order_list else empty_pos)
1805 if value.isnumeric():
1808 self.settings[field]['convert'] = 'string'
1811 def evaluate_params(self, params, sort_extractor):
1812 self._use_free_order = params.get('prefer_free_formats', False)
1813 self._sort_user = params.get('format_sort', [])
1814 self._sort_extractor = sort_extractor
1816 def add_item(field, reverse, closest, limit_text):
1817 field = field.lower()
1818 if field in self._order:
1820 self._order.append(field)
1821 limit = self._resolve_field_value(field, limit_text)
1824 'closest': False if limit is None else closest,
1825 'limit_text': limit_text,
1827 if field in self.settings:
1828 self.settings[field].update(data)
1830 self.settings[field] = data
1833 tuple(field for field in self.default if self._get_field_setting(field, 'forced'))
1834 + (tuple() if params.get('format_sort_force', False)
1835 else tuple(field for field in self.default if self._get_field_setting(field, 'priority')))
1836 + tuple(self._sort_user) + tuple(sort_extractor) + self.default)
1838 for item in sort_list:
1839 match = re.match(self.regex, item)
1841 raise ExtractorError('Invalid format sort string "%s" given by extractor' % item)
1842 field = match.group('field')
1845 if self._get_field_setting(field, 'type') == 'alias':
1846 alias, field = field, self._get_field_setting(field, 'field')
1847 if self._get_field_setting(alias, 'deprecated'):
1848 self.ydl.deprecation_warning(
1849 f'Format sorting alias {alias} is deprecated '
1850 f'and may be removed in a future version. Please use {field} instead')
1851 reverse = match.group('reverse') is not None
1852 closest = match.group('separator') == '~'
1853 limit_text = match.group('limit')
1855 has_limit = limit_text is not None
1856 has_multiple_fields = self._get_field_setting(field, 'type') == 'combined'
1857 has_multiple_limits = has_limit and has_multiple_fields and not self._get_field_setting(field, 'same_limit')
1859 fields = self._get_field_setting(field, 'field') if has_multiple_fields else (field,)
1860 limits = limit_text.split(':') if has_multiple_limits else (limit_text,) if has_limit else tuple()
1861 limit_count = len(limits)
1862 for (i, f) in enumerate(fields):
1863 add_item(f, reverse, closest,
1864 limits[i] if i < limit_count
1865 else limits[0] if has_limit and not has_multiple_limits
1868 def print_verbose_info(self, write_debug):
1870 write_debug('Sort order given by user: %s' % ', '.join(self._sort_user))
1871 if self._sort_extractor:
1872 write_debug('Sort order given by extractor: %s' % ', '.join(self._sort_extractor))
1873 write_debug('Formats sorted by: %s' % ', '.join(['%s%s%s' % (
1874 '+' if self._get_field_setting(field, 'reverse') else '', field,
1875 '%s%s(%s)' % ('~' if self._get_field_setting(field, 'closest') else ':',
1876 self._get_field_setting(field, 'limit_text'),
1877 self._get_field_setting(field, 'limit'))
1878 if self._get_field_setting(field, 'limit_text') is not None else '')
1879 for field in self._order if self._get_field_setting(field, 'visible')]))
1881 def _calculate_field_preference_from_value(self, format, field, type, value):
1882 reverse = self._get_field_setting(field, 'reverse')
1883 closest = self._get_field_setting(field, 'closest')
1884 limit = self._get_field_setting(field, 'limit')
1886 if type == 'extractor':
1887 maximum = self._get_field_setting(field, 'max')
1888 if value is None or (maximum is not None and value >= maximum):
1890 elif type == 'boolean':
1891 in_list = self._get_field_setting(field, 'in_list')
1892 not_in_list = self._get_field_setting(field, 'not_in_list')
1893 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
1894 elif type == 'ordered':
1895 value = self._resolve_field_value(field, value, True)
1897 # try to convert to number
1898 val_num = float_or_none(value, default=self._get_field_setting(field, 'default'))
1899 is_num = self._get_field_setting(field, 'convert') != 'string' and val_num is not None
1903 return ((-10, 0) if value is None
1904 else (1, value, 0) if not is_num # if a field has mixed strings and numbers, strings are sorted higher
1905 else (0, -abs(value - limit), value - limit if reverse else limit - value) if closest
1906 else (0, value, 0) if not reverse and (limit is None or value <= limit)
1907 else (0, -value, 0) if limit is None or (reverse and value == limit) or value > limit
1908 else (-1, value, 0))
1910 def _calculate_field_preference(self, format, field):
1911 type = self._get_field_setting(field, 'type') # extractor, boolean, ordered, field, multiple
1912 get_value = lambda f: format.get(self._get_field_setting(f, 'field'))
1913 if type == 'multiple':
1914 type = 'field' # Only 'field' is allowed in multiple for now
1915 actual_fields = self._get_field_setting(field, 'field')
1917 value = self._get_field_setting(field, 'function')(get_value(f) for f in actual_fields)
1919 value = get_value(field)
1920 return self._calculate_field_preference_from_value(format, field, type, value)
1922 def calculate_preference(self, format):
1923 # Determine missing protocol
1924 if not format.get('protocol'):
1925 format['protocol'] = determine_protocol(format)
1927 # Determine missing ext
1928 if not format.get('ext') and 'url' in format:
1929 format['ext'] = determine_ext(format['url'])
1930 if format.get('vcodec') == 'none':
1931 format['audio_ext'] = format['ext'] if format.get('acodec') != 'none' else 'none'
1932 format['video_ext'] = 'none'
1934 format['video_ext'] = format['ext']
1935 format['audio_ext'] = 'none'
1936 # if format.get('preference') is None and format.get('ext') in ('f4f', 'f4m'): # Not supported?
1937 # format['preference'] = -1000
1939 # Determine missing bitrates
1940 if format.get('tbr') is None:
1941 if format.get('vbr') is not None and format.get('abr') is not None:
1942 format['tbr'] = format.get('vbr', 0) + format.get('abr', 0)
1944 if format.get('vcodec') != 'none' and format.get('vbr') is None:
1945 format['vbr'] = format.get('tbr') - format.get('abr', 0)
1946 if format.get('acodec') != 'none' and format.get('abr') is None:
1947 format['abr'] = format.get('tbr') - format.get('vbr', 0)
1949 return tuple(self._calculate_field_preference(format, field) for field in self._order)
1951 def _sort_formats(self, formats, field_preference=[]):
1954 formats.sort(key=self.FormatSort(self, field_preference).calculate_preference)
1956 def _check_formats(self, formats, video_id):
1958 formats[:] = filter(
1959 lambda f: self._is_valid_url(
1961 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1965 def _remove_duplicate_formats(formats):
1969 if f['url'] not in format_urls:
1970 format_urls.add(f['url'])
1971 unique_formats.append(f)
1972 formats[:] = unique_formats
1974 def _is_valid_url(self, url, video_id, item='video', headers={}):
1975 url = self._proto_relative_url(url, scheme='http:')
1976 # For now assume non HTTP(S) URLs always valid
1977 if not (url.startswith('http://') or url.startswith('https://')):
1980 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1982 except ExtractorError as e:
1984 '%s: %s URL is invalid, skipping: %s'
1985 % (video_id, item, error_to_compat_str(e.cause)))
1988 def http_scheme(self):
1989 """ Either "http
:" or "https
:", depending on the user's preferences """
1992 if self.get_param('prefer_insecure', False)
1995 def _proto_relative_url(self, url, scheme=None):
1996 scheme = scheme or self.http_scheme()
1997 assert scheme.endswith(':')
1998 return sanitize_url(url, scheme=scheme[:-1])
2000 def _sleep(self, timeout, video_id, msg_template=None):
2001 if msg_template is None:
2002 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
2003 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
2007 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
2008 transform_source=lambda s: fix_xml_ampersands(s).strip(),
2009 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
2010 res = self._download_xml_handle(
2011 manifest_url, video_id, 'Downloading f4m manifest',
2012 'Unable to download f4m manifest',
2013 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
2014 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
2015 transform_source=transform_source,
2016 fatal=fatal, data=data, headers=headers, query=query)
2020 manifest, urlh = res
2021 manifest_url = urlh.geturl()
2023 return self._parse_f4m_formats(
2024 manifest, manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
2025 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
2027 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
2028 transform_source=lambda s: fix_xml_ampersands(s).strip(),
2029 fatal=True, m3u8_id=None):
2030 if not isinstance(manifest, xml.etree.ElementTree.Element) and not fatal:
2033 # currently yt-dlp cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
2034 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
2035 if akamai_pv is not None and ';' in akamai_pv.text:
2036 playerVerificationChallenge = akamai_pv.text.split(';')[0]
2037 if playerVerificationChallenge.strip() != '':
2041 manifest_version = '1.0'
2042 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
2044 manifest_version = '2.0'
2045 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
2046 # Remove unsupported DRM protected media from final formats
2047 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
2048 media_nodes = remove_encrypted_media(media_nodes)
2052 manifest_base_url = get_base_url(manifest)
2054 bootstrap_info = xpath_element(
2055 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
2056 'bootstrap info', default=None)
2059 mime_type = xpath_text(
2060 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
2061 'base URL', default=None)
2062 if mime_type and mime_type.startswith('audio/'):
2065 for i, media_el in enumerate(media_nodes):
2066 tbr = int_or_none(media_el.attrib.get('bitrate'))
2067 width = int_or_none(media_el.attrib.get('width'))
2068 height = int_or_none(media_el.attrib.get('height'))
2069 format_id = join_nonempty(f4m_id, tbr or i)
2070 # If <bootstrapInfo> is present, the specified f4m is a
2071 # stream-level manifest, and only set-level manifests may refer to
2072 # external resources. See section 11.4 and section 4 of F4M spec
2073 if bootstrap_info is None:
2075 # @href is introduced in 2.0, see section 11.6 of F4M spec
2076 if manifest_version == '2.0':
2077 media_url = media_el.attrib.get('href')
2078 if media_url is None:
2079 media_url = media_el.attrib.get('url')
2083 media_url if media_url.startswith('http://') or media_url.startswith('https://')
2084 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
2085 # If media_url is itself a f4m manifest do the recursive extraction
2086 # since bitrates in parent manifest (this one) and media_url manifest
2087 # may differ leading to inability to resolve the format by requested
2088 # bitrate in f4m downloader
2089 ext = determine_ext(manifest_url)
2091 f4m_formats = self._extract_f4m_formats(
2092 manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
2093 transform_source=transform_source, fatal=fatal)
2094 # Sometimes stream-level manifest contains single media entry that
2095 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
2096 # At the same time parent's media entry in set-level manifest may
2097 # contain it. We will copy it from parent in such cases.
