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
,
70 parse_m3u8_attributes
,
96 class InfoExtractor(object):
97 """Information Extractor class.
99 Information extractors are the classes that, given a URL, extract
100 information about the video (or videos) the URL refers to. This
101 information includes the real video URL, the video title, author and
102 others. The information is stored in a dictionary which is then
103 passed to the YoutubeDL. The YoutubeDL processes this
104 information possibly downloading the video to the file system, among
105 other possible outcomes.
107 The type field determines the type of the result.
108 By far the most common value (and the default if _type is missing) is
109 "video", which indicates a single video.
111 For a video, the dictionaries must include the following fields:
113 id: Video identifier.
114 title: Video title, unescaped.
116 Additionally, it must contain either a formats entry or a url one:
118 formats: A list of dictionaries for each format available, ordered
119 from worst to best quality.
122 * url The mandatory URL representing the media:
123 for plain file media - HTTP URL of this file,
125 for HLS - URL of the M3U8 media playlist,
126 for HDS - URL of the F4M manifest,
128 - HTTP URL to plain file media (in case of
130 - URL of the MPD manifest or base URL
131 representing the media if MPD manifest
132 is parsed from a string (in case of
134 for MSS - URL of the ISM manifest.
136 The URL of the manifest file in case of
138 for HLS - URL of the M3U8 master playlist,
139 for HDS - URL of the F4M manifest,
140 for DASH - URL of the MPD manifest,
141 for MSS - URL of the ISM manifest.
142 * manifest_stream_number (For internal use only)
143 The index of the stream in the manifest file
144 * ext Will be calculated from URL if missing
145 * format A human-readable description of the format
146 ("mp4 container with h264/opus").
147 Calculated from the format_id, width, height.
148 and format_note fields if missing.
149 * format_id A short description of the format
150 ("mp4_h264_opus" or "19").
151 Technically optional, but strongly recommended.
152 * format_note Additional info about the format
153 ("3D" or "DASH video")
154 * width Width of the video, if known
155 * height Height of the video, if known
156 * resolution Textual description of width and height
157 * dynamic_range The dynamic range of the video. One of:
158 "SDR" (None), "HDR10", "HDR10+, "HDR12", "HLG, "DV"
159 * tbr Average bitrate of audio and video in KBit/s
160 * abr Average audio bitrate in KBit/s
161 * acodec Name of the audio codec in use
162 * asr Audio sampling rate in Hertz
163 * vbr Average video bitrate in KBit/s
165 * vcodec Name of the video codec in use
166 * container Name of the container format
167 * filesize The number of bytes, if known in advance
168 * filesize_approx An estimate for the number of bytes
169 * player_url SWF Player URL (used for rtmpdump).
170 * protocol The protocol that will be used for the actual
171 download, lower-case. One of "http", "https" or
172 one of the protocols defined in downloader.PROTOCOL_MAP
174 Base URL for fragments. Each fragment's path
175 value (if present) will be relative to
177 * fragments A list of fragments of a fragmented media.
178 Each fragment entry must contain either an url
179 or a path. If an url is present it should be
180 considered by a client. Otherwise both path and
181 fragment_base_url must be present. Here is
182 the list of all potential fields:
183 * "url" - fragment's URL
184 * "path" - fragment's path relative to
186 * "duration" (optional, int or float)
187 * "filesize" (optional, int)
188 * is_from_start Is a live format that can be downloaded
189 from the start. Boolean
190 * preference Order number of this format. If this field is
191 present and not None, the formats get sorted
192 by this field, regardless of all other values.
193 -1 for default (order by other properties),
194 -2 or smaller for less than default.
195 < -1000 to hide the format (if there is
196 another one which is strictly better)
197 * language Language code, e.g. "de" or "en-US".
198 * language_preference Is this in the language mentioned in
200 10 if it's what the URL is about,
201 -1 for default (don't know),
202 -10 otherwise, other values reserved for now.
203 * quality Order number of the video quality of this
204 format, irrespective of the file format.
205 -1 for default (order by other properties),
206 -2 or smaller for less than default.
207 * source_preference Order number for this video source
208 (quality takes higher priority)
209 -1 for default (order by other properties),
210 -2 or smaller for less than default.
211 * http_headers A dictionary of additional HTTP headers
212 to add to the request.
213 * stretched_ratio If given and not 1, indicates that the
214 video's pixels are not square.
215 width : height ratio as float.
216 * no_resume The server does not support resuming the
217 (HTTP or RTMP) download. Boolean.
218 * has_drm The format has DRM and cannot be downloaded. Boolean
219 * downloader_options A dictionary of downloader options as
220 described in FileDownloader (For internal use only)
221 RTMP formats can also have the additional fields: page_url,
222 app, play_path, tc_url, flash_version, rtmp_live, rtmp_conn,
223 rtmp_protocol, rtmp_real_time
225 url: Final video URL.
226 ext: Video filename extension.
227 format: The video format, defaults to ext (used for --get-format)
228 player_url: SWF Player URL (used for rtmpdump).
230 The following fields are optional:
232 direct: True if a direct video file was given (must only be set by GenericIE)
233 alt_title: A secondary title of the video.
234 display_id An alternative identifier for the video, not necessarily
235 unique, but available before title. Typically, id is
236 something like "4234987", title "Dancing naked mole rats",
237 and display_id "dancing-naked-mole-rats"
238 thumbnails: A list of dictionaries, with the following entries:
239 * "id" (optional, string) - Thumbnail format ID
241 * "preference" (optional, int) - quality of the image
242 * "width" (optional, int)
243 * "height" (optional, int)
244 * "resolution" (optional, string "{width}x{height}",
246 * "filesize" (optional, int)
247 * "http_headers" (dict) - HTTP headers for the request
248 thumbnail: Full URL to a video thumbnail image.
249 description: Full video description.
250 uploader: Full name of the video uploader.
251 license: License name the video is licensed under.
252 creator: The creator of the video.
253 timestamp: UNIX timestamp of the moment the video was uploaded
254 upload_date: Video upload date in UTC (YYYYMMDD).
255 If not explicitly set, calculated from timestamp
256 release_timestamp: UNIX timestamp of the moment the video was released.
257 If it is not clear whether to use timestamp or this, use the former
258 release_date: The date (YYYYMMDD) when the video was released in UTC.
259 If not explicitly set, calculated from release_timestamp
260 modified_timestamp: UNIX timestamp of the moment the video was last modified.
261 modified_date: The date (YYYYMMDD) when the video was last modified in UTC.
262 If not explicitly set, calculated from modified_timestamp
263 uploader_id: Nickname or id of the video uploader.
264 uploader_url: Full URL to a personal webpage of the video uploader.
265 channel: Full name of the channel the video is uploaded on.
266 Note that channel fields may or may not repeat uploader
267 fields. This depends on a particular extractor.
268 channel_id: Id of the channel.
269 channel_url: Full URL to a channel webpage.
270 channel_follower_count: Number of followers of the channel.
271 location: Physical location where the video was filmed.
272 subtitles: The available subtitles as a dictionary in the format
273 {tag: subformats}. "tag" is usually a language code, and
274 "subformats" is a list sorted from lower to higher
275 preference, each element is a dictionary with the "ext"
277 * "data": The subtitles file contents
278 * "url": A URL pointing to the subtitles file
279 It can optionally also have:
280 * "name": Name or description of the subtitles
281 * "http_headers": A dictionary of additional HTTP headers
282 to add to the request.
283 "ext" will be calculated from URL if missing
284 automatic_captions: Like 'subtitles'; contains automatically generated
285 captions instead of normal subtitles
286 duration: Length of the video in seconds, as an integer or float.
287 view_count: How many users have watched the video on the platform.
288 like_count: Number of positive ratings of the video
289 dislike_count: Number of negative ratings of the video
290 repost_count: Number of reposts of the video
291 average_rating: Average rating give by users, the scale used depends on the webpage
292 comment_count: Number of comments on the video
293 comments: A list of comments, each with one or more of the following
294 properties (all but one of text or html optional):
295 * "author" - human-readable name of the comment author
296 * "author_id" - user ID of the comment author
297 * "author_thumbnail" - The thumbnail of the comment author
299 * "html" - Comment as HTML
300 * "text" - Plain text of the comment
301 * "timestamp" - UNIX timestamp of comment
302 * "parent" - ID of the comment this one is replying to.
303 Set to "root" to indicate that this is a
304 comment to the original video.
305 * "like_count" - Number of positive ratings of the comment
306 * "dislike_count" - Number of negative ratings of the comment
307 * "is_favorited" - Whether the comment is marked as
308 favorite by the video uploader
309 * "author_is_uploader" - Whether the comment is made by
311 age_limit: Age restriction for the video, as an integer (years)
312 webpage_url: The URL to the video webpage, if given to yt-dlp it
313 should allow to get the same result again. (It will be set
314 by YoutubeDL if it's missing)
315 categories: A list of categories that the video falls in, for example
317 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
318 cast: A list of the video cast
319 is_live: True, False, or None (=unknown). Whether this video is a
320 live stream that goes on instead of a fixed-length video.
321 was_live: True, False, or None (=unknown). Whether this video was
322 originally a live stream.
323 live_status: 'is_live', 'is_upcoming', 'was_live', 'not_live' or None (=unknown)
324 If absent, automatically set from is_live, was_live
325 start_time: Time in seconds where the reproduction should start, as
326 specified in the URL.
327 end_time: Time in seconds where the reproduction should end, as
328 specified in the URL.
329 chapters: A list of dictionaries, with the following entries:
330 * "start_time" - The start time of the chapter in seconds
331 * "end_time" - The end time of the chapter in seconds
332 * "title" (optional, string)
333 playable_in_embed: Whether this video is allowed to play in embedded
334 players on other sites. Can be True (=always allowed),
335 False (=never allowed), None (=unknown), or a string
336 specifying the criteria for embedability (Eg: 'whitelist')
337 availability: Under what condition the video is available. One of
338 'private', 'premium_only', 'subscriber_only', 'needs_auth',
339 'unlisted' or 'public'. Use 'InfoExtractor._availability'
341 __post_extractor: A function to be called just before the metadata is
342 written to either disk, logger or console. The function
343 must return a dict which will be added to the info_dict.
344 This is usefull for additional information that is
345 time-consuming to extract. Note that the fields thus
346 extracted will not be available to output template and
347 match_filter. So, only "comments" and "comment_count" are
348 currently allowed to be extracted via this method.
350 The following fields should only be used when the video belongs to some logical
353 chapter: Name or title of the chapter the video belongs to.
354 chapter_number: Number of the chapter the video belongs to, as an integer.
355 chapter_id: Id of the chapter the video belongs to, as a unicode string.
357 The following fields should only be used when the video is an episode of some
358 series, programme or podcast:
360 series: Title of the series or programme the video episode belongs to.
361 series_id: Id of the series or programme the video episode belongs to, as a unicode string.
362 season: Title of the season the video episode belongs to.
363 season_number: Number of the season the video episode belongs to, as an integer.
364 season_id: Id of the season the video episode belongs to, as a unicode string.
365 episode: Title of the video episode. Unlike mandatory video title field,
366 this field should denote the exact title of the video episode
367 without any kind of decoration.
368 episode_number: Number of the video episode within a season, as an integer.
369 episode_id: Id of the video episode, as a unicode string.
371 The following fields should only be used when the media is a track or a part of
374 track: Title of the track.
375 track_number: Number of the track within an album or a disc, as an integer.
376 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
378 artist: Artist(s) of the track.
379 genre: Genre(s) of the track.
380 album: Title of the album the track belongs to.
381 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
382 album_artist: List of all artists appeared on the album (e.g.
383 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
385 disc_number: Number of the disc or other physical medium the track belongs to,
387 release_year: Year (YYYY) when the album was released.
388 composer: Composer of the piece
390 Unless mentioned otherwise, the fields should be Unicode strings.
392 Unless mentioned otherwise, None is equivalent to absence of information.
395 _type "playlist" indicates multiple videos.
396 There must be a key "entries", which is a list, an iterable, or a PagedList
397 object, each element of which is a valid dictionary by this specification.
399 Additionally, playlists can have "id", "title", and any other relevent
400 attributes with the same semantics as videos (see above).
402 It can also have the following optional fields:
404 playlist_count: The total number of videos in a playlist. If not given,
405 YoutubeDL tries to calculate it from "entries"
408 _type "multi_video" indicates that there are multiple videos that
409 form a single show, for examples multiple acts of an opera or TV episode.
410 It must have an entries key like a playlist and contain all the keys
411 required for a video at the same time.
414 _type "url" indicates that the video must be extracted from another
415 location, possibly by a different extractor. Its only required key is:
416 "url" - the next URL to extract.
417 The key "ie_key" can be set to the class name (minus the trailing "IE",
418 e.g. "Youtube") if the extractor class is known in advance.
419 Additionally, the dictionary may have any properties of the resolved entity
420 known in advance, for example "title" if the title of the referred video is
424 _type "url_transparent" entities have the same specification as "url", but
425 indicate that the given additional information is more precise than the one
426 associated with the resolved URL.
427 This is useful when a site employs a video service that hosts the video and
428 its technical metadata, but that video service does not embed a useful
429 title, description etc.
432 Subclasses of this should define a _VALID_URL regexp and, re-define the
433 _real_extract() and (optionally) _real_initialize() methods.
434 Probably, they should also be added to the list of extractors.
436 Subclasses may also override suitable() if necessary, but ensure the function
437 signature is preserved and that this function imports everything it needs
438 (except other extractors), so that lazy_extractors works correctly.
440 To support username + password (or netrc) login, the extractor must define a
441 _NETRC_MACHINE and re-define _perform_login(username, password) and
442 (optionally) _initialize_pre_login() methods. The _perform_login method will
443 be called between _initialize_pre_login and _real_initialize if credentials
444 are passed by the user. In cases where it is necessary to have the login
445 process as part of the extraction rather than initialization, _perform_login
446 can be left undefined.
448 _GEO_BYPASS attribute may be set to False in order to disable
449 geo restriction bypass mechanisms for a particular extractor.
450 Though it won't disable explicit geo restriction bypass based on
451 country code provided with geo_bypass_country.
453 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
454 countries for this extractor. One of these countries will be used by
455 geo restriction bypass mechanism right away in order to bypass
456 geo restriction, of course, if the mechanism is not disabled.
458 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
459 IP blocks in CIDR notation for this extractor. One of these IP blocks
460 will be used by geo restriction bypass mechanism similarly
463 The _WORKING attribute should be set to False for broken IEs
464 in order to warn the users and skip the tests.
469 _x_forwarded_for_ip
= None
471 _GEO_COUNTRIES
= None
472 _GEO_IP_BLOCKS
= None
474 _NETRC_MACHINE
= None
478 'any': 'Use --cookies, --cookies-from-browser, --username and --password, or --netrc to provide account credentials',
480 'Use --cookies-from-browser or --cookies for the authentication. '
481 'See https://github.com/ytdl-org/youtube-dl#how-do-i-pass-cookies-to-youtube-dl for how to manually pass cookies'),
482 'password': 'Use --username and --password, or --netrc to provide account credentials',
485 def __init__(self
, downloader
=None):
486 """Constructor. Receives an optional downloader (a YoutubeDL instance).
487 If a downloader is not passed during initialization,
488 it must be set using "set_downloader()" before "extract()" is called"""
490 self
._x
_forwarded
_for
_ip
= None
491 self
._printed
_messages
= set()
492 self
.set_downloader(downloader
)
495 def _match_valid_url(cls
, url
):
496 # This does not use has/getattr intentionally - we want to know whether
497 # we have cached the regexp for *this* class, whereas getattr would also
498 # match the superclass
499 if '_VALID_URL_RE' not in cls
.__dict
__:
500 if '_VALID_URL' not in cls
.__dict
__:
501 cls
._VALID
_URL
= cls
._make
_valid
_url
()
502 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
503 return cls
._VALID
_URL
_RE
.match(url
)
506 def suitable(cls
, url
):
507 """Receives a URL and returns True if suitable for this IE."""
508 # This function must import everything it needs (except other extractors),
509 # so that lazy_extractors works correctly
510 return cls
._match
_valid
_url
(url
) is not None
513 def _match_id(cls
, url
):
514 return cls
._match
_valid
_url
(url
).group('id')
517 def get_temp_id(cls
, url
):
519 return cls
._match
_id
(url
)
520 except (IndexError, AttributeError):
525 """Getter method for _WORKING."""
529 def supports_login(cls
):
530 return bool(cls
._NETRC
_MACHINE
)
532 def initialize(self
):
533 """Initializes an instance (authentication, etc)."""
