2 from __future__
import unicode_literals
17 from ..compat
import (
18 compat_cookiejar_Cookie
,
19 compat_cookies_SimpleCookie
,
21 compat_etree_fromstring
,
28 compat_urllib_parse_unquote
,
29 compat_urllib_parse_urlencode
,
30 compat_urllib_request
,
32 compat_xml_parse_error
,
34 from ..downloader
import FileDownloader
35 from ..downloader
.f4m
import (
37 remove_encrypted_media
,
69 parse_m3u8_attributes
,
95 class InfoExtractor(object):
96 """Information Extractor class.
98 Information extractors are the classes that, given a URL, extract
99 information about the video (or videos) the URL refers to. This
100 information includes the real video URL, the video title, author and
101 others. The information is stored in a dictionary which is then
102 passed to the YoutubeDL. The YoutubeDL processes this
103 information possibly downloading the video to the file system, among
104 other possible outcomes.
106 The type field determines the type of the result.
107 By far the most common value (and the default if _type is missing) is
108 "video", which indicates a single video.
110 For a video, the dictionaries must include the following fields:
112 id: Video identifier.
113 title: Video title, unescaped.
115 Additionally, it must contain either a formats entry or a url one:
117 formats: A list of dictionaries for each format available, ordered
118 from worst to best quality.
121 * url The mandatory URL representing the media:
122 for plain file media - HTTP URL of this file,
124 for HLS - URL of the M3U8 media playlist,
125 for HDS - URL of the F4M manifest,
127 - HTTP URL to plain file media (in case of
129 - URL of the MPD manifest or base URL
130 representing the media if MPD manifest
131 is parsed from a string (in case of
133 for MSS - URL of the ISM manifest.
135 The URL of the manifest file in case of
137 for HLS - URL of the M3U8 master playlist,
138 for HDS - URL of the F4M manifest,
139 for DASH - URL of the MPD manifest,
140 for MSS - URL of the ISM manifest.
141 * ext Will be calculated from URL if missing
142 * format A human-readable description of the format
143 ("mp4 container with h264/opus").
144 Calculated from the format_id, width, height.
145 and format_note fields if missing.
146 * format_id A short description of the format
147 ("mp4_h264_opus" or "19").
148 Technically optional, but strongly recommended.
149 * format_note Additional info about the format
150 ("3D" or "DASH video")
151 * width Width of the video, if known
152 * height Height of the video, if known
153 * resolution Textual description of width and height
154 * dynamic_range The dynamic range of the video. One of:
155 "SDR" (None), "HDR10", "HDR10+, "HDR12", "HLG, "DV"
156 * tbr Average bitrate of audio and video in KBit/s
157 * abr Average audio bitrate in KBit/s
158 * acodec Name of the audio codec in use
159 * asr Audio sampling rate in Hertz
160 * vbr Average video bitrate in KBit/s
162 * vcodec Name of the video codec in use
163 * container Name of the container format
164 * filesize The number of bytes, if known in advance
165 * filesize_approx An estimate for the number of bytes
166 * player_url SWF Player URL (used for rtmpdump).
167 * protocol The protocol that will be used for the actual
168 download, lower-case. One of "http", "https" or
169 one of the protocols defined in downloader.PROTOCOL_MAP
171 Base URL for fragments. Each fragment's path
172 value (if present) will be relative to
174 * fragments A list of fragments of a fragmented media.
175 Each fragment entry must contain either an url
176 or a path. If an url is present it should be
177 considered by a client. Otherwise both path and
178 fragment_base_url must be present. Here is
179 the list of all potential fields:
180 * "url" - fragment's URL
181 * "path" - fragment's path relative to
183 * "duration" (optional, int or float)
184 * "filesize" (optional, int)
185 * is_from_start Is a live format that can be downloaded
186 from the start. Boolean
187 * preference Order number of this format. If this field is
188 present and not None, the formats get sorted
189 by this field, regardless of all other values.
190 -1 for default (order by other properties),
191 -2 or smaller for less than default.
192 < -1000 to hide the format (if there is
193 another one which is strictly better)
194 * language Language code, e.g. "de" or "en-US".
195 * language_preference Is this in the language mentioned in
197 10 if it's what the URL is about,
198 -1 for default (don't know),
199 -10 otherwise, other values reserved for now.
200 * quality Order number of the video quality of this
201 format, irrespective of the file format.
202 -1 for default (order by other properties),
203 -2 or smaller for less than default.
204 * source_preference Order number for this video source
205 (quality takes higher priority)
206 -1 for default (order by other properties),
207 -2 or smaller for less than default.
208 * http_headers A dictionary of additional HTTP headers
209 to add to the request.
210 * stretched_ratio If given and not 1, indicates that the
211 video's pixels are not square.
212 width : height ratio as float.
213 * no_resume The server does not support resuming the
214 (HTTP or RTMP) download. Boolean.
215 * has_drm The format has DRM and cannot be downloaded. Boolean
216 * downloader_options A dictionary of downloader options as
217 described in FileDownloader
218 RTMP formats can also have the additional fields: page_url,
219 app, play_path, tc_url, flash_version, rtmp_live, rtmp_conn,
220 rtmp_protocol, rtmp_real_time
222 url: Final video URL.
223 ext: Video filename extension.
224 format: The video format, defaults to ext (used for --get-format)
225 player_url: SWF Player URL (used for rtmpdump).
227 The following fields are optional:
229 direct: True if a direct video file was given (must only be set by GenericIE)
230 alt_title: A secondary title of the video.
231 display_id An alternative identifier for the video, not necessarily
232 unique, but available before title. Typically, id is
233 something like "4234987", title "Dancing naked mole rats",
234 and display_id "dancing-naked-mole-rats"
235 thumbnails: A list of dictionaries, with the following entries:
236 * "id" (optional, string) - Thumbnail format ID
238 * "preference" (optional, int) - quality of the image
239 * "width" (optional, int)
240 * "height" (optional, int)
241 * "resolution" (optional, string "{width}x{height}",
243 * "filesize" (optional, int)
244 * "http_headers" (dict) - HTTP headers for the request
245 thumbnail: Full URL to a video thumbnail image.
246 description: Full video description.
247 uploader: Full name of the video uploader.
248 license: License name the video is licensed under.
249 creator: The creator of the video.
250 timestamp: UNIX timestamp of the moment the video was uploaded
251 upload_date: Video upload date (YYYYMMDD).
252 If not explicitly set, calculated from timestamp
253 release_timestamp: UNIX timestamp of the moment the video was released.
254 If it is not clear whether to use timestamp or this, use the former
255 release_date: The date (YYYYMMDD) when the video was released.
256 If not explicitly set, calculated from release_timestamp
257 modified_timestamp: UNIX timestamp of the moment the video was last modified.
258 modified_date: The date (YYYYMMDD) when the video was last modified.
259 If not explicitly set, calculated from modified_timestamp
260 uploader_id: Nickname or id of the video uploader.
261 uploader_url: Full URL to a personal webpage of the video uploader.
262 channel: Full name of the channel the video is uploaded on.
263 Note that channel fields may or may not repeat uploader
264 fields. This depends on a particular extractor.
265 channel_id: Id of the channel.
266 channel_url: Full URL to a channel webpage.
267 channel_follower_count: Number of followers of the channel.
268 location: Physical location where the video was filmed.
269 subtitles: The available subtitles as a dictionary in the format
270 {tag: subformats}. "tag" is usually a language code, and
271 "subformats" is a list sorted from lower to higher
272 preference, each element is a dictionary with the "ext"
274 * "data": The subtitles file contents
275 * "url": A URL pointing to the subtitles file
276 It can optionally also have:
277 * "name": Name or description of the subtitles
278 * "http_headers": A dictionary of additional HTTP headers
279 to add to the request.
280 "ext" will be calculated from URL if missing
281 automatic_captions: Like 'subtitles'; contains automatically generated
282 captions instead of normal subtitles
283 duration: Length of the video in seconds, as an integer or float.
284 view_count: How many users have watched the video on the platform.
285 like_count: Number of positive ratings of the video
286 dislike_count: Number of negative ratings of the video
287 repost_count: Number of reposts of the video
288 average_rating: Average rating give by users, the scale used depends on the webpage
289 comment_count: Number of comments on the video
290 comments: A list of comments, each with one or more of the following
291 properties (all but one of text or html optional):
292 * "author" - human-readable name of the comment author
293 * "author_id" - user ID of the comment author
294 * "author_thumbnail" - The thumbnail of the comment author
296 * "html" - Comment as HTML
297 * "text" - Plain text of the comment
298 * "timestamp" - UNIX timestamp of comment
299 * "parent" - ID of the comment this one is replying to.
300 Set to "root" to indicate that this is a
301 comment to the original video.
302 * "like_count" - Number of positive ratings of the comment
303 * "dislike_count" - Number of negative ratings of the comment
304 * "is_favorited" - Whether the comment is marked as
305 favorite by the video uploader
306 * "author_is_uploader" - Whether the comment is made by
308 age_limit: Age restriction for the video, as an integer (years)
309 webpage_url: The URL to the video webpage, if given to yt-dlp it
310 should allow to get the same result again. (It will be set
311 by YoutubeDL if it's missing)
312 categories: A list of categories that the video falls in, for example
314 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
315 cast: A list of the video cast
316 is_live: True, False, or None (=unknown). Whether this video is a
317 live stream that goes on instead of a fixed-length video.
318 was_live: True, False, or None (=unknown). Whether this video was
319 originally a live stream.
320 live_status: 'is_live', 'is_upcoming', 'was_live', 'not_live' or None (=unknown)
321 If absent, automatically set from is_live, was_live
322 start_time: Time in seconds where the reproduction should start, as
323 specified in the URL.
324 end_time: Time in seconds where the reproduction should end, as
325 specified in the URL.
326 chapters: A list of dictionaries, with the following entries:
327 * "start_time" - The start time of the chapter in seconds
328 * "end_time" - The end time of the chapter in seconds
329 * "title" (optional, string)
330 playable_in_embed: Whether this video is allowed to play in embedded
331 players on other sites. Can be True (=always allowed),
332 False (=never allowed), None (=unknown), or a string
333 specifying the criteria for embedability (Eg: 'whitelist')
334 availability: Under what condition the video is available. One of
335 'private', 'premium_only', 'subscriber_only', 'needs_auth',
336 'unlisted' or 'public'. Use 'InfoExtractor._availability'
338 __post_extractor: A function to be called just before the metadata is
339 written to either disk, logger or console. The function
340 must return a dict which will be added to the info_dict.
341 This is usefull for additional information that is
342 time-consuming to extract. Note that the fields thus
343 extracted will not be available to output template and
344 match_filter. So, only "comments" and "comment_count" are
345 currently allowed to be extracted via this method.
347 The following fields should only be used when the video belongs to some logical
350 chapter: Name or title of the chapter the video belongs to.
351 chapter_number: Number of the chapter the video belongs to, as an integer.
352 chapter_id: Id of the chapter the video belongs to, as a unicode string.
354 The following fields should only be used when the video is an episode of some
355 series, programme or podcast:
357 series: Title of the series or programme the video episode belongs to.
358 series_id: Id of the series or programme the video episode belongs to, as a unicode string.
359 season: Title of the season the video episode belongs to.
360 season_number: Number of the season the video episode belongs to, as an integer.
361 season_id: Id of the season the video episode belongs to, as a unicode string.
362 episode: Title of the video episode. Unlike mandatory video title field,
363 this field should denote the exact title of the video episode
364 without any kind of decoration.
365 episode_number: Number of the video episode within a season, as an integer.
366 episode_id: Id of the video episode, as a unicode string.
368 The following fields should only be used when the media is a track or a part of
371 track: Title of the track.
372 track_number: Number of the track within an album or a disc, as an integer.
373 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
375 artist: Artist(s) of the track.
376 genre: Genre(s) of the track.
377 album: Title of the album the track belongs to.
378 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
379 album_artist: List of all artists appeared on the album (e.g.
380 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
382 disc_number: Number of the disc or other physical medium the track belongs to,
384 release_year: Year (YYYY) when the album was released.
385 composer: Composer of the piece
387 Unless mentioned otherwise, the fields should be Unicode strings.
389 Unless mentioned otherwise, None is equivalent to absence of information.
392 _type "playlist" indicates multiple videos.
393 There must be a key "entries", which is a list, an iterable, or a PagedList
394 object, each element of which is a valid dictionary by this specification.
396 Additionally, playlists can have "id", "title", and any other relevent
397 attributes with the same semantics as videos (see above).
399 It can also have the following optional fields:
401 playlist_count: The total number of videos in a playlist. If not given,
402 YoutubeDL tries to calculate it from "entries"
405 _type "multi_video" indicates that there are multiple videos that
406 form a single show, for examples multiple acts of an opera or TV episode.
407 It must have an entries key like a playlist and contain all the keys
408 required for a video at the same time.
411 _type "url" indicates that the video must be extracted from another
412 location, possibly by a different extractor. Its only required key is:
413 "url" - the next URL to extract.
414 The key "ie_key" can be set to the class name (minus the trailing "IE",
415 e.g. "Youtube") if the extractor class is known in advance.
416 Additionally, the dictionary may have any properties of the resolved entity
417 known in advance, for example "title" if the title of the referred video is
421 _type "url_transparent" entities have the same specification as "url", but
422 indicate that the given additional information is more precise than the one
423 associated with the resolved URL.
424 This is useful when a site employs a video service that hosts the video and
425 its technical metadata, but that video service does not embed a useful
426 title, description etc.
429 Subclasses of this should define a _VALID_URL regexp and, re-define the
430 _real_extract() and (optionally) _real_initialize() methods.
431 Probably, they should also be added to the list of extractors.
433 Subclasses may also override suitable() if necessary, but ensure the function
434 signature is preserved and that this function imports everything it needs
435 (except other extractors), so that lazy_extractors works correctly.
437 To support username + password (or netrc) login, the extractor must define a
438 _NETRC_MACHINE and re-define _perform_login(username, password) and
439 (optionally) _initialize_pre_login() methods. The _perform_login method will
440 be called between _initialize_pre_login and _real_initialize if credentials
441 are passed by the user. In cases where it is necessary to have the login
442 process as part of the extraction rather than initialization, _perform_login
443 can be left undefined.
445 _GEO_BYPASS attribute may be set to False in order to disable
446 geo restriction bypass mechanisms for a particular extractor.
447 Though it won't disable explicit geo restriction bypass based on
448 country code provided with geo_bypass_country.
450 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
451 countries for this extractor. One of these countries will be used by
452 geo restriction bypass mechanism right away in order to bypass
453 geo restriction, of course, if the mechanism is not disabled.
455 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
456 IP blocks in CIDR notation for this extractor. One of these IP blocks
457 will be used by geo restriction bypass mechanism similarly
460 The _WORKING attribute should be set to False for broken IEs
461 in order to warn the users and skip the tests.
466 _x_forwarded_for_ip
= None
468 _GEO_COUNTRIES
= None
469 _GEO_IP_BLOCKS
= None
471 _NETRC_MACHINE
= None
474 'any': 'Use --cookies, --cookies-from-browser, --username and --password, or --netrc to provide account credentials',
476 'Use --cookies-from-browser or --cookies for the authentication. '
477 'See https://github.com/ytdl-org/youtube-dl#how-do-i-pass-cookies-to-youtube-dl for how to manually pass cookies'),
478 'password': 'Use --username and --password, or --netrc to provide account credentials',
481 def __init__(self
, downloader
=None):
482 """Constructor. Receives an optional downloader (a YoutubeDL instance).
483 If a downloader is not passed during initialization,
484 it must be set using "set_downloader()" before "extract()" is called"""
486 self
._x
_forwarded
_for
_ip
= None
487 self
._printed
_messages
= set()
488 self
.set_downloader(downloader
)
491 def _match_valid_url(cls
, url
):
492 # This does not use has/getattr intentionally - we want to know whether
493 # we have cached the regexp for *this* class, whereas getattr would also
494 # match the superclass
495 if '_VALID_URL_RE' not in cls
.__dict
__:
496 if '_VALID_URL' not in cls
.__dict
__:
497 cls
._VALID
_URL
= cls
._make
_valid
_url
()
498 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
499 return cls
._VALID
_URL
_RE
.match(url
)
502 def suitable(cls
, url
):
503 """Receives a URL and returns True if suitable for this IE."""
504 # This function must import everything it needs (except other extractors),
505 # so that lazy_extractors works correctly
506 return cls
._match
_valid
_url
(url
) is not None
509 def _match_id(cls
, url
):
510 return cls
._match
_valid
_url
(url
).group('id')
513 def get_temp_id(cls
, url
):
515 return cls
._match
_id
(url
)
516 except (IndexError, AttributeError):
521 """Getter method for _WORKING."""
525 def supports_login(cls
):
526 return bool(cls
._NETRC
_MACHINE
)
528 def initialize(self
):
529 """Initializes an instance (authentication, etc)."""
