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
,
67 parse_m3u8_attributes
,
91 class InfoExtractor(object):
92 """Information Extractor class.
94 Information extractors are the classes that, given a URL, extract
95 information about the video (or videos) the URL refers to. This
96 information includes the real video URL, the video title, author and
97 others. The information is stored in a dictionary which is then
98 passed to the YoutubeDL. The YoutubeDL processes this
99 information possibly downloading the video to the file system, among
100 other possible outcomes.
102 The type field determines the type of the result.
103 By far the most common value (and the default if _type is missing) is
104 "video", which indicates a single video.
106 For a video, the dictionaries must include the following fields:
108 id: Video identifier.
109 title: Video title, unescaped.
111 Additionally, it must contain either a formats entry or a url one:
113 formats: A list of dictionaries for each format available, ordered
114 from worst to best quality.
117 * url The mandatory URL representing the media:
118 for plain file media - HTTP URL of this file,
120 for HLS - URL of the M3U8 media playlist,
121 for HDS - URL of the F4M manifest,
123 - HTTP URL to plain file media (in case of
125 - URL of the MPD manifest or base URL
126 representing the media if MPD manifest
127 is parsed from a string (in case of
129 for MSS - URL of the ISM manifest.
131 The URL of the manifest file in case of
133 for HLS - URL of the M3U8 master playlist,
134 for HDS - URL of the F4M manifest,
135 for DASH - URL of the MPD manifest,
136 for MSS - URL of the ISM manifest.
137 * ext Will be calculated from URL if missing
138 * format A human-readable description of the format
139 ("mp4 container with h264/opus").
140 Calculated from the format_id, width, height.
141 and format_note fields if missing.
142 * format_id A short description of the format
143 ("mp4_h264_opus" or "19").
144 Technically optional, but strongly recommended.
145 * format_note Additional info about the format
146 ("3D" or "DASH video")
147 * width Width of the video, if known
148 * height Height of the video, if known
149 * resolution Textual description of width and height
150 * dynamic_range The dynamic range of the video. One of:
151 "SDR" (None), "HDR10", "HDR10+, "HDR12", "HLG, "DV"
152 * tbr Average bitrate of audio and video in KBit/s
153 * abr Average audio bitrate in KBit/s
154 * acodec Name of the audio codec in use
155 * asr Audio sampling rate in Hertz
156 * vbr Average video bitrate in KBit/s
158 * vcodec Name of the video codec in use
159 * container Name of the container format
160 * filesize The number of bytes, if known in advance
161 * filesize_approx An estimate for the number of bytes
162 * player_url SWF Player URL (used for rtmpdump).
163 * protocol The protocol that will be used for the actual
164 download, lower-case.
165 "http", "https", "rtsp", "rtmp", "rtmp_ffmpeg", "rtmpe",
166 "m3u8", "m3u8_native" or "http_dash_segments".
168 Base URL for fragments. Each fragment's path
169 value (if present) will be relative to
171 * fragments A list of fragments of a fragmented media.
172 Each fragment entry must contain either an url
173 or a path. If an url is present it should be
174 considered by a client. Otherwise both path and
175 fragment_base_url must be present. Here is
176 the list of all potential fields:
177 * "url" - fragment's URL
178 * "path" - fragment's path relative to
180 * "duration" (optional, int or float)
181 * "filesize" (optional, int)
182 * preference Order number of this format. If this field is
183 present and not None, the formats get sorted
184 by this field, regardless of all other values.
185 -1 for default (order by other properties),
186 -2 or smaller for less than default.
187 < -1000 to hide the format (if there is
188 another one which is strictly better)
189 * language Language code, e.g. "de" or "en-US".
190 * language_preference Is this in the language mentioned in
192 10 if it's what the URL is about,
193 -1 for default (don't know),
194 -10 otherwise, other values reserved for now.
195 * quality Order number of the video quality of this
196 format, irrespective of the file format.
197 -1 for default (order by other properties),
198 -2 or smaller for less than default.
199 * source_preference Order number for this video source
200 (quality takes higher priority)
201 -1 for default (order by other properties),
202 -2 or smaller for less than default.
203 * http_headers A dictionary of additional HTTP headers
204 to add to the request.
205 * stretched_ratio If given and not 1, indicates that the
206 video's pixels are not square.
207 width : height ratio as float.
208 * no_resume The server does not support resuming the
209 (HTTP or RTMP) download. Boolean.
210 * has_drm The format has DRM and cannot be downloaded. Boolean
211 * downloader_options A dictionary of downloader options as
212 described in FileDownloader
213 RTMP formats can also have the additional fields: page_url,
214 app, play_path, tc_url, flash_version, rtmp_live, rtmp_conn,
215 rtmp_protocol, rtmp_real_time
217 url: Final video URL.
218 ext: Video filename extension.
219 format: The video format, defaults to ext (used for --get-format)
220 player_url: SWF Player URL (used for rtmpdump).
222 The following fields are optional:
224 alt_title: A secondary title of the video.
225 display_id An alternative identifier for the video, not necessarily
226 unique, but available before title. Typically, id is
227 something like "4234987", title "Dancing naked mole rats",
228 and display_id "dancing-naked-mole-rats"
229 thumbnails: A list of dictionaries, with the following entries:
230 * "id" (optional, string) - Thumbnail format ID
232 * "preference" (optional, int) - quality of the image
233 * "width" (optional, int)
234 * "height" (optional, int)
235 * "resolution" (optional, string "{width}x{height}",
237 * "filesize" (optional, int)
238 thumbnail: Full URL to a video thumbnail image.
239 description: Full video description.
240 uploader: Full name of the video uploader.
241 license: License name the video is licensed under.
242 creator: The creator of the video.
243 release_timestamp: UNIX timestamp of the moment the video was released.
244 release_date: The date (YYYYMMDD) when the video was released.
245 timestamp: UNIX timestamp of the moment the video was uploaded
246 upload_date: Video upload date (YYYYMMDD).
247 If not explicitly set, calculated from timestamp.
248 uploader_id: Nickname or id of the video uploader.
249 uploader_url: Full URL to a personal webpage of the video uploader.
250 channel: Full name of the channel the video is uploaded on.
251 Note that channel fields may or may not repeat uploader
252 fields. This depends on a particular extractor.
253 channel_id: Id of the channel.
254 channel_url: Full URL to a channel webpage.
255 location: Physical location where the video was filmed.
256 subtitles: The available subtitles as a dictionary in the format
257 {tag: subformats}. "tag" is usually a language code, and
258 "subformats" is a list sorted from lower to higher
259 preference, each element is a dictionary with the "ext"
261 * "data": The subtitles file contents
262 * "url": A URL pointing to the subtitles file
263 It can optionally also have:
264 * "name": Name or description of the subtitles
265 "ext" will be calculated from URL if missing
266 automatic_captions: Like 'subtitles'; contains automatically generated
267 captions instead of normal subtitles
268 duration: Length of the video in seconds, as an integer or float.
269 view_count: How many users have watched the video on the platform.
270 like_count: Number of positive ratings of the video
271 dislike_count: Number of negative ratings of the video
272 repost_count: Number of reposts of the video
273 average_rating: Average rating give by users, the scale used depends on the webpage
274 comment_count: Number of comments on the video
275 comments: A list of comments, each with one or more of the following
276 properties (all but one of text or html optional):
277 * "author" - human-readable name of the comment author
278 * "author_id" - user ID of the comment author
279 * "author_thumbnail" - The thumbnail of the comment author
281 * "html" - Comment as HTML
282 * "text" - Plain text of the comment
283 * "timestamp" - UNIX timestamp of comment
284 * "parent" - ID of the comment this one is replying to.
285 Set to "root" to indicate that this is a
286 comment to the original video.
287 * "like_count" - Number of positive ratings of the comment
288 * "dislike_count" - Number of negative ratings of the comment
289 * "is_favorited" - Whether the comment is marked as
290 favorite by the video uploader
291 * "author_is_uploader" - Whether the comment is made by
293 age_limit: Age restriction for the video, as an integer (years)
294 webpage_url: The URL to the video webpage, if given to yt-dlp it
295 should allow to get the same result again. (It will be set
296 by YoutubeDL if it's missing)
297 categories: A list of categories that the video falls in, for example
299 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
300 cast: A list of the video cast
301 is_live: True, False, or None (=unknown). Whether this video is a
302 live stream that goes on instead of a fixed-length video.
303 was_live: True, False, or None (=unknown). Whether this video was
304 originally a live stream.
305 live_status: 'is_live', 'is_upcoming', 'was_live', 'not_live' or None (=unknown)
306 If absent, automatically set from is_live, was_live
307 start_time: Time in seconds where the reproduction should start, as
308 specified in the URL.