534 self
._printed
_messages
= set()
535 self
._initialize
_geo
_bypass
({
536 'countries': self
._GEO
_COUNTRIES
,
537 'ip_blocks': self
._GEO
_IP
_BLOCKS
,
540 self
._initialize
_pre
_login
()
541 if self
.supports_login():
542 username
, password
= self
._get
_login
_info
()
544 self
._perform
_login
(username
, password
)
545 elif self
.get_param('username') and False not in (self
.IE_DESC
, self
._NETRC
_MACHINE
):
546 self
.report_warning(f
'Login with password is not supported for this website. {self._LOGIN_HINTS["cookies"]}')
547 self
._real
_initialize
()
550 def _initialize_geo_bypass(self
, geo_bypass_context
):
552 Initialize geo restriction bypass mechanism.
554 This method is used to initialize geo bypass mechanism based on faking
555 X-Forwarded-For HTTP header. A random country from provided country list
556 is selected and a random IP belonging to this country is generated. This
557 IP will be passed as X-Forwarded-For HTTP header in all subsequent
560 This method will be used for initial geo bypass mechanism initialization
561 during the instance initialization with _GEO_COUNTRIES and
564 You may also manually call it from extractor's code if geo bypass
565 information is not available beforehand (e.g. obtained during
566 extraction) or due to some other reason. In this case you should pass
567 this information in geo bypass context passed as first argument. It may
568 contain following fields:
570 countries: List of geo unrestricted countries (similar
572 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
573 (similar to _GEO_IP_BLOCKS)
576 if not self
._x
_forwarded
_for
_ip
:
578 # Geo bypass mechanism is explicitly disabled by user
579 if not self
.get_param('geo_bypass', True):
582 if not geo_bypass_context
:
583 geo_bypass_context
= {}
585 # Backward compatibility: previously _initialize_geo_bypass
586 # expected a list of countries, some 3rd party code may still use
588 if isinstance(geo_bypass_context
, (list, tuple)):
589 geo_bypass_context
= {
590 'countries': geo_bypass_context
,
593 # The whole point of geo bypass mechanism is to fake IP
594 # as X-Forwarded-For HTTP header based on some IP block or
597 # Path 1: bypassing based on IP block in CIDR notation
599 # Explicit IP block specified by user, use it right away
600 # regardless of whether extractor is geo bypassable or not
601 ip_block
= self
.get_param('geo_bypass_ip_block', None)
603 # Otherwise use random IP block from geo bypass context but only
604 # if extractor is known as geo bypassable
606 ip_blocks
= geo_bypass_context
.get('ip_blocks')
607 if self
._GEO
_BYPASS
and ip_blocks
:
608 ip_block
= random
.choice(ip_blocks
)
611 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(ip_block
)
612 self
._downloader
.write_debug(
613 '[debug] Using fake IP %s as X-Forwarded-For' % self
._x
_forwarded
_for
_ip
)
616 # Path 2: bypassing based on country code
618 # Explicit country code specified by user, use it right away
619 # regardless of whether extractor is geo bypassable or not
620 country
= self
.get_param('geo_bypass_country', None)
622 # Otherwise use random country code from geo bypass context but
623 # only if extractor is known as geo bypassable
625 countries
= geo_bypass_context
.get('countries')
626 if self
._GEO
_BYPASS
and countries
:
627 country
= random
.choice(countries
)
630 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country
)
631 self
._downloader
.write_debug(
632 'Using fake IP %s (%s) as X-Forwarded-For' % (self
._x
_forwarded
_for
_ip
, country
.upper()))
634 def extract(self
, url
):
635 """Extracts URL information and returns it in list of dicts."""
640 self
.write_debug('Extracting URL: %s' % url
)
641 ie_result
= self
._real
_extract
(url
)
642 if ie_result
is None:
644 if self
._x
_forwarded
_for
_ip
:
645 ie_result
['__x_forwarded_for_ip'] = self
._x
_forwarded
_for
_ip
646 subtitles
= ie_result
.get('subtitles')
647 if (subtitles
and 'live_chat' in subtitles
648 and 'no-live-chat' in self
.get_param('compat_opts', [])):
649 del subtitles
['live_chat']
651 except GeoRestrictedError
as e
:
652 if self
.__maybe
_fake
_ip
_and
_retry
(e
.countries
):
655 except UnsupportedError
:
657 except ExtractorError
as e
:
659 'video_id': e
.video_id
or self
.get_temp_id(url
),
661 'tb': e
.traceback
or sys
.exc_info()[2],
662 'expected': e
.expected
,
665 if hasattr(e
, 'countries'):
666 kwargs
['countries'] = e
.countries
667 raise type(e
)(e
.orig_msg
, **kwargs
)
668 except compat_http_client
.IncompleteRead
as e
:
669 raise ExtractorError('A network error has occurred.', cause
=e
, expected
=True, video_id
=self
.get_temp_id(url
))
670 except (KeyError, StopIteration) as e
:
671 raise ExtractorError('An extractor error has occurred.', cause
=e
, video_id
=self
.get_temp_id(url
))
673 def __maybe_fake_ip_and_retry(self
, countries
):
674 if (not self
.get_param('geo_bypass_country', None)
676 and self
.get_param('geo_bypass', True)
677 and not self
._x
_forwarded
_for
_ip
679 country_code
= random
.choice(countries
)
680 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country_code
)
681 if self
._x
_forwarded
_for
_ip
:
683 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
684 % (self
._x
_forwarded
_for
_ip
, country_code
.upper()))
688 def set_downloader(self
, downloader
):
689 """Sets a YoutubeDL instance as the downloader for this IE."""
690 self
._downloader
= downloader
692 def _initialize_pre_login(self
):
693 """ Intialization before login. Redefine in subclasses."""
696 def _perform_login(self
, username
, password
):
697 """ Login with username and password. Redefine in subclasses."""
700 def _real_initialize(self
):
701 """Real initialization process. Redefine in subclasses."""
704 def _real_extract(self
, url
):
705 """Real extraction process. Redefine in subclasses."""
706 raise NotImplementedError('This method must be implemented by subclasses')
710 """A string for getting the InfoExtractor with get_info_extractor"""
711 return cls
.__name
__[:-2]
715 return compat_str(type(self
).__name
__[:-2])
718 def __can_accept_status_code(err
, expected_status
):
719 assert isinstance(err
, compat_urllib_error
.HTTPError
)
720 if expected_status
is None:
722 elif callable(expected_status
):
723 return expected_status(err
.code
) is True
725 return err
.code
in variadic(expected_status
)
727 def _request_webpage(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
, expected_status
=None):
729 Return the response handle.
731 See _download_webpage docstring for arguments specification.
733 if not self
._downloader
._first
_webpage
_request
:
734 sleep_interval
= self
.get_param('sleep_interval_requests') or 0
735 if sleep_interval
> 0:
736 self
.to_screen('Sleeping %s seconds ...' % sleep_interval
)
737 time
.sleep(sleep_interval
)
739 self
._downloader
._first
_webpage
_request
= False
742 self
.report_download_webpage(video_id
)
743 elif note
is not False:
745 self
.to_screen('%s' % (note
,))
747 self
.to_screen('%s: %s' % (video_id
, note
))
749 # Some sites check X-Forwarded-For HTTP header in order to figure out
750 # the origin of the client behind proxy. This allows bypassing geo
751 # restriction by faking this header's value to IP that belongs to some
752 # geo unrestricted country. We will do so once we encounter any
753 # geo restriction error.
754 if self
._x
_forwarded
_for
_ip
:
755 if 'X-Forwarded-For' not in headers
:
756 headers
['X-Forwarded-For'] = self
._x
_forwarded
_for
_ip
758 if isinstance(url_or_request
, compat_urllib_request
.Request
):
759 url_or_request
= update_Request(
760 url_or_request
, data
=data
, headers
=headers
, query
=query
)
763 url_or_request
= update_url_query(url_or_request
, query
)
764 if data
is not None or headers
:
765 url_or_request
= sanitized_Request(url_or_request
, data
, headers
)
767 return self
._downloader
.urlopen(url_or_request
)
768 except network_exceptions
as err
:
769 if isinstance(err
, compat_urllib_error
.HTTPError
):
770 if self
.__can
_accept
_status
_code
(err
, expected_status
):
771 # Retain reference to error to prevent file object from
772 # being closed before it can be read. Works around the
773 # effects of <https://bugs.python.org/issue15002>
774 # introduced in Python 3.4.1.
781 errnote
= 'Unable to download webpage'
783 errmsg
= '%s: %s' % (errnote
, error_to_compat_str(err
))
785 raise ExtractorError(errmsg
, cause
=err
)
787 self
.report_warning(errmsg
)
790 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):
792 Return a tuple (page content as string, URL handle).
794 See _download_webpage docstring for arguments specification.
796 # Strip hashes from the URL (#1038)
797 if isinstance(url_or_request
, (compat_str
, str)):
798 url_or_request
= url_or_request
.partition('#')[0]
800 urlh
= self
._request
_webpage
(url_or_request
, video_id
, note
, errnote
, fatal
, data
=data
, headers
=headers
, query
=query
, expected_status
=expected_status
)
804 content
= self
._webpage
_read
_content
(urlh
, url_or_request
, video_id
, note
, errnote
, fatal
, encoding
=encoding
)
805 return (content
, urlh
)
808 def _guess_encoding_from_content(content_type
, webpage_bytes
):
809 m
= re
.match(r
'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type
)
811 encoding
= m
.group(1)
813 m
= re
.search(br
'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
814 webpage_bytes[:1024])
816 encoding = m.group(1).decode('ascii')
817 elif webpage_bytes.startswith(b'\xff\xfe'):
824 def __check_blocked(self, content):
825 first_block = content[:512]
826 if ('<title>Access to this site is blocked</title>' in content
827 and 'Websense' in first_block):
828 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
829 blocked_iframe = self._html_search_regex(
830 r'<iframe src="([^
"]+)"', content,
831 'Websense information URL
', default=None)
833 msg += ' Visit
%s for more details
' % blocked_iframe
834 raise ExtractorError(msg, expected=True)
835 if '<title
>The URL you requested has been blocked
</title
>' in first_block:
837 'Access to this webpage has been blocked by Indian censorship
. '
838 'Use a VPN
or proxy
server (with --proxy
) to route around it
.')
839 block_msg = self._html_search_regex(
840 r'</h1
><p
>(.*?
)</p
>',
841 content, 'block message
', default=None)
843 msg += ' (Message
: "%s")' % block_msg.replace('\n', ' ')
844 raise ExtractorError(msg, expected=True)
845 if ('<title
>TTK
:: Доступ к ресурсу ограничен
</title
>' in content
846 and 'blocklist
.rkn
.gov
.ru
' in content):
847 raise ExtractorError(
848 'Access to this webpage has been blocked by decision of the Russian government
. '
849 'Visit http
://blocklist
.rkn
.gov
.ru
/ for a block reason
.',
852 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
853 content_type = urlh.headers.get('Content
-Type
', '')
854 webpage_bytes = urlh.read()
855 if prefix is not None:
856 webpage_bytes = prefix + webpage_bytes
858 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
859 if self.get_param('dump_intermediate_pages
', False):
860 self.to_screen('Dumping request to
' + urlh.geturl())
861 dump = base64.b64encode(webpage_bytes).decode('ascii
')
862 self._downloader.to_screen(dump)
863 if self.get_param('write_pages
', False):
864 basen = '%s_%s' % (video_id, urlh.geturl())
865 trim_length = self.get_param('trim_file_name
') or 240
866 if len(basen) > trim_length:
867 h = '___
' + hashlib.md5(basen.encode('utf
-8')).hexdigest()
868 basen = basen[:trim_length - len(h)] + h
869 raw_filename = basen + '.dump
'
870 filename = sanitize_filename(raw_filename, restricted=True)
871 self.to_screen('Saving request to
' + filename)
872 # Working around MAX_PATH limitation on Windows (see
873 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
874 if compat_os_name == 'nt
':
875 absfilepath = os.path.abspath(filename)
876 if len(absfilepath) > 259:
877 filename = '\\\\?
\\' + absfilepath
878 with open(filename, 'wb
') as outf:
879 outf.write(webpage_bytes)
882 content = webpage_bytes.decode(encoding, 'replace
')
884 content = webpage_bytes.decode('utf
-8', 'replace
')
886 self.__check_blocked(content)
890 def _download_webpage(
891 self, url_or_request, video_id, note=None, errnote=None,
892 fatal=True, tries=1, timeout=5, encoding=None, data=None,
893 headers={}, query={}, expected_status=None):
895 Return the data of the page as a string.
898 url_or_request -- plain text URL as a string or
899 a compat_urllib_request.Requestobject
900 video_id -- Video/playlist/item identifier (string)
903 note -- note printed before downloading (string)
904 errnote -- note printed in case of an error (string)
905 fatal -- flag denoting whether error should be considered fatal,
906 i.e. whether it should cause ExtractionError to be raised,
907 otherwise a warning will be reported and extraction continued
908 tries -- number of tries
909 timeout -- sleep interval between tries
910 encoding -- encoding for a page content decoding, guessed automatically
911 when not explicitly specified
912 data -- POST data (bytes)
913 headers -- HTTP headers (dict)
914 query -- URL query (dict)
915 expected_status -- allows to accept failed HTTP requests (non 2xx
916 status code) by explicitly specifying a set of accepted status
917 codes. Can be any of the following entities:
918 - an integer type specifying an exact failed status code to
920 - a list or a tuple of integer types specifying a list of
921 failed status codes to accept
922 - a callable accepting an actual failed status code and
923 returning True if it should be accepted
924 Note that this argument does not affect success status codes (2xx)
925 which are always accepted.
930 while success is False:
932 res = self._download_webpage_handle(
933 url_or_request, video_id, note, errnote, fatal,
934 encoding=encoding, data=data, headers=headers, query=query,
935 expected_status=expected_status)
937 except compat_http_client.IncompleteRead as e:
939 if try_count >= tries:
941 self._sleep(timeout, video_id)
948 def _download_xml_handle(
949 self, url_or_request, video_id, note='Downloading XML
',
950 errnote='Unable to download XML
', transform_source=None,
951 fatal=True, encoding=None, data=None, headers={}, query={},
952 expected_status=None):
954 Return a tuple (xml as an compat_etree_Element, URL handle).
956 See _download_webpage docstring for arguments specification.
958 res = self._download_webpage_handle(
959 url_or_request, video_id, note, errnote, fatal=fatal,
960 encoding=encoding, data=data, headers=headers, query=query,
961 expected_status=expected_status)
964 xml_string, urlh = res
965 return self._parse_xml(
966 xml_string, video_id, transform_source=transform_source,
970 self, url_or_request, video_id,
971 note='Downloading XML
', errnote='Unable to download XML
',
972 transform_source=None, fatal=True, encoding=None,
973 data=None, headers={}, query={}, expected_status=None):
975 Return the xml as an compat_etree_Element.
977 See _download_webpage docstring for arguments specification.
979 res = self._download_xml_handle(
980 url_or_request, video_id, note=note, errnote=errnote,
981 transform_source=transform_source, fatal=fatal, encoding=encoding,
982 data=data, headers=headers, query=query,
983 expected_status=expected_status)
984 return res if res is False else res[0]
986 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
988 xml_string = transform_source(xml_string)
990 return compat_etree_fromstring(xml_string.encode('utf
-8'))
991 except compat_xml_parse_error as ve:
992 errmsg = '%s: Failed to parse XML
' % video_id
994 raise ExtractorError(errmsg, cause=ve)
996 self.report_warning(errmsg + str(ve))
998 def _download_json_handle(
999 self, url_or_request, video_id, note='Downloading JSON metadata
',
1000 errnote='Unable to download JSON metadata
', transform_source=None,
1001 fatal=True, encoding=None, data=None, headers={}, query={},
1002 expected_status=None):
1004 Return a tuple (JSON object, URL handle).
1006 See _download_webpage docstring for arguments specification.
1008 res = self._download_webpage_handle(
1009 url_or_request, video_id, note, errnote, fatal=fatal,
1010 encoding=encoding, data=data, headers=headers, query=query,
1011 expected_status=expected_status)
1014 json_string, urlh = res
1015 return self._parse_json(
1016 json_string, video_id, transform_source=transform_source,
1020 self, url_or_request, video_id, note='Downloading JSON metadata
',
1021 errnote='Unable to download JSON metadata
', transform_source=None,
1022 fatal=True, encoding=None, data=None, headers={}, query={},
1023 expected_status=None):
1025 Return the JSON object as a dict.
1027 See _download_webpage docstring for arguments specification.
1029 res = self._download_json_handle(
1030 url_or_request, video_id, note=note, errnote=errnote,
1031 transform_source=transform_source, fatal=fatal, encoding=encoding,
1032 data=data, headers=headers, query=query,
1033 expected_status=expected_status)
1034 return res if res is False else res[0]
1036 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
1037 if transform_source:
1038 json_string = transform_source(json_string)
1040 return json.loads(json_string, strict=False)
1041 except ValueError as ve:
1042 errmsg = '%s: Failed to parse JSON
' % video_id
1044 raise ExtractorError(errmsg, cause=ve)
1046 self.report_warning(errmsg + str(ve))
1048 def _parse_socket_response_as_json(self, data, video_id, transform_source=None, fatal=True):
1049 return self._parse_json(
1050 data[data.find('{'):data.rfind('}
') + 1],
1051 video_id, transform_source, fatal)
1053 def _download_socket_json_handle(
1054 self, url_or_request, video_id, note='Polling socket
',
1055 errnote='Unable to poll socket
', transform_source=None,
1056 fatal=True, encoding=None, data=None, headers={}, query={},
1057 expected_status=None):
1059 Return a tuple (JSON object, URL handle).
1061 See _download_webpage docstring for arguments specification.
1063 res = self._download_webpage_handle(
1064 url_or_request, video_id, note, errnote, fatal=fatal,
1065 encoding=encoding, data=data, headers=headers, query=query,
1066 expected_status=expected_status)
1070 return self._parse_socket_response_as_json(
1071 webpage, video_id, transform_source=transform_source,
1074 def _download_socket_json(
1075 self, url_or_request, video_id, note='Polling socket
',
1076 errnote='Unable to poll socket
', transform_source=None,
1077 fatal=True, encoding=None, data=None, headers={}, query={},
1078 expected_status=None):
1080 Return the JSON object as a dict.
1082 See _download_webpage docstring for arguments specification.
1084 res = self._download_socket_json_handle(
1085 url_or_request, video_id, note=note, errnote=errnote,
1086 transform_source=transform_source, fatal=fatal, encoding=encoding,
1087 data=data, headers=headers, query=query,
1088 expected_status=expected_status)
1089 return res if res is False else res[0]
1091 def report_warning(self, msg, video_id=None, *args, only_once=False, **kwargs):
1092 idstr = format_field(video_id, template='%s: ')
1093 msg = f'[{self.IE_NAME}
] {idstr}{msg}
'
1095 if f'WARNING
: {msg}
' in self._printed_messages:
1097 self._printed_messages.add(f'WARNING
: {msg}
')
1098 self._downloader.report_warning(msg, *args, **kwargs)
1100 def to_screen(self, msg, *args, **kwargs):
1101 """Print msg to screen, prefixing it with '[ie_name
]'"""
1102 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg), *args, **kwargs)
1104 def write_debug(self, msg, *args, **kwargs):
1105 self._downloader.write_debug('[%s] %s' % (self.IE_NAME, msg), *args, **kwargs)
1107 def get_param(self, name, default=None, *args, **kwargs):
1108 if self._downloader:
1109 return self._downloader.params.get(name, default, *args, **kwargs)
1112 def report_drm(self, video_id, partial=False):
1113 self.raise_no_formats('This video
is DRM protected
', expected=True, video_id=video_id)
1115 def report_extraction(self, id_or_name):
1116 """Report information extraction."""
