21 import xml
.etree
.ElementTree
23 from ..compat
import functools
# isort: split
24 from ..compat
import compat_etree_fromstring
, compat_expanduser
, compat_os_name
25 from ..cookies
import LenientSimpleCookie
26 from ..downloader
import FileDownloader
27 from ..downloader
.f4m
import get_base_url
, remove_encrypted_media
64 parse_m3u8_attributes
,
91 """Information Extractor class.
93 Information extractors are the classes that, given a URL, extract
94 information about the video (or videos) the URL refers to. This
95 information includes the real video URL, the video title, author and
96 others. The information is stored in a dictionary which is then
97 passed to the YoutubeDL. The YoutubeDL processes this
98 information possibly downloading the video to the file system, among
99 other possible outcomes.
101 The type field determines the type of the result.
102 By far the most common value (and the default if _type is missing) is
103 "video", which indicates a single video.
105 For a video, the dictionaries must include the following fields:
107 id: Video identifier.
108 title: Video title, unescaped. Set to an empty string if video has
109 no title as opposed to "None" which signifies that the
110 extractor failed to obtain a title
112 Additionally, it must contain either a formats entry or a url one:
114 formats: A list of dictionaries for each format available, ordered
115 from worst to best quality.
118 * url The mandatory URL representing the media:
119 for plain file media - HTTP URL of this file,
121 for HLS - URL of the M3U8 media playlist,
122 for HDS - URL of the F4M manifest,
124 - HTTP URL to plain file media (in case of
126 - URL of the MPD manifest or base URL
127 representing the media if MPD manifest
128 is parsed from a string (in case of
130 for MSS - URL of the ISM manifest.
132 The URL of the manifest file in case of
134 for HLS - URL of the M3U8 master playlist,
135 for HDS - URL of the F4M manifest,
136 for DASH - URL of the MPD manifest,
137 for MSS - URL of the ISM manifest.
138 * manifest_stream_number (For internal use only)
139 The index of the stream in the manifest file
140 * ext Will be calculated from URL if missing
141 * format A human-readable description of the format
142 ("mp4 container with h264/opus").
143 Calculated from the format_id, width, height.
144 and format_note fields if missing.
145 * format_id A short description of the format
146 ("mp4_h264_opus" or "19").
147 Technically optional, but strongly recommended.
148 * format_note Additional info about the format
149 ("3D" or "DASH video")
150 * width Width of the video, if known
151 * height Height of the video, if known
152 * resolution Textual description of width and height
153 * dynamic_range The dynamic range of the video. One of:
154 "SDR" (None), "HDR10", "HDR10+, "HDR12", "HLG, "DV"
155 * tbr Average bitrate of audio and video in KBit/s
156 * abr Average audio bitrate in KBit/s
157 * acodec Name of the audio codec in use
158 * asr Audio sampling rate in Hertz
159 * audio_channels Number of audio channels
160 * vbr Average video bitrate in KBit/s
162 * vcodec Name of the video codec in use
163 * container Name of the container format
164 * filesize The number of bytes, if known in advance
165 * filesize_approx An estimate for the number of bytes
166 * player_url SWF Player URL (used for rtmpdump).
167 * protocol The protocol that will be used for the actual
168 download, lower-case. One of "http", "https" or
169 one of the protocols defined in downloader.PROTOCOL_MAP
171 Base URL for fragments. Each fragment's path
172 value (if present) will be relative to
174 * fragments A list of fragments of a fragmented media.
175 Each fragment entry must contain either an url
176 or a path. If an url is present it should be
177 considered by a client. Otherwise both path and
178 fragment_base_url must be present. Here is
179 the list of all potential fields:
180 * "url" - fragment's URL
181 * "path" - fragment's path relative to
183 * "duration" (optional, int or float)
184 * "filesize" (optional, int)
185 * is_from_start Is a live format that can be downloaded
186 from the start. Boolean
187 * preference Order number of this format. If this field is
188 present and not None, the formats get sorted
189 by this field, regardless of all other values.
190 -1 for default (order by other properties),
191 -2 or smaller for less than default.
192 < -1000 to hide the format (if there is
193 another one which is strictly better)
194 * language Language code, e.g. "de" or "en-US".
195 * language_preference Is this in the language mentioned in
197 10 if it's what the URL is about,
198 -1 for default (don't know),
199 -10 otherwise, other values reserved for now.
200 * quality Order number of the video quality of this
201 format, irrespective of the file format.
202 -1 for default (order by other properties),
203 -2 or smaller for less than default.
204 * source_preference Order number for this video source
205 (quality takes higher priority)
206 -1 for default (order by other properties),
207 -2 or smaller for less than default.
208 * http_headers A dictionary of additional HTTP headers
209 to add to the request.
210 * stretched_ratio If given and not 1, indicates that the
211 video's pixels are not square.
212 width : height ratio as float.
213 * no_resume The server does not support resuming the
214 (HTTP or RTMP) download. Boolean.
215 * has_drm The format has DRM and cannot be downloaded. Boolean
216 * downloader_options A dictionary of downloader options
217 (For internal use only)
218 * http_chunk_size Chunk size for HTTP downloads
219 * ffmpeg_args Extra arguments for ffmpeg downloader
220 RTMP formats can also have the additional fields: page_url,
221 app, play_path, tc_url, flash_version, rtmp_live, rtmp_conn,
222 rtmp_protocol, rtmp_real_time
224 url: Final video URL.
225 ext: Video filename extension.
226 format: The video format, defaults to ext (used for --get-format)
227 player_url: SWF Player URL (used for rtmpdump).
229 The following fields are optional:
231 direct: True if a direct video file was given (must only be set by GenericIE)
232 alt_title: A secondary title of the video.
233 display_id An alternative identifier for the video, not necessarily
234 unique, but available before title. Typically, id is
235 something like "4234987", title "Dancing naked mole rats",
236 and display_id "dancing-naked-mole-rats"
237 thumbnails: A list of dictionaries, with the following entries:
238 * "id" (optional, string) - Thumbnail format ID
240 * "preference" (optional, int) - quality of the image
241 * "width" (optional, int)
242 * "height" (optional, int)
243 * "resolution" (optional, string "{width}x{height}",
245 * "filesize" (optional, int)
246 * "http_headers" (dict) - HTTP headers for the request
247 thumbnail: Full URL to a video thumbnail image.
248 description: Full video description.
249 uploader: Full name of the video uploader.
250 license: License name the video is licensed under.
251 creator: The creator of the video.
252 timestamp: UNIX timestamp of the moment the video was uploaded
253 upload_date: Video upload date in UTC (YYYYMMDD).
254 If not explicitly set, calculated from timestamp
255 release_timestamp: UNIX timestamp of the moment the video was released.
256 If it is not clear whether to use timestamp or this, use the former
257 release_date: The date (YYYYMMDD) when the video was released in UTC.
258 If not explicitly set, calculated from release_timestamp
259 modified_timestamp: UNIX timestamp of the moment the video was last modified.
260 modified_date: The date (YYYYMMDD) when the video was last modified in UTC.
261 If not explicitly set, calculated from modified_timestamp
262 uploader_id: Nickname or id of the video uploader.
263 uploader_url: Full URL to a personal webpage of the video uploader.
264 channel: Full name of the channel the video is uploaded on.
265 Note that channel fields may or may not repeat uploader
266 fields. This depends on a particular extractor.
267 channel_id: Id of the channel.
268 channel_url: Full URL to a channel webpage.
269 channel_follower_count: Number of followers of the channel.
270 location: Physical location where the video was filmed.
271 subtitles: The available subtitles as a dictionary in the format
272 {tag: subformats}. "tag" is usually a language code, and
273 "subformats" is a list sorted from lower to higher
274 preference, each element is a dictionary with the "ext"
276 * "data": The subtitles file contents
277 * "url": A URL pointing to the subtitles file
278 It can optionally also have:
279 * "name": Name or description of the subtitles
280 * "http_headers": A dictionary of additional HTTP headers
281 to add to the request.
282 "ext" will be calculated from URL if missing
283 automatic_captions: Like 'subtitles'; contains automatically generated
284 captions instead of normal subtitles
285 duration: Length of the video in seconds, as an integer or float.
286 view_count: How many users have watched the video on the platform.
287 like_count: Number of positive ratings of the video
288 dislike_count: Number of negative ratings of the video
289 repost_count: Number of reposts of the video
290 average_rating: Average rating give by users, the scale used depends on the webpage
291 comment_count: Number of comments on the video
292 comments: A list of comments, each with one or more of the following
293 properties (all but one of text or html optional):
294 * "author" - human-readable name of the comment author
295 * "author_id" - user ID of the comment author
296 * "author_thumbnail" - The thumbnail of the comment author
298 * "html" - Comment as HTML
299 * "text" - Plain text of the comment
300 * "timestamp" - UNIX timestamp of comment
301 * "parent" - ID of the comment this one is replying to.
302 Set to "root" to indicate that this is a
303 comment to the original video.
304 * "like_count" - Number of positive ratings of the comment
305 * "dislike_count" - Number of negative ratings of the comment
306 * "is_favorited" - Whether the comment is marked as
307 favorite by the video uploader
308 * "author_is_uploader" - Whether the comment is made by
310 age_limit: Age restriction for the video, as an integer (years)
311 webpage_url: The URL to the video webpage, if given to yt-dlp it
312 should allow to get the same result again. (It will be set
313 by YoutubeDL if it's missing)
314 categories: A list of categories that the video falls in, for example
316 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
317 cast: A list of the video cast
318 is_live: True, False, or None (=unknown). Whether this video is a
319 live stream that goes on instead of a fixed-length video.
320 was_live: True, False, or None (=unknown). Whether this video was
321 originally a live stream.
322 live_status: None (=unknown), 'is_live', 'is_upcoming', 'was_live', 'not_live',
323 or 'post_live' (was live, but VOD is not yet processed)
324 If absent, automatically set from is_live, was_live
325 start_time: Time in seconds where the reproduction should start, as
326 specified in the URL.
327 end_time: Time in seconds where the reproduction should end, as
328 specified in the URL.
329 chapters: A list of dictionaries, with the following entries:
330 * "start_time" - The start time of the chapter in seconds
331 * "end_time" - The end time of the chapter in seconds
332 * "title" (optional, string)
333 playable_in_embed: Whether this video is allowed to play in embedded
334 players on other sites. Can be True (=always allowed),
335 False (=never allowed), None (=unknown), or a string
336 specifying the criteria for embedability; e.g. 'whitelist'
337 availability: Under what condition the video is available. One of
338 'private', 'premium_only', 'subscriber_only', 'needs_auth',
339 'unlisted' or 'public'. Use 'InfoExtractor._availability'
341 _old_archive_ids: A list of old archive ids needed for backward compatibility
342 __post_extractor: A function to be called just before the metadata is
343 written to either disk, logger or console. The function
344 must return a dict which will be added to the info_dict.
345 This is usefull for additional information that is
346 time-consuming to extract. Note that the fields thus
347 extracted will not be available to output template and
348 match_filter. So, only "comments" and "comment_count" are
349 currently allowed to be extracted via this method.
351 The following fields should only be used when the video belongs to some logical
354 chapter: Name or title of the chapter the video belongs to.
355 chapter_number: Number of the chapter the video belongs to, as an integer.
356 chapter_id: Id of the chapter the video belongs to, as a unicode string.
358 The following fields should only be used when the video is an episode of some
359 series, programme or podcast:
361 series: Title of the series or programme the video episode belongs to.
362 series_id: Id of the series or programme the video episode belongs to, as a unicode string.
363 season: Title of the season the video episode belongs to.
364 season_number: Number of the season the video episode belongs to, as an integer.
365 season_id: Id of the season the video episode belongs to, as a unicode string.
366 episode: Title of the video episode. Unlike mandatory video title field,
367 this field should denote the exact title of the video episode
368 without any kind of decoration.
369 episode_number: Number of the video episode within a season, as an integer.
370 episode_id: Id of the video episode, as a unicode string.
372 The following fields should only be used when the media is a track or a part of
375 track: Title of the track.
376 track_number: Number of the track within an album or a disc, as an integer.
377 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
379 artist: Artist(s) of the track.
380 genre: Genre(s) of the track.
381 album: Title of the album the track belongs to.
382 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
383 album_artist: List of all artists appeared on the album (e.g.
384 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
386 disc_number: Number of the disc or other physical medium the track belongs to,
388 release_year: Year (YYYY) when the album was released.
389 composer: Composer of the piece
391 The following fields should only be set for clips that should be cut from the original video:
393 section_start: Start time of the section in seconds
394 section_end: End time of the section in seconds
396 The following fields should only be set for storyboards:
397 rows: Number of rows in each storyboard fragment, as an integer
398 columns: Number of columns in each storyboard fragment, as an integer
400 Unless mentioned otherwise, the fields should be Unicode strings.
402 Unless mentioned otherwise, None is equivalent to absence of information.
405 _type "playlist" indicates multiple videos.
406 There must be a key "entries", which is a list, an iterable, or a PagedList
407 object, each element of which is a valid dictionary by this specification.
409 Additionally, playlists can have "id", "title", and any other relevant
410 attributes with the same semantics as videos (see above).
412 It can also have the following optional fields:
414 playlist_count: The total number of videos in a playlist. If not given,
415 YoutubeDL tries to calculate it from "entries"
418 _type "multi_video" indicates that there are multiple videos that
419 form a single show, for examples multiple acts of an opera or TV episode.
420 It must have an entries key like a playlist and contain all the keys
421 required for a video at the same time.
424 _type "url" indicates that the video must be extracted from another
425 location, possibly by a different extractor. Its only required key is:
426 "url" - the next URL to extract.
427 The key "ie_key" can be set to the class name (minus the trailing "IE",
428 e.g. "Youtube") if the extractor class is known in advance.
429 Additionally, the dictionary may have any properties of the resolved entity
430 known in advance, for example "title" if the title of the referred video is
434 _type "url_transparent" entities have the same specification as "url", but
435 indicate that the given additional information is more precise than the one
436 associated with the resolved URL.
437 This is useful when a site employs a video service that hosts the video and
438 its technical metadata, but that video service does not embed a useful
439 title, description etc.
442 Subclasses of this should also be added to the list of extractors and
443 should define a _VALID_URL regexp and, re-define the _real_extract() and
444 (optionally) _real_initialize() methods.
446 Subclasses may also override suitable() if necessary, but ensure the function
447 signature is preserved and that this function imports everything it needs
448 (except other extractors), so that lazy_extractors works correctly.
450 Subclasses can define a list of _EMBED_REGEX, which will be searched for in
451 the HTML of Generic webpages. It may also override _extract_embed_urls
452 or _extract_from_webpage as necessary. While these are normally classmethods,
453 _extract_from_webpage is allowed to be an instance method.
455 _extract_from_webpage may raise self.StopExtraction() to stop further
456 processing of the webpage and obtain exclusive rights to it. This is useful
457 when the extractor cannot reliably be matched using just the URL,
458 e.g. invidious/peertube instances
460 Embed-only extractors can be defined by setting _VALID_URL = False.
462 To support username + password (or netrc) login, the extractor must define a
463 _NETRC_MACHINE and re-define _perform_login(username, password) and
464 (optionally) _initialize_pre_login() methods. The _perform_login method will
465 be called between _initialize_pre_login and _real_initialize if credentials
466 are passed by the user. In cases where it is necessary to have the login
467 process as part of the extraction rather than initialization, _perform_login
468 can be left undefined.
470 _GEO_BYPASS attribute may be set to False in order to disable
471 geo restriction bypass mechanisms for a particular extractor.
472 Though it won't disable explicit geo restriction bypass based on
473 country code provided with geo_bypass_country.
475 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
476 countries for this extractor. One of these countries will be used by
477 geo restriction bypass mechanism right away in order to bypass
478 geo restriction, of course, if the mechanism is not disabled.
480 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
481 IP blocks in CIDR notation for this extractor. One of these IP blocks
482 will be used by geo restriction bypass mechanism similarly
485 The _ENABLED attribute should be set to False for IEs that
486 are disabled by default and must be explicitly enabled.
488 The _WORKING attribute should be set to False for broken IEs
489 in order to warn the users and skip the tests.
494 _x_forwarded_for_ip
= None
496 _GEO_COUNTRIES
= None
497 _GEO_IP_BLOCKS
= None
500 _NETRC_MACHINE
= None
506 def _login_hint(self
, method
=NO_DEFAULT
, netrc
=None):
507 password_hint
= f
'--username and --password, or --netrc ({netrc or self._NETRC_MACHINE}) to provide account credentials'
510 'any': f
'Use --cookies, --cookies-from-browser, {password_hint}',
511 'password': f
'Use {password_hint}',
513 'Use --cookies-from-browser or --cookies for the authentication. '
514 'See https://github.com/yt-dlp/yt-dlp/wiki/FAQ#how-do-i-pass-cookies-to-yt-dlp for how to manually pass cookies'),
515 }[method
if method
is not NO_DEFAULT
else 'any' if self
.supports_login() else 'cookies']
517 def __init__(self
, downloader
=None):
518 """Constructor. Receives an optional downloader (a YoutubeDL instance).
519 If a downloader is not passed during initialization,
520 it must be set using "set_downloader()" before "extract()" is called"""
522 self
._x
_forwarded
_for
_ip
= None
523 self
._printed
_messages
= set()
524 self
.set_downloader(downloader
)
527 def _match_valid_url(cls
, url
):
528 if cls
._VALID
_URL
is False:
530 # This does not use has/getattr intentionally - we want to know whether
531 # we have cached the regexp for *this* class, whereas getattr would also
532 # match the superclass
533 if '_VALID_URL_RE' not in cls
.__dict
__:
534 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
535 return cls
._VALID
_URL
_RE
.match(url
)
538 def suitable(cls
, url
):
539 """Receives a URL and returns True if suitable for this IE."""
540 # This function must import everything it needs (except other extractors),
541 # so that lazy_extractors works correctly
542 return cls
._match
_valid
_url
(url
) is not None
545 def _match_id(cls
, url
):
546 return cls
._match
_valid
_url
(url
).group('id')
549 def get_temp_id(cls
, url
):
551 return cls
._match
_id
(url
)
552 except (IndexError, AttributeError):
557 """Getter method for _WORKING."""
561 def supports_login(cls
):
562 return bool(cls
._NETRC
_MACHINE
)
564 def initialize(self
):
565 """Initializes an instance (authentication, etc)."""
