12 import xml
.etree
.ElementTree
14 from ..compat
import (
15 compat_cookiejar_Cookie
,
16 compat_cookies_SimpleCookie
,
17 compat_etree_fromstring
,
24 compat_urllib_parse_unquote
,
25 compat_urllib_parse_urlencode
,
26 compat_urllib_request
,
30 from ..downloader
import FileDownloader
31 from ..downloader
.f4m
import get_base_url
, remove_encrypted_media
64 parse_m3u8_attributes
,
89 """Information Extractor class.
91 Information extractors are the classes that, given a URL, extract
92 information about the video (or videos) the URL refers to. This
93 information includes the real video URL, the video title, author and
94 others. The information is stored in a dictionary which is then
95 passed to the YoutubeDL. The YoutubeDL processes this
96 information possibly downloading the video to the file system, among
97 other possible outcomes.
99 The type field determines the type of the result.
100 By far the most common value (and the default if _type is missing) is
101 "video", which indicates a single video.
103 For a video, the dictionaries must include the following fields:
105 id: Video identifier.
106 title: Video title, unescaped.
108 Additionally, it must contain either a formats entry or a url one:
110 formats: A list of dictionaries for each format available, ordered
111 from worst to best quality.
114 * url The mandatory URL representing the media:
115 for plain file media - HTTP URL of this file,
117 for HLS - URL of the M3U8 media playlist,
118 for HDS - URL of the F4M manifest,
120 - HTTP URL to plain file media (in case of
122 - URL of the MPD manifest or base URL
123 representing the media if MPD manifest
124 is parsed from a string (in case of
126 for MSS - URL of the ISM manifest.
128 The URL of the manifest file in case of
130 for HLS - URL of the M3U8 master playlist,
131 for HDS - URL of the F4M manifest,
132 for DASH - URL of the MPD manifest,
133 for MSS - URL of the ISM manifest.
134 * manifest_stream_number (For internal use only)
135 The index of the stream in the manifest file
136 * ext Will be calculated from URL if missing
137 * format A human-readable description of the format
138 ("mp4 container with h264/opus").
139 Calculated from the format_id, width, height.
140 and format_note fields if missing.
141 * format_id A short description of the format
142 ("mp4_h264_opus" or "19").
143 Technically optional, but strongly recommended.
144 * format_note Additional info about the format
145 ("3D" or "DASH video")
146 * width Width of the video, if known
147 * height Height of the video, if known
148 * resolution Textual description of width and height
149 * dynamic_range The dynamic range of the video. One of:
150 "SDR" (None), "HDR10", "HDR10+, "HDR12", "HLG, "DV"
151 * tbr Average bitrate of audio and video in KBit/s
152 * abr Average audio bitrate in KBit/s
153 * acodec Name of the audio codec in use
154 * asr Audio sampling rate in Hertz
155 * vbr Average video bitrate in KBit/s
157 * vcodec Name of the video codec in use
158 * container Name of the container format
159 * filesize The number of bytes, if known in advance
160 * filesize_approx An estimate for the number of bytes
161 * player_url SWF Player URL (used for rtmpdump).
162 * protocol The protocol that will be used for the actual
163 download, lower-case. One of "http", "https" or
164 one of the protocols defined in downloader.PROTOCOL_MAP
166 Base URL for fragments. Each fragment's path
167 value (if present) will be relative to
169 * fragments A list of fragments of a fragmented media.
170 Each fragment entry must contain either an url
171 or a path. If an url is present it should be
172 considered by a client. Otherwise both path and
173 fragment_base_url must be present. Here is
174 the list of all potential fields:
175 * "url" - fragment's URL
176 * "path" - fragment's path relative to
178 * "duration" (optional, int or float)
179 * "filesize" (optional, int)
180 * is_from_start Is a live format that can be downloaded
181 from the start. Boolean
182 * preference Order number of this format. If this field is
183 present and not None, the formats get sorted
184 by this field, regardless of all other values.
185 -1 for default (order by other properties),
186 -2 or smaller for less than default.
187 < -1000 to hide the format (if there is
188 another one which is strictly better)
189 * language Language code, e.g. "de" or "en-US".
190 * language_preference Is this in the language mentioned in
192 10 if it's what the URL is about,
193 -1 for default (don't know),
194 -10 otherwise, other values reserved for now.
195 * quality Order number of the video quality of this
196 format, irrespective of the file format.
197 -1 for default (order by other properties),
198 -2 or smaller for less than default.
199 * source_preference Order number for this video source
200 (quality takes higher priority)
201 -1 for default (order by other properties),
202 -2 or smaller for less than default.
203 * http_headers A dictionary of additional HTTP headers
204 to add to the request.
205 * stretched_ratio If given and not 1, indicates that the
206 video's pixels are not square.
207 width : height ratio as float.
208 * no_resume The server does not support resuming the
209 (HTTP or RTMP) download. Boolean.
210 * has_drm The format has DRM and cannot be downloaded. Boolean
211 * downloader_options A dictionary of downloader options as
212 described in FileDownloader (For internal use only)
213 RTMP formats can also have the additional fields: page_url,
214 app, play_path, tc_url, flash_version, rtmp_live, rtmp_conn,
215 rtmp_protocol, rtmp_real_time
217 url: Final video URL.
218 ext: Video filename extension.
219 format: The video format, defaults to ext (used for --get-format)
220 player_url: SWF Player URL (used for rtmpdump).
222 The following fields are optional:
224 direct: True if a direct video file was given (must only be set by GenericIE)
225 alt_title: A secondary title of the video.
226 display_id An alternative identifier for the video, not necessarily
227 unique, but available before title. Typically, id is
228 something like "4234987", title "Dancing naked mole rats",
229 and display_id "dancing-naked-mole-rats"
230 thumbnails: A list of dictionaries, with the following entries:
231 * "id" (optional, string) - Thumbnail format ID
233 * "preference" (optional, int) - quality of the image
234 * "width" (optional, int)
235 * "height" (optional, int)
236 * "resolution" (optional, string "{width}x{height}",
238 * "filesize" (optional, int)
239 * "http_headers" (dict) - HTTP headers for the request
240 thumbnail: Full URL to a video thumbnail image.
241 description: Full video description.
242 uploader: Full name of the video uploader.
243 license: License name the video is licensed under.
244 creator: The creator of the video.
245 timestamp: UNIX timestamp of the moment the video was uploaded
246 upload_date: Video upload date in UTC (YYYYMMDD).
247 If not explicitly set, calculated from timestamp
248 release_timestamp: UNIX timestamp of the moment the video was released.
249 If it is not clear whether to use timestamp or this, use the former
250 release_date: The date (YYYYMMDD) when the video was released in UTC.
251 If not explicitly set, calculated from release_timestamp
252 modified_timestamp: UNIX timestamp of the moment the video was last modified.
253 modified_date: The date (YYYYMMDD) when the video was last modified in UTC.
254 If not explicitly set, calculated from modified_timestamp
255 uploader_id: Nickname or id of the video uploader.
256 uploader_url: Full URL to a personal webpage of the video uploader.
257 channel: Full name of the channel the video is uploaded on.
258 Note that channel fields may or may not repeat uploader
259 fields. This depends on a particular extractor.
260 channel_id: Id of the channel.
261 channel_url: Full URL to a channel webpage.
262 channel_follower_count: Number of followers of the channel.
263 location: Physical location where the video was filmed.
264 subtitles: The available subtitles as a dictionary in the format
265 {tag: subformats}. "tag" is usually a language code, and
266 "subformats" is a list sorted from lower to higher
267 preference, each element is a dictionary with the "ext"
269 * "data": The subtitles file contents
270 * "url": A URL pointing to the subtitles file
271 It can optionally also have:
272 * "name": Name or description of the subtitles
273 * "http_headers": A dictionary of additional HTTP headers
274 to add to the request.
275 "ext" will be calculated from URL if missing
276 automatic_captions: Like 'subtitles'; contains automatically generated
277 captions instead of normal subtitles
278 duration: Length of the video in seconds, as an integer or float.
279 view_count: How many users have watched the video on the platform.
280 like_count: Number of positive ratings of the video
281 dislike_count: Number of negative ratings of the video
282 repost_count: Number of reposts of the video
283 average_rating: Average rating give by users, the scale used depends on the webpage
284 comment_count: Number of comments on the video
285 comments: A list of comments, each with one or more of the following
286 properties (all but one of text or html optional):
287 * "author" - human-readable name of the comment author
288 * "author_id" - user ID of the comment author
289 * "author_thumbnail" - The thumbnail of the comment author
291 * "html" - Comment as HTML
292 * "text" - Plain text of the comment
293 * "timestamp" - UNIX timestamp of comment
294 * "parent" - ID of the comment this one is replying to.
295 Set to "root" to indicate that this is a
296 comment to the original video.
297 * "like_count" - Number of positive ratings of the comment
298 * "dislike_count" - Number of negative ratings of the comment
299 * "is_favorited" - Whether the comment is marked as
300 favorite by the video uploader
301 * "author_is_uploader" - Whether the comment is made by
303 age_limit: Age restriction for the video, as an integer (years)
304 webpage_url: The URL to the video webpage, if given to yt-dlp it
305 should allow to get the same result again. (It will be set
306 by YoutubeDL if it's missing)
307 categories: A list of categories that the video falls in, for example
309 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
310 cast: A list of the video cast
311 is_live: True, False, or None (=unknown). Whether this video is a
312 live stream that goes on instead of a fixed-length video.
313 was_live: True, False, or None (=unknown). Whether this video was
314 originally a live stream.
315 live_status: 'is_live', 'is_upcoming', 'was_live', 'not_live' or None (=unknown)
316 If absent, automatically set from is_live, was_live
317 start_time: Time in seconds where the reproduction should start, as
318 specified in the URL.
319 end_time: Time in seconds where the reproduction should end, as
320 specified in the URL.
321 chapters: A list of dictionaries, with the following entries:
322 * "start_time" - The start time of the chapter in seconds
323 * "end_time" - The end time of the chapter in seconds
324 * "title" (optional, string)
325 playable_in_embed: Whether this video is allowed to play in embedded
326 players on other sites. Can be True (=always allowed),
327 False (=never allowed), None (=unknown), or a string
328 specifying the criteria for embedability (Eg: 'whitelist')
329 availability: Under what condition the video is available. One of
330 'private', 'premium_only', 'subscriber_only', 'needs_auth',
331 'unlisted' or 'public'. Use 'InfoExtractor._availability'
333 __post_extractor: A function to be called just before the metadata is
334 written to either disk, logger or console. The function
335 must return a dict which will be added to the info_dict.
336 This is usefull for additional information that is
337 time-consuming to extract. Note that the fields thus
338 extracted will not be available to output template and
339 match_filter. So, only "comments" and "comment_count" are
340 currently allowed to be extracted via this method.
342 The following fields should only be used when the video belongs to some logical
345 chapter: Name or title of the chapter the video belongs to.
346 chapter_number: Number of the chapter the video belongs to, as an integer.
347 chapter_id: Id of the chapter the video belongs to, as a unicode string.
349 The following fields should only be used when the video is an episode of some
350 series, programme or podcast:
352 series: Title of the series or programme the video episode belongs to.
353 series_id: Id of the series or programme the video episode belongs to, as a unicode string.
354 season: Title of the season the video episode belongs to.
355 season_number: Number of the season the video episode belongs to, as an integer.
356 season_id: Id of the season the video episode belongs to, as a unicode string.
357 episode: Title of the video episode. Unlike mandatory video title field,
358 this field should denote the exact title of the video episode
359 without any kind of decoration.
360 episode_number: Number of the video episode within a season, as an integer.
361 episode_id: Id of the video episode, as a unicode string.
363 The following fields should only be used when the media is a track or a part of
366 track: Title of the track.
367 track_number: Number of the track within an album or a disc, as an integer.
368 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
370 artist: Artist(s) of the track.
371 genre: Genre(s) of the track.
372 album: Title of the album the track belongs to.
373 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
374 album_artist: List of all artists appeared on the album (e.g.
375 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
377 disc_number: Number of the disc or other physical medium the track belongs to,
379 release_year: Year (YYYY) when the album was released.
380 composer: Composer of the piece
382 Unless mentioned otherwise, the fields should be Unicode strings.
384 Unless mentioned otherwise, None is equivalent to absence of information.
387 _type "playlist" indicates multiple videos.
388 There must be a key "entries", which is a list, an iterable, or a PagedList
389 object, each element of which is a valid dictionary by this specification.
391 Additionally, playlists can have "id", "title", and any other relevent
392 attributes with the same semantics as videos (see above).
394 It can also have the following optional fields:
396 playlist_count: The total number of videos in a playlist. If not given,
397 YoutubeDL tries to calculate it from "entries"
400 _type "multi_video" indicates that there are multiple videos that
401 form a single show, for examples multiple acts of an opera or TV episode.
402 It must have an entries key like a playlist and contain all the keys
403 required for a video at the same time.
406 _type "url" indicates that the video must be extracted from another
407 location, possibly by a different extractor. Its only required key is:
408 "url" - the next URL to extract.
409 The key "ie_key" can be set to the class name (minus the trailing "IE",
410 e.g. "Youtube") if the extractor class is known in advance.
411 Additionally, the dictionary may have any properties of the resolved entity
412 known in advance, for example "title" if the title of the referred video is
416 _type "url_transparent" entities have the same specification as "url", but
417 indicate that the given additional information is more precise than the one
418 associated with the resolved URL.
419 This is useful when a site employs a video service that hosts the video and
420 its technical metadata, but that video service does not embed a useful
421 title, description etc.
424 Subclasses of this should define a _VALID_URL regexp and, re-define the
425 _real_extract() and (optionally) _real_initialize() methods.
426 Probably, they should also be added to the list of extractors.
428 Subclasses may also override suitable() if necessary, but ensure the function
429 signature is preserved and that this function imports everything it needs
430 (except other extractors), so that lazy_extractors works correctly.
432 To support username + password (or netrc) login, the extractor must define a
433 _NETRC_MACHINE and re-define _perform_login(username, password) and
434 (optionally) _initialize_pre_login() methods. The _perform_login method will
435 be called between _initialize_pre_login and _real_initialize if credentials
436 are passed by the user. In cases where it is necessary to have the login
437 process as part of the extraction rather than initialization, _perform_login
438 can be left undefined.
440 _GEO_BYPASS attribute may be set to False in order to disable
441 geo restriction bypass mechanisms for a particular extractor.
442 Though it won't disable explicit geo restriction bypass based on
443 country code provided with geo_bypass_country.
445 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
446 countries for this extractor. One of these countries will be used by
447 geo restriction bypass mechanism right away in order to bypass
448 geo restriction, of course, if the mechanism is not disabled.
450 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
451 IP blocks in CIDR notation for this extractor. One of these IP blocks
452 will be used by geo restriction bypass mechanism similarly
455 The _WORKING attribute should be set to False for broken IEs
456 in order to warn the users and skip the tests.
461 _x_forwarded_for_ip
= None
463 _GEO_COUNTRIES
= None
464 _GEO_IP_BLOCKS
= None
466 _NETRC_MACHINE
= None
470 'any': 'Use --cookies, --cookies-from-browser, --username and --password, or --netrc to provide account credentials',
472 'Use --cookies-from-browser or --cookies for the authentication. '
473 'See https://github.com/ytdl-org/youtube-dl#how-do-i-pass-cookies-to-youtube-dl for how to manually pass cookies'),
474 'password': 'Use --username and --password, or --netrc to provide account credentials',
477 def __init__(self
, downloader
=None):
478 """Constructor. Receives an optional downloader (a YoutubeDL instance).
479 If a downloader is not passed during initialization,
480 it must be set using "set_downloader()" before "extract()" is called"""
482 self
._x
_forwarded
_for
_ip
= None
483 self
._printed
_messages
= set()
484 self
.set_downloader(downloader
)
487 def _match_valid_url(cls
, url
):
488 # This does not use has/getattr intentionally - we want to know whether
489 # we have cached the regexp for *this* class, whereas getattr would also
490 # match the superclass
491 if '_VALID_URL_RE' not in cls
.__dict
__:
492 if '_VALID_URL' not in cls
.__dict
__:
493 cls
._VALID
_URL
= cls
._make
_valid
_url
()
494 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
495 return cls
._VALID
_URL
_RE
.match(url
)
498 def suitable(cls
, url
):
499 """Receives a URL and returns True if suitable for this IE."""
500 # This function must import everything it needs (except other extractors),
501 # so that lazy_extractors works correctly
502 return cls
._match
_valid
_url
(url
) is not None
505 def _match_id(cls
, url
):
506 return cls
._match
_valid
_url
(url
).group('id')
509 def get_temp_id(cls
, url
):
511 return cls
._match
_id
(url
)
512 except (IndexError, AttributeError):
517 """Getter method for _WORKING."""
521 def supports_login(cls
):
522 return bool(cls
._NETRC
_MACHINE
)
524 def initialize(self
):
525 """Initializes an instance (authentication, etc)."""