1117 self.to_screen('%s: Extracting information
' % id_or_name)
1119 def report_download_webpage(self, video_id):
1120 """Report webpage download."""
1121 self.to_screen('%s: Downloading webpage
' % video_id)
1123 def report_age_confirmation(self):
1124 """Report attempt to confirm age."""
1125 self.to_screen('Confirming age
')
1127 def report_login(self):
1128 """Report attempt to log in."""
1129 self.to_screen('Logging
in')
1131 def raise_login_required(
1132 self, msg='This video
is only available
for registered users
',
1133 metadata_available=False, method=NO_DEFAULT):
1134 if metadata_available and (
1135 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1136 self.report_warning(msg)
1138 if method is NO_DEFAULT:
1139 method = 'any
' if self.supports_login() else 'cookies
'
1140 if method is not None:
1141 assert method in self._LOGIN_HINTS, 'Invalid login method
'
1142 msg = '%s. %s' % (msg, self._LOGIN_HINTS[method])
1143 raise ExtractorError(msg, expected=True)
1145 def raise_geo_restricted(
1146 self, msg='This video
is not available
from your location due to geo restriction
',
1147 countries=None, metadata_available=False):
1148 if metadata_available and (
1149 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1150 self.report_warning(msg)
1152 raise GeoRestrictedError(msg, countries=countries)
1154 def raise_no_formats(self, msg, expected=False, video_id=None):
1156 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1157 self.report_warning(msg, video_id)
1158 elif isinstance(msg, ExtractorError):
1161 raise ExtractorError(msg, expected=expected, video_id=video_id)
1163 # Methods for following #608
1165 def url_result(url, ie=None, video_id=None, video_title=None, *, url_transparent=False, **kwargs):
1166 """Returns a URL that points to a page that should be processed"""
1168 kwargs['ie_key
'] = ie if isinstance(ie, str) else ie.ie_key()
1169 if video_id is not None:
1170 kwargs['id'] = video_id
1171 if video_title is not None:
1172 kwargs['title
'] = video_title
1175 '_type
': 'url_transparent
' if url_transparent else 'url
',
1179 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None, video_kwargs=None, **kwargs):
1180 urls = (self.url_result(self._proto_relative_url(m), ie, **(video_kwargs or {}))
1181 for m in orderedSet(map(getter, matches) if getter else matches))
1182 return self.playlist_result(urls, playlist_id, playlist_title, **kwargs)
1185 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None, *, multi_video=False, **kwargs):
1186 """Returns a playlist"""
1188 kwargs['id'] = playlist_id
1190 kwargs['title
'] = playlist_title
1191 if playlist_description is not None:
1192 kwargs['description
'] = playlist_description
1195 '_type
': 'multi_video
' if multi_video else 'playlist
',
1199 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1201 Perform a regex search on the given string, using a single or a list of
1202 patterns returning the first matching group.
1203 In case of failure return a default value or raise a WARNING or a
1204 RegexNotFoundError, depending on fatal, specifying the field name.
1206 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
1207 mobj = re.search(pattern, string, flags)
1210 mobj = re.search(p, string, flags)
1214 _name = self._downloader._format_err(name, self._downloader.Styles.EMPHASIS)
1218 # return the first matching group
1219 return next(g for g in mobj.groups() if g is not None)
1220 elif isinstance(group, (list, tuple)):
1221 return tuple(mobj.group(g) for g in group)
1223 return mobj.group(group)
1224 elif default is not NO_DEFAULT:
1227 raise RegexNotFoundError('Unable to extract
%s' % _name)
1229 self.report_warning('unable to extract
%s' % _name + bug_reports_message())
1232 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1234 Like _search_regex, but strips HTML tags and unescapes entities.
1236 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1238 return clean_html(res).strip()
1242 def _get_netrc_login_info(self, netrc_machine=None):
1245 netrc_machine = netrc_machine or self._NETRC_MACHINE
1247 if self.get_param('usenetrc
', False):
1249 netrc_file = compat_expanduser(self.get_param('netrc_location
') or '~
')
1250 if os.path.isdir(netrc_file):
1251 netrc_file = os.path.join(netrc_file, '.netrc
')
1252 info = netrc.netrc(file=netrc_file).authenticators(netrc_machine)
1253 if info is not None:
1257 raise netrc.NetrcParseError(
1258 'No authenticators
for %s' % netrc_machine)
1259 except (IOError, netrc.NetrcParseError) as err:
1260 self.report_warning(
1261 'parsing
.netrc
: %s' % error_to_compat_str(err))
1263 return username, password
1265 def _get_login_info(self, username_option='username
', password_option='password
', netrc_machine=None):
1267 Get the login info as (username, password)
1268 First look for the manually specified credentials using username_option
1269 and password_option as keys in params dictionary. If no such credentials
1270 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1272 If there's no info available
, return (None, None)
1275 # Attempt to use provided username and password or .netrc data
1276 username = self.get_param(username_option)
1277 if username is not None:
1278 password = self.get_param(password_option)
1280 username, password = self._get_netrc_login_info(netrc_machine)
1282 return username, password
1284 def _get_tfa_info(self, note='two-factor verification code'):
1286 Get the two
-factor authentication info
1287 TODO
- asking the user will be required
for sms
/phone verify
1288 currently just uses the command line option
1289 If there
's no info available, return None
1292 tfa = self.get_param('twofactor
')
1296 return compat_getpass('Type
%s and press
[Return
]: ' % note)
1298 # Helper functions for extracting OpenGraph info
1300 def _og_regexes(prop):
1301 content_re = r'content
=(?
:"([^"]+?
)"|\'([^\']+?)\'|\s*([^\s"\'=<>`
]+?
))'
1302 property_re = (r'(?
:name|
property)=(?
:\'og
%(sep)s%(prop)s\'|
"og%(sep)s%(prop)s"|\s
*og
%(sep)s%(prop)s\b)'
1303 % {'prop': re.escape(prop), 'sep': '(?::|[:-])'})
1304 template = r'<meta
[^
>]+?
%s[^
>]+?
%s'
1306 template % (property_re, content_re),
1307 template % (content_re, property_re),
1311 def _meta_regex(prop):
1312 return r'''(?isx)<meta
1313 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1314 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1316 def _og_search_property(self, prop, html, name=None, **kargs):
1317 prop = variadic(prop)
1319 name = 'OpenGraph
%s' % prop[0]
1322 og_regexes.extend(self._og_regexes(p))
1323 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1326 return unescapeHTML(escaped)
1328 def _og_search_thumbnail(self, html, **kargs):
1329 return self._og_search_property('image
', html, 'thumbnail URL
', fatal=False, **kargs)
1331 def _og_search_description(self, html, **kargs):
1332 return self._og_search_property('description
', html, fatal=False, **kargs)
1334 def _og_search_title(self, html, *, fatal=False, **kargs):
1335 return self._og_search_property('title
', html, fatal=fatal, **kargs)
1337 def _og_search_video_url(self, html, name='video url
', secure=True, **kargs):
1338 regexes = self._og_regexes('video
') + self._og_regexes('video
:url
')
1340 regexes = self._og_regexes('video
:secure_url
') + regexes
1341 return self._html_search_regex(regexes, html, name, **kargs)
1343 def _og_search_url(self, html, **kargs):
1344 return self._og_search_property('url
', html, **kargs)
1346 def _html_extract_title(self, html, name='title
', *, fatal=False, **kwargs):
1347 return self._html_search_regex(r'(?s
)<title
>([^
<]+)</title
>', html, name, fatal=fatal, **kwargs)
1349 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1350 name = variadic(name)
1351 if display_name is None:
1352 display_name = name[0]
1353 return self._html_search_regex(
1354 [self._meta_regex(n) for n in name],
1355 html, display_name, fatal=fatal, group='content
', **kwargs)
1357 def _dc_search_uploader(self, html):
1358 return self._html_search_meta('dc
.creator
', html, 'uploader
')
1360 def _rta_search(self, html):
1361 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1362 if re.search(r'(?ix
)<meta\s
+name
="rating"\s
+'
1363 r' content
="RTA-5042-1996-1400-1577-RTA"',
1368 def _media_rating_search(self, html):
1369 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1370 rating = self._html_search_meta('rating
', html)
1382 return RATING_TABLE.get(rating.lower())
1384 def _family_friendly_search(self, html):
1385 # See http://schema.org/VideoObject
1386 family_friendly = self._html_search_meta(
1387 'isFamilyFriendly
', html, default=None)
1389 if not family_friendly:
1398 return RATING_TABLE.get(family_friendly.lower())
1400 def _twitter_search_player(self, html):
1401 return self._html_search_meta('twitter
:player
', html,
1402 'twitter card player
')
1404 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
1405 json_ld_list = list(re.finditer(JSON_LD_RE, html))
1406 default = kwargs.get('default
', NO_DEFAULT)
1407 # JSON-LD may be malformed and thus `fatal` should be respected.
1408 # At the same time `default` may be passed that assumes `fatal=False`
1409 # for _search_regex. Let's simulate the same behavior here
as well
.
1410 fatal
= kwargs
.get('fatal', True) if default
is NO_DEFAULT
else False
1412 for mobj
in json_ld_list
:
1413 json_ld_item
= self
._parse
_json
(
1414 mobj
.group('json_ld'), video_id
, fatal
=fatal
)
1415 if not json_ld_item
:
1417 if isinstance(json_ld_item
, dict):
1418 json_ld
.append(json_ld_item
)
1419 elif isinstance(json_ld_item
, (list, tuple)):
1420 json_ld
.extend(json_ld_item
)
1422 json_ld
= self
._json
_ld
(json_ld
, video_id
, fatal
=fatal
, expected_type
=expected_type
)
1425 if default
is not NO_DEFAULT
:
1428 raise RegexNotFoundError('Unable to extract JSON-LD')
1430 self
.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1433 def _json_ld(self
, json_ld
, video_id
, fatal
=True, expected_type
=None):
1434 if isinstance(json_ld
, compat_str
):
1435 json_ld
= self
._parse
_json
(json_ld
, video_id
, fatal
=fatal
)
1439 if not isinstance(json_ld
, (list, tuple, dict)):
1441 if isinstance(json_ld
, dict):
1444 INTERACTION_TYPE_MAP
= {
1445 'CommentAction': 'comment',
1446 'AgreeAction': 'like',
1447 'DisagreeAction': 'dislike',
1448 'LikeAction': 'like',
1449 'DislikeAction': 'dislike',
1450 'ListenAction': 'view',
1451 'WatchAction': 'view',
1452 'ViewAction': 'view',
1455 def extract_interaction_type(e
):
1456 interaction_type
= e
.get('interactionType')
1457 if isinstance(interaction_type
, dict):
1458 interaction_type
= interaction_type
.get('@type')
1459 return str_or_none(interaction_type
)
1461 def extract_interaction_statistic(e
):
1462 interaction_statistic
= e
.get('interactionStatistic')
1463 if isinstance(interaction_statistic
, dict):
1464 interaction_statistic
= [interaction_statistic
]
1465 if not isinstance(interaction_statistic
, list):
1467 for is_e
in interaction_statistic
:
1468 if not isinstance(is_e
, dict):
1470 if is_e
.get('@type') != 'InteractionCounter':
1472 interaction_type
= extract_interaction_type(is_e
)
1473 if not interaction_type
:
1475 # For interaction count some sites provide string instead of
1476 # an integer (as per spec) with non digit characters (e.g. ",")
1477 # so extracting count with more relaxed str_to_int
1478 interaction_count
= str_to_int(is_e
.get('userInteractionCount'))
1479 if interaction_count
is None:
1481 count_kind
= INTERACTION_TYPE_MAP
.get(interaction_type
.split('/')[-1])
1484 count_key
= '%s_count' % count_kind
1485 if info
.get(count_key
) is not None:
1487 info
[count_key
] = interaction_count
1489 def extract_chapter_information(e
):
1491 'title': part
.get('name'),
1492 'start_time': part
.get('startOffset'),
1493 'end_time': part
.get('endOffset'),
1494 } for part
in variadic(e
.get('hasPart') or []) if part
.get('@type') == 'Clip']
1495 for idx
, (last_c
, current_c
, next_c
) in enumerate(zip(
1496 [{'end_time': 0}
] + chapters
, chapters
, chapters
[1:])):
1497 current_c
['end_time'] = current_c
['end_time'] or next_c
['start_time']
1498 current_c
['start_time'] = current_c
['start_time'] or last_c
['end_time']
1499 if None in current_c
.values():
1500 self
.report_warning(f
'Chapter {idx} contains broken data. Not extracting chapters')
1503 chapters
[-1]['end_time'] = chapters
[-1]['end_time'] or info
['duration']
1504 info
['chapters'] = chapters
1506 def extract_video_object(e
):
1507 assert e
['@type'] == 'VideoObject'
1508 author
= e
.get('author')
1510 'url': url_or_none(e
.get('contentUrl')),
1511 'title': unescapeHTML(e
.get('name')),
1512 'description': unescapeHTML(e
.get('description')),
1513 'thumbnails': [{'url': url_or_none(url)}
1514 for url
in variadic(traverse_obj(e
, 'thumbnailUrl', 'thumbnailURL'))],
1515 'duration': parse_duration(e
.get('duration')),
1516 'timestamp': unified_timestamp(e
.get('uploadDate')),
1517 # author can be an instance of 'Organization' or 'Person' types.
1518 # both types can have 'name' property(inherited from 'Thing' type). [1]
1519 # however some websites are using 'Text' type instead.
1520 # 1. https://schema.org/VideoObject
1521 'uploader': author
.get('name') if isinstance(author
, dict) else author
if isinstance(author
, compat_str
) else None,
1522 'filesize': float_or_none(e
.get('contentSize')),
1523 'tbr': int_or_none(e
.get('bitrate')),
1524 'width': int_or_none(e
.get('width')),
1525 'height': int_or_none(e
.get('height')),
1526 'view_count': int_or_none(e
.get('interactionCount')),
1528 extract_interaction_statistic(e
)
1529 extract_chapter_information(e
)
1531 def traverse_json_ld(json_ld
, at_top_level
=True):
1533 if at_top_level
and '@context' not in e
:
1535 if at_top_level
and set(e
.keys()) == {'@context', '@graph'}
:
1536 traverse_json_ld(variadic(e
['@graph'], allowed_types
=(dict,)), at_top_level
=False)
1538 item_type
= e
.get('@type')
1539 if expected_type
is not None and expected_type
!= item_type
:
1541 rating
= traverse_obj(e
, ('aggregateRating', 'ratingValue'), expected_type
=float_or_none
)
1542 if rating
is not None:
1543 info
['average_rating'] = rating
1544 if item_type
in ('TVEpisode', 'Episode'):
1545 episode_name
= unescapeHTML(e
.get('name'))
1547 'episode': episode_name
,
1548 'episode_number': int_or_none(e
.get('episodeNumber')),
1549 'description': unescapeHTML(e
.get('description')),
1551 if not info
.get('title') and episode_name
:
1552 info
['title'] = episode_name
1553 part_of_season
= e
.get('partOfSeason')
1554 if isinstance(part_of_season
, dict) and part_of_season
.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1556 'season': unescapeHTML(part_of_season
.get('name')),
1557 'season_number': int_or_none(part_of_season
.get('seasonNumber')),
1559 part_of_series
= e
.get('partOfSeries') or e
.get('partOfTVSeries')
1560 if isinstance(part_of_series
, dict) and part_of_series
.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1561 info
['series'] = unescapeHTML(part_of_series
.get('name'))
1562 elif item_type
== 'Movie':
1564 'title': unescapeHTML(e
.get('name')),
1565 'description': unescapeHTML(e
.get('description')),
1566 'duration': parse_duration(e
.get('duration')),
1567 'timestamp': unified_timestamp(e
.get('dateCreated')),
1569 elif item_type
in ('Article', 'NewsArticle'):
1571 'timestamp': parse_iso8601(e
.get('datePublished')),
1572 'title': unescapeHTML(e
.get('headline')),
1573 'description': unescapeHTML(e
.get('articleBody') or e
.get('description')),
1575 if traverse_obj(e
, ('video', 0, '@type')) == 'VideoObject':
1576 extract_video_object(e
['video'][0])
1577 elif item_type
== 'VideoObject':
1578 extract_video_object(e
)
1579 if expected_type
is None:
1583 video
= e
.get('video')
1584 if isinstance(video
, dict) and video
.get('@type') == 'VideoObject':
1585 extract_video_object(video
)
1586 if expected_type
is None:
1590 traverse_json_ld(json_ld
)
1592 return filter_dict(info
)
1594 def _search_nextjs_data(self
, webpage
, video_id
, *, transform_source
=None, fatal
=True, **kw
):
1595 return self
._parse
_json
(
1597 r
'(?s)<script[^>]+id=[\'"]__NEXT_DATA__[\'"][^
>]*>([^
<]+)</script
>',
1598 webpage, 'next
.js data
', fatal=fatal, **kw),
1599 video_id, transform_source=transform_source, fatal=fatal)
1601 def _search_nuxt_data(self, webpage, video_id, context_name='__NUXT__
'):
1602 ''' Parses Nuxt.js metadata. This works as long as the function __NUXT__ invokes is a pure function. '''
1603 # not all website do this, but it can be changed
1604 # https://stackoverflow.com/questions/67463109/how-to-change-or-hide-nuxt-and-nuxt-keyword-in-page-source
1605 rectx = re.escape(context_name)
1606 js, arg_keys, arg_vals = self._search_regex(
1607 (r'<script
>window\
.%s=\
(function\
((?P
<arg_keys
>.*?