309 end_time: Time in seconds where the reproduction should end, as
310 specified in the URL.
311 chapters: A list of dictionaries, with the following entries:
312 * "start_time" - The start time of the chapter in seconds
313 * "end_time" - The end time of the chapter in seconds
314 * "title" (optional, string)
315 playable_in_embed: Whether this video is allowed to play in embedded
316 players on other sites. Can be True (=always allowed),
317 False (=never allowed), None (=unknown), or a string
318 specifying the criteria for embedability (Eg: 'whitelist')
319 availability: Under what condition the video is available. One of
320 'private', 'premium_only', 'subscriber_only', 'needs_auth',
321 'unlisted' or 'public'. Use 'InfoExtractor._availability'
323 __post_extractor: A function to be called just before the metadata is
324 written to either disk, logger or console. The function
325 must return a dict which will be added to the info_dict.
326 This is usefull for additional information that is
327 time-consuming to extract. Note that the fields thus
328 extracted will not be available to output template and
329 match_filter. So, only "comments" and "comment_count" are
330 currently allowed to be extracted via this method.
332 The following fields should only be used when the video belongs to some logical
335 chapter: Name or title of the chapter the video belongs to.
336 chapter_number: Number of the chapter the video belongs to, as an integer.
337 chapter_id: Id of the chapter the video belongs to, as a unicode string.
339 The following fields should only be used when the video is an episode of some
340 series, programme or podcast:
342 series: Title of the series or programme the video episode belongs to.
343 season: Title of the season the video episode belongs to.
344 season_number: Number of the season the video episode belongs to, as an integer.
345 season_id: Id of the season the video episode belongs to, as a unicode string.
346 episode: Title of the video episode. Unlike mandatory video title field,
347 this field should denote the exact title of the video episode
348 without any kind of decoration.
349 episode_number: Number of the video episode within a season, as an integer.
350 episode_id: Id of the video episode, as a unicode string.
352 The following fields should only be used when the media is a track or a part of
355 track: Title of the track.
356 track_number: Number of the track within an album or a disc, as an integer.
357 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
359 artist: Artist(s) of the track.
360 genre: Genre(s) of the track.
361 album: Title of the album the track belongs to.
362 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
363 album_artist: List of all artists appeared on the album (e.g.
364 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
366 disc_number: Number of the disc or other physical medium the track belongs to,
368 release_year: Year (YYYY) when the album was released.
370 Unless mentioned otherwise, the fields should be Unicode strings.
372 Unless mentioned otherwise, None is equivalent to absence of information.
375 _type "playlist" indicates multiple videos.
376 There must be a key "entries", which is a list, an iterable, or a PagedList
377 object, each element of which is a valid dictionary by this specification.
379 Additionally, playlists can have "id", "title", and any other relevent
380 attributes with the same semantics as videos (see above).
383 _type "multi_video" indicates that there are multiple videos that
384 form a single show, for examples multiple acts of an opera or TV episode.
385 It must have an entries key like a playlist and contain all the keys
386 required for a video at the same time.
389 _type "url" indicates that the video must be extracted from another
390 location, possibly by a different extractor. Its only required key is:
391 "url" - the next URL to extract.
392 The key "ie_key" can be set to the class name (minus the trailing "IE",
393 e.g. "Youtube") if the extractor class is known in advance.
394 Additionally, the dictionary may have any properties of the resolved entity
395 known in advance, for example "title" if the title of the referred video is
399 _type "url_transparent" entities have the same specification as "url", but
400 indicate that the given additional information is more precise than the one
401 associated with the resolved URL.
402 This is useful when a site employs a video service that hosts the video and
403 its technical metadata, but that video service does not embed a useful
404 title, description etc.
407 Subclasses of this one should re-define the _real_initialize() and
408 _real_extract() methods and define a _VALID_URL regexp.
409 Probably, they should also be added to the list of extractors.
411 Subclasses may also override suitable() if necessary, but ensure the function
412 signature is preserved and that this function imports everything it needs
413 (except other extractors), so that lazy_extractors works correctly
415 _GEO_BYPASS attribute may be set to False in order to disable
416 geo restriction bypass mechanisms for a particular extractor.
417 Though it won't disable explicit geo restriction bypass based on
418 country code provided with geo_bypass_country.
420 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
421 countries for this extractor. One of these countries will be used by
422 geo restriction bypass mechanism right away in order to bypass
423 geo restriction, of course, if the mechanism is not disabled.
425 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
426 IP blocks in CIDR notation for this extractor. One of these IP blocks
427 will be used by geo restriction bypass mechanism similarly
430 The _WORKING attribute should be set to False for broken IEs
431 in order to warn the users and skip the tests.
436 _x_forwarded_for_ip
= None
438 _GEO_COUNTRIES
= None
439 _GEO_IP_BLOCKS
= None
443 'any': 'Use --cookies, --username and --password or --netrc to provide account credentials',
445 'Use --cookies-from-browser or --cookies for the authentication. '
446 'See https://github.com/ytdl-org/youtube-dl#how-do-i-pass-cookies-to-youtube-dl for how to manually pass cookies'),
447 'password': 'Use --username and --password or --netrc to provide account credentials',
450 def __init__(self
, downloader
=None):
451 """Constructor. Receives an optional downloader (a YoutubeDL instance).
452 If a downloader is not passed during initialization,
453 it must be set using "set_downloader()" before "extract()" is called"""
455 self
._x
_forwarded
_for
_ip
= None
456 self
._printed
_messages
= set()
457 self
.set_downloader(downloader
)
460 def _match_valid_url(cls
, url
):
461 # This does not use has/getattr intentionally - we want to know whether
462 # we have cached the regexp for *this* class, whereas getattr would also
463 # match the superclass
464 if '_VALID_URL_RE' not in cls
.__dict
__:
465 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
466 return cls
._VALID
_URL
_RE
.match(url
)
469 def suitable(cls
, url
):
470 """Receives a URL and returns True if suitable for this IE."""
471 # This function must import everything it needs (except other extractors),
472 # so that lazy_extractors works correctly
473 return cls
._match
_valid
_url
(url
) is not None
476 def _match_id(cls
, url
):
477 return cls
._match
_valid
_url
(url
).group('id')
480 def get_temp_id(cls
, url
):
482 return cls
._match
_id
(url
)
483 except (IndexError, AttributeError):
488 """Getter method for _WORKING."""
491 def initialize(self
):
492 """Initializes an instance (authentication, etc)."""
493 self
._printed
_messages
= set()
494 self
._initialize
_geo
_bypass
({
495 'countries': self
._GEO
_COUNTRIES
,
496 'ip_blocks': self
._GEO
_IP
_BLOCKS
,
499 self
._real
_initialize
()
502 def _initialize_geo_bypass(self
, geo_bypass_context
):
504 Initialize geo restriction bypass mechanism.
506 This method is used to initialize geo bypass mechanism based on faking
507 X-Forwarded-For HTTP header. A random country from provided country list
508 is selected and a random IP belonging to this country is generated. This
509 IP will be passed as X-Forwarded-For HTTP header in all subsequent
512 This method will be used for initial geo bypass mechanism initialization
513 during the instance initialization with _GEO_COUNTRIES and
516 You may also manually call it from extractor's code if geo bypass
517 information is not available beforehand (e.g. obtained during
518 extraction) or due to some other reason. In this case you should pass
519 this information in geo bypass context passed as first argument. It may
520 contain following fields:
522 countries: List of geo unrestricted countries (similar
524 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
525 (similar to _GEO_IP_BLOCKS)
528 if not self
._x
_forwarded
_for
_ip
:
530 # Geo bypass mechanism is explicitly disabled by user
531 if not self
.get_param('geo_bypass', True):
534 if not geo_bypass_context
:
535 geo_bypass_context
= {}
537 # Backward compatibility: previously _initialize_geo_bypass
538 # expected a list of countries, some 3rd party code may still use
540 if isinstance(geo_bypass_context
, (list, tuple)):
541 geo_bypass_context
= {
542 'countries': geo_bypass_context
,
545 # The whole point of geo bypass mechanism is to fake IP
546 # as X-Forwarded-For HTTP header based on some IP block or
549 # Path 1: bypassing based on IP block in CIDR notation
551 # Explicit IP block specified by user, use it right away
552 # regardless of whether extractor is geo bypassable or not
553 ip_block
= self
.get_param('geo_bypass_ip_block', None)
555 # Otherwise use random IP block from geo bypass context but only
556 # if extractor is known as geo bypassable
558 ip_blocks
= geo_bypass_context
.get('ip_blocks')
559 if self
._GEO
_BYPASS
and ip_blocks
:
560 ip_block
= random
.choice(ip_blocks
)
563 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(ip_block
)
564 self
._downloader
.write_debug(
565 '[debug] Using fake IP %s as X-Forwarded-For' % self
._x
_forwarded
_for
_ip
)
568 # Path 2: bypassing based on country code
570 # Explicit country code specified by user, use it right away
571 # regardless of whether extractor is geo bypassable or not
572 country
= self
.get_param('geo_bypass_country', None)
574 # Otherwise use random country code from geo bypass context but
575 # only if extractor is known as geo bypassable
577 countries
= geo_bypass_context
.get('countries')
578 if self
._GEO
_BYPASS
and countries
:
579 country
= random
.choice(countries
)
582 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country
)
583 self
._downloader
.write_debug(
584 'Using fake IP %s (%s) as X-Forwarded-For' % (self
._x
_forwarded
_for
_ip
, country
.upper()))
586 def extract(self
, url
):
587 """Extracts URL information and returns it in list of dicts."""