526 self
._printed
_messages
= set()
527 self
._initialize
_geo
_bypass
({
528 'countries': self
._GEO
_COUNTRIES
,
529 'ip_blocks': self
._GEO
_IP
_BLOCKS
,
532 self
._initialize
_pre
_login
()
533 if self
.supports_login():
534 username
, password
= self
._get
_login
_info
()
536 self
._perform
_login
(username
, password
)
537 elif self
.get_param('username') and False not in (self
.IE_DESC
, self
._NETRC
_MACHINE
):
538 self
.report_warning(f
'Login with password is not supported for this website. {self._LOGIN_HINTS["cookies"]}')
539 self
._real
_initialize
()
542 def _initialize_geo_bypass(self
, geo_bypass_context
):
544 Initialize geo restriction bypass mechanism.
546 This method is used to initialize geo bypass mechanism based on faking
547 X-Forwarded-For HTTP header. A random country from provided country list
548 is selected and a random IP belonging to this country is generated. This
549 IP will be passed as X-Forwarded-For HTTP header in all subsequent
552 This method will be used for initial geo bypass mechanism initialization
553 during the instance initialization with _GEO_COUNTRIES and
556 You may also manually call it from extractor's code if geo bypass
557 information is not available beforehand (e.g. obtained during
558 extraction) or due to some other reason. In this case you should pass
559 this information in geo bypass context passed as first argument. It may
560 contain following fields:
562 countries: List of geo unrestricted countries (similar
564 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
565 (similar to _GEO_IP_BLOCKS)
568 if not self
._x
_forwarded
_for
_ip
:
570 # Geo bypass mechanism is explicitly disabled by user
571 if not self
.get_param('geo_bypass', True):
574 if not geo_bypass_context
:
575 geo_bypass_context
= {}
577 # Backward compatibility: previously _initialize_geo_bypass
578 # expected a list of countries, some 3rd party code may still use
580 if isinstance(geo_bypass_context
, (list, tuple)):
581 geo_bypass_context
= {
582 'countries': geo_bypass_context
,
585 # The whole point of geo bypass mechanism is to fake IP
586 # as X-Forwarded-For HTTP header based on some IP block or
589 # Path 1: bypassing based on IP block in CIDR notation
591 # Explicit IP block specified by user, use it right away
592 # regardless of whether extractor is geo bypassable or not
593 ip_block
= self
.get_param('geo_bypass_ip_block', None)
595 # Otherwise use random IP block from geo bypass context but only
596 # if extractor is known as geo bypassable
598 ip_blocks
= geo_bypass_context
.get('ip_blocks')
599 if self
._GEO
_BYPASS
and ip_blocks
:
600 ip_block
= random
.choice(ip_blocks
)
603 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(ip_block
)
604 self
._downloader
.write_debug(
605 '[debug] Using fake IP %s as X-Forwarded-For' % self
._x
_forwarded
_for
_ip
)
608 # Path 2: bypassing based on country code
610 # Explicit country code specified by user, use it right away
611 # regardless of whether extractor is geo bypassable or not
612 country
= self
.get_param('geo_bypass_country', None)
614 # Otherwise use random country code from geo bypass context but
615 # only if extractor is known as geo bypassable
617 countries
= geo_bypass_context
.get('countries')
618 if self
._GEO
_BYPASS
and countries
:
619 country
= random
.choice(countries
)
622 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country
)
623 self
._downloader
.write_debug(
624 f
'Using fake IP {self._x_forwarded_for_ip} ({country.upper()}) as X-Forwarded-For')
626 def extract(self
, url
):
627 """Extracts URL information and returns it in list of dicts."""
632 self
.write_debug('Extracting URL: %s' % url
)
633 ie_result
= self
._real
_extract
(url
)
634 if ie_result
is None:
636 if self
._x
_forwarded
_for
_ip
:
637 ie_result
['__x_forwarded_for_ip'] = self
._x
_forwarded
_for
_ip
638 subtitles
= ie_result
.get('subtitles')
639 if (subtitles
and 'live_chat' in subtitles
640 and 'no-live-chat' in self
.get_param('compat_opts', [])):
641 del subtitles
['live_chat']
643 except GeoRestrictedError
as e
:
644 if self
.__maybe
_fake
_ip
_and
_retry
(e
.countries
):
647 except UnsupportedError
:
649 except ExtractorError
as e
:
651 'video_id': e
.video_id
or self
.get_temp_id(url
),
653 'tb': e
.traceback
or sys
.exc_info()[2],
654 'expected': e
.expected
,
657 if hasattr(e
, 'countries'):
658 kwargs
['countries'] = e
.countries
659 raise type(e
)(e
.orig_msg
, **kwargs
)
660 except compat_http_client
.IncompleteRead
as e
:
661 raise ExtractorError('A network error has occurred.', cause
=e
, expected
=True, video_id
=self
.get_temp_id(url
))
662 except (KeyError, StopIteration) as e
:
663 raise ExtractorError('An extractor error has occurred.', cause
=e
, video_id
=self
.get_temp_id(url
))
665 def __maybe_fake_ip_and_retry(self
, countries
):
666 if (not self
.get_param('geo_bypass_country', None)
668 and self
.get_param('geo_bypass', True)
669 and not self
._x
_forwarded
_for
_ip
671 country_code
= random
.choice(countries
)
672 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country_code
)
673 if self
._x
_forwarded
_for
_ip
:
675 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
676 % (self
._x
_forwarded
_for
_ip
, country_code
.upper()))
680 def set_downloader(self
, downloader
):
681 """Sets a YoutubeDL instance as the downloader for this IE."""
682 self
._downloader
= downloader
684 def _initialize_pre_login(self
):
685 """ Intialization before login. Redefine in subclasses."""
688 def _perform_login(self
, username
, password
):
689 """ Login with username and password. Redefine in subclasses."""
692 def _real_initialize(self
):
693 """Real initialization process. Redefine in subclasses."""
696 def _real_extract(self
, url
):
697 """Real extraction process. Redefine in subclasses."""
698 raise NotImplementedError('This method must be implemented by subclasses')
702 """A string for getting the InfoExtractor with get_info_extractor"""
703 return cls
.__name
__[:-2]
707 return compat_str(type(self
).__name
__[:-2])
710 def __can_accept_status_code(err
, expected_status
):
711 assert isinstance(err
, compat_urllib_error
.HTTPError
)
712 if expected_status
is None:
714 elif callable(expected_status
):
715 return expected_status(err
.code
) is True
717 return err
.code
in variadic(expected_status
)
719 def _request_webpage(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
, expected_status
=None):
721 Return the response handle.
723 See _download_webpage docstring for arguments specification.
725 if not self
._downloader
._first
_webpage
_request
:
726 sleep_interval
= self
.get_param('sleep_interval_requests') or 0
727 if sleep_interval
> 0:
728 self
.to_screen('Sleeping %s seconds ...' % sleep_interval
)
729 time
.sleep(sleep_interval
)
731 self
._downloader
._first
_webpage
_request
= False
734 self
.report_download_webpage(video_id
)
735 elif note
is not False:
737 self
.to_screen(str(note
))
739 self
.to_screen(f
'{video_id}: {note}')
741 # Some sites check X-Forwarded-For HTTP header in order to figure out
742 # the origin of the client behind proxy. This allows bypassing geo
743 # restriction by faking this header's value to IP that belongs to some
744 # geo unrestricted country. We will do so once we encounter any
745 # geo restriction error.
746 if self
._x
_forwarded
_for
_ip
:
747 if 'X-Forwarded-For' not in headers
:
748 headers
['X-Forwarded-For'] = self
._x
_forwarded
_for
_ip
750 if isinstance(url_or_request
, compat_urllib_request
.Request
):
751 url_or_request
= update_Request(
752 url_or_request
, data
=data
, headers
=headers
, query
=query
)
755 url_or_request
= update_url_query(url_or_request
, query
)
756 if data
is not None or headers
:
757 url_or_request
= sanitized_Request(url_or_request
, data
, headers
)
759 return self
._downloader
.urlopen(url_or_request
)
760 except network_exceptions
as err
:
761 if isinstance(err
, compat_urllib_error
.HTTPError
):
762 if self
.__can
_accept
_status
_code
(err
, expected_status
):
763 # Retain reference to error to prevent file object from
764 # being closed before it can be read. Works around the
765 # effects of <https://bugs.python.org/issue15002>
766 # introduced in Python 3.4.1.
773 errnote
= 'Unable to download webpage'
775 errmsg
= f
'{errnote}: {error_to_compat_str(err)}'
777 raise ExtractorError(errmsg
, cause
=err
)
779 self
.report_warning(errmsg
)
782 def _download_webpage_handle(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, encoding
=None, data
=None, headers
={}, query={}
, expected_status
=None):
784 Return a tuple (page content as string, URL handle).
786 See _download_webpage docstring for arguments specification.
788 # Strip hashes from the URL (#1038)
789 if isinstance(url_or_request
, (compat_str
, str)):
790 url_or_request
= url_or_request
.partition('#')[0]
792 urlh
= self
._request
_webpage
(url_or_request
, video_id
, note
, errnote
, fatal
, data
=data
, headers
=headers
, query
=query
, expected_status
=expected_status
)
796 content
= self
._webpage
_read
_content
(urlh
, url_or_request
, video_id
, note
, errnote
, fatal
, encoding
=encoding
)
797 return (content
, urlh
)
800 def _guess_encoding_from_content(content_type
, webpage_bytes
):
801 m
= re
.match(r
'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type
)
803 encoding
= m
.group(1)
805 m
= re
.search(br
'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
806 webpage_bytes[:1024])
808 encoding = m.group(1).decode('ascii')
809 elif webpage_bytes.startswith(b'\xff\xfe'):
816 def __check_blocked(self, content):
817 first_block = content[:512]
818 if ('<title>Access to this site is blocked</title>' in content
819 and 'Websense' in first_block):
820 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
821 blocked_iframe = self._html_search_regex(
822 r'<iframe src="([^
"]+)"', content,
823 'Websense information URL
', default=None)
825 msg += ' Visit
%s for more details
' % blocked_iframe
826 raise ExtractorError(msg, expected=True)
827 if '<title
>The URL you requested has been blocked
</title
>' in first_block:
829 'Access to this webpage has been blocked by Indian censorship
. '
830 'Use a VPN
or proxy
server (with --proxy
) to route around it
.')
831 block_msg = self._html_search_regex(
832 r'</h1
><p
>(.*?
)</p
>',
833 content, 'block message
', default=None)
835 msg += ' (Message
: "%s")' % block_msg.replace('\n', ' ')
836 raise ExtractorError(msg, expected=True)
837 if ('<title
>TTK
:: Доступ к ресурсу ограничен
</title
>' in content
838 and 'blocklist
.rkn
.gov
.ru
' in content):
839 raise ExtractorError(
840 'Access to this webpage has been blocked by decision of the Russian government
. '
841 'Visit http
://blocklist
.rkn
.gov
.ru
/ for a block reason
.',
844 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
845 content_type = urlh.headers.get('Content
-Type
', '')
846 webpage_bytes = urlh.read()
847 if prefix is not None:
848 webpage_bytes = prefix + webpage_bytes
850 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
851 if self.get_param('dump_intermediate_pages
', False):
852 self.to_screen('Dumping request to
' + urlh.geturl())
853 dump = base64.b64encode(webpage_bytes).decode('ascii
')
854 self._downloader.to_screen(dump)
855 if self.get_param('write_pages
', False):
856 basen = f'{video_id}_{urlh.geturl()}
'
857 trim_length = self.get_param('trim_file_name
') or 240
858 if len(basen) > trim_length:
859 h = '___
' + hashlib.md5(basen.encode('utf
-8')).hexdigest()
860 basen = basen[:trim_length - len(h)] + h
861 raw_filename = basen + '.dump
'
862 filename = sanitize_filename(raw_filename, restricted=True)
863 self.to_screen('Saving request to
' + filename)
864 # Working around MAX_PATH limitation on Windows (see
865 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
866 if compat_os_name == 'nt
':
867 absfilepath = os.path.abspath(filename)
868 if len(absfilepath) > 259:
869 filename = '\\\\?
\\' + absfilepath
870 with open(filename, 'wb
') as outf:
871 outf.write(webpage_bytes)
874 content = webpage_bytes.decode(encoding, 'replace
')
876 content = webpage_bytes.decode('utf
-8', 'replace
')
878 self.__check_blocked(content)
882 def _download_webpage(
883 self, url_or_request, video_id, note=None, errnote=None,
884 fatal=True, tries=1, timeout=5, encoding=None, data=None,
885 headers={}, query={}, expected_status=None):
887 Return the data of the page as a string.
890 url_or_request -- plain text URL as a string or
891 a compat_urllib_request.Requestobject
892 video_id -- Video/playlist/item identifier (string)
895 note -- note printed before downloading (string)
896 errnote -- note printed in case of an error (string)
897 fatal -- flag denoting whether error should be considered fatal,
898 i.e. whether it should cause ExtractionError to be raised,
899 otherwise a warning will be reported and extraction continued
900 tries -- number of tries
901 timeout -- sleep interval between tries
902 encoding -- encoding for a page content decoding, guessed automatically
903 when not explicitly specified
904 data -- POST data (bytes)
905 headers -- HTTP headers (dict)
906 query -- URL query (dict)
907 expected_status -- allows to accept failed HTTP requests (non 2xx
908 status code) by explicitly specifying a set of accepted status
909 codes. Can be any of the following entities:
910 - an integer type specifying an exact failed status code to
912 - a list or a tuple of integer types specifying a list of
913 failed status codes to accept
914 - a callable accepting an actual failed status code and
915 returning True if it should be accepted
916 Note that this argument does not affect success status codes (2xx)
917 which are always accepted.
922 while success is False:
924 res = self._download_webpage_handle(
925 url_or_request, video_id, note, errnote, fatal,
926 encoding=encoding, data=data, headers=headers, query=query,
927 expected_status=expected_status)
929 except compat_http_client.IncompleteRead as e:
931 if try_count >= tries:
933 self._sleep(timeout, video_id)
940 def _download_xml_handle(
941 self, url_or_request, video_id, note='Downloading XML
',
942 errnote='Unable to download XML
', transform_source=None,
943 fatal=True, encoding=None, data=None, headers={}, query={},
944 expected_status=None):
946 Return a tuple (xml as an xml.etree.ElementTree.Element, URL handle).
948 See _download_webpage docstring for arguments specification.
950 res = self._download_webpage_handle(
951 url_or_request, video_id, note, errnote, fatal=fatal,
952 encoding=encoding, data=data, headers=headers, query=query,
953 expected_status=expected_status)
956 xml_string, urlh = res
957 return self._parse_xml(
958 xml_string, video_id, transform_source=transform_source,
962 self, url_or_request, video_id,
963 note='Downloading XML
', errnote='Unable to download XML
',
964 transform_source=None, fatal=True, encoding=None,
965 data=None, headers={}, query={}, expected_status=None):
967 Return the xml as an xml.etree.ElementTree.Element.
969 See _download_webpage docstring for arguments specification.
971 res = self._download_xml_handle(
972 url_or_request, video_id, note=note, errnote=errnote,
973 transform_source=transform_source, fatal=fatal, encoding=encoding,
974 data=data, headers=headers, query=query,
975 expected_status=expected_status)
976 return res if res is False else res[0]
978 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
980 xml_string = transform_source(xml_string)
982 return compat_etree_fromstring(xml_string.encode('utf
-8'))
983 except xml.etree.ElementTree.ParseError as ve:
984 errmsg = '%s: Failed to parse XML
' % video_id
986 raise ExtractorError(errmsg, cause=ve)
988 self.report_warning(errmsg + str(ve))
990 def _download_json_handle(
991 self, url_or_request, video_id, note='Downloading JSON metadata
',
992 errnote='Unable to download JSON metadata
', transform_source=None,
993 fatal=True, encoding=None, data=None, headers={}, query={},
994 expected_status=None):
996 Return a tuple (JSON object, URL handle).
998 See _download_webpage docstring for arguments specification.
1000 res = self._download_webpage_handle(
1001 url_or_request, video_id, note, errnote, fatal=fatal,
1002 encoding=encoding, data=data, headers=headers, query=query,
1003 expected_status=expected_status)
1006 json_string, urlh = res
1007 return self._parse_json(
1008 json_string, video_id, transform_source=transform_source,
1012 self, url_or_request, video_id, note='Downloading JSON metadata
',
1013 errnote='Unable to download JSON metadata
', transform_source=None,
1014 fatal=True, encoding=None, data=None, headers={}, query={},
1015 expected_status=None):
1017 Return the JSON object as a dict.
1019 See _download_webpage docstring for arguments specification.
1021 res = self._download_json_handle(
1022 url_or_request, video_id, note=note, errnote=errnote,
1023 transform_source=transform_source, fatal=fatal, encoding=encoding,
1024 data=data, headers=headers, query=query,
1025 expected_status=expected_status)
1026 return res if res is False else res[0]
1028 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
1029 if transform_source:
1030 json_string = transform_source(json_string)
1032 return json.loads(json_string, strict=False)
1033 except ValueError as ve:
1034 errmsg = '%s: Failed to parse JSON
' % video_id
1036 raise ExtractorError(errmsg, cause=ve)
1038 self.report_warning(errmsg + str(ve))
1040 def _parse_socket_response_as_json(self, data, video_id, transform_source=None, fatal=True):
1041 return self._parse_json(
1042 data[data.find('{'):data.rfind('}
') + 1],
1043 video_id, transform_source, fatal)
1045 def _download_socket_json_handle(
1046 self, url_or_request, video_id, note='Polling socket
',
1047 errnote='Unable to poll socket
', transform_source=None,
1048 fatal=True, encoding=None, data=None, headers={}, query={},
1049 expected_status=None):
1051 Return a tuple (JSON object, URL handle).
1053 See _download_webpage docstring for arguments specification.
1055 res = self._download_webpage_handle(
1056 url_or_request, video_id, note, errnote, fatal=fatal,
1057 encoding=encoding, data=data, headers=headers, query=query,
1058 expected_status=expected_status)
1062 return self._parse_socket_response_as_json(
1063 webpage, video_id, transform_source=transform_source,
1066 def _download_socket_json(
1067 self, url_or_request, video_id, note='Polling socket
',
1068 errnote='Unable to poll socket
', transform_source=None,
1069 fatal=True, encoding=None, data=None, headers={}, query={},
1070 expected_status=None):
1072 Return the JSON object as a dict.
1074 See _download_webpage docstring for arguments specification.
1076 res = self._download_socket_json_handle(
1077 url_or_request, video_id, note=note, errnote=errnote,
1078 transform_source=transform_source, fatal=fatal, encoding=encoding,
1079 data=data, headers=headers, query=query,
1080 expected_status=expected_status)
1081 return res if res is False else res[0]
1083 def report_warning(self, msg, video_id=None, *args, only_once=False, **kwargs):
1084 idstr = format_field(video_id, template='%s: ')
1085 msg = f'[{self.IE_NAME}
] {idstr}{msg}
'
1087 if f'WARNING
: {msg}
' in self._printed_messages:
1089 self._printed_messages.add(f'WARNING
: {msg}
')
1090 self._downloader.report_warning(msg, *args, **kwargs)
1092 def to_screen(self, msg, *args, **kwargs):
1093 """Print msg to screen, prefixing it with '[ie_name
]'"""
1094 self._downloader.to_screen(f'[{self.IE_NAME}
] {msg}
', *args, **kwargs)
1096 def write_debug(self, msg, *args, **kwargs):
1097 self._downloader.write_debug(f'[{self.IE_NAME}
] {msg}
', *args, **kwargs)
1099 def get_param(self, name, default=None, *args, **kwargs):
1100 if self._downloader:
1101 return self._downloader.params.get(name, default, *args, **kwargs)
1104 def report_drm(self, video_id, partial=False):
1105 self.raise_no_formats('This video
is DRM protected
', expected=True, video_id=video_id)
1107 def report_extraction(self, id_or_name):
1108 """Report information extraction."""