530 self
._printed
_messages
= set()
531 self
._initialize
_geo
_bypass
({
532 'countries': self
._GEO
_COUNTRIES
,
533 'ip_blocks': self
._GEO
_IP
_BLOCKS
,
536 self
._initialize
_pre
_login
()
537 if self
.supports_login():
538 username
, password
= self
._get
_login
_info
()
540 self
._perform
_login
(username
, password
)
541 elif self
.get_param('username') and False not in (self
.IE_DESC
, self
._NETRC
_MACHINE
):
542 self
.report_warning(f
'Login with password is not supported for this website. {self._LOGIN_HINTS["cookies"]}')
543 self
._real
_initialize
()
546 def _initialize_geo_bypass(self
, geo_bypass_context
):
548 Initialize geo restriction bypass mechanism.
550 This method is used to initialize geo bypass mechanism based on faking
551 X-Forwarded-For HTTP header. A random country from provided country list
552 is selected and a random IP belonging to this country is generated. This
553 IP will be passed as X-Forwarded-For HTTP header in all subsequent
556 This method will be used for initial geo bypass mechanism initialization
557 during the instance initialization with _GEO_COUNTRIES and
560 You may also manually call it from extractor's code if geo bypass
561 information is not available beforehand (e.g. obtained during
562 extraction) or due to some other reason. In this case you should pass
563 this information in geo bypass context passed as first argument. It may
564 contain following fields:
566 countries: List of geo unrestricted countries (similar
568 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
569 (similar to _GEO_IP_BLOCKS)
572 if not self
._x
_forwarded
_for
_ip
:
574 # Geo bypass mechanism is explicitly disabled by user
575 if not self
.get_param('geo_bypass', True):
578 if not geo_bypass_context
:
579 geo_bypass_context
= {}
581 # Backward compatibility: previously _initialize_geo_bypass
582 # expected a list of countries, some 3rd party code may still use
584 if isinstance(geo_bypass_context
, (list, tuple)):
585 geo_bypass_context
= {
586 'countries': geo_bypass_context
,
589 # The whole point of geo bypass mechanism is to fake IP
590 # as X-Forwarded-For HTTP header based on some IP block or
593 # Path 1: bypassing based on IP block in CIDR notation
595 # Explicit IP block specified by user, use it right away
596 # regardless of whether extractor is geo bypassable or not
597 ip_block
= self
.get_param('geo_bypass_ip_block', None)
599 # Otherwise use random IP block from geo bypass context but only
600 # if extractor is known as geo bypassable
602 ip_blocks
= geo_bypass_context
.get('ip_blocks')
603 if self
._GEO
_BYPASS
and ip_blocks
:
604 ip_block
= random
.choice(ip_blocks
)
607 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(ip_block
)
608 self
._downloader
.write_debug(
609 '[debug] Using fake IP %s as X-Forwarded-For' % self
._x
_forwarded
_for
_ip
)
612 # Path 2: bypassing based on country code
614 # Explicit country code specified by user, use it right away
615 # regardless of whether extractor is geo bypassable or not
616 country
= self
.get_param('geo_bypass_country', None)
618 # Otherwise use random country code from geo bypass context but
619 # only if extractor is known as geo bypassable
621 countries
= geo_bypass_context
.get('countries')
622 if self
._GEO
_BYPASS
and countries
:
623 country
= random
.choice(countries
)
626 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country
)
627 self
._downloader
.write_debug(
628 'Using fake IP %s (%s) as X-Forwarded-For' % (self
._x
_forwarded
_for
_ip
, country
.upper()))
630 def extract(self
, url
):
631 """Extracts URL information and returns it in list of dicts."""
636 self
.write_debug('Extracting URL: %s' % url
)
637 ie_result
= self
._real
_extract
(url
)
638 if ie_result
is None:
640 if self
._x
_forwarded
_for
_ip
:
641 ie_result
['__x_forwarded_for_ip'] = self
._x
_forwarded
_for
_ip
642 subtitles
= ie_result
.get('subtitles')
643 if (subtitles
and 'live_chat' in subtitles
644 and 'no-live-chat' in self
.get_param('compat_opts', [])):
645 del subtitles
['live_chat']
647 except GeoRestrictedError
as e
:
648 if self
.__maybe
_fake
_ip
_and
_retry
(e
.countries
):
651 except UnsupportedError
:
653 except ExtractorError
as e
:
655 'video_id': e
.video_id
or self
.get_temp_id(url
),
657 'tb': e
.traceback
or sys
.exc_info()[2],
658 'expected': e
.expected
,
661 if hasattr(e
, 'countries'):
662 kwargs
['countries'] = e
.countries
663 raise type(e
)(e
.orig_msg
, **kwargs
)
664 except compat_http_client
.IncompleteRead
as e
:
665 raise ExtractorError('A network error has occurred.', cause
=e
, expected
=True, video_id
=self
.get_temp_id(url
))
666 except (KeyError, StopIteration) as e
:
667 raise ExtractorError('An extractor error has occurred.', cause
=e
, video_id
=self
.get_temp_id(url
))
669 def __maybe_fake_ip_and_retry(self
, countries
):
670 if (not self
.get_param('geo_bypass_country', None)
672 and self
.get_param('geo_bypass', True)
673 and not self
._x
_forwarded
_for
_ip
675 country_code
= random
.choice(countries
)
676 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country_code
)
677 if self
._x
_forwarded
_for
_ip
:
679 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
680 % (self
._x
_forwarded
_for
_ip
, country_code
.upper()))
684 def set_downloader(self
, downloader
):
685 """Sets a YoutubeDL instance as the downloader for this IE."""
686 self
._downloader
= downloader
688 def _initialize_pre_login(self
):
689 """ Intialization before login. Redefine in subclasses."""
692 def _perform_login(self
, username
, password
):
693 """ Login with username and password. Redefine in subclasses."""
696 def _real_initialize(self
):
697 """Real initialization process. Redefine in subclasses."""
700 def _real_extract(self
, url
):
701 """Real extraction process. Redefine in subclasses."""
702 raise NotImplementedError('This method must be implemented by subclasses')
706 """A string for getting the InfoExtractor with get_info_extractor"""
707 return cls
.__name
__[:-2]
711 return compat_str(type(self
).__name
__[:-2])
714 def __can_accept_status_code(err
, expected_status
):
715 assert isinstance(err
, compat_urllib_error
.HTTPError
)
716 if expected_status
is None:
718 elif callable(expected_status
):
719 return expected_status(err
.code
) is True
721 return err
.code
in variadic(expected_status
)
723 def _request_webpage(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
, expected_status
=None):
725 Return the response handle.
727 See _download_webpage docstring for arguments specification.
729 if not self
._downloader
._first
_webpage
_request
:
730 sleep_interval
= self
.get_param('sleep_interval_requests') or 0
731 if sleep_interval
> 0:
732 self
.to_screen('Sleeping %s seconds ...' % sleep_interval
)
733 time
.sleep(sleep_interval
)
735 self
._downloader
._first
_webpage
_request
= False
738 self
.report_download_webpage(video_id
)
739 elif note
is not False:
741 self
.to_screen('%s' % (note
,))
743 self
.to_screen('%s: %s' % (video_id
, note
))
745 # Some sites check X-Forwarded-For HTTP header in order to figure out
746 # the origin of the client behind proxy. This allows bypassing geo
747 # restriction by faking this header's value to IP that belongs to some
748 # geo unrestricted country. We will do so once we encounter any
749 # geo restriction error.
750 if self
._x
_forwarded
_for
_ip
:
751 if 'X-Forwarded-For' not in headers
:
752 headers
['X-Forwarded-For'] = self
._x
_forwarded
_for
_ip
754 if isinstance(url_or_request
, compat_urllib_request
.Request
):
755 url_or_request
= update_Request(
756 url_or_request
, data
=data
, headers
=headers
, query
=query
)
759 url_or_request
= update_url_query(url_or_request
, query
)
760 if data
is not None or headers
:
761 url_or_request
= sanitized_Request(url_or_request
, data
, headers
)
763 return self
._downloader
.urlopen(url_or_request
)
764 except network_exceptions
as err
:
765 if isinstance(err
, compat_urllib_error
.HTTPError
):
766 if self
.__can
_accept
_status
_code
(err
, expected_status
):
767 # Retain reference to error to prevent file object from
768 # being closed before it can be read. Works around the
769 # effects of <https://bugs.python.org/issue15002>
770 # introduced in Python 3.4.1.
777 errnote
= 'Unable to download webpage'
779 errmsg
= '%s: %s' % (errnote
, error_to_compat_str(err
))
781 raise ExtractorError(errmsg
, cause
=err
)
783 self
.report_warning(errmsg
)
786 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):
788 Return a tuple (page content as string, URL handle).
790 See _download_webpage docstring for arguments specification.
792 # Strip hashes from the URL (#1038)
793 if isinstance(url_or_request
, (compat_str
, str)):
794 url_or_request
= url_or_request
.partition('#')[0]
796 urlh
= self
._request
_webpage
(url_or_request
, video_id
, note
, errnote
, fatal
, data
=data
, headers
=headers
, query
=query
, expected_status
=expected_status
)
800 content
= self
._webpage
_read
_content
(urlh
, url_or_request
, video_id
, note
, errnote
, fatal
, encoding
=encoding
)
801 return (content
, urlh
)
804 def _guess_encoding_from_content(content_type
, webpage_bytes
):
805 m
= re
.match(r
'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type
)
807 encoding
= m
.group(1)
809 m
= re
.search(br
'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
810 webpage_bytes[:1024])
812 encoding = m.group(1).decode('ascii')
813 elif webpage_bytes.startswith(b'\xff\xfe'):
820 def __check_blocked(self, content):
821 first_block = content[:512]
822 if ('<title>Access to this site is blocked</title>' in content
823 and 'Websense' in first_block):
824 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
825 blocked_iframe = self._html_search_regex(
826 r'<iframe src="([^
"]+)"', content,
827 'Websense information URL
', default=None)
829 msg += ' Visit
%s for more details
' % blocked_iframe
830 raise ExtractorError(msg, expected=True)
831 if '<title
>The URL you requested has been blocked
</title
>' in first_block:
833 'Access to this webpage has been blocked by Indian censorship
. '
834 'Use a VPN
or proxy
server (with --proxy
) to route around it
.')
835 block_msg = self._html_search_regex(
836 r'</h1
><p
>(.*?
)</p
>',
837 content, 'block message
', default=None)
839 msg += ' (Message
: "%s")' % block_msg.replace('\n', ' ')
840 raise ExtractorError(msg, expected=True)
841 if ('<title
>TTK
:: Доступ к ресурсу ограничен
</title
>' in content
842 and 'blocklist
.rkn
.gov
.ru
' in content):
843 raise ExtractorError(
844 'Access to this webpage has been blocked by decision of the Russian government
. '
845 'Visit http
://blocklist
.rkn
.gov
.ru
/ for a block reason
.',
848 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
849 content_type = urlh.headers.get('Content
-Type
', '')
850 webpage_bytes = urlh.read()
851 if prefix is not None:
852 webpage_bytes = prefix + webpage_bytes
854 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
855 if self.get_param('dump_intermediate_pages
', False):
856 self.to_screen('Dumping request to
' + urlh.geturl())
857 dump = base64.b64encode(webpage_bytes).decode('ascii
')
858 self._downloader.to_screen(dump)
859 if self.get_param('write_pages
', False):
860 basen = '%s_%s' % (video_id, urlh.geturl())
861 trim_length = self.get_param('trim_file_name
') or 240
862 if len(basen) > trim_length:
863 h = '___
' + hashlib.md5(basen.encode('utf
-8')).hexdigest()
864 basen = basen[:trim_length - len(h)] + h
865 raw_filename = basen + '.dump
'
866 filename = sanitize_filename(raw_filename, restricted=True)
867 self.to_screen('Saving request to
' + filename)
868 # Working around MAX_PATH limitation on Windows (see
869 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
870 if compat_os_name == 'nt
':
871 absfilepath = os.path.abspath(filename)
872 if len(absfilepath) > 259:
873 filename = '\\\\?
\\' + absfilepath
874 with open(filename, 'wb
') as outf:
875 outf.write(webpage_bytes)
878 content = webpage_bytes.decode(encoding, 'replace
')
880 content = webpage_bytes.decode('utf
-8', 'replace
')
882 self.__check_blocked(content)
886 def _download_webpage(
887 self, url_or_request, video_id, note=None, errnote=None,
888 fatal=True, tries=1, timeout=5, encoding=None, data=None,
889 headers={}, query={}, expected_status=None):
891 Return the data of the page as a string.
894 url_or_request -- plain text URL as a string or
895 a compat_urllib_request.Requestobject
896 video_id -- Video/playlist/item identifier (string)
899 note -- note printed before downloading (string)
900 errnote -- note printed in case of an error (string)
901 fatal -- flag denoting whether error should be considered fatal,
902 i.e. whether it should cause ExtractionError to be raised,
903 otherwise a warning will be reported and extraction continued
904 tries -- number of tries
905 timeout -- sleep interval between tries
906 encoding -- encoding for a page content decoding, guessed automatically
907 when not explicitly specified
908 data -- POST data (bytes)
909 headers -- HTTP headers (dict)
910 query -- URL query (dict)
911 expected_status -- allows to accept failed HTTP requests (non 2xx
912 status code) by explicitly specifying a set of accepted status
913 codes. Can be any of the following entities:
914 - an integer type specifying an exact failed status code to
916 - a list or a tuple of integer types specifying a list of
917 failed status codes to accept
918 - a callable accepting an actual failed status code and
919 returning True if it should be accepted
920 Note that this argument does not affect success status codes (2xx)
921 which are always accepted.
926 while success is False:
928 res = self._download_webpage_handle(
929 url_or_request, video_id, note, errnote, fatal,
930 encoding=encoding, data=data, headers=headers, query=query,
931 expected_status=expected_status)
933 except compat_http_client.IncompleteRead as e:
935 if try_count >= tries:
937 self._sleep(timeout, video_id)
944 def _download_xml_handle(
945 self, url_or_request, video_id, note='Downloading XML
',
946 errnote='Unable to download XML
', transform_source=None,
947 fatal=True, encoding=None, data=None, headers={}, query={},
948 expected_status=None):
950 Return a tuple (xml as an compat_etree_Element, URL handle).
952 See _download_webpage docstring for arguments specification.
954 res = self._download_webpage_handle(
955 url_or_request, video_id, note, errnote, fatal=fatal,
956 encoding=encoding, data=data, headers=headers, query=query,
957 expected_status=expected_status)
960 xml_string, urlh = res
961 return self._parse_xml(
962 xml_string, video_id, transform_source=transform_source,
966 self, url_or_request, video_id,
967 note='Downloading XML
', errnote='Unable to download XML
',
968 transform_source=None, fatal=True, encoding=None,
969 data=None, headers={}, query={}, expected_status=None):
971 Return the xml as an compat_etree_Element.
973 See _download_webpage docstring for arguments specification.
975 res = self._download_xml_handle(
976 url_or_request, video_id, note=note, errnote=errnote,
977 transform_source=transform_source, fatal=fatal, encoding=encoding,
978 data=data, headers=headers, query=query,
979 expected_status=expected_status)
980 return res if res is False else res[0]
982 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
984 xml_string = transform_source(xml_string)
986 return compat_etree_fromstring(xml_string.encode('utf
-8'))
987 except compat_xml_parse_error as ve:
988 errmsg = '%s: Failed to parse XML
' % video_id
990 raise ExtractorError(errmsg, cause=ve)
992 self.report_warning(errmsg + str(ve))
994 def _download_json_handle(
995 self, url_or_request, video_id, note='Downloading JSON metadata
',
996 errnote='Unable to download JSON metadata
', transform_source=None,
997 fatal=True, encoding=None, data=None, headers={}, query={},
998 expected_status=None):
1000 Return a tuple (JSON object, URL handle).
1002 See _download_webpage docstring for arguments specification.
1004 res = self._download_webpage_handle(
1005 url_or_request, video_id, note, errnote, fatal=fatal,
1006 encoding=encoding, data=data, headers=headers, query=query,
1007 expected_status=expected_status)
1010 json_string, urlh = res
1011 return self._parse_json(
1012 json_string, video_id, transform_source=transform_source,
1016 self, url_or_request, video_id, note='Downloading JSON metadata
',
1017 errnote='Unable to download JSON metadata
', transform_source=None,
1018 fatal=True, encoding=None, data=None, headers={}, query={},
1019 expected_status=None):
1021 Return the JSON object as a dict.
1023 See _download_webpage docstring for arguments specification.
1025 res = self._download_json_handle(
1026 url_or_request, video_id, note=note, errnote=errnote,
1027 transform_source=transform_source, fatal=fatal, encoding=encoding,
1028 data=data, headers=headers, query=query,
1029 expected_status=expected_status)
1030 return res if res is False else res[0]
1032 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
1033 if transform_source:
1034 json_string = transform_source(json_string)
1036 return json.loads(json_string, strict=False)
1037 except ValueError as ve:
1038 errmsg = '%s: Failed to parse JSON
' % video_id
1040 raise ExtractorError(errmsg, cause=ve)
1042 self.report_warning(errmsg + str(ve))
1044 def _parse_socket_response_as_json(self, data, video_id, transform_source=None, fatal=True):
1045 return self._parse_json(
1046 data[data.find('{'):data.rfind('}
') + 1],
1047 video_id, transform_source, fatal)
1049 def _download_socket_json_handle(
1050 self, url_or_request, video_id, note='Polling socket
',
1051 errnote='Unable to poll socket
', transform_source=None,
1052 fatal=True, encoding=None, data=None, headers={}, query={},
1053 expected_status=None):
1055 Return a tuple (JSON object, URL handle).
1057 See _download_webpage docstring for arguments specification.
1059 res = self._download_webpage_handle(
1060 url_or_request, video_id, note, errnote, fatal=fatal,
1061 encoding=encoding, data=data, headers=headers, query=query,
1062 expected_status=expected_status)
1066 return self._parse_socket_response_as_json(
1067 webpage, video_id, transform_source=transform_source,
1070 def _download_socket_json(
1071 self, url_or_request, video_id, note='Polling socket
',
1072 errnote='Unable to poll socket
', transform_source=None,
1073 fatal=True, encoding=None, data=None, headers={}, query={},
1074 expected_status=None):
1076 Return the JSON object as a dict.
1078 See _download_webpage docstring for arguments specification.
1080 res = self._download_socket_json_handle(
1081 url_or_request, video_id, note=note, errnote=errnote,
1082 transform_source=transform_source, fatal=fatal, encoding=encoding,
1083 data=data, headers=headers, query=query,
1084 expected_status=expected_status)
1085 return res if res is False else res[0]
1087 def report_warning(self, msg, video_id=None, *args, only_once=False, **kwargs):
1088 idstr = format_field(video_id, template='%s: ')
1089 msg = f'[{self.IE_NAME}
] {idstr}{msg}
'
1091 if f'WARNING
: {msg}
' in self._printed_messages:
1093 self._printed_messages.add(f'WARNING
: {msg}
')
1094 self._downloader.report_warning(msg, *args, **kwargs)
1096 def to_screen(self, msg, *args, **kwargs):
1097 """Print msg to screen, prefixing it with '[ie_name
]'"""
1098 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg), *args, **kwargs)
1100 def write_debug(self, msg, *args, **kwargs):
1101 self._downloader.write_debug('[%s] %s' % (self.IE_NAME, msg), *args, **kwargs)
1103 def get_param(self, name, default=None, *args, **kwargs):
1104 if self._downloader:
1105 return self._downloader.params.get(name, default, *args, **kwargs)
1108 def report_drm(self, video_id, partial=False):
1109 self.raise_no_formats('This video
is DRM protected
', expected=True, video_id=video_id)
1111 def report_extraction(self, id_or_name):
1112 """Report information extraction."""