592 self
.write_debug('Extracting URL: %s' % url
)
593 ie_result
= self
._real
_extract
(url
)
594 if ie_result
is None:
596 if self
._x
_forwarded
_for
_ip
:
597 ie_result
['__x_forwarded_for_ip'] = self
._x
_forwarded
_for
_ip
598 subtitles
= ie_result
.get('subtitles')
599 if (subtitles
and 'live_chat' in subtitles
600 and 'no-live-chat' in self
.get_param('compat_opts', [])):
601 del subtitles
['live_chat']
603 except GeoRestrictedError
as e
:
604 if self
.__maybe
_fake
_ip
_and
_retry
(e
.countries
):
607 except ExtractorError
as e
:
608 video_id
= e
.video_id
or self
.get_temp_id(url
)
609 raise ExtractorError(
610 e
.msg
, video_id
=video_id
, ie
=self
.IE_NAME
, tb
=e
.traceback
, expected
=e
.expected
, cause
=e
.cause
)
611 except compat_http_client
.IncompleteRead
as e
:
612 raise ExtractorError('A network error has occurred.', cause
=e
, expected
=True, video_id
=self
.get_temp_id(url
))
613 except (KeyError, StopIteration) as e
:
614 raise ExtractorError('An extractor error has occurred.', cause
=e
, video_id
=self
.get_temp_id(url
))
616 def __maybe_fake_ip_and_retry(self
, countries
):
617 if (not self
.get_param('geo_bypass_country', None)
619 and self
.get_param('geo_bypass', True)
620 and not self
._x
_forwarded
_for
_ip
622 country_code
= random
.choice(countries
)
623 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country_code
)
624 if self
._x
_forwarded
_for
_ip
:
626 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
627 % (self
._x
_forwarded
_for
_ip
, country_code
.upper()))
631 def set_downloader(self
, downloader
):
632 """Sets the downloader for this IE."""
633 self
._downloader
= downloader
635 def _real_initialize(self
):
636 """Real initialization process. Redefine in subclasses."""
639 def _real_extract(self
, url
):
640 """Real extraction process. Redefine in subclasses."""
645 """A string for getting the InfoExtractor with get_info_extractor"""
646 return cls
.__name
__[:-2]
650 return compat_str(type(self
).__name
__[:-2])
653 def __can_accept_status_code(err
, expected_status
):
654 assert isinstance(err
, compat_urllib_error
.HTTPError
)
655 if expected_status
is None:
657 elif callable(expected_status
):
658 return expected_status(err
.code
) is True
660 return err
.code
in variadic(expected_status
)
662 def _request_webpage(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
, expected_status
=None):
664 Return the response handle.
666 See _download_webpage docstring for arguments specification.
668 if not self
._downloader
._first
_webpage
_request
:
669 sleep_interval
= self
.get_param('sleep_interval_requests') or 0
670 if sleep_interval
> 0:
671 self
.to_screen('Sleeping %s seconds ...' % sleep_interval
)
672 time
.sleep(sleep_interval
)
674 self
._downloader
._first
_webpage
_request
= False
677 self
.report_download_webpage(video_id
)
678 elif note
is not False:
680 self
.to_screen('%s' % (note
,))
682 self
.to_screen('%s: %s' % (video_id
, note
))
684 # Some sites check X-Forwarded-For HTTP header in order to figure out
685 # the origin of the client behind proxy. This allows bypassing geo
686 # restriction by faking this header's value to IP that belongs to some
687 # geo unrestricted country. We will do so once we encounter any
688 # geo restriction error.
689 if self
._x
_forwarded
_for
_ip
:
690 if 'X-Forwarded-For' not in headers
:
691 headers
['X-Forwarded-For'] = self
._x
_forwarded
_for
_ip
693 if isinstance(url_or_request
, compat_urllib_request
.Request
):
694 url_or_request
= update_Request(
695 url_or_request
, data
=data
, headers
=headers
, query
=query
)
698 url_or_request
= update_url_query(url_or_request
, query
)
699 if data
is not None or headers
:
700 url_or_request
= sanitized_Request(url_or_request
, data
, headers
)
702 return self
._downloader
.urlopen(url_or_request
)
703 except network_exceptions
as err
:
704 if isinstance(err
, compat_urllib_error
.HTTPError
):
705 if self
.__can
_accept
_status
_code
(err
, expected_status
):
706 # Retain reference to error to prevent file object from
707 # being closed before it can be read. Works around the
708 # effects of <https://bugs.python.org/issue15002>
709 # introduced in Python 3.4.1.
716 errnote
= 'Unable to download webpage'
718 errmsg
= '%s: %s' % (errnote
, error_to_compat_str(err
))
720 raise ExtractorError(errmsg
, sys
.exc_info()[2], cause
=err
)
722 self
.report_warning(errmsg
)
725 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):
727 Return a tuple (page content as string, URL handle).
729 See _download_webpage docstring for arguments specification.
731 # Strip hashes from the URL (#1038)
732 if isinstance(url_or_request
, (compat_str
, str)):
733 url_or_request
= url_or_request
.partition('#')[0]
735 urlh
= self
._request
_webpage
(url_or_request
, video_id
, note
, errnote
, fatal
, data
=data
, headers
=headers
, query
=query
, expected_status
=expected_status
)
739 content
= self
._webpage
_read
_content
(urlh
, url_or_request
, video_id
, note
, errnote
, fatal
, encoding
=encoding
)
740 return (content
, urlh
)
743 def _guess_encoding_from_content(content_type
, webpage_bytes
):
744 m
= re
.match(r
'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type
)
746 encoding
= m
.group(1)
748 m
= re
.search(br
'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
749 webpage_bytes[:1024])
751 encoding = m.group(1).decode('ascii')
752 elif webpage_bytes.startswith(b'\xff\xfe'):
759 def __check_blocked(self, content):
760 first_block = content[:512]
761 if ('<title>Access to this site is blocked</title>' in content
762 and 'Websense' in first_block):
763 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
764 blocked_iframe = self._html_search_regex(
765 r'<iframe src="([^
"]+)"', content,
766 'Websense information URL
', default=None)
768 msg += ' Visit
%s for more details
' % blocked_iframe
769 raise ExtractorError(msg, expected=True)
770 if '<title
>The URL you requested has been blocked
</title
>' in first_block:
772 'Access to this webpage has been blocked by Indian censorship
. '
773 'Use a VPN
or proxy
server (with --proxy
) to route around it
.')
774 block_msg = self._html_search_regex(
775 r'</h1
><p
>(.*?
)</p
>',
776 content, 'block message
', default=None)
778 msg += ' (Message
: "%s")' % block_msg.replace('\n', ' ')
779 raise ExtractorError(msg, expected=True)
780 if ('<title
>TTK
:: Доступ к ресурсу ограничен
</title
>' in content
781 and 'blocklist
.rkn
.gov
.ru
' in content):
782 raise ExtractorError(
783 'Access to this webpage has been blocked by decision of the Russian government
. '
784 'Visit http
://blocklist
.rkn
.gov
.ru
/ for a block reason
.',
787 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
788 content_type = urlh.headers.get('Content
-Type
', '')
789 webpage_bytes = urlh.read()
790 if prefix is not None:
791 webpage_bytes = prefix + webpage_bytes
793 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
794 if self.get_param('dump_intermediate_pages
', False):
795 self.to_screen('Dumping request to
' + urlh.geturl())
796 dump = base64.b64encode(webpage_bytes).decode('ascii
')
797 self._downloader.to_screen(dump)
798 if self.get_param('write_pages
', False):
799 basen = '%s_%s' % (video_id, urlh.geturl())
800 trim_length = self.get_param('trim_file_name
') or 240
801 if len(basen) > trim_length:
802 h = '___
' + hashlib.md5(basen.encode('utf
-8')).hexdigest()
803 basen = basen[:trim_length - len(h)] + h
804 raw_filename = basen + '.dump
'
805 filename = sanitize_filename(raw_filename, restricted=True)
806 self.to_screen('Saving request to
' + filename)
807 # Working around MAX_PATH limitation on Windows (see
808 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
809 if compat_os_name == 'nt
':
810 absfilepath = os.path.abspath(filename)
811 if len(absfilepath) > 259:
812 filename = '\\\\?