2098 if len(f4m_formats) == 1:
2101 'tbr': f.get('tbr') or tbr,
2102 'width': f.get('width') or width,
2103 'height': f.get('height') or height,
2104 'format_id': f.get('format_id') if not tbr else format_id,
2107 formats.extend(f4m_formats)
2110 formats.extend(self._extract_m3u8_formats(
2111 manifest_url, video_id, 'mp4', preference=preference,
2112 quality=quality, m3u8_id=m3u8_id, fatal=fatal))
2115 'format_id': format_id,
2116 'url': manifest_url,
2117 'manifest_url': manifest_url,
2118 'ext': 'flv' if bootstrap_info is not None else None,
2124 'preference': preference,
2129 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, quality=None, m3u8_id=None):
2131 'format_id': join_nonempty(m3u8_id, 'meta'),
2135 'preference': preference - 100 if preference else -100,
2137 'resolution': 'multiple',
2138 'format_note': 'Quality selection URL',
2141 def _report_ignoring_subs(self, name):
2142 self.report_warning(bug_reports_message(
2143 f'Ignoring subtitle tracks found in the {name} manifest; '
2144 'if any subtitle tracks are missing,'
2147 def _extract_m3u8_formats(self, *args, **kwargs):
2148 fmts, subs = self._extract_m3u8_formats_and_subtitles(*args, **kwargs)
2150 self._report_ignoring_subs('HLS')
2153 def _extract_m3u8_formats_and_subtitles(
2154 self, m3u8_url, video_id, ext=None, entry_protocol='m3u8_native',
2155 preference=None, quality=None, m3u8_id=None, note=None,
2156 errnote=None, fatal=True, live=False, data=None, headers={},
2159 res = self._download_webpage_handle(
2161 note='Downloading m3u8 information' if note is None else note,
2162 errnote='Failed to download m3u8 information' if errnote is None else errnote,
2163 fatal=fatal, data=data, headers=headers, query=query)
2168 m3u8_doc, urlh = res
2169 m3u8_url = urlh.geturl()
2171 return self._parse_m3u8_formats_and_subtitles(
2172 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
2173 preference=preference, quality=quality, m3u8_id=m3u8_id,
2174 note=note, errnote=errnote, fatal=fatal, live=live, data=data,
2175 headers=headers, query=query, video_id=video_id)
2177 def _parse_m3u8_formats_and_subtitles(
2178 self, m3u8_doc, m3u8_url=None, ext=None, entry_protocol='m3u8_native',
2179 preference=None, quality=None, m3u8_id=None, live=False, note=None,
2180 errnote=None, fatal=True, data=None, headers={}, query={},
2182 formats, subtitles = [], {}
2184 has_drm = re.search('|'.join([
2185 r'#EXT-X-FAXS-CM:', # Adobe Flash Access
2186 r'#EXT-X-(?:SESSION-)?KEY:.*?URI="skd
://', # Apple FairPlay
2189 def format_url(url):
2190 return url if re.match(r'^https?
://', url) else urllib.parse.urljoin(m3u8_url, url)
2192 if self.get_param('hls_split_discontinuity
', False):
2193 def _extract_m3u8_playlist_indices(manifest_url=None, m3u8_doc=None):
2195 if not manifest_url:
2197 m3u8_doc = self._download_webpage(
2198 manifest_url, video_id, fatal=fatal, data=data, headers=headers,
2199 note=False, errnote='Failed to download m3u8 playlist information
')
2200 if m3u8_doc is False:
2202 return range(1 + sum(line.startswith('#EXT-X-DISCONTINUITY') for line in m3u8_doc.splitlines()))
2205 def _extract_m3u8_playlist_indices(*args
, **kwargs
):
2209 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
2210 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
2211 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
2213 # We should try extracting formats only from master playlists [1, 4.3.4],
2214 # i.e. playlists that describe available qualities. On the other hand
2215 # media playlists [1, 4.3.3] should be returned as is since they contain
2216 # just the media without qualities renditions.
2217 # Fortunately, master playlist can be easily distinguished from media
2218 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
2219 # master playlist tags MUST NOT appear in a media playlist and vice versa.
2220 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
2221 # media playlist and MUST NOT appear in master playlist thus we can
2222 # clearly detect media playlist with this criterion.
2224 if '#EXT-X-TARGETDURATION' in m3u8_doc
: # media playlist, return as is
2226 'format_id': join_nonempty(m3u8_id
, idx
),
2227 'format_index': idx
,
2228 'url': m3u8_url
or encode_data_uri(m3u8_doc
.encode('utf-8'), 'application/x-mpegurl'),
2230 'protocol': entry_protocol
,
2231 'preference': preference
,
2234 } for idx
in _extract_m3u8_playlist_indices(m3u8_doc
=m3u8_doc
)]
2236 return formats
, subtitles
2239 last_stream_inf
= {}
2241 def extract_media(x_media_line
):
2242 media
= parse_m3u8_attributes(x_media_line
)
2243 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
2244 media_type
, group_id
, name
= media
.get('TYPE'), media
.get('GROUP-ID'), media
.get('NAME')
2245 if not (media_type
and group_id
and name
):
2247 groups
.setdefault(group_id
, []).append(media
)
2248 # <https://tools.ietf.org/html/rfc8216#section-4.3.4.1>
2249 if media_type
== 'SUBTITLES':
2250 # According to RFC 8216 §4.3.4.2.1, URI is REQUIRED in the
2251 # EXT-X-MEDIA tag if the media type is SUBTITLES.
2252 # However, lack of URI has been spotted in the wild.
2253 # e.g. NebulaIE; see https://github.com/yt-dlp/yt-dlp/issues/339
2254 if not media
.get('URI'):
2256 url
= format_url(media
['URI'])
2259 'ext': determine_ext(url
),
2261 if sub_info
['ext'] == 'm3u8':
2262 # Per RFC 8216 §3.1, the only possible subtitle format m3u8
2263 # files may contain is WebVTT:
2264 # <https://tools.ietf.org/html/rfc8216#section-3.1>
2265 sub_info
['ext'] = 'vtt'
2266 sub_info
['protocol'] = 'm3u8_native'
2267 lang
= media
.get('LANGUAGE') or 'und'
2268 subtitles
.setdefault(lang
, []).append(sub_info
)
2269 if media_type
not in ('VIDEO', 'AUDIO'):
2271 media_url
= media
.get('URI')
2273 manifest_url
= format_url(media_url
)
2275 'format_id': join_nonempty(m3u8_id
, group_id
, name
, idx
),
2276 'format_note': name
,
2277 'format_index': idx
,
2278 'url': manifest_url
,
2279 'manifest_url': m3u8_url
,
2280 'language': media
.get('LANGUAGE'),
2282 'protocol': entry_protocol
,
2283 'preference': preference
,
2285 'vcodec': 'none' if media_type
== 'AUDIO' else None,
2286 } for idx
in _extract_m3u8_playlist_indices(manifest_url
))
2288 def build_stream_name():
2289 # Despite specification does not mention NAME attribute for
2290 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
2291 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
2292 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
2293 stream_name
= last_stream_inf
.get('NAME')
2296 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
2297 # from corresponding rendition group
2298 stream_group_id
= last_stream_inf
.get('VIDEO')
2299 if not stream_group_id
:
2301 stream_group
= groups
.get(stream_group_id
)
2302 if not stream_group
:
2303 return stream_group_id
2304 rendition
= stream_group
[0]
2305 return rendition
.get('NAME') or stream_group_id
2307 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
2308 # chance to detect video only formats when EXT-X-STREAM-INF tags
2309 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
2310 for line
in m3u8_doc
.splitlines():
2311 if line
.startswith('#EXT-X-MEDIA:'):
2314 for line
in m3u8_doc
.splitlines():
2315 if line
.startswith('#EXT-X-STREAM-INF:'):
2316 last_stream_inf
= parse_m3u8_attributes(line
)
2317 elif line
.startswith('#') or not line
.strip():
2320 tbr
= float_or_none(
2321 last_stream_inf
.get('AVERAGE-BANDWIDTH')
2322 or last_stream_inf
.get('BANDWIDTH'), scale
=1000)
2323 manifest_url
= format_url(line
.strip())
2325 for idx
in _extract_m3u8_playlist_indices(manifest_url
):
2326 format_id
= [m3u8_id
, None, idx
]
2327 # Bandwidth of live streams may differ over time thus making
2328 # format_id unpredictable. So it's better to keep provided
2331 stream_name
= build_stream_name()
2332 format_id
[1] = stream_name
or '%d' % (tbr
or len(formats
))
2334 'format_id': join_nonempty(*format_id
),
2335 'format_index': idx
,
2336 'url': manifest_url
,
2337 'manifest_url': m3u8_url
,
2340 'fps': float_or_none(last_stream_inf
.get('FRAME-RATE')),
2341 'protocol': entry_protocol
,
2342 'preference': preference
,
2345 resolution
= last_stream_inf
.get('RESOLUTION')
2347 mobj
= re
.search(r
'(?P<width>\d+)[xX](?P<height>\d+)', resolution
)
2349 f
['width'] = int(mobj
.group('width'))
2350 f
['height'] = int(mobj
.group('height'))
2351 # Unified Streaming Platform
2353 r
'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f
['url'])
2355 abr
, vbr
= mobj
.groups()
2356 abr
, vbr
= float_or_none(abr
, 1000), float_or_none(vbr
, 1000)
2361 codecs
= parse_codecs(last_stream_inf
.get('CODECS'))
2363 audio_group_id
= last_stream_inf
.get('AUDIO')
2364 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
2365 # references a rendition group MUST have a CODECS attribute.