1113 self.to_screen('%s: Extracting information
' % id_or_name)
1115 def report_download_webpage(self, video_id):
1116 """Report webpage download."""
1117 self.to_screen('%s: Downloading webpage
' % video_id)
1119 def report_age_confirmation(self):
1120 """Report attempt to confirm age."""
1121 self.to_screen('Confirming age
')
1123 def report_login(self):
1124 """Report attempt to log in."""
1125 self.to_screen('Logging
in')
1127 def raise_login_required(
1128 self, msg='This video
is only available
for registered users
',
1129 metadata_available=False, method=NO_DEFAULT):
1130 if metadata_available and (
1131 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1132 self.report_warning(msg)
1134 if method is NO_DEFAULT:
1135 method = 'any
' if self.supports_login() else 'cookies
'
1136 if method is not None:
1137 assert method in self._LOGIN_HINTS, 'Invalid login method
'
1138 msg = '%s. %s' % (msg, self._LOGIN_HINTS[method])
1139 raise ExtractorError(msg, expected=True)
1141 def raise_geo_restricted(
1142 self, msg='This video
is not available
from your location due to geo restriction
',
1143 countries=None, metadata_available=False):
1144 if metadata_available and (
1145 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1146 self.report_warning(msg)
1148 raise GeoRestrictedError(msg, countries=countries)
1150 def raise_no_formats(self, msg, expected=False, video_id=None):
1152 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1153 self.report_warning(msg, video_id)
1154 elif isinstance(msg, ExtractorError):
1157 raise ExtractorError(msg, expected=expected, video_id=video_id)
1159 # Methods for following #608
1161 def url_result(url, ie=None, video_id=None, video_title=None, *, url_transparent=False, **kwargs):
1162 """Returns a URL that points to a page that should be processed"""
1164 kwargs['ie_key
'] = ie if isinstance(ie, str) else ie.ie_key()
1165 if video_id is not None:
1166 kwargs['id'] = video_id
1167 if video_title is not None:
1168 kwargs['title
'] = video_title
1171 '_type
': 'url_transparent
' if url_transparent else 'url
',
1175 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None, video_kwargs=None, **kwargs):
1176 urls = (self.url_result(self._proto_relative_url(m), ie, **(video_kwargs or {}))
1177 for m in orderedSet(map(getter, matches) if getter else matches))
1178 return self.playlist_result(urls, playlist_id, playlist_title, **kwargs)
1181 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None, *, multi_video=False, **kwargs):
1182 """Returns a playlist"""
1184 kwargs['id'] = playlist_id
1186 kwargs['title
'] = playlist_title
1187 if playlist_description is not None:
1188 kwargs['description
'] = playlist_description
1191 '_type
': 'multi_video
' if multi_video else 'playlist
',
1195 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1197 Perform a regex search on the given string, using a single or a list of
1198 patterns returning the first matching group.
1199 In case of failure return a default value or raise a WARNING or a
1200 RegexNotFoundError, depending on fatal, specifying the field name.
1202 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
1203 mobj = re.search(pattern, string, flags)
1206 mobj = re.search(p, string, flags)
1210 _name = self._downloader._format_err(name, self._downloader.Styles.EMPHASIS)
1214 # return the first matching group
1215 return next(g for g in mobj.groups() if g is not None)
1216 elif isinstance(group, (list, tuple)):
1217 return tuple(mobj.group(g) for g in group)
1219 return mobj.group(group)
1220 elif default is not NO_DEFAULT:
1223 raise RegexNotFoundError('Unable to extract
%s' % _name)
1225 self.report_warning('unable to extract
%s' % _name + bug_reports_message())
1228 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1230 Like _search_regex, but strips HTML tags and unescapes entities.
1232 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1234 return clean_html(res).strip()
1238 def _get_netrc_login_info(self, netrc_machine=None):
1241 netrc_machine = netrc_machine or self._NETRC_MACHINE
1243 if self.get_param('usenetrc
', False):
1245 netrc_file = compat_expanduser(self.get_param('netrc_location
') or '~
')
1246 if os.path.isdir(netrc_file):
1247 netrc_file = os.path.join(netrc_file, '.netrc
')
1248 info = netrc.netrc(file=netrc_file).authenticators(netrc_machine)
1249 if info is not None:
1253 raise netrc.NetrcParseError(
1254 'No authenticators
for %s' % netrc_machine)
1255 except (IOError, netrc.NetrcParseError) as err:
1256 self.report_warning(
1257 'parsing
.netrc
: %s' % error_to_compat_str(err))
1259 return username, password
1261 def _get_login_info(self, username_option='username
', password_option='password
', netrc_machine=None):
1263 Get the login info as (username, password)
1264 First look for the manually specified credentials using username_option
1265 and password_option as keys in params dictionary. If no such credentials
1266 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1268 If there's no info available
, return (None, None)
1271 # Attempt to use provided username and password or .netrc data
1272 username = self.get_param(username_option)
1273 if username is not None:
1274 password = self.get_param(password_option)
1276 username, password = self._get_netrc_login_info(netrc_machine)
1278 return username, password
1280 def _get_tfa_info(self, note='two-factor verification code'):
1282 Get the two
-factor authentication info
1283 TODO
- asking the user will be required
for sms
/phone verify
1284 currently just uses the command line option
1285 If there
's no info available, return None
1288 tfa = self.get_param('twofactor
')
1292 return compat_getpass('Type
%s and press
[Return
]: ' % note)
1294 # Helper functions for extracting OpenGraph info
1296 def _og_regexes(prop):
1297 content_re = r'content
=(?
:"([^"]+?
)"|\'([^\']+?)\'|\s*([^\s"\'=<>`
]+?
))'
1298 property_re = (r'(?
:name|
property)=(?
:\'og
[:-]%(prop)s\'|
"og[:-]%(prop)s"|\s
*og
[:-]%(prop)s\b)'
1299 % {'prop': re.escape(prop)})
1300 template = r'<meta
[^
>]+?
%s[^
>]+?
%s'
1302 template % (property_re, content_re),
1303 template % (content_re, property_re),
1307 def _meta_regex(prop):
1308 return r'''(?isx)<meta
1309 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1310 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1312 def _og_search_property(self, prop, html, name=None, **kargs):
1313 prop = variadic(prop)
1315 name = 'OpenGraph
%s' % prop[0]
1318 og_regexes.extend(self._og_regexes(p))
1319 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1322 return unescapeHTML(escaped)
1324 def _og_search_thumbnail(self, html, **kargs):
1325 return self._og_search_property('image
', html, 'thumbnail URL
', fatal=False, **kargs)
1327 def _og_search_description(self, html, **kargs):
1328 return self._og_search_property('description
', html, fatal=False, **kargs)
1330 def _og_search_title(self, html, **kargs):
1331 kargs.setdefault('fatal
', False)
1332 return self._og_search_property('title
', html, **kargs)
1334 def _og_search_video_url(self, html, name='video url
', secure=True, **kargs):
1335 regexes = self._og_regexes('video
') + self._og_regexes('video
:url
')
1337 regexes = self._og_regexes('video
:secure_url
') + regexes
1338 return self._html_search_regex(regexes, html, name, **kargs)
1340 def _og_search_url(self, html, **kargs):
1341 return self._og_search_property('url
', html, **kargs)
1343 def _html_extract_title(self, html, name, **kwargs):
1344 return self._html_search_regex(
1345 r'(?s
)<title
>(.*?
)</title
>', html, name, **kwargs)
1347 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1348 name = variadic(name)
1349 if display_name is None:
1350 display_name = name[0]
1351 return self._html_search_regex(
1352 [self._meta_regex(n) for n in name],
1353 html, display_name, fatal=fatal, group='content
', **kwargs)
1355 def _dc_search_uploader(self, html):
1356 return self._html_search_meta('dc
.creator
', html, 'uploader
')
1358 def _rta_search(self, html):
1359 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1360 if re.search(r'(?ix
)<meta\s
+name
="rating"\s
+'
1361 r' content
="RTA-5042-1996-1400-1577-RTA"',
1366 def _media_rating_search(self, html):
1367 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1368 rating = self._html_search_meta('rating
', html)
1380 return RATING_TABLE.get(rating.lower())
1382 def _family_friendly_search(self, html):
1383 # See http://schema.org/VideoObject
1384 family_friendly = self._html_search_meta(
1385 'isFamilyFriendly
', html, default=None)
1387 if not family_friendly:
1396 return RATING_TABLE.get(family_friendly.lower())
1398 def _twitter_search_player(self, html):
1399 return self._html_search_meta('twitter
:player
', html,
1400 'twitter card player
')
1402 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
1403 json_ld_list = list(re.finditer(JSON_LD_RE, html))
1404 default = kwargs.get('default
', NO_DEFAULT)
1405 # JSON-LD may be malformed and thus `fatal` should be respected.
1406 # At the same time `default` may be passed that assumes `fatal=False`
1407 # for _search_regex. Let's simulate the same behavior here
as well
.
1408 fatal
= kwargs
.get('fatal', True) if default
is NO_DEFAULT
else False
1410 for mobj
in json_ld_list
:
1411 json_ld_item
= self
._parse
_json
(
1412 mobj
.group('json_ld'), video_id
, fatal
=fatal
)
1413 if not json_ld_item
:
1415 if isinstance(json_ld_item
, dict):
1416 json_ld
.append(json_ld_item
)
1417 elif isinstance(json_ld_item
, (list, tuple)):
1418 json_ld
.extend(json_ld_item
)
1420 json_ld
= self
._json
_ld
(json_ld
, video_id
, fatal
=fatal
, expected_type
=expected_type
)
1423 if default
is not NO_DEFAULT
:
1426 raise RegexNotFoundError('Unable to extract JSON-LD')
1428 self
.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1431 def _json_ld(self
, json_ld
, video_id
, fatal
=True, expected_type
=None):
1432 if isinstance(json_ld
, compat_str
):
1433 json_ld
= self
._parse
_json
(json_ld
, video_id
, fatal
=fatal
)
1437 if not isinstance(json_ld
, (list, tuple, dict)):
1439 if isinstance(json_ld
, dict):
1442 INTERACTION_TYPE_MAP
= {
1443 'CommentAction': 'comment',
1444 'AgreeAction': 'like',
1445 'DisagreeAction': 'dislike',
1446 'LikeAction': 'like',
1447 'DislikeAction': 'dislike',
1448 'ListenAction': 'view',
1449 'WatchAction': 'view',
1450 'ViewAction': 'view',
1453 def extract_interaction_type(e
):
1454 interaction_type
= e
.get('interactionType')
1455 if isinstance(interaction_type
, dict):
1456 interaction_type
= interaction_type
.get('@type')
1457 return str_or_none(interaction_type
)
1459 def extract_interaction_statistic(e
):
1460 interaction_statistic
= e
.get('interactionStatistic')
1461 if isinstance(interaction_statistic
, dict):
1462 interaction_statistic
= [interaction_statistic
]
1463 if not isinstance(interaction_statistic
, list):
1465 for is_e
in interaction_statistic
:
1466 if not isinstance(is_e
, dict):
1468 if is_e
.get('@type') != 'InteractionCounter':
1470 interaction_type
= extract_interaction_type(is_e
)
1471 if not interaction_type
:
1473 # For interaction count some sites provide string instead of
1474 # an integer (as per spec) with non digit characters (e.g. ",")
1475 # so extracting count with more relaxed str_to_int
1476 interaction_count
= str_to_int(is_e
.get('userInteractionCount'))
1477 if interaction_count
is None:
1479 count_kind
= INTERACTION_TYPE_MAP
.get(interaction_type
.split('/')[-1])
1482 count_key
= '%s_count' % count_kind
1483 if info
.get(count_key
) is not None:
1485 info
[count_key
] = interaction_count
1487 def extract_chapter_information(e
):
1489 'title': part
.get('name'),
1490 'start_time': part
.get('startOffset'),
1491 'end_time': part
.get('endOffset'),
1492 } for part
in variadic(e
.get('hasPart') or []) if part
.get('@type') == 'Clip']
1493 for idx
, (last_c
, current_c
, next_c
) in enumerate(zip(
1494 [{'end_time': 0}
] + chapters
, chapters
, chapters
[1:])):
1495 current_c
['end_time'] = current_c
['end_time'] or next_c
['start_time']
1496 current_c
['start_time'] = current_c
['start_time'] or last_c
['end_time']
1497 if None in current_c
.values():
1498 self
.report_warning(f
'Chapter {idx} contains broken data. Not extracting chapters')
1501 chapters
[-1]['end_time'] = chapters
[-1]['end_time'] or info
['duration']
1502 info
['chapters'] = chapters
1504 def extract_video_object(e
):
1505 assert e
['@type'] == 'VideoObject'
1506 author
= e
.get('author')
1508 'url': url_or_none(e
.get('contentUrl')),
1509 'title': unescapeHTML(e
.get('name')),
1510 'description': unescapeHTML(e
.get('description')),
1511 'thumbnails': [{'url': url_or_none(url)}
1512 for url
in variadic(traverse_obj(e
, 'thumbnailUrl', 'thumbnailURL'))],
1513 'duration': parse_duration(e
.get('duration')),
1514 'timestamp': unified_timestamp(e
.get('uploadDate')),
1515 # author can be an instance of 'Organization' or 'Person' types.
1516 # both types can have 'name' property(inherited from 'Thing' type). [1]
1517 # however some websites are using 'Text' type instead.