566 self
._printed
_messages
= set()
567 self
._initialize
_geo
_bypass
({
568 'countries': self
._GEO
_COUNTRIES
,
569 'ip_blocks': self
._GEO
_IP
_BLOCKS
,
572 self
._initialize
_pre
_login
()
573 if self
.supports_login():
574 username
, password
= self
._get
_login
_info
()
576 self
._perform
_login
(username
, password
)
577 elif self
.get_param('username') and False not in (self
.IE_DESC
, self
._NETRC
_MACHINE
):
578 self
.report_warning(f
'Login with password is not supported for this website. {self._login_hint("cookies")}')
579 self
._real
_initialize
()
582 def _initialize_geo_bypass(self
, geo_bypass_context
):
584 Initialize geo restriction bypass mechanism.
586 This method is used to initialize geo bypass mechanism based on faking
587 X-Forwarded-For HTTP header. A random country from provided country list
588 is selected and a random IP belonging to this country is generated. This
589 IP will be passed as X-Forwarded-For HTTP header in all subsequent
592 This method will be used for initial geo bypass mechanism initialization
593 during the instance initialization with _GEO_COUNTRIES and
596 You may also manually call it from extractor's code if geo bypass
597 information is not available beforehand (e.g. obtained during
598 extraction) or due to some other reason. In this case you should pass
599 this information in geo bypass context passed as first argument. It may
600 contain following fields:
602 countries: List of geo unrestricted countries (similar
604 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
605 (similar to _GEO_IP_BLOCKS)
608 if not self
._x
_forwarded
_for
_ip
:
610 # Geo bypass mechanism is explicitly disabled by user
611 if not self
.get_param('geo_bypass', True):
614 if not geo_bypass_context
:
615 geo_bypass_context
= {}
617 # Backward compatibility: previously _initialize_geo_bypass
618 # expected a list of countries, some 3rd party code may still use
620 if isinstance(geo_bypass_context
, (list, tuple)):
621 geo_bypass_context
= {
622 'countries': geo_bypass_context
,
625 # The whole point of geo bypass mechanism is to fake IP
626 # as X-Forwarded-For HTTP header based on some IP block or
629 # Path 1: bypassing based on IP block in CIDR notation
631 # Explicit IP block specified by user, use it right away
632 # regardless of whether extractor is geo bypassable or not
633 ip_block
= self
.get_param('geo_bypass_ip_block', None)
635 # Otherwise use random IP block from geo bypass context but only
636 # if extractor is known as geo bypassable
638 ip_blocks
= geo_bypass_context
.get('ip_blocks')
639 if self
._GEO
_BYPASS
and ip_blocks
:
640 ip_block
= random
.choice(ip_blocks
)
643 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(ip_block
)
644 self
.write_debug(f
'Using fake IP {self._x_forwarded_for_ip} as X-Forwarded-For')
647 # Path 2: bypassing based on country code
649 # Explicit country code specified by user, use it right away
650 # regardless of whether extractor is geo bypassable or not
651 country
= self
.get_param('geo_bypass_country', None)
653 # Otherwise use random country code from geo bypass context but
654 # only if extractor is known as geo bypassable
656 countries
= geo_bypass_context
.get('countries')
657 if self
._GEO
_BYPASS
and countries
:
658 country
= random
.choice(countries
)
661 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country
)
662 self
._downloader
.write_debug(
663 f
'Using fake IP {self._x_forwarded_for_ip} ({country.upper()}) as X-Forwarded-For')
665 def extract(self
, url
):
666 """Extracts URL information and returns it in list of dicts."""
671 self
.write_debug('Extracting URL: %s' % url
)
672 ie_result
= self
._real
_extract
(url
)
673 if ie_result
is None:
675 if self
._x
_forwarded
_for
_ip
:
676 ie_result
['__x_forwarded_for_ip'] = self
._x
_forwarded
_for
_ip
677 subtitles
= ie_result
.get('subtitles') or {}
678 if 'no-live-chat' in self
.get_param('compat_opts'):
679 for lang
in ('live_chat', 'comments', 'danmaku'):
680 subtitles
.pop(lang
, None)
682 except GeoRestrictedError
as e
:
683 if self
.__maybe
_fake
_ip
_and
_retry
(e
.countries
):
686 except UnsupportedError
:
688 except ExtractorError
as e
:
690 'video_id': e
.video_id
or self
.get_temp_id(url
),
692 'tb': e
.traceback
or sys
.exc_info()[2],
693 'expected': e
.expected
,
696 if hasattr(e
, 'countries'):
697 kwargs
['countries'] = e
.countries
698 raise type(e
)(e
.orig_msg
, **kwargs
)
699 except http
.client
.IncompleteRead
as e
:
700 raise ExtractorError('A network error has occurred.', cause
=e
, expected
=True, video_id
=self
.get_temp_id(url
))
701 except (KeyError, StopIteration) as e
:
702 raise ExtractorError('An extractor error has occurred.', cause
=e
, video_id
=self
.get_temp_id(url
))
704 def __maybe_fake_ip_and_retry(self
, countries
):
705 if (not self
.get_param('geo_bypass_country', None)
707 and self
.get_param('geo_bypass', True)
708 and not self
._x
_forwarded
_for
_ip
710 country_code
= random
.choice(countries
)
711 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country_code
)
712 if self
._x
_forwarded
_for
_ip
:
714 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
715 % (self
._x
_forwarded
_for
_ip
, country_code
.upper()))
719 def set_downloader(self
, downloader
):
720 """Sets a YoutubeDL instance as the downloader for this IE."""
721 self
._downloader
= downloader
725 return self
._downloader
.cache
729 return self
._downloader
.cookiejar
731 def _initialize_pre_login(self
):
732 """ Initialization before login. Redefine in subclasses."""
735 def _perform_login(self
, username
, password
):
736 """ Login with username and password. Redefine in subclasses."""
739 def _real_initialize(self
):
740 """Real initialization process. Redefine in subclasses."""
743 def _real_extract(self
, url
):
744 """Real extraction process. Redefine in subclasses."""
745 raise NotImplementedError('This method must be implemented by subclasses')
749 """A string for getting the InfoExtractor with get_info_extractor"""
750 return cls
.__name
__[:-2]
754 return cls
.__name
__[:-2]
757 def __can_accept_status_code(err
, expected_status
):
758 assert isinstance(err
, urllib
.error
.HTTPError
)
759 if expected_status
is None:
761 elif callable(expected_status
):
762 return expected_status(err
.code
) is True
764 return err
.code
in variadic(expected_status
)
766 def _create_request(self
, url_or_request
, data
=None, headers
=None, query
=None):
767 if isinstance(url_or_request
, urllib
.request
.Request
):
768 return update_Request(url_or_request
, data
=data
, headers
=headers
, query
=query
)
770 url_or_request
= update_url_query(url_or_request
, query
)
771 return sanitized_Request(url_or_request
, data
, headers
or {})
773 def _request_webpage(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, data
=None, headers
=None, query
=None, expected_status
=None):
775 Return the response handle.
777 See _download_webpage docstring for arguments specification.
779 if not self
._downloader
._first
_webpage
_request
:
780 sleep_interval
= self
.get_param('sleep_interval_requests') or 0
781 if sleep_interval
> 0:
782 self
.to_screen('Sleeping %s seconds ...' % sleep_interval
)
783 time
.sleep(sleep_interval
)
785 self
._downloader
._first
_webpage
_request
= False
788 self
.report_download_webpage(video_id
)
789 elif note
is not False:
791 self
.to_screen(str(note
))
793 self
.to_screen(f
'{video_id}: {note}')
795 # Some sites check X-Forwarded-For HTTP header in order to figure out
796 # the origin of the client behind proxy. This allows bypassing geo
797 # restriction by faking this header's value to IP that belongs to some
798 # geo unrestricted country. We will do so once we encounter any
799 # geo restriction error.
800 if self
._x
_forwarded
_for
_ip
:
801 headers
= (headers
or {}).copy()
802 headers
.setdefault('X-Forwarded-For', self
._x
_forwarded
_for
_ip
)
805 return self
._downloader
.urlopen(self
._create
_request
(url_or_request
, data
, headers
, query
))
806 except network_exceptions
as err
:
807 if isinstance(err
, urllib
.error
.HTTPError
):
808 if self
.__can
_accept
_status
_code
(err
, expected_status
):
809 # Retain reference to error to prevent file object from
810 # being closed before it can be read. Works around the
811 # effects of <https://bugs.python.org/issue15002>
812 # introduced in Python 3.4.1.
819 errnote
= 'Unable to download webpage'
821 errmsg
= f
'{errnote}: {error_to_compat_str(err)}'
823 raise ExtractorError(errmsg
, cause
=err
)
825 self
.report_warning(errmsg
)
828 def _download_webpage_handle(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True,
829 encoding
=None, data
=None, headers
={}, query={}
, expected_status
=None):
831 Return a tuple (page content as string, URL handle).
834 url_or_request -- plain text URL as a string or
835 a urllib.request.Request object
836 video_id -- Video/playlist/item identifier (string)
839 note -- note printed before downloading (string)
840 errnote -- note printed in case of an error (string)
841 fatal -- flag denoting whether error should be considered fatal,
842 i.e. whether it should cause ExtractionError to be raised,
843 otherwise a warning will be reported and extraction continued
844 encoding -- encoding for a page content decoding, guessed automatically
845 when not explicitly specified
846 data -- POST data (bytes)
847 headers -- HTTP headers (dict)
848 query -- URL query (dict)
849 expected_status -- allows to accept failed HTTP requests (non 2xx
850 status code) by explicitly specifying a set of accepted status
851 codes. Can be any of the following entities:
852 - an integer type specifying an exact failed status code to
854 - a list or a tuple of integer types specifying a list of
855 failed status codes to accept
856 - a callable accepting an actual failed status code and
857 returning True if it should be accepted
858 Note that this argument does not affect success status codes (2xx)
859 which are always accepted.
862 # Strip hashes from the URL (#1038)
863 if isinstance(url_or_request
, str):
864 url_or_request
= url_or_request
.partition('#')[0]
866 urlh
= self
._request
_webpage
(url_or_request
, video_id
, note
, errnote
, fatal
, data
=data
, headers
=headers
, query
=query
, expected_status
=expected_status
)
870 content
= self
._webpage
_read
_content
(urlh
, url_or_request
, video_id
, note
, errnote
, fatal
, encoding
=encoding
)
871 return (content
, urlh
)
874 def _guess_encoding_from_content(content_type
, webpage_bytes
):
875 m
= re
.match(r
'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type
)
877 encoding
= m
.group(1)
879 m
= re
.search(br
'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
880 webpage_bytes[:1024])
882 encoding = m.group(1).decode('ascii')
883 elif webpage_bytes.startswith(b'\xff\xfe'):
890 def __check_blocked(self, content):
891 first_block = content[:512]
892 if ('<title>Access to this site is blocked</title>' in content
893 and 'Websense' in first_block):
894 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
895 blocked_iframe = self._html_search_regex(
896 r'<iframe src="([^
"]+)"', content,
897 'Websense information URL
', default=None)
899 msg += ' Visit
%s for more details
' % blocked_iframe
900 raise ExtractorError(msg, expected=True)
901 if '<title
>The URL you requested has been blocked
</title
>' in first_block:
903 'Access to this webpage has been blocked by Indian censorship
. '
904 'Use a VPN
or proxy
server (with --proxy
) to route around it
.')
905 block_msg = self._html_search_regex(
906 r'</h1
><p
>(.*?
)</p
>',
907 content, 'block message
', default=None)
909 msg += ' (Message
: "%s")' % block_msg.replace('\n', ' ')
910 raise ExtractorError(msg, expected=True)
911 if ('<title
>TTK
:: Доступ к ресурсу ограничен
</title
>' in content
912 and 'blocklist
.rkn
.gov
.ru
' in content):
913 raise ExtractorError(
914 'Access to this webpage has been blocked by decision of the Russian government
. '
915 'Visit http
://blocklist
.rkn
.gov
.ru
/ for a block reason
.',
918 def _request_dump_filename(self, url, video_id):
919 basen = f'{video_id}_{url}
'
920 trim_length = self.get_param('trim_file_name
') or 240
921 if len(basen) > trim_length:
922 h = '___
' + hashlib.md5(basen.encode('utf
-8')).hexdigest()
923 basen = basen[:trim_length - len(h)] + h
924 filename = sanitize_filename(f'{basen}
.dump
', restricted=True)
925 # Working around MAX_PATH limitation on Windows (see
926 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
927 if compat_os_name == 'nt
':
928 absfilepath = os.path.abspath(filename)
929 if len(absfilepath) > 259:
930 filename = fR'\\?\{absfilepath}
'
933 def __decode_webpage(self, webpage_bytes, encoding, headers):
935 encoding = self._guess_encoding_from_content(headers.get('Content
-Type
', ''), webpage_bytes)
937 return webpage_bytes.decode(encoding, 'replace
')
939 return webpage_bytes.decode('utf
-8', 'replace
')
941 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
942 webpage_bytes = urlh.read()
943 if prefix is not None:
944 webpage_bytes = prefix + webpage_bytes
945 if self.get_param('dump_intermediate_pages
', False):
946 self.to_screen('Dumping request to
' + urlh.geturl())
947 dump = base64.b64encode(webpage_bytes).decode('ascii
')
948 self._downloader.to_screen(dump)
949 if self.get_param('write_pages
'):
950 filename = self._request_dump_filename(urlh.geturl(), video_id)
951 self.to_screen(f'Saving request to {filename}
')
952 with open(filename, 'wb
') as outf:
953 outf.write(webpage_bytes)
955 content = self.__decode_webpage(webpage_bytes, encoding, urlh.headers)
956 self.__check_blocked(content)
960 def __print_error(self, errnote, fatal, video_id, err):
962 raise ExtractorError(f'{video_id}
: {errnote}
', cause=err)
964 self.report_warning(f'{video_id}
: {errnote}
: {err}
')
966 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True, errnote=None):
968 xml_string = transform_source(xml_string)
970 return compat_etree_fromstring(xml_string.encode('utf
-8'))
971 except xml.etree.ElementTree.ParseError as ve:
972 self.__print_error('Failed to parse XML
' if errnote is None else errnote, fatal, video_id, ve)
974 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True, errnote=None, **parser_kwargs):
977 json_string, cls=LenientJSONDecoder, strict=False, transform_source=transform_source, **parser_kwargs)
978 except ValueError as ve:
979 self.__print_error('Failed to parse JSON
' if errnote is None else errnote, fatal, video_id, ve)
981 def _parse_socket_response_as_json(self, data, *args, **kwargs):
982 return self._parse_json(data[data.find('{'):data.rfind('}
') + 1], *args, **kwargs)
984 def __create_download_methods(name, parser, note, errnote, return_value):
986 def parse(ie, content, *args, errnote=errnote, **kwargs):
990 kwargs['errnote
'] = errnote
991 # parser is fetched by name so subclasses can override it
992 return getattr(ie, parser)(content, *args, **kwargs)
994 def download_handle(self, url_or_request, video_id, note=note, errnote=errnote, transform_source=None,
995 fatal=True, encoding=None, data=None, headers={}, query={}, expected_status=None):
996 res = self._download_webpage_handle(
997 url_or_request, video_id, note=note, errnote=errnote, fatal=fatal, encoding=encoding,
998 data=data, headers=headers, query=query, expected_status=expected_status)
1002 return parse(self, content, video_id, transform_source=transform_source, fatal=fatal, errnote=errnote), urlh
1004 def download_content(self, url_or_request, video_id, note=note, errnote=errnote, transform_source=None,
1005 fatal=True, encoding=None, data=None, headers={}, query={}, expected_status=None):
1006 if self.get_param('load_pages
'):
1007 url_or_request = self._create_request(url_or_request, data, headers, query)
1008 filename = self._request_dump_filename(url_or_request.full_url, video_id)
1009 self.to_screen(f'Loading request
from {filename}
')
1011 with open(filename, 'rb
') as dumpf:
1012 webpage_bytes = dumpf.read()
1013 except OSError as e:
1014 self.report_warning(f'Unable to load request
from disk
: {e}
')
1016 content = self.__decode_webpage(webpage_bytes, encoding, url_or_request.headers)
1017 return parse(self, content, video_id, transform_source=transform_source, fatal=fatal, errnote=errnote)
1021 'transform_source
': transform_source,
1023 'encoding
': encoding,
1027 'expected_status
': expected_status,
1030 kwargs.pop('transform_source
')
1031 # The method is fetched by name so subclasses can override _download_..._handle
1032 res = getattr(self, download_handle.__name__)(url_or_request, video_id, **kwargs)
1033 return res if res is False else res[0]
1035 def impersonate(func, name, return_value):
1036 func.__name__, func.__qualname__ = name, f'InfoExtractor
.{name}
'
1038 @param transform_source Apply this transformation before parsing
1039 @returns {return_value}
1041 See _download_webpage_handle docstring for other arguments specification
1044 impersonate(download_handle, f'_download_{name}_handle
', f'({return_value}
, URL handle
)')
1045 impersonate(download_content, f'_download_{name}
', f'{return_value}
')
1046 return download_handle, download_content
1048 _download_xml_handle, _download_xml = __create_download_methods(
1049 'xml
', '_parse_xml
', 'Downloading XML
', 'Unable to download XML
', 'xml
as an xml
.etree
.ElementTree
.Element
')
1050 _download_json_handle, _download_json = __create_download_methods(
1051 'json
', '_parse_json
', 'Downloading JSON metadata
', 'Unable to download JSON metadata
', 'JSON
object as a
dict')
1052 _download_socket_json_handle, _download_socket_json = __create_download_methods(
1053 'socket_json
', '_parse_socket_response_as_json
', 'Polling socket
', 'Unable to poll socket
', 'JSON
object as a
dict')
1054 __download_webpage = __create_download_methods('webpage
', None, None, None, 'data of the page
as a string
')[1]
1056 def _download_webpage(
1057 self, url_or_request, video_id, note=None, errnote=None,
1058 fatal=True, tries=1, timeout=NO_DEFAULT, *args, **kwargs):
1060 Return the data of the page as a string.
1063 tries -- number of tries
1064 timeout -- sleep interval between tries
1066 See _download_webpage_handle docstring for other arguments specification.
1069 R''' # NB: These are unused; should they be deprecated?
1071 self._downloader.deprecation_warning('tries argument
is deprecated
in InfoExtractor
._download
_webpage
')
1072 if timeout is NO_DEFAULT:
1075 self._downloader.deprecation_warning('timeout argument
is deprecated
in InfoExtractor
._download
_webpage
')
1081 return self.__download_webpage(url_or_request, video_id, note, errnote, None, fatal, *args, **kwargs)
1082 except http.client.IncompleteRead as e:
1084 if try_count >= tries:
1086 self._sleep(timeout, video_id)
1088 def report_warning(self, msg, video_id=None, *args, only_once=False, **kwargs):
1089 idstr = format_field(video_id, None, '%s: ')
1090 msg = f'[{self.IE_NAME}
] {idstr}{msg}
'
1092 if f'WARNING
: {msg}
' in self._printed_messages:
1094 self._printed_messages.add(f'WARNING
: {msg}
')
1095 self._downloader.report_warning(msg, *args, **kwargs)
1097 def to_screen(self, msg, *args, **kwargs):
1098 """Print msg to screen, prefixing it with '[ie_name
]'"""
1099 self._downloader.to_screen(f'[{self.IE_NAME}
] {msg}
', *args, **kwargs)
1101 def write_debug(self, msg, *args, **kwargs):
1102 self._downloader.write_debug(f'[{self.IE_NAME}
] {msg}
', *args, **kwargs)
1104 def get_param(self, name, default=None, *args, **kwargs):
1105 if self._downloader:
1106 return self._downloader.params.get(name, default, *args, **kwargs)
1109 def report_drm(self, video_id, partial=False):
1110 self.raise_no_formats('This video
is DRM protected
', expected=True, video_id=video_id)
1112 def report_extraction(self, id_or_name):
1113 """Report information extraction."""