2366 # However, this is not always respected, for example, [2]
2367 # contains EXT-X-STREAM-INF tag which references AUDIO
2368 # rendition group but does not have CODECS and despite
2369 # referencing an audio group it represents a complete
2370 # (with audio and video) format. So, for such cases we will
2371 # ignore references to rendition groups and treat them
2372 # as complete formats.
2373 if audio_group_id
and codecs
and f
.get('vcodec') != 'none':
2374 audio_group
= groups
.get(audio_group_id
)
2375 if audio_group
and audio_group
[0].get('URI'):
2376 # TODO: update acodec for audio only formats with
2378 f
['acodec'] = 'none'
2379 if not f
.get('ext'):
2380 f
['ext'] = 'm4a' if f
.get('vcodec') == 'none' else 'mp4'
2384 progressive_uri
= last_stream_inf
.get('PROGRESSIVE-URI')
2387 del http_f
['manifest_url']
2389 'format_id': f
['format_id'].replace('hls-', 'http-'),
2391 'url': progressive_uri
,
2393 formats
.append(http_f
)
2395 last_stream_inf
= {}
2396 return formats
, subtitles
2398 def _extract_m3u8_vod_duration(
2399 self
, m3u8_vod_url
, video_id
, note
=None, errnote
=None, data
=None, headers
={}, query={}
):
2401 m3u8_vod
= self
._download
_webpage
(
2402 m3u8_vod_url
, video_id
,
2403 note
='Downloading m3u8 VOD manifest' if note
is None else note
,
2404 errnote
='Failed to download VOD manifest' if errnote
is None else errnote
,
2405 fatal
=False, data
=data
, headers
=headers
, query
=query
)
2407 return self
._parse
_m
3u8_vod
_duration
(m3u8_vod
or '', video_id
)
2409 def _parse_m3u8_vod_duration(self
, m3u8_vod
, video_id
):
2410 if '#EXT-X-PLAYLIST-TYPE:VOD' not in m3u8_vod
:
2414 float(line
[len('#EXTINF:'):].split(',')[0])
2415 for line
in m3u8_vod
.splitlines() if line
.startswith('#EXTINF:'))) or None
2418 def _xpath_ns(path
, namespace
=None):
2422 for c
in path
.split('/'):
2423 if not c
or c
== '.':
2426 out
.append('{%s}%s' % (namespace
, c
))
2427 return '/'.join(out
)
2429 def _extract_smil_formats_and_subtitles(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None, transform_source
=None):
2430 res
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
, transform_source
=transform_source
)
2436 smil_url
= urlh
.geturl()
2438 namespace
= self
._parse
_smil
_namespace
(smil
)
2440 fmts
= self
._parse
_smil
_formats
(
2441 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2442 subs
= self
._parse
_smil
_subtitles
(
2443 smil
, namespace
=namespace
)
2447 def _extract_smil_formats(self
, *args
, **kwargs
):
2448 fmts
, subs
= self
._extract
_smil
_formats
_and
_subtitles
(*args
, **kwargs
)
2450 self
._report
_ignoring
_subs
('SMIL')
2453 def _extract_smil_info(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None):
2454 res
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
)
2459 smil_url
= urlh
.geturl()
2461 return self
._parse
_smil
(smil
, smil_url
, video_id
, f4m_params
=f4m_params
)
2463 def _download_smil(self
, smil_url
, video_id
, fatal
=True, transform_source
=None):
2464 return self
._download
_xml
_handle
(
2465 smil_url
, video_id
, 'Downloading SMIL file',
2466 'Unable to download SMIL file', fatal
=fatal
, transform_source
=transform_source
)
2468 def _parse_smil(self
, smil
, smil_url
, video_id
, f4m_params
=None):
2469 namespace
= self
._parse
_smil
_namespace
(smil
)
2471 formats
= self
._parse
_smil
_formats
(
2472 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2473 subtitles
= self
._parse
_smil
_subtitles
(smil
, namespace
=namespace
)
2475 video_id
= os
.path
.splitext(url_basename(smil_url
))[0]
2479 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2480 name
= meta
.attrib
.get('name')
2481 content
= meta
.attrib
.get('content')
2482 if not name
or not content
:
2484 if not title
and name
== 'title':
2486 elif not description
and name
in ('description', 'abstract'):
2487 description
= content
2488 elif not upload_date
and name
== 'date':
2489 upload_date
= unified_strdate(content
)
2492 'id': image
.get('type'),
2493 'url': image
.get('src'),
2494 'width': int_or_none(image
.get('width')),
2495 'height': int_or_none(image
.get('height')),
2496 } for image
in smil
.findall(self
._xpath
_ns
('.//image', namespace
)) if image
.get('src')]
2500 'title': title
or video_id
,
2501 'description': description
,
2502 'upload_date': upload_date
,
2503 'thumbnails': thumbnails
,
2505 'subtitles': subtitles
,
2508 def _parse_smil_namespace(self
, smil
):
2509 return self
._search
_regex
(
2510 r
'(?i)^{([^}]+)?}smil$', smil
.tag
, 'namespace', default
=None)
2512 def _parse_smil_formats(self
, smil
, smil_url
, video_id
, namespace
=None, f4m_params
=None, transform_rtmp_url
=None):
2514 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2515 b
= meta
.get('base') or meta
.get('httpBase')
2527 media
= smil
.findall(self
._xpath
_ns
('.//video', namespace
)) + smil
.findall(self
._xpath
_ns
('.//audio', namespace
))
2528 for medium
in media
:
2529 src
= medium
.get('src')
2530 if not src
or src
in srcs
:
2534 bitrate
= float_or_none(medium
.get('system-bitrate') or medium
.get('systemBitrate'), 1000)
2535 filesize
= int_or_none(medium
.get('size') or medium
.get('fileSize'))
2536 width
= int_or_none(medium
.get('width'))
2537 height
= int_or_none(medium
.get('height'))
2538 proto
= medium
.get('proto')
2539 ext
= medium
.get('ext')
2540 src_ext
= determine_ext(src
)
2541 streamer
= medium
.get('streamer') or base
2543 if proto
== 'rtmp' or streamer
.startswith('rtmp'):
2549 'format_id': 'rtmp-%d' % (rtmp_count
if bitrate
is None else bitrate
),
2551 'filesize': filesize
,
2555 if transform_rtmp_url
:
2556 streamer
, src
= transform_rtmp_url(streamer
, src
)
2557 formats
[-1].update({
2563 src_url
= src
if src
.startswith('http') else urllib
.parse
.urljoin(base
, src
)
2564 src_url
= src_url
.strip()
2566 if proto
== 'm3u8' or src_ext
== 'm3u8':
2567 m3u8_formats
= self
._extract
_m
3u8_formats
(
2568 src_url
, video_id
, ext
or 'mp4', m3u8_id
='hls', fatal
=False)
2569 if len(m3u8_formats
) == 1:
2571 m3u8_formats
[0].update({
2572 'format_id': 'hls-%d' % (m3u8_count
if bitrate
is None else bitrate
),
2577 formats
.extend(m3u8_formats
)
2578 elif src_ext
== 'f4m':
2583 'plugin': 'flowplayer-3.2.0.1',
2585 f4m_url
+= '&' if '?' in f4m_url
else '?'