)\
)\{return\s(?P<js>\{.*?\}
)\
}\
((?P
<arg_vals
>.+?
)\
)\
);?
</script
>' % rectx,
1608 r'%s\
(.*?\
(function\
((?P
<arg_keys
>.*?
)\
)\{return\s(?P<js>\{.*?\}
)\
}\
((?P
<arg_vals
>.*?
)\
)' % rectx),
1609 webpage, context_name, group=['js
', 'arg_keys
', 'arg_vals
'])
1611 args = dict(zip(arg_keys.split(','), arg_vals.split(',')))
1613 for key, val in args.items():
1614 if val in ('undefined
', 'void
0'):
1617 return self._parse_json(js_to_json(js, args), video_id)['data
'][0]
1620 def _hidden_inputs(html):
1621 html = re.sub(r'<!--(?
:(?
!<!--).)*-->', '', html)
1623 for input in re.findall(r'(?i
)(<input[^
>]+>)', html):
1624 attrs = extract_attributes(input)
1627 if attrs.get('type') not in ('hidden
', 'submit
'):
1629 name = attrs.get('name
') or attrs.get('id')
1630 value = attrs.get('value
')
1631 if name and value is not None:
1632 hidden_inputs[name] = value
1633 return hidden_inputs
1635 def _form_hidden_inputs(self, form_id, html):
1636 form = self._search_regex(
1637 r'(?
is)<form
[^
>]+?
id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1638 html, '%s form' % form_id, group='form')
1639 return self._hidden_inputs(form)
1642 regex = r' *((?P<reverse>\+)?(?P<field>[a-zA-Z0-9_]+)((?P<separator>[~:])(?P<limit>.*?))?)? *$'
1644 default = ('hidden', 'aud_or_vid', 'hasvid', 'ie_pref', 'lang', 'quality',
1645 'res', 'fps', 'hdr:12', 'codec:vp9.2', 'size', 'br', 'asr',
1646 'proto', 'ext', 'hasaud', 'source', 'id') # These must not be aliases
1647 ytdl_default = ('hasaud', 'lang', 'quality', 'tbr', 'filesize', 'vbr',
1648 'height', 'width', 'proto', 'vext', 'abr', 'aext',
1649 'fps', 'fs_approx', 'source', 'id')
1652 'vcodec': {'type': 'ordered', 'regex': True,
1653 'order': ['av0?1', 'vp0?9.2', 'vp0?9', '[hx]265|he?vc?', '[hx]264|avc', 'vp0?8', 'mp4v|h263', 'theora', '', None, 'none']},
1654 'acodec': {'type': 'ordered', 'regex': True,
1655 'order': ['[af]lac', 'wav|aiff', 'opus', 'vorbis|ogg', 'aac', 'mp?4a?', 'mp3', 'e-?a?c-?3', 'ac-?3', 'dts', '', None, 'none']},
1656 'hdr': {'type': 'ordered', 'regex': True, 'field': 'dynamic_range',
1657 'order': ['dv', '(hdr)?12', r'(hdr)?10\+', '(hdr)?10', 'hlg', '', 'sdr', None]},
1658 'proto': {'type': 'ordered', 'regex': True, 'field': 'protocol',
1659 'order': ['(ht|f)tps', '(ht|f)tp$', 'm3u8.*', '.*dash', 'websocket_frag', 'rtmpe?', '', 'mms|rtsp', 'ws|websocket', 'f4']},
1660 'vext': {'type': 'ordered', 'field': 'video_ext',
1661 'order': ('mp4', 'webm', 'flv', '', 'none'),
1662 'order_free': ('webm', 'mp4', 'flv', '', 'none')},
1663 'aext': {'type': 'ordered', 'field': 'audio_ext',
1664 'order': ('m4a', 'aac', 'mp3', 'ogg', 'opus', 'webm', '', 'none'),
1665 'order_free': ('opus', 'ogg', 'webm', 'm4a', 'mp3', 'aac', '', 'none')},
1666 'hidden': {'visible': False, 'forced': True, 'type': 'extractor', 'max': -1000},
1667 'aud_or_vid': {'visible': False, 'forced': True, 'type': 'multiple',
1668 'field': ('vcodec', 'acodec'),
1669 'function': lambda it: int(any(v != 'none' for v in it))},
1670 'ie_pref': {'priority': True, 'type': 'extractor'},
1671 'hasvid': {'priority': True, 'field': 'vcodec', 'type': 'boolean', 'not_in_list': ('none',)},
1672 'hasaud': {'field': 'acodec', 'type': 'boolean', 'not_in_list': ('none',)},
1673 'lang': {'convert': 'float', 'field': 'language_preference', 'default': -1},
1674 'quality': {'convert': 'float', 'default': -1},
1675 'filesize': {'convert': 'bytes'},
1676 'fs_approx': {'convert': 'bytes', 'field': 'filesize_approx'},
1677 'id': {'convert': 'string', 'field': 'format_id'},
1678 'height': {'convert': 'float_none'},
1679 'width': {'convert': 'float_none'},
1680 'fps': {'convert': 'float_none'},
1681 'tbr': {'convert': 'float_none'},
1682 'vbr': {'convert': 'float_none'},
1683 'abr': {'convert': 'float_none'},
1684 'asr': {'convert': 'float_none'},
1685 'source': {'convert': 'float', 'field': 'source_preference', 'default': -1},
1687 'codec': {'type': 'combined', 'field': ('vcodec', 'acodec')},
1688 'br': {'type': 'combined', 'field': ('tbr', 'vbr', 'abr'), 'same_limit': True},
1689 'size': {'type': 'combined', 'same_limit': True, 'field': ('filesize', 'fs_approx')},
1690 'ext': {'type': 'combined', 'field': ('vext', 'aext')},
1691 'res': {'type': 'multiple', 'field': ('height', 'width'),
1692 'function': lambda it: (lambda l: min(l) if l else 0)(tuple(filter(None, it)))},
1694 # For compatibility with youtube-dl
1695 'format_id': {'type': 'alias', 'field': 'id'},
1696 'preference': {'type': 'alias', 'field': 'ie_pref'},
1697 'language_preference': {'type': 'alias', 'field': 'lang'},
1698 'source_preference': {'type': 'alias', 'field': 'source'},
1699 'protocol': {'type': 'alias', 'field': 'proto'},
1700 'filesize_approx': {'type': 'alias', 'field': 'fs_approx'},
1703 'dimension': {'type': 'alias', 'field': 'res', 'deprecated': True},
1704 'resolution': {'type': 'alias', 'field': 'res', 'deprecated': True},
1705 'extension': {'type': 'alias', 'field': 'ext', 'deprecated': True},
1706 'bitrate': {'type': 'alias', 'field': 'br', 'deprecated': True},
1707 'total_bitrate': {'type': 'alias', 'field': 'tbr', 'deprecated': True},
1708 'video_bitrate': {'type': 'alias', 'field': 'vbr', 'deprecated': True},
1709 'audio_bitrate': {'type': 'alias', 'field': 'abr', 'deprecated': True},
1710 'framerate': {'type': 'alias', 'field': 'fps', 'deprecated': True},
1711 'filesize_estimate': {'type': 'alias', 'field': 'size', 'deprecated': True},
1712 'samplerate': {'type': 'alias', 'field': 'asr', 'deprecated': True},
1713 'video_ext': {'type': 'alias', 'field': 'vext', 'deprecated': True},
1714 'audio_ext': {'type': 'alias', 'field': 'aext', 'deprecated': True},
1715 'video_codec': {'type': 'alias', 'field': 'vcodec', 'deprecated': True},
1716 'audio_codec': {'type': 'alias', 'field': 'acodec', 'deprecated': True},
1717 'video': {'type': 'alias', 'field': 'hasvid', 'deprecated': True},
1718 'has_video': {'type': 'alias', 'field': 'hasvid', 'deprecated': True},
1719 'audio': {'type': 'alias', 'field': 'hasaud', 'deprecated': True},
1720 'has_audio': {'type': 'alias', 'field': 'hasaud', 'deprecated': True},
1721 'extractor': {'type': 'alias', 'field': 'ie_pref', 'deprecated': True},
1722 'extractor_preference': {'type': 'alias', 'field': 'ie_pref', 'deprecated': True},
1725 def __init__(self, ie, field_preference):
1727 self.ydl = ie._downloader
1728 self.evaluate_params(self.ydl.params, field_preference)
1729 if ie.get_param('verbose'):
1730 self.print_verbose_info(self.ydl.write_debug)
1732 def _get_field_setting(self, field, key):
1733 if field not in self.settings:
1734 if key in ('forced', 'priority'):
1736 self.ydl.deprecation_warning(
1737 f'Using arbitrary fields ({field}) for format sorting is deprecated '
1738 'and may be removed in a future version')
1739 self.settings[field] = {}
1740 propObj = self.settings[field]
1741 if key not in propObj:
1742 type = propObj.get('type')
1744 default = 'preference' if type == 'extractor' else (field,) if type in ('combined', 'multiple') else field
1745 elif key == 'convert':
1746 default = 'order' if type == 'ordered' else 'float_string' if field else 'ignore'
1748 default = {'type': 'field', 'visible': True, 'order': [], 'not_in_list': (None,)}.get(key, None)
1749 propObj[key] = default
1752 def _resolve_field_value(self, field, value, convertNone=False):
1757 value = value.lower()
1758 conversion = self._get_field_setting(field, 'convert')
1759 if conversion == 'ignore':
1761 if conversion == 'string':
1763 elif conversion == 'float_none':
1764 return float_or_none(value)
1765 elif conversion == 'bytes':
1766 return FileDownloader.parse_bytes(value)
1767 elif conversion == 'order':
1768 order_list = (self._use_free_order and self._get_field_setting(field, 'order_free')) or self._get_field_setting(field, 'order')
1769 use_regex = self._get_field_setting(field, 'regex')
1770 list_length = len(order_list)
1771 empty_pos = order_list.index('') if '' in order_list else list_length + 1
1772 if use_regex and value is not None:
1773 for i, regex in enumerate(order_list):
1774 if regex and re.match(regex, value):
1775 return list_length - i
1776 return list_length - empty_pos # not in list
1777 else: # not regex or value = None
1778 return list_length - (order_list.index(value) if value in order_list else empty_pos)
1780 if value.isnumeric():
1783 self.settings[field]['convert'] = 'string'
1786 def evaluate_params(self, params, sort_extractor):
1787 self._use_free_order = params.get('prefer_free_formats', False)
1788 self._sort_user = params.get('format_sort', [])
1789 self._sort_extractor = sort_extractor
1791 def add_item(field, reverse, closest, limit_text):
1792 field = field.lower()
1793 if field in self._order:
1795 self._order.append(field)
1796 limit = self._resolve_field_value(field, limit_text)
1799 'closest': False if limit is None else closest,
1800 'limit_text': limit_text,
1802 if field in self.settings:
1803 self.settings[field].update(data)
1805 self.settings[field] = data
1808 tuple(field for field in self.default if self._get_field_setting(field, 'forced'))
1809 + (tuple() if params.get('format_sort_force', False)
1810 else tuple(field for field in self.default if self._get_field_setting(field, 'priority')))
1811 + tuple(self._sort_user) + tuple(sort_extractor) + self.default)
1813 for item in sort_list:
1814 match = re.match(self.regex, item)
1816 raise ExtractorError('Invalid format sort string "%s" given by extractor' % item)
1817 field = match.group('field')
1820 if self._get_field_setting(field, 'type') == 'alias':
1821 alias, field = field, self._get_field_setting(field, 'field')
1822 if self._get_field_setting(alias, 'deprecated'):
1823 self.ydl.deprecation_warning(
1824 f'Format sorting alias {alias} is deprecated '
1825 f'and may be removed in a future version. Please use {field} instead')
1826 reverse = match.group('reverse') is not None
1827 closest = match.group('separator') == '~'
1828 limit_text = match.group('limit')
1830 has_limit = limit_text is not None
1831 has_multiple_fields = self._get_field_setting(field, 'type') == 'combined'
1832 has_multiple_limits = has_limit and has_multiple_fields and not self._get_field_setting(field, 'same_limit')
1834 fields = self._get_field_setting(field, 'field') if has_multiple_fields else (field,)
1835 limits = limit_text.split(':') if has_multiple_limits else (limit_text,) if has_limit else tuple()
1836 limit_count = len(limits)
1837 for (i, f) in enumerate(fields):
1838 add_item(f, reverse, closest,
1839 limits[i] if i < limit_count
1840 else limits[0] if has_limit and not has_multiple_limits
1843 def print_verbose_info(self, write_debug):
1845 write_debug('Sort order given by user: %s' % ', '.join(self._sort_user))
1846 if self._sort_extractor:
1847 write_debug('Sort order given by extractor: %s' % ', '.join(self._sort_extractor))
1848 write_debug('Formats sorted by: %s' % ', '.join(['%s%s%s' % (
1849 '+' if self._get_field_setting(field, 'reverse') else '', field,
1850 '%s%s(%s)' % ('~' if self._get_field_setting(field, 'closest') else ':',
1851 self._get_field_setting(field, 'limit_text'),
1852 self._get_field_setting(field, 'limit'))
1853 if self._get_field_setting(field, 'limit_text') is not None else '')
1854 for field in self._order if self._get_field_setting(field, 'visible')]))
1856 def _calculate_field_preference_from_value(self, format, field, type, value):
1857 reverse = self._get_field_setting(field, 'reverse')
1858 closest = self._get_field_setting(field, 'closest')
1859 limit = self._get_field_setting(field, 'limit')
1861 if type == 'extractor':
1862 maximum = self._get_field_setting(field, 'max')
1863 if value is None or (maximum is not None and value >= maximum):
1865 elif type == 'boolean':
1866 in_list = self._get_field_setting(field, 'in_list')
1867 not_in_list = self._get_field_setting(field, 'not_in_list')
1868 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
1869 elif type == 'ordered':
1870 value = self._resolve_field_value(field, value, True)
1872 # try to convert to number
1873 val_num = float_or_none(value, default=self._get_field_setting(field, 'default'))
1874 is_num = self._get_field_setting(field, 'convert') != 'string' and val_num is not None
1878 return ((-10, 0) if value is None
1879 else (1, value, 0) if not is_num # if a field has mixed strings and numbers, strings are sorted higher
1880 else (0, -abs(value - limit), value - limit if reverse else limit - value) if closest
1881 else (0, value, 0) if not reverse and (limit is None or value <= limit)
1882 else (0, -value, 0) if limit is None or (reverse and value == limit) or value > limit
1883 else (-1, value, 0))
1885 def _calculate_field_preference(self, format, field):
1886 type = self._get_field_setting(field, 'type') # extractor, boolean, ordered, field, multiple
1887 get_value = lambda f: format.get(self._get_field_setting(f, 'field'))
1888 if type == 'multiple':
1889 type = 'field' # Only 'field' is allowed in multiple for now
1890 actual_fields = self._get_field_setting(field, 'field')
1892 value = self._get_field_setting(field, 'function')(get_value(f) for f in actual_fields)
1894 value = get_value(field)
1895 return self._calculate_field_preference_from_value(format, field, type, value)
1897 def calculate_preference(self, format):
1898 # Determine missing protocol
1899 if not format.get('protocol'):
1900 format['protocol'] = determine_protocol(format)
1902 # Determine missing ext
1903 if not format.get('ext') and 'url' in format:
1904 format['ext'] = determine_ext(format['url'])
1905 if format.get('vcodec') == 'none':
1906 format['audio_ext'] = format['ext'] if format.get('acodec') != 'none' else 'none'
1907 format['video_ext'] = 'none'
1909 format['video_ext'] = format['ext']
1910 format['audio_ext'] = 'none'
1911 # if format.get('preference') is None and format.get('ext') in ('f4f', 'f4m'): # Not supported?