1518 # 1. https://schema.org/VideoObject
1519 'uploader': author
.get('name') if isinstance(author
, dict) else author
if isinstance(author
, compat_str
) else None,
1520 'filesize': float_or_none(e
.get('contentSize')),
1521 'tbr': int_or_none(e
.get('bitrate')),
1522 'width': int_or_none(e
.get('width')),
1523 'height': int_or_none(e
.get('height')),
1524 'view_count': int_or_none(e
.get('interactionCount')),
1526 extract_interaction_statistic(e
)
1527 extract_chapter_information(e
)
1529 def traverse_json_ld(json_ld
, at_top_level
=True):
1531 if at_top_level
and '@context' not in e
:
1533 if at_top_level
and set(e
.keys()) == {'@context', '@graph'}
:
1534 traverse_json_ld(variadic(e
['@graph'], allowed_types
=(dict,)), at_top_level
=False)
1536 item_type
= e
.get('@type')
1537 if expected_type
is not None and expected_type
!= item_type
:
1539 rating
= traverse_obj(e
, ('aggregateRating', 'ratingValue'), expected_type
=float_or_none
)
1540 if rating
is not None:
1541 info
['average_rating'] = rating
1542 if item_type
in ('TVEpisode', 'Episode'):
1543 episode_name
= unescapeHTML(e
.get('name'))
1545 'episode': episode_name
,
1546 'episode_number': int_or_none(e
.get('episodeNumber')),
1547 'description': unescapeHTML(e
.get('description')),
1549 if not info
.get('title') and episode_name
:
1550 info
['title'] = episode_name
1551 part_of_season
= e
.get('partOfSeason')
1552 if isinstance(part_of_season
, dict) and part_of_season
.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1554 'season': unescapeHTML(part_of_season
.get('name')),
1555 'season_number': int_or_none(part_of_season
.get('seasonNumber')),
1557 part_of_series
= e
.get('partOfSeries') or e
.get('partOfTVSeries')
1558 if isinstance(part_of_series
, dict) and part_of_series
.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1559 info
['series'] = unescapeHTML(part_of_series
.get('name'))
1560 elif item_type
== 'Movie':
1562 'title': unescapeHTML(e
.get('name')),
1563 'description': unescapeHTML(e
.get('description')),
1564 'duration': parse_duration(e
.get('duration')),
1565 'timestamp': unified_timestamp(e
.get('dateCreated')),
1567 elif item_type
in ('Article', 'NewsArticle'):
1569 'timestamp': parse_iso8601(e
.get('datePublished')),
1570 'title': unescapeHTML(e
.get('headline')),
1571 'description': unescapeHTML(e
.get('articleBody') or e
.get('description')),
1573 if traverse_obj(e
, ('video', 0, '@type')) == 'VideoObject':
1574 extract_video_object(e
['video'][0])
1575 elif item_type
== 'VideoObject':
1576 extract_video_object(e
)
1577 if expected_type
is None:
1581 video
= e
.get('video')
1582 if isinstance(video
, dict) and video
.get('@type') == 'VideoObject':
1583 extract_video_object(video
)
1584 if expected_type
is None:
1588 traverse_json_ld(json_ld
)
1590 return dict((k
, v
) for k
, v
in info
.items() if v
is not None)
1592 def _search_nextjs_data(self
, webpage
, video_id
, *, transform_source
=None, fatal
=True, **kw
):
1593 return self
._parse
_json
(
1595 r
'(?s)<script[^>]+id=[\'"]__NEXT_DATA__[\'"][^
>]*>([^
<]+)</script
>',
1596 webpage, 'next
.js data
', fatal=fatal, **kw),
1597 video_id, transform_source=transform_source, fatal=fatal)
1599 def _search_nuxt_data(self, webpage, video_id, context_name='__NUXT__
'):
1600 ''' Parses Nuxt.js metadata. This works as long as the function __NUXT__ invokes is a pure function. '''
1601 # not all website do this, but it can be changed
1602 # https://stackoverflow.com/questions/67463109/how-to-change-or-hide-nuxt-and-nuxt-keyword-in-page-source
1603 rectx = re.escape(context_name)
1604 js, arg_keys, arg_vals = self._search_regex(
1605 (r'<script
>window\
.%s=\
(function\
((?P
<arg_keys
>.*?
)\
)\{return\s(?P<js>\{.*?\}
)\
}\
((?P
<arg_vals
>.+?
)\
)\
);?
</script
>' % rectx,
1606 r'%s\
(.*?\
(function\
((?P
<arg_keys
>.*?
)\
)\{return\s(?P<js>\{.*?\}
)\
}\
((?P
<arg_vals
>.*?
)\
)' % rectx),
1607 webpage, context_name, group=['js
', 'arg_keys
', 'arg_vals
'])
1609 args = dict(zip(arg_keys.split(','), arg_vals.split(',')))
1611 for key, val in args.items():
1612 if val in ('undefined
', 'void
0'):
1615 return self._parse_json(js_to_json(js, args), video_id)['data
'][0]
1618 def _hidden_inputs(html):
1619 html = re.sub(r'<!--(?
:(?
!<!--).)*-->', '', html)
1621 for input in re.findall(r'(?i
)(<input[^
>]+>)', html):
1622 attrs = extract_attributes(input)
1625 if attrs.get('type') not in ('hidden
', 'submit
'):
1627 name = attrs.get('name
') or attrs.get('id')
1628 value = attrs.get('value
')
1629 if name and value is not None:
1630 hidden_inputs[name] = value
1631 return hidden_inputs
1633 def _form_hidden_inputs(self, form_id, html):
1634 form = self._search_regex(
1635 r'(?
is)<form
[^
>]+?
id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1636 html, '%s form' % form_id, group='form')
1637 return self._hidden_inputs(form)
1640 regex = r' *((?P<reverse>\+)?(?P<field>[a-zA-Z0-9_]+)((?P<separator>[~:])(?P<limit>.*?))?)? *$'
1642 default = ('hidden', 'aud_or_vid', 'hasvid', 'ie_pref', 'lang', 'quality',
1643 'res', 'fps', 'hdr:12', 'codec:vp9.2', 'size', 'br', 'asr',
1644 'proto', 'ext', 'hasaud', 'source', 'id') # These must not be aliases
1645 ytdl_default = ('hasaud', 'lang', 'quality', 'tbr', 'filesize', 'vbr',
1646 'height', 'width', 'proto', 'vext', 'abr', 'aext',
1647 'fps', 'fs_approx', 'source', 'id')
1650 'vcodec': {'type': 'ordered', 'regex': True,
1651 'order': ['av0?1', 'vp0?9.2', 'vp0?9', '[hx]265|he?vc?', '[hx]264|avc', 'vp0?8', 'mp4v|h263', 'theora', '', None, 'none']},
1652 'acodec': {'type': 'ordered', 'regex': True,
1653 'order': ['[af]lac', 'wav|aiff', 'opus', 'vorbis|ogg', 'aac', 'mp?4a?', 'mp3', 'e-?a?c-?3', 'ac-?3', 'dts', '', None, 'none']},
1654 'hdr': {'type': 'ordered', 'regex': True, 'field': 'dynamic_range',
1655 'order': ['dv', '(hdr)?12', r'(hdr)?10\+', '(hdr)?10', 'hlg', '', 'sdr', None]},
1656 'proto': {'type': 'ordered', 'regex': True, 'field': 'protocol',
1657 'order': ['(ht|f)tps', '(ht|f)tp$', 'm3u8.*', '.*dash', 'websocket_frag', 'rtmpe?', '', 'mms|rtsp', 'ws|websocket', 'f4']},
1658 'vext': {'type': 'ordered', 'field': 'video_ext',
1659 'order': ('mp4', 'webm', 'flv', '', 'none'),
1660 'order_free': ('webm', 'mp4', 'flv', '', 'none')},
1661 'aext': {'type': 'ordered', 'field': 'audio_ext',
1662 'order': ('m4a', 'aac', 'mp3', 'ogg', 'opus', 'webm', '', 'none'),
1663 'order_free': ('opus', 'ogg', 'webm', 'm4a', 'mp3', 'aac', '', 'none')},
1664 'hidden': {'visible': False, 'forced': True, 'type': 'extractor', 'max': -1000},
1665 'aud_or_vid': {'visible': False, 'forced': True, 'type': 'multiple',
1666 'field': ('vcodec', 'acodec'),
1667 'function': lambda it: int(any(v != 'none' for v in it))},
1668 'ie_pref': {'priority': True, 'type': 'extractor'},
1669 'hasvid': {'priority': True, 'field': 'vcodec', 'type': 'boolean', 'not_in_list': ('none',)},
1670 'hasaud': {'field': 'acodec', 'type': 'boolean', 'not_in_list': ('none',)},
1671 'lang': {'convert': 'float', 'field': 'language_preference', 'default': -1},
1672 'quality': {'convert': 'float', 'default': -1},
1673 'filesize': {'convert': 'bytes'},
1674 'fs_approx': {'convert': 'bytes', 'field': 'filesize_approx'},
1675 'id': {'convert': 'string', 'field': 'format_id'},
1676 'height': {'convert': 'float_none'},
1677 'width': {'convert': 'float_none'},
1678 'fps': {'convert': 'float_none'},
1679 'tbr': {'convert': 'float_none'},
1680 'vbr': {'convert': 'float_none'},
1681 'abr': {'convert': 'float_none'},
1682 'asr': {'convert': 'float_none'},
1683 'source': {'convert': 'float', 'field': 'source_preference', 'default': -1},
1685 'codec': {'type': 'combined', 'field': ('vcodec', 'acodec')},
1686 'br': {'type': 'combined', 'field': ('tbr', 'vbr', 'abr'), 'same_limit': True},
1687 'size': {'type': 'combined', 'same_limit': True, 'field': ('filesize', 'fs_approx')},
1688 'ext': {'type': 'combined', 'field': ('vext', 'aext')},
1689 'res': {'type': 'multiple', 'field': ('height', 'width'),
1690 'function': lambda it: (lambda l: min(l) if l else 0)(tuple(filter(None, it)))},
1692 # For compatibility with youtube-dl
1693 'format_id': {'type': 'alias', 'field': 'id'},
1694 'preference': {'type': 'alias', 'field': 'ie_pref'},
1695 'language_preference': {'type': 'alias', 'field': 'lang'},
1696 'source_preference': {'type': 'alias', 'field': 'source'},
1697 'protocol': {'type': 'alias', 'field': 'proto'},
1698 'filesize_approx': {'type': 'alias', 'field': 'fs_approx'},
1701 'dimension': {'type': 'alias', 'field': 'res', 'deprecated': True},
1702 'resolution': {'type': 'alias', 'field': 'res', 'deprecated': True},
1703 'extension': {'type': 'alias', 'field': 'ext', 'deprecated': True},
1704 'bitrate': {'type': 'alias', 'field': 'br', 'deprecated': True},
1705 'total_bitrate': {'type': 'alias', 'field': 'tbr', 'deprecated': True},
1706 'video_bitrate': {'type': 'alias', 'field': 'vbr', 'deprecated': True},
1707 'audio_bitrate': {'type': 'alias', 'field': 'abr', 'deprecated': True},
1708 'framerate': {'type': 'alias', 'field': 'fps', 'deprecated': True},
1709 'filesize_estimate': {'type': 'alias', 'field': 'size', 'deprecated': True},
1710 'samplerate': {'type': 'alias', 'field': 'asr', 'deprecated': True},
1711 'video_ext': {'type': 'alias', 'field': 'vext', 'deprecated': True},
1712 'audio_ext': {'type': 'alias', 'field': 'aext', 'deprecated': True},
1713 'video_codec': {'type': 'alias', 'field': 'vcodec', 'deprecated': True},
1714 'audio_codec': {'type': 'alias', 'field': 'acodec', 'deprecated': True},
1715 'video': {'type': 'alias', 'field': 'hasvid', 'deprecated': True},
1716 'has_video': {'type': 'alias', 'field': 'hasvid', 'deprecated': True},
1717 'audio': {'type': 'alias', 'field': 'hasaud', 'deprecated': True},
1718 'has_audio': {'type': 'alias', 'field': 'hasaud', 'deprecated': True},
1719 'extractor': {'type': 'alias', 'field': 'ie_pref', 'deprecated': True},
1720 'extractor_preference': {'type': 'alias', 'field': 'ie_pref', 'deprecated': True},
1723 def __init__(self, ie, field_preference):
1725 self.ydl = ie._downloader
1726 self.evaluate_params(self.ydl.params, field_preference)
1727 if ie.get_param('verbose'):
1728 self.print_verbose_info(self.ydl.write_debug)
1730 def _get_field_setting(self, field, key):
1731 if field not in self.settings:
1732 if key in ('forced', 'priority'):
1734 self.ydl.deprecation_warning(
1735 f'Using arbitrary fields ({field}) for format sorting is deprecated '
1736 'and may be removed in a future version')
1737 self.settings[field] = {}
1738 propObj = self.settings[field]
1739 if key not in propObj:
1740 type = propObj.get('type')
1742 default = 'preference' if type == 'extractor' else (field,) if type in ('combined', 'multiple') else field
1743 elif key == 'convert':
1744 default = 'order' if type == 'ordered' else 'float_string' if field else 'ignore'
1746 default = {'type': 'field', 'visible': True, 'order': [], 'not_in_list': (None,)}.get(key, None)
1747 propObj[key] = default
1750 def _resolve_field_value(self, field, value, convertNone=False):
1755 value = value.lower()
1756 conversion = self._get_field_setting(field, 'convert')
1757 if conversion == 'ignore':
1759 if conversion == 'string':
1761 elif conversion == 'float_none':
1762 return float_or_none(value)
1763 elif conversion == 'bytes':
1764 return FileDownloader.parse_bytes(value)
1765 elif conversion == 'order':
1766 order_list = (self._use_free_order and self._get_field_setting(field, 'order_free')) or self._get_field_setting(field, 'order')
1767 use_regex = self._get_field_setting(field, 'regex')
1768 list_length = len(order_list)
1769 empty_pos = order_list.index('') if '' in order_list else list_length + 1
1770 if use_regex and value is not None:
1771 for i, regex in enumerate(order_list):
1772 if regex and re.match(regex, value):
1773 return list_length - i
1774 return list_length - empty_pos # not in list
1775 else: # not regex or value = None
1776 return list_length - (order_list.index(value) if value in order_list else empty_pos)
1778 if value.isnumeric():
1781 self.settings[field]['convert'] = 'string'
1784 def evaluate_params(self, params, sort_extractor):
1785 self._use_free_order = params.get('prefer_free_formats', False)
1786 self._sort_user = params.get('format_sort', [])
1787 self._sort_extractor = sort_extractor
1789 def add_item(field, reverse, closest, limit_text):
1790 field = field.lower()
1791 if field in self._order:
1793 self._order.append(field)
1794 limit = self._resolve_field_value(field, limit_text)
1797 'closest': False if limit is None else closest,
1798 'limit_text': limit_text,
1800 if field in self.settings:
1801 self.settings[field].update(data)
1803 self.settings[field] = data
1806 tuple(field for field in self.default if self._get_field_setting(field, 'forced'))
1807 + (tuple() if params.get('format_sort_force', False)
1808 else tuple(field for field in self.default if self._get_field_setting(field, 'priority')))
1809 + tuple(self._sort_user) + tuple(sort_extractor) + self.default)
1811 for item in sort_list:
1812 match = re.match(self.regex, item)
1814 raise ExtractorError('Invalid format sort string "%s" given by extractor' % item)
1815 field = match.group('field')
1818 if self._get_field_setting(field, 'type') == 'alias':
1819 alias, field = field, self._get_field_setting(field, 'field')
1820 if self._get_field_setting(alias, 'deprecated'):
1821 self.ydl.deprecation_warning(
1822 f'Format sorting alias {alias} is deprecated '
1823 f'and may be removed in a future version. Please use {field} instead')
1824 reverse = match.group('reverse') is not None
1825 closest = match.group('separator') == '~'
1826 limit_text = match.group('limit')
1828 has_limit = limit_text is not None
1829 has_multiple_fields = self._get_field_setting(field, 'type') == 'combined'
1830 has_multiple_limits = has_limit and has_multiple_fields and not self._get_field_setting(field, 'same_limit')
1832 fields = self._get_field_setting(field, 'field') if has_multiple_fields else (field,)
1833 limits = limit_text.split(':') if has_multiple_limits else (limit_text,) if has_limit else tuple()
1834 limit_count = len(limits)
1835 for (i, f) in enumerate(fields):
1836 add_item(f, reverse, closest,
1837 limits[i] if i < limit_count
1838 else limits[0] if has_limit and not has_multiple_limits
1841 def print_verbose_info(self, write_debug):
1843 write_debug('Sort order given by user: %s' % ', '.join(self._sort_user))
1844 if self._sort_extractor:
1845 write_debug('Sort order given by extractor: %s' % ', '.join(self._sort_extractor))
1846 write_debug('Formats sorted by: %s' % ', '.join(['%s%s%s' % (
1847 '+' if self._get_field_setting(field, 'reverse') else '', field,
1848 '%s%s(%s)' % ('~' if self._get_field_setting(field, 'closest') else ':',
1849 self._get_field_setting(field, 'limit_text'),
1850 self._get_field_setting(field, 'limit'))
1851 if self._get_field_setting(field, 'limit_text') is not None else '')
1852 for field in self._order if self._get_field_setting(field, 'visible')]))
1854 def _calculate_field_preference_from_value(self, format, field, type, value):
1855 reverse = self._get_field_setting(field, 'reverse')
1856 closest = self._get_field_setting(field, 'closest')
1857 limit = self._get_field_setting(field, 'limit')
1859 if type == 'extractor':
1860 maximum = self._get_field_setting(field, 'max')
1861 if value is None or (maximum is not None and value >= maximum):
1863 elif type == 'boolean':
1864 in_list = self._get_field_setting(field, 'in_list')
1865 not_in_list = self._get_field_setting(field, 'not_in_list')
1866 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
1867 elif type == 'ordered':
1868 value = self._resolve_field_value(field, value, True)
1870 # try to convert to number
1871 val_num = float_or_none(value, default=self._get_field_setting(field, 'default'))
1872 is_num = self._get_field_setting(field, 'convert') != 'string' and val_num is not None
1876 return ((-10, 0) if value is None
1877 else (1, value, 0) if not is_num # if a field has mixed strings and numbers, strings are sorted higher
1878 else (0, -abs(value - limit), value - limit if reverse else limit - value) if closest
1879 else (0, value, 0) if not reverse and (limit is None or value <= limit)
1880 else (0, -value, 0) if limit is None or (reverse and value == limit) or value > limit
1881 else (-1, value, 0))
1883 def _calculate_field_preference(self, format, field):
1884 type = self._get_field_setting(field, 'type') # extractor, boolean, ordered, field, multiple
1885 get_value = lambda f: format.get(self._get_field_setting(f, 'field'))
1886 if type == 'multiple':
1887 type = 'field' # Only 'field' is allowed in multiple for now
1888 actual_fields = self._get_field_setting(field, 'field')
1890 value = self._get_field_setting(field, 'function')(get_value(f) for f in actual_fields)
1892 value = get_value(field)
1893 return self._calculate_field_preference_from_value(format, field, type, value)
1895 def calculate_preference(self, format):
1896 # Determine missing protocol
1897 if not format.get('protocol'):
1898 format['protocol'] = determine_protocol(format)
1900 # Determine missing ext
1901 if not format.get('ext') and 'url' in format:
1902 format['ext'] = determine_ext(format['url'])
1903 if format.get('vcodec') == 'none':
1904 format['audio_ext'] = format['ext'] if format.get('acodec') != 'none' else 'none'
1905 format['video_ext'] = 'none'
1907 format['video_ext'] = format['ext']
1908 format['audio_ext'] = 'none'
1909 # if format.get('preference') is None and format.get('ext') in ('f4f', 'f4m'): # Not supported?