2586 f4m_url
+= urllib
.parse
.urlencode(f4m_params
)
2587 formats
.extend(self
._extract
_f
4m
_formats
(f4m_url
, video_id
, f4m_id
='hds', fatal
=False))
2588 elif src_ext
== 'mpd':
2589 formats
.extend(self
._extract
_mpd
_formats
(
2590 src_url
, video_id
, mpd_id
='dash', fatal
=False))
2591 elif re
.search(r
'\.ism/[Mm]anifest', src_url
):
2592 formats
.extend(self
._extract
_ism
_formats
(
2593 src_url
, video_id
, ism_id
='mss', fatal
=False))
2594 elif src_url
.startswith('http') and self
._is
_valid
_url
(src
, video_id
):
2598 'ext': ext
or src_ext
or 'flv',
2599 'format_id': 'http-%d' % (bitrate
or http_count
),
2601 'filesize': filesize
,
2606 for medium
in smil
.findall(self
._xpath
_ns
('.//imagestream', namespace
)):
2607 src
= medium
.get('src')
2608 if not src
or src
in srcs
:
2614 'format_id': 'imagestream-%d' % (imgs_count
),
2616 'ext': mimetype2ext(medium
.get('type')),
2619 'width': int_or_none(medium
.get('width')),
2620 'height': int_or_none(medium
.get('height')),
2621 'format_note': 'SMIL storyboards',
2626 def _parse_smil_subtitles(self
, smil
, namespace
=None, subtitles_lang
='en'):
2629 for num
, textstream
in enumerate(smil
.findall(self
._xpath
_ns
('.//textstream', namespace
))):
2630 src
= textstream
.get('src')
2631 if not src
or src
in urls
:
2634 ext
= textstream
.get('ext') or mimetype2ext(textstream
.get('type')) or determine_ext(src
)
2635 lang
= textstream
.get('systemLanguage') or textstream
.get('systemLanguageName') or textstream
.get('lang') or subtitles_lang
2636 subtitles
.setdefault(lang
, []).append({
2642 def _extract_xspf_playlist(self
, xspf_url
, playlist_id
, fatal
=True):
2643 res
= self
._download
_xml
_handle
(
2644 xspf_url
, playlist_id
, 'Downloading xpsf playlist',
2645 'Unable to download xspf manifest', fatal
=fatal
)
2650 xspf_url
= urlh
.geturl()
2652 return self
._parse
_xspf
(
2653 xspf
, playlist_id
, xspf_url
=xspf_url
,
2654 xspf_base_url
=base_url(xspf_url
))
2656 def _parse_xspf(self
, xspf_doc
, playlist_id
, xspf_url
=None, xspf_base_url
=None):
2658 'xspf': 'http://xspf.org/ns/0/',
2659 's1': 'http://static.streamone.nl/player/ns/0',
2663 for track
in xspf_doc
.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP
)):
2665 track
, xpath_with_ns('./xspf:title', NS_MAP
), 'title', default
=playlist_id
)
2666 description
= xpath_text(
2667 track
, xpath_with_ns('./xspf:annotation', NS_MAP
), 'description')
2668 thumbnail
= xpath_text(
2669 track
, xpath_with_ns('./xspf:image', NS_MAP
), 'thumbnail')
2670 duration
= float_or_none(
2671 xpath_text(track
, xpath_with_ns('./xspf:duration', NS_MAP
), 'duration'), 1000)
2674 for location
in track
.findall(xpath_with_ns('./xspf:location', NS_MAP
)):
2675 format_url
= urljoin(xspf_base_url
, location
.text
)
2680 'manifest_url': xspf_url
,
2681 'format_id': location
.get(xpath_with_ns('s1:label', NS_MAP
)),
2682 'width': int_or_none(location
.get(xpath_with_ns('s1:width', NS_MAP
))),
2683 'height': int_or_none(location
.get(xpath_with_ns('s1:height', NS_MAP
))),
2685 self
._sort
_formats
(formats
)
2690 'description': description
,
2691 'thumbnail': thumbnail
,
2692 'duration': duration
,
2697 def _extract_mpd_formats(self
, *args
, **kwargs
):
2698 fmts
, subs
= self
._extract
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2700 self
._report
_ignoring
_subs
('DASH')
2703 def _extract_mpd_formats_and_subtitles(
2704 self
, mpd_url
, video_id
, mpd_id
=None, note
=None, errnote
=None,
2705 fatal
=True, data
=None, headers
={}, query={}
):
2706 res
= self
._download
_xml
_handle
(
2708 note
='Downloading MPD manifest' if note
is None else note
,
2709 errnote
='Failed to download MPD manifest' if errnote
is None else errnote
,
2710 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
2717 # We could have been redirected to a new url when we retrieved our mpd file.
2718 mpd_url
= urlh
.geturl()
2719 mpd_base_url
= base_url(mpd_url
)
2721 return self
._parse
_mpd
_formats
_and
_subtitles
(
2722 mpd_doc
, mpd_id
, mpd_base_url
, mpd_url
)
2724 def _parse_mpd_formats(self
, *args
, **kwargs
):
2725 fmts
, subs
= self
._parse
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2727 self
._report
_ignoring
_subs
('DASH')
2730 def _parse_mpd_formats_and_subtitles(
2731 self
, mpd_doc
, mpd_id
=None, mpd_base_url
='', mpd_url
=None):
2733 Parse formats from MPD manifest.
2735 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
2736 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
2737 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
2739 if not self
.get_param('dynamic_mpd', True):
2740 if mpd_doc
.get('type') == 'dynamic':
2743 namespace
= self
._search
_regex
(r
'(?i)^{([^}]+)?}MPD$', mpd_doc
.tag
, 'namespace', default
=None)
2746 return self
._xpath
_ns
(path
, namespace
)
2748 def is_drm_protected(element
):
2749 return element
.find(_add_ns('ContentProtection')) is not None
2751 def extract_multisegment_info(element
, ms_parent_info
):
2752 ms_info
= ms_parent_info
.copy()
2754 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2755 # common attributes and elements. We will only extract relevant
2757 def extract_common(source
):
2758 segment_timeline
= source
.find(_add_ns('SegmentTimeline'))
2759 if segment_timeline
is not None:
2760 s_e
= segment_timeline
.findall(_add_ns('S'))
2762 ms_info
['total_number'] = 0
2765 r
= int(s
.get('r', 0))
2766 ms_info
['total_number'] += 1 + r
2767 ms_info
['s'].append({
2768 't': int(s
.get('t', 0)),
2769 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2770 'd': int(s
.attrib
['d']),
2773 start_number
= source
.get('startNumber')
2775 ms_info
['start_number'] = int(start_number
)
2776 timescale
= source
.get('timescale')
2778 ms_info
['timescale'] = int(timescale
)
2779 segment_duration
= source
.get('duration')
2780 if segment_duration
:
2781 ms_info
['segment_duration'] = float(segment_duration
)
2783 def extract_Initialization(source
):
2784 initialization
= source
.find(_add_ns('Initialization'))
2785 if initialization
is not None:
2786 ms_info
['initialization_url'] = initialization
.attrib
['sourceURL']
2788 segment_list
= element
.find(_add_ns('SegmentList'))
2789 if segment_list
is not None:
2790 extract_common(segment_list
)
2791 extract_Initialization(segment_list
)
2792 segment_urls_e
= segment_list
.findall(_add_ns('SegmentURL'))
2794 ms_info
['segment_urls'] = [segment
.attrib
['media'] for segment
in segment_urls_e
]
2796 segment_template
= element
.find(_add_ns('SegmentTemplate'))
2797 if segment_template
is not None:
2798 extract_common(segment_template
)
2799 media
= segment_template
.get('media')
2801 ms_info
['media'] = media
2802 initialization
= segment_template
.get('initialization')
2804 ms_info
['initialization'] = initialization
2806 extract_Initialization(segment_template
)
2809 mpd_duration
= parse_duration(mpd_doc
.get('mediaPresentationDuration'))
2810 formats
, subtitles
= [], {}
2811 stream_numbers
= collections
.defaultdict(int)
2812 for period
in mpd_doc
.findall(_add_ns('Period')):
2813 period_duration
= parse_duration(period
.get('duration')) or mpd_duration
2814 period_ms_info
= extract_multisegment_info(period
, {
2818 for adaptation_set
in period
.findall(_add_ns('AdaptationSet')):
2819 adaption_set_ms_info
= extract_multisegment_info(adaptation_set
, period_ms_info
)
2820 for representation
in adaptation_set
.findall(_add_ns('Representation')):
2821 representation_attrib
= adaptation_set
.attrib
.copy()
2822 representation_attrib
.update(representation
.attrib
)
2823 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2824 mime_type
= representation_attrib
['mimeType']
2825 content_type
= representation_attrib
.get('contentType', mime_type
.split('/')[0])
2827 codec_str
= representation_attrib
.get('codecs', '')
2828 # Some kind of binary subtitle found in some youtube livestreams
2829 if mime_type
== 'application/x-rawcc':
2830 codecs
= {'scodec': codec_str}
2832 codecs
= parse_codecs(codec_str
)
2833 if content_type
not in ('video', 'audio', 'text'):
2834 if mime_type
== 'image/jpeg':
2835 content_type
= mime_type
2836 elif codecs
.get('vcodec', 'none') != 'none':
2837 content_type
= 'video'
2838 elif codecs
.get('acodec', 'none') != 'none':
2839 content_type
= 'audio'
2840 elif codecs
.get('scodec', 'none') != 'none':
2841 content_type
= 'text'
2842 elif mimetype2ext(mime_type
) in ('tt', 'dfxp', 'ttml', 'xml', 'json'):
2843 content_type
= 'text'
2845 self
.report_warning('Unknown MIME type %s in DASH manifest' % mime_type
)
2849 for element
in (representation
, adaptation_set
, period
, mpd_doc
):
2850 base_url_e
= element
.find(_add_ns('BaseURL'))
2851 if try_call(lambda: base_url_e
.text
) is not None:
2852 base_url
= base_url_e
.text
+ base_url
2853 if re
.match(r
'^https?://', base_url
):
2855 if mpd_base_url
and base_url