1109 self.to_screen('%s: Extracting information
' % id_or_name)
1111 def report_download_webpage(self, video_id):
1112 """Report webpage download."""
1113 self.to_screen('%s: Downloading webpage
' % video_id)
1115 def report_age_confirmation(self):
1116 """Report attempt to confirm age."""
1117 self.to_screen('Confirming age
')
1119 def report_login(self):
1120 """Report attempt to log in."""
1121 self.to_screen('Logging
in')
1123 def raise_login_required(
1124 self, msg='This video
is only available
for registered users
',
1125 metadata_available=False, method=NO_DEFAULT):
1126 if metadata_available and (
1127 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1128 self.report_warning(msg)
1130 if method is NO_DEFAULT:
1131 method = 'any
' if self.supports_login() else 'cookies
'
1132 if method is not None:
1133 assert method in self._LOGIN_HINTS, 'Invalid login method
'
1134 msg = f'{msg}
. {self._LOGIN_HINTS[method]}
'
1135 raise ExtractorError(msg, expected=True)
1137 def raise_geo_restricted(
1138 self, msg='This video
is not available
from your location due to geo restriction
',
1139 countries=None, metadata_available=False):
1140 if metadata_available and (
1141 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1142 self.report_warning(msg)
1144 raise GeoRestrictedError(msg, countries=countries)
1146 def raise_no_formats(self, msg, expected=False, video_id=None):
1148 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1149 self.report_warning(msg, video_id)
1150 elif isinstance(msg, ExtractorError):
1153 raise ExtractorError(msg, expected=expected, video_id=video_id)
1155 # Methods for following #608
1157 def url_result(url, ie=None, video_id=None, video_title=None, *, url_transparent=False, **kwargs):
1158 """Returns a URL that points to a page that should be processed"""
1160 kwargs['ie_key
'] = ie if isinstance(ie, str) else ie.ie_key()
1161 if video_id is not None:
1162 kwargs['id'] = video_id
1163 if video_title is not None:
1164 kwargs['title
'] = video_title
1167 '_type
': 'url_transparent
' if url_transparent else 'url
',
1171 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None, video_kwargs=None, **kwargs):
1172 urls = (self.url_result(self._proto_relative_url(m), ie, **(video_kwargs or {}))
1173 for m in orderedSet(map(getter, matches) if getter else matches))
1174 return self.playlist_result(urls, playlist_id, playlist_title, **kwargs)
1177 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None, *, multi_video=False, **kwargs):
1178 """Returns a playlist"""
1180 kwargs['id'] = playlist_id
1182 kwargs['title
'] = playlist_title
1183 if playlist_description is not None:
1184 kwargs['description
'] = playlist_description
1187 '_type
': 'multi_video
' if multi_video else 'playlist
',
1191 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1193 Perform a regex search on the given string, using a single or a list of
1194 patterns returning the first matching group.
1195 In case of failure return a default value or raise a WARNING or a
1196 RegexNotFoundError, depending on fatal, specifying the field name.
1200 elif isinstance(pattern, (str, re.Pattern)):
1201 mobj = re.search(pattern, string, flags)
1204 mobj = re.search(p, string, flags)
1208 _name = self._downloader._format_err(name, self._downloader.Styles.EMPHASIS)
1212 # return the first matching group
1213 return next(g for g in mobj.groups() if g is not None)
1214 elif isinstance(group, (list, tuple)):
1215 return tuple(mobj.group(g) for g in group)
1217 return mobj.group(group)
1218 elif default is not NO_DEFAULT:
1221 raise RegexNotFoundError('Unable to extract
%s' % _name)
1223 self.report_warning('unable to extract
%s' % _name + bug_reports_message())
1226 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1228 Like _search_regex, but strips HTML tags and unescapes entities.
1230 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1232 return clean_html(res).strip()
1236 def _get_netrc_login_info(self, netrc_machine=None):
1239 netrc_machine = netrc_machine or self._NETRC_MACHINE
1241 if self.get_param('usenetrc
', False):
1243 netrc_file = compat_expanduser(self.get_param('netrc_location
') or '~
')
1244 if os.path.isdir(netrc_file):
1245 netrc_file = os.path.join(netrc_file, '.netrc
')
1246 info = netrc.netrc(file=netrc_file).authenticators(netrc_machine)
1247 if info is not None:
1251 raise netrc.NetrcParseError(
1252 'No authenticators
for %s' % netrc_machine)
1253 except (OSError, netrc.NetrcParseError) as err:
1254 self.report_warning(
1255 'parsing
.netrc
: %s' % error_to_compat_str(err))
1257 return username, password
1259 def _get_login_info(self, username_option='username
', password_option='password
', netrc_machine=None):
1261 Get the login info as (username, password)
1262 First look for the manually specified credentials using username_option
1263 and password_option as keys in params dictionary. If no such credentials
1264 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1266 If there's no info available
, return (None, None)
1269 # Attempt to use provided username and password or .netrc data
1270 username = self.get_param(username_option)
1271 if username is not None:
1272 password = self.get_param(password_option)
1274 username, password = self._get_netrc_login_info(netrc_machine)
1276 return username, password
1278 def _get_tfa_info(self, note='two-factor verification code'):
1280 Get the two
-factor authentication info
1281 TODO
- asking the user will be required
for sms
/phone verify
1282 currently just uses the command line option
1283 If there
's no info available, return None
1286 tfa = self.get_param('twofactor
')
1290 return compat_getpass('Type
%s and press
[Return
]: ' % note)
1292 # Helper functions for extracting OpenGraph info
1294 def _og_regexes(prop):
1295 content_re = r'content
=(?
:"([^"]+?
)"|\'([^\']+?)\'|\s*([^\s"\'=<>`
]+?
))'
1296 property_re = (r'(?
:name|
property)=(?
:\'og
%(sep)s%(prop)s\'|
"og%(sep)s%(prop)s"|\s
*og
%(sep)s%(prop)s\b)'
1297 % {'prop': re.escape(prop), 'sep': '(?::|[:-])'})
1298 template = r'<meta
[^
>]+?
%s[^
>]+?
%s'
1300 template % (property_re, content_re),
1301 template % (content_re, property_re),
1305 def _meta_regex(prop):
1306 return r'''(?isx)<meta
1307 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1308 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1310 def _og_search_property(self, prop, html, name=None, **kargs):
1311 prop = variadic(prop)
1313 name = 'OpenGraph
%s' % prop[0]
1316 og_regexes.extend(self._og_regexes(p))
1317 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1320 return unescapeHTML(escaped)
1322 def _og_search_thumbnail(self, html, **kargs):
1323 return self._og_search_property('image
', html, 'thumbnail URL
', fatal=False, **kargs)
1325 def _og_search_description(self, html, **kargs):
1326 return self._og_search_property('description
', html, fatal=False, **kargs)
1328 def _og_search_title(self, html, *, fatal=False, **kargs):
1329 return self._og_search_property('title
', html, fatal=fatal, **kargs)
1331 def _og_search_video_url(self, html, name='video url
', secure=True, **kargs):
1332 regexes = self._og_regexes('video
') + self._og_regexes('video
:url
')
1334 regexes = self._og_regexes('video
:secure_url
') + regexes
1335 return self._html_search_regex(regexes, html, name, **kargs)
1337 def _og_search_url(self, html, **kargs):
1338 return self._og_search_property('url
', html, **kargs)
1340 def _html_extract_title(self, html, name='title
', *, fatal=False, **kwargs):
1341 return self._html_search_regex(r'(?s
)<title
>([^
<]+)</title
>', html, name, fatal=fatal, **kwargs)
1343 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1344 name = variadic(name)
1345 if display_name is None:
1346 display_name = name[0]
1347 return self._html_search_regex(
1348 [self._meta_regex(n) for n in name],
1349 html, display_name, fatal=fatal, group='content
', **kwargs)
1351 def _dc_search_uploader(self, html):
1352 return self._html_search_meta('dc
.creator
', html, 'uploader
')
1354 def _rta_search(self, html):
1355 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1356 if re.search(r'(?ix
)<meta\s
+name
="rating"\s
+'
1357 r' content
="RTA-5042-1996-1400-1577-RTA"',
1362 def _media_rating_search(self, html):
1363 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1364 rating = self._html_search_meta('rating
', html)
1376 return RATING_TABLE.get(rating.lower())
1378 def _family_friendly_search(self, html):
1379 # See http://schema.org/VideoObject
1380 family_friendly = self._html_search_meta(
1381 'isFamilyFriendly
', html, default=None)
1383 if not family_friendly:
1392 return RATING_TABLE.get(family_friendly.lower())
1394 def _twitter_search_player(self, html):
1395 return self._html_search_meta('twitter
:player
', html,
1396 'twitter card player
')
1398 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
1399 json_ld_list = list(re.finditer(JSON_LD_RE, html))
1400 default = kwargs.get('default
', NO_DEFAULT)
1401 # JSON-LD may be malformed and thus `fatal` should be respected.
1402 # At the same time `default` may be passed that assumes `fatal=False`
1403 # for _search_regex. Let's simulate the same behavior here
as well
.
1404 fatal
= kwargs
.get('fatal', True) if default
is NO_DEFAULT
else False
1406 for mobj
in json_ld_list
:
1407 json_ld_item
= self
._parse
_json
(
1408 mobj
.group('json_ld'), video_id
, fatal
=fatal
)
1409 if not json_ld_item
:
1411 if isinstance(json_ld_item
, dict):
1412 json_ld
.append(json_ld_item
)
1413 elif isinstance(json_ld_item
, (list, tuple)):
1414 json_ld
.extend(json_ld_item
)
1416 json_ld
= self
._json
_ld
(json_ld
, video_id
, fatal
=fatal
, expected_type
=expected_type
)
1419 if default
is not NO_DEFAULT
:
1422 raise RegexNotFoundError('Unable to extract JSON-LD')
1424 self
.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1427 def _json_ld(self
, json_ld
, video_id
, fatal
=True, expected_type
=None):
1428 if isinstance(json_ld
, compat_str
):
1429 json_ld
= self
._parse
_json
(json_ld
, video_id
, fatal
=fatal
)
1433 if not isinstance(json_ld
, (list, tuple, dict)):
1435 if isinstance(json_ld
, dict):
1438 INTERACTION_TYPE_MAP
= {
1439 'CommentAction': 'comment',
1440 'AgreeAction': 'like',
1441 'DisagreeAction': 'dislike',
1442 'LikeAction': 'like',
1443 'DislikeAction': 'dislike',
1444 'ListenAction': 'view',
1445 'WatchAction': 'view',
1446 'ViewAction': 'view',
1449 def extract_interaction_type(e
):
1450 interaction_type
= e
.get('interactionType')
1451 if isinstance(interaction_type
, dict):
1452 interaction_type
= interaction_type
.get('@type')
1453 return str_or_none(interaction_type
)
1455 def extract_interaction_statistic(e
):
1456 interaction_statistic
= e
.get('interactionStatistic')
1457 if isinstance(interaction_statistic
, dict):
1458 interaction_statistic
= [interaction_statistic
]
1459 if not isinstance(interaction_statistic
, list):
1461 for is_e
in interaction_statistic
:
1462 if not isinstance(is_e
, dict):
1464 if is_e
.get('@type') != 'InteractionCounter':
1466 interaction_type
= extract_interaction_type(is_e
)
1467 if not interaction_type
:
1469 # For interaction count some sites provide string instead of
1470 # an integer (as per spec) with non digit characters (e.g. ",")
1471 # so extracting count with more relaxed str_to_int
1472 interaction_count
= str_to_int(is_e
.get('userInteractionCount'))
1473 if interaction_count
is None:
1475 count_kind
= INTERACTION_TYPE_MAP
.get(interaction_type
.split('/')[-1])
1478 count_key
= '%s_count' % count_kind
1479 if info
.get(count_key
) is not None:
1481 info
[count_key
] = interaction_count
1483 def extract_chapter_information(e
):
1485 'title': part
.get('name'),
1486 'start_time': part
.get('startOffset'),
1487 'end_time': part
.get('endOffset'),
1488 } for part
in variadic(e
.get('hasPart') or []) if part
.get('@type') == 'Clip']
1489 for idx
, (last_c
, current_c
, next_c
) in enumerate(zip(
1490 [{'end_time': 0}
] + chapters
, chapters
, chapters
[1:])):
1491 current_c
['end_time'] = current_c
['end_time'] or next_c
['start_time']
1492 current_c
['start_time'] = current_c
['start_time'] or last_c
['end_time']
1493 if None in current_c
.values():
1494 self
.report_warning(f
'Chapter {idx} contains broken data. Not extracting chapters')
1497 chapters
[-1]['end_time'] = chapters
[-1]['end_time'] or info
['duration']
1498 info
['chapters'] = chapters
1500 def extract_video_object(e
):
1501 assert e
['@type'] == 'VideoObject'
1502 author
= e
.get('author')
1504 'url': url_or_none(e
.get('contentUrl')),
1505 'title': unescapeHTML(e
.get('name')),
1506 'description': unescapeHTML(e
.get('description')),
1507 'thumbnails': [{'url': url_or_none(url)}
1508 for url
in variadic(traverse_obj(e
, 'thumbnailUrl', 'thumbnailURL'))],
1509 'duration': parse_duration(e
.get('duration')),
1510 'timestamp': unified_timestamp(e
.get('uploadDate')),
1511 # author can be an instance of 'Organization' or 'Person' types.
1512 # both types can have 'name' property(inherited from 'Thing' type). [1]
1513 # however some websites are using 'Text' type instead.