1910 # format['preference'] = -1000
1912 # Determine missing bitrates
1913 if format.get('tbr') is None:
1914 if format.get('vbr') is not None and format.get('abr') is not None:
1915 format['tbr'] = format.get('vbr', 0) + format.get('abr', 0)
1917 if format.get('vcodec') != 'none' and format.get('vbr') is None:
1918 format['vbr'] = format.get('tbr') - format.get('abr', 0)
1919 if format.get('acodec') != 'none' and format.get('abr') is None:
1920 format['abr'] = format.get('tbr') - format.get('vbr', 0)
1922 return tuple(self._calculate_field_preference(format, field) for field in self._order)
1924 def _sort_formats(self, formats, field_preference=[]):
1927 format_sort = self.FormatSort(self, field_preference)
1928 formats.sort(key=lambda f: format_sort.calculate_preference(f))
1930 def _check_formats(self, formats, video_id):
1932 formats[:] = filter(
1933 lambda f: self._is_valid_url(
1935 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1939 def _remove_duplicate_formats(formats):
1943 if f['url'] not in format_urls:
1944 format_urls.add(f['url'])
1945 unique_formats.append(f)
1946 formats[:] = unique_formats
1948 def _is_valid_url(self, url, video_id, item='video', headers={}):
1949 url = self._proto_relative_url(url, scheme='http:')
1950 # For now assume non HTTP(S) URLs always valid
1951 if not (url.startswith('http://') or url.startswith('https://')):
1954 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1956 except ExtractorError as e:
1958 '%s: %s URL is invalid, skipping: %s'
1959 % (video_id, item, error_to_compat_str(e.cause)))
1962 def http_scheme(self):
1963 """ Either "http
:" or "https
:", depending on the user's preferences """
1966 if self.get_param('prefer_insecure', False)
1969 def _proto_relative_url(self, url, scheme=None):
1972 if url.startswith('//'):
1974 scheme = self.http_scheme()
1979 def _sleep(self, timeout, video_id, msg_template=None):
1980 if msg_template is None:
1981 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1982 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1986 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1987 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1988 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
1989 manifest = self._download_xml(
1990 manifest_url, video_id, 'Downloading f4m manifest',
1991 'Unable to download f4m manifest',
1992 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1993 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
1994 transform_source=transform_source,
1995 fatal=fatal, data=data, headers=headers, query=query)
1997 if manifest is False:
2000 return self._parse_f4m_formats(
2001 manifest, manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
2002 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
2004 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
2005 transform_source=lambda s: fix_xml_ampersands(s).strip(),
2006 fatal=True, m3u8_id=None):
2007 if not isinstance(manifest, compat_etree_Element) and not fatal:
2010 # currently yt-dlp cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
2011 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
2012 if akamai_pv is not None and ';' in akamai_pv.text:
2013 playerVerificationChallenge = akamai_pv.text.split(';')[0]
2014 if playerVerificationChallenge.strip() != '':
2018 manifest_version = '1.0'
2019 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
2021 manifest_version = '2.0'
2022 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
2023 # Remove unsupported DRM protected media from final formats
2024 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
2025 media_nodes = remove_encrypted_media(media_nodes)
2029 manifest_base_url = get_base_url(manifest)
2031 bootstrap_info = xpath_element(
2032 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
2033 'bootstrap info', default=None)
2036 mime_type = xpath_text(
2037 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
2038 'base URL', default=None)
2039 if mime_type and mime_type.startswith('audio/'):
2042 for i, media_el in enumerate(media_nodes):
2043 tbr = int_or_none(media_el.attrib.get('bitrate'))
2044 width = int_or_none(media_el.attrib.get('width'))
2045 height = int_or_none(media_el.attrib.get('height'))
2046 format_id = join_nonempty(f4m_id, tbr or i)
2047 # If <bootstrapInfo> is present, the specified f4m is a
2048 # stream-level manifest, and only set-level manifests may refer to
2049 # external resources. See section 11.4 and section 4 of F4M spec
2050 if bootstrap_info is None:
2052 # @href is introduced in 2.0, see section 11.6 of F4M spec
2053 if manifest_version == '2.0':
2054 media_url = media_el.attrib.get('href')
2055 if media_url is None:
2056 media_url = media_el.attrib.get('url')
2060 media_url if media_url.startswith('http://') or media_url.startswith('https://')
2061 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
2062 # If media_url is itself a f4m manifest do the recursive extraction
2063 # since bitrates in parent manifest (this one) and media_url manifest
2064 # may differ leading to inability to resolve the format by requested
2065 # bitrate in f4m downloader
2066 ext = determine_ext(manifest_url)
2068 f4m_formats = self._extract_f4m_formats(
2069 manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
2070 transform_source=transform_source, fatal=fatal)
2071 # Sometimes stream-level manifest contains single media entry that
2072 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
2073 # At the same time parent's media entry in set-level manifest may
2074 # contain it. We will copy it from parent in such cases.
2075 if len(f4m_formats) == 1:
2078 'tbr': f.get('tbr') or tbr,
2079 'width': f.get('width') or width,
2080 'height': f.get('height') or height,
2081 'format_id': f.get('format_id') if not tbr else format_id,
2084 formats.extend(f4m_formats)
2087 formats.extend(self._extract_m3u8_formats(
2088 manifest_url, video_id, 'mp4', preference=preference,
2089 quality=quality, m3u8_id=m3u8_id, fatal=fatal))
2092 'format_id': format_id,
2093 'url': manifest_url,
2094 'manifest_url': manifest_url,
2095 'ext': 'flv' if bootstrap_info is not None else None,
2101 'preference': preference,
2106 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, quality=None, m3u8_id=None):
2108 'format_id': join_nonempty(m3u8_id, 'meta'),
2112 'preference': preference - 100 if preference else -100,
2114 'resolution': 'multiple',
2115 'format_note': 'Quality selection URL',
2118 def _report_ignoring_subs(self, name):
2119 self.report_warning(bug_reports_message(
2120 f'Ignoring subtitle tracks found in the {name} manifest; '
2121 'if any subtitle tracks are missing,'
2124 def _extract_m3u8_formats(self, *args, **kwargs):
2125 fmts, subs = self._extract_m3u8_formats_and_subtitles(*args, **kwargs)
2127 self._report_ignoring_subs('HLS')
2130 def _extract_m3u8_formats_and_subtitles(
2131 self, m3u8_url, video_id, ext=None, entry_protocol='m3u8_native',
2132 preference=None, quality=None, m3u8_id=None, note=None,
2133 errnote=None, fatal=True, live=False, data=None, headers={},
2136 res = self._download_webpage_handle(
2138 note='Downloading m3u8 information' if note is None else note,
2139 errnote='Failed to download m3u8 information' if errnote is None else errnote,
2140 fatal=fatal, data=data, headers=headers, query=query)
2145 m3u8_doc, urlh = res
2146 m3u8_url = urlh.geturl()
2148 return self._parse_m3u8_formats_and_subtitles(
2149 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
2150 preference=preference, quality=quality, m3u8_id=m3u8_id,
2151 note=note, errnote=errnote, fatal=fatal, live=live, data=data,
2152 headers=headers, query=query, video_id=video_id)
2154 def _parse_m3u8_formats_and_subtitles(
2155 self, m3u8_doc, m3u8_url=None, ext=None, entry_protocol='m3u8_native',
2156 preference=None, quality=None, m3u8_id=None, live=False, note=None,
2157 errnote=None, fatal=True, data=None, headers={}, query={},
2159 formats, subtitles = [], {}
2161 has_drm = re.search('|'.join([
2162 r'#EXT-X-FAXS-CM:', # Adobe Flash Access
2163 r'#EXT-X-(?:SESSION-)?KEY:.*?URI="skd
://', # Apple FairPlay
2166 def format_url(url):
2167 return url if re.match(r'^https?
://', url) else compat_urlparse.urljoin(m3u8_url, url)
2169 if self.get_param('hls_split_discontinuity
', False):
2170 def _extract_m3u8_playlist_indices(manifest_url=None, m3u8_doc=None):
2172 if not manifest_url:
2174 m3u8_doc = self._download_webpage(
2175 manifest_url, video_id, fatal=fatal, data=data, headers=headers,
2176 note=False, errnote='Failed to download m3u8 playlist information
')
2177 if m3u8_doc is False:
2179 return range(1 + sum(line.startswith('#EXT-X-DISCONTINUITY') for line in m3u8_doc.splitlines()))
2182 def _extract_m3u8_playlist_indices(*args
, **kwargs
):
2186 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
2187 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
2188 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
2190 # We should try extracting formats only from master playlists [1, 4.3.4],
2191 # i.e. playlists that describe available qualities. On the other hand
2192 # media playlists [1, 4.3.3] should be returned as is since they contain
2193 # just the media without qualities renditions.
2194 # Fortunately, master playlist can be easily distinguished from media
2195 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
2196 # master playlist tags MUST NOT appear in a media playlist and vice versa.
2197 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
2198 # media playlist and MUST NOT appear in master playlist thus we can
2199 # clearly detect media playlist with this criterion.
2201 if '#EXT-X-TARGETDURATION' in m3u8_doc
: # media playlist, return as is
2203 'format_id': join_nonempty(m3u8_id
, idx
),
2204 'format_index': idx
,
2205 'url': m3u8_url
or encode_data_uri(m3u8_doc
.encode('utf-8'), 'application/x-mpegurl'),
2207 'protocol': entry_protocol
,
2208 'preference': preference
,
2211 } for idx
in _extract_m3u8_playlist_indices(m3u8_doc
=m3u8_doc
)]
2213 return formats
, subtitles
2216 last_stream_inf
= {}
2218 def extract_media(x_media_line
):
2219 media
= parse_m3u8_attributes(x_media_line
)
2220 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
2221 media_type
, group_id
, name
= media
.get('TYPE'), media
.get('GROUP-ID'), media
.get('NAME')
2222 if not (media_type
and group_id
and name
):
2224 groups
.setdefault(group_id
, []).append(media
)
2225 # <https://tools.ietf.org/html/rfc8216#section-4.3.4.1>
2226 if media_type
== 'SUBTITLES':
2227 # According to RFC 8216 §4.3.4.2.1, URI is REQUIRED in the
2228 # EXT-X-MEDIA tag if the media type is SUBTITLES.
2229 # However, lack of URI has been spotted in the wild.
2230 # e.g. NebulaIE; see https://github.com/yt-dlp/yt-dlp/issues/339
2231 if not media
.get('URI'):
2233 url
= format_url(media
['URI'])
2236 'ext': determine_ext(url
),
2238 if sub_info
['ext'] == 'm3u8':
2239 # Per RFC 8216 §3.1, the only possible subtitle format m3u8
2240 # files may contain is WebVTT:
2241 # <https://tools.ietf.org/html/rfc8216#section-3.1>
2242 sub_info
['ext'] = 'vtt'
2243 sub_info
['protocol'] = 'm3u8_native'
2244 lang
= media
.get('LANGUAGE') or 'und'
2245 subtitles
.setdefault(lang
, []).append(sub_info
)
2246 if media_type
not in ('VIDEO', 'AUDIO'):
2248 media_url
= media
.get('URI')
2250 manifest_url
= format_url(media_url
)
2252 'format_id': join_nonempty(m3u8_id
, group_id
, name
, idx
),
2253 'format_note': name
,
2254 'format_index': idx
,
2255 'url': manifest_url
,
2256 'manifest_url': m3u8_url
,
2257 'language': media
.get('LANGUAGE'),
2259 'protocol': entry_protocol
,
2260 'preference': preference
,
2262 'vcodec': 'none' if media_type
== 'AUDIO' else None,
2263 } for idx
in _extract_m3u8_playlist_indices(manifest_url
))
2265 def build_stream_name():
2266 # Despite specification does not mention NAME attribute for
2267 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
2268 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
2269 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
2270 stream_name
= last_stream_inf
.get('NAME')
2273 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
2274 # from corresponding rendition group
2275 stream_group_id
= last_stream_inf
.get('VIDEO')
2276 if not stream_group_id
:
2278 stream_group
= groups
.get(stream_group_id
)
2279 if not stream_group
:
2280 return stream_group_id
2281 rendition
= stream_group
[0]
2282 return rendition
.get('NAME') or stream_group_id
2284 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
2285 # chance to detect video only formats when EXT-X-STREAM-INF tags
2286 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
2287 for line
in m3u8_doc
.splitlines():
2288 if line
.startswith('#EXT-X-MEDIA:'):
2291 for line
in m3u8_doc
.splitlines():
2292 if line
.startswith('#EXT-X-STREAM-INF:'):
2293 last_stream_inf
= parse_m3u8_attributes(line
)
2294 elif line
.startswith('#') or not line
.strip():
2297 tbr
= float_or_none(
2298 last_stream_inf
.get('AVERAGE-BANDWIDTH')
2299 or last_stream_inf
.get('BANDWIDTH'), scale
=1000)
2300 manifest_url
= format_url(line
.strip())
2302 for idx
in _extract_m3u8_playlist_indices(manifest_url
):
2303 format_id
= [m3u8_id
, None, idx
]
2304 # Bandwidth of live streams may differ over time thus making
2305 # format_id unpredictable. So it's better to keep provided
2308 stream_name
= build_stream_name()
2309 format_id
[1] = stream_name
or '%d' % (tbr
or len(formats
))
2311 'format_id': join_nonempty(*format_id
),
2312 'format_index': idx
,
2313 'url': manifest_url
,
2314 'manifest_url': m3u8_url
,
2317 'fps': float_or_none(last_stream_inf
.get('FRAME-RATE')),
2318 'protocol': entry_protocol
,
2319 'preference': preference
,
2322 resolution
= last_stream_inf
.get('RESOLUTION')
2324 mobj
= re
.search(r
'(?P<width>\d+)[xX](?P<height>\d+)', resolution
)
2326 f
['width'] = int(mobj
.group('width'))
2327 f
['height'] = int(mobj
.group('height'))
2328 # Unified Streaming Platform
2330 r
'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f
['url'])
2332 abr
, vbr
= mobj
.groups()
2333 abr
, vbr
= float_or_none(abr
, 1000), float_or_none(vbr
, 1000)
2338 codecs
= parse_codecs(last_stream_inf
.get('CODECS'))
2340 audio_group_id
= last_stream_inf
.get('AUDIO')
2341 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
2342 # references a rendition group MUST have a CODECS attribute.
2343 # However, this is not always respected, for example, [2]
2344 # contains EXT-X-STREAM-INF tag which references AUDIO
2345 # rendition group but does not have CODECS and despite
2346 # referencing an audio group it represents a complete
2347 # (with audio and video) format. So, for such cases we will
2348 # ignore references to rendition groups and treat them
2349 # as complete formats.