1912 # format['preference'] = -1000
1914 # Determine missing bitrates
1915 if format.get('tbr') is None:
1916 if format.get('vbr') is not None and format.get('abr') is not None:
1917 format['tbr'] = format.get('vbr', 0) + format.get('abr', 0)
1919 if format.get('vcodec') != 'none' and format.get('vbr') is None:
1920 format['vbr'] = format.get('tbr') - format.get('abr', 0)
1921 if format.get('acodec') != 'none' and format.get('abr') is None:
1922 format['abr'] = format.get('tbr') - format.get('vbr', 0)
1924 return tuple(self._calculate_field_preference(format, field) for field in self._order)
1926 def _sort_formats(self, formats, field_preference=[]):
1929 format_sort = self.FormatSort(self, field_preference)
1930 formats.sort(key=lambda f: format_sort.calculate_preference(f))
1932 def _check_formats(self, formats, video_id):
1934 formats[:] = filter(
1935 lambda f: self._is_valid_url(
1937 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1941 def _remove_duplicate_formats(formats):
1945 if f['url'] not in format_urls:
1946 format_urls.add(f['url'])
1947 unique_formats.append(f)
1948 formats[:] = unique_formats
1950 def _is_valid_url(self, url, video_id, item='video', headers={}):
1951 url = self._proto_relative_url(url, scheme='http:')
1952 # For now assume non HTTP(S) URLs always valid
1953 if not (url.startswith('http://') or url.startswith('https://')):
1956 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1958 except ExtractorError as e:
1960 '%s: %s URL is invalid, skipping: %s'
1961 % (video_id, item, error_to_compat_str(e.cause)))
1964 def http_scheme(self):
1965 """ Either "http
:" or "https
:", depending on the user's preferences """
1968 if self.get_param('prefer_insecure', False)
1971 def _proto_relative_url(self, url, scheme=None):
1974 if url.startswith('//'):
1976 scheme = self.http_scheme()
1981 def _sleep(self, timeout, video_id, msg_template=None):
1982 if msg_template is None:
1983 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1984 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1988 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1989 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1990 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
1991 manifest = self._download_xml(
1992 manifest_url, video_id, 'Downloading f4m manifest',
1993 'Unable to download f4m manifest',
1994 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1995 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
1996 transform_source=transform_source,
1997 fatal=fatal, data=data, headers=headers, query=query)
1999 if manifest is False:
2002 return self._parse_f4m_formats(
2003 manifest, manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
2004 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
2006 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
2007 transform_source=lambda s: fix_xml_ampersands(s).strip(),
2008 fatal=True, m3u8_id=None):
2009 if not isinstance(manifest, compat_etree_Element) and not fatal:
2012 # currently yt-dlp cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
2013 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
2014 if akamai_pv is not None and ';' in akamai_pv.text:
2015 playerVerificationChallenge = akamai_pv.text.split(';')[0]
2016 if playerVerificationChallenge.strip() != '':
2020 manifest_version = '1.0'
2021 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
2023 manifest_version = '2.0'
2024 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
2025 # Remove unsupported DRM protected media from final formats
2026 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
2027 media_nodes = remove_encrypted_media(media_nodes)
2031 manifest_base_url = get_base_url(manifest)
2033 bootstrap_info = xpath_element(
2034 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
2035 'bootstrap info', default=None)
2038 mime_type = xpath_text(
2039 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
2040 'base URL', default=None)
2041 if mime_type and mime_type.startswith('audio/'):
2044 for i, media_el in enumerate(media_nodes):
2045 tbr = int_or_none(media_el.attrib.get('bitrate'))
2046 width = int_or_none(media_el.attrib.get('width'))
2047 height = int_or_none(media_el.attrib.get('height'))
2048 format_id = join_nonempty(f4m_id, tbr or i)
2049 # If <bootstrapInfo> is present, the specified f4m is a
2050 # stream-level manifest, and only set-level manifests may refer to
2051 # external resources. See section 11.4 and section 4 of F4M spec
2052 if bootstrap_info is None:
2054 # @href is introduced in 2.0, see section 11.6 of F4M spec
2055 if manifest_version == '2.0':
2056 media_url = media_el.attrib.get('href')
2057 if media_url is None:
2058 media_url = media_el.attrib.get('url')
2062 media_url if media_url.startswith('http://') or media_url.startswith('https://')
2063 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
2064 # If media_url is itself a f4m manifest do the recursive extraction
2065 # since bitrates in parent manifest (this one) and media_url manifest
2066 # may differ leading to inability to resolve the format by requested
2067 # bitrate in f4m downloader
2068 ext = determine_ext(manifest_url)
2070 f4m_formats = self._extract_f4m_formats(
2071 manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
2072 transform_source=transform_source, fatal=fatal)
2073 # Sometimes stream-level manifest contains single media entry that
2074 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
2075 # At the same time parent's media entry in set-level manifest may
2076 # contain it. We will copy it from parent in such cases.
2077 if len(f4m_formats) == 1:
2080 'tbr': f.get('tbr') or tbr,
2081 'width': f.get('width') or width,
2082 'height': f.get('height') or height,
2083 'format_id': f.get('format_id') if not tbr else format_id,
2086 formats.extend(f4m_formats)
2089 formats.extend(self._extract_m3u8_formats(
2090 manifest_url, video_id, 'mp4', preference=preference,
2091 quality=quality, m3u8_id=m3u8_id, fatal=fatal))
2094 'format_id': format_id,
2095 'url': manifest_url,
2096 'manifest_url': manifest_url,
2097 'ext': 'flv' if bootstrap_info is not None else None,
2103 'preference': preference,
2108 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, quality=None, m3u8_id=None):
2110 'format_id': join_nonempty(m3u8_id, 'meta'),
2114 'preference': preference - 100 if preference else -100,
2116 'resolution': 'multiple',
2117 'format_note': 'Quality selection URL',
2120 def _report_ignoring_subs(self, name):
2121 self.report_warning(bug_reports_message(
2122 f'Ignoring subtitle tracks found in the {name} manifest; '
2123 'if any subtitle tracks are missing,'
2126 def _extract_m3u8_formats(self, *args, **kwargs):
2127 fmts, subs = self._extract_m3u8_formats_and_subtitles(*args, **kwargs)
2129 self._report_ignoring_subs('HLS')
2132 def _extract_m3u8_formats_and_subtitles(
2133 self, m3u8_url, video_id, ext=None, entry_protocol='m3u8_native',
2134 preference=None, quality=None, m3u8_id=None, note=None,
2135 errnote=None, fatal=True, live=False, data=None, headers={},
2138 res = self._download_webpage_handle(
2140 note='Downloading m3u8 information' if note is None else note,
2141 errnote='Failed to download m3u8 information' if errnote is None else errnote,
2142 fatal=fatal, data=data, headers=headers, query=query)
2147 m3u8_doc, urlh = res
2148 m3u8_url = urlh.geturl()
2150 return self._parse_m3u8_formats_and_subtitles(
2151 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
2152 preference=preference, quality=quality, m3u8_id=m3u8_id,
2153 note=note, errnote=errnote, fatal=fatal, live=live, data=data,
2154 headers=headers, query=query, video_id=video_id)
2156 def _parse_m3u8_formats_and_subtitles(
2157 self, m3u8_doc, m3u8_url=None, ext=None, entry_protocol='m3u8_native',
2158 preference=None, quality=None, m3u8_id=None, live=False, note=None,
2159 errnote=None, fatal=True, data=None, headers={}, query={},
2161 formats, subtitles = [], {}
2163 has_drm = re.search('|'.join([
2164 r'#EXT-X-FAXS-CM:', # Adobe Flash Access
2165 r'#EXT-X-(?:SESSION-)?KEY:.*?URI="skd
://', # Apple FairPlay
2168 def format_url(url):
2169 return url if re.match(r'^https?
://', url) else compat_urlparse.urljoin(m3u8_url, url)
2171 if self.get_param('hls_split_discontinuity
', False):
2172 def _extract_m3u8_playlist_indices(manifest_url=None, m3u8_doc=None):
2174 if not manifest_url:
2176 m3u8_doc = self._download_webpage(
2177 manifest_url, video_id, fatal=fatal, data=data, headers=headers,
2178 note=False, errnote='Failed to download m3u8 playlist information
')
2179 if m3u8_doc is False:
2181 return range(1 + sum(line.startswith('#EXT-X-DISCONTINUITY') for line in m3u8_doc.splitlines()))
2184 def _extract_m3u8_playlist_indices(*args
, **kwargs
):
2188 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
2189 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
2190 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
2192 # We should try extracting formats only from master playlists [1, 4.3.4],
2193 # i.e. playlists that describe available qualities. On the other hand
2194 # media playlists [1, 4.3.3] should be returned as is since they contain
2195 # just the media without qualities renditions.
2196 # Fortunately, master playlist can be easily distinguished from media
2197 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
2198 # master playlist tags MUST NOT appear in a media playlist and vice versa.
2199 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
2200 # media playlist and MUST NOT appear in master playlist thus we can
2201 # clearly detect media playlist with this criterion.
2203 if '#EXT-X-TARGETDURATION' in m3u8_doc
: # media playlist, return as is
2205 'format_id': join_nonempty(m3u8_id
, idx
),
2206 'format_index': idx
,
2207 'url': m3u8_url
or encode_data_uri(m3u8_doc
.encode('utf-8'), 'application/x-mpegurl'),
2209 'protocol': entry_protocol
,
2210 'preference': preference
,
2213 } for idx
in _extract_m3u8_playlist_indices(m3u8_doc
=m3u8_doc
)]
2215 return formats
, subtitles
2218 last_stream_inf
= {}
2220 def extract_media(x_media_line
):
2221 media
= parse_m3u8_attributes(x_media_line
)
2222 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
2223 media_type
, group_id
, name
= media
.get('TYPE'), media
.get('GROUP-ID'), media
.get('NAME')
2224 if not (media_type
and group_id
and name
):
2226 groups
.setdefault(group_id
, []).append(media
)
2227 # <https://tools.ietf.org/html/rfc8216#section-4.3.4.1>
2228 if media_type
== 'SUBTITLES':
2229 # According to RFC 8216 §4.3.4.2.1, URI is REQUIRED in the
2230 # EXT-X-MEDIA tag if the media type is SUBTITLES.
2231 # However, lack of URI has been spotted in the wild.
2232 # e.g. NebulaIE; see https://github.com/yt-dlp/yt-dlp/issues/339
2233 if not media
.get('URI'):
2235 url
= format_url(media
['URI'])
2238 'ext': determine_ext(url
),
2240 if sub_info
['ext'] == 'm3u8':
2241 # Per RFC 8216 §3.1, the only possible subtitle format m3u8
2242 # files may contain is WebVTT:
2243 # <https://tools.ietf.org/html/rfc8216#section-3.1>
2244 sub_info
['ext'] = 'vtt'
2245 sub_info
['protocol'] = 'm3u8_native'
2246 lang
= media
.get('LANGUAGE') or 'und'
2247 subtitles
.setdefault(lang
, []).append(sub_info
)
2248 if media_type
not in ('VIDEO', 'AUDIO'):
2250 media_url
= media
.get('URI')
2252 manifest_url
= format_url(media_url
)
2254 'format_id': join_nonempty(m3u8_id
, group_id
, name
, idx
),
2255 'format_note': name
,
2256 'format_index': idx
,
2257 'url': manifest_url
,
2258 'manifest_url': m3u8_url
,
2259 'language': media
.get('LANGUAGE'),
2261 'protocol': entry_protocol
,
2262 'preference': preference
,
2264 'vcodec': 'none' if media_type
== 'AUDIO' else None,
2265 } for idx
in _extract_m3u8_playlist_indices(manifest_url
))
2267 def build_stream_name():
2268 # Despite specification does not mention NAME attribute for
2269 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
2270 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
2271 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
2272 stream_name
= last_stream_inf
.get('NAME')
2275 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
2276 # from corresponding rendition group
2277 stream_group_id
= last_stream_inf
.get('VIDEO')
2278 if not stream_group_id
:
2280 stream_group
= groups
.get(stream_group_id
)
2281 if not stream_group
:
2282 return stream_group_id
2283 rendition
= stream_group
[0]
2284 return rendition
.get('NAME') or stream_group_id
2286 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
2287 # chance to detect video only formats when EXT-X-STREAM-INF tags
2288 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
2289 for line
in m3u8_doc
.splitlines():
2290 if line
.startswith('#EXT-X-MEDIA:'):
2293 for line
in m3u8_doc
.splitlines():
2294 if line
.startswith('#EXT-X-STREAM-INF:'):
2295 last_stream_inf
= parse_m3u8_attributes(line
)
2296 elif line
.startswith('#') or not line
.strip():
2299 tbr
= float_or_none(
2300 last_stream_inf
.get('AVERAGE-BANDWIDTH')
2301 or last_stream_inf
.get('BANDWIDTH'), scale
=1000)
2302 manifest_url
= format_url(line
.strip())
2304 for idx
in _extract_m3u8_playlist_indices(manifest_url
):
2305 format_id
= [m3u8_id
, None, idx
]
2306 # Bandwidth of live streams may differ over time thus making
2307 # format_id unpredictable. So it's better to keep provided
2310 stream_name
= build_stream_name()
2311 format_id
[1] = stream_name
or '%d' % (tbr
or len(formats
))
2313 'format_id': join_nonempty(*format_id
),
2314 'format_index': idx
,
2315 'url': manifest_url
,
2316 'manifest_url': m3u8_url
,
2319 'fps': float_or_none(last_stream_inf
.get('FRAME-RATE')),
2320 'protocol': entry_protocol
,
2321 'preference': preference
,
2324 resolution
= last_stream_inf
.get('RESOLUTION')
2326 mobj
= re
.search(r
'(?P<width>\d+)[xX](?P<height>\d+)', resolution
)
2328 f
['width'] = int(mobj
.group('width'))
2329 f
['height'] = int(mobj
.group('height'))
2330 # Unified Streaming Platform
2332 r
'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f
['url'])
2334 abr
, vbr
= mobj
.groups()
2335 abr
, vbr
= float_or_none(abr
, 1000), float_or_none(vbr
, 1000)
2340 codecs
= parse_codecs(last_stream_inf
.get('CODECS'))
2342 audio_group_id
= last_stream_inf
.get('AUDIO')
2343 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
2344 # references a rendition group MUST have a CODECS attribute.
2345 # However, this is not always respected, for example, [2]
2346 # contains EXT-X-STREAM-INF tag which references AUDIO
2347 # rendition group but does not have CODECS and despite
2348 # referencing an audio group it represents a complete
2349 # (with audio and video) format. So, for such cases we will
2350 # ignore references to rendition groups and treat them
2351 # as complete formats.