1514 # 1. https://schema.org/VideoObject
1515 'uploader': author
.get('name') if isinstance(author
, dict) else author
if isinstance(author
, compat_str
) else None,
1516 'filesize': float_or_none(e
.get('contentSize')),
1517 'tbr': int_or_none(e
.get('bitrate')),
1518 'width': int_or_none(e
.get('width')),
1519 'height': int_or_none(e
.get('height')),
1520 'view_count': int_or_none(e
.get('interactionCount')),
1522 extract_interaction_statistic(e
)
1523 extract_chapter_information(e
)
1525 def traverse_json_ld(json_ld
, at_top_level
=True):
1527 if at_top_level
and '@context' not in e
:
1529 if at_top_level
and set(e
.keys()) == {'@context', '@graph'}
:
1530 traverse_json_ld(variadic(e
['@graph'], allowed_types
=(dict,)), at_top_level
=False)
1532 item_type
= e
.get('@type')
1533 if expected_type
is not None and expected_type
!= item_type
:
1535 rating
= traverse_obj(e
, ('aggregateRating', 'ratingValue'), expected_type
=float_or_none
)
1536 if rating
is not None:
1537 info
['average_rating'] = rating
1538 if item_type
in ('TVEpisode', 'Episode'):
1539 episode_name
= unescapeHTML(e
.get('name'))
1541 'episode': episode_name
,
1542 'episode_number': int_or_none(e
.get('episodeNumber')),
1543 'description': unescapeHTML(e
.get('description')),
1545 if not info
.get('title') and episode_name
:
1546 info
['title'] = episode_name
1547 part_of_season
= e
.get('partOfSeason')
1548 if isinstance(part_of_season
, dict) and part_of_season
.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1550 'season': unescapeHTML(part_of_season
.get('name')),
1551 'season_number': int_or_none(part_of_season
.get('seasonNumber')),
1553 part_of_series
= e
.get('partOfSeries') or e
.get('partOfTVSeries')
1554 if isinstance(part_of_series
, dict) and part_of_series
.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1555 info
['series'] = unescapeHTML(part_of_series
.get('name'))
1556 elif item_type
== 'Movie':
1558 'title': unescapeHTML(e
.get('name')),
1559 'description': unescapeHTML(e
.get('description')),
1560 'duration': parse_duration(e
.get('duration')),
1561 'timestamp': unified_timestamp(e
.get('dateCreated')),
1563 elif item_type
in ('Article', 'NewsArticle'):
1565 'timestamp': parse_iso8601(e
.get('datePublished')),
1566 'title': unescapeHTML(e
.get('headline')),
1567 'description': unescapeHTML(e
.get('articleBody') or e
.get('description')),
1569 if traverse_obj(e
, ('video', 0, '@type')) == 'VideoObject':
1570 extract_video_object(e
['video'][0])
1571 elif item_type
== 'VideoObject':
1572 extract_video_object(e
)
1573 if expected_type
is None:
1577 video
= e
.get('video')
1578 if isinstance(video
, dict) and video
.get('@type') == 'VideoObject':
1579 extract_video_object(video
)
1580 if expected_type
is None:
1584 traverse_json_ld(json_ld
)
1586 return filter_dict(info
)
1588 def _search_nextjs_data(self
, webpage
, video_id
, *, transform_source
=None, fatal
=True, **kw
):
1589 return self
._parse
_json
(
1591 r
'(?s)<script[^>]+id=[\'"]__NEXT_DATA__[\'"][^
>]*>([^
<]+)</script
>',
1592 webpage, 'next
.js data
', fatal=fatal, **kw),
1593 video_id, transform_source=transform_source, fatal=fatal)
1595 def _search_nuxt_data(self, webpage, video_id, context_name='__NUXT__
'):
1596 ''' Parses Nuxt.js metadata. This works as long as the function __NUXT__ invokes is a pure function. '''
1597 # not all website do this, but it can be changed
1598 # https://stackoverflow.com/questions/67463109/how-to-change-or-hide-nuxt-and-nuxt-keyword-in-page-source
1599 rectx = re.escape(context_name)
1600 js, arg_keys, arg_vals = self._search_regex(
1601 (r'<script
>window\
.%s=\
(function\
((?P
<arg_keys
>.*?
)\
)\{return\s(?P<js>\{.*?\}
)\
}\
((?P
<arg_vals
>.+?
)\
)\
);?
</script
>' % rectx,
1602 r'%s\
(.*?\
(function\
((?P
<arg_keys
>.*?
)\
)\{return\s(?P<js>\{.*?\}
)\
}\
((?P
<arg_vals
>.*?
)\
)' % rectx),
1603 webpage, context_name, group=['js
', 'arg_keys
', 'arg_vals
'])
1605 args = dict(zip(arg_keys.split(','), arg_vals.split(',')))
1607 for key, val in args.items():
1608 if val in ('undefined
', 'void
0'):
1611 return self._parse_json(js_to_json(js, args), video_id)['data
'][0]
1614 def _hidden_inputs(html):
1615 html = re.sub(r'<!--(?
:(?
!<!--).)*-->', '', html)
1617 for input in re.findall(r'(?i
)(<input[^
>]+>)', html):
1618 attrs = extract_attributes(input)
1621 if attrs.get('type') not in ('hidden
', 'submit
'):
1623 name = attrs.get('name
') or attrs.get('id')
1624 value = attrs.get('value
')
1625 if name and value is not None:
1626 hidden_inputs[name] = value
1627 return hidden_inputs
1629 def _form_hidden_inputs(self, form_id, html):
1630 form = self._search_regex(
1631 r'(?
is)<form
[^
>]+?
id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1632 html, '%s form' % form_id, group='form')
1633 return self._hidden_inputs(form)
1636 regex = r' *((?P<reverse>\+)?(?P<field>[a-zA-Z0-9_]+)((?P<separator>[~:])(?P<limit>.*?))?)? *$'
1638 default = ('hidden', 'aud_or_vid', 'hasvid', 'ie_pref', 'lang', 'quality',
1639 'res', 'fps', 'hdr:12', 'codec:vp9.2', 'size', 'br', 'asr',
1640 'proto', 'ext', 'hasaud', 'source', 'id') # These must not be aliases
1641 ytdl_default = ('hasaud', 'lang', 'quality', 'tbr', 'filesize', 'vbr',
1642 'height', 'width', 'proto', 'vext', 'abr', 'aext',
1643 'fps', 'fs_approx', 'source', 'id')
1646 'vcodec': {'type': 'ordered', 'regex': True,
1647 'order': ['av0?1', 'vp0?9.2', 'vp0?9', '[hx]265|he?vc?', '[hx]264|avc', 'vp0?8', 'mp4v|h263', 'theora', '', None, 'none']},
1648 'acodec': {'type': 'ordered', 'regex': True,
1649 'order': ['[af]lac', 'wav|aiff', 'opus', 'vorbis|ogg', 'aac', 'mp?4a?', 'mp3', 'e-?a?c-?3', 'ac-?3', 'dts', '', None, 'none']},
1650 'hdr': {'type': 'ordered', 'regex': True, 'field': 'dynamic_range',
1651 'order': ['dv', '(hdr)?12', r'(hdr)?10\+', '(hdr)?10', 'hlg', '', 'sdr', None]},
1652 'proto': {'type': 'ordered', 'regex': True, 'field': 'protocol',
1653 'order': ['(ht|f)tps', '(ht|f)tp$', 'm3u8.*', '.*dash', 'websocket_frag', 'rtmpe?', '', 'mms|rtsp', 'ws|websocket', 'f4']},
1654 'vext': {'type': 'ordered', 'field': 'video_ext',
1655 'order': ('mp4', 'webm', 'flv', '', 'none'),
1656 'order_free': ('webm', 'mp4', 'flv', '', 'none')},
1657 'aext': {'type': 'ordered', 'field': 'audio_ext',
1658 'order': ('m4a', 'aac', 'mp3', 'ogg', 'opus', 'webm', '', 'none'),
1659 'order_free': ('opus', 'ogg', 'webm', 'm4a', 'mp3', 'aac', '', 'none')},
1660 'hidden': {'visible': False, 'forced': True, 'type': 'extractor', 'max': -1000},
1661 'aud_or_vid': {'visible': False, 'forced': True, 'type': 'multiple',
1662 'field': ('vcodec', 'acodec'),
1663 'function': lambda it: int(any(v != 'none' for v in it))},
1664 'ie_pref': {'priority': True, 'type': 'extractor'},
1665 'hasvid': {'priority': True, 'field': 'vcodec', 'type': 'boolean', 'not_in_list': ('none',)},
1666 'hasaud': {'field': 'acodec', 'type': 'boolean', 'not_in_list': ('none',)},
1667 'lang': {'convert': 'float', 'field': 'language_preference', 'default': -1},
1668 'quality': {'convert': 'float', 'default': -1},
1669 'filesize': {'convert': 'bytes'},
1670 'fs_approx': {'convert': 'bytes', 'field': 'filesize_approx'},
1671 'id': {'convert': 'string', 'field': 'format_id'},
1672 'height': {'convert': 'float_none'},
1673 'width': {'convert': 'float_none'},
1674 'fps': {'convert': 'float_none'},
1675 'tbr': {'convert': 'float_none'},
1676 'vbr': {'convert': 'float_none'},
1677 'abr': {'convert': 'float_none'},
1678 'asr': {'convert': 'float_none'},
1679 'source': {'convert': 'float', 'field': 'source_preference', 'default': -1},
1681 'codec': {'type': 'combined', 'field': ('vcodec', 'acodec')},
1682 'br': {'type': 'combined', 'field': ('tbr', 'vbr', 'abr'), 'same_limit': True},
1683 'size': {'type': 'combined', 'same_limit': True, 'field': ('filesize', 'fs_approx')},
1684 'ext': {'type': 'combined', 'field': ('vext', 'aext')},
1685 'res': {'type': 'multiple', 'field': ('height', 'width'),
1686 'function': lambda it: (lambda l: min(l) if l else 0)(tuple(filter(None, it)))},
1688 # For compatibility with youtube-dl
1689 'format_id': {'type': 'alias', 'field': 'id'},
1690 'preference': {'type': 'alias', 'field': 'ie_pref'},
1691 'language_preference': {'type': 'alias', 'field': 'lang'},
1692 'source_preference': {'type': 'alias', 'field': 'source'},
1693 'protocol': {'type': 'alias', 'field': 'proto'},
1694 'filesize_approx': {'type': 'alias', 'field': 'fs_approx'},
1697 'dimension': {'type': 'alias', 'field': 'res', 'deprecated': True},
1698 'resolution': {'type': 'alias', 'field': 'res', 'deprecated': True},
1699 'extension': {'type': 'alias', 'field': 'ext', 'deprecated': True},
1700 'bitrate': {'type': 'alias', 'field': 'br', 'deprecated': True},
1701 'total_bitrate': {'type': 'alias', 'field': 'tbr', 'deprecated': True},
1702 'video_bitrate': {'type': 'alias', 'field': 'vbr', 'deprecated': True},
1703 'audio_bitrate': {'type': 'alias', 'field': 'abr', 'deprecated': True},
1704 'framerate': {'type': 'alias', 'field': 'fps', 'deprecated': True},
1705 'filesize_estimate': {'type': 'alias', 'field': 'size', 'deprecated': True},
1706 'samplerate': {'type': 'alias', 'field': 'asr', 'deprecated': True},
1707 'video_ext': {'type': 'alias', 'field': 'vext', 'deprecated': True},
1708 'audio_ext': {'type': 'alias', 'field': 'aext', 'deprecated': True},
1709 'video_codec': {'type': 'alias', 'field': 'vcodec', 'deprecated': True},
1710 'audio_codec': {'type': 'alias', 'field': 'acodec', 'deprecated': True},
1711 'video': {'type': 'alias', 'field': 'hasvid', 'deprecated': True},
1712 'has_video': {'type': 'alias', 'field': 'hasvid', 'deprecated': True},
1713 'audio': {'type': 'alias', 'field': 'hasaud', 'deprecated': True},
1714 'has_audio': {'type': 'alias', 'field': 'hasaud', 'deprecated': True},
1715 'extractor': {'type': 'alias', 'field': 'ie_pref', 'deprecated': True},
1716 'extractor_preference': {'type': 'alias', 'field': 'ie_pref', 'deprecated': True},
1719 def __init__(self, ie, field_preference):
1721 self.ydl = ie._downloader
1722 self.evaluate_params(self.ydl.params, field_preference)
1723 if ie.get_param('verbose'):
1724 self.print_verbose_info(self.ydl.write_debug)
1726 def _get_field_setting(self, field, key):
1727 if field not in self.settings:
1728 if key in ('forced', 'priority'):
1730 self.ydl.deprecation_warning(
1731 f'Using arbitrary fields ({field}) for format sorting is deprecated '
1732 'and may be removed in a future version')
1733 self.settings[field] = {}
1734 propObj = self.settings[field]
1735 if key not in propObj:
1736 type = propObj.get('type')
1738 default = 'preference' if type == 'extractor' else (field,) if type in ('combined', 'multiple') else field
1739 elif key == 'convert':
1740 default = 'order' if type == 'ordered' else 'float_string' if field else 'ignore'
1742 default = {'type': 'field', 'visible': True, 'order': [], 'not_in_list': (None,)}.get(key, None)
1743 propObj[key] = default
1746 def _resolve_field_value(self, field, value, convertNone=False):
1751 value = value.lower()
1752 conversion = self._get_field_setting(field, 'convert')
1753 if conversion == 'ignore':
1755 if conversion == 'string':
1757 elif conversion == 'float_none':
1758 return float_or_none(value)
1759 elif conversion == 'bytes':
1760 return FileDownloader.parse_bytes(value)
1761 elif conversion == 'order':
1762 order_list = (self._use_free_order and self._get_field_setting(field, 'order_free')) or self._get_field_setting(field, 'order')
1763 use_regex = self._get_field_setting(field, 'regex')
1764 list_length = len(order_list)
1765 empty_pos = order_list.index('') if '' in order_list else list_length + 1
1766 if use_regex and value is not None:
1767 for i, regex in enumerate(order_list):
1768 if regex and re.match(regex, value):
1769 return list_length - i
1770 return list_length - empty_pos # not in list
1771 else: # not regex or value = None
1772 return list_length - (order_list.index(value) if value in order_list else empty_pos)
1774 if value.isnumeric():
1777 self.settings[field]['convert'] = 'string'
1780 def evaluate_params(self, params, sort_extractor):
1781 self._use_free_order = params.get('prefer_free_formats', False)
1782 self._sort_user = params.get('format_sort', [])
1783 self._sort_extractor = sort_extractor
1785 def add_item(field, reverse, closest, limit_text):
1786 field = field.lower()
1787 if field in self._order:
1789 self._order.append(field)
1790 limit = self._resolve_field_value(field, limit_text)
1793 'closest': False if limit is None else closest,
1794 'limit_text': limit_text,
1796 if field in self.settings:
1797 self.settings[field].update(data)
1799 self.settings[field] = data
1802 tuple(field for field in self.default if self._get_field_setting(field, 'forced'))
1803 + (tuple() if params.get('format_sort_force', False)
1804 else tuple(field for field in self.default if self._get_field_setting(field, 'priority')))
1805 + tuple(self._sort_user) + tuple(sort_extractor) + self.default)
1807 for item in sort_list:
1808 match = re.match(self.regex, item)
1810 raise ExtractorError('Invalid format sort string "%s" given by extractor' % item)
1811 field = match.group('field')
1814 if self._get_field_setting(field, 'type') == 'alias':
1815 alias, field = field, self._get_field_setting(field, 'field')
1816 if self._get_field_setting(alias, 'deprecated'):
1817 self.ydl.deprecation_warning(
1818 f'Format sorting alias {alias} is deprecated '
1819 f'and may be removed in a future version. Please use {field} instead')
1820 reverse = match.group('reverse') is not None
1821 closest = match.group('separator') == '~'
1822 limit_text = match.group('limit')
1824 has_limit = limit_text is not None
1825 has_multiple_fields = self._get_field_setting(field, 'type') == 'combined'
1826 has_multiple_limits = has_limit and has_multiple_fields and not self._get_field_setting(field, 'same_limit')
1828 fields = self._get_field_setting(field, 'field') if has_multiple_fields else (field,)
1829 limits = limit_text.split(':') if has_multiple_limits else (limit_text,) if has_limit else tuple()
1830 limit_count = len(limits)
1831 for (i, f) in enumerate(fields):
1832 add_item(f, reverse, closest,
1833 limits[i] if i < limit_count
1834 else limits[0] if has_limit and not has_multiple_limits
1837 def print_verbose_info(self, write_debug):
1839 write_debug('Sort order given by user: %s' % ', '.join(self._sort_user))
1840 if self._sort_extractor:
1841 write_debug('Sort order given by extractor: %s' % ', '.join(self._sort_extractor))
1842 write_debug('Formats sorted by: %s' % ', '.join(['%s%s%s' % (
1843 '+' if self._get_field_setting(field, 'reverse') else '', field,
1844 '%s%s(%s)' % ('~' if self._get_field_setting(field, 'closest') else ':',
1845 self._get_field_setting(field, 'limit_text'),
1846 self._get_field_setting(field, 'limit'))
1847 if self._get_field_setting(field, 'limit_text') is not None else '')
1848 for field in self._order if self._get_field_setting(field, 'visible')]))
1850 def _calculate_field_preference_from_value(self, format, field, type, value):
1851 reverse = self._get_field_setting(field, 'reverse')
1852 closest = self._get_field_setting(field, 'closest')
1853 limit = self._get_field_setting(field, 'limit')
1855 if type == 'extractor':
1856 maximum = self._get_field_setting(field, 'max')
1857 if value is None or (maximum is not None and value >= maximum):
1859 elif type == 'boolean':
1860 in_list = self._get_field_setting(field, 'in_list')
1861 not_in_list = self._get_field_setting(field, 'not_in_list')
1862 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
1863 elif type == 'ordered':
1864 value = self._resolve_field_value(field, value, True)
1866 # try to convert to number
1867 val_num = float_or_none(value, default=self._get_field_setting(field, 'default'))
1868 is_num = self._get_field_setting(field, 'convert') != 'string' and val_num is not None
1872 return ((-10, 0) if value is None
1873 else (1, value, 0) if not is_num # if a field has mixed strings and numbers, strings are sorted higher
1874 else (0, -abs(value - limit), value - limit if reverse else limit - value) if closest
1875 else (0, value, 0) if not reverse and (limit is None or value <= limit)
1876 else (0, -value, 0) if limit is None or (reverse and value == limit) or value > limit
1877 else (-1, value, 0))
1879 def _calculate_field_preference(self, format, field):
1880 type = self._get_field_setting(field, 'type') # extractor, boolean, ordered, field, multiple
1881 get_value = lambda f: format.get(self._get_field_setting(f, 'field'))
1882 if type == 'multiple':
1883 type = 'field' # Only 'field' is allowed in multiple for now
1884 actual_fields = self._get_field_setting(field, 'field')
1886 value = self._get_field_setting(field, 'function')(get_value(f) for f in actual_fields)
1888 value = get_value(field)
1889 return self._calculate_field_preference_from_value(format, field, type, value)
1891 def calculate_preference(self, format):
1892 # Determine missing protocol
1893 if not format.get('protocol'):
1894 format['protocol'] = determine_protocol(format)
1896 # Determine missing ext
1897 if not format.get('ext') and 'url' in format:
1898 format['ext'] = determine_ext(format['url'])
1899 if format.get('vcodec') == 'none':
1900 format['audio_ext'] = format['ext'] if format.get('acodec') != 'none' else 'none'
1901 format['video_ext'] = 'none'
1903 format['video_ext'] = format['ext']
1904 format['audio_ext'] = 'none'
1905 # if format.get('preference') is None and format.get('ext') in ('f4f', 'f4m'): # Not supported?