2350 if audio_group_id
and codecs
and f
.get('vcodec') != 'none':
2351 audio_group
= groups
.get(audio_group_id
)
2352 if audio_group
and audio_group
[0].get('URI'):
2353 # TODO: update acodec for audio only formats with
2355 f
['acodec'] = 'none'
2356 if not f
.get('ext'):
2357 f
['ext'] = 'm4a' if f
.get('vcodec') == 'none' else 'mp4'
2361 progressive_uri
= last_stream_inf
.get('PROGRESSIVE-URI')
2364 del http_f
['manifest_url']
2366 'format_id': f
['format_id'].replace('hls-', 'http-'),
2368 'url': progressive_uri
,
2370 formats
.append(http_f
)
2372 last_stream_inf
= {}
2373 return formats
, subtitles
2375 def _extract_m3u8_vod_duration(
2376 self
, m3u8_vod_url
, video_id
, note
=None, errnote
=None, data
=None, headers
={}, query={}
):
2378 m3u8_vod
= self
._download
_webpage
(
2379 m3u8_vod_url
, video_id
,
2380 note
='Downloading m3u8 VOD manifest' if note
is None else note
,
2381 errnote
='Failed to download VOD manifest' if errnote
is None else errnote
,
2382 fatal
=False, data
=data
, headers
=headers
, query
=query
)
2384 return self
._parse
_m
3u8_vod
_duration
(m3u8_vod
or '', video_id
)
2386 def _parse_m3u8_vod_duration(self
, m3u8_vod
, video_id
):
2387 if '#EXT-X-PLAYLIST-TYPE:VOD' not in m3u8_vod
:
2391 float(line
[len('#EXTINF:'):].split(',')[0])
2392 for line
in m3u8_vod
.splitlines() if line
.startswith('#EXTINF:'))) or None
2395 def _xpath_ns(path
, namespace
=None):
2399 for c
in path
.split('/'):
2400 if not c
or c
== '.':
2403 out
.append('{%s}%s' % (namespace
, c
))
2404 return '/'.join(out
)
2406 def _extract_smil_formats_and_subtitles(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None, transform_source
=None):
2407 smil
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
, transform_source
=transform_source
)
2413 namespace
= self
._parse
_smil
_namespace
(smil
)
2415 fmts
= self
._parse
_smil
_formats
(
2416 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2417 subs
= self
._parse
_smil
_subtitles
(
2418 smil
, namespace
=namespace
)
2422 def _extract_smil_formats(self
, *args
, **kwargs
):
2423 fmts
, subs
= self
._extract
_smil
_formats
_and
_subtitles
(*args
, **kwargs
)
2425 self
._report
_ignoring
_subs
('SMIL')
2428 def _extract_smil_info(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None):
2429 smil
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
)
2432 return self
._parse
_smil
(smil
, smil_url
, video_id
, f4m_params
=f4m_params
)
2434 def _download_smil(self
, smil_url
, video_id
, fatal
=True, transform_source
=None):
2435 return self
._download
_xml
(
2436 smil_url
, video_id
, 'Downloading SMIL file',
2437 'Unable to download SMIL file', fatal
=fatal
, transform_source
=transform_source
)
2439 def _parse_smil(self
, smil
, smil_url
, video_id
, f4m_params
=None):
2440 namespace
= self
._parse
_smil
_namespace
(smil
)
2442 formats
= self
._parse
_smil
_formats
(
2443 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2444 subtitles
= self
._parse
_smil
_subtitles
(smil
, namespace
=namespace
)
2446 video_id
= os
.path
.splitext(url_basename(smil_url
))[0]
2450 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2451 name
= meta
.attrib
.get('name')
2452 content
= meta
.attrib
.get('content')
2453 if not name
or not content
:
2455 if not title
and name
== 'title':
2457 elif not description
and name
in ('description', 'abstract'):
2458 description
= content
2459 elif not upload_date
and name
== 'date':
2460 upload_date
= unified_strdate(content
)
2463 'id': image
.get('type'),
2464 'url': image
.get('src'),
2465 'width': int_or_none(image
.get('width')),
2466 'height': int_or_none(image
.get('height')),
2467 } for image
in smil
.findall(self
._xpath
_ns
('.//image', namespace
)) if image
.get('src')]
2471 'title': title
or video_id
,
2472 'description': description
,
2473 'upload_date': upload_date
,
2474 'thumbnails': thumbnails
,
2476 'subtitles': subtitles
,
2479 def _parse_smil_namespace(self
, smil
):
2480 return self
._search
_regex
(
2481 r
'(?i)^{([^}]+)?}smil$', smil
.tag
, 'namespace', default
=None)
2483 def _parse_smil_formats(self
, smil
, smil_url
, video_id
, namespace
=None, f4m_params
=None, transform_rtmp_url
=None):
2485 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2486 b
= meta
.get('base') or meta
.get('httpBase')
2498 media
= smil
.findall(self
._xpath
_ns
('.//video', namespace
)) + smil
.findall(self
._xpath
_ns
('.//audio', namespace
))
2499 for medium
in media
:
2500 src
= medium
.get('src')
2501 if not src
or src
in srcs
:
2505 bitrate
= float_or_none(medium
.get('system-bitrate') or medium
.get('systemBitrate'), 1000)
2506 filesize
= int_or_none(medium
.get('size') or medium
.get('fileSize'))
2507 width
= int_or_none(medium
.get('width'))
2508 height
= int_or_none(medium
.get('height'))
2509 proto
= medium
.get('proto')
2510 ext
= medium
.get('ext')
2511 src_ext
= determine_ext(src
)
2512 streamer
= medium
.get('streamer') or base
2514 if proto
== 'rtmp' or streamer
.startswith('rtmp'):
2520 'format_id': 'rtmp-%d' % (rtmp_count
if bitrate
is None else bitrate
),
2522 'filesize': filesize
,
2526 if transform_rtmp_url
:
2527 streamer
, src
= transform_rtmp_url(streamer
, src
)
2528 formats
[-1].update({
2534 src_url
= src
if src
.startswith('http') else compat_urlparse
.urljoin(base
, src
)
2535 src_url
= src_url
.strip()
2537 if proto
== 'm3u8' or src_ext
== 'm3u8':
2538 m3u8_formats
= self
._extract
_m
3u8_formats
(
2539 src_url
, video_id
, ext
or 'mp4', m3u8_id
='hls', fatal
=False)
2540 if len(m3u8_formats
) == 1:
2542 m3u8_formats
[0].update({
2543 'format_id': 'hls-%d' % (m3u8_count
if bitrate
is None else bitrate
),
2548 formats
.extend(m3u8_formats
)
2549 elif src_ext
== 'f4m':
2554 'plugin': 'flowplayer-3.2.0.1',
2556 f4m_url
+= '&' if '?' in f4m_url
else '?'
2557 f4m_url
+= compat_urllib_parse_urlencode(f4m_params
)
2558 formats
.extend(self
._extract
_f
4m
_formats
(f4m_url
, video_id
, f4m_id
='hds', fatal
=False))
2559 elif src_ext
== 'mpd':
2560 formats
.extend(self
._extract
_mpd
_formats
(
2561 src_url
, video_id
, mpd_id
='dash', fatal
=False))
2562 elif re
.search(r
'\.ism/[Mm]anifest', src_url
):
2563 formats
.extend(self
._extract
_ism
_formats
(
2564 src_url
, video_id
, ism_id
='mss', fatal
=False))
2565 elif src_url
.startswith('http') and self
._is
_valid
_url
(src
, video_id
):
2569 'ext': ext
or src_ext
or 'flv',
2570 'format_id': 'http-%d' % (bitrate
or http_count
),
2572 'filesize': filesize
,
2577 for medium
in smil
.findall(self
._xpath
_ns
('.//imagestream', namespace
)):
2578 src
= medium
.get('src')
2579 if not src
or src
in srcs
:
2585 'format_id': 'imagestream-%d' % (imgs_count
),
2587 'ext': mimetype2ext(medium
.get('type')),
2590 'width': int_or_none(medium
.get('width')),
2591 'height': int_or_none(medium
.get('height')),
2592 'format_note': 'SMIL storyboards',
2597 def _parse_smil_subtitles(self
, smil
, namespace
=None, subtitles_lang
='en'):
2600 for num
, textstream
in enumerate(smil
.findall(self
._xpath
_ns
('.//textstream', namespace
))):
2601 src
= textstream
.get('src')
2602 if not src
or src
in urls
:
2605 ext
= textstream
.get('ext') or mimetype2ext(textstream
.get('type')) or determine_ext(src
)
2606 lang
= textstream
.get('systemLanguage') or textstream
.get('systemLanguageName') or textstream
.get('lang') or subtitles_lang
2607 subtitles
.setdefault(lang
, []).append({
2613 def _extract_xspf_playlist(self
, xspf_url
, playlist_id
, fatal
=True):
2614 xspf
= self
._download
_xml
(
2615 xspf_url
, playlist_id
, 'Downloading xpsf playlist',
2616 'Unable to download xspf manifest', fatal
=fatal
)
2619 return self
._parse
_xspf
(
2620 xspf
, playlist_id
, xspf_url
=xspf_url
,
2621 xspf_base_url
=base_url(xspf_url
))
2623 def _parse_xspf(self
, xspf_doc
, playlist_id
, xspf_url
=None, xspf_base_url
=None):
2625 'xspf': 'http://xspf.org/ns/0/',
2626 's1': 'http://static.streamone.nl/player/ns/0',
2630 for track
in xspf_doc
.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP
)):
2632 track
, xpath_with_ns('./xspf:title', NS_MAP
), 'title', default
=playlist_id
)
2633 description
= xpath_text(
2634 track
, xpath_with_ns('./xspf:annotation', NS_MAP
), 'description')
2635 thumbnail
= xpath_text(
2636 track
, xpath_with_ns('./xspf:image', NS_MAP
), 'thumbnail')
2637 duration
= float_or_none(
2638 xpath_text(track
, xpath_with_ns('./xspf:duration', NS_MAP
), 'duration'), 1000)
2641 for location
in track
.findall(xpath_with_ns('./xspf:location', NS_MAP
)):
2642 format_url
= urljoin(xspf_base_url
, location
.text
)
2647 'manifest_url': xspf_url
,
2648 'format_id': location
.get(xpath_with_ns('s1:label', NS_MAP
)),
2649 'width': int_or_none(location
.get(xpath_with_ns('s1:width', NS_MAP
))),
2650 'height': int_or_none(location
.get(xpath_with_ns('s1:height', NS_MAP
))),
2652 self
._sort
_formats
(formats
)
2657 'description': description
,
2658 'thumbnail': thumbnail
,
2659 'duration': duration
,
2664 def _extract_mpd_formats(self
, *args
, **kwargs
):
2665 fmts
, subs
= self
._extract
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2667 self
._report
_ignoring
_subs
('DASH')
2670 def _extract_mpd_formats_and_subtitles(
2671 self
, mpd_url
, video_id
, mpd_id
=None, note
=None, errnote
=None,
2672 fatal
=True, data
=None, headers
={}, query={}
):
2673 res
= self
._download
_xml
_handle
(
2675 note
='Downloading MPD manifest' if note
is None else note
,
2676 errnote
='Failed to download MPD manifest' if errnote
is None else errnote
,
2677 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
2683 mpd_base_url
= base_url(urlh
.geturl())
2685 return self
._parse
_mpd
_formats
_and
_subtitles
(
2686 mpd_doc
, mpd_id
, mpd_base_url
, mpd_url
)
2688 def _parse_mpd_formats(self
, *args
, **kwargs
):
2689 fmts
, subs
= self
._parse
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2691 self
._report
_ignoring
_subs
('DASH')
2694 def _parse_mpd_formats_and_subtitles(
2695 self
, mpd_doc
, mpd_id
=None, mpd_base_url
='', mpd_url
=None):
2697 Parse formats from MPD manifest.
2699 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
2700 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
2701 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
2703 if not self
.get_param('dynamic_mpd', True):
2704 if mpd_doc
.get('type') == 'dynamic':
2707 namespace
= self
._search
_regex
(r
'(?i)^{([^}]+)?}MPD$', mpd_doc
.tag
, 'namespace', default
=None)
2710 return self
._xpath
_ns
(path
, namespace
)
2712 def is_drm_protected(element
):
2713 return element
.find(_add_ns('ContentProtection')) is not None
2715 def extract_multisegment_info(element
, ms_parent_info
):
2716 ms_info
= ms_parent_info
.copy()
2718 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2719 # common attributes and elements. We will only extract relevant
2721 def extract_common(source
):
2722 segment_timeline
= source
.find(_add_ns('SegmentTimeline'))
2723 if segment_timeline
is not None:
2724 s_e
= segment_timeline
.findall(_add_ns('S'))
2726 ms_info
['total_number'] = 0
2729 r
= int(s
.get('r', 0))
2730 ms_info
['total_number'] += 1 + r
2731 ms_info
['s'].append({
2732 't': int(s
.get('t', 0)),
2733 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2734 'd': int(s
.attrib
['d']),
2737 start_number
= source
.get('startNumber')
2739 ms_info
['start_number'] = int(start_number
)
2740 timescale
= source
.get('timescale')
2742 ms_info
['timescale'] = int(timescale
)
2743 segment_duration
= source
.get('duration')
2744 if segment_duration
:
2745 ms_info
['segment_duration'] = float(segment_duration
)
2747 def extract_Initialization(source
):
2748 initialization
= source
.find(_add_ns('Initialization'))
2749 if initialization
is not None:
2750 ms_info
['initialization_url'] = initialization
.attrib
['sourceURL']
2752 segment_list
= element
.find(_add_ns('SegmentList'))
2753 if segment_list
is not None:
2754 extract_common(segment_list
)
2755 extract_Initialization(segment_list
)
2756 segment_urls_e
= segment_list
.findall(_add_ns('SegmentURL'))
2758 ms_info
['segment_urls'] = [segment
.attrib
['media'] for segment
in segment_urls_e
]
2760 segment_template
= element
.find(_add_ns('SegmentTemplate'))
2761 if segment_template
is not None:
2762 extract_common(segment_template
)
2763 media
= segment_template
.get('media')
2765 ms_info
['media'] = media
2766 initialization
= segment_template
.get('initialization')
2768 ms_info
['initialization'] = initialization
2770 extract_Initialization(segment_template
)
2773 mpd_duration
= parse_duration(mpd_doc
.get('mediaPresentationDuration'))
2774 formats
, subtitles
= [], {}
2775 stream_numbers
= collections
.defaultdict(int)
2776 for period
in mpd_doc
.findall(_add_ns('Period')):
2777 period_duration
= parse_duration(period
.get('duration')) or mpd_duration
2778 period_ms_info
= extract_multisegment_info(period
, {
2782 for adaptation_set
in period
.findall(_add_ns('AdaptationSet')):
2783 adaption_set_ms_info
= extract_multisegment_info(adaptation_set
, period_ms_info
)
2784 for representation
in adaptation_set
.findall(_add_ns('Representation')):
2785 representation_attrib
= adaptation_set
.attrib
.copy()
2786 representation_attrib
.update(representation
.attrib
)
2787 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2788 mime_type
= representation_attrib
['mimeType']
2789 content_type
= representation_attrib
.get('contentType', mime_type
.split('/')[0])
2791 codecs
= parse_codecs(representation_attrib
.get('codecs', ''))
2792 if content_type
not in ('video', 'audio', 'text'):
2793 if mime_type
== 'image/jpeg':
2794 content_type
= mime_type
2795 elif codecs
['vcodec'] != 'none':
2796 content_type
= 'video'
2797 elif codecs
['acodec'] != 'none':
2798 content_type
= 'audio'
2799 elif codecs
.get('tcodec', 'none') != 'none':
2800 content_type
= 'text'
2801 elif mimetype2ext(mime_type
) in ('tt', 'dfxp', 'ttml', 'xml', 'json'):
2802 content_type
= 'text'
2804 self
.report_warning('Unknown MIME type %s in DASH manifest' % mime_type
)
2808 for element
in (representation
, adaptation_set
, period
, mpd_doc
):
2809 base_url_e
= element
.find(_add_ns('BaseURL'))
2810 if base_url_e
is not None:
2811 base_url
= base_url_e
.text
+ base_url
2812 if re
.match(r
'^https?://', base_url
):
2814 if mpd_base_url
and base_url
.startswith('/'):
2815 base_url
= compat_urlparse
.urljoin(mpd_base_url
, base_url
)
2816 elif mpd_base_url
and not re
.match(r
'^https?://', base_url
):
2817 if not mpd_base_url
.endswith('/'):
2819 base_url
= mpd_base_url
+ base_url
2820 representation_id
= representation_attrib
.get('id')
2821 lang
= representation_attrib
.get('lang')
2822 url_el
= representation
.find(_add_ns('BaseURL'))
2823 filesize
= int_or_none(url_el
.attrib
.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el
is not None else None)
2824 bandwidth
= int_or_none(representation_attrib
.get('bandwidth'))
2825 if representation_id
is not None:
2826 format_id
= representation_id
2828 format_id
= content_type
2830 format_id
= mpd_id
+ '-' + format_id
2831 if content_type
in ('video', 'audio'):
2833 'format_id': format_id
,
2834 'manifest_url': mpd_url
,
2835 'ext': mimetype2ext(mime_type
),
2836 'width': int_or_none(representation_attrib
.get('width')),
2837 'height': int_or_none(representation_attrib
.get('height')),
2838 'tbr': float_or_none(bandwidth
, 1000),
2839 'asr': int_or_none(representation_attrib
.get('audioSamplingRate')),
2840 'fps': int_or_none(representation_attrib
.get('frameRate')),
2841 'language': lang
if lang
not in ('mul', 'und', 'zxx', 'mis') else None,
2842 'format_note': 'DASH %s' % content_type
,
2843 'filesize': filesize
,
2844 'container': mimetype2ext(mime_type
) + '_dash',
2847 elif content_type
== 'text':
2849 'ext': mimetype2ext(mime_type
),
2850 'manifest_url': mpd_url
,
2851 'filesize': filesize
,
2853 elif content_type
== 'image/jpeg':
2854 # See test case in VikiIE
2855 # https://www.viki.com/videos/1175236v-choosing-spouse-by-lottery-episode-1
2857 'format_id': format_id
,
2859 'manifest_url': mpd_url
,
2860 'format_note': 'DASH storyboards (jpeg)',
2864 if is_drm_protected(adaptation_set
) or is_drm_protected(representation
):
2866 representation_ms_info
= extract_multisegment_info(representation
, adaption_set_ms_info
)
2868 def prepare_template(template_name
, identifiers
):
2869 tmpl
= representation_ms_info
[template_name
]
2870 # First of, % characters outside $...$ templates
2871 # must be escaped by doubling for proper processing
2872 # by % operator string formatting used further (see
2873 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2879 in_template
= not in_template
2880 elif c
== '%' and not in_template
:
2882 # Next, $...$ templates are translated to their
2883 # %(...) counterparts to be used with % operator
2884 if representation_id
is not None:
2885 t
= t
.replace('$RepresentationID$', representation_id
)
2886 t
= re
.sub(r
'\$(%s)\$' % '|'.join(identifiers
), r
'%(\1)d', t
)
2887 t
= re
.sub(r
'\$(%s)%%([^$]+)\$' % '|'.join(identifiers
), r
'%(\1)\2', t
)
2888 t
.replace('$$', '$')
2891 # @initialization is a regular template like @media one
2892 # so it should be handled just the same way (see
2893 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2894 if 'initialization' in representation_ms_info
:
2895 initialization_template
= prepare_template(
2897 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2898 # $Time$ shall not be included for @initialization thus
2899 # only $Bandwidth$ remains
2901 representation_ms_info
['initialization_url'] = initialization_template
% {
2902 'Bandwidth': bandwidth
,
2905 def location_key(location
):
2906 return 'url' if re
.match(r
'^https?://', location
) else 'path'
2908 if 'segment_urls' not in representation_ms_info
and 'media' in representation_ms_info
:
2910 media_template
= prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2911 media_location_key
= location_key(media_template
)
2913 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2914 # can't be used at the same time
2915 if '%(Number' in media_template
and 's' not in representation_ms_info
:
2916 segment_duration
= None
2917 if 'total_number' not in representation_ms_info
and 'segment_duration' in representation_ms_info
:
2918 segment_duration
= float_or_none(representation_ms_info
['segment_duration'], representation_ms_info
['timescale'])
2919 representation_ms_info
['total_number'] = int(math
.ceil(
2920 float_or_none(period_duration
, segment_duration
, default
=0)))
2921 representation_ms_info
['fragments'] = [{
2922 media_location_key
: media_template
% {
2923 'Number': segment_number
,
2924 'Bandwidth': bandwidth
,
2926 'duration': segment_duration
,
2927 } for segment_number
in range(
2928 representation_ms_info
['start_number'],
2929 representation_ms_info
['total_number'] + representation_ms_info
['start_number'])]
2931 # $Number*$ or $Time$ in media template with S list available
2932 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2933 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2934 representation_ms_info
['fragments'] = []
2937 segment_number
= representation_ms_info
['start_number']
2939 def add_segment_url():
2940 segment_url
= media_template
% {
2941 'Time': segment_time
,
2942 'Bandwidth': bandwidth
,
2943 'Number': segment_number
,
2945 representation_ms_info
['fragments'].append({
2946 media_location_key
: segment_url
,
2947 'duration': float_or_none(segment_d
, representation_ms_info
['timescale']),
2950 for num
, s
in enumerate(representation_ms_info
['s']):
2951 segment_time
= s
.get('t') or segment_time
2955 for r
in range(s
.get('r', 0)):
2956 segment_time
+= segment_d
2959 segment_time
+= segment_d
2960 elif 'segment_urls' in representation_ms_info
and 's' in representation_ms_info
:
2962 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2963 # or any YouTube dashsegments video
2966 timescale
= representation_ms_info
['timescale']
2967 for s
in representation_ms_info
['s']:
2968 duration
= float_or_none(s
['d'], timescale
)
2969 for r
in range(s
.get('r', 0) + 1):
2970 segment_uri
= representation_ms_info
['segment_urls'][segment_index
]
2972 location_key(segment_uri
): segment_uri
,
2973 'duration': duration
,
2976 representation_ms_info
['fragments'] = fragments
2977 elif 'segment_urls' in representation_ms_info
:
2978 # Segment URLs with no SegmentTimeline
2979 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2980 # https://github.com/ytdl-org/youtube-dl/pull/14844
2982 segment_duration
= float_or_none(
2983 representation_ms_info
['segment_duration'],
2984 representation_ms_info
['timescale']) if 'segment_duration' in representation_ms_info
else None
2985 for segment_url
in representation_ms_info
['segment_urls']:
2987 location_key(segment_url
): segment_url
,
2989 if segment_duration
:
2990 fragment
['duration'] = segment_duration
2991 fragments
.append(fragment
)
2992 representation_ms_info
['fragments'] = fragments
2993 # If there is a fragments key available then we correctly recognized fragmented media.