1114 self.to_screen('%s: Extracting information
' % id_or_name)
1116 def report_download_webpage(self, video_id):
1117 """Report webpage download."""
1118 self.to_screen('%s: Downloading webpage
' % video_id)
1120 def report_age_confirmation(self):
1121 """Report attempt to confirm age."""
1122 self.to_screen('Confirming age
')
1124 def report_login(self):
1125 """Report attempt to log in."""
1126 self.to_screen('Logging
in')
1128 def raise_login_required(
1129 self, msg='This video
is only available
for registered users
',
1130 metadata_available=False, method=NO_DEFAULT):
1131 if metadata_available and (
1132 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1133 self.report_warning(msg)
1135 msg += format_field(self._login_hint(method), None, '. %s')
1136 raise ExtractorError(msg, expected=True)
1138 def raise_geo_restricted(
1139 self, msg='This video
is not available
from your location due to geo restriction
',
1140 countries=None, metadata_available=False):
1141 if metadata_available and (
1142 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1143 self.report_warning(msg)
1145 raise GeoRestrictedError(msg, countries=countries)
1147 def raise_no_formats(self, msg, expected=False, video_id=None):
1149 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1150 self.report_warning(msg, video_id)
1151 elif isinstance(msg, ExtractorError):
1154 raise ExtractorError(msg, expected=expected, video_id=video_id)
1156 # Methods for following #608
1158 def url_result(url, ie=None, video_id=None, video_title=None, *, url_transparent=False, **kwargs):
1159 """Returns a URL that points to a page that should be processed"""
1161 kwargs['ie_key
'] = ie if isinstance(ie, str) else ie.ie_key()
1162 if video_id is not None:
1163 kwargs['id'] = video_id
1164 if video_title is not None:
1165 kwargs['title
'] = video_title
1168 '_type
': 'url_transparent
' if url_transparent else 'url
',
1173 def playlist_from_matches(cls, matches, playlist_id=None, playlist_title=None,
1174 getter=IDENTITY, ie=None, video_kwargs=None, **kwargs):
1175 return cls.playlist_result(
1176 (cls.url_result(m, ie, **(video_kwargs or {})) for m in orderedSet(map(getter, matches), lazy=True)),
1177 playlist_id, playlist_title, **kwargs)
1180 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None, *, multi_video=False, **kwargs):
1181 """Returns a playlist"""
1183 kwargs['id'] = playlist_id
1185 kwargs['title
'] = playlist_title
1186 if playlist_description is not None:
1187 kwargs['description
'] = playlist_description
1190 '_type
': 'multi_video
' if multi_video else 'playlist
',
1194 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1196 Perform a regex search on the given string, using a single or a list of
1197 patterns returning the first matching group.
1198 In case of failure return a default value or raise a WARNING or a
1199 RegexNotFoundError, depending on fatal, specifying the field name.
1203 elif isinstance(pattern, (str, re.Pattern)):
1204 mobj = re.search(pattern, string, flags)
1207 mobj = re.search(p, string, flags)
1211 _name = self._downloader._format_err(name, self._downloader.Styles.EMPHASIS)
1215 # return the first matching group
1216 return next(g for g in mobj.groups() if g is not None)
1217 elif isinstance(group, (list, tuple)):
1218 return tuple(mobj.group(g) for g in group)
1220 return mobj.group(group)
1221 elif default is not NO_DEFAULT:
1224 raise RegexNotFoundError('Unable to extract
%s' % _name)
1226 self.report_warning('unable to extract
%s' % _name + bug_reports_message())
1229 def _search_json(self, start_pattern, string, name, video_id, *, end_pattern='',
1230 contains_pattern=r'{(?s:.+)}
', fatal=True, default=NO_DEFAULT, **kwargs):
1231 """Searches string for the JSON object specified by start_pattern"""
1232 # NB: end_pattern is only used to reduce the size of the initial match
1233 if default is NO_DEFAULT:
1234 default, has_default = {}, False
1236 fatal, has_default = False, True
1238 json_string = self._search_regex(
1239 rf'(?
:{start_pattern}
)\s
*(?P
<json
>{contains_pattern}
)\s
*(?
:{end_pattern}
)',
1240 string, name, group='json
', fatal=fatal, default=None if has_default else NO_DEFAULT)
1244 _name = self._downloader._format_err(name, self._downloader.Styles.EMPHASIS)
1246 return self._parse_json(json_string, video_id, ignore_extra=True, **kwargs)
1247 except ExtractorError as e:
1249 raise ExtractorError(
1250 f'Unable to extract {_name}
- Failed to parse JSON
', cause=e.cause, video_id=video_id)
1251 elif not has_default:
1252 self.report_warning(
1253 f'Unable to extract {_name}
- Failed to parse JSON
: {e}
', video_id=video_id)
1256 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1258 Like _search_regex, but strips HTML tags and unescapes entities.
1260 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1262 return clean_html(res).strip()
1266 def _get_netrc_login_info(self, netrc_machine=None):
1269 netrc_machine = netrc_machine or self._NETRC_MACHINE
1271 if self.get_param('usenetrc
', False):
1273 netrc_file = compat_expanduser(self.get_param('netrc_location
') or '~
')
1274 if os.path.isdir(netrc_file):
1275 netrc_file = os.path.join(netrc_file, '.netrc
')
1276 info = netrc.netrc(file=netrc_file).authenticators(netrc_machine)
1277 if info is not None:
1281 raise netrc.NetrcParseError(
1282 'No authenticators
for %s' % netrc_machine)
1283 except (OSError, netrc.NetrcParseError) as err:
1284 self.report_warning(
1285 'parsing
.netrc
: %s' % error_to_compat_str(err))
1287 return username, password
1289 def _get_login_info(self, username_option='username
', password_option='password
', netrc_machine=None):
1291 Get the login info as (username, password)
1292 First look for the manually specified credentials using username_option
1293 and password_option as keys in params dictionary. If no such credentials
1294 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1296 If there's no info available
, return (None, None)
1299 # Attempt to use provided username and password or .netrc data
1300 username = self.get_param(username_option)
1301 if username is not None:
1302 password = self.get_param(password_option)
1304 username, password = self._get_netrc_login_info(netrc_machine)
1306 return username, password
1308 def _get_tfa_info(self, note='two-factor verification code'):
1310 Get the two
-factor authentication info
1311 TODO
- asking the user will be required
for sms
/phone verify
1312 currently just uses the command line option
1313 If there
's no info available, return None
1316 tfa = self.get_param('twofactor
')
1320 return getpass.getpass('Type
%s and press
[Return
]: ' % note)
1322 # Helper functions for extracting OpenGraph info
1324 def _og_regexes(prop):
1325 content_re = r'content
=(?
:"([^"]+?
)"|\'([^\']+?)\'|\s*([^\s"\'=<>`
]+?
))'
1326 property_re = (r'(?
:name|
property)=(?
:\'og
%(sep)s%(prop)s\'|
"og%(sep)s%(prop)s"|\s
*og
%(sep)s%(prop)s\b)'
1327 % {'prop': re.escape(prop), 'sep': '(?::|[:-])'})
1328 template = r'<meta
[^
>]+?
%s[^
>]+?
%s'
1330 template % (property_re, content_re),
1331 template % (content_re, property_re),
1335 def _meta_regex(prop):
1336 return r'''(?isx)<meta
1337 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1338 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1340 def _og_search_property(self, prop, html, name=None, **kargs):
1341 prop = variadic(prop)
1343 name = 'OpenGraph
%s' % prop[0]
1346 og_regexes.extend(self._og_regexes(p))
1347 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1350 return unescapeHTML(escaped)
1352 def _og_search_thumbnail(self, html, **kargs):
1353 return self._og_search_property('image
', html, 'thumbnail URL
', fatal=False, **kargs)
1355 def _og_search_description(self, html, **kargs):
1356 return self._og_search_property('description
', html, fatal=False, **kargs)
1358 def _og_search_title(self, html, *, fatal=False, **kargs):
1359 return self._og_search_property('title
', html, fatal=fatal, **kargs)
1361 def _og_search_video_url(self, html, name='video url
', secure=True, **kargs):
1362 regexes = self._og_regexes('video
') + self._og_regexes('video
:url
')
1364 regexes = self._og_regexes('video
:secure_url
') + regexes
1365 return self._html_search_regex(regexes, html, name, **kargs)
1367 def _og_search_url(self, html, **kargs):
1368 return self._og_search_property('url
', html, **kargs)
1370 def _html_extract_title(self, html, name='title
', *, fatal=False, **kwargs):
1371 return self._html_search_regex(r'(?s
)<title
\b[^
>]*>([^
<]+)</title
>', html, name, fatal=fatal, **kwargs)
1373 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1374 name = variadic(name)
1375 if display_name is None:
1376 display_name = name[0]
1377 return self._html_search_regex(
1378 [self._meta_regex(n) for n in name],
1379 html, display_name, fatal=fatal, group='content
', **kwargs)
1381 def _dc_search_uploader(self, html):
1382 return self._html_search_meta('dc
.creator
', html, 'uploader
')
1385 def _rta_search(html):
1386 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1387 if re.search(r'(?ix
)<meta\s
+name
="rating"\s
+'
1388 r' content
="RTA-5042-1996-1400-1577-RTA"',
1392 # And then there are the jokers who advertise that they use RTA, but actually don't
.
1393 AGE_LIMIT_MARKERS
= [
1394 r
'Proudly Labeled <a href="http://www\.rtalabel\.org/" title="Restricted to Adults">RTA</a>',
1396 if any(re
.search(marker
, html
) for marker
in AGE_LIMIT_MARKERS
):
1400 def _media_rating_search(self
, html
):
1401 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1402 rating
= self
._html
_search
_meta
('rating', html
)
1414 return RATING_TABLE
.get(rating
.lower())
1416 def _family_friendly_search(self
, html
):
1417 # See http://schema.org/VideoObject
1418 family_friendly
= self
._html
_search
_meta
(
1419 'isFamilyFriendly', html
, default
=None)
1421 if not family_friendly
:
1430 return RATING_TABLE
.get(family_friendly
.lower())
1432 def _twitter_search_player(self
, html
):
1433 return self
._html
_search
_meta
('twitter:player', html
,
1434 'twitter card player')
1436 def _yield_json_ld(self
, html
, video_id
, *, fatal
=True, default
=NO_DEFAULT
):
1437 """Yield all json ld objects in the html"""
1438 if default
is not NO_DEFAULT
:
1440 for mobj
in re
.finditer(JSON_LD_RE
, html
):
1441 json_ld_item
= self
._parse
_json
(mobj
.group('json_ld'), video_id
, fatal
=fatal
)
1442 for json_ld
in variadic(json_ld_item
):
1443 if isinstance(json_ld
, dict):
1446 def _search_json_ld(self
, html
, video_id
, expected_type
=None, *, fatal
=True, default
=NO_DEFAULT
):
1447 """Search for a video in any json ld in the html"""
1448 if default
is not NO_DEFAULT
:
1450 info
= self
._json
_ld
(
1451 list(self
._yield
_json
_ld
(html
, video_id
, fatal
=fatal
, default
=default
)),
1452 video_id
, fatal
=fatal
, expected_type
=expected_type
)
1455 if default
is not NO_DEFAULT
:
1458 raise RegexNotFoundError('Unable to extract JSON-LD')
1460 self
.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1463 def _json_ld(self
, json_ld
, video_id
, fatal
=True, expected_type
=None):
1464 if isinstance(json_ld
, str):
1465 json_ld
= self
._parse
_json
(json_ld
, video_id
, fatal
=fatal
)
1469 if not isinstance(json_ld
, (list, tuple, dict)):
1471 if isinstance(json_ld
, dict):
1474 INTERACTION_TYPE_MAP
= {
1475 'CommentAction': 'comment',
1476 'AgreeAction': 'like',
1477 'DisagreeAction': 'dislike',
1478 'LikeAction': 'like',
1479 'DislikeAction': 'dislike',
1480 'ListenAction': 'view',
1481 'WatchAction': 'view',
1482 'ViewAction': 'view',
1485 def is_type(e
, *expected_types
):
1486 type = variadic(traverse_obj(e
, '@type'))
1487 return any(x
in type for x
in expected_types
)
1489 def extract_interaction_type(e
):
1490 interaction_type
= e
.get('interactionType')
1491 if isinstance(interaction_type
, dict):
1492 interaction_type
= interaction_type
.get('@type')
1493 return str_or_none(interaction_type
)
1495 def extract_interaction_statistic(e
):
1496 interaction_statistic
= e
.get('interactionStatistic')
1497 if isinstance(interaction_statistic
, dict):
1498 interaction_statistic
= [interaction_statistic
]
1499 if not isinstance(interaction_statistic
, list):
1501 for is_e
in interaction_statistic
:
1502 if not is_type(is_e
, 'InteractionCounter'):
1504 interaction_type
= extract_interaction_type(is_e
)
1505 if not interaction_type
:
1507 # For interaction count some sites provide string instead of
1508 # an integer (as per spec) with non digit characters (e.g. ",")
1509 # so extracting count with more relaxed str_to_int
1510 interaction_count
= str_to_int(is_e
.get('userInteractionCount'))
1511 if interaction_count
is None:
1513 count_kind
= INTERACTION_TYPE_MAP
.get(interaction_type
.split('/')[-1])
1516 count_key
= '%s_count' % count_kind
1517 if info
.get(count_key
) is not None:
1519 info
[count_key
] = interaction_count
1521 def extract_chapter_information(e
):
1523 'title': part
.get('name'),
1524 'start_time': part
.get('startOffset'),
1525 'end_time': part
.get('endOffset'),
1526 } for part
in variadic(e
.get('hasPart') or []) if part
.get('@type') == 'Clip']
1527 for idx
, (last_c
, current_c
, next_c
) in enumerate(zip(
1528 [{'end_time': 0}
] + chapters
, chapters
, chapters
[1:])):
1529 current_c
['end_time'] = current_c
['end_time'] or next_c
['start_time']
1530 current_c
['start_time'] = current_c
['start_time'] or last_c
['end_time']
1531 if None in current_c
.values():
1532 self
.report_warning(f
'Chapter {idx} contains broken data. Not extracting chapters')
1535 chapters
[-1]['end_time'] = chapters
[-1]['end_time'] or info
['duration']
1536 info
['chapters'] = chapters
1538 def extract_video_object(e
):
1539 author
= e
.get('author')
1541 'url': url_or_none(e
.get('contentUrl')),
1542 'ext': mimetype2ext(e
.get('encodingFormat')),
1543 'title': unescapeHTML(e
.get('name')),
1544 'description': unescapeHTML(e
.get('description')),
1545 'thumbnails': [{'url': unescapeHTML(url)}
1546 for url
in variadic(traverse_obj(e
, 'thumbnailUrl', 'thumbnailURL'))
1547 if url_or_none(url
)],
1548 'duration': parse_duration(e
.get('duration')),
1549 'timestamp': unified_timestamp(e
.get('uploadDate')),
1550 # author can be an instance of 'Organization' or 'Person' types.
1551 # both types can have 'name' property(inherited from 'Thing' type). [1]
1552 # however some websites are using 'Text' type instead.