1906 # format['preference'] = -1000
1908 # Determine missing bitrates
1909 if format.get('tbr') is None:
1910 if format.get('vbr') is not None and format.get('abr') is not None:
1911 format['tbr'] = format.get('vbr', 0) + format.get('abr', 0)
1913 if format.get('vcodec') != 'none' and format.get('vbr') is None:
1914 format['vbr'] = format.get('tbr') - format.get('abr', 0)
1915 if format.get('acodec') != 'none' and format.get('abr') is None:
1916 format['abr'] = format.get('tbr') - format.get('vbr', 0)
1918 return tuple(self._calculate_field_preference(format, field) for field in self._order)
1920 def _sort_formats(self, formats, field_preference=[]):
1923 format_sort = self.FormatSort(self, field_preference)
1924 formats.sort(key=lambda f: format_sort.calculate_preference(f))
1926 def _check_formats(self, formats, video_id):
1928 formats[:] = filter(
1929 lambda f: self._is_valid_url(
1931 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1935 def _remove_duplicate_formats(formats):
1939 if f['url'] not in format_urls:
1940 format_urls.add(f['url'])
1941 unique_formats.append(f)
1942 formats[:] = unique_formats
1944 def _is_valid_url(self, url, video_id, item='video', headers={}):
1945 url = self._proto_relative_url(url, scheme='http:')
1946 # For now assume non HTTP(S) URLs always valid
1947 if not (url.startswith('http://') or url.startswith('https://')):
1950 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1952 except ExtractorError as e:
1954 '%s: %s URL is invalid, skipping: %s'
1955 % (video_id, item, error_to_compat_str(e.cause)))
1958 def http_scheme(self):
1959 """ Either "http
:" or "https
:", depending on the user's preferences """
1962 if self.get_param('prefer_insecure', False)
1965 def _proto_relative_url(self, url, scheme=None):
1968 if url.startswith('//'):
1970 scheme = self.http_scheme()
1975 def _sleep(self, timeout, video_id, msg_template=None):
1976 if msg_template is None:
1977 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1978 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1982 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1983 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1984 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
1985 manifest = self._download_xml(
1986 manifest_url, video_id, 'Downloading f4m manifest',
1987 'Unable to download f4m manifest',
1988 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1989 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
1990 transform_source=transform_source,
1991 fatal=fatal, data=data, headers=headers, query=query)
1993 if manifest is False:
1996 return self._parse_f4m_formats(
1997 manifest, manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1998 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
2000 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
2001 transform_source=lambda s: fix_xml_ampersands(s).strip(),
2002 fatal=True, m3u8_id=None):
2003 if not isinstance(manifest, xml.etree.ElementTree.Element) and not fatal:
2006 # currently yt-dlp cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
2007 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
2008 if akamai_pv is not None and ';' in akamai_pv.text:
2009 playerVerificationChallenge = akamai_pv.text.split(';')[0]
2010 if playerVerificationChallenge.strip() != '':
2014 manifest_version = '1.0'
2015 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
2017 manifest_version = '2.0'
2018 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
2019 # Remove unsupported DRM protected media from final formats
2020 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
2021 media_nodes = remove_encrypted_media(media_nodes)
2025 manifest_base_url = get_base_url(manifest)
2027 bootstrap_info = xpath_element(
2028 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
2029 'bootstrap info', default=None)
2032 mime_type = xpath_text(
2033 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
2034 'base URL', default=None)
2035 if mime_type and mime_type.startswith('audio/'):
2038 for i, media_el in enumerate(media_nodes):
2039 tbr = int_or_none(media_el.attrib.get('bitrate'))
2040 width = int_or_none(media_el.attrib.get('width'))
2041 height = int_or_none(media_el.attrib.get('height'))
2042 format_id = join_nonempty(f4m_id, tbr or i)
2043 # If <bootstrapInfo> is present, the specified f4m is a
2044 # stream-level manifest, and only set-level manifests may refer to
2045 # external resources. See section 11.4 and section 4 of F4M spec
2046 if bootstrap_info is None:
2048 # @href is introduced in 2.0, see section 11.6 of F4M spec
2049 if manifest_version == '2.0':
2050 media_url = media_el.attrib.get('href')
2051 if media_url is None:
2052 media_url = media_el.attrib.get('url')
2056 media_url if media_url.startswith('http://') or media_url.startswith('https://')
2057 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
2058 # If media_url is itself a f4m manifest do the recursive extraction
2059 # since bitrates in parent manifest (this one) and media_url manifest
2060 # may differ leading to inability to resolve the format by requested
2061 # bitrate in f4m downloader
2062 ext = determine_ext(manifest_url)
2064 f4m_formats = self._extract_f4m_formats(
2065 manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
2066 transform_source=transform_source, fatal=fatal)
2067 # Sometimes stream-level manifest contains single media entry that
2068 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
2069 # At the same time parent's media entry in set-level manifest may
2070 # contain it. We will copy it from parent in such cases.
2071 if len(f4m_formats) == 1:
2074 'tbr': f.get('tbr') or tbr,
2075 'width': f.get('width') or width,
2076 'height': f.get('height') or height,
2077 'format_id': f.get('format_id') if not tbr else format_id,
2080 formats.extend(f4m_formats)
2083 formats.extend(self._extract_m3u8_formats(
2084 manifest_url, video_id, 'mp4', preference=preference,
2085 quality=quality, m3u8_id=m3u8_id, fatal=fatal))
2088 'format_id': format_id,
2089 'url': manifest_url,
2090 'manifest_url': manifest_url,
2091 'ext': 'flv' if bootstrap_info is not None else None,
2097 'preference': preference,
2102 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, quality=None, m3u8_id=None):
2104 'format_id': join_nonempty(m3u8_id, 'meta'),
2108 'preference': preference - 100 if preference else -100,
2110 'resolution': 'multiple',
2111 'format_note': 'Quality selection URL',
2114 def _report_ignoring_subs(self, name):
2115 self.report_warning(bug_reports_message(
2116 f'Ignoring subtitle tracks found in the {name} manifest; '
2117 'if any subtitle tracks are missing,'
2120 def _extract_m3u8_formats(self, *args, **kwargs):
2121 fmts, subs = self._extract_m3u8_formats_and_subtitles(*args, **kwargs)
2123 self._report_ignoring_subs('HLS')
2126 def _extract_m3u8_formats_and_subtitles(
2127 self, m3u8_url, video_id, ext=None, entry_protocol='m3u8_native',
2128 preference=None, quality=None, m3u8_id=None, note=None,
2129 errnote=None, fatal=True, live=False, data=None, headers={},
2132 res = self._download_webpage_handle(
2134 note='Downloading m3u8 information' if note is None else note,
2135 errnote='Failed to download m3u8 information' if errnote is None else errnote,
2136 fatal=fatal, data=data, headers=headers, query=query)
2141 m3u8_doc, urlh = res
2142 m3u8_url = urlh.geturl()
2144 return self._parse_m3u8_formats_and_subtitles(
2145 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
2146 preference=preference, quality=quality, m3u8_id=m3u8_id,
2147 note=note, errnote=errnote, fatal=fatal, live=live, data=data,
2148 headers=headers, query=query, video_id=video_id)
2150 def _parse_m3u8_formats_and_subtitles(
2151 self, m3u8_doc, m3u8_url=None, ext=None, entry_protocol='m3u8_native',
2152 preference=None, quality=None, m3u8_id=None, live=False, note=None,
2153 errnote=None, fatal=True, data=None, headers={}, query={},
2155 formats, subtitles = [], {}
2157 has_drm = re.search('|'.join([
2158 r'#EXT-X-FAXS-CM:', # Adobe Flash Access
2159 r'#EXT-X-(?:SESSION-)?KEY:.*?URI="skd
://', # Apple FairPlay
2162 def format_url(url):
2163 return url if re.match(r'^https?
://', url) else compat_urlparse.urljoin(m3u8_url, url)
2165 if self.get_param('hls_split_discontinuity
', False):
2166 def _extract_m3u8_playlist_indices(manifest_url=None, m3u8_doc=None):
2168 if not manifest_url:
2170 m3u8_doc = self._download_webpage(
2171 manifest_url, video_id, fatal=fatal, data=data, headers=headers,
2172 note=False, errnote='Failed to download m3u8 playlist information
')
2173 if m3u8_doc is False:
2175 return range(1 + sum(line.startswith('#EXT-X-DISCONTINUITY') for line in m3u8_doc.splitlines()))
2178 def _extract_m3u8_playlist_indices(*args
, **kwargs
):
2182 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
2183 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
2184 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
2186 # We should try extracting formats only from master playlists [1, 4.3.4],
2187 # i.e. playlists that describe available qualities. On the other hand
2188 # media playlists [1, 4.3.3] should be returned as is since they contain
2189 # just the media without qualities renditions.
2190 # Fortunately, master playlist can be easily distinguished from media
2191 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
2192 # master playlist tags MUST NOT appear in a media playlist and vice versa.
2193 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
2194 # media playlist and MUST NOT appear in master playlist thus we can
2195 # clearly detect media playlist with this criterion.
2197 if '#EXT-X-TARGETDURATION' in m3u8_doc
: # media playlist, return as is
2199 'format_id': join_nonempty(m3u8_id
, idx
),
2200 'format_index': idx
,
2201 'url': m3u8_url
or encode_data_uri(m3u8_doc
.encode('utf-8'), 'application/x-mpegurl'),
2203 'protocol': entry_protocol
,
2204 'preference': preference
,
2207 } for idx
in _extract_m3u8_playlist_indices(m3u8_doc
=m3u8_doc
)]
2209 return formats
, subtitles
2212 last_stream_inf
= {}
2214 def extract_media(x_media_line
):
2215 media
= parse_m3u8_attributes(x_media_line
)
2216 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
2217 media_type
, group_id
, name
= media
.get('TYPE'), media
.get('GROUP-ID'), media
.get('NAME')
2218 if not (media_type
and group_id
and name
):
2220 groups
.setdefault(group_id
, []).append(media
)
2221 # <https://tools.ietf.org/html/rfc8216#section-4.3.4.1>
2222 if media_type
== 'SUBTITLES':
2223 # According to RFC 8216 §4.3.4.2.1, URI is REQUIRED in the
2224 # EXT-X-MEDIA tag if the media type is SUBTITLES.
2225 # However, lack of URI has been spotted in the wild.
2226 # e.g. NebulaIE; see https://github.com/yt-dlp/yt-dlp/issues/339
2227 if not media
.get('URI'):
2229 url
= format_url(media
['URI'])
2232 'ext': determine_ext(url
),
2234 if sub_info
['ext'] == 'm3u8':
2235 # Per RFC 8216 §3.1, the only possible subtitle format m3u8
2236 # files may contain is WebVTT:
2237 # <https://tools.ietf.org/html/rfc8216#section-3.1>
2238 sub_info
['ext'] = 'vtt'
2239 sub_info
['protocol'] = 'm3u8_native'
2240 lang
= media
.get('LANGUAGE') or 'und'
2241 subtitles
.setdefault(lang
, []).append(sub_info
)
2242 if media_type
not in ('VIDEO', 'AUDIO'):
2244 media_url
= media
.get('URI')
2246 manifest_url
= format_url(media_url
)
2248 'format_id': join_nonempty(m3u8_id
, group_id
, name
, idx
),
2249 'format_note': name
,
2250 'format_index': idx
,
2251 'url': manifest_url
,
2252 'manifest_url': m3u8_url
,
2253 'language': media
.get('LANGUAGE'),
2255 'protocol': entry_protocol
,
2256 'preference': preference
,
2258 'vcodec': 'none' if media_type
== 'AUDIO' else None,
2259 } for idx
in _extract_m3u8_playlist_indices(manifest_url
))
2261 def build_stream_name():
2262 # Despite specification does not mention NAME attribute for
2263 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
2264 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
2265 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
2266 stream_name
= last_stream_inf
.get('NAME')
2269 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
2270 # from corresponding rendition group
2271 stream_group_id
= last_stream_inf
.get('VIDEO')
2272 if not stream_group_id
:
2274 stream_group
= groups
.get(stream_group_id
)
2275 if not stream_group
:
2276 return stream_group_id
2277 rendition
= stream_group
[0]
2278 return rendition
.get('NAME') or stream_group_id
2280 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
2281 # chance to detect video only formats when EXT-X-STREAM-INF tags
2282 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
2283 for line
in m3u8_doc
.splitlines():
2284 if line
.startswith('#EXT-X-MEDIA:'):
2287 for line
in m3u8_doc
.splitlines():
2288 if line
.startswith('#EXT-X-STREAM-INF:'):
2289 last_stream_inf
= parse_m3u8_attributes(line
)
2290 elif line
.startswith('#') or not line
.strip():
2293 tbr
= float_or_none(
2294 last_stream_inf
.get('AVERAGE-BANDWIDTH')
2295 or last_stream_inf
.get('BANDWIDTH'), scale
=1000)
2296 manifest_url
= format_url(line
.strip())
2298 for idx
in _extract_m3u8_playlist_indices(manifest_url
):
2299 format_id
= [m3u8_id
, None, idx
]
2300 # Bandwidth of live streams may differ over time thus making
2301 # format_id unpredictable. So it's better to keep provided
2304 stream_name
= build_stream_name()
2305 format_id
[1] = stream_name
or '%d' % (tbr
or len(formats
))
2307 'format_id': join_nonempty(*format_id
),
2308 'format_index': idx
,
2309 'url': manifest_url
,
2310 'manifest_url': m3u8_url
,
2313 'fps': float_or_none(last_stream_inf
.get('FRAME-RATE')),
2314 'protocol': entry_protocol
,
2315 'preference': preference
,
2318 resolution
= last_stream_inf
.get('RESOLUTION')
2320 mobj
= re
.search(r
'(?P<width>\d+)[xX](?P<height>\d+)', resolution
)
2322 f
['width'] = int(mobj
.group('width'))
2323 f
['height'] = int(mobj
.group('height'))
2324 # Unified Streaming Platform
2326 r
'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f
['url'])
2328 abr
, vbr
= mobj
.groups()
2329 abr
, vbr
= float_or_none(abr
, 1000), float_or_none(vbr
, 1000)
2334 codecs
= parse_codecs(last_stream_inf
.get('CODECS'))
2336 audio_group_id
= last_stream_inf
.get('AUDIO')
2337 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
2338 # references a rendition group MUST have a CODECS attribute.
2339 # However, this is not always respected, for example, [2]
2340 # contains EXT-X-STREAM-INF tag which references AUDIO
2341 # rendition group but does not have CODECS and despite
2342 # referencing an audio group it represents a complete
2343 # (with audio and video) format. So, for such cases we will
2344 # ignore references to rendition groups and treat them
2345 # as complete formats.