.startswith('/'):
2856 base_url
= urllib
.parse
.urljoin(mpd_base_url
, base_url
)
2857 elif mpd_base_url
and not re
.match(r
'^https?://', base_url
):
2858 if not mpd_base_url
.endswith('/'):
2860 base_url
= mpd_base_url
+ base_url
2861 representation_id
= representation_attrib
.get('id')
2862 lang
= representation_attrib
.get('lang')
2863 url_el
= representation
.find(_add_ns('BaseURL'))
2864 filesize
= int_or_none(url_el
.attrib
.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el
is not None else None)
2865 bandwidth
= int_or_none(representation_attrib
.get('bandwidth'))
2866 if representation_id
is not None:
2867 format_id
= representation_id
2869 format_id
= content_type
2871 format_id
= mpd_id
+ '-' + format_id
2872 if content_type
in ('video', 'audio'):
2874 'format_id': format_id
,
2875 'manifest_url': mpd_url
,
2876 'ext': mimetype2ext(mime_type
),
2877 'width': int_or_none(representation_attrib
.get('width')),
2878 'height': int_or_none(representation_attrib
.get('height')),
2879 'tbr': float_or_none(bandwidth
, 1000),
2880 'asr': int_or_none(representation_attrib
.get('audioSamplingRate')),
2881 'fps': int_or_none(representation_attrib
.get('frameRate')),
2882 'language': lang
if lang
not in ('mul', 'und', 'zxx', 'mis') else None,
2883 'format_note': 'DASH %s' % content_type
,
2884 'filesize': filesize
,
2885 'container': mimetype2ext(mime_type
) + '_dash',
2888 elif content_type
== 'text':
2890 'ext': mimetype2ext(mime_type
),
2891 'manifest_url': mpd_url
,
2892 'filesize': filesize
,
2894 elif content_type
== 'image/jpeg':
2895 # See test case in VikiIE
2896 # https://www.viki.com/videos/1175236v-choosing-spouse-by-lottery-episode-1
2898 'format_id': format_id
,
2900 'manifest_url': mpd_url
,
2901 'format_note': 'DASH storyboards (jpeg)',
2905 if is_drm_protected(adaptation_set
) or is_drm_protected(representation
):
2907 representation_ms_info
= extract_multisegment_info(representation
, adaption_set_ms_info
)
2909 def prepare_template(template_name
, identifiers
):
2910 tmpl
= representation_ms_info
[template_name
]
2911 # First of, % characters outside $...$ templates
2912 # must be escaped by doubling for proper processing
2913 # by % operator string formatting used further (see
2914 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2920 in_template
= not in_template
2921 elif c
== '%' and not in_template
:
2923 # Next, $...$ templates are translated to their
2924 # %(...) counterparts to be used with % operator
2925 if representation_id
is not None:
2926 t
= t
.replace('$RepresentationID$', representation_id
)
2927 t
= re
.sub(r
'\$(%s)\$' % '|'.join(identifiers
), r
'%(\1)d', t
)
2928 t
= re
.sub(r
'\$(%s)%%([^$]+)\$' % '|'.join(identifiers
), r
'%(\1)\2', t
)
2929 t
.replace('$$', '$')
2932 # @initialization is a regular template like @media one
2933 # so it should be handled just the same way (see
2934 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2935 if 'initialization' in representation_ms_info
:
2936 initialization_template
= prepare_template(
2938 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2939 # $Time$ shall not be included for @initialization thus
2940 # only $Bandwidth$ remains
2942 representation_ms_info
['initialization_url'] = initialization_template
% {
2943 'Bandwidth': bandwidth
,
2946 def location_key(location
):
2947 return 'url' if re
.match(r
'^https?://', location
) else 'path'
2949 if 'segment_urls' not in representation_ms_info
and 'media' in representation_ms_info
:
2951 media_template
= prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2952 media_location_key
= location_key(media_template
)
2954 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2955 # can't be used at the same time
2956 if '%(Number' in media_template
and 's' not in representation_ms_info
:
2957 segment_duration
= None
2958 if 'total_number' not in representation_ms_info
and 'segment_duration' in representation_ms_info
:
2959 segment_duration
= float_or_none(representation_ms_info
['segment_duration'], representation_ms_info
['timescale'])
2960 representation_ms_info
['total_number'] = int(math
.ceil(
2961 float_or_none(period_duration
, segment_duration
, default
=0)))
2962 representation_ms_info
['fragments'] = [{
2963 media_location_key
: media_template
% {
2964 'Number': segment_number
,
2965 'Bandwidth': bandwidth
,
2967 'duration': segment_duration
,
2968 } for segment_number
in range(
2969 representation_ms_info
['start_number'],
2970 representation_ms_info
['total_number'] + representation_ms_info
['start_number'])]
2972 # $Number*$ or $Time$ in media template with S list available
2973 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2974 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2975 representation_ms_info
['fragments'] = []
2978 segment_number
= representation_ms_info
['start_number']
2980 def add_segment_url():
2981 segment_url
= media_template
% {
2982 'Time': segment_time
,
2983 'Bandwidth': bandwidth
,
2984 'Number': segment_number
,
2986 representation_ms_info
['fragments'].append({
2987 media_location_key
: segment_url
,
2988 'duration': float_or_none(segment_d
, representation_ms_info
['timescale']),
2991 for num
, s
in enumerate(representation_ms_info
['s']):
2992 segment_time
= s
.get('t') or segment_time
2996 for r
in range(s
.get('r', 0)):
2997 segment_time
+= segment_d
3000 segment_time
+= segment_d
3001 elif 'segment_urls' in representation_ms_info
and 's' in representation_ms_info
:
3003 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
3004 # or any YouTube dashsegments video
3007 timescale
= representation_ms_info
['timescale']
3008 for s
in representation_ms_info
['s']:
3009 duration
= float_or_none(s
['d'], timescale
)
3010 for r
in range(s
.get('r', 0) + 1):
3011 segment_uri
= representation_ms_info
['segment_urls'][segment_index
]
3013 location_key(segment_uri
): segment_uri
,
3014 'duration': duration
,
3017 representation_ms_info
['fragments'] = fragments
3018 elif 'segment_urls' in representation_ms_info
:
3019 # Segment URLs with no SegmentTimeline
3020 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
3021 # https://github.com/ytdl-org/youtube-dl/pull/14844
3023 segment_duration
= float_or_none(
3024 representation_ms_info
['segment_duration'],
3025 representation_ms_info
['timescale']) if 'segment_duration' in representation_ms_info
else None
3026 for segment_url
in representation_ms_info
['segment_urls']:
3028 location_key(segment_url
): segment_url
,
3030 if segment_duration
:
3031 fragment
['duration'] = segment_duration
3032 fragments
.append(fragment
)
3033 representation_ms_info
['fragments'] = fragments
3034 # If there is a fragments key available then we correctly recognized fragmented media.
3035 # Otherwise we will assume unfragmented media with direct access. Technically, such
3036 # assumption is not necessarily correct since we may simply have no support for
3037 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
3038 if 'fragments' in representation_ms_info
:
3040 # NB: mpd_url may be empty when MPD manifest is parsed from a string
3041 'url': mpd_url
or base_url
,
3042 'fragment_base_url': base_url
,
3044 'protocol': 'http_dash_segments' if mime_type
!= 'image/jpeg' else 'mhtml',
3046 if 'initialization_url' in representation_ms_info
:
3047 initialization_url
= representation_ms_info
['initialization_url']
3048 if not f
.get('url'):
3049 f
['url'] = initialization_url
3050 f
['fragments'].append({location_key(initialization_url): initialization_url}
)
3051 f
['fragments'].extend(representation_ms_info
['fragments'])
3052 if not period_duration
:
3053 period_duration
= try_get(
3054 representation_ms_info
,
3055 lambda r
: sum(frag
['duration'] for frag
in r
['fragments']), float)
3057 # Assuming direct URL to unfragmented media.
3059 if content_type
in ('video', 'audio', 'image/jpeg'):
3060 f
['manifest_stream_number'] = stream_numbers
[f
['url']]
3061 stream_numbers
[f
['url']] += 1
3063 elif content_type
== 'text':
3064 subtitles
.setdefault(lang
or 'und', []).append(f
)
3066 return formats
, subtitles
3068 def _extract_ism_formats(self
, *args
, **kwargs
):
3069 fmts
, subs
= self
._extract
_ism
_formats
_and
_subtitles
(*args
, **kwargs
)
3071 self
._report
_ignoring
_subs
('ISM')
3074 def _extract_ism_formats_and_subtitles(self
, ism_url
, video_id
, ism_id
=None, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
):
3075 res
= self
._download
_xml
_handle
(
3077 note
='Downloading ISM manifest' if note
is None else note
,
3078 errnote
='Failed to download ISM manifest' if errnote
is None else errnote
,
3079 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
3086 return self
._parse
_ism
_formats
_and
_subtitles
(ism_doc
, urlh
.geturl(), ism_id
)
3088 def _parse_ism_formats_and_subtitles(self
, ism_doc
, ism_url
, ism_id
=None):