1553 # 1. https://schema.org/VideoObject
1554 'uploader': author
.get('name') if isinstance(author
, dict) else author
if isinstance(author
, str) else None,
1555 'artist': traverse_obj(e
, ('byArtist', 'name'), expected_type
=str),
1556 'filesize': int_or_none(float_or_none(e
.get('contentSize'))),
1557 'tbr': int_or_none(e
.get('bitrate')),
1558 'width': int_or_none(e
.get('width')),
1559 'height': int_or_none(e
.get('height')),
1560 'view_count': int_or_none(e
.get('interactionCount')),
1561 'tags': try_call(lambda: e
.get('keywords').split(',')),
1563 if is_type(e
, 'AudioObject'):
1566 'abr': int_or_none(e
.get('bitrate')),
1568 extract_interaction_statistic(e
)
1569 extract_chapter_information(e
)
1571 def traverse_json_ld(json_ld
, at_top_level
=True):
1573 if at_top_level
and '@context' not in e
:
1575 if at_top_level
and set(e
.keys()) == {'@context', '@graph'}
:
1576 traverse_json_ld(variadic(e
['@graph'], allowed_types
=(dict,)), at_top_level
=False)
1578 if expected_type
is not None and not is_type(e
, expected_type
):
1580 rating
= traverse_obj(e
, ('aggregateRating', 'ratingValue'), expected_type
=float_or_none
)
1581 if rating
is not None:
1582 info
['average_rating'] = rating
1583 if is_type(e
, 'TVEpisode', 'Episode'):
1584 episode_name
= unescapeHTML(e
.get('name'))
1586 'episode': episode_name
,
1587 'episode_number': int_or_none(e
.get('episodeNumber')),
1588 'description': unescapeHTML(e
.get('description')),
1590 if not info
.get('title') and episode_name
:
1591 info
['title'] = episode_name
1592 part_of_season
= e
.get('partOfSeason')
1593 if is_type(part_of_season
, 'TVSeason', 'Season', 'CreativeWorkSeason'):
1595 'season': unescapeHTML(part_of_season
.get('name')),
1596 'season_number': int_or_none(part_of_season
.get('seasonNumber')),
1598 part_of_series
= e
.get('partOfSeries') or e
.get('partOfTVSeries')
1599 if is_type(part_of_series
, 'TVSeries', 'Series', 'CreativeWorkSeries'):
1600 info
['series'] = unescapeHTML(part_of_series
.get('name'))
1601 elif is_type(e
, 'Movie'):
1603 'title': unescapeHTML(e
.get('name')),
1604 'description': unescapeHTML(e
.get('description')),
1605 'duration': parse_duration(e
.get('duration')),
1606 'timestamp': unified_timestamp(e
.get('dateCreated')),
1608 elif is_type(e
, 'Article', 'NewsArticle'):
1610 'timestamp': parse_iso8601(e
.get('datePublished')),
1611 'title': unescapeHTML(e
.get('headline')),
1612 'description': unescapeHTML(e
.get('articleBody') or e
.get('description')),
1614 if is_type(traverse_obj(e
, ('video', 0)), 'VideoObject'):
1615 extract_video_object(e
['video'][0])
1616 elif is_type(traverse_obj(e
, ('subjectOf', 0)), 'VideoObject'):
1617 extract_video_object(e
['subjectOf'][0])
1618 elif is_type(e
, 'VideoObject', 'AudioObject'):
1619 extract_video_object(e
)
1620 if expected_type
is None:
1624 video
= e
.get('video')
1625 if is_type(video
, 'VideoObject'):
1626 extract_video_object(video
)
1627 if expected_type
is None:
1631 traverse_json_ld(json_ld
)
1633 return filter_dict(info
)
1635 def _search_nextjs_data(self
, webpage
, video_id
, *, transform_source
=None, fatal
=True, **kw
):
1636 return self
._parse
_json
(
1638 r
'(?s)<script[^>]+id=[\'"]__NEXT_DATA__[\'"][^
>]*>([^
<]+)</script
>',
1639 webpage, 'next
.js data
', fatal=fatal, **kw),
1640 video_id, transform_source=transform_source, fatal=fatal)
1642 def _search_nuxt_data(self, webpage, video_id, context_name='__NUXT__
', *, fatal=True, traverse=('data
', 0)):
1643 """Parses Nuxt.js metadata. This works as long as the function __NUXT__ invokes is a pure function"""
1644 rectx = re.escape(context_name)
1645 FUNCTION_RE = r'\
(function\
((?P
<arg_keys
>.*?
)\
){return\s+(?P<js>{.*?}
)\s
*;?\s
*}\
((?P
<arg_vals
>.*?
)\
)'
1646 js, arg_keys, arg_vals = self._search_regex(
1647 (rf'<script
>\s
*window\
.{rectx}
={FUNCTION_RE}\s
*\
)\s
*;?\s
*</script
>', rf'{rectx}\
(.*?{FUNCTION_RE}
'),
1648 webpage, context_name, group=('js
', 'arg_keys
', 'arg_vals
'), fatal=fatal)
1650 args = dict(zip(arg_keys.split(','), arg_vals.split(',')))
1652 for key, val in args.items():
1653 if val in ('undefined
', 'void
0'):
1656 ret = self._parse_json(js, video_id, transform_source=functools.partial(js_to_json, vars=args), fatal=fatal)
1657 return traverse_obj(ret, traverse) or {}
1660 def _hidden_inputs(html):
1661 html = re.sub(r'<!--(?
:(?
!<!--).)*-->', '', html)
1663 for input in re.findall(r'(?i
)(<input[^
>]+>)', html):
1664 attrs = extract_attributes(input)
1667 if attrs.get('type') not in ('hidden
', 'submit
'):
1669 name = attrs.get('name
') or attrs.get('id')
1670 value = attrs.get('value
')
1671 if name and value is not None:
1672 hidden_inputs[name] = value
1673 return hidden_inputs
1675 def _form_hidden_inputs(self, form_id, html):
1676 form = self._search_regex(
1677 r'(?
is)<form
[^
>]+?
id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1678 html, '%s form' % form_id, group='form')
1679 return self._hidden_inputs(form)
1682 regex = r' *((?P<reverse>\+)?(?P<field>[a-zA-Z0-9_]+)((?P<separator>[~:])(?P<limit>.*?))?)? *$'
1684 default = ('hidden', 'aud_or_vid', 'hasvid', 'ie_pref', 'lang', 'quality',
1685 'res', 'fps', 'hdr:12', 'vcodec:vp9.2', 'channels', 'acodec',
1686 'size', 'br', 'asr', 'proto', 'ext', 'hasaud', 'source', 'id') # These must not be aliases
1687 ytdl_default = ('hasaud', 'lang', 'quality', 'tbr', 'filesize', 'vbr',
1688 'height', 'width', 'proto', 'vext', 'abr', 'aext',
1689 'fps', 'fs_approx', 'source', 'id')
1692 'vcodec': {'type': 'ordered', 'regex': True,
1693 'order': ['av0?1', 'vp0?9.2', 'vp0?9', '[hx]265|he?vc?', '[hx]264|avc', 'vp0?8', 'mp4v|h263', 'theora', '', None, 'none']},
1694 'acodec': {'type': 'ordered', 'regex': True,
1695 'order': ['[af]lac', 'wav|aiff', 'opus', 'vorbis|ogg', 'aac', 'mp?4a?', 'mp3', 'e-?a?c-?3', 'ac-?3', 'dts', '', None, 'none']},
1696 'hdr': {'type': 'ordered', 'regex': True, 'field': 'dynamic_range',
1697 'order': ['dv', '(hdr)?12', r'(hdr)?10\+', '(hdr)?10', 'hlg', '', 'sdr', None]},
1698 'proto': {'type': 'ordered', 'regex': True, 'field': 'protocol',
1699 'order': ['(ht|f)tps', '(ht|f)tp$', 'm3u8.*', '.*dash', 'websocket_frag', 'rtmpe?', '', 'mms|rtsp', 'ws|websocket', 'f4']},
1700 'vext': {'type': 'ordered', 'field': 'video_ext',
1701 'order': ('mp4', 'webm', 'flv', '', 'none'),
1702 'order_free': ('webm', 'mp4', 'flv', '', 'none')},
1703 'aext': {'type': 'ordered', 'field': 'audio_ext',
1704 'order': ('m4a', 'aac', 'mp3', 'ogg', 'opus', 'webm', '', 'none'),
1705 'order_free': ('ogg', 'opus', 'webm', 'mp3', 'm4a', 'aac', '', 'none')},
1706 'hidden': {'visible': False, 'forced': True, 'type': 'extractor', 'max': -1000},
1707 'aud_or_vid': {'visible': False, 'forced': True, 'type': 'multiple',
1708 'field': ('vcodec', 'acodec'),
1709 'function': lambda it: int(any(v != 'none' for v in it))},
1710 'ie_pref': {'priority': True, 'type': 'extractor'},
1711 'hasvid': {'priority': True, 'field': 'vcodec', 'type': 'boolean', 'not_in_list': ('none',)},
1712 'hasaud': {'field': 'acodec', 'type': 'boolean', 'not_in_list': ('none',)},
1713 'lang': {'convert': 'float', 'field': 'language_preference', 'default': -1},
1714 'quality': {'convert': 'float', 'default': -1},
1715 'filesize': {'convert': 'bytes'},
1716 'fs_approx': {'convert': 'bytes', 'field': 'filesize_approx'},
1717 'id': {'convert': 'string', 'field': 'format_id'},
1718 'height': {'convert': 'float_none'},
1719 'width': {'convert': 'float_none'},
1720 'fps': {'convert': 'float_none'},
1721 'channels': {'convert': 'float_none', 'field': 'audio_channels'},
1722 'tbr': {'convert': 'float_none'},
1723 'vbr': {'convert': 'float_none'},
1724 'abr': {'convert': 'float_none'},
1725 'asr': {'convert': 'float_none'},
1726 'source': {'convert': 'float', 'field': 'source_preference', 'default': -1},
1728 'codec': {'type': 'combined', 'field': ('vcodec', 'acodec')},
1729 'br': {'type': 'combined', 'field': ('tbr', 'vbr', 'abr'), 'same_limit': True},
1730 'size': {'type': 'combined', 'same_limit': True, 'field': ('filesize', 'fs_approx')},
1731 'ext': {'type': 'combined', 'field': ('vext', 'aext')},
1732 'res': {'type': 'multiple', 'field': ('height', 'width'),
1733 'function': lambda it: (lambda l: min(l) if l else 0)(tuple(filter(None, it)))},
1735 # Actual field names
1736 'format_id': {'type': 'alias', 'field': 'id'},
1737 'preference': {'type': 'alias', 'field': 'ie_pref'},
1738 'language_preference': {'type': 'alias', 'field': 'lang'},
1739 'source_preference': {'type': 'alias', 'field': 'source'},
1740 'protocol': {'type': 'alias', 'field': 'proto'},
1741 'filesize_approx': {'type': 'alias', 'field': 'fs_approx'},
1742 'audio_channels': {'type': 'alias', 'field': 'channels'},
1745 'dimension': {'type': 'alias', 'field': 'res', 'deprecated': True},
1746 'resolution': {'type': 'alias', 'field': 'res', 'deprecated': True},
1747 'extension': {'type': 'alias', 'field': 'ext', 'deprecated': True},
1748 'bitrate': {'type': 'alias', 'field': 'br', 'deprecated': True},
1749 'total_bitrate': {'type': 'alias', 'field': 'tbr', 'deprecated': True},
1750 'video_bitrate': {'type': 'alias', 'field': 'vbr', 'deprecated': True},
1751 'audio_bitrate': {'type': 'alias', 'field': 'abr', 'deprecated': True},
1752 'framerate': {'type': 'alias', 'field': 'fps', 'deprecated': True},
1753 'filesize_estimate': {'type': 'alias', 'field': 'size', 'deprecated': True},
1754 'samplerate': {'type': 'alias', 'field': 'asr', 'deprecated': True},
1755 'video_ext': {'type': 'alias', 'field': 'vext', 'deprecated': True},
1756 'audio_ext': {'type': 'alias', 'field': 'aext', 'deprecated': True},
1757 'video_codec': {'type': 'alias', 'field': 'vcodec', 'deprecated': True},
1758 'audio_codec': {'type': 'alias', 'field': 'acodec', 'deprecated': True},
1759 'video': {'type': 'alias', 'field': 'hasvid', 'deprecated': True},
1760 'has_video': {'type': 'alias', 'field': 'hasvid', 'deprecated': True},
1761 'audio': {'type': 'alias', 'field': 'hasaud', 'deprecated': True},
1762 'has_audio': {'type': 'alias', 'field': 'hasaud', 'deprecated': True},
1763 'extractor': {'type': 'alias', 'field': 'ie_pref', 'deprecated': True},
1764 'extractor_preference': {'type': 'alias', 'field': 'ie_pref', 'deprecated': True},
1767 def __init__(self, ie, field_preference):
1769 self.ydl = ie._downloader
1770 self.evaluate_params(self.ydl.params, field_preference)
1771 if ie.get_param('verbose'):
1772 self.print_verbose_info(self.ydl.write_debug)
1774 def _get_field_setting(self, field, key):
1775 if field not in self.settings:
1776 if key in ('forced', 'priority'):
1778 self.ydl.deprecated_feature(f'Using arbitrary fields ({field}) for format sorting is '
1779 'deprecated and may be removed in a future version')
1780 self.settings[field] = {}
1781 propObj = self.settings[field]
1782 if key not in propObj:
1783 type = propObj.get('type')
1785 default = 'preference' if type == 'extractor' else (field,) if type in ('combined', 'multiple') else field
1786 elif key == 'convert':
1787 default = 'order' if type == 'ordered' else 'float_string' if field else 'ignore'
1789 default = {'type': 'field', 'visible': True, 'order': [], 'not_in_list': (None,)}.get(key, None)
1790 propObj[key] = default
1793 def _resolve_field_value(self, field, value, convertNone=False):
1798 value = value.lower()
1799 conversion = self._get_field_setting(field, 'convert')
1800 if conversion == 'ignore':
1802 if conversion == 'string':
1804 elif conversion == 'float_none':
1805 return float_or_none(value)
1806 elif conversion == 'bytes':
1807 return FileDownloader.parse_bytes(value)
1808 elif conversion == 'order':
1809 order_list = (self._use_free_order and self._get_field_setting(field, 'order_free')) or self._get_field_setting(field, 'order')
1810 use_regex = self._get_field_setting(field, 'regex')
1811 list_length = len(order_list)
1812 empty_pos = order_list.index('') if '' in order_list else list_length + 1
1813 if use_regex and value is not None:
1814 for i, regex in enumerate(order_list):
1815 if regex and re.match(regex, value):
1816 return list_length - i
1817 return list_length - empty_pos # not in list
1818 else: # not regex or value = None
1819 return list_length - (order_list.index(value) if value in order_list else empty_pos)
1821 if value.isnumeric():
1824 self.settings[field]['convert'] = 'string'
1827 def evaluate_params(self, params, sort_extractor):
1828 self._use_free_order = params.get('prefer_free_formats', False)
1829 self._sort_user = params.get('format_sort', [])
1830 self._sort_extractor = sort_extractor
1832 def add_item(field, reverse, closest, limit_text):
1833 field = field.lower()
1834 if field in self._order:
1836 self._order.append(field)
1837 limit = self._resolve_field_value(field, limit_text)
1840 'closest': False if limit is None else closest,
1841 'limit_text': limit_text,
1843 if field in self.settings:
1844 self.settings[field].update(data)
1846 self.settings[field] = data
1849 tuple(field for field in self.default if self._get_field_setting(field, 'forced'))
1850 + (tuple() if params.get('format_sort_force', False)
1851 else tuple(field for field in self.default if self._get_field_setting(field, 'priority')))
1852 + tuple(self._sort_user) + tuple(sort_extractor) + self.default)
1854 for item in sort_list:
1855 match = re.match(self.regex, item)
1857 raise ExtractorError('Invalid format sort string "%s" given by extractor' % item)
1858 field = match.group('field')
1861 if self._get_field_setting(field, 'type') == 'alias':
1862 alias, field = field, self._get_field_setting(field, 'field')
1863 if self._get_field_setting(alias, 'deprecated'):
1864 self.ydl.deprecated_feature(f'Format sorting alias {alias} is deprecated and may '
1865 f'be removed in a future version. Please use {field} instead')
1866 reverse = match.group('reverse') is not None
1867 closest = match.group('separator') == '~'
1868 limit_text = match.group('limit')
1870 has_limit = limit_text is not None
1871 has_multiple_fields = self._get_field_setting(field, 'type') == 'combined'
1872 has_multiple_limits = has_limit and has_multiple_fields and not self._get_field_setting(field, 'same_limit')
1874 fields = self._get_field_setting(field, 'field') if has_multiple_fields else (field,)
1875 limits = limit_text.split(':') if has_multiple_limits else (limit_text,) if has_limit else tuple()
1876 limit_count = len(limits)
1877 for (i, f) in enumerate(fields):
1878 add_item(f, reverse, closest,
1879 limits[i] if i < limit_count
1880 else limits[0] if has_limit and not has_multiple_limits
1883 def print_verbose_info(self, write_debug):
1885 write_debug('Sort order given by user: %s' % ', '.join(self._sort_user))
1886 if self._sort_extractor:
1887 write_debug('Sort order given by extractor: %s' % ', '.join(self._sort_extractor))
1888 write_debug('Formats sorted by: %s' % ', '.join(['%s%s%s' % (
1889 '+' if self._get_field_setting(field, 'reverse') else '', field,
1890 '%s%s(%s)' % ('~' if self._get_field_setting(field, 'closest') else ':',
1891 self._get_field_setting(field, 'limit_text'),
1892 self._get_field_setting(field, 'limit'))
1893 if self._get_field_setting(field, 'limit_text') is not None else '')
1894 for field in self._order if self._get_field_setting(field, 'visible')]))
1896 def _calculate_field_preference_from_value(self, format, field, type, value):
1897 reverse = self._get_field_setting(field, 'reverse')
1898 closest = self._get_field_setting(field, 'closest')
1899 limit = self._get_field_setting(field, 'limit')
1901 if type == 'extractor':
1902 maximum = self._get_field_setting(field, 'max')
1903 if value is None or (maximum is not None and value >= maximum):
1905 elif type == 'boolean':
1906 in_list = self._get_field_setting(field, 'in_list')
1907 not_in_list = self._get_field_setting(field, 'not_in_list')
1908 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
1909 elif type == 'ordered':
1910 value = self._resolve_field_value(field, value, True)
1912 # try to convert to number
1913 val_num = float_or_none(value, default=self._get_field_setting(field, 'default'))
1914 is_num = self._get_field_setting(field, 'convert') != 'string' and val_num is not None
1918 return ((-10, 0) if value is None
1919 else (1, value, 0) if not is_num # if a field has mixed strings and numbers, strings are sorted higher
1920 else (0, -abs(value - limit), value - limit if reverse else limit - value) if closest
1921 else (0, value, 0) if not reverse and (limit is None or value <= limit)
1922 else (0, -value, 0) if limit is None or (reverse and value == limit) or value > limit
1923 else (-1, value, 0))
1925 def _calculate_field_preference(self, format, field):
1926 type = self._get_field_setting(field, 'type') # extractor, boolean, ordered, field, multiple
1927 get_value = lambda f: format.get(self._get_field_setting(f, 'field'))
1928 if type == 'multiple':
1929 type = 'field' # Only 'field' is allowed in multiple for now
1930 actual_fields = self._get_field_setting(field, 'field')
1932 value = self._get_field_setting(field, 'function')(get_value(f) for f in actual_fields)
1934 value = get_value(field)
1935 return self._calculate_field_preference_from_value(format, field, type, value)
1937 def calculate_preference(self, format):
1938 # Determine missing protocol
1939 if not format.get('protocol'):
1940 format['protocol'] = determine_protocol(format)
1942 # Determine missing ext
1943 if not format.get('ext') and 'url' in format:
1944 format['ext'] = determine_ext(format['url'])
1945 if format.get('vcodec') == 'none':
1946 format['audio_ext'] = format['ext'] if format.get('acodec') != 'none' else 'none'
1947 format['video_ext'] = 'none'
1949 format['video_ext'] = format['ext']
1950 format['audio_ext'] = 'none'
1951 # if format.get('preference') is None and format.get('ext') in ('f4f', 'f4m'): # Not supported?