\\' + absfilepath
813 with open(filename, 'wb
') as outf:
814 outf.write(webpage_bytes)
817 content = webpage_bytes.decode(encoding, 'replace
')
819 content = webpage_bytes.decode('utf
-8', 'replace
')
821 self.__check_blocked(content)
825 def _download_webpage(
826 self, url_or_request, video_id, note=None, errnote=None,
827 fatal=True, tries=1, timeout=5, encoding=None, data=None,
828 headers={}, query={}, expected_status=None):
830 Return the data of the page as a string.
833 url_or_request -- plain text URL as a string or
834 a compat_urllib_request.Requestobject
835 video_id -- Video/playlist/item identifier (string)
838 note -- note printed before downloading (string)
839 errnote -- note printed in case of an error (string)
840 fatal -- flag denoting whether error should be considered fatal,
841 i.e. whether it should cause ExtractionError to be raised,
842 otherwise a warning will be reported and extraction continued
843 tries -- number of tries
844 timeout -- sleep interval between tries
845 encoding -- encoding for a page content decoding, guessed automatically
846 when not explicitly specified
847 data -- POST data (bytes)
848 headers -- HTTP headers (dict)
849 query -- URL query (dict)
850 expected_status -- allows to accept failed HTTP requests (non 2xx
851 status code) by explicitly specifying a set of accepted status
852 codes. Can be any of the following entities:
853 - an integer type specifying an exact failed status code to
855 - a list or a tuple of integer types specifying a list of
856 failed status codes to accept
857 - a callable accepting an actual failed status code and
858 returning True if it should be accepted
859 Note that this argument does not affect success status codes (2xx)
860 which are always accepted.
865 while success is False:
867 res = self._download_webpage_handle(
868 url_or_request, video_id, note, errnote, fatal,
869 encoding=encoding, data=data, headers=headers, query=query,
870 expected_status=expected_status)
872 except compat_http_client.IncompleteRead as e:
874 if try_count >= tries:
876 self._sleep(timeout, video_id)
883 def _download_xml_handle(
884 self, url_or_request, video_id, note='Downloading XML
',
885 errnote='Unable to download XML
', transform_source=None,
886 fatal=True, encoding=None, data=None, headers={}, query={},
887 expected_status=None):
889 Return a tuple (xml as an compat_etree_Element, URL handle).
891 See _download_webpage docstring for arguments specification.
893 res = self._download_webpage_handle(
894 url_or_request, video_id, note, errnote, fatal=fatal,
895 encoding=encoding, data=data, headers=headers, query=query,
896 expected_status=expected_status)
899 xml_string, urlh = res
900 return self._parse_xml(
901 xml_string, video_id, transform_source=transform_source,
905 self, url_or_request, video_id,
906 note='Downloading XML
', errnote='Unable to download XML
',
907 transform_source=None, fatal=True, encoding=None,
908 data=None, headers={}, query={}, expected_status=None):
910 Return the xml as an compat_etree_Element.
912 See _download_webpage docstring for arguments specification.
914 res = self._download_xml_handle(
915 url_or_request, video_id, note=note, errnote=errnote,
916 transform_source=transform_source, fatal=fatal, encoding=encoding,
917 data=data, headers=headers, query=query,
918 expected_status=expected_status)
919 return res if res is False else res[0]
921 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
923 xml_string = transform_source(xml_string)
925 return compat_etree_fromstring(xml_string.encode('utf
-8'))
926 except compat_xml_parse_error as ve:
927 errmsg = '%s: Failed to parse XML
' % video_id
929 raise ExtractorError(errmsg, cause=ve)
931 self.report_warning(errmsg + str(ve))
933 def _download_json_handle(
934 self, url_or_request, video_id, note='Downloading JSON metadata
',
935 errnote='Unable to download JSON metadata
', transform_source=None,
936 fatal=True, encoding=None, data=None, headers={}, query={},
937 expected_status=None):
939 Return a tuple (JSON object, URL handle).
941 See _download_webpage docstring for arguments specification.
943 res = self._download_webpage_handle(
944 url_or_request, video_id, note, errnote, fatal=fatal,
945 encoding=encoding, data=data, headers=headers, query=query,
946 expected_status=expected_status)
949 json_string, urlh = res
950 return self._parse_json(
951 json_string, video_id, transform_source=transform_source,
955 self, url_or_request, video_id, note='Downloading JSON metadata
',
956 errnote='Unable to download JSON metadata
', transform_source=None,
957 fatal=True, encoding=None, data=None, headers={}, query={},
958 expected_status=None):
960 Return the JSON object as a dict.
962 See _download_webpage docstring for arguments specification.
964 res = self._download_json_handle(
965 url_or_request, video_id, note=note, errnote=errnote,
966 transform_source=transform_source, fatal=fatal, encoding=encoding,
967 data=data, headers=headers, query=query,
968 expected_status=expected_status)
969 return res if res is False else res[0]
971 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
973 json_string = transform_source(json_string)
975 return json.loads(json_string)
976 except ValueError as ve:
977 errmsg = '%s: Failed to parse JSON
' % video_id
979 raise ExtractorError(errmsg, cause=ve)
981 self.report_warning(errmsg + str(ve))
983 def _parse_socket_response_as_json(self, data, video_id, transform_source=None, fatal=True):
984 return self._parse_json(
985 data[data.find('{'):data.rfind('}
') + 1],
986 video_id, transform_source, fatal)
988 def _download_socket_json_handle(
989 self, url_or_request, video_id, note='Polling socket
',
990 errnote='Unable to poll socket
', transform_source=None,
991 fatal=True, encoding=None, data=None, headers={}, query={},
992 expected_status=None):
994 Return a tuple (JSON object, URL handle).
996 See _download_webpage docstring for arguments specification.
998 res = self._download_webpage_handle(
999 url_or_request, video_id, note, errnote, fatal=fatal,
1000 encoding=encoding, data=data, headers=headers, query=query,
1001 expected_status=expected_status)
1005 return self._parse_socket_response_as_json(
1006 webpage, video_id, transform_source=transform_source,
1009 def _download_socket_json(
1010 self, url_or_request, video_id, note='Polling socket
',
1011 errnote='Unable to poll socket
', transform_source=None,
1012 fatal=True, encoding=None, data=None, headers={}, query={},
1013 expected_status=None):
1015 Return the JSON object as a dict.
1017 See _download_webpage docstring for arguments specification.
1019 res = self._download_socket_json_handle(
1020 url_or_request, video_id, note=note, errnote=errnote,
1021 transform_source=transform_source, fatal=fatal, encoding=encoding,
1022 data=data, headers=headers, query=query,
1023 expected_status=expected_status)
1024 return res if res is False else res[0]
1026 def report_warning(self, msg, video_id=None, *args, only_once=False, **kwargs):
1027 idstr = format_field(video_id, template='%s: ')
1028 msg = f'[{self.IE_NAME}
] {idstr}{msg}
'
1030 if f'WARNING
: {msg}
' in self._printed_messages:
1032 self._printed_messages.add(f'WARNING
: {msg}
')
1033 self._downloader.report_warning(msg, *args, **kwargs)
1035 def to_screen(self, msg, *args, **kwargs):
1036 """Print msg to screen, prefixing it with '[ie_name
]'"""
1037 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg), *args, **kwargs)
1039 def write_debug(self, msg, *args, **kwargs):
1040 self._downloader.write_debug('[%s] %s' % (self.IE_NAME, msg), *args, **kwargs)
1042 def get_param(self, name, default=None, *args, **kwargs):
1043 if self._downloader:
1044 return self._downloader.params.get(name, default, *args, **kwargs)
1047 def report_drm(self, video_id, partial=False):
1048 self.raise_no_formats('This video
is DRM protected
', expected=True, video_id=video_id)
1050 def report_extraction(self, id_or_name):
1051 """Report information extraction."""
1052 self.to_screen('%s: Extracting information
' % id_or_name)
1054 def report_download_webpage(self, video_id):
1055 """Report webpage download."""
1056 self.to_screen('%s: Downloading webpage
' % video_id)
1058 def report_age_confirmation(self):
1059 """Report attempt to confirm age."""
1060 self.to_screen('Confirming age
')
1062 def report_login(self):
1063 """Report attempt to log in."""