2352 if audio_group_id
and codecs
and f
.get('vcodec') != 'none':
2353 audio_group
= groups
.get(audio_group_id
)
2354 if audio_group
and audio_group
[0].get('URI'):
2355 # TODO: update acodec for audio only formats with
2357 f
['acodec'] = 'none'
2358 if not f
.get('ext'):
2359 f
['ext'] = 'm4a' if f
.get('vcodec') == 'none' else 'mp4'
2363 progressive_uri
= last_stream_inf
.get('PROGRESSIVE-URI')
2366 del http_f
['manifest_url']
2368 'format_id': f
['format_id'].replace('hls-', 'http-'),
2370 'url': progressive_uri
,
2372 formats
.append(http_f
)
2374 last_stream_inf
= {}
2375 return formats
, subtitles
2377 def _extract_m3u8_vod_duration(
2378 self
, m3u8_vod_url
, video_id
, note
=None, errnote
=None, data
=None, headers
={}, query={}
):
2380 m3u8_vod
= self
._download
_webpage
(
2381 m3u8_vod_url
, video_id
,
2382 note
='Downloading m3u8 VOD manifest' if note
is None else note
,
2383 errnote
='Failed to download VOD manifest' if errnote
is None else errnote
,
2384 fatal
=False, data
=data
, headers
=headers
, query
=query
)
2386 return self
._parse
_m
3u8_vod
_duration
(m3u8_vod
or '', video_id
)
2388 def _parse_m3u8_vod_duration(self
, m3u8_vod
, video_id
):
2389 if '#EXT-X-PLAYLIST-TYPE:VOD' not in m3u8_vod
:
2393 float(line
[len('#EXTINF:'):].split(',')[0])
2394 for line
in m3u8_vod
.splitlines() if line
.startswith('#EXTINF:'))) or None
2397 def _xpath_ns(path
, namespace
=None):
2401 for c
in path
.split('/'):
2402 if not c
or c
== '.':
2405 out
.append('{%s}%s' % (namespace
, c
))
2406 return '/'.join(out
)
2408 def _extract_smil_formats_and_subtitles(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None, transform_source
=None):
2409 smil
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
, transform_source
=transform_source
)
2415 namespace
= self
._parse
_smil
_namespace
(smil
)
2417 fmts
= self
._parse
_smil
_formats
(
2418 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2419 subs
= self
._parse
_smil
_subtitles
(
2420 smil
, namespace
=namespace
)
2424 def _extract_smil_formats(self
, *args
, **kwargs
):
2425 fmts
, subs
= self
._extract
_smil
_formats
_and
_subtitles
(*args
, **kwargs
)
2427 self
._report
_ignoring
_subs
('SMIL')
2430 def _extract_smil_info(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None):
2431 smil
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
)
2434 return self
._parse
_smil
(smil
, smil_url
, video_id
, f4m_params
=f4m_params
)
2436 def _download_smil(self
, smil_url
, video_id
, fatal
=True, transform_source
=None):
2437 return self
._download
_xml
(
2438 smil_url
, video_id
, 'Downloading SMIL file',
2439 'Unable to download SMIL file', fatal
=fatal
, transform_source
=transform_source
)
2441 def _parse_smil(self
, smil
, smil_url
, video_id
, f4m_params
=None):
2442 namespace
= self
._parse
_smil
_namespace
(smil
)
2444 formats
= self
._parse
_smil
_formats
(
2445 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2446 subtitles
= self
._parse
_smil
_subtitles
(smil
, namespace
=namespace
)
2448 video_id
= os
.path
.splitext(url_basename(smil_url
))[0]
2452 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2453 name
= meta
.attrib
.get('name')
2454 content
= meta
.attrib
.get('content')
2455 if not name
or not content
:
2457 if not title
and name
== 'title':
2459 elif not description
and name
in ('description', 'abstract'):
2460 description
= content
2461 elif not upload_date
and name
== 'date':
2462 upload_date
= unified_strdate(content
)
2465 'id': image
.get('type'),
2466 'url': image
.get('src'),
2467 'width': int_or_none(image
.get('width')),
2468 'height': int_or_none(image
.get('height')),
2469 } for image
in smil
.findall(self
._xpath
_ns
('.//image', namespace
)) if image
.get('src')]
2473 'title': title
or video_id
,
2474 'description': description
,
2475 'upload_date': upload_date
,
2476 'thumbnails': thumbnails
,
2478 'subtitles': subtitles
,
2481 def _parse_smil_namespace(self
, smil
):
2482 return self
._search
_regex
(
2483 r
'(?i)^{([^}]+)?}smil$', smil
.tag
, 'namespace', default
=None)
2485 def _parse_smil_formats(self
, smil
, smil_url
, video_id
, namespace
=None, f4m_params
=None, transform_rtmp_url
=None):
2487 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2488 b
= meta
.get('base') or meta
.get('httpBase')
2500 media
= smil
.findall(self
._xpath
_ns
('.//video', namespace
)) + smil
.findall(self
._xpath
_ns
('.//audio', namespace
))
2501 for medium
in media
:
2502 src
= medium
.get('src')
2503 if not src
or src
in srcs
:
2507 bitrate
= float_or_none(medium
.get('system-bitrate') or medium
.get('systemBitrate'), 1000)
2508 filesize
= int_or_none(medium
.get('size') or medium
.get('fileSize'))
2509 width
= int_or_none(medium
.get('width'))
2510 height
= int_or_none(medium
.get('height'))
2511 proto
= medium
.get('proto')
2512 ext
= medium
.get('ext')
2513 src_ext
= determine_ext(src
)
2514 streamer
= medium
.get('streamer') or base
2516 if proto
== 'rtmp' or streamer
.startswith('rtmp'):
2522 'format_id': 'rtmp-%d' % (rtmp_count
if bitrate
is None else bitrate
),
2524 'filesize': filesize
,
2528 if transform_rtmp_url
:
2529 streamer
, src
= transform_rtmp_url(streamer
, src
)
2530 formats
[-1].update({
2536 src_url
= src
if src
.startswith('http') else compat_urlparse
.urljoin(base
, src
)
2537 src_url
= src_url
.strip()
2539 if proto
== 'm3u8' or src_ext
== 'm3u8':
2540 m3u8_formats
= self
._extract
_m
3u8_formats
(
2541 src_url
, video_id
, ext
or 'mp4', m3u8_id
='hls', fatal
=False)
2542 if len(m3u8_formats
) == 1:
2544 m3u8_formats
[0].update({
2545 'format_id': 'hls-%d' % (m3u8_count
if bitrate
is None else bitrate
),
2550 formats
.extend(m3u8_formats
)
2551 elif src_ext
== 'f4m':
2556 'plugin': 'flowplayer-3.2.0.1',
2558 f4m_url
+= '&' if '?' in f4m_url
else '?'
2559 f4m_url
+= compat_urllib_parse_urlencode(f4m_params
)
2560 formats
.extend(self
._extract
_f
4m
_formats
(f4m_url
, video_id
, f4m_id
='hds', fatal
=False))
2561 elif src_ext
== 'mpd':
2562 formats
.extend(self
._extract
_mpd
_formats
(
2563 src_url
, video_id
, mpd_id
='dash', fatal
=False))
2564 elif re
.search(r
'\.ism/[Mm]anifest', src_url
):
2565 formats
.extend(self
._extract
_ism
_formats
(
2566 src_url
, video_id
, ism_id
='mss', fatal
=False))
2567 elif src_url
.startswith('http') and self
._is
_valid
_url
(src
, video_id
):
2571 'ext': ext
or src_ext
or 'flv',
2572 'format_id': 'http-%d' % (bitrate
or http_count
),
2574 'filesize': filesize
,
2579 for medium
in smil
.findall(self
._xpath
_ns
('.//imagestream', namespace
)):
2580 src
= medium
.get('src')
2581 if not src
or src
in srcs
:
2587 'format_id': 'imagestream-%d' % (imgs_count
),
2589 'ext': mimetype2ext(medium
.get('type')),
2592 'width': int_or_none(medium
.get('width')),
2593 'height': int_or_none(medium
.get('height')),
2594 'format_note': 'SMIL storyboards',
2599 def _parse_smil_subtitles(self
, smil
, namespace
=None, subtitles_lang
='en'):
2602 for num
, textstream
in enumerate(smil
.findall(self
._xpath
_ns
('.//textstream', namespace
))):
2603 src
= textstream
.get('src')
2604 if not src
or src
in urls
:
2607 ext
= textstream
.get('ext') or mimetype2ext(textstream
.get('type')) or determine_ext(src
)
2608 lang
= textstream
.get('systemLanguage') or textstream
.get('systemLanguageName') or textstream
.get('lang') or subtitles_lang
2609 subtitles
.setdefault(lang
, []).append({
2615 def _extract_xspf_playlist(self
, xspf_url
, playlist_id
, fatal
=True):
2616 xspf
= self
._download
_xml
(
2617 xspf_url
, playlist_id
, 'Downloading xpsf playlist',
2618 'Unable to download xspf manifest', fatal
=fatal
)
2621 return self
._parse
_xspf
(
2622 xspf
, playlist_id
, xspf_url
=xspf_url
,
2623 xspf_base_url
=base_url(xspf_url
))
2625 def _parse_xspf(self
, xspf_doc
, playlist_id
, xspf_url
=None, xspf_base_url
=None):
2627 'xspf': 'http://xspf.org/ns/0/',
2628 's1': 'http://static.streamone.nl/player/ns/0',
2632 for track
in xspf_doc
.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP
)):
2634 track
, xpath_with_ns('./xspf:title', NS_MAP
), 'title', default
=playlist_id
)
2635 description
= xpath_text(
2636 track
, xpath_with_ns('./xspf:annotation', NS_MAP
), 'description')
2637 thumbnail
= xpath_text(
2638 track
, xpath_with_ns('./xspf:image', NS_MAP
), 'thumbnail')
2639 duration
= float_or_none(
2640 xpath_text(track
, xpath_with_ns('./xspf:duration', NS_MAP
), 'duration'), 1000)
2643 for location
in track
.findall(xpath_with_ns('./xspf:location', NS_MAP
)):
2644 format_url
= urljoin(xspf_base_url
, location
.text
)
2649 'manifest_url': xspf_url
,
2650 'format_id': location
.get(xpath_with_ns('s1:label', NS_MAP
)),
2651 'width': int_or_none(location
.get(xpath_with_ns('s1:width', NS_MAP
))),
2652 'height': int_or_none(location
.get(xpath_with_ns('s1:height', NS_MAP
))),
2654 self
._sort
_formats
(formats
)
2659 'description': description
,
2660 'thumbnail': thumbnail
,
2661 'duration': duration
,
2666 def _extract_mpd_formats(self
, *args
, **kwargs
):
2667 fmts
, subs
= self
._extract
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2669 self
._report
_ignoring
_subs
('DASH')
2672 def _extract_mpd_formats_and_subtitles(
2673 self
, mpd_url
, video_id
, mpd_id
=None, note
=None, errnote
=None,
2674 fatal
=True, data
=None, headers
={}, query={}
):
2675 res
= self
._download
_xml
_handle
(
2677 note
='Downloading MPD manifest' if note
is None else note
,
2678 errnote
='Failed to download MPD manifest' if errnote
is None else errnote
,
2679 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
2685 mpd_base_url
= base_url(urlh
.geturl())
2687 return self
._parse
_mpd
_formats
_and
_subtitles
(
2688 mpd_doc
, mpd_id
, mpd_base_url
, mpd_url
)
2690 def _parse_mpd_formats(self
, *args
, **kwargs
):
2691 fmts
, subs
= self
._parse
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2693 self
._report
_ignoring
_subs
('DASH')
2696 def _parse_mpd_formats_and_subtitles(
2697 self
, mpd_doc
, mpd_id
=None, mpd_base_url
='', mpd_url
=None):
2699 Parse formats from MPD manifest.
2701 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
2702 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
2703 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
2705 if not self
.get_param('dynamic_mpd', True):
2706 if mpd_doc
.get('type') == 'dynamic':
2709 namespace
= self
._search
_regex
(r
'(?i)^{([^}]+)?}MPD$', mpd_doc
.tag
, 'namespace', default
=None)
2712 return self
._xpath
_ns
(path
, namespace
)
2714 def is_drm_protected(element
):
2715 return element
.find(_add_ns('ContentProtection')) is not None
2717 def extract_multisegment_info(element
, ms_parent_info
):
2718 ms_info
= ms_parent_info
.copy()
2720 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2721 # common attributes and elements. We will only extract relevant
2723 def extract_common(source
):
2724 segment_timeline
= source
.find(_add_ns('SegmentTimeline'))
2725 if segment_timeline
is not None:
2726 s_e
= segment_timeline
.findall(_add_ns('S'))
2728 ms_info
['total_number'] = 0
2731 r
= int(s
.get('r', 0))
2732 ms_info
['total_number'] += 1 + r
2733 ms_info
['s'].append({
2734 't': int(s
.get('t', 0)),
2735 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2736 'd': int(s
.attrib
['d']),
2739 start_number
= source
.get('startNumber')
2741 ms_info
['start_number'] = int(start_number
)
2742 timescale
= source
.get('timescale')
2744 ms_info
['timescale'] = int(timescale
)
2745 segment_duration
= source
.get('duration')
2746 if segment_duration
:
2747 ms_info
['segment_duration'] = float(segment_duration
)
2749 def extract_Initialization(source
):
2750 initialization
= source
.find(_add_ns('Initialization'))
2751 if initialization
is not None:
2752 ms_info
['initialization_url'] = initialization
.attrib
['sourceURL']
2754 segment_list
= element
.find(_add_ns('SegmentList'))
2755 if segment_list
is not None:
2756 extract_common(segment_list
)
2757 extract_Initialization(segment_list
)
2758 segment_urls_e
= segment_list
.findall(_add_ns('SegmentURL'))
2760 ms_info
['segment_urls'] = [segment
.attrib
['media'] for segment
in segment_urls_e
]
2762 segment_template
= element
.find(_add_ns('SegmentTemplate'))
2763 if segment_template
is not None:
2764 extract_common(segment_template
)
2765 media
= segment_template
.get('media')
2767 ms_info
['media'] = media
2768 initialization
= segment_template
.get('initialization')
2770 ms_info
['initialization'] = initialization
2772 extract_Initialization(segment_template
)
2775 mpd_duration
= parse_duration(mpd_doc
.get('mediaPresentationDuration'))
2776 formats
, subtitles
= [], {}
2777 stream_numbers
= collections
.defaultdict(int)
2778 for period
in mpd_doc
.findall(_add_ns('Period')):
2779 period_duration
= parse_duration(period
.get('duration')) or mpd_duration
2780 period_ms_info
= extract_multisegment_info(period
, {
2784 for adaptation_set
in period
.findall(_add_ns('AdaptationSet')):
2785 adaption_set_ms_info
= extract_multisegment_info(adaptation_set
, period_ms_info
)
2786 for representation
in adaptation_set
.findall(_add_ns('Representation')):
2787 representation_attrib
= adaptation_set
.attrib
.copy()
2788 representation_attrib
.update(representation
.attrib
)
2789 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2790 mime_type
= representation_attrib
['mimeType']
2791 content_type
= representation_attrib
.get('contentType', mime_type
.split('/')[0])
2793 codecs
= parse_codecs(representation_attrib
.get('codecs', ''))
2794 if content_type
not in ('video', 'audio', 'text'):
2795 if mime_type
== 'image/jpeg':
2796 content_type
= mime_type
2797 elif codecs
['vcodec'] != 'none':
2798 content_type
= 'video'
2799 elif codecs
['acodec'] != 'none':
2800 content_type
= 'audio'
2801 elif codecs
.get('tcodec', 'none') != 'none':
2802 content_type
= 'text'
2803 elif mimetype2ext(mime_type
) in ('tt', 'dfxp', 'ttml', 'xml', 'json'):
2804 content_type
= 'text'
2806 self
.report_warning('Unknown MIME type %s in DASH manifest' % mime_type
)
2810 for element
in (representation
, adaptation_set
, period
, mpd_doc
):
2811 base_url_e
= element
.find(_add_ns('BaseURL'))
2812 if base_url_e
is not None:
2813 base_url
= base_url_e
.text
+ base_url
2814 if re
.match(r
'^https?://', base_url
):
2816 if mpd_base_url
and base_url
.startswith('/'):
2817 base_url
= compat_urlparse
.urljoin(mpd_base_url
, base_url
)
2818 elif mpd_base_url
and not re
.match(r
'^https?://', base_url
):
2819 if not mpd_base_url
.endswith('/'):
2821 base_url
= mpd_base_url
+ base_url
2822 representation_id
= representation_attrib
.get('id')
2823 lang
= representation_attrib
.get('lang')
2824 url_el
= representation
.find(_add_ns('BaseURL'))
2825 filesize
= int_or_none(url_el
.attrib
.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el
is not None else None)
2826 bandwidth
= int_or_none(representation_attrib
.get('bandwidth'))
2827 if representation_id
is not None:
2828 format_id
= representation_id
2830 format_id
= content_type
2832 format_id
= mpd_id
+ '-' + format_id
2833 if content_type
in ('video', 'audio'):
2835 'format_id': format_id
,
2836 'manifest_url': mpd_url
,
2837 'ext': mimetype2ext(mime_type
),
2838 'width': int_or_none(representation_attrib
.get('width')),
2839 'height': int_or_none(representation_attrib
.get('height')),
2840 'tbr': float_or_none(bandwidth
, 1000),
2841 'asr': int_or_none(representation_attrib
.get('audioSamplingRate')),
2842 'fps': int_or_none(representation_attrib
.get('frameRate')),
2843 'language': lang
if lang
not in ('mul', 'und', 'zxx', 'mis') else None,
2844 'format_note': 'DASH %s' % content_type
,
2845 'filesize': filesize
,
2846 'container': mimetype2ext(mime_type
) + '_dash',
2849 elif content_type
== 'text':
2851 'ext': mimetype2ext(mime_type
),
2852 'manifest_url': mpd_url
,
2853 'filesize': filesize
,
2855 elif content_type
== 'image/jpeg':
2856 # See test case in VikiIE
2857 # https://www.viki.com/videos/1175236v-choosing-spouse-by-lottery-episode-1
2859 'format_id': format_id
,
2861 'manifest_url': mpd_url
,
2862 'format_note': 'DASH storyboards (jpeg)',
2866 if is_drm_protected(adaptation_set
) or is_drm_protected(representation
):
2868 representation_ms_info
= extract_multisegment_info(representation
, adaption_set_ms_info
)
2870 def prepare_template(template_name
, identifiers
):
2871 tmpl
= representation_ms_info
[template_name
]
2872 # First of, % characters outside $...$ templates
2873 # must be escaped by doubling for proper processing
2874 # by % operator string formatting used further (see
2875 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2881 in_template
= not in_template
2882 elif c
== '%' and not in_template
:
2884 # Next, $...$ templates are translated to their
2885 # %(...) counterparts to be used with % operator
2886 if representation_id
is not None:
2887 t
= t
.replace('$RepresentationID$', representation_id
)
2888 t
= re
.sub(r
'\$(%s)\$' % '|'.join(identifiers
), r
'%(\1)d', t
)
2889 t
= re
.sub(r
'\$(%s)%%([^$]+)\$' % '|'.join(identifiers
), r
'%(\1)\2', t
)
2890 t
.replace('$$', '$')
2893 # @initialization is a regular template like @media one
2894 # so it should be handled just the same way (see
2895 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2896 if 'initialization' in representation_ms_info
:
2897 initialization_template
= prepare_template(
2899 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2900 # $Time$ shall not be included for @initialization thus
2901 # only $Bandwidth$ remains
2903 representation_ms_info
['initialization_url'] = initialization_template
% {
2904 'Bandwidth': bandwidth
,
2907 def location_key(location
):
2908 return 'url' if re
.match(r
'^https?://', location
) else 'path'
2910 if 'segment_urls' not in representation_ms_info
and 'media' in representation_ms_info
:
2912 media_template
= prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2913 media_location_key
= location_key(media_template
)
2915 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2916 # can't be used at the same time
2917 if '%(Number' in media_template
and 's' not in representation_ms_info
:
2918 segment_duration
= None
2919 if 'total_number' not in representation_ms_info
and 'segment_duration' in representation_ms_info
:
2920 segment_duration
= float_or_none(representation_ms_info
['segment_duration'], representation_ms_info
['timescale'])
2921 representation_ms_info
['total_number'] = int(math
.ceil(
2922 float_or_none(period_duration
, segment_duration
, default
=0)))
2923 representation_ms_info
['fragments'] = [{
2924 media_location_key
: media_template
% {
2925 'Number': segment_number
,
2926 'Bandwidth': bandwidth
,
2928 'duration': segment_duration
,
2929 } for segment_number
in range(
2930 representation_ms_info
['start_number'],
2931 representation_ms_info
['total_number'] + representation_ms_info
['start_number'])]
2933 # $Number*$ or $Time$ in media template with S list available
2934 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2935 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2936 representation_ms_info
['fragments'] = []
2939 segment_number
= representation_ms_info
['start_number']
2941 def add_segment_url():
2942 segment_url
= media_template
% {
2943 'Time': segment_time
,
2944 'Bandwidth': bandwidth
,
2945 'Number': segment_number
,
2947 representation_ms_info
['fragments'].append({
2948 media_location_key
: segment_url
,
2949 'duration': float_or_none(segment_d
, representation_ms_info
['timescale']),
2952 for num
, s
in enumerate(representation_ms_info
['s']):
2953 segment_time
= s
.get('t') or segment_time
2957 for r
in range(s
.get('r', 0)):
2958 segment_time
+= segment_d
2961 segment_time
+= segment_d
2962 elif 'segment_urls' in representation_ms_info
and 's' in representation_ms_info
:
2964 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2965 # or any YouTube dashsegments video
2968 timescale
= representation_ms_info
['timescale']
2969 for s
in representation_ms_info
['s']:
2970 duration
= float_or_none(s
['d'], timescale
)
2971 for r
in range(s
.get('r', 0) + 1):
2972 segment_uri
= representation_ms_info
['segment_urls'][segment_index
]
2974 location_key(segment_uri
): segment_uri
,
2975 'duration': duration
,
2978 representation_ms_info
['fragments'] = fragments
2979 elif 'segment_urls' in representation_ms_info
:
2980 # Segment URLs with no SegmentTimeline
2981 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2982 # https://github.com/ytdl-org/youtube-dl/pull/14844
2984 segment_duration
= float_or_none(
2985 representation_ms_info
['segment_duration'],
2986 representation_ms_info
['timescale']) if 'segment_duration' in representation_ms_info
else None
2987 for segment_url
in representation_ms_info
['segment_urls']:
2989 location_key(segment_url
): segment_url
,
2991 if segment_duration
:
2992 fragment
['duration'] = segment_duration
2993 fragments
.append(fragment
)
2994 representation_ms_info
['fragments'] = fragments
2995 # If there is a fragments key available then we correctly recognized fragmented media.