3090 Parse formats from ISM manifest.
3092 1. [MS-SSTR]: Smooth Streaming Protocol,
3093 https://msdn.microsoft.com/en-us/library/ff469518.aspx
3095 if ism_doc
.get('IsLive') == 'TRUE':
3098 duration
= int(ism_doc
.attrib
['Duration'])
3099 timescale
= int_or_none(ism_doc
.get('TimeScale')) or 10000000
3103 for stream
in ism_doc
.findall('StreamIndex'):
3104 stream_type
= stream
.get('Type')
3105 if stream_type
not in ('video', 'audio', 'text'):
3107 url_pattern
= stream
.attrib
['Url']
3108 stream_timescale
= int_or_none(stream
.get('TimeScale')) or timescale
3109 stream_name
= stream
.get('Name')
3110 stream_language
= stream
.get('Language', 'und')
3111 for track
in stream
.findall('QualityLevel'):
3112 fourcc
= track
.get('FourCC') or ('AACL' if track
.get('AudioTag') == '255' else None)
3113 # TODO: add support for WVC1 and WMAP
3114 if fourcc
not in ('H264', 'AVC1', 'AACL', 'TTML'):
3115 self
.report_warning('%s is not a supported codec' % fourcc
)
3117 tbr
= int(track
.attrib
['Bitrate']) // 1000
3118 # [1] does not mention Width and Height attributes. However,
3119 # they're often present while MaxWidth and MaxHeight are
3120 # missing, so should be used as fallbacks
3121 width
= int_or_none(track
.get('MaxWidth') or track
.get('Width'))
3122 height
= int_or_none(track
.get('MaxHeight') or track
.get('Height'))
3123 sampling_rate
= int_or_none(track
.get('SamplingRate'))
3125 track_url_pattern
= re
.sub(r
'{[Bb]itrate}', track
.attrib
['Bitrate'], url_pattern
)
3126 track_url_pattern
= urllib
.parse
.urljoin(ism_url
, track_url_pattern
)
3132 stream_fragments
= stream
.findall('c')
3133 for stream_fragment_index
, stream_fragment
in enumerate(stream_fragments
):
3134 fragment_ctx
['time'] = int_or_none(stream_fragment
.get('t')) or fragment_ctx
['time']
3135 fragment_repeat
= int_or_none(stream_fragment
.get('r')) or 1
3136 fragment_ctx
['duration'] = int_or_none(stream_fragment
.get('d'))
3137 if not fragment_ctx
['duration']:
3139 next_fragment_time
= int(stream_fragment
[stream_fragment_index
+ 1].attrib
['t'])
3141 next_fragment_time
= duration
3142 fragment_ctx
['duration'] = (next_fragment_time
- fragment_ctx
['time']) / fragment_repeat
3143 for _
in range(fragment_repeat
):
3145 'url': re
.sub(r
'{start[ _]time}', str(fragment_ctx
['time']), track_url_pattern
),
3146 'duration': fragment_ctx
['duration'] / stream_timescale
,
3148 fragment_ctx
['time'] += fragment_ctx
['duration']
3150 if stream_type
== 'text':
3151 subtitles
.setdefault(stream_language
, []).append({
3155 'manifest_url': ism_url
,
3156 'fragments': fragments
,
3157 '_download_params': {
3158 'stream_type': stream_type
,
3159 'duration': duration
,
3160 'timescale': stream_timescale
,
3162 'language': stream_language
,
3163 'codec_private_data': track
.get('CodecPrivateData'),
3166 elif stream_type
in ('video', 'audio'):
3168 'format_id': join_nonempty(ism_id
, stream_name
, tbr
),
3170 'manifest_url': ism_url
,
3171 'ext': 'ismv' if stream_type
== 'video' else 'isma',
3175 'asr': sampling_rate
,
3176 'vcodec': 'none' if stream_type
== 'audio' else fourcc
,
3177 'acodec': 'none' if stream_type
== 'video' else fourcc
,
3179 'fragments': fragments
,
3180 'has_drm': ism_doc
.find('Protection') is not None,
3181 '_download_params': {
3182 'stream_type': stream_type
,
3183 'duration': duration
,
3184 'timescale': stream_timescale
,
3185 'width': width
or 0,
3186 'height': height
or 0,
3188 'language': stream_language
,
3189 'codec_private_data': track
.get('CodecPrivateData'),
3190 'sampling_rate': sampling_rate
,
3191 'channels': int_or_none(track
.get('Channels', 2)),
3192 'bits_per_sample': int_or_none(track
.get('BitsPerSample', 16)),
3193 'nal_unit_length_field': int_or_none(track
.get('NALUnitLengthField', 4)),
3196 return formats
, subtitles
3198 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):
3199 def absolute_url(item_url
):
3200 return urljoin(base_url
, item_url
)
3202 def parse_content_type(content_type
):
3203 if not content_type
:
3205 ctr
= re
.search(r
'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type
)
3207 mimetype
, codecs
= ctr
.groups()
3208 f
= parse_codecs(codecs
)
3209 f
['ext'] = mimetype2ext(mimetype
)
3213 def _media_formats(src
, cur_media_type
, type_info
=None):
3214 type_info
= type_info
or {}
3215 full_url
= absolute_url(src
)
3216 ext
= type_info
.get('ext') or determine_ext(full_url
)
3218 is_plain_url
= False
3219 formats
= self
._extract
_m
3u8_formats
(
3220 full_url
, video_id
, ext
='mp4',
3221 entry_protocol
=m3u8_entry_protocol
, m3u8_id
=m3u8_id
,
3222 preference
=preference
, quality
=quality
, fatal
=False)
3224 is_plain_url
= False
3225 formats
= self
._extract
_mpd
_formats
(
3226 full_url
, video_id
, mpd_id
=mpd_id
, fatal
=False)
3231 'vcodec': 'none' if cur_media_type
== 'audio' else None,
3234 return is_plain_url
, formats
3237 # amp-video and amp-audio are very similar to their HTML5 counterparts
3238 # so we will include them right here (see
3239 # https://www.ampproject.org/docs/reference/components/amp-video)
3240 # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/
3241 _MEDIA_TAG_NAME_RE
= r
'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)'
3242 media_tags
= [(media_tag
, media_tag_name
, media_type
, '')
3243 for media_tag
, media_tag_name
, media_type
3244 in re
.findall(r
'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE
, webpage
)]
3245 media_tags
.extend(re
.findall(
3246 # We only allow video|audio followed by a whitespace or '>'.
3247 # Allowing more characters may end up in significant slow down (see
3248 # https://github.com/ytdl-org/youtube-dl/issues/11979, example URL:
3249 # http://www.porntrex.com/maps/videositemap.xml).
3250 r
'(?s)(<(?P<tag>%s)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>' % _MEDIA_TAG_NAME_RE
, webpage
))
3251 for media_tag
, _
, media_type
, media_content
in media_tags
:
3256 media_attributes
= extract_attributes(media_tag
)
3257 src
= strip_or_none(media_attributes
.get('src'))
3259 f
= parse_content_type(media_attributes
.get('type'))
3260 _
, formats
= _media_formats(src
, media_type
, f
)
3261 media_info
['formats'].extend(formats
)
3262 media_info
['thumbnail'] = absolute_url(media_attributes
.get('poster'))
3264 for source_tag
in re
.findall(r
'<source[^>]+>', media_content
):
3265 s_attr
= extract_attributes(source_tag
)
3266 # data-video-src and data-src are non standard but seen
3267 # several times in the wild
3268 src
= strip_or_none(dict_get(s_attr
, ('src', 'data-video-src', 'data-src')))
3271 f
= parse_content_type(s_attr
.get('type'))
3272 is_plain_url
, formats
= _media_formats(src
, media_type
, f
)
3274 # width, height, res, label and title attributes are
3275 # all not standard but seen several times in the wild
3278 for lbl
in ('label', 'title')
3279 if str_or_none(s_attr
.get(lbl
))
3281 width
= int_or_none(s_attr
.get('width'))
3282 height
= (int_or_none(s_attr
.get('height'))
3283 or int_or_none(s_attr
.get('res')))
3284 if not width
or not height
:
3286 resolution
= parse_resolution(lbl
)
3289 width
= width
or resolution
.get('width')
3290 height
= height
or resolution
.get('height')
3292 tbr
= parse_bitrate(lbl
)
3301 'format_id': s_attr
.get('label') or s_attr
.get('title'),
3303 f
.update(formats
[0])
3304 media_info
['formats'].append(f
)
3306 media_info
['formats'].extend(formats
)
3307 for track_tag
in re
.findall(r
'<track[^>]+>', media_content
):
3308 track_attributes
= extract_attributes(track_tag
)
3309 kind
= track_attributes
.get('kind')
3310 if not kind
or kind
in ('subtitles', 'captions'):
3311 src
= strip_or_none(track_attributes
.get('src'))
3314 lang
= track_attributes
.get('srclang') or track_attributes
.get('lang') or track_attributes
.get('label')
3315 media_info
['subtitles'].setdefault(lang
, []).append({
3316 'url': absolute_url(src
),
3318 for f
in media_info
['formats']:
3319 f
.setdefault('http_headers', {})['Referer'] = base_url
3320 if media_info
['formats'] or media_info
['subtitles']:
3321 entries
.append(media_info
)
3324 def _extract_akamai_formats(self
, *args
, **kwargs
):
3325 fmts
, subs
= self
._extract
_akamai
_formats
_and
_subtitles
(*args
, **kwargs
)
3327 self
._report
_ignoring
_subs
('akamai')
3330 def _extract_akamai_formats_and_subtitles(self
, manifest_url
, video_id
, hosts
={}):
3331 signed
= 'hdnea=' in manifest_url
3333 # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
3334 manifest_url
= re
.sub(
3335 r
'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
3336 '', manifest_url
).strip('?')
3341 hdcore_sign
= 'hdcore=3.7.0'
3342 f4m_url
= re
.sub(r
'(https?://[^/]+)/i/', r
'\1/z/', manifest_url
).replace('/master.m3u8', '/manifest.f4m')
3343 hds_host
= hosts
.get('hds')
3345 f4m_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hds_host
, f4m_url
)
3346 if 'hdcore=' not in f4m_url
:
3347 f4m_url
+= ('&' if '?' in f4m_url
else '?') + hdcore_sign
3348 f4m_formats
= self
._extract
_f
4m
_formats
(
3349 f4m_url
, video_id
, f4m_id
='hds', fatal
=False)
3350 for entry
in f4m_formats
:
3351 entry
.update({'extra_param_to_segment_url': hdcore_sign}
)
3352 formats
.extend(f4m_formats
)
3354 m3u8_url
= re
.sub(r
'(https?://[^/]+)/z/', r
'\1/i/', manifest_url
).replace('/manifest.f4m', '/master.m3u8')
3355 hls_host
= hosts
.get('hls')
3357 m3u8_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hls_host
, m3u8_url
)
3358 m3u8_formats
, m3u8_subtitles
= self
._extract
_m
3u8_formats
_and
_subtitles
(
3359 m3u8_url
, video_id
, 'mp4', 'm3u8_native',
3360 m3u8_id
='hls', fatal
=False)
3361 formats
.extend(m3u8_formats
)
3362 subtitles
= self
._merge
_subtitles
(subtitles
, m3u8_subtitles
)
3364 http_host
= hosts
.get('http')
3365 if http_host
and m3u8_formats
and not signed
:
3366 REPL_REGEX
= r
'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
3367 qualities
= re
.match(REPL_REGEX
, m3u8_url
).group(2).split(',')
3368 qualities_length
= len(qualities
)
3369 if len(m3u8_formats
) in (qualities_length
, qualities_length
+ 1):
3371 for f
in m3u8_formats
:
3372 if f
['vcodec'] != 'none':
3373 for protocol
in ('http', 'https'):
3375 del http_f
['manifest_url']
3377 REPL_REGEX
, protocol
+ fr
'://{http_host}/\g<1>{qualities[i]}\3', f
['url'])
3379 'format_id': http_f
['format_id'].replace('hls-', protocol
+ '-'),
3381 'protocol': protocol
,
3383 formats
.append(http_f
)
3386 return formats
, subtitles
3388 def _extract_wowza_formats(self
, url
, video_id
, m3u8_entry_protocol
='m3u8_native', skip_protocols
=[]):
3389 query
= urllib
.parse
.urlparse(url
).query
3390 url
= re
.sub(r
'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url
)
3392 r
'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url
)
3393 url_base
= mobj
.group('url')
3394 http_base_url
= '%s%s:%s' % ('http', mobj
.group('s') or '', url_base
)
3397 def manifest_url(manifest
):
3398 m_url
= f
'{http_base_url}/{manifest}'
3400 m_url
+= '?%s' % query
3403 if 'm3u8' not in skip_protocols
:
3404 formats
.extend(self
._extract
_m
3u8_formats
(
3405 manifest_url('playlist.m3u8'), video_id
, 'mp4',
3406 m3u8_entry_protocol
, m3u8_id
='hls', fatal
=False))
3407 if 'f4m' not in skip_protocols
:
3408 formats
.extend(self
._extract
_f
4m
_formats
(
3409 manifest_url('manifest.f4m'),
3410 video_id
, f4m_id
='hds', fatal
=False))
3411 if 'dash' not in skip_protocols
:
3412 formats
.extend(self
._extract
_mpd
_formats
(
3413 manifest_url('manifest.mpd'),
3414 video_id
, mpd_id
='dash', fatal
=False))
3415 if re
.search(r
'(?:/smil:|\.smil)', url_base
):
3416 if 'smil' not in skip_protocols
:
3417 rtmp_formats
= self
._extract
_smil
_formats
(
3418 manifest_url('jwplayer.smil'),
3419 video_id
, fatal
=False)
3420 for rtmp_format
in rtmp_formats
:
3421 rtsp_format
= rtmp_format
.copy()
3422 rtsp_format
['url'] = '%s/%s' % (rtmp_format
['url'], rtmp_format
['play_path'])
3423 del rtsp_format
['play_path']
3424 del rtsp_format
['ext']
3425 rtsp_format
.update({
3426 'url': rtsp_format
['url'].replace('rtmp://', 'rtsp://'),
3427 'format_id': rtmp_format
['format_id'].replace('rtmp', 'rtsp'),
3430 formats
.extend([rtmp_format
, rtsp_format
])
3432 for protocol
in ('rtmp', 'rtsp'):
3433 if protocol
not in skip_protocols
:
3435 'url': f
'{protocol}:{url_base}',
3436 'format_id': protocol
,
3437 'protocol': protocol
,
3441 def _find_jwplayer_data(self
, webpage
, video_id
=None, transform_source
=js_to_json
):
3443 r
'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P
=quote
)\
)(?