1064 self.to_screen('Logging
in')
1066 def raise_login_required(
1067 self, msg='This video
is only available
for registered users
',
1068 metadata_available=False, method='any
'):
1069 if metadata_available and self.get_param('ignore_no_formats_error
'):
1070 self.report_warning(msg)
1071 if method is not None:
1072 msg = '%s. %s' % (msg, self._LOGIN_HINTS[method])
1073 raise ExtractorError(msg, expected=True)
1075 def raise_geo_restricted(
1076 self, msg='This video
is not available
from your location due to geo restriction
',
1077 countries=None, metadata_available=False):
1078 if metadata_available and self.get_param('ignore_no_formats_error
'):
1079 self.report_warning(msg)
1081 raise GeoRestrictedError(msg, countries=countries)
1083 def raise_no_formats(self, msg, expected=False, video_id=None):
1084 if expected and self.get_param('ignore_no_formats_error
'):
1085 self.report_warning(msg, video_id)
1086 elif isinstance(msg, ExtractorError):
1089 raise ExtractorError(msg, expected=expected, video_id=video_id)
1091 # Methods for following #608
1093 def url_result(url, ie=None, video_id=None, video_title=None, **kwargs):
1094 """Returns a URL that points to a page that should be processed"""
1095 # TODO: ie should be the class used for getting the info
1096 video_info = {'_type
': 'url
',
1099 video_info.update(kwargs)
1100 if video_id is not None:
1101 video_info['id'] = video_id
1102 if video_title is not None:
1103 video_info['title
'] = video_title
1106 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None):
1108 self.url_result(self._proto_relative_url(getter(m) if getter else m), ie)
1110 return self.playlist_result(
1111 urls, playlist_id=playlist_id, playlist_title=playlist_title)
1114 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None, **kwargs):
1115 """Returns a playlist"""
1116 video_info = {'_type
': 'playlist
',
1118 video_info.update(kwargs)
1120 video_info['id'] = playlist_id
1122 video_info['title
'] = playlist_title
1123 if playlist_description is not None:
1124 video_info['description
'] = playlist_description
1127 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1129 Perform a regex search on the given string, using a single or a list of
1130 patterns returning the first matching group.
1131 In case of failure return a default value or raise a WARNING or a
1132 RegexNotFoundError, depending on fatal, specifying the field name.
1134 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
1135 mobj = re.search(pattern, string, flags)
1138 mobj = re.search(p, string, flags)
1142 _name = self._downloader._format_err(name, self._downloader.Styles.EMPHASIS)
1146 # return the first matching group
1147 return next(g for g in mobj.groups() if g is not None)
1148 elif isinstance(group, (list, tuple)):
1149 return tuple(mobj.group(g) for g in group)
1151 return mobj.group(group)
1152 elif default is not NO_DEFAULT:
1155 raise RegexNotFoundError('Unable to extract
%s' % _name)
1157 self.report_warning('unable to extract
%s' % _name + bug_reports_message())
1160 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1162 Like _search_regex, but strips HTML tags and unescapes entities.
1164 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1166 return clean_html(res).strip()
1170 def _get_netrc_login_info(self, netrc_machine=None):
1173 netrc_machine = netrc_machine or self._NETRC_MACHINE
1175 if self.get_param('usenetrc
', False):
1177 netrc_file = compat_expanduser(self.get_param('netrc_location
') or '~
')
1178 if os.path.isdir(netrc_file):
1179 netrc_file = os.path.join(netrc_file, '.netrc
')
1180 info = netrc.netrc(file=netrc_file).authenticators(netrc_machine)
1181 if info is not None:
1185 raise netrc.NetrcParseError(
1186 'No authenticators
for %s' % netrc_machine)
1187 except (IOError, netrc.NetrcParseError) as err:
1188 self.report_warning(
1189 'parsing
.netrc
: %s' % error_to_compat_str(err))
1191 return username, password
1193 def _get_login_info(self, username_option='username
', password_option='password
', netrc_machine=None):
1195 Get the login info as (username, password)
1196 First look for the manually specified credentials using username_option
1197 and password_option as keys in params dictionary. If no such credentials
1198 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1200 If there's no info available
, return (None, None)
1203 # Attempt to use provided username and password or .netrc data
1204 username = self.get_param(username_option)
1205 if username is not None:
1206 password = self.get_param(password_option)
1208 username, password = self._get_netrc_login_info(netrc_machine)
1210 return username, password
1212 def _get_tfa_info(self, note='two-factor verification code'):
1214 Get the two
-factor authentication info
1215 TODO
- asking the user will be required
for sms
/phone verify
1216 currently just uses the command line option
1217 If there
's no info available, return None
1220 tfa = self.get_param('twofactor
')
1224 return compat_getpass('Type
%s and press
[Return
]: ' % note)
1226 # Helper functions for extracting OpenGraph info
1228 def _og_regexes(prop):
1229 content_re = r'content
=(?
:"([^"]+?
)"|\'([^\']+?)\'|\s*([^\s"\'=<>`
]+?
))'
1230 property_re = (r'(?
:name|
property)=(?
:\'og
[:-]%(prop)s\'|
"og[:-]%(prop)s"|\s
*og
[:-]%(prop)s\b)'
1231 % {'prop': re.escape(prop)})
1232 template = r'<meta
[^
>]+?
%s[^
>]+?
%s'
1234 template % (property_re, content_re),
1235 template % (content_re, property_re),
1239 def _meta_regex(prop):
1240 return r'''(?isx)<meta
1241 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1242 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1244 def _og_search_property(self, prop, html, name=None, **kargs):
1245 prop = variadic(prop)
1247 name = 'OpenGraph
%s' % prop[0]
1250 og_regexes.extend(self._og_regexes(p))
1251 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1254 return unescapeHTML(escaped)
1256 def _og_search_thumbnail(self, html, **kargs):
1257 return self._og_search_property('image
', html, 'thumbnail URL
', fatal=False, **kargs)
1259 def _og_search_description(self, html, **kargs):
1260 return self._og_search_property('description
', html, fatal=False, **kargs)
1262 def _og_search_title(self, html, **kargs):
1263 return self._og_search_property('title
', html, **kargs)
1265 def _og_search_video_url(self, html, name='video url
', secure=True, **kargs):
1266 regexes = self._og_regexes('video
') + self._og_regexes('video
:url
')
1268 regexes = self._og_regexes('video
:secure_url
') + regexes
1269 return self._html_search_regex(regexes, html, name, **kargs)
1271 def _og_search_url(self, html, **kargs):
1272 return self._og_search_property('url
', html, **kargs)
1274 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1275 name = variadic(name)
1276 if display_name is None:
1277 display_name = name[0]
1278 return self._html_search_regex(
1279 [self._meta_regex(n) for n in name],
1280 html, display_name, fatal=fatal, group='content
', **kwargs)
1282 def _dc_search_uploader(self, html):
1283 return self._html_search_meta('dc
.creator
', html, 'uploader
')
1285 def _rta_search(self, html):
1286 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1287 if re.search(r'(?ix
)<meta\s
+name
="rating"\s
+'
1288 r' content
="RTA-5042-1996-1400-1577-RTA"',
1293 def _media_rating_search(self, html):
1294 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1295 rating = self._html_search_meta('rating
', html)
1307 return RATING_TABLE.get(rating.lower())
1309 def _family_friendly_search(self, html):
1310 # See http://schema.org/VideoObject
1311 family_friendly = self._html_search_meta(
1312 'isFamilyFriendly
', html, default=None)
1314 if not family_friendly:
1323 return RATING_TABLE.get(family_friendly.lower())
1325 def _twitter_search_player(self, html):
1326 return self._html_search_meta('twitter
:player
', html,
1327 'twitter card player
')
1329 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
1330 json_ld_list = list(re.finditer(JSON_LD_RE, html))
1331 default = kwargs.get('default
', NO_DEFAULT)
1332 # JSON-LD may be malformed and thus `fatal` should be respected.
1333 # At the same time `default` may be passed that assumes `fatal=False`
1334 # for _search_regex. Let's simulate the same behavior here
as well
.