1952 # format['preference'] = -1000
1954 # Determine missing bitrates
1955 if format.get('tbr') is None:
1956 if format.get('vbr') is not None and format.get('abr') is not None:
1957 format['tbr'] = format.get('vbr', 0) + format.get('abr', 0)
1959 if format.get('vcodec') != 'none' and format.get('vbr') is None:
1960 format['vbr'] = format.get('tbr') - format.get('abr', 0)
1961 if format.get('acodec') != 'none' and format.get('abr') is None:
1962 format['abr'] = format.get('tbr') - format.get('vbr', 0)
1964 return tuple(self._calculate_field_preference(format, field) for field in self._order)
1966 def _sort_formats(self, formats, field_preference=[]):
1969 formats.sort(key=self.FormatSort(self, field_preference).calculate_preference)
1971 def _check_formats(self, formats, video_id):
1973 formats[:] = filter(
1974 lambda f: self._is_valid_url(
1976 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1980 def _remove_duplicate_formats(formats):
1984 if f['url'] not in format_urls:
1985 format_urls.add(f['url'])
1986 unique_formats.append(f)
1987 formats[:] = unique_formats
1989 def _is_valid_url(self, url, video_id, item='video', headers={}):
1990 url = self._proto_relative_url(url, scheme='http:')
1991 # For now assume non HTTP(S) URLs always valid
1992 if not (url.startswith('http://') or url.startswith('https://')):
1995 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1997 except ExtractorError as e:
1999 '%s: %s URL is invalid, skipping: %s'
2000 % (video_id, item, error_to_compat_str(e.cause)))
2003 def http_scheme(self):
2004 """ Either "http
:" or "https
:", depending on the user's preferences """
2007 if self.get_param('prefer_insecure', False)
2010 def _proto_relative_url(self, url, scheme=None):
2011 scheme = scheme or self.http_scheme()
2012 assert scheme.endswith(':')
2013 return sanitize_url(url, scheme=scheme[:-1])
2015 def _sleep(self, timeout, video_id, msg_template=None):
2016 if msg_template is None:
2017 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
2018 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
2022 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
2023 transform_source=lambda s: fix_xml_ampersands(s).strip(),
2024 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
2025 res = self._download_xml_handle(
2026 manifest_url, video_id, 'Downloading f4m manifest',
2027 'Unable to download f4m manifest',
2028 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
2029 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
2030 transform_source=transform_source,
2031 fatal=fatal, data=data, headers=headers, query=query)
2035 manifest, urlh = res
2036 manifest_url = urlh.geturl()
2038 return self._parse_f4m_formats(
2039 manifest, manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
2040 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
2042 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
2043 transform_source=lambda s: fix_xml_ampersands(s).strip(),
2044 fatal=True, m3u8_id=None):
2045 if not isinstance(manifest, xml.etree.ElementTree.Element) and not fatal:
2048 # currently yt-dlp cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
2049 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
2050 if akamai_pv is not None and ';' in akamai_pv.text:
2051 playerVerificationChallenge = akamai_pv.text.split(';')[0]
2052 if playerVerificationChallenge.strip() != '':
2056 manifest_version = '1.0'
2057 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
2059 manifest_version = '2.0'
2060 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
2061 # Remove unsupported DRM protected media from final formats
2062 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
2063 media_nodes = remove_encrypted_media(media_nodes)
2067 manifest_base_url = get_base_url(manifest)
2069 bootstrap_info = xpath_element(
2070 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
2071 'bootstrap info', default=None)
2074 mime_type = xpath_text(
2075 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
2076 'base URL', default=None)
2077 if mime_type and mime_type.startswith('audio/'):
2080 for i, media_el in enumerate(media_nodes):
2081 tbr = int_or_none(media_el.attrib.get('bitrate'))
2082 width = int_or_none(media_el.attrib.get('width'))
2083 height = int_or_none(media_el.attrib.get('height'))
2084 format_id = join_nonempty(f4m_id, tbr or i)
2085 # If <bootstrapInfo> is present, the specified f4m is a
2086 # stream-level manifest, and only set-level manifests may refer to
2087 # external resources. See section 11.4 and section 4 of F4M spec
2088 if bootstrap_info is None:
2090 # @href is introduced in 2.0, see section 11.6 of F4M spec
2091 if manifest_version == '2.0':
2092 media_url = media_el.attrib.get('href')
2093 if media_url is None:
2094 media_url = media_el.attrib.get('url')
2098 media_url if media_url.startswith('http://') or media_url.startswith('https://')
2099 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
2100 # If media_url is itself a f4m manifest do the recursive extraction
2101 # since bitrates in parent manifest (this one) and media_url manifest
2102 # may differ leading to inability to resolve the format by requested
2103 # bitrate in f4m downloader
2104 ext = determine_ext(manifest_url)
2106 f4m_formats = self._extract_f4m_formats(
2107 manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
2108 transform_source=transform_source, fatal=fatal)
2109 # Sometimes stream-level manifest contains single media entry that
2110 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
2111 # At the same time parent's media entry in set-level manifest may
2112 # contain it. We will copy it from parent in such cases.
2113 if len(f4m_formats) == 1:
2116 'tbr': f.get('tbr') or tbr,
2117 'width': f.get('width') or width,
2118 'height': f.get('height') or height,
2119 'format_id': f.get('format_id') if not tbr else format_id,
2122 formats.extend(f4m_formats)
2125 formats.extend(self._extract_m3u8_formats(
2126 manifest_url, video_id, 'mp4', preference=preference,
2127 quality=quality, m3u8_id=m3u8_id, fatal=fatal))
2130 'format_id': format_id,
2131 'url': manifest_url,
2132 'manifest_url': manifest_url,
2133 'ext': 'flv' if bootstrap_info is not None else None,
2139 'preference': preference,
2144 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, quality=None, m3u8_id=None):
2146 'format_id': join_nonempty(m3u8_id, 'meta'),
2150 'preference': preference - 100 if preference else -100,
2152 'resolution': 'multiple',
2153 'format_note': 'Quality selection URL',
2156 def _report_ignoring_subs(self, name):
2157 self.report_warning(bug_reports_message(
2158 f'Ignoring subtitle tracks found in the {name} manifest; '
2159 'if any subtitle tracks are missing,'
2162 def _extract_m3u8_formats(self, *args, **kwargs):
2163 fmts, subs = self._extract_m3u8_formats_and_subtitles(*args, **kwargs)
2165 self._report_ignoring_subs('HLS')
2168 def _extract_m3u8_formats_and_subtitles(
2169 self, m3u8_url, video_id, ext=None, entry_protocol='m3u8_native',
2170 preference=None, quality=None, m3u8_id=None, note=None,
2171 errnote=None, fatal=True, live=False, data=None, headers={},
2174 res = self._download_webpage_handle(
2176 note='Downloading m3u8 information' if note is None else note,
2177 errnote='Failed to download m3u8 information' if errnote is None else errnote,
2178 fatal=fatal, data=data, headers=headers, query=query)
2183 m3u8_doc, urlh = res
2184 m3u8_url = urlh.geturl()
2186 return self._parse_m3u8_formats_and_subtitles(
2187 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
2188 preference=preference, quality=quality, m3u8_id=m3u8_id,
2189 note=note, errnote=errnote, fatal=fatal, live=live, data=data,
2190 headers=headers, query=query, video_id=video_id)
2192 def _parse_m3u8_formats_and_subtitles(
2193 self, m3u8_doc, m3u8_url=None, ext=None, entry_protocol='m3u8_native',
2194 preference=None, quality=None, m3u8_id=None, live=False, note=None,
2195 errnote=None, fatal=True, data=None, headers={}, query={},
2197 formats, subtitles = [], {}
2199 has_drm = re.search('|'.join([
2200 r'#EXT-X-FAXS-CM:', # Adobe Flash Access
2201 r'#EXT-X-(?:SESSION-)?KEY:.*?URI="skd
://', # Apple FairPlay
2204 def format_url(url):
2205 return url if re.match(r'^https?
://', url) else urllib.parse.urljoin(m3u8_url, url)
2207 if self.get_param('hls_split_discontinuity
', False):
2208 def _extract_m3u8_playlist_indices(manifest_url=None, m3u8_doc=None):
2210 if not manifest_url:
2212 m3u8_doc = self._download_webpage(
2213 manifest_url, video_id, fatal=fatal, data=data, headers=headers,
2214 note=False, errnote='Failed to download m3u8 playlist information
')
2215 if m3u8_doc is False:
2217 return range(1 + sum(line.startswith('#EXT-X-DISCONTINUITY') for line in m3u8_doc.splitlines()))
2220 def _extract_m3u8_playlist_indices(*args
, **kwargs
):
2224 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
2225 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
2226 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
2228 # We should try extracting formats only from master playlists [1, 4.3.4],
2229 # i.e. playlists that describe available qualities. On the other hand
2230 # media playlists [1, 4.3.3] should be returned as is since they contain
2231 # just the media without qualities renditions.
2232 # Fortunately, master playlist can be easily distinguished from media
2233 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
2234 # master playlist tags MUST NOT appear in a media playlist and vice versa.
2235 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
2236 # media playlist and MUST NOT appear in master playlist thus we can
2237 # clearly detect media playlist with this criterion.
2239 if '#EXT-X-TARGETDURATION' in m3u8_doc
: # media playlist, return as is
2241 'format_id': join_nonempty(m3u8_id
, idx
),
2242 'format_index': idx
,
2243 'url': m3u8_url
or encode_data_uri(m3u8_doc
.encode('utf-8'), 'application/x-mpegurl'),
2245 'protocol': entry_protocol
,
2246 'preference': preference
,
2249 } for idx
in _extract_m3u8_playlist_indices(m3u8_doc
=m3u8_doc
)]
2251 return formats
, subtitles
2254 last_stream_inf
= {}
2256 def extract_media(x_media_line
):
2257 media
= parse_m3u8_attributes(x_media_line
)
2258 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
2259 media_type
, group_id
, name
= media
.get('TYPE'), media
.get('GROUP-ID'), media
.get('NAME')
2260 if not (media_type
and group_id
and name
):
2262 groups
.setdefault(group_id
, []).append(media
)
2263 # <https://tools.ietf.org/html/rfc8216#section-4.3.4.1>
2264 if media_type
== 'SUBTITLES':
2265 # According to RFC 8216 §4.3.4.2.1, URI is REQUIRED in the
2266 # EXT-X-MEDIA tag if the media type is SUBTITLES.
2267 # However, lack of URI has been spotted in the wild.
2268 # e.g. NebulaIE; see https://github.com/yt-dlp/yt-dlp/issues/339
2269 if not media
.get('URI'):
2271 url
= format_url(media
['URI'])
2274 'ext': determine_ext(url
),
2276 if sub_info
['ext'] == 'm3u8':
2277 # Per RFC 8216 §3.1, the only possible subtitle format m3u8
2278 # files may contain is WebVTT:
2279 # <https://tools.ietf.org/html/rfc8216#section-3.1>
2280 sub_info
['ext'] = 'vtt'
2281 sub_info
['protocol'] = 'm3u8_native'
2282 lang
= media
.get('LANGUAGE') or 'und'
2283 subtitles
.setdefault(lang
, []).append(sub_info
)
2284 if media_type
not in ('VIDEO', 'AUDIO'):
2286 media_url
= media
.get('URI')
2288 manifest_url
= format_url(media_url
)
2290 'format_id': join_nonempty(m3u8_id
, group_id
, name
, idx
),
2291 'format_note': name
,
2292 'format_index': idx
,
2293 'url': manifest_url
,
2294 'manifest_url': m3u8_url
,
2295 'language': media
.get('LANGUAGE'),
2297 'protocol': entry_protocol
,
2298 'preference': preference
,
2300 'vcodec': 'none' if media_type
== 'AUDIO' else None,
2301 } for idx
in _extract_m3u8_playlist_indices(manifest_url
))
2303 def build_stream_name():
2304 # Despite specification does not mention NAME attribute for
2305 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
2306 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
2307 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
2308 stream_name
= last_stream_inf
.get('NAME')
2311 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
2312 # from corresponding rendition group
2313 stream_group_id
= last_stream_inf
.get('VIDEO')
2314 if not stream_group_id
:
2316 stream_group
= groups
.get(stream_group_id
)
2317 if not stream_group
:
2318 return stream_group_id
2319 rendition
= stream_group
[0]
2320 return rendition
.get('NAME') or stream_group_id
2322 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
2323 # chance to detect video only formats when EXT-X-STREAM-INF tags
2324 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
2325 for line
in m3u8_doc
.splitlines():
2326 if line
.startswith('#EXT-X-MEDIA:'):
2329 for line
in m3u8_doc
.splitlines():
2330 if line
.startswith('#EXT-X-STREAM-INF:'):
2331 last_stream_inf
= parse_m3u8_attributes(line
)
2332 elif line
.startswith('#') or not line
.strip():
2335 tbr
= float_or_none(
2336 last_stream_inf
.get('AVERAGE-BANDWIDTH')
2337 or last_stream_inf
.get('BANDWIDTH'), scale
=1000)
2338 manifest_url
= format_url(line
.strip())
2340 for idx
in _extract_m3u8_playlist_indices(manifest_url
):
2341 format_id
= [m3u8_id
, None, idx
]
2342 # Bandwidth of live streams may differ over time thus making
2343 # format_id unpredictable. So it's better to keep provided
2346 stream_name
= build_stream_name()
2347 format_id
[1] = stream_name
or '%d' % (tbr
or len(formats
))
2349 'format_id': join_nonempty(*format_id
),
2350 'format_index': idx
,
2351 'url': manifest_url
,
2352 'manifest_url': m3u8_url
,
2355 'fps': float_or_none(last_stream_inf
.get('FRAME-RATE')),
2356 'protocol': entry_protocol
,
2357 'preference': preference
,
2360 resolution
= last_stream_inf
.get('RESOLUTION')
2362 mobj
= re
.search(r
'(?P<width>\d+)[xX](?P<height>\d+)', resolution
)
2364 f
['width'] = int(mobj
.group('width'))
2365 f
['height'] = int(mobj
.group('height'))
2366 # Unified Streaming Platform
2368 r
'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f
['url'])
2370 abr
, vbr
= mobj
.groups()
2371 abr
, vbr
= float_or_none(abr
, 1000), float_or_none(vbr
, 1000)
2376 codecs
= parse_codecs(last_stream_inf
.get('CODECS'))
2378 audio_group_id
= last_stream_inf
.get('AUDIO')
2379 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
2380 # references a rendition group MUST have a CODECS attribute.
2381 # However, this is not always respected. E.g. [2]
2382 # contains EXT-X-STREAM-INF tag which references AUDIO
2383 # rendition group but does not have CODECS and despite
2384 # referencing an audio group it represents a complete
2385 # (with audio and video) format. So, for such cases we will
2386 # ignore references to rendition groups and treat them
2387 # as complete formats.
2388 if audio_group_id
and codecs
and f
.get('vcodec') != 'none':
2389 audio_group
= groups
.get(audio_group_id
)
2390 if audio_group
and audio_group
[0].get('URI'):
2391 # TODO: update acodec for audio only formats with
2393 f
['acodec'] = 'none'
2394 if not f
.get('ext'):
2395 f
['ext'] = 'm4a' if f
.get('vcodec') == 'none' else 'mp4'
2399 progressive_uri
= last_stream_inf
.get('PROGRESSIVE-URI')
2402 del http_f
['manifest_url']
2404 'format_id': f
['format_id'].replace('hls-', 'http-'),
2406 'url': progressive_uri
,
2408 formats
.append(http_f
)
2410 last_stream_inf
= {}
2411 return formats
, subtitles
2413 def _extract_m3u8_vod_duration(
2414 self
, m3u8_vod_url
, video_id
, note
=None, errnote
=None, data
=None, headers
={}, query={}
):
2416 m3u8_vod
= self
._download
_webpage
(
2417 m3u8_vod_url
, video_id
,
2418 note
='Downloading m3u8 VOD manifest' if note
is None else note
,
2419 errnote
='Failed to download VOD manifest' if errnote
is None else errnote
,
2420 fatal
=False, data
=data
, headers
=headers
, query
=query
)
2422 return self
._parse
_m
3u8_vod
_duration
(m3u8_vod
or '', video_id
)
2424 def _parse_m3u8_vod_duration(self
, m3u8_vod
, video_id
):
2425 if '#EXT-X-PLAYLIST-TYPE:VOD' not in m3u8_vod
:
2429 float(line
[len('#EXTINF:'):].split(',')[0])
2430 for line
in m3u8_vod
.splitlines() if line
.startswith('#EXTINF:'))) or None
2433 def _xpath_ns(path
, namespace
=None):
2437 for c
in path
.split('/'):
2438 if not c
or c
== '.':
2441 out
.append('{%s}%s' % (namespace
, c
))
2442 return '/'.join(out
)
2444 def _extract_smil_formats_and_subtitles(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None, transform_source
=None):
2445 res
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
, transform_source
=transform_source
)
2451 smil_url
= urlh
.geturl()
2453 namespace
= self
._parse
_smil
_namespace
(smil
)
2455 fmts
= self
._parse
_smil
_formats
(
2456 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2457 subs
= self
._parse
_smil
_subtitles
(
2458 smil
, namespace
=namespace
)
2462 def _extract_smil_formats(self
, *args
, **kwargs
):
2463 fmts
, subs
= self
._extract
_smil
_formats
_and
_subtitles
(*args
, **kwargs
)
2465 self
._report
_ignoring
_subs
('SMIL')
2468 def _extract_smil_info(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None):
2469 res
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
)
2474 smil_url
= urlh
.geturl()
2476 return self
._parse
_smil
(smil
, smil_url
, video_id
, f4m_params
=f4m_params
)
2478 def _download_smil(self
, smil_url
, video_id
, fatal
=True, transform_source
=None):
2479 return self
._download
_xml
_handle
(
2480 smil_url
, video_id
, 'Downloading SMIL file',
2481 'Unable to download SMIL file', fatal
=fatal
, transform_source
=transform_source
)
2483 def _parse_smil(self
, smil
, smil_url
, video_id
, f4m_params
=None):
2484 namespace
= self
._parse
_smil
_namespace
(smil
)
2486 formats
= self
._parse
_smil
_formats
(
2487 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2488 subtitles
= self
._parse
_smil
_subtitles
(smil
, namespace
=namespace
)
2490 video_id
= os
.path
.splitext(url_basename(smil_url
))[0]
2494 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2495 name
= meta
.attrib
.get('name')
2496 content
= meta
.attrib
.get('content')
2497 if not name
or not content
:
2499 if not title
and name
== 'title':
2501 elif not description
and name
in ('description', 'abstract'):
2502 description
= content
2503 elif not upload_date
and name
== 'date':
2504 upload_date
= unified_strdate(content
)
2507 'id': image
.get('type'),
2508 'url': image
.get('src'),
2509 'width': int_or_none(image
.get('width')),
2510 'height': int_or_none(image
.get('height')),
2511 } for image
in smil
.findall(self
._xpath
_ns
('.//image', namespace
)) if image
.get('src')]
2515 'title': title
or video_id
,
2516 'description': description
,
2517 'upload_date': upload_date
,
2518 'thumbnails': thumbnails
,
2520 'subtitles': subtitles
,
2523 def _parse_smil_namespace(self
, smil
):
2524 return self
._search
_regex
(
2525 r
'(?i)^{([^}]+)?}smil$', smil
.tag
, 'namespace', default
=None)
2527 def _parse_smil_formats(self
, smil
, smil_url
, video_id
, namespace
=None, f4m_params
=None, transform_rtmp_url
=None):
2529 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2530 b
= meta
.get('base') or meta
.get('httpBase')
2542 media
= smil
.findall(self
._xpath
_ns
('.//video', namespace
)) + smil
.findall(self
._xpath
_ns
('.//audio', namespace
))
2543 for medium
in media
:
2544 src
= medium
.get('src')
2545 if not src
or src
in srcs
:
2549 bitrate
= float_or_none(medium
.get('system-bitrate') or medium
.get('systemBitrate'), 1000)
2550 filesize
= int_or_none(medium
.get('size') or medium
.get('fileSize'))
2551 width
= int_or_none(medium
.get('width'))
2552 height
= int_or_none(medium
.get('height'))
2553 proto
= medium
.get('proto')
2554 ext
= medium
.get('ext')
2555 src_ext
= determine_ext(src
)
2556 streamer
= medium
.get('streamer') or base
2558 if proto
== 'rtmp' or streamer
.startswith('rtmp'):
2564 'format_id': 'rtmp-%d' % (rtmp_count
if bitrate
is None else bitrate
),
2566 'filesize': filesize
,
2570 if transform_rtmp_url
:
2571 streamer
, src
= transform_rtmp_url(streamer
, src
)
2572 formats
[-1].update({
2578 src_url
= src
if src
.startswith('http') else urllib
.parse
.urljoin(base
, src
)
2579 src_url
= src_url
.strip()
2581 if proto
== 'm3u8' or src_ext
== 'm3u8':
2582 m3u8_formats
= self
._extract
_m
3u8_formats
(
2583 src_url
, video_id
, ext
or 'mp4', m3u8_id
='hls', fatal
=False)
2584 if len(m3u8_formats
) == 1:
2586 m3u8_formats
[0].update({
2587 'format_id': 'hls-%d' % (m3u8_count
if bitrate
is None else bitrate
),
2592 formats
.extend(m3u8_formats
)
2593 elif src_ext
== 'f4m':
2598 'plugin': 'flowplayer-3.2.0.1',
2600 f4m_url
+= '&' if '?' in f4m_url
else '?'