!</script
>).*?\
.setup\s
*\
((?P
<options
>[^
)]+)\
)',
3447 jwplayer_data = self._parse_json(mobj.group('options
'),
3449 transform_source=transform_source)
3450 except ExtractorError:
3453 if isinstance(jwplayer_data, dict):
3454 return jwplayer_data
3456 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
3457 jwplayer_data = self._find_jwplayer_data(
3458 webpage, video_id, transform_source=js_to_json)
3459 return self._parse_jwplayer_data(
3460 jwplayer_data, video_id, *args, **kwargs)
3462 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
3463 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3464 # JWPlayer backward compatibility: flattened playlists
3465 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
3466 if 'playlist
' not in jwplayer_data:
3467 jwplayer_data = {'playlist': [jwplayer_data]}
3471 # JWPlayer backward compatibility: single playlist item
3472 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
3473 if not isinstance(jwplayer_data['playlist
'], list):
3474 jwplayer_data['playlist
'] = [jwplayer_data['playlist
']]
3476 for video_data in jwplayer_data['playlist
']:
3477 # JWPlayer backward compatibility: flattened sources
3478 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
3479 if 'sources
' not in video_data:
3480 video_data['sources
'] = [video_data]
3482 this_video_id = video_id or video_data['mediaid
']
3484 formats = self._parse_jwplayer_formats(
3485 video_data['sources
'], video_id=this_video_id, m3u8_id=m3u8_id,
3486 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
3489 tracks = video_data.get('tracks
')
3490 if tracks and isinstance(tracks, list):
3491 for track in tracks:
3492 if not isinstance(track, dict):
3494 track_kind = track.get('kind
')
3495 if not track_kind or not isinstance(track_kind, str):
3497 if track_kind.lower() not in ('captions
', 'subtitles
'):
3499 track_url = urljoin(base_url, track.get('file'))
3502 subtitles.setdefault(track.get('label
') or 'en
', []).append({
3503 'url
': self._proto_relative_url(track_url)
3507 'id': this_video_id,
3508 'title
': unescapeHTML(video_data['title
'] if require_title else video_data.get('title
')),
3509 'description
': clean_html(video_data.get('description
')),
3510 'thumbnail
': urljoin(base_url, self._proto_relative_url(video_data.get('image
'))),
3511 'timestamp
': int_or_none(video_data.get('pubdate
')),
3512 'duration
': float_or_none(jwplayer_data.get('duration
') or video_data.get('duration
')),
3513 'subtitles
': subtitles,
3515 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
3516 if len(formats) == 1 and re.search(r'^
(?
:http|
//).*(?
:youtube\
.com|youtu\
.be
)/.+', formats[0]['url
']):
3518 '_type
': 'url_transparent
',
3519 'url
': formats[0]['url
'],
3522 self._sort_formats(formats)
3523 entry['formats
'] = formats
3524 entries.append(entry)
3525 if len(entries) == 1:
3528 return self.playlist_result(entries)
3530 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
3531 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3534 for source in jwplayer_sources_data:
3535 if not isinstance(source, dict):
3537 source_url = urljoin(
3538 base_url, self._proto_relative_url(source.get('file')))
3539 if not source_url or source_url in urls:
3541 urls.append(source_url)
3542 source_type = source.get('type') or ''
3543 ext = mimetype2ext(source_type) or determine_ext(source_url)
3544 if source_type == 'hls
' or ext == 'm3u8
':
3545 formats.extend(self._extract_m3u8_formats(
3546 source_url, video_id, 'mp4
', entry_protocol='m3u8_native
',
3547 m3u8_id=m3u8_id, fatal=False))
3548 elif source_type == 'dash
' or ext == 'mpd
':
3549 formats.extend(self._extract_mpd_formats(
3550 source_url, video_id, mpd_id=mpd_id, fatal=False))
3552 formats.extend(self._extract_smil_formats(
3553 source_url, video_id, fatal=False))
3554 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3555 elif source_type.startswith('audio
') or ext in (
3556 'oga
', 'aac
', 'mp3
', 'mpeg
', 'vorbis
'):
3563 height = int_or_none(source.get('height
'))
3565 # Often no height is provided but there is a label in
3566 # format like "1080p", "720p SD", or 1080.
3567 height = int_or_none(self._search_regex(
3568 r'^
(\d{3,4}
)[pP
]?
(?
:\b|$
)', str(source.get('label
') or ''),
3569 'height
', default=None))
3572 'width
': int_or_none(source.get('width
')),
3574 'tbr
': int_or_none(source.get('bitrate
')),
3577 if source_url.startswith('rtmp
'):
3578 a_format['ext
'] = 'flv
'
3579 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3580 # of jwplayer.flash.swf
3581 rtmp_url_parts = re.split(
3582 r'((?
:mp4|mp3|flv
):)', source_url, 1)
3583 if len(rtmp_url_parts) == 3:
3584 rtmp_url, prefix, play_path = rtmp_url_parts
3587 'play_path
': prefix + play_path,
3590 a_format.update(rtmp_params)
3591 formats.append(a_format)
3594 def _live_title(self, name):
3595 self._downloader.deprecation_warning('yt_dlp
.InfoExtractor
._live
_title
is deprecated
and does
not work
as expected
')
3598 def _int(self, v, name, fatal=False, **kwargs):
3599 res = int_or_none(v, **kwargs)
3601 msg = f'Failed to extract {name}
: Could
not parse value {v!r}
'
3603 raise ExtractorError(msg)
3605 self.report_warning(msg)
3608 def _float(self, v, name, fatal=False, **kwargs):
3609 res = float_or_none(v, **kwargs)
3611 msg = f'Failed to extract {name}
: Could
not parse value {v!r}
'
3613 raise ExtractorError(msg)
3615 self.report_warning(msg)
3618 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
3619 path='/', secure=False, discard=False, rest={}, **kwargs):
3620 cookie = http.cookiejar.Cookie(
3621 0, name, value, port, port is not None, domain, True,
3622 domain.startswith('.'), path, True, secure, expire_time,
3623 discard, None, None, rest)
3624 self.cookiejar.set_cookie(cookie)
3626 def _get_cookies(self, url):
3627 """ Return a http.cookies.SimpleCookie with the cookies for the url """
3628 return http.cookies.SimpleCookie(self._downloader._calc_cookies(url))
3630 def _apply_first_set_cookie_header(self, url_handle, cookie):
3632 Apply first Set-Cookie header instead of the last. Experimental.
3634 Some sites (e.g. [1-3]) may serve two cookies under the same name
3635 in Set-Cookie header and expect the first (old) one to be set rather
3636 than second (new). However, as of RFC6265 the newer one cookie
3637 should be set into cookie store what actually happens.
3638 We will workaround this issue by resetting the cookie to
3639 the first one manually.
3640 1. https://new.vk.com/
3641 2. https://github.com/ytdl-org/youtube-dl/issues/9841#issuecomment-227871201
3642 3. https://learning.oreilly.com/
3644 for header, cookies in url_handle.headers.items():
3645 if header.lower() != 'set-cookie
':
3647 cookies = cookies.encode('iso
-8859-1').decode('utf
-8')
3648 cookie_value = re.search(
3649 r'%s=(.+?
);.*?
\b[Dd
]omain
=(.+?
)(?