1335 fatal
= kwargs
.get('fatal', True) if default
is NO_DEFAULT
else False
1337 for mobj
in json_ld_list
:
1338 json_ld_item
= self
._parse
_json
(
1339 mobj
.group('json_ld'), video_id
, fatal
=fatal
)
1340 if not json_ld_item
:
1342 if isinstance(json_ld_item
, dict):
1343 json_ld
.append(json_ld_item
)
1344 elif isinstance(json_ld_item
, (list, tuple)):
1345 json_ld
.extend(json_ld_item
)
1347 json_ld
= self
._json
_ld
(json_ld
, video_id
, fatal
=fatal
, expected_type
=expected_type
)
1350 if default
is not NO_DEFAULT
:
1353 raise RegexNotFoundError('Unable to extract JSON-LD')
1355 self
.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1358 def _json_ld(self
, json_ld
, video_id
, fatal
=True, expected_type
=None):
1359 if isinstance(json_ld
, compat_str
):
1360 json_ld
= self
._parse
_json
(json_ld
, video_id
, fatal
=fatal
)
1364 if not isinstance(json_ld
, (list, tuple, dict)):
1366 if isinstance(json_ld
, dict):
1369 INTERACTION_TYPE_MAP
= {
1370 'CommentAction': 'comment',
1371 'AgreeAction': 'like',
1372 'DisagreeAction': 'dislike',
1373 'LikeAction': 'like',
1374 'DislikeAction': 'dislike',
1375 'ListenAction': 'view',
1376 'WatchAction': 'view',
1377 'ViewAction': 'view',
1380 def extract_interaction_type(e
):
1381 interaction_type
= e
.get('interactionType')
1382 if isinstance(interaction_type
, dict):
1383 interaction_type
= interaction_type
.get('@type')
1384 return str_or_none(interaction_type
)
1386 def extract_interaction_statistic(e
):
1387 interaction_statistic
= e
.get('interactionStatistic')
1388 if isinstance(interaction_statistic
, dict):
1389 interaction_statistic
= [interaction_statistic
]
1390 if not isinstance(interaction_statistic
, list):
1392 for is_e
in interaction_statistic
:
1393 if not isinstance(is_e
, dict):
1395 if is_e
.get('@type') != 'InteractionCounter':
1397 interaction_type
= extract_interaction_type(is_e
)
1398 if not interaction_type
:
1400 # For interaction count some sites provide string instead of
1401 # an integer (as per spec) with non digit characters (e.g. ",")
1402 # so extracting count with more relaxed str_to_int
1403 interaction_count
= str_to_int(is_e
.get('userInteractionCount'))
1404 if interaction_count
is None:
1406 count_kind
= INTERACTION_TYPE_MAP
.get(interaction_type
.split('/')[-1])
1409 count_key
= '%s_count' % count_kind
1410 if info
.get(count_key
) is not None:
1412 info
[count_key
] = interaction_count
1414 def extract_video_object(e
):
1415 assert e
['@type'] == 'VideoObject'
1416 author
= e
.get('author')
1418 'url': url_or_none(e
.get('contentUrl')),
1419 'title': unescapeHTML(e
.get('name')),
1420 'description': unescapeHTML(e
.get('description')),
1421 'thumbnail': url_or_none(e
.get('thumbnailUrl') or e
.get('thumbnailURL')),
1422 'duration': parse_duration(e
.get('duration')),
1423 'timestamp': unified_timestamp(e
.get('uploadDate')),
1424 # author can be an instance of 'Organization' or 'Person' types.
1425 # both types can have 'name' property(inherited from 'Thing' type). [1]
1426 # however some websites are using 'Text' type instead.
1427 # 1. https://schema.org/VideoObject
1428 'uploader': author
.get('name') if isinstance(author
, dict) else author
if isinstance(author
, compat_str
) else None,
1429 'filesize': float_or_none(e
.get('contentSize')),
1430 'tbr': int_or_none(e
.get('bitrate')),
1431 'width': int_or_none(e
.get('width')),
1432 'height': int_or_none(e
.get('height')),
1433 'view_count': int_or_none(e
.get('interactionCount')),
1435 extract_interaction_statistic(e
)
1439 item_type
= e
.get('@type')
1440 if expected_type
is not None and expected_type
!= item_type
:
1442 if item_type
in ('TVEpisode', 'Episode'):
1443 episode_name
= unescapeHTML(e
.get('name'))
1445 'episode': episode_name
,
1446 'episode_number': int_or_none(e
.get('episodeNumber')),
1447 'description': unescapeHTML(e
.get('description')),
1449 if not info
.get('title') and episode_name
:
1450 info
['title'] = episode_name
1451 part_of_season
= e
.get('partOfSeason')
1452 if isinstance(part_of_season
, dict) and part_of_season
.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1454 'season': unescapeHTML(part_of_season
.get('name')),
1455 'season_number': int_or_none(part_of_season
.get('seasonNumber')),
1457 part_of_series
= e
.get('partOfSeries') or e
.get('partOfTVSeries')
1458 if isinstance(part_of_series
, dict) and part_of_series
.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1459 info
['series'] = unescapeHTML(part_of_series
.get('name'))
1460 elif item_type
== 'Movie':
1462 'title': unescapeHTML(e
.get('name')),
1463 'description': unescapeHTML(e
.get('description')),
1464 'duration': parse_duration(e
.get('duration')),
1465 'timestamp': unified_timestamp(e
.get('dateCreated')),
1467 elif item_type
in ('Article', 'NewsArticle'):
1469 'timestamp': parse_iso8601(e
.get('datePublished')),
1470 'title': unescapeHTML(e
.get('headline')),
1471 'description': unescapeHTML(e
.get('articleBody')),
1473 elif item_type
== 'VideoObject':
1474 extract_video_object(e
)
1475 if expected_type
is None:
1479 video
= e
.get('video')
1480 if isinstance(video
, dict) and video
.get('@type') == 'VideoObject':
1481 extract_video_object(video
)
1482 if expected_type
is None:
1486 return dict((k
, v
) for k
, v
in info
.items() if v
is not None)
1489 def _hidden_inputs(html
):
1490 html
= re
.sub(r
'<!--(?:(?!<!--).)*-->', '', html
)
1492 for input in re
.findall(r
'(?i)(<input[^>]+>)', html
):
1493 attrs
= extract_attributes(input)
1496 if attrs
.get('type') not in ('hidden', 'submit'):
1498 name
= attrs
.get('name') or attrs
.get('id')
1499 value
= attrs
.get('value')
1500 if name
and value
is not None:
1501 hidden_inputs
[name
] = value
1502 return hidden_inputs
1504 def _form_hidden_inputs(self
, form_id
, html
):
1505 form
= self
._search
_regex
(
1506 r
'(?is)<form[^>]+?id=(["\'])%s\
1[^
>]*>(?P
<form
>.+?
)</form
>' % form_id,
1507 html, '%s form
' % form_id, group='form
')
1508 return self._hidden_inputs(form)
1511 regex = r' *((?P
<reverse
>\
+)?
(?P
<field
>[a
-zA
-Z0
-9_]+)((?P
<separator
>[~
:])(?P
<limit
>.*?
))?
)?
*$
'
1513 default = ('hidden
', 'aud_or_vid
', 'hasvid
', 'ie_pref
', 'lang
', 'quality
',
1514 'res
', 'fps
', 'hdr
:12', 'codec
:vp9
.2
', 'size
', 'br
', 'asr
',
1515 'proto
', 'ext
', 'hasaud
', 'source
', 'format_id
') # These must not be aliases
1516 ytdl_default = ('hasaud
', 'lang
', 'quality
', 'tbr
', 'filesize
', 'vbr
',
1517 'height
', 'width
', 'proto
', 'vext
', 'abr
', 'aext
',
1518 'fps
', 'fs_approx
', 'source
', 'format_id
')
1521 'vcodec
': {'type': 'ordered
', 'regex
': True,
1522 'order
': ['av0?
1', 'vp0?
9.2', 'vp0?
9', '[hx
]265|he?vc?
', '[hx
]264|avc
', 'vp0?
8', 'mp4v|h263
', 'theora
', '', None, 'none
']},
1523 'acodec
': {'type': 'ordered
', 'regex
': True,
1524 'order
': ['opus
', 'vorbis
', 'aac
', 'mp?
4a?
', 'mp3
', 'e?a?c
-?
3', 'dts
', '', None, 'none
']},
1525 'hdr
': {'type': 'ordered
', 'regex
': True, 'field
': 'dynamic_range
',
1526 'order
': ['dv
', '(hdr
)?
12', r'(hdr
)?
10\
+', '(hdr
)?