2601 f4m_url
+= urllib
.parse
.urlencode(f4m_params
)
2602 formats
.extend(self
._extract
_f
4m
_formats
(f4m_url
, video_id
, f4m_id
='hds', fatal
=False))
2603 elif src_ext
== 'mpd':
2604 formats
.extend(self
._extract
_mpd
_formats
(
2605 src_url
, video_id
, mpd_id
='dash', fatal
=False))
2606 elif re
.search(r
'\.ism/[Mm]anifest', src_url
):
2607 formats
.extend(self
._extract
_ism
_formats
(
2608 src_url
, video_id
, ism_id
='mss', fatal
=False))
2609 elif src_url
.startswith('http') and self
._is
_valid
_url
(src
, video_id
):
2613 'ext': ext
or src_ext
or 'flv',
2614 'format_id': 'http-%d' % (bitrate
or http_count
),
2616 'filesize': filesize
,
2621 for medium
in smil
.findall(self
._xpath
_ns
('.//imagestream', namespace
)):
2622 src
= medium
.get('src')
2623 if not src
or src
in srcs
:
2629 'format_id': 'imagestream-%d' % (imgs_count
),
2631 'ext': mimetype2ext(medium
.get('type')),
2634 'width': int_or_none(medium
.get('width')),
2635 'height': int_or_none(medium
.get('height')),
2636 'format_note': 'SMIL storyboards',
2641 def _parse_smil_subtitles(self
, smil
, namespace
=None, subtitles_lang
='en'):
2644 for num
, textstream
in enumerate(smil
.findall(self
._xpath
_ns
('.//textstream', namespace
))):
2645 src
= textstream
.get('src')
2646 if not src
or src
in urls
:
2649 ext
= textstream
.get('ext') or mimetype2ext(textstream
.get('type')) or determine_ext(src
)
2650 lang
= textstream
.get('systemLanguage') or textstream
.get('systemLanguageName') or textstream
.get('lang') or subtitles_lang
2651 subtitles
.setdefault(lang
, []).append({
2657 def _extract_xspf_playlist(self
, xspf_url
, playlist_id
, fatal
=True):
2658 res
= self
._download
_xml
_handle
(
2659 xspf_url
, playlist_id
, 'Downloading xpsf playlist',
2660 'Unable to download xspf manifest', fatal
=fatal
)
2665 xspf_url
= urlh
.geturl()
2667 return self
._parse
_xspf
(
2668 xspf
, playlist_id
, xspf_url
=xspf_url
,
2669 xspf_base_url
=base_url(xspf_url
))
2671 def _parse_xspf(self
, xspf_doc
, playlist_id
, xspf_url
=None, xspf_base_url
=None):
2673 'xspf': 'http://xspf.org/ns/0/',
2674 's1': 'http://static.streamone.nl/player/ns/0',
2678 for track
in xspf_doc
.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP
)):
2680 track
, xpath_with_ns('./xspf:title', NS_MAP
), 'title', default
=playlist_id
)
2681 description
= xpath_text(
2682 track
, xpath_with_ns('./xspf:annotation', NS_MAP
), 'description')
2683 thumbnail
= xpath_text(
2684 track
, xpath_with_ns('./xspf:image', NS_MAP
), 'thumbnail')
2685 duration
= float_or_none(
2686 xpath_text(track
, xpath_with_ns('./xspf:duration', NS_MAP
), 'duration'), 1000)
2689 for location
in track
.findall(xpath_with_ns('./xspf:location', NS_MAP
)):
2690 format_url
= urljoin(xspf_base_url
, location
.text
)
2695 'manifest_url': xspf_url
,
2696 'format_id': location
.get(xpath_with_ns('s1:label', NS_MAP
)),
2697 'width': int_or_none(location
.get(xpath_with_ns('s1:width', NS_MAP
))),
2698 'height': int_or_none(location
.get(xpath_with_ns('s1:height', NS_MAP
))),
2700 self
._sort
_formats
(formats
)
2705 'description': description
,
2706 'thumbnail': thumbnail
,
2707 'duration': duration
,
2712 def _extract_mpd_formats(self
, *args
, **kwargs
):
2713 fmts
, subs
= self
._extract
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2715 self
._report
_ignoring
_subs
('DASH')
2718 def _extract_mpd_formats_and_subtitles(
2719 self
, mpd_url
, video_id
, mpd_id
=None, note
=None, errnote
=None,
2720 fatal
=True, data
=None, headers
={}, query={}
):
2721 res
= self
._download
_xml
_handle
(
2723 note
='Downloading MPD manifest' if note
is None else note
,
2724 errnote
='Failed to download MPD manifest' if errnote
is None else errnote
,
2725 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
2732 # We could have been redirected to a new url when we retrieved our mpd file.
2733 mpd_url
= urlh
.geturl()
2734 mpd_base_url
= base_url(mpd_url
)
2736 return self
._parse
_mpd
_formats
_and
_subtitles
(
2737 mpd_doc
, mpd_id
, mpd_base_url
, mpd_url
)
2739 def _parse_mpd_formats(self
, *args
, **kwargs
):
2740 fmts
, subs
= self
._parse
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2742 self
._report
_ignoring
_subs
('DASH')
2745 def _parse_mpd_formats_and_subtitles(
2746 self
, mpd_doc
, mpd_id
=None, mpd_base_url
='', mpd_url
=None):
2748 Parse formats from MPD manifest.
2750 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
2751 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
2752 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
2754 if not self
.get_param('dynamic_mpd', True):
2755 if mpd_doc
.get('type') == 'dynamic':
2758 namespace
= self
._search
_regex
(r
'(?i)^{([^}]+)?}MPD$', mpd_doc
.tag
, 'namespace', default
=None)
2761 return self
._xpath
_ns
(path
, namespace
)
2763 def is_drm_protected(element
):
2764 return element
.find(_add_ns('ContentProtection')) is not None
2766 def extract_multisegment_info(element
, ms_parent_info
):
2767 ms_info
= ms_parent_info
.copy()
2769 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2770 # common attributes and elements. We will only extract relevant
2772 def extract_common(source
):
2773 segment_timeline
= source
.find(_add_ns('SegmentTimeline'))
2774 if segment_timeline
is not None:
2775 s_e
= segment_timeline
.findall(_add_ns('S'))
2777 ms_info
['total_number'] = 0
2780 r
= int(s
.get('r', 0))
2781 ms_info
['total_number'] += 1 + r
2782 ms_info
['s'].append({
2783 't': int(s
.get('t', 0)),
2784 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2785 'd': int(s
.attrib
['d']),
2788 start_number
= source
.get('startNumber')
2790 ms_info
['start_number'] = int(start_number
)
2791 timescale
= source
.get('timescale')
2793 ms_info
['timescale'] = int(timescale
)
2794 segment_duration
= source
.get('duration')
2795 if segment_duration
:
2796 ms_info
['segment_duration'] = float(segment_duration
)
2798 def extract_Initialization(source
):
2799 initialization
= source
.find(_add_ns('Initialization'))
2800 if initialization
is not None:
2801 ms_info
['initialization_url'] = initialization
.attrib
['sourceURL']
2803 segment_list
= element
.find(_add_ns('SegmentList'))
2804 if segment_list
is not None:
2805 extract_common(segment_list
)
2806 extract_Initialization(segment_list
)
2807 segment_urls_e
= segment_list
.findall(_add_ns('SegmentURL'))
2809 ms_info
['segment_urls'] = [segment
.attrib
['media'] for segment
in segment_urls_e
]
2811 segment_template
= element
.find(_add_ns('SegmentTemplate'))
2812 if segment_template
is not None:
2813 extract_common(segment_template
)
2814 media
= segment_template
.get('media')
2816 ms_info
['media'] = media
2817 initialization
= segment_template
.get('initialization')
2819 ms_info
['initialization'] = initialization
2821 extract_Initialization(segment_template
)
2824 mpd_duration
= parse_duration(mpd_doc
.get('mediaPresentationDuration'))
2825 formats
, subtitles
= [], {}
2826 stream_numbers
= collections
.defaultdict(int)
2827 for period
in mpd_doc
.findall(_add_ns('Period')):
2828 period_duration
= parse_duration(period
.get('duration')) or mpd_duration
2829 period_ms_info
= extract_multisegment_info(period
, {
2833 for adaptation_set
in period
.findall(_add_ns('AdaptationSet')):
2834 adaption_set_ms_info
= extract_multisegment_info(adaptation_set
, period_ms_info
)
2835 for representation
in adaptation_set
.findall(_add_ns('Representation')):
2836 representation_attrib
= adaptation_set
.attrib
.copy()
2837 representation_attrib
.update(representation
.attrib
)
2838 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2839 mime_type
= representation_attrib
['mimeType']
2840 content_type
= representation_attrib
.get('contentType', mime_type
.split('/')[0])
2842 codec_str
= representation_attrib
.get('codecs', '')
2843 # Some kind of binary subtitle found in some youtube livestreams
2844 if mime_type
== 'application/x-rawcc':
2845 codecs
= {'scodec': codec_str}
2847 codecs
= parse_codecs(codec_str
)
2848 if content_type
not in ('video', 'audio', 'text'):
2849 if mime_type
== 'image/jpeg':
2850 content_type
= mime_type
2851 elif codecs
.get('vcodec', 'none') != 'none':
2852 content_type
= 'video'
2853 elif codecs
.get('acodec', 'none') != 'none':
2854 content_type
= 'audio'
2855 elif codecs
.get('scodec', 'none') != 'none':
2856 content_type
= 'text'
2857 elif mimetype2ext(mime_type
) in ('tt', 'dfxp', 'ttml', 'xml', 'json'):
2858 content_type
= 'text'
2860 self
.report_warning('Unknown MIME type %s in DASH manifest' % mime_type
)
2864 for element
in (representation
, adaptation_set
, period
, mpd_doc
):
2865 base_url_e
= element
.find(_add_ns('BaseURL'))
2866 if try_call(lambda: base_url_e
.text
) is not None:
2867 base_url
= base_url_e
.text
+ base_url
2868 if re
.match(r
'^https?://', base_url
):
2870 if mpd_base_url
and base_url
.startswith('/'):
2871 base_url
= urllib
.parse
.urljoin(mpd_base_url
, base_url
)
2872 elif mpd_base_url
and not re
.match(r
'^https?://', base_url
):
2873 if not mpd_base_url
.endswith('/'):
2875 base_url
= mpd_base_url
+ base_url
2876 representation_id
= representation_attrib
.get('id')
2877 lang
= representation_attrib
.get('lang')
2878 url_el
= representation
.find(_add_ns('BaseURL'))
2879 filesize
= int_or_none(url_el
.attrib
.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el
is not None else None)
2880 bandwidth
= int_or_none(representation_attrib
.get('bandwidth'))
2881 if representation_id
is not None:
2882 format_id
= representation_id
2884 format_id
= content_type
2886 format_id
= mpd_id
+ '-' + format_id
2887 if content_type
in ('video', 'audio'):
2889 'format_id': format_id
,
2890 'manifest_url': mpd_url
,
2891 'ext': mimetype2ext(mime_type
),
2892 'width': int_or_none(representation_attrib
.get('width')),
2893 'height': int_or_none(representation_attrib
.get('height')),
2894 'tbr': float_or_none(bandwidth
, 1000),
2895 'asr': int_or_none(representation_attrib
.get('audioSamplingRate')),
2896 'fps': int_or_none(representation_attrib
.get('frameRate')),
2897 'language': lang
if lang
not in ('mul', 'und', 'zxx', 'mis') else None,
2898 'format_note': 'DASH %s' % content_type
,
2899 'filesize': filesize
,
2900 'container': mimetype2ext(mime_type
) + '_dash',
2903 elif content_type
== 'text':
2905 'ext': mimetype2ext(mime_type
),
2906 'manifest_url': mpd_url
,
2907 'filesize': filesize
,
2909 elif content_type
== 'image/jpeg':
2910 # See test case in VikiIE
2911 # https://www.viki.com/videos/1175236v-choosing-spouse-by-lottery-episode-1
2913 'format_id': format_id
,
2915 'manifest_url': mpd_url
,
2916 'format_note': 'DASH storyboards (jpeg)',
2920 if is_drm_protected(adaptation_set
) or is_drm_protected(representation
):
2922 representation_ms_info
= extract_multisegment_info(representation
, adaption_set_ms_info
)
2924 def prepare_template(template_name
, identifiers
):
2925 tmpl
= representation_ms_info
[template_name
]
2926 if representation_id
is not None:
2927 tmpl
= tmpl
.replace('$RepresentationID$', representation_id
)
2928 # First of, % characters outside $...$ templates
2929 # must be escaped by doubling for proper processing
2930 # by % operator string formatting used further (see
2931 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2937 in_template
= not in_template
2938 elif c
== '%' and not in_template
:
2940 # Next, $...$ templates are translated to their
2941 # %(...) counterparts to be used with % operator
2942 t
= re
.sub(r
'\$(%s)\$' % '|'.join(identifiers
), r
'%(\1)d', t
)
2943 t
= re
.sub(r
'\$(%s)%%([^$]+)\$' % '|'.join(identifiers
), r
'%(\1)\2', t
)
2944 t
.replace('$$', '$')
2947 # @initialization is a regular template like @media one
2948 # so it should be handled just the same way (see
2949 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2950 if 'initialization' in representation_ms_info
:
2951 initialization_template
= prepare_template(
2953 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2954 # $Time$ shall not be included for @initialization thus
2955 # only $Bandwidth$ remains
2957 representation_ms_info
['initialization_url'] = initialization_template
% {
2958 'Bandwidth': bandwidth
,
2961 def location_key(location
):
2962 return 'url' if re
.match(r
'^https?://', location
) else 'path'
2964 if 'segment_urls' not in representation_ms_info
and 'media' in representation_ms_info
:
2966 media_template
= prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2967 media_location_key
= location_key(media_template
)
2969 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2970 # can't be used at the same time
2971 if '%(Number' in media_template
and 's' not in representation_ms_info
:
2972 segment_duration
= None
2973 if 'total_number' not in representation_ms_info
and 'segment_duration' in representation_ms_info
:
2974 segment_duration
= float_or_none(representation_ms_info
['segment_duration'], representation_ms_info
['timescale'])
2975 representation_ms_info
['total_number'] = int(math
.ceil(
2976 float_or_none(period_duration
, segment_duration
, default
=0)))
2977 representation_ms_info
['fragments'] = [{
2978 media_location_key
: media_template
% {
2979 'Number': segment_number
,
2980 'Bandwidth': bandwidth
,
2982 'duration': segment_duration
,
2983 } for segment_number
in range(
2984 representation_ms_info
['start_number'],
2985 representation_ms_info
['total_number'] + representation_ms_info
['start_number'])]
2987 # $Number*$ or $Time$ in media template with S list available
2988 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2989 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2990 representation_ms_info
['fragments'] = []
2993 segment_number
= representation_ms_info
['start_number']
2995 def add_segment_url():
2996 segment_url
= media_template
% {
2997 'Time': segment_time
,
2998 'Bandwidth': bandwidth
,
2999 'Number': segment_number
,
3001 representation_ms_info
['fragments'].append({
3002 media_location_key
: segment_url
,
3003 'duration': float_or_none(segment_d
, representation_ms_info
['timescale']),
3006 for num
, s
in enumerate(representation_ms_info
['s']):
3007 segment_time
= s
.get('t') or segment_time
3011 for r
in range(s
.get('r', 0)):
3012 segment_time
+= segment_d
3015 segment_time
+= segment_d
3016 elif 'segment_urls' in representation_ms_info
and 's' in representation_ms_info
:
3017 # No media template,
3018 # e.g. https://www.youtube.com/watch?v=iXZV5uAYMJI
3019 # or any YouTube dashsegments video
3022 timescale
= representation_ms_info
['timescale']
3023 for s
in representation_ms_info
['s']:
3024 duration
= float_or_none(s
['d'], timescale
)
3025 for r
in range(s
.get('r', 0) + 1):
3026 segment_uri
= representation_ms_info
['segment_urls'][segment_index
]
3028 location_key(segment_uri
): segment_uri
,
3029 'duration': duration
,
3032 representation_ms_info
['fragments'] = fragments
3033 elif 'segment_urls' in representation_ms_info
:
3034 # Segment URLs with no SegmentTimeline
3035 # E.g. https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
3036 # https://github.com/ytdl-org/youtube-dl/pull/14844
3038 segment_duration
= float_or_none(
3039 representation_ms_info
['segment_duration'],
3040 representation_ms_info
['timescale']) if 'segment_duration' in representation_ms_info
else None
3041 for segment_url
in representation_ms_info
['segment_urls']:
3043 location_key(segment_url
): segment_url
,
3045 if segment_duration
:
3046 fragment
['duration'] = segment_duration
3047 fragments
.append(fragment
)
3048 representation_ms_info
['fragments'] = fragments
3049 # If there is a fragments key available then we correctly recognized fragmented media.