2996 # Otherwise we will assume unfragmented media with direct access. Technically, such
2997 # assumption is not necessarily correct since we may simply have no support for
2998 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2999 if 'fragments' in representation_ms_info
:
3001 # NB: mpd_url may be empty when MPD manifest is parsed from a string
3002 'url': mpd_url
or base_url
,
3003 'fragment_base_url': base_url
,
3005 'protocol': 'http_dash_segments' if mime_type
!= 'image/jpeg' else 'mhtml',
3007 if 'initialization_url' in representation_ms_info
:
3008 initialization_url
= representation_ms_info
['initialization_url']
3009 if not f
.get('url'):
3010 f
['url'] = initialization_url
3011 f
['fragments'].append({location_key(initialization_url): initialization_url}
)
3012 f
['fragments'].extend(representation_ms_info
['fragments'])
3013 if not period_duration
:
3014 period_duration
= try_get(
3015 representation_ms_info
,
3016 lambda r
: sum(frag
['duration'] for frag
in r
['fragments']), float)
3018 # Assuming direct URL to unfragmented media.
3020 if content_type
in ('video', 'audio', 'image/jpeg'):
3021 f
['manifest_stream_number'] = stream_numbers
[f
['url']]
3022 stream_numbers
[f
['url']] += 1
3024 elif content_type
== 'text':
3025 subtitles
.setdefault(lang
or 'und', []).append(f
)
3027 return formats
, subtitles
3029 def _extract_ism_formats(self
, *args
, **kwargs
):
3030 fmts
, subs
= self
._extract
_ism
_formats
_and
_subtitles
(*args
, **kwargs
)
3032 self
._report
_ignoring
_subs
('ISM')
3035 def _extract_ism_formats_and_subtitles(self
, ism_url
, video_id
, ism_id
=None, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
):
3036 res
= self
._download
_xml
_handle
(
3038 note
='Downloading ISM manifest' if note
is None else note
,
3039 errnote
='Failed to download ISM manifest' if errnote
is None else errnote
,
3040 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
3047 return self
._parse
_ism
_formats
_and
_subtitles
(ism_doc
, urlh
.geturl(), ism_id
)
3049 def _parse_ism_formats_and_subtitles(self
, ism_doc
, ism_url
, ism_id
=None):
3051 Parse formats from ISM manifest.
3053 1. [MS-SSTR]: Smooth Streaming Protocol,
3054 https://msdn.microsoft.com/en-us/library/ff469518.aspx
3056 if ism_doc
.get('IsLive') == 'TRUE':
3059 duration
= int(ism_doc
.attrib
['Duration'])
3060 timescale
= int_or_none(ism_doc
.get('TimeScale')) or 10000000
3064 for stream
in ism_doc
.findall('StreamIndex'):
3065 stream_type
= stream
.get('Type')
3066 if stream_type
not in ('video', 'audio', 'text'):
3068 url_pattern
= stream
.attrib
['Url']
3069 stream_timescale
= int_or_none(stream
.get('TimeScale')) or timescale
3070 stream_name
= stream
.get('Name')
3071 stream_language
= stream
.get('Language', 'und')
3072 for track
in stream
.findall('QualityLevel'):
3073 fourcc
= track
.get('FourCC') or ('AACL' if track
.get('AudioTag') == '255' else None)
3074 # TODO: add support for WVC1 and WMAP
3075 if fourcc
not in ('H264', 'AVC1', 'AACL', 'TTML'):
3076 self
.report_warning('%s is not a supported codec' % fourcc
)
3078 tbr
= int(track
.attrib
['Bitrate']) // 1000
3079 # [1] does not mention Width and Height attributes. However,
3080 # they're often present while MaxWidth and MaxHeight are
3081 # missing, so should be used as fallbacks
3082 width
= int_or_none(track
.get('MaxWidth') or track
.get('Width'))
3083 height
= int_or_none(track
.get('MaxHeight') or track
.get('Height'))
3084 sampling_rate
= int_or_none(track
.get('SamplingRate'))
3086 track_url_pattern
= re
.sub(r
'{[Bb]itrate}', track
.attrib
['Bitrate'], url_pattern
)
3087 track_url_pattern
= compat_urlparse
.urljoin(ism_url
, track_url_pattern
)
3093 stream_fragments
= stream
.findall('c')
3094 for stream_fragment_index
, stream_fragment
in enumerate(stream_fragments
):
3095 fragment_ctx
['time'] = int_or_none(stream_fragment
.get('t')) or fragment_ctx
['time']
3096 fragment_repeat
= int_or_none(stream_fragment
.get('r')) or 1
3097 fragment_ctx
['duration'] = int_or_none(stream_fragment
.get('d'))
3098 if not fragment_ctx
['duration']:
3100 next_fragment_time
= int(stream_fragment
[stream_fragment_index
+ 1].attrib
['t'])
3102 next_fragment_time
= duration
3103 fragment_ctx
['duration'] = (next_fragment_time
- fragment_ctx
['time']) / fragment_repeat
3104 for _
in range(fragment_repeat
):
3106 'url': re
.sub(r
'{start[ _]time}', compat_str(fragment_ctx
['time']), track_url_pattern
),
3107 'duration': fragment_ctx
['duration'] / stream_timescale
,
3109 fragment_ctx
['time'] += fragment_ctx
['duration']
3111 if stream_type
== 'text':
3112 subtitles
.setdefault(stream_language
, []).append({
3116 'manifest_url': ism_url
,
3117 'fragments': fragments
,
3118 '_download_params': {
3119 'stream_type': stream_type
,
3120 'duration': duration
,
3121 'timescale': stream_timescale
,
3123 'language': stream_language
,
3124 'codec_private_data': track
.get('CodecPrivateData'),
3127 elif stream_type
in ('video', 'audio'):
3129 'format_id': join_nonempty(ism_id
, stream_name
, tbr
),
3131 'manifest_url': ism_url
,
3132 'ext': 'ismv' if stream_type
== 'video' else 'isma',
3136 'asr': sampling_rate
,
3137 'vcodec': 'none' if stream_type
== 'audio' else fourcc
,
3138 'acodec': 'none' if stream_type
== 'video' else fourcc
,
3140 'fragments': fragments
,
3141 'has_drm': ism_doc
.find('Protection') is not None,
3142 '_download_params': {
3143 'stream_type': stream_type
,
3144 'duration': duration
,
3145 'timescale': stream_timescale
,
3146 'width': width
or 0,
3147 'height': height
or 0,
3149 'language': stream_language
,
3150 'codec_private_data': track
.get('CodecPrivateData'),
3151 'sampling_rate': sampling_rate
,
3152 'channels': int_or_none(track
.get('Channels', 2)),
3153 'bits_per_sample': int_or_none(track
.get('BitsPerSample', 16)),
3154 'nal_unit_length_field': int_or_none(track
.get('NALUnitLengthField', 4)),
3157 return formats
, subtitles
3159 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):
3160 def absolute_url(item_url
):
3161 return urljoin(base_url
, item_url
)
3163 def parse_content_type(content_type
):
3164 if not content_type
:
3166 ctr
= re
.search(r
'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type
)
3168 mimetype
, codecs
= ctr
.groups()
3169 f
= parse_codecs(codecs
)
3170 f
['ext'] = mimetype2ext(mimetype
)
3174 def _media_formats(src
, cur_media_type
, type_info
={}):
3175 full_url
= absolute_url(src
)
3176 ext
= type_info
.get('ext') or determine_ext(full_url
)
3178 is_plain_url
= False
3179 formats
= self
._extract
_m
3u8_formats
(
3180 full_url
, video_id
, ext
='mp4',
3181 entry_protocol
=m3u8_entry_protocol
, m3u8_id
=m3u8_id
,
3182 preference
=preference
, quality
=quality
, fatal
=False)
3184 is_plain_url
= False
3185 formats
= self
._extract
_mpd
_formats
(
3186 full_url
, video_id
, mpd_id
=mpd_id
, fatal
=False)
3191 'vcodec': 'none' if cur_media_type
== 'audio' else None,
3193 return is_plain_url
, formats
3196 # amp-video and amp-audio are very similar to their HTML5 counterparts
3197 # so we wll include them right here (see
3198 # https://www.ampproject.org/docs/reference/components/amp-video)
3199 # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/
3200 _MEDIA_TAG_NAME_RE
= r
'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)'
3201 media_tags
= [(media_tag
, media_tag_name
, media_type
, '')
3202 for media_tag
, media_tag_name
, media_type
3203 in re
.findall(r
'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE
, webpage
)]
3204 media_tags
.extend(re
.findall(
3205 # We only allow video|audio followed by a whitespace or '>'.
3206 # Allowing more characters may end up in significant slow down (see
3207 # https://github.com/ytdl-org/youtube-dl/issues/11979, example URL:
3208 # http://www.porntrex.com/maps/videositemap.xml).
3209 r
'(?s)(<(?P<tag>%s)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>' % _MEDIA_TAG_NAME_RE
, webpage
))
3210 for media_tag
, _
, media_type
, media_content
in media_tags
:
3215 media_attributes
= extract_attributes(media_tag
)
3216 src
= strip_or_none(media_attributes
.get('src'))
3218 _
, formats
= _media_formats(src
, media_type
)
3219 media_info
['formats'].extend(formats
)
3220 media_info
['thumbnail'] = absolute_url(media_attributes
.get('poster'))
3222 for source_tag
in re
.findall(r
'<source[^>]+>', media_content
):
3223 s_attr
= extract_attributes(source_tag
)
3224 # data-video-src and data-src are non standard but seen
3225 # several times in the wild
3226 src
= strip_or_none(dict_get(s_attr
, ('src', 'data-video-src', 'data-src')))
3229 f
= parse_content_type(s_attr
.get('type'))
3230 is_plain_url
, formats
= _media_formats(src
, media_type
, f
)
3232 # width, height, res, label and title attributes are
3233 # all not standard but seen several times in the wild
3236 for lbl
in ('label', 'title')
3237 if str_or_none(s_attr
.get(lbl
))
3239 width
= int_or_none(s_attr
.get('width'))
3240 height
= (int_or_none(s_attr
.get('height'))
3241 or int_or_none(s_attr
.get('res')))
3242 if not width
or not height
:
3244 resolution
= parse_resolution(lbl
)
3247 width
= width
or resolution
.get('width')
3248 height
= height
or resolution
.get('height')
3250 tbr
= parse_bitrate(lbl
)
3259 'format_id': s_attr
.get('label') or s_attr
.get('title'),
3261 f
.update(formats
[0])
3262 media_info
['formats'].append(f
)
3264 media_info
['formats'].extend(formats
)
3265 for track_tag
in re
.findall(r
'<track[^>]+>', media_content
):
3266 track_attributes
= extract_attributes(track_tag
)
3267 kind
= track_attributes
.get('kind')
3268 if not kind
or kind
in ('subtitles', 'captions'):
3269 src
= strip_or_none(track_attributes
.get('src'))
3272 lang
= track_attributes
.get('srclang') or track_attributes
.get('lang') or track_attributes
.get('label')
3273 media_info
['subtitles'].setdefault(lang
, []).append({
3274 'url': absolute_url(src
),
3276 for f
in media_info
['formats']:
3277 f
.setdefault('http_headers', {})['Referer'] = base_url
3278 if media_info
['formats'] or media_info
['subtitles']:
3279 entries
.append(media_info
)
3282 def _extract_akamai_formats(self
, *args
, **kwargs
):
3283 fmts
, subs
= self
._extract
_akamai
_formats
_and
_subtitles
(*args
, **kwargs
)
3285 self
._report
_ignoring
_subs
('akamai')
3288 def _extract_akamai_formats_and_subtitles(self
, manifest_url
, video_id
, hosts
={}):
3289 signed
= 'hdnea=' in manifest_url
3291 # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
3292 manifest_url
= re
.sub(
3293 r
'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
3294 '', manifest_url
).strip('?')