2994 # Otherwise we will assume unfragmented media with direct access. Technically, such
2995 # assumption is not necessarily correct since we may simply have no support for
2996 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2997 if 'fragments' in representation_ms_info
:
2999 # NB: mpd_url may be empty when MPD manifest is parsed from a string
3000 'url': mpd_url
or base_url
,
3001 'fragment_base_url': base_url
,
3003 'protocol': 'http_dash_segments' if mime_type
!= 'image/jpeg' else 'mhtml',
3005 if 'initialization_url' in representation_ms_info
:
3006 initialization_url
= representation_ms_info
['initialization_url']
3007 if not f
.get('url'):
3008 f
['url'] = initialization_url
3009 f
['fragments'].append({location_key(initialization_url): initialization_url}
)
3010 f
['fragments'].extend(representation_ms_info
['fragments'])
3011 if not period_duration
:
3012 period_duration
= try_get(
3013 representation_ms_info
,
3014 lambda r
: sum(frag
['duration'] for frag
in r
['fragments']), float)
3016 # Assuming direct URL to unfragmented media.
3018 if content_type
in ('video', 'audio', 'image/jpeg'):
3019 f
['manifest_stream_number'] = stream_numbers
[f
['url']]
3020 stream_numbers
[f
['url']] += 1
3022 elif content_type
== 'text':
3023 subtitles
.setdefault(lang
or 'und', []).append(f
)
3025 return formats
, subtitles
3027 def _extract_ism_formats(self
, *args
, **kwargs
):
3028 fmts
, subs
= self
._extract
_ism
_formats
_and
_subtitles
(*args
, **kwargs
)
3030 self
._report
_ignoring
_subs
('ISM')
3033 def _extract_ism_formats_and_subtitles(self
, ism_url
, video_id
, ism_id
=None, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
):
3034 res
= self
._download
_xml
_handle
(
3036 note
='Downloading ISM manifest' if note
is None else note
,
3037 errnote
='Failed to download ISM manifest' if errnote
is None else errnote
,
3038 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
3045 return self
._parse
_ism
_formats
_and
_subtitles
(ism_doc
, urlh
.geturl(), ism_id
)
3047 def _parse_ism_formats_and_subtitles(self
, ism_doc
, ism_url
, ism_id
=None):
3049 Parse formats from ISM manifest.
3051 1. [MS-SSTR]: Smooth Streaming Protocol,
3052 https://msdn.microsoft.com/en-us/library/ff469518.aspx
3054 if ism_doc
.get('IsLive') == 'TRUE':
3057 duration
= int(ism_doc
.attrib
['Duration'])
3058 timescale
= int_or_none(ism_doc
.get('TimeScale')) or 10000000
3062 for stream
in ism_doc
.findall('StreamIndex'):
3063 stream_type
= stream
.get('Type')
3064 if stream_type
not in ('video', 'audio', 'text'):
3066 url_pattern
= stream
.attrib
['Url']
3067 stream_timescale
= int_or_none(stream
.get('TimeScale')) or timescale
3068 stream_name
= stream
.get('Name')
3069 stream_language
= stream
.get('Language', 'und')
3070 for track
in stream
.findall('QualityLevel'):
3071 fourcc
= track
.get('FourCC') or ('AACL' if track
.get('AudioTag') == '255' else None)
3072 # TODO: add support for WVC1 and WMAP
3073 if fourcc
not in ('H264', 'AVC1', 'AACL', 'TTML'):
3074 self
.report_warning('%s is not a supported codec' % fourcc
)
3076 tbr
= int(track
.attrib
['Bitrate']) // 1000
3077 # [1] does not mention Width and Height attributes. However,
3078 # they're often present while MaxWidth and MaxHeight are
3079 # missing, so should be used as fallbacks
3080 width
= int_or_none(track
.get('MaxWidth') or track
.get('Width'))
3081 height
= int_or_none(track
.get('MaxHeight') or track
.get('Height'))
3082 sampling_rate
= int_or_none(track
.get('SamplingRate'))
3084 track_url_pattern
= re
.sub(r
'{[Bb]itrate}', track
.attrib
['Bitrate'], url_pattern
)
3085 track_url_pattern
= compat_urlparse
.urljoin(ism_url
, track_url_pattern
)
3091 stream_fragments
= stream
.findall('c')
3092 for stream_fragment_index
, stream_fragment
in enumerate(stream_fragments
):
3093 fragment_ctx
['time'] = int_or_none(stream_fragment
.get('t')) or fragment_ctx
['time']
3094 fragment_repeat
= int_or_none(stream_fragment
.get('r')) or 1
3095 fragment_ctx
['duration'] = int_or_none(stream_fragment
.get('d'))
3096 if not fragment_ctx
['duration']:
3098 next_fragment_time
= int(stream_fragment
[stream_fragment_index
+ 1].attrib
['t'])
3100 next_fragment_time
= duration
3101 fragment_ctx
['duration'] = (next_fragment_time
- fragment_ctx
['time']) / fragment_repeat
3102 for _
in range(fragment_repeat
):
3104 'url': re
.sub(r
'{start[ _]time}', compat_str(fragment_ctx
['time']), track_url_pattern
),
3105 'duration': fragment_ctx
['duration'] / stream_timescale
,
3107 fragment_ctx
['time'] += fragment_ctx
['duration']
3109 if stream_type
== 'text':
3110 subtitles
.setdefault(stream_language
, []).append({
3114 'manifest_url': ism_url
,
3115 'fragments': fragments
,
3116 '_download_params': {
3117 'stream_type': stream_type
,
3118 'duration': duration
,
3119 'timescale': stream_timescale
,
3121 'language': stream_language
,
3122 'codec_private_data': track
.get('CodecPrivateData'),
3125 elif stream_type
in ('video', 'audio'):
3127 'format_id': join_nonempty(ism_id
, stream_name
, tbr
),
3129 'manifest_url': ism_url
,
3130 'ext': 'ismv' if stream_type
== 'video' else 'isma',
3134 'asr': sampling_rate
,
3135 'vcodec': 'none' if stream_type
== 'audio' else fourcc
,
3136 'acodec': 'none' if stream_type
== 'video' else fourcc
,
3138 'fragments': fragments
,
3139 'has_drm': ism_doc
.find('Protection') is not None,
3140 '_download_params': {
3141 'stream_type': stream_type
,
3142 'duration': duration
,
3143 'timescale': stream_timescale
,
3144 'width': width
or 0,
3145 'height': height
or 0,
3147 'language': stream_language
,
3148 'codec_private_data': track
.get('CodecPrivateData'),
3149 'sampling_rate': sampling_rate
,
3150 'channels': int_or_none(track
.get('Channels', 2)),
3151 'bits_per_sample': int_or_none(track
.get('BitsPerSample', 16)),
3152 'nal_unit_length_field': int_or_none(track
.get('NALUnitLengthField', 4)),
3155 return formats
, subtitles
3157 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):
3158 def absolute_url(item_url
):
3159 return urljoin(base_url
, item_url
)
3161 def parse_content_type(content_type
):
3162 if not content_type
:
3164 ctr
= re
.search(r
'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type
)
3166 mimetype
, codecs
= ctr
.groups()
3167 f
= parse_codecs(codecs
)
3168 f
['ext'] = mimetype2ext(mimetype
)
3172 def _media_formats(src
, cur_media_type
, type_info
={}):
3173 full_url
= absolute_url(src
)
3174 ext
= type_info
.get('ext') or determine_ext(full_url
)
3176 is_plain_url
= False
3177 formats
= self
._extract
_m
3u8_formats
(
3178 full_url
, video_id
, ext
='mp4',
3179 entry_protocol
=m3u8_entry_protocol
, m3u8_id
=m3u8_id
,
3180 preference
=preference
, quality
=quality
, fatal
=False)
3182 is_plain_url
= False
3183 formats
= self
._extract
_mpd
_formats
(
3184 full_url
, video_id
, mpd_id
=mpd_id
, fatal
=False)
3189 'vcodec': 'none' if cur_media_type
== 'audio' else None,
3191 return is_plain_url
, formats
3194 # amp-video and amp-audio are very similar to their HTML5 counterparts
3195 # so we wll include them right here (see
3196 # https://www.ampproject.org/docs/reference/components/amp-video)
3197 # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/
3198 _MEDIA_TAG_NAME_RE
= r
'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)'
3199 media_tags
= [(media_tag
, media_tag_name
, media_type
, '')
3200 for media_tag
, media_tag_name
, media_type
3201 in re
.findall(r
'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE
, webpage
)]
3202 media_tags
.extend(re
.findall(
3203 # We only allow video|audio followed by a whitespace or '>'.
3204 # Allowing more characters may end up in significant slow down (see
3205 # https://github.com/ytdl-org/youtube-dl/issues/11979, example URL:
3206 # http://www.porntrex.com/maps/videositemap.xml).
3207 r
'(?s)(<(?P<tag>%s)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>' % _MEDIA_TAG_NAME_RE
, webpage
))
3208 for media_tag
, _
, media_type
, media_content
in media_tags
:
3213 media_attributes
= extract_attributes(media_tag
)
3214 src
= strip_or_none(media_attributes
.get('src'))
3216 _
, formats
= _media_formats(src
, media_type
)
3217 media_info
['formats'].extend(formats
)
3218 media_info
['thumbnail'] = absolute_url(media_attributes
.get('poster'))
3220 for source_tag
in re
.findall(r
'<source[^>]+>', media_content
):
3221 s_attr
= extract_attributes(source_tag
)
3222 # data-video-src and data-src are non standard but seen
3223 # several times in the wild
3224 src
= strip_or_none(dict_get(s_attr
, ('src', 'data-video-src', 'data-src')))
3227 f
= parse_content_type(s_attr
.get('type'))
3228 is_plain_url
, formats
= _media_formats(src
, media_type
, f
)
3230 # width, height, res, label and title attributes are
3231 # all not standard but seen several times in the wild
3234 for lbl
in ('label', 'title')
3235 if str_or_none(s_attr
.get(lbl
))
3237 width
= int_or_none(s_attr
.get('width'))
3238 height
= (int_or_none(s_attr
.get('height'))
3239 or int_or_none(s_attr
.get('res')))
3240 if not width
or not height
:
3242 resolution
= parse_resolution(lbl
)
3245 width
= width
or resolution
.get('width')
3246 height
= height
or resolution
.get('height')
3248 tbr
= parse_bitrate(lbl
)
3257 'format_id': s_attr
.get('label') or s_attr
.get('title'),
3259 f
.update(formats
[0])
3260 media_info
['formats'].append(f
)
3262 media_info
['formats'].extend(formats
)
3263 for track_tag
in re
.findall(r
'<track[^>]+>', media_content
):
3264 track_attributes
= extract_attributes(track_tag
)
3265 kind
= track_attributes
.get('kind')
3266 if not kind
or kind
in ('subtitles', 'captions'):
3267 src
= strip_or_none(track_attributes
.get('src'))
3270 lang
= track_attributes
.get('srclang') or track_attributes
.get('lang') or track_attributes
.get('label')
3271 media_info
['subtitles'].setdefault(lang
, []).append({
3272 'url': absolute_url(src
),
3274 for f
in media_info
['formats']:
3275 f
.setdefault('http_headers', {})['Referer'] = base_url
3276 if media_info
['formats'] or media_info
['subtitles']:
3277 entries
.append(media_info
)
3280 def _extract_akamai_formats(self
, *args
, **kwargs
):
3281 fmts
, subs
= self
._extract
_akamai
_formats
_and
_subtitles
(*args
, **kwargs
)
3283 self
._report
_ignoring
_subs
('akamai')
3286 def _extract_akamai_formats_and_subtitles(self
, manifest_url
, video_id
, hosts
={}):
3287 signed
= 'hdnea=' in manifest_url
3289 # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
3290 manifest_url
= re
.sub(
3291 r
'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
3292 '', manifest_url
).strip('?')