2346 if audio_group_id
and codecs
and f
.get('vcodec') != 'none':
2347 audio_group
= groups
.get(audio_group_id
)
2348 if audio_group
and audio_group
[0].get('URI'):
2349 # TODO: update acodec for audio only formats with
2351 f
['acodec'] = 'none'
2352 if not f
.get('ext'):
2353 f
['ext'] = 'm4a' if f
.get('vcodec') == 'none' else 'mp4'
2357 progressive_uri
= last_stream_inf
.get('PROGRESSIVE-URI')
2360 del http_f
['manifest_url']
2362 'format_id': f
['format_id'].replace('hls-', 'http-'),
2364 'url': progressive_uri
,
2366 formats
.append(http_f
)
2368 last_stream_inf
= {}
2369 return formats
, subtitles
2371 def _extract_m3u8_vod_duration(
2372 self
, m3u8_vod_url
, video_id
, note
=None, errnote
=None, data
=None, headers
={}, query={}
):
2374 m3u8_vod
= self
._download
_webpage
(
2375 m3u8_vod_url
, video_id
,
2376 note
='Downloading m3u8 VOD manifest' if note
is None else note
,
2377 errnote
='Failed to download VOD manifest' if errnote
is None else errnote
,
2378 fatal
=False, data
=data
, headers
=headers
, query
=query
)
2380 return self
._parse
_m
3u8_vod
_duration
(m3u8_vod
or '', video_id
)
2382 def _parse_m3u8_vod_duration(self
, m3u8_vod
, video_id
):
2383 if '#EXT-X-PLAYLIST-TYPE:VOD' not in m3u8_vod
:
2387 float(line
[len('#EXTINF:'):].split(',')[0])
2388 for line
in m3u8_vod
.splitlines() if line
.startswith('#EXTINF:'))) or None
2391 def _xpath_ns(path
, namespace
=None):
2395 for c
in path
.split('/'):
2396 if not c
or c
== '.':
2399 out
.append('{%s}%s' % (namespace
, c
))
2400 return '/'.join(out
)
2402 def _extract_smil_formats_and_subtitles(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None, transform_source
=None):
2403 smil
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
, transform_source
=transform_source
)
2409 namespace
= self
._parse
_smil
_namespace
(smil
)
2411 fmts
= self
._parse
_smil
_formats
(
2412 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2413 subs
= self
._parse
_smil
_subtitles
(
2414 smil
, namespace
=namespace
)
2418 def _extract_smil_formats(self
, *args
, **kwargs
):
2419 fmts
, subs
= self
._extract
_smil
_formats
_and
_subtitles
(*args
, **kwargs
)
2421 self
._report
_ignoring
_subs
('SMIL')
2424 def _extract_smil_info(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None):
2425 smil
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
)
2428 return self
._parse
_smil
(smil
, smil_url
, video_id
, f4m_params
=f4m_params
)
2430 def _download_smil(self
, smil_url
, video_id
, fatal
=True, transform_source
=None):
2431 return self
._download
_xml
(
2432 smil_url
, video_id
, 'Downloading SMIL file',
2433 'Unable to download SMIL file', fatal
=fatal
, transform_source
=transform_source
)
2435 def _parse_smil(self
, smil
, smil_url
, video_id
, f4m_params
=None):
2436 namespace
= self
._parse
_smil
_namespace
(smil
)
2438 formats
= self
._parse
_smil
_formats
(
2439 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2440 subtitles
= self
._parse
_smil
_subtitles
(smil
, namespace
=namespace
)
2442 video_id
= os
.path
.splitext(url_basename(smil_url
))[0]
2446 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2447 name
= meta
.attrib
.get('name')
2448 content
= meta
.attrib
.get('content')
2449 if not name
or not content
:
2451 if not title
and name
== 'title':
2453 elif not description
and name
in ('description', 'abstract'):
2454 description
= content
2455 elif not upload_date
and name
== 'date':
2456 upload_date
= unified_strdate(content
)
2459 'id': image
.get('type'),
2460 'url': image
.get('src'),
2461 'width': int_or_none(image
.get('width')),
2462 'height': int_or_none(image
.get('height')),
2463 } for image
in smil
.findall(self
._xpath
_ns
('.//image', namespace
)) if image
.get('src')]
2467 'title': title
or video_id
,
2468 'description': description
,
2469 'upload_date': upload_date
,
2470 'thumbnails': thumbnails
,
2472 'subtitles': subtitles
,
2475 def _parse_smil_namespace(self
, smil
):
2476 return self
._search
_regex
(
2477 r
'(?i)^{([^}]+)?}smil$', smil
.tag
, 'namespace', default
=None)
2479 def _parse_smil_formats(self
, smil
, smil_url
, video_id
, namespace
=None, f4m_params
=None, transform_rtmp_url
=None):
2481 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2482 b
= meta
.get('base') or meta
.get('httpBase')
2494 media
= smil
.findall(self
._xpath
_ns
('.//video', namespace
)) + smil
.findall(self
._xpath
_ns
('.//audio', namespace
))
2495 for medium
in media
:
2496 src
= medium
.get('src')
2497 if not src
or src
in srcs
:
2501 bitrate
= float_or_none(medium
.get('system-bitrate') or medium
.get('systemBitrate'), 1000)
2502 filesize
= int_or_none(medium
.get('size') or medium
.get('fileSize'))
2503 width
= int_or_none(medium
.get('width'))
2504 height
= int_or_none(medium
.get('height'))
2505 proto
= medium
.get('proto')
2506 ext
= medium
.get('ext')
2507 src_ext
= determine_ext(src
)
2508 streamer
= medium
.get('streamer') or base
2510 if proto
== 'rtmp' or streamer
.startswith('rtmp'):
2516 'format_id': 'rtmp-%d' % (rtmp_count
if bitrate
is None else bitrate
),
2518 'filesize': filesize
,
2522 if transform_rtmp_url
:
2523 streamer
, src
= transform_rtmp_url(streamer
, src
)
2524 formats
[-1].update({
2530 src_url
= src
if src
.startswith('http') else compat_urlparse
.urljoin(base
, src
)
2531 src_url
= src_url
.strip()
2533 if proto
== 'm3u8' or src_ext
== 'm3u8':
2534 m3u8_formats
= self
._extract
_m
3u8_formats
(
2535 src_url
, video_id
, ext
or 'mp4', m3u8_id
='hls', fatal
=False)
2536 if len(m3u8_formats
) == 1:
2538 m3u8_formats
[0].update({
2539 'format_id': 'hls-%d' % (m3u8_count
if bitrate
is None else bitrate
),
2544 formats
.extend(m3u8_formats
)
2545 elif src_ext
== 'f4m':
2550 'plugin': 'flowplayer-3.2.0.1',
2552 f4m_url
+= '&' if '?' in f4m_url
else '?'
2553 f4m_url
+= compat_urllib_parse_urlencode(f4m_params
)
2554 formats
.extend(self
._extract
_f
4m
_formats
(f4m_url
, video_id
, f4m_id
='hds', fatal
=False))
2555 elif src_ext
== 'mpd':
2556 formats
.extend(self
._extract
_mpd
_formats
(
2557 src_url
, video_id
, mpd_id
='dash', fatal
=False))
2558 elif re
.search(r
'\.ism/[Mm]anifest', src_url
):
2559 formats
.extend(self
._extract
_ism
_formats
(
2560 src_url
, video_id
, ism_id
='mss', fatal
=False))
2561 elif src_url
.startswith('http') and self
._is
_valid
_url
(src
, video_id
):
2565 'ext': ext
or src_ext
or 'flv',
2566 'format_id': 'http-%d' % (bitrate
or http_count
),
2568 'filesize': filesize
,
2573 for medium
in smil
.findall(self
._xpath
_ns
('.//imagestream', namespace
)):
2574 src
= medium
.get('src')
2575 if not src
or src
in srcs
:
2581 'format_id': 'imagestream-%d' % (imgs_count
),
2583 'ext': mimetype2ext(medium
.get('type')),
2586 'width': int_or_none(medium
.get('width')),
2587 'height': int_or_none(medium
.get('height')),
2588 'format_note': 'SMIL storyboards',
2593 def _parse_smil_subtitles(self
, smil
, namespace
=None, subtitles_lang
='en'):
2596 for num
, textstream
in enumerate(smil
.findall(self
._xpath
_ns
('.//textstream', namespace
))):
2597 src
= textstream
.get('src')
2598 if not src
or src
in urls
:
2601 ext
= textstream
.get('ext') or mimetype2ext(textstream
.get('type')) or determine_ext(src
)
2602 lang
= textstream
.get('systemLanguage') or textstream
.get('systemLanguageName') or textstream
.get('lang') or subtitles_lang
2603 subtitles
.setdefault(lang
, []).append({
2609 def _extract_xspf_playlist(self
, xspf_url
, playlist_id
, fatal
=True):
2610 xspf
= self
._download
_xml
(
2611 xspf_url
, playlist_id
, 'Downloading xpsf playlist',
2612 'Unable to download xspf manifest', fatal
=fatal
)
2615 return self
._parse
_xspf
(
2616 xspf
, playlist_id
, xspf_url
=xspf_url
,
2617 xspf_base_url
=base_url(xspf_url
))
2619 def _parse_xspf(self
, xspf_doc
, playlist_id
, xspf_url
=None, xspf_base_url
=None):
2621 'xspf': 'http://xspf.org/ns/0/',
2622 's1': 'http://static.streamone.nl/player/ns/0',
2626 for track
in xspf_doc
.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP
)):
2628 track
, xpath_with_ns('./xspf:title', NS_MAP
), 'title', default
=playlist_id
)
2629 description
= xpath_text(
2630 track
, xpath_with_ns('./xspf:annotation', NS_MAP
), 'description')
2631 thumbnail
= xpath_text(
2632 track
, xpath_with_ns('./xspf:image', NS_MAP
), 'thumbnail')
2633 duration
= float_or_none(
2634 xpath_text(track
, xpath_with_ns('./xspf:duration', NS_MAP
), 'duration'), 1000)
2637 for location
in track
.findall(xpath_with_ns('./xspf:location', NS_MAP
)):
2638 format_url
= urljoin(xspf_base_url
, location
.text
)
2643 'manifest_url': xspf_url
,
2644 'format_id': location
.get(xpath_with_ns('s1:label', NS_MAP
)),
2645 'width': int_or_none(location
.get(xpath_with_ns('s1:width', NS_MAP
))),
2646 'height': int_or_none(location
.get(xpath_with_ns('s1:height', NS_MAP
))),
2648 self
._sort
_formats
(formats
)
2653 'description': description
,
2654 'thumbnail': thumbnail
,
2655 'duration': duration
,
2660 def _extract_mpd_formats(self
, *args
, **kwargs
):
2661 fmts
, subs
= self
._extract
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2663 self
._report
_ignoring
_subs
('DASH')
2666 def _extract_mpd_formats_and_subtitles(
2667 self
, mpd_url
, video_id
, mpd_id
=None, note
=None, errnote
=None,
2668 fatal
=True, data
=None, headers
={}, query={}
):
2669 res
= self
._download
_xml
_handle
(
2671 note
='Downloading MPD manifest' if note
is None else note
,
2672 errnote
='Failed to download MPD manifest' if errnote
is None else errnote
,
2673 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
2679 mpd_base_url
= base_url(urlh
.geturl())
2681 return self
._parse
_mpd
_formats
_and
_subtitles
(
2682 mpd_doc
, mpd_id
, mpd_base_url
, mpd_url
)
2684 def _parse_mpd_formats(self
, *args
, **kwargs
):
2685 fmts
, subs
= self
._parse
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2687 self
._report
_ignoring
_subs
('DASH')
2690 def _parse_mpd_formats_and_subtitles(
2691 self
, mpd_doc
, mpd_id
=None, mpd_base_url
='', mpd_url
=None):
2693 Parse formats from MPD manifest.
2695 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
2696 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
2697 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
2699 if not self
.get_param('dynamic_mpd', True):
2700 if mpd_doc
.get('type') == 'dynamic':
2703 namespace
= self
._search
_regex
(r
'(?i)^{([^}]+)?}MPD$', mpd_doc
.tag
, 'namespace', default
=None)
2706 return self
._xpath
_ns
(path
, namespace
)
2708 def is_drm_protected(element
):
2709 return element
.find(_add_ns('ContentProtection')) is not None
2711 def extract_multisegment_info(element
, ms_parent_info
):
2712 ms_info
= ms_parent_info
.copy()
2714 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2715 # common attributes and elements. We will only extract relevant
2717 def extract_common(source
):
2718 segment_timeline
= source
.find(_add_ns('SegmentTimeline'))
2719 if segment_timeline
is not None:
2720 s_e
= segment_timeline
.findall(_add_ns('S'))
2722 ms_info
['total_number'] = 0
2725 r
= int(s
.get('r', 0))
2726 ms_info
['total_number'] += 1 + r
2727 ms_info
['s'].append({
2728 't': int(s
.get('t', 0)),
2729 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2730 'd': int(s
.attrib
['d']),
2733 start_number
= source
.get('startNumber')
2735 ms_info
['start_number'] = int(start_number
)
2736 timescale
= source
.get('timescale')
2738 ms_info
['timescale'] = int(timescale
)
2739 segment_duration
= source
.get('duration')
2740 if segment_duration
:
2741 ms_info
['segment_duration'] = float(segment_duration
)
2743 def extract_Initialization(source
):
2744 initialization
= source
.find(_add_ns('Initialization'))
2745 if initialization
is not None:
2746 ms_info
['initialization_url'] = initialization
.attrib
['sourceURL']
2748 segment_list
= element
.find(_add_ns('SegmentList'))
2749 if segment_list
is not None:
2750 extract_common(segment_list
)
2751 extract_Initialization(segment_list
)
2752 segment_urls_e
= segment_list
.findall(_add_ns('SegmentURL'))
2754 ms_info
['segment_urls'] = [segment
.attrib
['media'] for segment
in segment_urls_e
]
2756 segment_template
= element
.find(_add_ns('SegmentTemplate'))
2757 if segment_template
is not None:
2758 extract_common(segment_template
)
2759 media
= segment_template
.get('media')
2761 ms_info
['media'] = media
2762 initialization
= segment_template
.get('initialization')
2764 ms_info
['initialization'] = initialization
2766 extract_Initialization(segment_template
)
2769 mpd_duration
= parse_duration(mpd_doc
.get('mediaPresentationDuration'))
2770 formats
, subtitles
= [], {}
2771 stream_numbers
= collections
.defaultdict(int)
2772 for period
in mpd_doc
.findall(_add_ns('Period')):
2773 period_duration
= parse_duration(period
.get('duration')) or mpd_duration
2774 period_ms_info
= extract_multisegment_info(period
, {
2778 for adaptation_set
in period
.findall(_add_ns('AdaptationSet')):
2779 adaption_set_ms_info
= extract_multisegment_info(adaptation_set
, period_ms_info
)
2780 for representation
in adaptation_set
.findall(_add_ns('Representation')):
2781 representation_attrib
= adaptation_set
.attrib
.copy()
2782 representation_attrib
.update(representation
.attrib
)
2783 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2784 mime_type
= representation_attrib
['mimeType']
2785 content_type
= representation_attrib
.get('contentType', mime_type
.split('/')[0])
2787 codecs
= parse_codecs(representation_attrib
.get('codecs', ''))
2788 if content_type
not in ('video', 'audio', 'text'):
2789 if mime_type
== 'image/jpeg':
2790 content_type
= mime_type
2791 elif codecs
['vcodec'] != 'none':
2792 content_type
= 'video'
2793 elif codecs
['acodec'] != 'none':
2794 content_type
= 'audio'
2795 elif codecs
.get('tcodec', 'none') != 'none':
2796 content_type
= 'text'
2797 elif mimetype2ext(mime_type
) in ('tt', 'dfxp', 'ttml', 'xml', 'json'):
2798 content_type
= 'text'
2800 self
.report_warning('Unknown MIME type %s in DASH manifest' % mime_type
)
2804 for element
in (representation
, adaptation_set
, period
, mpd_doc
):
2805 base_url_e
= element
.find(_add_ns('BaseURL'))
2806 if base_url_e
is not None:
2807 base_url
= base_url_e
.text
+ base_url
2808 if re
.match(r
'^https?://', base_url
):
2810 if mpd_base_url
and base_url
.startswith('/'):
2811 base_url
= compat_urlparse
.urljoin(mpd_base_url
, base_url
)
2812 elif mpd_base_url
and not re
.match(r
'^https?://', base_url
):
2813 if not mpd_base_url
.endswith('/'):
2815 base_url
= mpd_base_url
+ base_url
2816 representation_id
= representation_attrib
.get('id')
2817 lang
= representation_attrib
.get('lang')
2818 url_el
= representation
.find(_add_ns('BaseURL'))
2819 filesize
= int_or_none(url_el
.attrib
.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el
is not None else None)
2820 bandwidth
= int_or_none(representation_attrib
.get('bandwidth'))
2821 if representation_id
is not None:
2822 format_id
= representation_id
2824 format_id
= content_type
2826 format_id
= mpd_id
+ '-' + format_id
2827 if content_type
in ('video', 'audio'):
2829 'format_id': format_id
,
2830 'manifest_url': mpd_url
,
2831 'ext': mimetype2ext(mime_type
),
2832 'width': int_or_none(representation_attrib
.get('width')),
2833 'height': int_or_none(representation_attrib
.get('height')),
2834 'tbr': float_or_none(bandwidth
, 1000),
2835 'asr': int_or_none(representation_attrib
.get('audioSamplingRate')),
2836 'fps': int_or_none(representation_attrib
.get('frameRate')),
2837 'language': lang
if lang
not in ('mul', 'und', 'zxx', 'mis') else None,
2838 'format_note': 'DASH %s' % content_type
,
2839 'filesize': filesize
,
2840 'container': mimetype2ext(mime_type
) + '_dash',
2843 elif content_type
== 'text':
2845 'ext': mimetype2ext(mime_type
),
2846 'manifest_url': mpd_url
,
2847 'filesize': filesize
,
2849 elif content_type
== 'image/jpeg':
2850 # See test case in VikiIE
2851 # https://www.viki.com/videos/1175236v-choosing-spouse-by-lottery-episode-1
2853 'format_id': format_id
,
2855 'manifest_url': mpd_url
,
2856 'format_note': 'DASH storyboards (jpeg)',
2860 if is_drm_protected(adaptation_set
) or is_drm_protected(representation
):
2862 representation_ms_info
= extract_multisegment_info(representation
, adaption_set_ms_info
)
2864 def prepare_template(template_name
, identifiers
):
2865 tmpl
= representation_ms_info
[template_name
]
2866 # First of, % characters outside $...$ templates
2867 # must be escaped by doubling for proper processing
2868 # by % operator string formatting used further (see
2869 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2875 in_template
= not in_template
2876 elif c
== '%' and not in_template
:
2878 # Next, $...$ templates are translated to their
2879 # %(...) counterparts to be used with % operator
2880 if representation_id
is not None:
2881 t
= t
.replace('$RepresentationID$', representation_id
)
2882 t
= re
.sub(r
'\$(%s)\$' % '|'.join(identifiers
), r
'%(\1)d', t
)
2883 t
= re
.sub(r
'\$(%s)%%([^$]+)\$' % '|'.join(identifiers
), r
'%(\1)\2', t
)
2884 t
.replace('$$', '$')
2887 # @initialization is a regular template like @media one
2888 # so it should be handled just the same way (see
2889 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2890 if 'initialization' in representation_ms_info
:
2891 initialization_template
= prepare_template(
2893 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2894 # $Time$ shall not be included for @initialization thus
2895 # only $Bandwidth$ remains
2897 representation_ms_info
['initialization_url'] = initialization_template
% {
2898 'Bandwidth': bandwidth
,
2901 def location_key(location
):
2902 return 'url' if re
.match(r
'^https?://', location
) else 'path'
2904 if 'segment_urls' not in representation_ms_info
and 'media' in representation_ms_info
:
2906 media_template
= prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2907 media_location_key
= location_key(media_template
)
2909 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2910 # can't be used at the same time
2911 if '%(Number' in media_template
and 's' not in representation_ms_info
:
2912 segment_duration
= None
2913 if 'total_number' not in representation_ms_info
and 'segment_duration' in representation_ms_info
:
2914 segment_duration
= float_or_none(representation_ms_info
['segment_duration'], representation_ms_info
['timescale'])
2915 representation_ms_info
['total_number'] = int(math
.ceil(
2916 float_or_none(period_duration
, segment_duration
, default
=0)))
2917 representation_ms_info
['fragments'] = [{
2918 media_location_key
: media_template
% {
2919 'Number': segment_number
,
2920 'Bandwidth': bandwidth
,
2922 'duration': segment_duration
,
2923 } for segment_number
in range(
2924 representation_ms_info
['start_number'],
2925 representation_ms_info
['total_number'] + representation_ms_info
['start_number'])]
2927 # $Number*$ or $Time$ in media template with S list available
2928 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2929 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2930 representation_ms_info
['fragments'] = []
2933 segment_number
= representation_ms_info
['start_number']
2935 def add_segment_url():
2936 segment_url
= media_template
% {
2937 'Time': segment_time
,
2938 'Bandwidth': bandwidth
,
2939 'Number': segment_number
,
2941 representation_ms_info
['fragments'].append({
2942 media_location_key
: segment_url
,
2943 'duration': float_or_none(segment_d
, representation_ms_info
['timescale']),
2946 for num
, s
in enumerate(representation_ms_info
['s']):
2947 segment_time
= s
.get('t') or segment_time
2951 for r
in range(s
.get('r', 0)):
2952 segment_time
+= segment_d
2955 segment_time
+= segment_d
2956 elif 'segment_urls' in representation_ms_info
and 's' in representation_ms_info
:
2958 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2959 # or any YouTube dashsegments video
2962 timescale
= representation_ms_info
['timescale']
2963 for s
in representation_ms_info
['s']:
2964 duration
= float_or_none(s
['d'], timescale
)
2965 for r
in range(s
.get('r', 0) + 1):
2966 segment_uri
= representation_ms_info
['segment_urls'][segment_index
]
2968 location_key(segment_uri
): segment_uri
,
2969 'duration': duration
,
2972 representation_ms_info
['fragments'] = fragments
2973 elif 'segment_urls' in representation_ms_info
:
2974 # Segment URLs with no SegmentTimeline
2975 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2976 # https://github.com/ytdl-org/youtube-dl/pull/14844
2978 segment_duration
= float_or_none(
2979 representation_ms_info
['segment_duration'],
2980 representation_ms_info
['timescale']) if 'segment_duration' in representation_ms_info
else None
2981 for segment_url
in representation_ms_info
['segment_urls']:
2983 location_key(segment_url
): segment_url
,
2985 if segment_duration
:
2986 fragment
['duration'] = segment_duration
2987 fragments
.append(fragment
)
2988 representation_ms_info
['fragments'] = fragments
2989 # If there is a fragments key available then we correctly recognized fragmented media.