3299 hdcore_sign
= 'hdcore=3.7.0'
3300 f4m_url
= re
.sub(r
'(https?://[^/]+)/i/', r
'\1/z/', manifest_url
).replace('/master.m3u8', '/manifest.f4m')
3301 hds_host
= hosts
.get('hds')
3303 f4m_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hds_host
, f4m_url
)
3304 if 'hdcore=' not in f4m_url
:
3305 f4m_url
+= ('&' if '?' in f4m_url
else '?') + hdcore_sign
3306 f4m_formats
= self
._extract
_f
4m
_formats
(
3307 f4m_url
, video_id
, f4m_id
='hds', fatal
=False)
3308 for entry
in f4m_formats
:
3309 entry
.update({'extra_param_to_segment_url': hdcore_sign}
)
3310 formats
.extend(f4m_formats
)
3312 m3u8_url
= re
.sub(r
'(https?://[^/]+)/z/', r
'\1/i/', manifest_url
).replace('/manifest.f4m', '/master.m3u8')
3313 hls_host
= hosts
.get('hls')
3315 m3u8_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hls_host
, m3u8_url
)
3316 m3u8_formats
, m3u8_subtitles
= self
._extract
_m
3u8_formats
_and
_subtitles
(
3317 m3u8_url
, video_id
, 'mp4', 'm3u8_native',
3318 m3u8_id
='hls', fatal
=False)
3319 formats
.extend(m3u8_formats
)
3320 subtitles
= self
._merge
_subtitles
(subtitles
, m3u8_subtitles
)
3322 http_host
= hosts
.get('http')
3323 if http_host
and m3u8_formats
and not signed
:
3324 REPL_REGEX
= r
'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
3325 qualities
= re
.match(REPL_REGEX
, m3u8_url
).group(2).split(',')
3326 qualities_length
= len(qualities
)
3327 if len(m3u8_formats
) in (qualities_length
, qualities_length
+ 1):
3329 for f
in m3u8_formats
:
3330 if f
['vcodec'] != 'none':
3331 for protocol
in ('http', 'https'):
3333 del http_f
['manifest_url']
3335 REPL_REGEX
, protocol
+ r
'://%s/\g<1>%s\3' % (http_host
, qualities
[i
]), f
['url'])
3337 'format_id': http_f
['format_id'].replace('hls-', protocol
+ '-'),
3339 'protocol': protocol
,
3341 formats
.append(http_f
)
3344 return formats
, subtitles
3346 def _extract_wowza_formats(self
, url
, video_id
, m3u8_entry_protocol
='m3u8_native', skip_protocols
=[]):
3347 query
= compat_urlparse
.urlparse(url
).query
3348 url
= re
.sub(r
'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url
)
3350 r
'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url
)
3351 url_base
= mobj
.group('url')
3352 http_base_url
= '%s%s:%s' % ('http', mobj
.group('s') or '', url_base
)
3355 def manifest_url(manifest
):
3356 m_url
= '%s/%s' % (http_base_url
, manifest
)
3358 m_url
+= '?%s' % query
3361 if 'm3u8' not in skip_protocols
:
3362 formats
.extend(self
._extract
_m
3u8_formats
(
3363 manifest_url('playlist.m3u8'), video_id
, 'mp4',
3364 m3u8_entry_protocol
, m3u8_id
='hls', fatal
=False))
3365 if 'f4m' not in skip_protocols
:
3366 formats
.extend(self
._extract
_f
4m
_formats
(
3367 manifest_url('manifest.f4m'),
3368 video_id
, f4m_id
='hds', fatal
=False))
3369 if 'dash' not in skip_protocols
:
3370 formats
.extend(self
._extract
_mpd
_formats
(
3371 manifest_url('manifest.mpd'),
3372 video_id
, mpd_id
='dash', fatal
=False))
3373 if re
.search(r
'(?:/smil:|\.smil)', url_base
):
3374 if 'smil' not in skip_protocols
:
3375 rtmp_formats
= self
._extract
_smil
_formats
(
3376 manifest_url('jwplayer.smil'),
3377 video_id
, fatal
=False)
3378 for rtmp_format
in rtmp_formats
:
3379 rtsp_format
= rtmp_format
.copy()
3380 rtsp_format
['url'] = '%s/%s' % (rtmp_format
['url'], rtmp_format
['play_path'])
3381 del rtsp_format
['play_path']
3382 del rtsp_format
['ext']
3383 rtsp_format
.update({
3384 'url': rtsp_format
['url'].replace('rtmp://', 'rtsp://'),
3385 'format_id': rtmp_format
['format_id'].replace('rtmp', 'rtsp'),
3388 formats
.extend([rtmp_format
, rtsp_format
])
3390 for protocol
in ('rtmp', 'rtsp'):
3391 if protocol
not in skip_protocols
:
3393 'url': '%s:%s' % (protocol
, url_base
),
3394 'format_id': protocol
,
3395 'protocol': protocol
,
3399 def _find_jwplayer_data(self
, webpage
, video_id
=None, transform_source
=js_to_json
):
3401 r
'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P
=quote
)\
)(?
!</script
>).*?\
.setup\s
*\
((?P
<options
>[^
)]+)\
)',
3405 jwplayer_data = self._parse_json(mobj.group('options
'),
3407 transform_source=transform_source)
3408 except ExtractorError:
3411 if isinstance(jwplayer_data, dict):
3412 return jwplayer_data
3414 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
3415 jwplayer_data = self._find_jwplayer_data(
3416 webpage, video_id, transform_source=js_to_json)
3417 return self._parse_jwplayer_data(
3418 jwplayer_data, video_id, *args, **kwargs)
3420 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
3421 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3422 # JWPlayer backward compatibility: flattened playlists
3423 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
3424 if 'playlist
' not in jwplayer_data:
3425 jwplayer_data = {'playlist': [jwplayer_data]}
3429 # JWPlayer backward compatibility: single playlist item
3430 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
3431 if not isinstance(jwplayer_data['playlist
'], list):
3432 jwplayer_data['playlist
'] = [jwplayer_data['playlist
']]
3434 for video_data in jwplayer_data['playlist
']:
3435 # JWPlayer backward compatibility: flattened sources
3436 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
3437 if 'sources
' not in video_data:
3438 video_data['sources
'] = [video_data]
3440 this_video_id = video_id or video_data['mediaid
']
3442 formats = self._parse_jwplayer_formats(
3443 video_data['sources
'], video_id=this_video_id, m3u8_id=m3u8_id,
3444 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
3447 tracks = video_data.get('tracks
')
3448 if tracks and isinstance(tracks, list):
3449 for track in tracks:
3450 if not isinstance(track, dict):
3452 track_kind = track.get('kind
')
3453 if not track_kind or not isinstance(track_kind, compat_str):
3455 if track_kind.lower() not in ('captions
', 'subtitles
'):
3457 track_url = urljoin(base_url, track.get('file'))
3460 subtitles.setdefault(track.get('label
') or 'en
', []).append({
3461 'url
': self._proto_relative_url(track_url)
3465 'id': this_video_id,
3466 'title
': unescapeHTML(video_data['title
'] if require_title else video_data.get('title
')),
3467 'description
': clean_html(video_data.get('description
')),
3468 'thumbnail
': urljoin(base_url, self._proto_relative_url(video_data.get('image
'))),
3469 'timestamp
': int_or_none(video_data.get('pubdate
')),
3470 'duration
': float_or_none(jwplayer_data.get('duration
') or video_data.get('duration
')),
3471 'subtitles
': subtitles,
3473 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
3474 if len(formats) == 1 and re.search(r'^
(?
:http|
//).*(?
:youtube\
.com|youtu\
.be
)/.+', formats[0]['url
']):
3476 '_type
': 'url_transparent
',
3477 'url
': formats[0]['url
'],
3480 self._sort_formats(formats)
3481 entry['formats
'] = formats
3482 entries.append(entry)
3483 if len(entries) == 1:
3486 return self.playlist_result(entries)
3488 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
3489 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3492 for source in jwplayer_sources_data:
3493 if not isinstance(source, dict):
3495 source_url = urljoin(
3496 base_url, self._proto_relative_url(source.get('file')))
3497 if not source_url or source_url in urls:
3499 urls.append(source_url)
3500 source_type = source.get('type') or ''
3501 ext = mimetype2ext(source_type) or determine_ext(source_url)
3502 if source_type == 'hls
' or ext == 'm3u8
':
3503 formats.extend(self._extract_m3u8_formats(
3504 source_url, video_id, 'mp4
', entry_protocol='m3u8_native
',
3505 m3u8_id=m3u8_id, fatal=False))
3506 elif source_type == 'dash
' or ext == 'mpd
':
3507 formats.extend(self._extract_mpd_formats(
3508 source_url, video_id, mpd_id=mpd_id, fatal=False))
3510 formats.extend(self._extract_smil_formats(
3511 source_url, video_id, fatal=False))
3512 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3513 elif source_type.startswith('audio
') or ext in (
3514 'oga
', 'aac
', 'mp3
', 'mpeg
', 'vorbis
'):
3521 height = int_or_none(source.get('height
'))
3523 # Often no height is provided but there is a label in
3524 # format like "1080p", "720p SD", or 1080.
3525 height = int_or_none(self._search_regex(
3526 r'^
(\d{3,4}
)[pP
]?
(?
:\b|$
)', compat_str(source.get('label
') or ''),
3527 'height
', default=None))
3530 'width
': int_or_none(source.get('width
')),
3532 'tbr
': int_or_none(source.get('bitrate
')),
3535 if source_url.startswith('rtmp
'):
3536 a_format['ext
'] = 'flv
'
3537 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3538 # of jwplayer.flash.swf
3539 rtmp_url_parts = re.split(
3540 r'((?
:mp4|mp3|flv
):)', source_url, 1)
3541 if len(rtmp_url_parts) == 3:
3542 rtmp_url, prefix, play_path = rtmp_url_parts
3545 'play_path
': prefix + play_path,
3548 a_format.update(rtmp_params)
3549 formats.append(a_format)
3552 def _live_title(self, name):
3553 self._downloader.deprecation_warning('yt_dlp
.InfoExtractor
._live
_title
is deprecated
and does
not work
as expected
')
3556 def _int(self, v, name, fatal=False, **kwargs):
3557 res = int_or_none(v, **kwargs)
3559 msg = 'Failed to extract
%s: Could
not parse value
%r' % (name, v)
3561 raise ExtractorError(msg)
3563 self.report_warning(msg)
3566 def _float(self, v, name, fatal=False, **kwargs):
3567 res = float_or_none(v, **kwargs)
3569 msg = 'Failed to extract
%s: Could
not parse value
%r' % (name, v)
3571 raise ExtractorError(msg)
3573 self.report_warning(msg)
3576 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
3577 path='/', secure=False, discard=False, rest={}, **kwargs):
3578 cookie = compat_cookiejar_Cookie(
3579 0, name, value, port, port is not None, domain, True,
3580 domain.startswith('.'), path, True, secure, expire_time,
3581 discard, None, None, rest)
3582 self._downloader.cookiejar.set_cookie(cookie)
3584 def _get_cookies(self, url):
3585 """ Return a compat_cookies_SimpleCookie with the cookies for the url """
3586 req = sanitized_Request(url)
3587 self._downloader.cookiejar.add_cookie_header(req)
3588 return compat_cookies_SimpleCookie(req.get_header('Cookie
'))
3590 def _apply_first_set_cookie_header(self, url_handle, cookie):
3592 Apply first Set-Cookie header instead of the last. Experimental.
3594 Some sites (e.g. [1-3]) may serve two cookies under the same name
3595 in Set-Cookie header and expect the first (old) one to be set rather
3596 than second (new). However, as of RFC6265 the newer one cookie
3597 should be set into cookie store what actually happens.
3598 We will workaround this issue by resetting the cookie to
3599 the first one manually.
3600 1. https://new.vk.com/
3601 2. https://github.com/ytdl-org/youtube-dl/issues/9841#issuecomment-227871201
3602 3. https://learning.oreilly.com/
3604 for header, cookies in url_handle.headers.items():
3605 if header.lower() != 'set-cookie
':
3607 if sys.version_info[0] >= 3:
3608 cookies = cookies.encode('iso
-8859-1')
3609 cookies = cookies.decode('utf
-8')
3610 cookie_value = re.search(
3611 r'%s=(.+?
);.*?
\b[Dd
]omain
=(.+?
)(?
:[,;]|$
)' % cookie, cookies)
3613 value, domain = cookie_value.groups()
3614 self._set_cookie(domain, cookie, value)
3617 def get_testcases(self, include_onlymatching=False):
3618 t = getattr(self, '_TEST
', None)
3620 assert not hasattr(self, '_TESTS
'), \
3621 '%s has _TEST
and _TESTS
' % type(self).__name__
3624 tests = getattr(self, '_TESTS
', [])
3626 if not include_onlymatching and t.get('only_matching
', False):
3628 t['name
'] = type(self).__name__[:-len('IE
')]
3631 def is_suitable(self, age_limit):
3632 """ Test whether the extractor is generally suitable for the given
3633 age limit (i.e. pornographic sites are not, all others usually are) """
3635 any_restricted = False
3636 for tc in self.get_testcases(include_onlymatching=False):
3637 if tc.get('playlist
', []):
3638 tc = tc['playlist
'][0]
3639 is_restricted = age_restricted(
3640 tc.get('info_dict
', {}).get('age_limit
'), age_limit)
3641 if not is_restricted:
3643 any_restricted = any_restricted or is_restricted
3644 return not any_restricted
3646 def extract_subtitles(self, *args, **kwargs):
3647 if (self.get_param('writesubtitles
', False)
3648 or self.get_param('listsubtitles
')):
3649 return self._get_subtitles(*args, **kwargs)
3652 def _get_subtitles(self, *args, **kwargs):
3653 raise NotImplementedError('This method must be implemented by subclasses
')
3655 def extract_comments(self, *args, **kwargs):
3656 if not self.get_param('getcomments
'):
3658 generator = self._get_comments(*args, **kwargs)
3665 comments.append(next(generator))
3666 except StopIteration:
3668 except KeyboardInterrupt:
3669 self.to_screen('Interrupted by user
')
3670 except Exception as e:
3671 if self.get_param('ignoreerrors
') is not True:
3673 self._downloader.report_error(e)
3674 comment_count = len(comments)
3675 self.to_screen(f'Extracted {comment_count} comments
')
3677 'comments
': comments,
3678 'comment_count
': None if interrupted else comment_count
3682 def _get_comments(self, *args, **kwargs):
3683 raise NotImplementedError('This method must be implemented by subclasses
')
3686 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
3687 """ Merge subtitle items for one language. Items with duplicated URLs/data
3688 will be dropped. """
3689 list1_data = set((item.get('url
'), item.get('data
')) for item in subtitle_list1)
3690 ret = list(subtitle_list1)
3691 ret.extend(item for item in subtitle_list2 if (item.get('url
'), item.get('data
')) not in list1_data)
3695 def _merge_subtitles(cls, *dicts, target=None):
3696 """ Merge subtitle dictionaries, language by language. """
3700 for lang, subs in d.items():
3701 target[lang] = cls._merge_subtitle_items(target.get(lang, []), subs)
3704 def extract_automatic_captions(self, *args, **kwargs):
3705 if (self.get_param('writeautomaticsub
', False)
3706 or self.get_param('listsubtitles
')):
3707 return self._get_automatic_captions(*args, **kwargs)
3710 def _get_automatic_captions(self, *args, **kwargs):
3711 raise NotImplementedError('This method must be implemented by subclasses
')
3713 def mark_watched(self, *args, **kwargs):
3714 if not self.get_param('mark_watched
', False):
3716 if (self.supports_login() and self._get_login_info()[0] is not None
3717 or self.get_param('cookiefile
') or self.get_param('cookiesfrombrowser
')):
3718 self._mark_watched(*args, **kwargs)
3720 def _mark_watched(self, *args, **kwargs):
3721 raise NotImplementedError('This method must be implemented by subclasses
')
3723 def geo_verification_headers(self):
3725 geo_verification_proxy = self.get_param('geo_verification_proxy
')
3726 if geo_verification_proxy:
3727 headers['Ytdl
-request
-proxy
'] = geo_verification_proxy
3730 def _generic_id(self, url):
3731 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
3733 def _generic_title(self, url):
3734 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
3737 def _availability(is_private=None, needs_premium=None, needs_subscription=None, needs_auth=None, is_unlisted=None):
3738 all_known = all(map(
3739 lambda x: x is not None,
3740 (is_private, needs_premium, needs_subscription, needs_auth, is_unlisted)))
3742 'private
' if is_private
3743 else 'premium_only
' if needs_premium
3744 else 'subscriber_only
' if needs_subscription
3745 else 'needs_auth
' if needs_auth
3746 else 'unlisted
' if is_unlisted
3747 else 'public
' if all_known
3750 def _configuration_arg(self, key, default=NO_DEFAULT, *, ie_key=None, casesense=False):
3752 @returns A list of values for the extractor argument given by "key"
3753 or "default" if no such key is present
3754 @param default The default value to return when the key is not present (default: [])
3755 @param casesense When false, the values are converted to lower case
3758 self._downloader.params, ('extractor_args
', (ie_key or self.ie_key()).lower(), key))
3760 return [] if default is NO_DEFAULT else default
3761 return list(val) if casesense else [x.lower() for x in val]
3763 def _yes_playlist(self, playlist_id, video_id, smuggled_data=None, *, playlist_label='playlist
', video_label='video
'):
3764 if not playlist_id or not video_id:
3767 no_playlist = (smuggled_data or {}).get('force_noplaylist
')
3768 if no_playlist is not None:
3769 return not no_playlist
3771 video_id = '' if video_id is True else f' {video_id}
'
3772 playlist_id = '' if playlist_id is True else f' {playlist_id}
'
3773 if self.get_param('noplaylist
'):
3774 self.to_screen(f'Downloading just the {video_label}{video_id} because of
--no
-playlist
')
3776 self.to_screen(f'Downloading {playlist_label}{playlist_id}
- add
--no
-playlist to download just the {video_label}{video_id}
')
3780 class SearchInfoExtractor(InfoExtractor):
3782 Base class for paged search queries extractors.
3783 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
3784 Instances should define _SEARCH_KEY and optionally _MAX_RESULTS
3787 _MAX_RESULTS = float('inf
')
3790 def _make_valid_url(cls):
3791 return r'%s(?P
<prefix
>|
[1-9][0-9]*|all
):(?P
<query
>[\s\S
]+)' % cls._SEARCH_KEY
3793 def _real_extract(self, query):
3794 prefix, query = self._match_valid_url(query).group('prefix
', 'query
')
3796 return self._get_n_results(query, 1)
3797 elif prefix == 'all
':
3798 return self._get_n_results(query, self._MAX_RESULTS)
3802 raise ExtractorError('invalid download number
%s for query
"%s"' % (n, query))
3803 elif n > self._MAX_RESULTS:
3804 self.report_warning('%s returns
max %i results (you requested
%i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3805 n = self._MAX_RESULTS
3806 return self._get_n_results(query, n)
3808 def _get_n_results(self, query, n):
3809 """Get a specified number of results for a query.
3810 Either this function or _search_results must be overridden by subclasses """
3811 return self.playlist_result(
3812 itertools.islice(self._search_results(query), 0, None if n == float('inf
') else n),
3815 def _search_results(self, query):
3816 """Returns an iterator of search results"""
3817 raise NotImplementedError('This method must be implemented by subclasses
')
3820 def SEARCH_KEY(self):
3821 return self._SEARCH_KEY