3297 hdcore_sign
= 'hdcore=3.7.0'
3298 f4m_url
= re
.sub(r
'(https?://[^/]+)/i/', r
'\1/z/', manifest_url
).replace('/master.m3u8', '/manifest.f4m')
3299 hds_host
= hosts
.get('hds')
3301 f4m_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hds_host
, f4m_url
)
3302 if 'hdcore=' not in f4m_url
:
3303 f4m_url
+= ('&' if '?' in f4m_url
else '?') + hdcore_sign
3304 f4m_formats
= self
._extract
_f
4m
_formats
(
3305 f4m_url
, video_id
, f4m_id
='hds', fatal
=False)
3306 for entry
in f4m_formats
:
3307 entry
.update({'extra_param_to_segment_url': hdcore_sign}
)
3308 formats
.extend(f4m_formats
)
3310 m3u8_url
= re
.sub(r
'(https?://[^/]+)/z/', r
'\1/i/', manifest_url
).replace('/manifest.f4m', '/master.m3u8')
3311 hls_host
= hosts
.get('hls')
3313 m3u8_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hls_host
, m3u8_url
)
3314 m3u8_formats
, m3u8_subtitles
= self
._extract
_m
3u8_formats
_and
_subtitles
(
3315 m3u8_url
, video_id
, 'mp4', 'm3u8_native',
3316 m3u8_id
='hls', fatal
=False)
3317 formats
.extend(m3u8_formats
)
3318 subtitles
= self
._merge
_subtitles
(subtitles
, m3u8_subtitles
)
3320 http_host
= hosts
.get('http')
3321 if http_host
and m3u8_formats
and not signed
:
3322 REPL_REGEX
= r
'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
3323 qualities
= re
.match(REPL_REGEX
, m3u8_url
).group(2).split(',')
3324 qualities_length
= len(qualities
)
3325 if len(m3u8_formats
) in (qualities_length
, qualities_length
+ 1):
3327 for f
in m3u8_formats
:
3328 if f
['vcodec'] != 'none':
3329 for protocol
in ('http', 'https'):
3331 del http_f
['manifest_url']
3333 REPL_REGEX
, protocol
+ r
'://%s/\g<1>%s\3' % (http_host
, qualities
[i
]), f
['url'])
3335 'format_id': http_f
['format_id'].replace('hls-', protocol
+ '-'),
3337 'protocol': protocol
,
3339 formats
.append(http_f
)
3342 return formats
, subtitles
3344 def _extract_wowza_formats(self
, url
, video_id
, m3u8_entry_protocol
='m3u8_native', skip_protocols
=[]):
3345 query
= compat_urlparse
.urlparse(url
).query
3346 url
= re
.sub(r
'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url
)
3348 r
'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url
)
3349 url_base
= mobj
.group('url')
3350 http_base_url
= '%s%s:%s' % ('http', mobj
.group('s') or '', url_base
)
3353 def manifest_url(manifest
):
3354 m_url
= '%s/%s' % (http_base_url
, manifest
)
3356 m_url
+= '?%s' % query
3359 if 'm3u8' not in skip_protocols
:
3360 formats
.extend(self
._extract
_m
3u8_formats
(
3361 manifest_url('playlist.m3u8'), video_id
, 'mp4',
3362 m3u8_entry_protocol
, m3u8_id
='hls', fatal
=False))
3363 if 'f4m' not in skip_protocols
:
3364 formats
.extend(self
._extract
_f
4m
_formats
(
3365 manifest_url('manifest.f4m'),
3366 video_id
, f4m_id
='hds', fatal
=False))
3367 if 'dash' not in skip_protocols
:
3368 formats
.extend(self
._extract
_mpd
_formats
(
3369 manifest_url('manifest.mpd'),
3370 video_id
, mpd_id
='dash', fatal
=False))
3371 if re
.search(r
'(?:/smil:|\.smil)', url_base
):
3372 if 'smil' not in skip_protocols
:
3373 rtmp_formats
= self
._extract
_smil
_formats
(
3374 manifest_url('jwplayer.smil'),
3375 video_id
, fatal
=False)
3376 for rtmp_format
in rtmp_formats
:
3377 rtsp_format
= rtmp_format
.copy()
3378 rtsp_format
['url'] = '%s/%s' % (rtmp_format
['url'], rtmp_format
['play_path'])
3379 del rtsp_format
['play_path']
3380 del rtsp_format
['ext']
3381 rtsp_format
.update({
3382 'url': rtsp_format
['url'].replace('rtmp://', 'rtsp://'),
3383 'format_id': rtmp_format
['format_id'].replace('rtmp', 'rtsp'),
3386 formats
.extend([rtmp_format
, rtsp_format
])
3388 for protocol
in ('rtmp', 'rtsp'):
3389 if protocol
not in skip_protocols
:
3391 'url': '%s:%s' % (protocol
, url_base
),
3392 'format_id': protocol
,
3393 'protocol': protocol
,
3397 def _find_jwplayer_data(self
, webpage
, video_id
=None, transform_source
=js_to_json
):
3399 r
'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P
=quote
)\
)(?
!</script
>).*?\
.setup\s
*\
((?P
<options
>[^
)]+)\
)',
3403 jwplayer_data = self._parse_json(mobj.group('options
'),
3405 transform_source=transform_source)
3406 except ExtractorError:
3409 if isinstance(jwplayer_data, dict):
3410 return jwplayer_data
3412 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
3413 jwplayer_data = self._find_jwplayer_data(
3414 webpage, video_id, transform_source=js_to_json)
3415 return self._parse_jwplayer_data(
3416 jwplayer_data, video_id, *args, **kwargs)
3418 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
3419 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3420 # JWPlayer backward compatibility: flattened playlists
3421 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
3422 if 'playlist
' not in jwplayer_data:
3423 jwplayer_data = {'playlist': [jwplayer_data]}
3427 # JWPlayer backward compatibility: single playlist item
3428 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
3429 if not isinstance(jwplayer_data['playlist
'], list):
3430 jwplayer_data['playlist
'] = [jwplayer_data['playlist
']]
3432 for video_data in jwplayer_data['playlist
']:
3433 # JWPlayer backward compatibility: flattened sources
3434 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
3435 if 'sources
' not in video_data:
3436 video_data['sources
'] = [video_data]
3438 this_video_id = video_id or video_data['mediaid
']
3440 formats = self._parse_jwplayer_formats(
3441 video_data['sources
'], video_id=this_video_id, m3u8_id=m3u8_id,
3442 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
3445 tracks = video_data.get('tracks
')
3446 if tracks and isinstance(tracks, list):
3447 for track in tracks:
3448 if not isinstance(track, dict):
3450 track_kind = track.get('kind
')
3451 if not track_kind or not isinstance(track_kind, compat_str):
3453 if track_kind.lower() not in ('captions
', 'subtitles
'):
3455 track_url = urljoin(base_url, track.get('file'))
3458 subtitles.setdefault(track.get('label
') or 'en
', []).append({
3459 'url
': self._proto_relative_url(track_url)
3463 'id': this_video_id,
3464 'title
': unescapeHTML(video_data['title
'] if require_title else video_data.get('title
')),
3465 'description
': clean_html(video_data.get('description
')),
3466 'thumbnail
': urljoin(base_url, self._proto_relative_url(video_data.get('image
'))),
3467 'timestamp
': int_or_none(video_data.get('pubdate
')),
3468 'duration
': float_or_none(jwplayer_data.get('duration
') or video_data.get('duration
')),
3469 'subtitles
': subtitles,
3471 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
3472 if len(formats) == 1 and re.search(r'^
(?
:http|
//).*(?
:youtube\
.com|youtu\
.be
)/.+', formats[0]['url
']):
3474 '_type
': 'url_transparent
',
3475 'url
': formats[0]['url
'],
3478 self._sort_formats(formats)
3479 entry['formats
'] = formats
3480 entries.append(entry)
3481 if len(entries) == 1:
3484 return self.playlist_result(entries)
3486 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
3487 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3490 for source in jwplayer_sources_data:
3491 if not isinstance(source, dict):
3493 source_url = urljoin(
3494 base_url, self._proto_relative_url(source.get('file')))
3495 if not source_url or source_url in urls:
3497 urls.append(source_url)
3498 source_type = source.get('type') or ''
3499 ext = mimetype2ext(source_type) or determine_ext(source_url)
3500 if source_type == 'hls
' or ext == 'm3u8
':
3501 formats.extend(self._extract_m3u8_formats(
3502 source_url, video_id, 'mp4
', entry_protocol='m3u8_native
',
3503 m3u8_id=m3u8_id, fatal=False))
3504 elif source_type == 'dash
' or ext == 'mpd
':
3505 formats.extend(self._extract_mpd_formats(
3506 source_url, video_id, mpd_id=mpd_id, fatal=False))
3508 formats.extend(self._extract_smil_formats(
3509 source_url, video_id, fatal=False))
3510 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3511 elif source_type.startswith('audio
') or ext in (
3512 'oga
', 'aac
', 'mp3
', 'mpeg
', 'vorbis
'):
3519 height = int_or_none(source.get('height
'))
3521 # Often no height is provided but there is a label in
3522 # format like "1080p", "720p SD", or 1080.
3523 height = int_or_none(self._search_regex(
3524 r'^
(\d{3,4}
)[pP
]?
(?
:\b|$
)', compat_str(source.get('label
') or ''),
3525 'height
', default=None))
3528 'width
': int_or_none(source.get('width
')),
3530 'tbr
': int_or_none(source.get('bitrate
')),
3533 if source_url.startswith('rtmp
'):
3534 a_format['ext
'] = 'flv
'
3535 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3536 # of jwplayer.flash.swf
3537 rtmp_url_parts = re.split(
3538 r'((?
:mp4|mp3|flv
):)', source_url, 1)
3539 if len(rtmp_url_parts) == 3:
3540 rtmp_url, prefix, play_path = rtmp_url_parts
3543 'play_path
': prefix + play_path,
3546 a_format.update(rtmp_params)
3547 formats.append(a_format)
3550 def _live_title(self, name):
3551 self._downloader.deprecation_warning('yt_dlp
.InfoExtractor
._live
_title
is deprecated
and does
not work
as expected
')
3554 def _int(self, v, name, fatal=False, **kwargs):
3555 res = int_or_none(v, **kwargs)
3557 msg = 'Failed to extract
%s: Could
not parse value
%r' % (name, v)
3559 raise ExtractorError(msg)
3561 self.report_warning(msg)
3564 def _float(self, v, name, fatal=False, **kwargs):
3565 res = float_or_none(v, **kwargs)
3567 msg = 'Failed to extract
%s: Could
not parse value
%r' % (name, v)
3569 raise ExtractorError(msg)
3571 self.report_warning(msg)
3574 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
3575 path='/', secure=False, discard=False, rest={}, **kwargs):
3576 cookie = compat_cookiejar_Cookie(
3577 0, name, value, port, port is not None, domain, True,
3578 domain.startswith('.'), path, True, secure, expire_time,
3579 discard, None, None, rest)
3580 self._downloader.cookiejar.set_cookie(cookie)
3582 def _get_cookies(self, url):
3583 """ Return a compat_cookies_SimpleCookie with the cookies for the url """
3584 req = sanitized_Request(url)
3585 self._downloader.cookiejar.add_cookie_header(req)
3586 return compat_cookies_SimpleCookie(req.get_header('Cookie
'))
3588 def _apply_first_set_cookie_header(self, url_handle, cookie):
3590 Apply first Set-Cookie header instead of the last. Experimental.
3592 Some sites (e.g. [1-3]) may serve two cookies under the same name
3593 in Set-Cookie header and expect the first (old) one to be set rather
3594 than second (new). However, as of RFC6265 the newer one cookie
3595 should be set into cookie store what actually happens.
3596 We will workaround this issue by resetting the cookie to
3597 the first one manually.
3598 1. https://new.vk.com/
3599 2. https://github.com/ytdl-org/youtube-dl/issues/9841#issuecomment-227871201
3600 3. https://learning.oreilly.com/
3602 for header, cookies in url_handle.headers.items():
3603 if header.lower() != 'set-cookie
':
3605 if sys.version_info[0] >= 3:
3606 cookies = cookies.encode('iso
-8859-1')
3607 cookies = cookies.decode('utf
-8')
3608 cookie_value = re.search(
3609 r'%s=(.+?
);.*?
\b[Dd
]omain
=(.+?
)(?
:[,;]|$
)' % cookie, cookies)
3611 value, domain = cookie_value.groups()
3612 self._set_cookie(domain, cookie, value)
3615 def get_testcases(self, include_onlymatching=False):
3616 t = getattr(self, '_TEST
', None)
3618 assert not hasattr(self, '_TESTS
'), \
3619 '%s has _TEST
and _TESTS
' % type(self).__name__
3622 tests = getattr(self, '_TESTS
', [])
3624 if not include_onlymatching and t.get('only_matching
', False):
3626 t['name
'] = type(self).__name__[:-len('IE
')]
3629 def is_suitable(self, age_limit):
3630 """ Test whether the extractor is generally suitable for the given
3631 age limit (i.e. pornographic sites are not, all others usually are) """
3633 any_restricted = False
3634 for tc in self.get_testcases(include_onlymatching=False):
3635 if tc.get('playlist
', []):
3636 tc = tc['playlist
'][0]
3637 is_restricted = age_restricted(
3638 tc.get('info_dict
', {}).get('age_limit
'), age_limit)
3639 if not is_restricted:
3641 any_restricted = any_restricted or is_restricted
3642 return not any_restricted
3644 def extract_subtitles(self, *args, **kwargs):
3645 if (self.get_param('writesubtitles
', False)
3646 or self.get_param('listsubtitles
')):
3647 return self._get_subtitles(*args, **kwargs)
3650 def _get_subtitles(self, *args, **kwargs):
3651 raise NotImplementedError('This method must be implemented by subclasses
')
3653 def extract_comments(self, *args, **kwargs):
3654 if not self.get_param('getcomments
'):
3656 generator = self._get_comments(*args, **kwargs)
3663 comments.append(next(generator))
3664 except StopIteration:
3666 except KeyboardInterrupt:
3667 self.to_screen('Interrupted by user
')
3668 except Exception as e:
3669 if self.get_param('ignoreerrors
') is not True:
3671 self._downloader.report_error(e)
3672 comment_count = len(comments)
3673 self.to_screen(f'Extracted {comment_count} comments
')
3675 'comments
': comments,
3676 'comment_count
': None if interrupted else comment_count
3680 def _get_comments(self, *args, **kwargs):
3681 raise NotImplementedError('This method must be implemented by subclasses
')
3684 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
3685 """ Merge subtitle items for one language. Items with duplicated URLs/data
3686 will be dropped. """
3687 list1_data = set([item.get('url
') or item['data
'] for item in subtitle_list1])
3688 ret = list(subtitle_list1)
3689 ret.extend([item for item in subtitle_list2 if (item.get('url
') or item['data
']) not in list1_data])
3693 def _merge_subtitles(cls, *dicts, target=None):
3694 """ Merge subtitle dictionaries, language by language. """
3698 for lang, subs in d.items():
3699 target[lang] = cls._merge_subtitle_items(target.get(lang, []), subs)
3702 def extract_automatic_captions(self, *args, **kwargs):
3703 if (self.get_param('writeautomaticsub
', False)
3704 or self.get_param('listsubtitles
')):
3705 return self._get_automatic_captions(*args, **kwargs)
3708 def _get_automatic_captions(self, *args, **kwargs):
3709 raise NotImplementedError('This method must be implemented by subclasses
')
3711 def mark_watched(self, *args, **kwargs):
3712 if not self.get_param('mark_watched
', False):
3714 if (self.supports_login() and self._get_login_info()[0] is not None
3715 or self.get_param('cookiefile
') or self.get_param('cookiesfrombrowser
')):
3716 self._mark_watched(*args, **kwargs)
3718 def _mark_watched(self, *args, **kwargs):
3719 raise NotImplementedError('This method must be implemented by subclasses
')
3721 def geo_verification_headers(self):
3723 geo_verification_proxy = self.get_param('geo_verification_proxy
')
3724 if geo_verification_proxy:
3725 headers['Ytdl
-request
-proxy
'] = geo_verification_proxy
3728 def _generic_id(self, url):
3729 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
3731 def _generic_title(self, url):
3732 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
3735 def _availability(is_private=None, needs_premium=None, needs_subscription=None, needs_auth=None, is_unlisted=None):
3736 all_known = all(map(
3737 lambda x: x is not None,
3738 (is_private, needs_premium, needs_subscription, needs_auth, is_unlisted)))
3740 'private
' if is_private
3741 else 'premium_only
' if needs_premium
3742 else 'subscriber_only
' if needs_subscription
3743 else 'needs_auth
' if needs_auth
3744 else 'unlisted
' if is_unlisted
3745 else 'public
' if all_known
3748 def _configuration_arg(self, key, default=NO_DEFAULT, *, ie_key=None, casesense=False):
3750 @returns A list of values for the extractor argument given by "key"
3751 or "default" if no such key is present
3752 @param default The default value to return when the key is not present (default: [])
3753 @param casesense When false, the values are converted to lower case
3756 self._downloader.params, ('extractor_args
', (ie_key or self.ie_key()).lower(), key))
3758 return [] if default is NO_DEFAULT else default
3759 return list(val) if casesense else [x.lower() for x in val]
3761 def _yes_playlist(self, playlist_id, video_id, smuggled_data=None, *, playlist_label='playlist
', video_label='video
'):
3762 if not playlist_id or not video_id:
3765 no_playlist = (smuggled_data or {}).get('force_noplaylist
')
3766 if no_playlist is not None:
3767 return not no_playlist
3769 video_id = '' if video_id is True else f' {video_id}
'
3770 playlist_id = '' if playlist_id is True else f' {playlist_id}
'
3771 if self.get_param('noplaylist
'):
3772 self.to_screen(f'Downloading just the {video_label}{video_id} because of
--no
-playlist
')
3774 self.to_screen(f'Downloading {playlist_label}{playlist_id}
- add
--no
-playlist to download just the {video_label}{video_id}
')
3778 class SearchInfoExtractor(InfoExtractor):
3780 Base class for paged search queries extractors.
3781 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
3782 Instances should define _SEARCH_KEY and optionally _MAX_RESULTS
3785 _MAX_RESULTS = float('inf
')
3788 def _make_valid_url(cls):
3789 return r'%s(?P
<prefix
>|
[1-9][0-9]*|all
):(?P
<query
>[\s\S
]+)' % cls._SEARCH_KEY
3791 def _real_extract(self, query):
3792 prefix, query = self._match_valid_url(query).group('prefix
', 'query
')
3794 return self._get_n_results(query, 1)
3795 elif prefix == 'all
':
3796 return self._get_n_results(query, self._MAX_RESULTS)
3800 raise ExtractorError('invalid download number
%s for query
"%s"' % (n, query))
3801 elif n > self._MAX_RESULTS:
3802 self.report_warning('%s returns
max %i results (you requested
%i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3803 n = self._MAX_RESULTS
3804 return self._get_n_results(query, n)
3806 def _get_n_results(self, query, n):
3807 """Get a specified number of results for a query.
3808 Either this function or _search_results must be overridden by subclasses """
3809 return self.playlist_result(
3810 itertools.islice(self._search_results(query), 0, None if n == float('inf
') else n),
3813 def _search_results(self, query):
3814 """Returns an iterator of search results"""
3815 raise NotImplementedError('This method must be implemented by subclasses
')
3818 def SEARCH_KEY(self):
3819 return self._SEARCH_KEY