3050 # Otherwise we will assume unfragmented media with direct access. Technically, such
3051 # assumption is not necessarily correct since we may simply have no support for
3052 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
3053 if 'fragments' in representation_ms_info
:
3055 # NB: mpd_url may be empty when MPD manifest is parsed from a string
3056 'url': mpd_url
or base_url
,
3057 'fragment_base_url': base_url
,
3059 'protocol': 'http_dash_segments' if mime_type
!= 'image/jpeg' else 'mhtml',
3061 if 'initialization_url' in representation_ms_info
:
3062 initialization_url
= representation_ms_info
['initialization_url']
3063 if not f
.get('url'):
3064 f
['url'] = initialization_url
3065 f
['fragments'].append({location_key(initialization_url): initialization_url}
)
3066 f
['fragments'].extend(representation_ms_info
['fragments'])
3067 if not period_duration
:
3068 period_duration
= try_get(
3069 representation_ms_info
,
3070 lambda r
: sum(frag
['duration'] for frag
in r
['fragments']), float)
3072 # Assuming direct URL to unfragmented media.
3074 if content_type
in ('video', 'audio', 'image/jpeg'):
3075 f
['manifest_stream_number'] = stream_numbers
[f
['url']]
3076 stream_numbers
[f
['url']] += 1
3078 elif content_type
== 'text':
3079 subtitles
.setdefault(lang
or 'und', []).append(f
)
3081 return formats
, subtitles
3083 def _extract_ism_formats(self
, *args
, **kwargs
):
3084 fmts
, subs
= self
._extract
_ism
_formats
_and
_subtitles
(*args
, **kwargs
)
3086 self
._report
_ignoring
_subs
('ISM')
3089 def _extract_ism_formats_and_subtitles(self
, ism_url
, video_id
, ism_id
=None, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
):
3090 res
= self
._download
_xml
_handle
(
3092 note
='Downloading ISM manifest' if note
is None else note
,
3093 errnote
='Failed to download ISM manifest' if errnote
is None else errnote
,
3094 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
3101 return self
._parse
_ism
_formats
_and
_subtitles
(ism_doc
, urlh
.geturl(), ism_id
)
3103 def _parse_ism_formats_and_subtitles(self
, ism_doc
, ism_url
, ism_id
=None):
3105 Parse formats from ISM manifest.
3107 1. [MS-SSTR]: Smooth Streaming Protocol,
3108 https://msdn.microsoft.com/en-us/library/ff469518.aspx
3110 if ism_doc
.get('IsLive') == 'TRUE':
3113 duration
= int(ism_doc
.attrib
['Duration'])
3114 timescale
= int_or_none(ism_doc
.get('TimeScale')) or 10000000
3118 for stream
in ism_doc
.findall('StreamIndex'):
3119 stream_type
= stream
.get('Type')
3120 if stream_type
not in ('video', 'audio', 'text'):
3122 url_pattern
= stream
.attrib
['Url']
3123 stream_timescale
= int_or_none(stream
.get('TimeScale')) or timescale
3124 stream_name
= stream
.get('Name')
3125 stream_language
= stream
.get('Language', 'und')
3126 for track
in stream
.findall('QualityLevel'):
3127 KNOWN_TAGS
= {'255': 'AACL', '65534': 'EC-3'}
3128 fourcc
= track
.get('FourCC') or KNOWN_TAGS
.get(track
.get('AudioTag'))
3129 # TODO: add support for WVC1 and WMAP
3130 if fourcc
not in ('H264', 'AVC1', 'AACL', 'TTML', 'EC-3'):
3131 self
.report_warning('%s is not a supported codec' % fourcc
)
3133 tbr
= int(track
.attrib
['Bitrate']) // 1000
3134 # [1] does not mention Width and Height attributes. However,
3135 # they're often present while MaxWidth and MaxHeight are
3136 # missing, so should be used as fallbacks
3137 width
= int_or_none(track
.get('MaxWidth') or track
.get('Width'))
3138 height
= int_or_none(track
.get('MaxHeight') or track
.get('Height'))
3139 sampling_rate
= int_or_none(track
.get('SamplingRate'))
3141 track_url_pattern
= re
.sub(r
'{[Bb]itrate}', track
.attrib
['Bitrate'], url_pattern
)
3142 track_url_pattern
= urllib
.parse
.urljoin(ism_url
, track_url_pattern
)
3148 stream_fragments
= stream
.findall('c')
3149 for stream_fragment_index
, stream_fragment
in enumerate(stream_fragments
):
3150 fragment_ctx
['time'] = int_or_none(stream_fragment
.get('t')) or fragment_ctx
['time']
3151 fragment_repeat
= int_or_none(stream_fragment
.get('r')) or 1
3152 fragment_ctx
['duration'] = int_or_none(stream_fragment
.get('d'))
3153 if not fragment_ctx
['duration']:
3155 next_fragment_time
= int(stream_fragment
[stream_fragment_index
+ 1].attrib
['t'])
3157 next_fragment_time
= duration
3158 fragment_ctx
['duration'] = (next_fragment_time
- fragment_ctx
['time']) / fragment_repeat
3159 for _
in range(fragment_repeat
):
3161 'url': re
.sub(r
'{start[ _]time}', str(fragment_ctx
['time']), track_url_pattern
),
3162 'duration': fragment_ctx
['duration'] / stream_timescale
,
3164 fragment_ctx
['time'] += fragment_ctx
['duration']
3166 if stream_type
== 'text':
3167 subtitles
.setdefault(stream_language
, []).append({
3171 'manifest_url': ism_url
,
3172 'fragments': fragments
,
3173 '_download_params': {
3174 'stream_type': stream_type
,
3175 'duration': duration
,
3176 'timescale': stream_timescale
,
3178 'language': stream_language
,
3179 'codec_private_data': track
.get('CodecPrivateData'),
3182 elif stream_type
in ('video', 'audio'):
3184 'format_id': join_nonempty(ism_id
, stream_name
, tbr
),
3186 'manifest_url': ism_url
,
3187 'ext': 'ismv' if stream_type
== 'video' else 'isma',
3191 'asr': sampling_rate
,
3192 'vcodec': 'none' if stream_type
== 'audio' else fourcc
,
3193 'acodec': 'none' if stream_type
== 'video' else fourcc
,
3195 'fragments': fragments
,
3196 'has_drm': ism_doc
.find('Protection') is not None,
3197 '_download_params': {
3198 'stream_type': stream_type
,
3199 'duration': duration
,
3200 'timescale': stream_timescale
,
3201 'width': width
or 0,
3202 'height': height
or 0,
3204 'language': stream_language
,
3205 'codec_private_data': track
.get('CodecPrivateData'),
3206 'sampling_rate': sampling_rate
,
3207 'channels': int_or_none(track
.get('Channels', 2)),
3208 'bits_per_sample': int_or_none(track
.get('BitsPerSample', 16)),
3209 'nal_unit_length_field': int_or_none(track
.get('NALUnitLengthField', 4)),
3212 return formats
, subtitles
3214 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):
3215 def absolute_url(item_url
):
3216 return urljoin(base_url
, item_url
)
3218 def parse_content_type(content_type
):
3219 if not content_type
:
3221 ctr
= re
.search(r
'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type
)
3223 mimetype
, codecs
= ctr
.groups()
3224 f
= parse_codecs(codecs
)
3225 f
['ext'] = mimetype2ext(mimetype
)
3229 def _media_formats(src
, cur_media_type
, type_info
=None):
3230 type_info
= type_info
or {}
3231 full_url
= absolute_url(src
)
3232 ext
= type_info
.get('ext') or determine_ext(full_url
)
3234 is_plain_url
= False
3235 formats
= self
._extract
_m
3u8_formats
(
3236 full_url
, video_id
, ext
='mp4',
3237 entry_protocol
=m3u8_entry_protocol
, m3u8_id
=m3u8_id
,
3238 preference
=preference
, quality
=quality
, fatal
=False)
3240 is_plain_url
= False
3241 formats
= self
._extract
_mpd
_formats
(
3242 full_url
, video_id
, mpd_id
=mpd_id
, fatal
=False)
3247 'vcodec': 'none' if cur_media_type
== 'audio' else None,
3250 return is_plain_url
, formats
3253 # amp-video and amp-audio are very similar to their HTML5 counterparts
3254 # so we will include them right here (see
3255 # https://www.ampproject.org/docs/reference/components/amp-video)
3256 # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/
3257 _MEDIA_TAG_NAME_RE
= r
'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)'
3258 media_tags
= [(media_tag
, media_tag_name
, media_type
, '')
3259 for media_tag
, media_tag_name
, media_type
3260 in re
.findall(r
'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE
, webpage
)]
3261 media_tags
.extend(re
.findall(
3262 # We only allow video|audio followed by a whitespace or '>'.
3263 # Allowing more characters may end up in significant slow down (see
3264 # https://github.com/ytdl-org/youtube-dl/issues/11979,
3265 # e.g. http://www.porntrex.com/maps/videositemap.xml).
3266 r
'(?s)(<(?P<tag>%s)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>' % _MEDIA_TAG_NAME_RE
, webpage
))
3267 for media_tag
, _
, media_type
, media_content
in media_tags
:
3272 media_attributes
= extract_attributes(media_tag
)
3273 src
= strip_or_none(dict_get(media_attributes
, ('src', 'data-video-src', 'data-src', 'data-source')))
3275 f
= parse_content_type(media_attributes
.get('type'))
3276 _
, formats
= _media_formats(src
, media_type
, f
)
3277 media_info
['formats'].extend(formats
)
3278 media_info
['thumbnail'] = absolute_url(media_attributes
.get('poster'))
3280 for source_tag
in re
.findall(r
'<source[^>]+>', media_content
):
3281 s_attr
= extract_attributes(source_tag
)
3282 # data-video-src and data-src are non standard but seen
3283 # several times in the wild
3284 src
= strip_or_none(dict_get(s_attr
, ('src', 'data-video-src', 'data-src', 'data-source')))
3287 f
= parse_content_type(s_attr
.get('type'))
3288 is_plain_url
, formats
= _media_formats(src
, media_type
, f
)
3290 # width, height, res, label and title attributes are
3291 # all not standard but seen several times in the wild
3294 for lbl
in ('label', 'title')
3295 if str_or_none(s_attr
.get(lbl
))
3297 width
= int_or_none(s_attr
.get('width'))
3298 height
= (int_or_none(s_attr
.get('height'))
3299 or int_or_none(s_attr
.get('res')))
3300 if not width
or not height
:
3302 resolution
= parse_resolution(lbl
)
3305 width
= width
or resolution
.get('width')
3306 height
= height
or resolution
.get('height')
3308 tbr
= parse_bitrate(lbl
)
3317 'format_id': s_attr
.get('label') or s_attr
.get('title'),
3319 f
.update(formats
[0])
3320 media_info
['formats'].append(f
)
3322 media_info
['formats'].extend(formats
)
3323 for track_tag
in re
.findall(r
'<track[^>]+>', media_content
):
3324 track_attributes
= extract_attributes(track_tag
)
3325 kind
= track_attributes
.get('kind')
3326 if not kind
or kind
in ('subtitles', 'captions'):
3327 src
= strip_or_none(track_attributes
.get('src'))
3330 lang
= track_attributes
.get('srclang') or track_attributes
.get('lang') or track_attributes
.get('label')
3331 media_info
['subtitles'].setdefault(lang
, []).append({
3332 'url': absolute_url(src
),
3334 for f
in media_info
['formats']:
3335 f
.setdefault('http_headers', {})['Referer'] = base_url
3336 if media_info
['formats'] or media_info
['subtitles']:
3337 entries
.append(media_info
)
3340 def _extract_akamai_formats(self
, *args
, **kwargs
):
3341 fmts
, subs
= self
._extract
_akamai
_formats
_and
_subtitles
(*args
, **kwargs
)
3343 self
._report
_ignoring
_subs
('akamai')
3346 def _extract_akamai_formats_and_subtitles(self
, manifest_url
, video_id
, hosts
={}):
3347 signed
= 'hdnea=' in manifest_url
3349 # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
3350 manifest_url
= re
.sub(
3351 r
'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
3352 '', manifest_url
).strip('?')
3357 hdcore_sign
= 'hdcore=3.7.0'
3358 f4m_url
= re
.sub(r
'(https?://[^/]+)/i/', r
'\1/z/', manifest_url
).replace('/master.m3u8', '/manifest.f4m')
3359 hds_host
= hosts
.get('hds')
3361 f4m_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hds_host
, f4m_url
)
3362 if 'hdcore=' not in f4m_url
:
3363 f4m_url
+= ('&' if '?' in f4m_url
else '?') + hdcore_sign
3364 f4m_formats
= self
._extract
_f
4m
_formats
(
3365 f4m_url
, video_id
, f4m_id
='hds', fatal
=False)
3366 for entry
in f4m_formats
:
3367 entry
.update({'extra_param_to_segment_url': hdcore_sign}
)
3368 formats
.extend(f4m_formats
)
3370 m3u8_url
= re
.sub(r
'(https?://[^/]+)/z/', r
'\1/i/', manifest_url
).replace('/manifest.f4m', '/master.m3u8')
3371 hls_host
= hosts
.get('hls')
3373 m3u8_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hls_host
, m3u8_url
)
3374 m3u8_formats
, m3u8_subtitles
= self
._extract
_m
3u8_formats
_and
_subtitles
(
3375 m3u8_url
, video_id
, 'mp4', 'm3u8_native',
3376 m3u8_id
='hls', fatal
=False)
3377 formats
.extend(m3u8_formats
)
3378 subtitles
= self
._merge
_subtitles
(subtitles
, m3u8_subtitles
)
3380 http_host
= hosts
.get('http')
3381 if http_host
and m3u8_formats
and not signed
:
3382 REPL_REGEX
= r
'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
3383 qualities
= re
.match(REPL_REGEX
, m3u8_url
).group(2).split(',')
3384 qualities_length
= len(qualities
)
3385 if len(m3u8_formats
) in (qualities_length
, qualities_length
+ 1):
3387 for f
in m3u8_formats
:
3388 if f
['vcodec'] != 'none':
3389 for protocol
in ('http', 'https'):
3391 del http_f
['manifest_url']
3393 REPL_REGEX
, protocol
+ fr
'://{http_host}/\g<1>{qualities[i]}\3', f
['url'])
3395 'format_id': http_f
['format_id'].replace('hls-', protocol
+ '-'),
3397 'protocol': protocol
,
3399 formats
.append(http_f
)
3402 return formats
, subtitles
3404 def _extract_wowza_formats(self
, url
, video_id
, m3u8_entry_protocol
='m3u8_native', skip_protocols
=[]):
3405 query
= urllib
.parse
.urlparse(url
).query
3406 url
= re
.sub(r
'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url
)
3408 r
'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url
)
3409 url_base
= mobj
.group('url')
3410 http_base_url
= '%s%s:%s' % ('http', mobj
.group('s') or '', url_base
)
3413 def manifest_url(manifest
):
3414 m_url
= f
'{http_base_url}/{manifest}'
3416 m_url
+= '?%s' % query
3419 if 'm3u8' not in skip_protocols
:
3420 formats
.extend(self
._extract
_m
3u8_formats
(
3421 manifest_url('playlist.m3u8'), video_id
, 'mp4',
3422 m3u8_entry_protocol
, m3u8_id
='hls', fatal
=False))
3423 if 'f4m' not in skip_protocols
:
3424 formats
.extend(self
._extract
_f
4m
_formats
(
3425 manifest_url('manifest.f4m'),
3426 video_id
, f4m_id
='hds', fatal
=False))
3427 if 'dash' not in skip_protocols
:
3428 formats
.extend(self
._extract
_mpd
_formats
(
3429 manifest_url('manifest.mpd'),
3430 video_id
, mpd_id
='dash', fatal
=False))
3431 if re
.search(r
'(?:/smil:|\.smil)', url_base
):
3432 if 'smil' not in skip_protocols
:
3433 rtmp_formats
= self
._extract
_smil
_formats
(
3434 manifest_url('jwplayer.smil'),
3435 video_id
, fatal
=False)
3436 for rtmp_format
in rtmp_formats
:
3437 rtsp_format
= rtmp_format
.copy()
3438 rtsp_format
['url'] = '%s/%s' % (rtmp_format
['url'], rtmp_format
['play_path'])
3439 del rtsp_format
['play_path']
3440 del rtsp_format
['ext']
3441 rtsp_format
.update({
3442 'url': rtsp_format
['url'].replace('rtmp://', 'rtsp://'),
3443 'format_id': rtmp_format
['format_id'].replace('rtmp', 'rtsp'),
3446 formats
.extend([rtmp_format
, rtsp_format
])
3448 for protocol
in ('rtmp', 'rtsp'):
3449 if protocol
not in skip_protocols
:
3451 'url': f
'{protocol}:{url_base}',
3452 'format_id': protocol
,
3453 'protocol': protocol
,
3457 def _find_jwplayer_data(self
, webpage
, video_id
=None, transform_source
=js_to_json
):
3459 r
'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P
=quote
)\
)(?