10', 'hlg
', '', 'sdr
', None]},
1527 'proto
': {'type': 'ordered
', 'regex
': True, 'field
': 'protocol
',
1528 'order
': ['(ht|f
)tps
', '(ht|f
)tp$
', 'm3u8
.+', '.*dash
', 'ws|websocket
', '', 'mms|rtsp
', 'none
', 'f4
']},
1529 'vext
': {'type': 'ordered
', 'field
': 'video_ext
',
1530 'order
': ('mp4
', 'webm
', 'flv
', '', 'none
'),
1531 'order_free
': ('webm
', 'mp4
', 'flv
', '', 'none
')},
1532 'aext
': {'type': 'ordered
', 'field
': 'audio_ext
',
1533 'order
': ('m4a
', 'aac
', 'mp3
', 'ogg
', 'opus
', 'webm
', '', 'none
'),
1534 'order_free
': ('opus
', 'ogg
', 'webm
', 'm4a
', 'mp3
', 'aac
', '', 'none
')},
1535 'hidden
': {'visible': False, 'forced': True, 'type': 'extractor', 'max': -1000},
1536 'aud_or_vid
': {'visible
': False, 'forced
': True, 'type': 'multiple
',
1537 'field
': ('vcodec
', 'acodec
'),
1538 'function
': lambda it: int(any(v != 'none
' for v in it))},
1539 'ie_pref
': {'priority': True, 'type': 'extractor'},
1540 'hasvid
': {'priority': True, 'field': 'vcodec', 'type': 'boolean', 'not_in_list': ('none',)},
1541 'hasaud
': {'field': 'acodec', 'type': 'boolean', 'not_in_list': ('none',)},
1542 'lang
': {'convert': 'ignore', 'field': 'language_preference'},
1543 'quality
': {'convert': 'float_none', 'default': -1},
1544 'filesize
': {'convert': 'bytes'},
1545 'fs_approx
': {'convert': 'bytes', 'field': 'filesize_approx'},
1546 'id': {'convert': 'string', 'field': 'format_id'},
1547 'height
': {'convert': 'float_none'},
1548 'width
': {'convert': 'float_none'},
1549 'fps
': {'convert': 'float_none'},
1550 'tbr
': {'convert': 'float_none'},
1551 'vbr
': {'convert': 'float_none'},
1552 'abr
': {'convert': 'float_none'},
1553 'asr
': {'convert': 'float_none'},
1554 'source
': {'convert': 'ignore', 'field': 'source_preference'},
1556 'codec
': {'type': 'combined', 'field': ('vcodec', 'acodec')},
1557 'br
': {'type': 'combined', 'field': ('tbr', 'vbr', 'abr'), 'same_limit': True},
1558 'size
': {'type': 'combined', 'same_limit': True, 'field': ('filesize', 'fs_approx')},
1559 'ext
': {'type': 'combined', 'field': ('vext', 'aext')},
1560 'res
': {'type': 'multiple
', 'field
': ('height
', 'width
'),
1561 'function
': lambda it: (lambda l: min(l) if l else 0)(tuple(filter(None, it)))},
1563 # Most of these exist only for compatibility reasons
1564 'dimension
': {'type': 'alias', 'field': 'res'},
1565 'resolution
': {'type': 'alias', 'field': 'res'},
1566 'extension
': {'type': 'alias', 'field': 'ext'},
1567 'bitrate
': {'type': 'alias', 'field': 'br'},
1568 'total_bitrate
': {'type': 'alias', 'field': 'tbr'},
1569 'video_bitrate
': {'type': 'alias', 'field': 'vbr'},
1570 'audio_bitrate
': {'type': 'alias', 'field': 'abr'},
1571 'framerate
': {'type': 'alias', 'field': 'fps'},
1572 'language_preference
': {'type': 'alias', 'field': 'lang'}, # not named as 'language
' because such a field exists
1573 'protocol
': {'type': 'alias', 'field': 'proto'},
1574 'source_preference
': {'type': 'alias', 'field': 'source'},
1575 'filesize_approx
': {'type': 'alias', 'field': 'fs_approx'},
1576 'filesize_estimate
': {'type': 'alias', 'field': 'size'},
1577 'samplerate
': {'type': 'alias', 'field': 'asr'},
1578 'video_ext
': {'type': 'alias', 'field': 'vext'},
1579 'audio_ext
': {'type': 'alias', 'field': 'aext'},
1580 'video_codec
': {'type': 'alias', 'field': 'vcodec'},
1581 'audio_codec
': {'type': 'alias', 'field': 'acodec'},
1582 'video
': {'type': 'alias', 'field': 'hasvid'},
1583 'has_video
': {'type': 'alias', 'field': 'hasvid'},
1584 'audio
': {'type': 'alias', 'field': 'hasaud'},
1585 'has_audio
': {'type': 'alias', 'field': 'hasaud'},
1586 'extractor
': {'type': 'alias', 'field': 'ie_pref'},
1587 'preference
': {'type': 'alias', 'field': 'ie_pref'},
1588 'extractor_preference
': {'type': 'alias', 'field': 'ie_pref'},
1589 'format_id
': {'type': 'alias', 'field': 'id'},
1594 def _get_field_setting(self, field, key):
1595 if field not in self.settings:
1596 self.settings[field] = {}
1597 propObj = self.settings[field]
1598 if key not in propObj:
1599 type = propObj.get('type')
1601 default = 'preference
' if type == 'extractor
' else (field,) if type in ('combined
', 'multiple
') else field
1602 elif key == 'convert
':
1603 default = 'order
' if type == 'ordered
' else 'float_string
' if field else 'ignore
'
1605 default = {'type': 'field', 'visible': True, 'order': [], 'not_in_list': (None,)}.get(key, None)
1606 propObj[key] = default
1609 def _resolve_field_value(self, field, value, convertNone=False):
1614 value = value.lower()
1615 conversion = self._get_field_setting(field, 'convert
')
1616 if conversion == 'ignore
':
1618 if conversion == 'string
':
1620 elif conversion == 'float_none
':
1621 return float_or_none(value)
1622 elif conversion == 'bytes':
1623 return FileDownloader.parse_bytes(value)
1624 elif conversion == 'order
':
1625 order_list = (self._use_free_order and self._get_field_setting(field, 'order_free
')) or self._get_field_setting(field, 'order
')
1626 use_regex = self._get_field_setting(field, 'regex
')
1627 list_length = len(order_list)
1628 empty_pos = order_list.index('') if '' in order_list else list_length + 1
1629 if use_regex and value is not None:
1630 for i, regex in enumerate(order_list):
1631 if regex and re.match(regex, value):
1632 return list_length - i
1633 return list_length - empty_pos # not in list
1634 else: # not regex or value = None
1635 return list_length - (order_list.index(value) if value in order_list else empty_pos)
1637 if value.isnumeric():
1640 self.settings[field]['convert
'] = 'string
'
1643 def evaluate_params(self, params, sort_extractor):
1644 self._use_free_order = params.get('prefer_free_formats
', False)
1645 self._sort_user = params.get('format_sort
', [])
1646 self._sort_extractor = sort_extractor
1648 def add_item(field, reverse, closest, limit_text):
1649 field = field.lower()
1650 if field in self._order:
1652 self._order.append(field)
1653 limit = self._resolve_field_value(field, limit_text)
1656 'closest
': False if limit is None else closest,
1657 'limit_text
': limit_text,
1659 if field in self.settings:
1660 self.settings[field].update(data)
1662 self.settings[field] = data
1665 tuple(field for field in self.default if self._get_field_setting(field, 'forced
'))
1666 + (tuple() if params.get('format_sort_force
', False)
1667 else tuple(field for field in self.default if self._get_field_setting(field, 'priority
')))
1668 + tuple(self._sort_user) + tuple(sort_extractor) + self.default)
1670 for item in sort_list:
1671 match = re.match(self.regex, item)
1673 raise ExtractorError('Invalid format sort string
"%s" given by extractor
' % item)
1674 field = match.group('field
')
1677 if self._get_field_setting(field, 'type') == 'alias
':
1678 field = self._get_field_setting(field, 'field
')
1679 reverse = match.group('reverse
') is not None
1680 closest = match.group('separator
') == '~
'
1681 limit_text = match.group('limit
')
1683 has_limit = limit_text is not None
1684 has_multiple_fields = self._get_field_setting(field, 'type') == 'combined
'
1685 has_multiple_limits = has_limit and has_multiple_fields and not self._get_field_setting(field, 'same_limit
')
1687 fields = self._get_field_setting(field, 'field
') if has_multiple_fields else (field,)
1688 limits = limit_text.split(':') if has_multiple_limits else (limit_text,) if has_limit else tuple()
1689 limit_count = len(limits)
1690 for (i, f) in enumerate(fields):
1691 add_item(f, reverse, closest,
1692 limits[i] if i < limit_count
1693 else limits[0] if has_limit and not has_multiple_limits
1696 def print_verbose_info(self, write_debug):
1698 write_debug('Sort order given by user
: %s' % ', '.join(self._sort_user))
1699 if self._sort_extractor:
1700 write_debug('Sort order given by extractor
: %s' % ', '.join(self._sort_extractor))
1701 write_debug('Formats
sorted by
: %s' % ', '.join(['%s%s%s' % (
1702 '+' if self._get_field_setting(field, 'reverse
') else '', field,
1703 '%s%s(%s)' % ('~
' if self._get_field_setting(field, 'closest
') else ':',
1704 self._get_field_setting(field, 'limit_text
'),
1705 self._get_field_setting(field, 'limit
'))
1706 if self._get_field_setting(field, 'limit_text
') is not None else '')
1707 for field in self._order if self._get_field_setting(field, 'visible
')]))
1709 def _calculate_field_preference_from_value(self, format, field, type, value):
1710 reverse = self._get_field_setting(field, 'reverse
')
1711 closest = self._get_field_setting(field, 'closest
')
1712 limit = self._get_field_setting(field, 'limit
')
1714 if type == 'extractor
':
1715 maximum = self._get_field_setting(field, 'max')
1716 if value is None or (maximum is not None and value >= maximum):
1718 elif type == 'boolean
':
1719 in_list = self._get_field_setting(field, 'in_list