:[,;]|$
)' % cookie, cookies)
3651 value, domain = cookie_value.groups()
3652 self._set_cookie(domain, cookie, value)
3656 def get_testcases(cls, include_onlymatching=False):
3657 t = getattr(cls, '_TEST
', None)
3659 assert not hasattr(cls, '_TESTS
'), f'{cls.ie_key()}IE has _TEST
and _TESTS
'
3662 tests = getattr(cls, '_TESTS
', [])
3664 if not include_onlymatching and t.get('only_matching
', False):
3666 t['name
'] = cls.ie_key()
3670 def get_webpage_testcases(cls):
3671 tests = getattr(cls, '_WEBPAGE_TESTS
', [])
3673 t['name
'] = cls.ie_key()
3678 """Get age limit from the testcases"""
3679 return max(traverse_obj(
3680 (*cls.get_testcases(include_onlymatching=False), *cls.get_webpage_testcases()),
3681 (..., (('playlist
', 0), None), 'info_dict
', 'age_limit
')) or [0])
3684 def is_suitable(cls, age_limit):
3685 """Test whether the extractor is generally suitable for the given age limit"""
3686 return not age_restricted(cls.age_limit, age_limit)
3689 def description(cls, *, markdown=True, search_examples=None):
3690 """Description of the extractor"""
3692 if cls._NETRC_MACHINE:
3694 desc += f' [<abbr title
="netrc machine"><em
>{cls._NETRC_MACHINE}
</em
></abbr
>]'
3696 desc += f' [{cls._NETRC_MACHINE}
]'
3697 if cls.IE_DESC is False:
3700 desc += f' {cls.IE_DESC}
'
3702 desc += f'; "{cls.SEARCH_KEY}:" prefix
'
3704 _COUNTS = ('', '5', '10', 'all
')
3705 desc += f' (Example
: "{cls.SEARCH_KEY}{random.choice(_COUNTS)}:{random.choice(search_examples)}")'
3706 if not cls.working():
3707 desc += ' (**Currently broken
**)' if markdown else ' (Currently broken
)'
3709 name = f' - **{cls.IE_NAME}
**' if markdown else cls.IE_NAME
3710 return f'{name}
:{desc}
' if desc else name
3712 def extract_subtitles(self, *args, **kwargs):
3713 if (self.get_param('writesubtitles
', False)
3714 or self.get_param('listsubtitles
')):
3715 return self._get_subtitles(*args, **kwargs)
3718 def _get_subtitles(self, *args, **kwargs):
3719 raise NotImplementedError('This method must be implemented by subclasses
')
3721 def extract_comments(self, *args, **kwargs):
3722 if not self.get_param('getcomments
'):
3724 generator = self._get_comments(*args, **kwargs)
3731 comments.append(next(generator))
3732 except StopIteration:
3734 except KeyboardInterrupt:
3735 self.to_screen('Interrupted by user
')
3736 except Exception as e:
3737 if self.get_param('ignoreerrors
') is not True:
3739 self._downloader.report_error(e)
3740 comment_count = len(comments)
3741 self.to_screen(f'Extracted {comment_count} comments
')
3743 'comments
': comments,
3744 'comment_count
': None if interrupted else comment_count
3748 def _get_comments(self, *args, **kwargs):
3749 raise NotImplementedError('This method must be implemented by subclasses
')
3752 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
3753 """ Merge subtitle items for one language. Items with duplicated URLs/data
3754 will be dropped. """
3755 list1_data = {(item.get('url'), item.get('data')) for item in subtitle_list1}
3756 ret = list(subtitle_list1)
3757 ret.extend(item for item in subtitle_list2 if (item.get('url
'), item.get('data
')) not in list1_data)
3761 def _merge_subtitles(cls, *dicts, target=None):
3762 """ Merge subtitle dictionaries, language by language. """
3766 for lang, subs in d.items():
3767 target[lang] = cls._merge_subtitle_items(target.get(lang, []), subs)
3770 def extract_automatic_captions(self, *args, **kwargs):
3771 if (self.get_param('writeautomaticsub
', False)
3772 or self.get_param('listsubtitles
')):
3773 return self._get_automatic_captions(*args, **kwargs)
3776 def _get_automatic_captions(self, *args, **kwargs):
3777 raise NotImplementedError('This method must be implemented by subclasses
')
3779 @functools.cached_property
3780 def _cookies_passed(self):
3781 """Whether cookies have been passed to YoutubeDL"""
3782 return self.get_param('cookiefile
') is not None or self.get_param('cookiesfrombrowser
') is not None
3784 def mark_watched(self, *args, **kwargs):
3785 if not self.get_param('mark_watched
', False):
3787 if self.supports_login() and self._get_login_info()[0] is not None or self._cookies_passed:
3788 self._mark_watched(*args, **kwargs)
3790 def _mark_watched(self, *args, **kwargs):
3791 raise NotImplementedError('This method must be implemented by subclasses
')
3793 def geo_verification_headers(self):
3795 geo_verification_proxy = self.get_param('geo_verification_proxy
')
3796 if geo_verification_proxy:
3797 headers['Ytdl
-request
-proxy
'] = geo_verification_proxy
3801 def _generic_id(url):
3802 return urllib.parse.unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
3805 def _generic_title(url):
3806 return urllib.parse.unquote(os.path.splitext(url_basename(url))[0])
3809 def _availability(is_private=None, needs_premium=None, needs_subscription=None, needs_auth=None, is_unlisted=None):
3810 all_known = all(map(
3811 lambda x: x is not None,
3812 (is_private, needs_premium, needs_subscription, needs_auth, is_unlisted)))
3814 'private
' if is_private
3815 else 'premium_only
' if needs_premium
3816 else 'subscriber_only
' if needs_subscription
3817 else 'needs_auth
' if needs_auth
3818 else 'unlisted
' if is_unlisted
3819 else 'public
' if all_known
3822 def _configuration_arg(self, key, default=NO_DEFAULT, *, ie_key=None, casesense=False):
3824 @returns A list of values for the extractor argument given by "key"
3825 or "default" if no such key is present
3826 @param default The default value to return when the key is not present (default: [])
3827 @param casesense When false, the values are converted to lower case
3830 self._downloader.params, ('extractor_args
', (ie_key or self.ie_key()).lower(), key))
3832 return [] if default is NO_DEFAULT else default
3833 return list(val) if casesense else [x.lower() for x in val]
3835 def _yes_playlist(self, playlist_id, video_id, smuggled_data=None, *, playlist_label='playlist
', video_label='video
'):
3836 if not playlist_id or not video_id:
3839 no_playlist = (smuggled_data or {}).get('force_noplaylist
')
3840 if no_playlist is not None:
3841 return not no_playlist
3843 video_id = '' if video_id is True else f' {video_id}
'
3844 playlist_id = '' if playlist_id is True else f' {playlist_id}
'
3845 if self.get_param('noplaylist
'):
3846 self.to_screen(f'Downloading just the {video_label}{video_id} because of
--no
-playlist
')
3848 self.to_screen(f'Downloading {playlist_label}{playlist_id}
- add
--no
-playlist to download just the {video_label}{video_id}
')
3852 def extract_from_webpage(cls, ydl, url, webpage):
3853 ie = (cls if isinstance(cls._extract_from_webpage, types.MethodType)
3854 else ydl.get_info_extractor(cls.ie_key()))
3855 for info in ie._extract_from_webpage(url, webpage) or []:
3856 # url = None since we do not want to set (webpage/original)_url
3857 ydl.add_default_extra_info(info, ie, None)
3861 def _extract_from_webpage(cls, url, webpage):
3862 for embed_url in orderedSet(
3863 cls._extract_embed_urls(url, webpage) or [], lazy=True):
3864 yield cls.url_result(embed_url, cls)
3867 def _extract_embed_urls(cls, url, webpage):
3868 """@returns all the embed urls on the webpage"""
3869 if '_EMBED_URL_RE
' not in cls.__dict__:
3870 assert isinstance(cls._EMBED_REGEX, (list, tuple))
3871 for idx, regex in enumerate(cls._EMBED_REGEX):
3872 assert regex.count('(?P
<url
>') == 1, \
3873 f'{cls.__name__}
._EMBED
_REGEX
[{idx}
] must have exactly
1 url group
\n\t{regex}
'
3874 cls._EMBED_URL_RE = tuple(map(re.compile, cls._EMBED_REGEX))
3876 for regex in cls._EMBED_URL_RE:
3877 for mobj in regex.finditer(webpage):
3878 embed_url = urllib.parse.urljoin(url, unescapeHTML(mobj.group('url
')))
3879 if cls._VALID_URL is False or cls.suitable(embed_url):
3882 class StopExtraction(Exception):
3886 def _extract_url(cls, webpage): # TODO: Remove
3887 """Only for compatibility with some older extractors"""
3888 return next(iter(cls._extract_embed_urls(None, webpage) or []), None)
3891 class SearchInfoExtractor(InfoExtractor):
3893 Base class for paged search queries extractors.
3894 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
3895 Instances should define _SEARCH_KEY and optionally _MAX_RESULTS
3898 _MAX_RESULTS = float('inf
')
3901 def _VALID_URL(cls):
3902 return r'%s(?P
<prefix
>|
[1-9][0-9]*|all
):(?P
<query
>[\s\S
]+)' % cls._SEARCH_KEY
3904 def _real_extract(self, query):
3905 prefix, query = self._match_valid_url(query).group('prefix
', 'query
')
3907 return self._get_n_results(query, 1)
3908 elif prefix == 'all
':
3909 return self._get_n_results(query, self._MAX_RESULTS)
3913 raise ExtractorError(f'invalid download number {n}
for query
"{query}"')
3914 elif n > self._MAX_RESULTS:
3915 self.report_warning('%s returns
max %i results (you requested
%i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3916 n = self._MAX_RESULTS
3917 return self._get_n_results(query, n)
3919 def _get_n_results(self, query, n):
3920 """Get a specified number of results for a query.
3921 Either this function or _search_results must be overridden by subclasses """
3922 return self.playlist_result(
3923 itertools.islice(self._search_results(query), 0, None if n == float('inf
') else n),
3926 def _search_results(self, query):
3927 """Returns an iterator of search results"""
3928 raise NotImplementedError('This method must be implemented by subclasses
')
3931 def SEARCH_KEY(cls):
3932 return cls._SEARCH_KEY