!</script
>).*?\
.setup\s
*\
((?P
<options
>[^
)]+)\
)',
3463 jwplayer_data = self._parse_json(mobj.group('options
'),
3465 transform_source=transform_source)
3466 except ExtractorError:
3469 if isinstance(jwplayer_data, dict):
3470 return jwplayer_data
3472 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
3473 jwplayer_data = self._find_jwplayer_data(
3474 webpage, video_id, transform_source=js_to_json)
3475 return self._parse_jwplayer_data(
3476 jwplayer_data, video_id, *args, **kwargs)
3478 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
3479 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3480 # JWPlayer backward compatibility: flattened playlists
3481 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
3482 if 'playlist
' not in jwplayer_data:
3483 jwplayer_data = {'playlist': [jwplayer_data]}
3487 # JWPlayer backward compatibility: single playlist item
3488 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
3489 if not isinstance(jwplayer_data['playlist
'], list):
3490 jwplayer_data['playlist
'] = [jwplayer_data['playlist
']]
3492 for video_data in jwplayer_data['playlist
']:
3493 # JWPlayer backward compatibility: flattened sources
3494 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
3495 if 'sources
' not in video_data:
3496 video_data['sources
'] = [video_data]
3498 this_video_id = video_id or video_data['mediaid
']
3500 formats = self._parse_jwplayer_formats(
3501 video_data['sources
'], video_id=this_video_id, m3u8_id=m3u8_id,
3502 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
3505 tracks = video_data.get('tracks
')
3506 if tracks and isinstance(tracks, list):
3507 for track in tracks:
3508 if not isinstance(track, dict):
3510 track_kind = track.get('kind
')
3511 if not track_kind or not isinstance(track_kind, str):
3513 if track_kind.lower() not in ('captions
', 'subtitles
'):
3515 track_url = urljoin(base_url, track.get('file'))
3518 subtitles.setdefault(track.get('label
') or 'en
', []).append({
3519 'url
': self._proto_relative_url(track_url)
3523 'id': this_video_id,
3524 'title
': unescapeHTML(video_data['title
'] if require_title else video_data.get('title
')),
3525 'description
': clean_html(video_data.get('description
')),
3526 'thumbnail
': urljoin(base_url, self._proto_relative_url(video_data.get('image
'))),
3527 'timestamp
': int_or_none(video_data.get('pubdate
')),
3528 'duration
': float_or_none(jwplayer_data.get('duration
') or video_data.get('duration
')),
3529 'subtitles
': subtitles,
3531 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
3532 if len(formats) == 1 and re.search(r'^
(?
:http|
//).*(?
:youtube\
.com|youtu\
.be
)/.+', formats[0]['url
']):
3534 '_type
': 'url_transparent
',
3535 'url
': formats[0]['url
'],
3538 self._sort_formats(formats)
3539 entry['formats
'] = formats
3540 entries.append(entry)
3541 if len(entries) == 1:
3544 return self.playlist_result(entries)
3546 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
3547 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3550 for source in jwplayer_sources_data:
3551 if not isinstance(source, dict):
3553 source_url = urljoin(
3554 base_url, self._proto_relative_url(source.get('file')))
3555 if not source_url or source_url in urls:
3557 urls.append(source_url)
3558 source_type = source.get('type') or ''
3559 ext = mimetype2ext(source_type) or determine_ext(source_url)
3560 if source_type == 'hls
' or ext == 'm3u8
':
3561 formats.extend(self._extract_m3u8_formats(
3562 source_url, video_id, 'mp4
', entry_protocol='m3u8_native
',
3563 m3u8_id=m3u8_id, fatal=False))
3564 elif source_type == 'dash
' or ext == 'mpd
':
3565 formats.extend(self._extract_mpd_formats(
3566 source_url, video_id, mpd_id=mpd_id, fatal=False))
3568 formats.extend(self._extract_smil_formats(
3569 source_url, video_id, fatal=False))
3570 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3571 elif source_type.startswith('audio
') or ext in (
3572 'oga
', 'aac
', 'mp3
', 'mpeg
', 'vorbis
'):
3579 height = int_or_none(source.get('height
'))
3581 # Often no height is provided but there is a label in
3582 # format like "1080p", "720p SD", or 1080.
3583 height = int_or_none(self._search_regex(
3584 r'^
(\d{3,4}
)[pP
]?
(?
:\b|$
)', str(source.get('label
') or ''),
3585 'height
', default=None))
3588 'width
': int_or_none(source.get('width
')),
3590 'tbr
': int_or_none(source.get('bitrate
'), scale=1000),
3591 'filesize
': int_or_none(source.get('filesize
')),
3594 if source_url.startswith('rtmp
'):
3595 a_format['ext
'] = 'flv
'
3596 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3597 # of jwplayer.flash.swf
3598 rtmp_url_parts = re.split(
3599 r'((?
:mp4|mp3|flv
):)', source_url, 1)
3600 if len(rtmp_url_parts) == 3:
3601 rtmp_url, prefix, play_path = rtmp_url_parts
3604 'play_path
': prefix + play_path,
3607 a_format.update(rtmp_params)
3608 formats.append(a_format)
3611 def _live_title(self, name):
3612 self._downloader.deprecation_warning('yt_dlp
.InfoExtractor
._live
_title
is deprecated
and does
not work
as expected
')
3615 def _int(self, v, name, fatal=False, **kwargs):
3616 res = int_or_none(v, **kwargs)
3618 msg = f'Failed to extract {name}
: Could
not parse value {v!r}
'
3620 raise ExtractorError(msg)
3622 self.report_warning(msg)
3625 def _float(self, v, name, fatal=False, **kwargs):
3626 res = float_or_none(v, **kwargs)
3628 msg = f'Failed to extract {name}
: Could
not parse value {v!r}
'
3630 raise ExtractorError(msg)
3632 self.report_warning(msg)
3635 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
3636 path='/', secure=False, discard=False, rest={}, **kwargs):
3637 cookie = http.cookiejar.Cookie(
3638 0, name, value, port, port is not None, domain, True,
3639 domain.startswith('.'), path, True, secure, expire_time,
3640 discard, None, None, rest)
3641 self.cookiejar.set_cookie(cookie)
3643 def _get_cookies(self, url):
3644 """ Return a http.cookies.SimpleCookie with the cookies for the url """
3645 return LenientSimpleCookie(self._downloader._calc_cookies(url))
3647 def _apply_first_set_cookie_header(self, url_handle, cookie):
3649 Apply first Set-Cookie header instead of the last. Experimental.
3651 Some sites (e.g. [1-3]) may serve two cookies under the same name
3652 in Set-Cookie header and expect the first (old) one to be set rather
3653 than second (new). However, as of RFC6265 the newer one cookie
3654 should be set into cookie store what actually happens.
3655 We will workaround this issue by resetting the cookie to
3656 the first one manually.
3657 1. https://new.vk.com/
3658 2. https://github.com/ytdl-org/youtube-dl/issues/9841#issuecomment-227871201
3659 3. https://learning.oreilly.com/
3661 for header, cookies in url_handle.headers.items():
3662 if header.lower() != 'set-cookie
':
3664 cookies = cookies.encode('iso
-8859-1').decode('utf
-8')
3665 cookie_value = re.search(
3666 r'%s=(.+?
);.*?
\b[Dd
]omain
=(.+?
)(?
:[,;]|$
)' % cookie, cookies)
3668 value, domain = cookie_value.groups()
3669 self._set_cookie(domain, cookie, value)
3673 def get_testcases(cls, include_onlymatching=False):
3674 t = getattr(cls, '_TEST
', None)
3676 assert not hasattr(cls, '_TESTS
'), f'{cls.ie_key()}IE has _TEST
and _TESTS
'
3679 tests = getattr(cls, '_TESTS
', [])
3681 if not include_onlymatching and t.get('only_matching
', False):
3683 t['name
'] = cls.ie_key()
3687 def get_webpage_testcases(cls):
3688 tests = getattr(cls, '_WEBPAGE_TESTS
', [])
3690 t['name
'] = cls.ie_key()
3695 """Get age limit from the testcases"""
3696 return max(traverse_obj(
3697 (*cls.get_testcases(include_onlymatching=False), *cls.get_webpage_testcases()),
3698 (..., (('playlist
', 0), None), 'info_dict
', 'age_limit
')) or [0])
3701 def is_suitable(cls, age_limit):
3702 """Test whether the extractor is generally suitable for the given age limit"""
3703 return not age_restricted(cls.age_limit, age_limit)
3706 def description(cls, *, markdown=True, search_examples=None):
3707 """Description of the extractor"""
3709 if cls._NETRC_MACHINE:
3711 desc += f' [<abbr title
="netrc machine"><em
>{cls._NETRC_MACHINE}
</em
></abbr
>]'
3713 desc += f' [{cls._NETRC_MACHINE}
]'
3714 if cls.IE_DESC is False:
3717 desc += f' {cls.IE_DESC}
'
3719 desc += f'; "{cls.SEARCH_KEY}:" prefix
'
3721 _COUNTS = ('', '5', '10', 'all
')
3722 desc += f' (e
.g
. "{cls.SEARCH_KEY}{random.choice(_COUNTS)}:{random.choice(search_examples)}")'
3723 if not cls.working():
3724 desc += ' (**Currently broken
**)' if markdown else ' (Currently broken
)'
3726 name = f' - **{cls.IE_NAME}
**' if markdown else cls.IE_NAME
3727 return f'{name}
:{desc}
' if desc else name
3729 def extract_subtitles(self, *args, **kwargs):
3730 if (self.get_param('writesubtitles
', False)
3731 or self.get_param('listsubtitles
')):
3732 return self._get_subtitles(*args, **kwargs)
3735 def _get_subtitles(self, *args, **kwargs):
3736 raise NotImplementedError('This method must be implemented by subclasses
')
3738 def extract_comments(self, *args, **kwargs):
3739 if not self.get_param('getcomments
'):
3741 generator = self._get_comments(*args, **kwargs)
3748 comments.append(next(generator))
3749 except StopIteration:
3751 except KeyboardInterrupt:
3752 self.to_screen('Interrupted by user
')
3753 except Exception as e:
3754 if self.get_param('ignoreerrors
') is not True:
3756 self._downloader.report_error(e)
3757 comment_count = len(comments)
3758 self.to_screen(f'Extracted {comment_count} comments
')
3760 'comments
': comments,
3761 'comment_count
': None if interrupted else comment_count
3765 def _get_comments(self, *args, **kwargs):
3766 raise NotImplementedError('This method must be implemented by subclasses
')
3769 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
3770 """ Merge subtitle items for one language. Items with duplicated URLs/data
3771 will be dropped. """
3772 list1_data = {(item.get('url'), item.get('data')) for item in subtitle_list1}
3773 ret = list(subtitle_list1)
3774 ret.extend(item for item in subtitle_list2 if (item.get('url
'), item.get('data
')) not in list1_data)
3778 def _merge_subtitles(cls, *dicts, target=None):
3779 """ Merge subtitle dictionaries, language by language. """
3783 for lang, subs in d.items():
3784 target[lang] = cls._merge_subtitle_items(target.get(lang, []), subs)
3787 def extract_automatic_captions(self, *args, **kwargs):
3788 if (self.get_param('writeautomaticsub
', False)
3789 or self.get_param('listsubtitles
')):
3790 return self._get_automatic_captions(*args, **kwargs)
3793 def _get_automatic_captions(self, *args, **kwargs):
3794 raise NotImplementedError('This method must be implemented by subclasses
')
3796 @functools.cached_property
3797 def _cookies_passed(self):
3798 """Whether cookies have been passed to YoutubeDL"""
3799 return self.get_param('cookiefile
') is not None or self.get_param('cookiesfrombrowser
') is not None
3801 def mark_watched(self, *args, **kwargs):
3802 if not self.get_param('mark_watched
', False):
3804 if self.supports_login() and self._get_login_info()[0] is not None or self._cookies_passed:
3805 self._mark_watched(*args, **kwargs)
3807 def _mark_watched(self, *args, **kwargs):
3808 raise NotImplementedError('This method must be implemented by subclasses
')
3810 def geo_verification_headers(self):
3812 geo_verification_proxy = self.get_param('geo_verification_proxy
')
3813 if geo_verification_proxy:
3814 headers['Ytdl
-request
-proxy
'] = geo_verification_proxy
3818 def _generic_id(url):
3819 return urllib.parse.unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
3822 def _generic_title(url):
3823 return urllib.parse.unquote(os.path.splitext(url_basename(url))[0])
3826 def _availability(is_private=None, needs_premium=None, needs_subscription=None, needs_auth=None, is_unlisted=None):
3827 all_known = all(map(
3828 lambda x: x is not None,
3829 (is_private, needs_premium, needs_subscription, needs_auth, is_unlisted)))
3831 'private
' if is_private
3832 else 'premium_only
' if needs_premium
3833 else 'subscriber_only
' if needs_subscription
3834 else 'needs_auth
' if needs_auth
3835 else 'unlisted
' if is_unlisted
3836 else 'public
' if all_known
3839 def _configuration_arg(self, key, default=NO_DEFAULT, *, ie_key=None, casesense=False):
3841 @returns A list of values for the extractor argument given by "key"
3842 or "default" if no such key is present
3843 @param default The default value to return when the key is not present (default: [])
3844 @param casesense When false, the values are converted to lower case
3847 self._downloader.params, ('extractor_args
', (ie_key or self.ie_key()).lower(), key))
3849 return [] if default is NO_DEFAULT else default
3850 return list(val) if casesense else [x.lower() for x in val]
3852 def _yes_playlist(self, playlist_id, video_id, smuggled_data=None, *, playlist_label='playlist
', video_label='video
'):
3853 if not playlist_id or not video_id:
3856 no_playlist = (smuggled_data or {}).get('force_noplaylist
')
3857 if no_playlist is not None:
3858 return not no_playlist
3860 video_id = '' if video_id is True else f' {video_id}
'
3861 playlist_id = '' if playlist_id is True else f' {playlist_id}
'
3862 if self.get_param('noplaylist
'):
3863 self.to_screen(f'Downloading just the {video_label}{video_id} because of
--no
-playlist
')
3865 self.to_screen(f'Downloading {playlist_label}{playlist_id}
- add
--no
-playlist to download just the {video_label}{video_id}
')
3868 def _error_or_warning(self, err, _count=None, _retries=0, *, fatal=True):
3869 RetryManager.report_retry(
3870 err, _count or int(fatal), _retries,
3871 info=self.to_screen, warn=self.report_warning, error=None if fatal else self.report_warning,
3872 sleep_func=self.get_param('retry_sleep_functions
', {}).get('extractor
'))
3874 def RetryManager(self, **kwargs):
3875 return RetryManager(self.get_param('extractor_retries
', 3), self._error_or_warning, **kwargs)
3878 def extract_from_webpage(cls, ydl, url, webpage):
3879 ie = (cls if isinstance(cls._extract_from_webpage, types.MethodType)
3880 else ydl.get_info_extractor(cls.ie_key()))
3881 for info in ie._extract_from_webpage(url, webpage) or []:
3882 # url = None since we do not want to set (webpage/original)_url
3883 ydl.add_default_extra_info(info, ie, None)
3887 def _extract_from_webpage(cls, url, webpage):
3888 for embed_url in orderedSet(
3889 cls._extract_embed_urls(url, webpage) or [], lazy=True):
3890 yield cls.url_result(embed_url, None if cls._VALID_URL is False else cls)
3893 def _extract_embed_urls(cls, url, webpage):
3894 """@returns all the embed urls on the webpage"""
3895 if '_EMBED_URL_RE
' not in cls.__dict__:
3896 assert isinstance(cls._EMBED_REGEX, (list, tuple))
3897 for idx, regex in enumerate(cls._EMBED_REGEX):
3898 assert regex.count('(?P
<url
>') == 1, \
3899 f'{cls.__name__}
._EMBED
_REGEX
[{idx}
] must have exactly
1 url group
\n\t{regex}
'
3900 cls._EMBED_URL_RE = tuple(map(re.compile, cls._EMBED_REGEX))
3902 for regex in cls._EMBED_URL_RE:
3903 for mobj in regex.finditer(webpage):
3904 embed_url = urllib.parse.urljoin(url, unescapeHTML(mobj.group('url
')))
3905 if cls._VALID_URL is False or cls.suitable(embed_url):
3908 class StopExtraction(Exception):
3912 def _extract_url(cls, webpage): # TODO: Remove
3913 """Only for compatibility with some older extractors"""
3914 return next(iter(cls._extract_embed_urls(None, webpage) or []), None)
3917 def __init_subclass__(cls, *, plugin_name=None, **kwargs):
3919 mro = inspect.getmro(cls)
3920 super_class = cls.__wrapped__ = mro[mro.index(cls) + 1]
3921 cls.IE_NAME, cls.ie_key = f'{super_class.IE_NAME}
+{plugin_name}
', super_class.ie_key
3922 while getattr(super_class, '__wrapped__
', None):
3923 super_class = super_class.__wrapped__
3924 setattr(sys.modules[super_class.__module__], super_class.__name__, cls)
3926 return super().__init_subclass__(**kwargs)
3929 class SearchInfoExtractor(InfoExtractor):
3931 Base class for paged search queries extractors.
3932 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
3933 Instances should define _SEARCH_KEY and optionally _MAX_RESULTS
3936 _MAX_RESULTS = float('inf
')
3939 def _VALID_URL(cls):
3940 return r'%s(?P
<prefix
>|
[1-9][0-9]*|all
):(?P
<query
>[\s\S
]+)' % cls._SEARCH_KEY
3942 def _real_extract(self, query):
3943 prefix, query = self._match_valid_url(query).group('prefix
', 'query
')
3945 return self._get_n_results(query, 1)
3946 elif prefix == 'all
':
3947 return self._get_n_results(query, self._MAX_RESULTS)
3951 raise ExtractorError(f'invalid download number {n}
for query
"{query}"')
3952 elif n > self._MAX_RESULTS:
3953 self.report_warning('%s returns
max %i results (you requested
%i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3954 n = self._MAX_RESULTS
3955 return self._get_n_results(query, n)
3957 def _get_n_results(self, query, n):
3958 """Get a specified number of results for a query.
3959 Either this function or _search_results must be overridden by subclasses """
3960 return self.playlist_result(
3961 itertools.islice(self._search_results(query), 0, None if n == float('inf
') else n),
3964 def _search_results(self, query):
3965 """Returns an iterator of search results"""
3966 raise NotImplementedError('This method must be implemented by subclasses
')
3969 def SEARCH_KEY(cls):
3970 return cls._SEARCH_KEY
3973 class UnsupportedURLIE(InfoExtractor):
3978 def _real_extract(self, url):
3979 raise UnsupportedError(url)