2990 # Otherwise we will assume unfragmented media with direct access. Technically, such
2991 # assumption is not necessarily correct since we may simply have no support for
2992 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2993 if 'fragments' in representation_ms_info
:
2995 # NB: mpd_url may be empty when MPD manifest is parsed from a string
2996 'url': mpd_url
or base_url
,
2997 'fragment_base_url': base_url
,
2999 'protocol': 'http_dash_segments' if mime_type
!= 'image/jpeg' else 'mhtml',
3001 if 'initialization_url' in representation_ms_info
:
3002 initialization_url
= representation_ms_info
['initialization_url']
3003 if not f
.get('url'):
3004 f
['url'] = initialization_url
3005 f
['fragments'].append({location_key(initialization_url): initialization_url}
)
3006 f
['fragments'].extend(representation_ms_info
['fragments'])
3007 if not period_duration
:
3008 period_duration
= try_get(
3009 representation_ms_info
,
3010 lambda r
: sum(frag
['duration'] for frag
in r
['fragments']), float)
3012 # Assuming direct URL to unfragmented media.
3014 if content_type
in ('video', 'audio', 'image/jpeg'):
3015 f
['manifest_stream_number'] = stream_numbers
[f
['url']]
3016 stream_numbers
[f
['url']] += 1
3018 elif content_type
== 'text':
3019 subtitles
.setdefault(lang
or 'und', []).append(f
)
3021 return formats
, subtitles
3023 def _extract_ism_formats(self
, *args
, **kwargs
):
3024 fmts
, subs
= self
._extract
_ism
_formats
_and
_subtitles
(*args
, **kwargs
)
3026 self
._report
_ignoring
_subs
('ISM')
3029 def _extract_ism_formats_and_subtitles(self
, ism_url
, video_id
, ism_id
=None, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
):
3030 res
= self
._download
_xml
_handle
(
3032 note
='Downloading ISM manifest' if note
is None else note
,
3033 errnote
='Failed to download ISM manifest' if errnote
is None else errnote
,
3034 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
3041 return self
._parse
_ism
_formats
_and
_subtitles
(ism_doc
, urlh
.geturl(), ism_id
)
3043 def _parse_ism_formats_and_subtitles(self
, ism_doc
, ism_url
, ism_id
=None):
3045 Parse formats from ISM manifest.
3047 1. [MS-SSTR]: Smooth Streaming Protocol,
3048 https://msdn.microsoft.com/en-us/library/ff469518.aspx
3050 if ism_doc
.get('IsLive') == 'TRUE':
3053 duration
= int(ism_doc
.attrib
['Duration'])
3054 timescale
= int_or_none(ism_doc
.get('TimeScale')) or 10000000
3058 for stream
in ism_doc
.findall('StreamIndex'):
3059 stream_type
= stream
.get('Type')
3060 if stream_type
not in ('video', 'audio', 'text'):
3062 url_pattern
= stream
.attrib
['Url']
3063 stream_timescale
= int_or_none(stream
.get('TimeScale')) or timescale
3064 stream_name
= stream
.get('Name')
3065 stream_language
= stream
.get('Language', 'und')
3066 for track
in stream
.findall('QualityLevel'):
3067 fourcc
= track
.get('FourCC') or ('AACL' if track
.get('AudioTag') == '255' else None)
3068 # TODO: add support for WVC1 and WMAP
3069 if fourcc
not in ('H264', 'AVC1', 'AACL', 'TTML'):
3070 self
.report_warning('%s is not a supported codec' % fourcc
)
3072 tbr
= int(track
.attrib
['Bitrate']) // 1000
3073 # [1] does not mention Width and Height attributes. However,
3074 # they're often present while MaxWidth and MaxHeight are
3075 # missing, so should be used as fallbacks
3076 width
= int_or_none(track
.get('MaxWidth') or track
.get('Width'))
3077 height
= int_or_none(track
.get('MaxHeight') or track
.get('Height'))
3078 sampling_rate
= int_or_none(track
.get('SamplingRate'))
3080 track_url_pattern
= re
.sub(r
'{[Bb]itrate}', track
.attrib
['Bitrate'], url_pattern
)
3081 track_url_pattern
= compat_urlparse
.urljoin(ism_url
, track_url_pattern
)
3087 stream_fragments
= stream
.findall('c')
3088 for stream_fragment_index
, stream_fragment
in enumerate(stream_fragments
):
3089 fragment_ctx
['time'] = int_or_none(stream_fragment
.get('t')) or fragment_ctx
['time']
3090 fragment_repeat
= int_or_none(stream_fragment
.get('r')) or 1
3091 fragment_ctx
['duration'] = int_or_none(stream_fragment
.get('d'))
3092 if not fragment_ctx
['duration']:
3094 next_fragment_time
= int(stream_fragment
[stream_fragment_index
+ 1].attrib
['t'])
3096 next_fragment_time
= duration
3097 fragment_ctx
['duration'] = (next_fragment_time
- fragment_ctx
['time']) / fragment_repeat
3098 for _
in range(fragment_repeat
):
3100 'url': re
.sub(r
'{start[ _]time}', compat_str(fragment_ctx
['time']), track_url_pattern
),
3101 'duration': fragment_ctx
['duration'] / stream_timescale
,
3103 fragment_ctx
['time'] += fragment_ctx
['duration']
3105 if stream_type
== 'text':
3106 subtitles
.setdefault(stream_language
, []).append({
3110 'manifest_url': ism_url
,
3111 'fragments': fragments
,
3112 '_download_params': {
3113 'stream_type': stream_type
,
3114 'duration': duration
,
3115 'timescale': stream_timescale
,
3117 'language': stream_language
,
3118 'codec_private_data': track
.get('CodecPrivateData'),
3121 elif stream_type
in ('video', 'audio'):
3123 'format_id': join_nonempty(ism_id
, stream_name
, tbr
),
3125 'manifest_url': ism_url
,
3126 'ext': 'ismv' if stream_type
== 'video' else 'isma',
3130 'asr': sampling_rate
,
3131 'vcodec': 'none' if stream_type
== 'audio' else fourcc
,
3132 'acodec': 'none' if stream_type
== 'video' else fourcc
,
3134 'fragments': fragments
,
3135 'has_drm': ism_doc
.find('Protection') is not None,
3136 '_download_params': {
3137 'stream_type': stream_type
,
3138 'duration': duration
,
3139 'timescale': stream_timescale
,
3140 'width': width
or 0,
3141 'height': height
or 0,
3143 'language': stream_language
,
3144 'codec_private_data': track
.get('CodecPrivateData'),
3145 'sampling_rate': sampling_rate
,
3146 'channels': int_or_none(track
.get('Channels', 2)),
3147 'bits_per_sample': int_or_none(track
.get('BitsPerSample', 16)),
3148 'nal_unit_length_field': int_or_none(track
.get('NALUnitLengthField', 4)),
3151 return formats
, subtitles
3153 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):
3154 def absolute_url(item_url
):
3155 return urljoin(base_url
, item_url
)
3157 def parse_content_type(content_type
):
3158 if not content_type
:
3160 ctr
= re
.search(r
'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type
)
3162 mimetype
, codecs
= ctr
.groups()
3163 f
= parse_codecs(codecs
)
3164 f
['ext'] = mimetype2ext(mimetype
)
3168 def _media_formats(src
, cur_media_type
, type_info
={}):
3169 full_url
= absolute_url(src
)
3170 ext
= type_info
.get('ext') or determine_ext(full_url
)
3172 is_plain_url
= False
3173 formats
= self
._extract
_m
3u8_formats
(
3174 full_url
, video_id
, ext
='mp4',
3175 entry_protocol
=m3u8_entry_protocol
, m3u8_id
=m3u8_id
,
3176 preference
=preference
, quality
=quality
, fatal
=False)
3178 is_plain_url
= False
3179 formats
= self
._extract
_mpd
_formats
(
3180 full_url
, video_id
, mpd_id
=mpd_id
, fatal
=False)
3185 'vcodec': 'none' if cur_media_type
== 'audio' else None,
3187 return is_plain_url
, formats
3190 # amp-video and amp-audio are very similar to their HTML5 counterparts
3191 # so we wll include them right here (see
3192 # https://www.ampproject.org/docs/reference/components/amp-video)
3193 # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/
3194 _MEDIA_TAG_NAME_RE
= r
'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)'
3195 media_tags
= [(media_tag
, media_tag_name
, media_type
, '')
3196 for media_tag
, media_tag_name
, media_type
3197 in re
.findall(r
'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE
, webpage
)]
3198 media_tags
.extend(re
.findall(
3199 # We only allow video|audio followed by a whitespace or '>'.
3200 # Allowing more characters may end up in significant slow down (see
3201 # https://github.com/ytdl-org/youtube-dl/issues/11979, example URL:
3202 # http://www.porntrex.com/maps/videositemap.xml).
3203 r
'(?s)(<(?P<tag>%s)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>' % _MEDIA_TAG_NAME_RE
, webpage
))
3204 for media_tag
, _
, media_type
, media_content
in media_tags
:
3209 media_attributes
= extract_attributes(media_tag
)
3210 src
= strip_or_none(media_attributes
.get('src'))
3212 _
, formats
= _media_formats(src
, media_type
)
3213 media_info
['formats'].extend(formats
)
3214 media_info
['thumbnail'] = absolute_url(media_attributes
.get('poster'))
3216 for source_tag
in re
.findall(r
'<source[^>]+>', media_content
):
3217 s_attr
= extract_attributes(source_tag
)
3218 # data-video-src and data-src are non standard but seen
3219 # several times in the wild
3220 src
= strip_or_none(dict_get(s_attr
, ('src', 'data-video-src', 'data-src')))
3223 f
= parse_content_type(s_attr
.get('type'))
3224 is_plain_url
, formats
= _media_formats(src
, media_type
, f
)
3226 # width, height, res, label and title attributes are
3227 # all not standard but seen several times in the wild
3230 for lbl
in ('label', 'title')
3231 if str_or_none(s_attr
.get(lbl
))
3233 width
= int_or_none(s_attr
.get('width'))
3234 height
= (int_or_none(s_attr
.get('height'))
3235 or int_or_none(s_attr
.get('res')))
3236 if not width
or not height
:
3238 resolution
= parse_resolution(lbl
)
3241 width
= width
or resolution
.get('width')
3242 height
= height
or resolution
.get('height')
3244 tbr
= parse_bitrate(lbl
)
3253 'format_id': s_attr
.get('label') or s_attr
.get('title'),
3255 f
.update(formats
[0])
3256 media_info
['formats'].append(f
)
3258 media_info
['formats'].extend(formats
)
3259 for track_tag
in re
.findall(r
'<track[^>]+>', media_content
):
3260 track_attributes
= extract_attributes(track_tag
)
3261 kind
= track_attributes
.get('kind')
3262 if not kind
or kind
in ('subtitles', 'captions'):
3263 src
= strip_or_none(track_attributes
.get('src'))
3266 lang
= track_attributes
.get('srclang') or track_attributes
.get('lang') or track_attributes
.get('label')
3267 media_info
['subtitles'].setdefault(lang
, []).append({
3268 'url': absolute_url(src
),
3270 for f
in media_info
['formats']:
3271 f
.setdefault('http_headers', {})['Referer'] = base_url
3272 if media_info
['formats'] or media_info
['subtitles']:
3273 entries
.append(media_info
)
3276 def _extract_akamai_formats(self
, *args
, **kwargs
):
3277 fmts
, subs
= self
._extract
_akamai
_formats
_and
_subtitles
(*args
, **kwargs
)
3279 self
._report
_ignoring
_subs
('akamai')
3282 def _extract_akamai_formats_and_subtitles(self
, manifest_url
, video_id
, hosts
={}):
3283 signed
= 'hdnea=' in manifest_url
3285 # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
3286 manifest_url
= re
.sub(
3287 r
'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
3288 '', manifest_url
).strip('?')
3293 hdcore_sign
= 'hdcore=3.7.0'
3294 f4m_url
= re
.sub(r
'(https?://[^/]+)/i/', r
'\1/z/', manifest_url
).replace('/master.m3u8', '/manifest.f4m')
3295 hds_host
= hosts
.get('hds')
3297 f4m_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hds_host
, f4m_url
)
3298 if 'hdcore=' not in f4m_url
:
3299 f4m_url
+= ('&' if '?' in f4m_url
else '?') + hdcore_sign
3300 f4m_formats
= self
._extract
_f
4m
_formats
(
3301 f4m_url
, video_id
, f4m_id
='hds', fatal
=False)
3302 for entry
in f4m_formats
:
3303 entry
.update({'extra_param_to_segment_url': hdcore_sign}
)
3304 formats
.extend(f4m_formats
)
3306 m3u8_url
= re
.sub(r
'(https?://[^/]+)/z/', r
'\1/i/', manifest_url
).replace('/manifest.f4m', '/master.m3u8')
3307 hls_host
= hosts
.get('hls')
3309 m3u8_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hls_host
, m3u8_url
)
3310 m3u8_formats
, m3u8_subtitles
= self
._extract
_m
3u8_formats
_and
_subtitles
(
3311 m3u8_url
, video_id
, 'mp4', 'm3u8_native',
3312 m3u8_id
='hls', fatal
=False)
3313 formats
.extend(m3u8_formats
)
3314 subtitles
= self
._merge
_subtitles
(subtitles
, m3u8_subtitles
)
3316 http_host
= hosts
.get('http')
3317 if http_host
and m3u8_formats
and not signed
:
3318 REPL_REGEX
= r
'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
3319 qualities
= re
.match(REPL_REGEX
, m3u8_url
).group(2).split(',')
3320 qualities_length
= len(qualities
)
3321 if len(m3u8_formats
) in (qualities_length
, qualities_length
+ 1):
3323 for f
in m3u8_formats
:
3324 if f
['vcodec'] != 'none':
3325 for protocol
in ('http', 'https'):
3327 del http_f
['manifest_url']
3329 REPL_REGEX
, protocol
+ fr
'://{http_host}/\g<1>{qualities[i]}\3', f
['url'])
3331 'format_id': http_f
['format_id'].replace('hls-', protocol
+ '-'),
3333 'protocol': protocol
,
3335 formats
.append(http_f
)
3338 return formats
, subtitles
3340 def _extract_wowza_formats(self
, url
, video_id
, m3u8_entry_protocol
='m3u8_native', skip_protocols
=[]):
3341 query
= compat_urlparse
.urlparse(url
).query
3342 url
= re
.sub(r
'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url
)
3344 r
'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url
)
3345 url_base
= mobj
.group('url')
3346 http_base_url
= '%s%s:%s' % ('http', mobj
.group('s') or '', url_base
)
3349 def manifest_url(manifest
):
3350 m_url
= f
'{http_base_url}/{manifest}'
3352 m_url
+= '?%s' % query
3355 if 'm3u8' not in skip_protocols
:
3356 formats
.extend(self
._extract
_m
3u8_formats
(
3357 manifest_url('playlist.m3u8'), video_id
, 'mp4',
3358 m3u8_entry_protocol
, m3u8_id
='hls', fatal
=False))
3359 if 'f4m' not in skip_protocols
:
3360 formats
.extend(self
._extract
_f
4m
_formats
(
3361 manifest_url('manifest.f4m'),
3362 video_id
, f4m_id
='hds', fatal
=False))
3363 if 'dash' not in skip_protocols
:
3364 formats
.extend(self
._extract
_mpd
_formats
(
3365 manifest_url('manifest.mpd'),
3366 video_id
, mpd_id
='dash', fatal
=False))
3367 if re
.search(r
'(?:/smil:|\.smil)', url_base
):
3368 if 'smil' not in skip_protocols
:
3369 rtmp_formats
= self
._extract
_smil
_formats
(
3370 manifest_url('jwplayer.smil'),
3371 video_id
, fatal
=False)
3372 for rtmp_format
in rtmp_formats
:
3373 rtsp_format
= rtmp_format
.copy()
3374 rtsp_format
['url'] = '%s/%s' % (rtmp_format
['url'], rtmp_format
['play_path'])
3375 del rtsp_format
['play_path']
3376 del rtsp_format
['ext']
3377 rtsp_format
.update({
3378 'url': rtsp_format
['url'].replace('rtmp://', 'rtsp://'),
3379 'format_id': rtmp_format
['format_id'].replace('rtmp', 'rtsp'),
3382 formats
.extend([rtmp_format
, rtsp_format
])
3384 for protocol
in ('rtmp', 'rtsp'):
3385 if protocol
not in skip_protocols
:
3387 'url': f
'{protocol}:{url_base}',
3388 'format_id': protocol
,
3389 'protocol': protocol
,
3393 def _find_jwplayer_data(self
, webpage
, video_id
=None, transform_source
=js_to_json
):
3395 r
'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P
=quote
)\
)(?
!</script
>).*?\
.setup\s
*\
((?P
<options
>[^
)]+)\
)',
3399 jwplayer_data = self._parse_json(mobj.group('options
'),
3401 transform_source=transform_source)
3402 except ExtractorError:
3405 if isinstance(jwplayer_data, dict):
3406 return jwplayer_data
3408 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
3409 jwplayer_data = self._find_jwplayer_data(
3410 webpage, video_id, transform_source=js_to_json)
3411 return self._parse_jwplayer_data(
3412 jwplayer_data, video_id, *args, **kwargs)
3414 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
3415 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3416 # JWPlayer backward compatibility: flattened playlists
3417 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
3418 if 'playlist
' not in jwplayer_data:
3419 jwplayer_data = {'playlist': [jwplayer_data]}
3423 # JWPlayer backward compatibility: single playlist item
3424 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
3425 if not isinstance(jwplayer_data['playlist
'], list):
3426 jwplayer_data['playlist
'] = [jwplayer_data['playlist
']]
3428 for video_data in jwplayer_data['playlist
']:
3429 # JWPlayer backward compatibility: flattened sources
3430 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
3431 if 'sources
' not in video_data:
3432 video_data['sources
'] = [video_data]
3434 this_video_id = video_id or video_data['mediaid
']
3436 formats = self._parse_jwplayer_formats(
3437 video_data['sources
'], video_id=this_video_id, m3u8_id=m3u8_id,
3438 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
3441 tracks = video_data.get('tracks
')
3442 if tracks and isinstance(tracks, list):
3443 for track in tracks:
3444 if not isinstance(track, dict):
3446 track_kind = track.get('kind
')
3447 if not track_kind or not isinstance(track_kind, compat_str):
3449 if track_kind.lower() not in ('captions
', 'subtitles
'):
3451 track_url = urljoin(base_url, track.get('file'))
3454 subtitles.setdefault(track.get('label
') or 'en
', []).append({
3455 'url
': self._proto_relative_url(track_url)
3459 'id': this_video_id,
3460 'title
': unescapeHTML(video_data['title
'] if require_title else video_data.get('title
')),
3461 'description
': clean_html(video_data.get('description
')),
3462 'thumbnail
': urljoin(base_url, self._proto_relative_url(video_data.get('image
'))),
3463 'timestamp
': int_or_none(video_data.get('pubdate
')),
3464 'duration
': float_or_none(jwplayer_data.get('duration
') or video_data.get('duration
')),
3465 'subtitles
': subtitles,
3467 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
3468 if len(formats) == 1 and re.search(r'^
(?
:http|
//).*(?
:youtube\
.com|youtu\
.be
)/.+', formats[0]['url
']):
3470 '_type
': 'url_transparent
',
3471 'url
': formats[0]['url
'],
3474 self._sort_formats(formats)
3475 entry['formats
'] = formats
3476 entries.append(entry)
3477 if len(entries) == 1:
3480 return self.playlist_result(entries)
3482 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
3483 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3486 for source in jwplayer_sources_data:
3487 if not isinstance(source, dict):
3489 source_url = urljoin(
3490 base_url, self._proto_relative_url(source.get('file')))
3491 if not source_url or source_url in urls:
3493 urls.append(source_url)
3494 source_type = source.get('type') or ''
3495 ext = mimetype2ext(source_type) or determine_ext(source_url)
3496 if source_type == 'hls
' or ext == 'm3u8
':
3497 formats.extend(self._extract_m3u8_formats(
3498 source_url, video_id, 'mp4
', entry_protocol='m3u8_native
',
3499 m3u8_id=m3u8_id, fatal=False))
3500 elif source_type == 'dash
' or ext == 'mpd
':
3501 formats.extend(self._extract_mpd_formats(
3502 source_url, video_id, mpd_id=mpd_id, fatal=False))
3504 formats.extend(self._extract_smil_formats(
3505 source_url, video_id, fatal=False))
3506 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3507 elif source_type.startswith('audio
') or ext in (
3508 'oga
', 'aac
', 'mp3
', 'mpeg
', 'vorbis
'):
3515 height = int_or_none(source.get('height
'))
3517 # Often no height is provided but there is a label in
3518 # format like "1080p", "720p SD", or 1080.
3519 height = int_or_none(self._search_regex(
3520 r'^
(\d{3,4}
)[pP
]?
(?
:\b|$
)', compat_str(source.get('label
') or ''),
3521 'height
', default=None))
3524 'width
': int_or_none(source.get('width
')),
3526 'tbr
': int_or_none(source.get('bitrate
')),
3529 if source_url.startswith('rtmp
'):
3530 a_format['ext
'] = 'flv
'
3531 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3532 # of jwplayer.flash.swf
3533 rtmp_url_parts = re.split(
3534 r'((?
:mp4|mp3|flv
):)', source_url, 1)
3535 if len(rtmp_url_parts) == 3:
3536 rtmp_url, prefix, play_path = rtmp_url_parts
3539 'play_path
': prefix + play_path,
3542 a_format.update(rtmp_params)
3543 formats.append(a_format)
3546 def _live_title(self, name):
3547 self._downloader.deprecation_warning('yt_dlp
.InfoExtractor
._live
_title
is deprecated
and does
not work
as expected
')
3550 def _int(self, v, name, fatal=False, **kwargs):
3551 res = int_or_none(v, **kwargs)
3553 msg = f'Failed to extract {name}
: Could
not parse value {v!r}
'
3555 raise ExtractorError(msg)
3557 self.report_warning(msg)
3560 def _float(self, v, name, fatal=False, **kwargs):
3561 res = float_or_none(v, **kwargs)
3563 msg = f'Failed to extract {name}
: Could
not parse value {v!r}
'
3565 raise ExtractorError(msg)
3567 self.report_warning(msg)
3570 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
3571 path='/', secure=False, discard=False, rest={}, **kwargs):
3572 cookie = compat_cookiejar_Cookie(
3573 0, name, value, port, port is not None, domain, True,
3574 domain.startswith('.'), path, True, secure, expire_time,
3575 discard, None, None, rest)
3576 self._downloader.cookiejar.set_cookie(cookie)
3578 def _get_cookies(self, url):
3579 """ Return a compat_cookies_SimpleCookie with the cookies for the url """
3580 req = sanitized_Request(url)
3581 self._downloader.cookiejar.add_cookie_header(req)
3582 return compat_cookies_SimpleCookie(req.get_header('Cookie
'))
3584 def _apply_first_set_cookie_header(self, url_handle, cookie):
3586 Apply first Set-Cookie header instead of the last. Experimental.
3588 Some sites (e.g. [1-3]) may serve two cookies under the same name
3589 in Set-Cookie header and expect the first (old) one to be set rather
3590 than second (new). However, as of RFC6265 the newer one cookie
3591 should be set into cookie store what actually happens.
3592 We will workaround this issue by resetting the cookie to
3593 the first one manually.
3594 1. https://new.vk.com/
3595 2. https://github.com/ytdl-org/youtube-dl/issues/9841#issuecomment-227871201
3596 3. https://learning.oreilly.com/
3598 for header, cookies in url_handle.headers.items():
3599 if header.lower() != 'set-cookie
':
3601 cookies = cookies.encode('iso
-8859-1').decode('utf
-8')
3602 cookie_value = re.search(
3603 r'%s=(.+?
);.*?
\b[Dd
]omain
=(.+?
)(?
:[,;]|$
)' % cookie, cookies)
3605 value, domain = cookie_value.groups()
3606 self._set_cookie(domain, cookie, value)
3609 def get_testcases(self, include_onlymatching=False):
3610 t = getattr(self, '_TEST
', None)
3612 assert not hasattr(self, '_TESTS
'), \
3613 '%s has _TEST
and _TESTS
' % type(self).__name__
3616 tests = getattr(self, '_TESTS
', [])
3618 if not include_onlymatching and t.get('only_matching
', False):
3620 t['name
'] = type(self).__name__[:-len('IE
')]
3623 def is_suitable(self, age_limit):
3624 """ Test whether the extractor is generally suitable for the given
3625 age limit (i.e. pornographic sites are not, all others usually are) """
3627 any_restricted = False
3628 for tc in self.get_testcases(include_onlymatching=False):
3629 if tc.get('playlist
', []):
3630 tc = tc['playlist
'][0]
3631 is_restricted = age_restricted(
3632 tc.get('info_dict
', {}).get('age_limit
'), age_limit)
3633 if not is_restricted:
3635 any_restricted = any_restricted or is_restricted
3636 return not any_restricted
3638 def extract_subtitles(self, *args, **kwargs):
3639 if (self.get_param('writesubtitles
', False)
3640 or self.get_param('listsubtitles
')):
3641 return self._get_subtitles(*args, **kwargs)
3644 def _get_subtitles(self, *args, **kwargs):
3645 raise NotImplementedError('This method must be implemented by subclasses
')
3647 def extract_comments(self, *args, **kwargs):
3648 if not self.get_param('getcomments
'):
3650 generator = self._get_comments(*args, **kwargs)
3657 comments.append(next(generator))
3658 except StopIteration:
3660 except KeyboardInterrupt:
3661 self.to_screen('Interrupted by user
')
3662 except Exception as e:
3663 if self.get_param('ignoreerrors
') is not True:
3665 self._downloader.report_error(e)
3666 comment_count = len(comments)
3667 self.to_screen(f'Extracted {comment_count} comments
')
3669 'comments
': comments,
3670 'comment_count
': None if interrupted else comment_count
3674 def _get_comments(self, *args, **kwargs):
3675 raise NotImplementedError('This method must be implemented by subclasses
')
3678 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
3679 """ Merge subtitle items for one language. Items with duplicated URLs/data
3680 will be dropped. """
3681 list1_data = {(item.get('url'), item.get('data')) for item in subtitle_list1}
3682 ret = list(subtitle_list1)
3683 ret.extend(item for item in subtitle_list2 if (item.get('url
'), item.get('data
')) not in list1_data)
3687 def _merge_subtitles(cls, *dicts, target=None):
3688 """ Merge subtitle dictionaries, language by language. """
3692 for lang, subs in d.items():
3693 target[lang] = cls._merge_subtitle_items(target.get(lang, []), subs)
3696 def extract_automatic_captions(self, *args, **kwargs):
3697 if (self.get_param('writeautomaticsub
', False)
3698 or self.get_param('listsubtitles
')):
3699 return self._get_automatic_captions(*args, **kwargs)
3702 def _get_automatic_captions(self, *args, **kwargs):
3703 raise NotImplementedError('This method must be implemented by subclasses
')
3705 def mark_watched(self, *args, **kwargs):
3706 if not self.get_param('mark_watched
', False):
3708 if (self.supports_login() and self._get_login_info()[0] is not None
3709 or self.get_param('cookiefile
') or self.get_param('cookiesfrombrowser
')):
3710 self._mark_watched(*args, **kwargs)
3712 def _mark_watched(self, *args, **kwargs):
3713 raise NotImplementedError('This method must be implemented by subclasses
')
3715 def geo_verification_headers(self):
3717 geo_verification_proxy = self.get_param('geo_verification_proxy
')
3718 if geo_verification_proxy:
3719 headers['Ytdl
-request
-proxy
'] = geo_verification_proxy
3722 def _generic_id(self, url):
3723 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
3725 def _generic_title(self, url):
3726 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
3729 def _availability(is_private=None, needs_premium=None, needs_subscription=None, needs_auth=None, is_unlisted=None):
3730 all_known = all(map(
3731 lambda x: x is not None,
3732 (is_private, needs_premium, needs_subscription, needs_auth, is_unlisted)))
3734 'private
' if is_private
3735 else 'premium_only
' if needs_premium
3736 else 'subscriber_only
' if needs_subscription
3737 else 'needs_auth
' if needs_auth
3738 else 'unlisted
' if is_unlisted
3739 else 'public
' if all_known
3742 def _configuration_arg(self, key, default=NO_DEFAULT, *, ie_key=None, casesense=False):
3744 @returns A list of values for the extractor argument given by "key"
3745 or "default" if no such key is present
3746 @param default The default value to return when the key is not present (default: [])
3747 @param casesense When false, the values are converted to lower case
3750 self._downloader.params, ('extractor_args
', (ie_key or self.ie_key()).lower(), key))
3752 return [] if default is NO_DEFAULT else default
3753 return list(val) if casesense else [x.lower() for x in val]
3755 def _yes_playlist(self, playlist_id, video_id, smuggled_data=None, *, playlist_label='playlist
', video_label='video
'):
3756 if not playlist_id or not video_id:
3759 no_playlist = (smuggled_data or {}).get('force_noplaylist
')
3760 if no_playlist is not None:
3761 return not no_playlist
3763 video_id = '' if video_id is True else f' {video_id}
'
3764 playlist_id = '' if playlist_id is True else f' {playlist_id}
'
3765 if self.get_param('noplaylist
'):
3766 self.to_screen(f'Downloading just the {video_label}{video_id} because of
--no
-playlist
')
3768 self.to_screen(f'Downloading {playlist_label}{playlist_id}
- add
--no
-playlist to download just the {video_label}{video_id}
')
3772 class SearchInfoExtractor(InfoExtractor):
3774 Base class for paged search queries extractors.
3775 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
3776 Instances should define _SEARCH_KEY and optionally _MAX_RESULTS
3779 _MAX_RESULTS = float('inf
')
3782 def _make_valid_url(cls):
3783 return r'%s(?P
<prefix
>|
[1-9][0-9]*|all
):(?P
<query
>[\s\S
]+)' % cls._SEARCH_KEY
3785 def _real_extract(self, query):
3786 prefix, query = self._match_valid_url(query).group('prefix
', 'query
')
3788 return self._get_n_results(query, 1)
3789 elif prefix == 'all
':
3790 return self._get_n_results(query, self._MAX_RESULTS)
3794 raise ExtractorError(f'invalid download number {n}
for query
"{query}"')
3795 elif n > self._MAX_RESULTS:
3796 self.report_warning('%s returns
max %i results (you requested
%i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3797 n = self._MAX_RESULTS
3798 return self._get_n_results(query, n)
3800 def _get_n_results(self, query, n):
3801 """Get a specified number of results for a query.
3802 Either this function or _search_results must be overridden by subclasses """
3803 return self.playlist_result(
3804 itertools.islice(self._search_results(query), 0, None if n == float('inf
') else n),
3807 def _search_results(self, query):
3808 """Returns an iterator of search results"""
3809 raise NotImplementedError('This method must be implemented by subclasses
')
3812 def SEARCH_KEY(self):
3813 return self._SEARCH_KEY