')
1720 not_in_list = self._get_field_setting(field, 'not_in_list
')
1721 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
1722 elif type == 'ordered
':
1723 value = self._resolve_field_value(field, value, True)
1725 # try to convert to number
1726 val_num = float_or_none(value, default=self._get_field_setting(field, 'default
'))
1727 is_num = self._get_field_setting(field, 'convert
') != 'string
' and val_num is not None
1731 return ((-10, 0) if value is None
1732 else (1, value, 0) if not is_num # if a field has mixed strings and numbers, strings are sorted higher
1733 else (0, -abs(value - limit), value - limit if reverse else limit - value) if closest
1734 else (0, value, 0) if not reverse and (limit is None or value <= limit)
1735 else (0, -value, 0) if limit is None or (reverse and value == limit) or value > limit
1736 else (-1, value, 0))
1738 def _calculate_field_preference(self, format, field):
1739 type = self._get_field_setting(field, 'type') # extractor, boolean, ordered, field, multiple
1740 get_value = lambda f: format.get(self._get_field_setting(f, 'field
'))
1741 if type == 'multiple
':
1742 type = 'field
' # Only 'field
' is allowed in multiple for now
1743 actual_fields = self._get_field_setting(field, 'field
')
1745 value = self._get_field_setting(field, 'function
')(get_value(f) for f in actual_fields)
1747 value = get_value(field)
1748 return self._calculate_field_preference_from_value(format, field, type, value)
1750 def calculate_preference(self, format):
1751 # Determine missing protocol
1752 if not format.get('protocol
'):
1753 format['protocol
'] = determine_protocol(format)
1755 # Determine missing ext
1756 if not format.get('ext
') and 'url
' in format:
1757 format['ext
'] = determine_ext(format['url
'])
1758 if format.get('vcodec
') == 'none
':
1759 format['audio_ext
'] = format['ext
'] if format.get('acodec
') != 'none
' else 'none
'
1760 format['video_ext
'] = 'none
'
1762 format['video_ext
'] = format['ext
']
1763 format['audio_ext
'] = 'none
'
1764 # if format.get('preference
') is None and format.get('ext
') in ('f4f
', 'f4m
'): # Not supported?
1765 # format['preference
'] = -1000
1767 # Determine missing bitrates
1768 if format.get('tbr
') is None:
1769 if format.get('vbr
') is not None and format.get('abr
') is not None:
1770 format['tbr
'] = format.get('vbr
', 0) + format.get('abr
', 0)
1772 if format.get('vcodec
') != 'none
' and format.get('vbr
') is None:
1773 format['vbr
'] = format.get('tbr
') - format.get('abr
', 0)
1774 if format.get('acodec
') != 'none
' and format.get('abr
') is None:
1775 format['abr
'] = format.get('tbr
') - format.get('vbr
', 0)
1777 return tuple(self._calculate_field_preference(format, field) for field in self._order)
1779 def _sort_formats(self, formats, field_preference=[]):
1782 format_sort = self.FormatSort() # params and to_screen are taken from the downloader
1783 format_sort.evaluate_params(self._downloader.params, field_preference)
1784 if self.get_param('verbose
', False):
1785 format_sort.print_verbose_info(self._downloader.write_debug)
1786 formats.sort(key=lambda f: format_sort.calculate_preference(f))
1788 def _check_formats(self, formats, video_id):
1790 formats[:] = filter(
1791 lambda f: self._is_valid_url(
1793 item='%s video format
' % f.get('format_id
') if f.get('format_id
') else 'video
'),
1797 def _remove_duplicate_formats(formats):
1801 if f['url
'] not in format_urls:
1802 format_urls.add(f['url
'])
1803 unique_formats.append(f)
1804 formats[:] = unique_formats
1806 def _is_valid_url(self, url, video_id, item='video
', headers={}):
1807 url = self._proto_relative_url(url, scheme='http
:')
1808 # For now assume non HTTP(S) URLs always valid
1809 if not (url.startswith('http
://') or url.startswith('https
://')):
1812 self._request_webpage(url, video_id, 'Checking
%s URL
' % item, headers=headers)
1814 except ExtractorError as e:
1816 '%s: %s URL
is invalid
, skipping
: %s'
1817 % (video_id, item, error_to_compat_str(e.cause)))
1820 def http_scheme(self):
1821 """ Either "http:" or "https:", depending on the user's preferences
"""
1824 if self.get_param('prefer_insecure', False)
1827 def _proto_relative_url(self, url, scheme=None):
1830 if url.startswith('//'):
1832 scheme = self.http_scheme()
1837 def _sleep(self, timeout, video_id, msg_template=None):
1838 if msg_template is None:
1839 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1840 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1844 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1845 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1846 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
1847 manifest = self._download_xml(
1848 manifest_url, video_id, 'Downloading f4m manifest',
1849 'Unable to download f4m manifest',
1850 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1851 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
1852 transform_source=transform_source,
1853 fatal=fatal, data=data, headers=headers, query=query)
1855 if manifest is False:
1858 return self._parse_f4m_formats(
1859 manifest, manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1860 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1862 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1863 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1864 fatal=True, m3u8_id=None):
1865 if not isinstance(manifest, compat_etree_Element) and not fatal:
1868 # currently yt-dlp cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1869 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1870 if akamai_pv is not None and ';' in akamai_pv.text:
1871 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1872 if playerVerificationChallenge.strip() != '':
1876 manifest_version = '1.0'
1877 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1879 manifest_version = '2.0'
1880 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1881 # Remove unsupported DRM protected media from final formats
1882 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
1883 media_nodes = remove_encrypted_media(media_nodes)
1887 manifest_base_url = get_base_url(manifest)
1889 bootstrap_info = xpath_element(
1890 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1891 'bootstrap info', default=None)
1894 mime_type = xpath_text(
1895 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1896 'base URL', default=None)
1897 if mime_type and mime_type.startswith('audio/'):
1900 for i, media_el in enumerate(media_nodes):
1901 tbr = int_or_none(media_el.attrib.get('bitrate'))
1902 width = int_or_none(media_el.attrib.get('width'))
1903 height = int_or_none(media_el.attrib.get('height'))
1904 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1905 # If <bootstrapInfo> is present, the specified f4m is a
1906 # stream-level manifest, and only set-level manifests may refer to
1907 # external resources. See section 11.4 and section 4 of F4M spec
1908 if bootstrap_info is None:
1910 # @href is introduced in 2.0, see section 11.6 of F4M spec
1911 if manifest_version == '2.0':
1912 media_url = media_el.attrib.get('href')
1913 if media_url is None:
1914 media_url = media_el.attrib.get('url')
1918 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1919 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1920 # If media_url is itself a f4m manifest do the recursive extraction
1921 # since bitrates in parent manifest (this one) and media_url manifest
1922 # may differ leading to inability to resolve the format by requested
1923 # bitrate in f4m downloader
1924 ext = determine_ext(manifest_url)
1926 f4m_formats = self._extract_f4m_formats(
1927 manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1928 transform_source=transform_source, fatal=fatal)
1929 # Sometimes stream-level manifest contains single media entry that
1930 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1931 # At the same time parent's media entry in set-level manifest may
1932 # contain it. We will copy it from parent in such cases.
1933 if len(f4m_formats) == 1:
1936 'tbr': f.get('tbr') or tbr,
1937 'width': f.get('width') or width,
1938 'height': f.get('height') or height,
1939 'format_id': f.get('format_id') if not tbr else format_id,
1942 formats.extend(f4m_formats)
1945 formats.extend(self._extract_m3u8_formats(
1946 manifest_url, video_id, 'mp4', preference=preference,
1947 quality=quality, m3u8_id=m3u8_id, fatal=fatal))
1950 'format_id': format_id,
1951 'url': manifest_url,
1952 'manifest_url': manifest_url,
1953 'ext': 'flv' if bootstrap_info is not None else None,
1959 'preference': preference,
1964 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, quality=None, m3u8_id=None):
1966 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1970 'preference': preference - 100 if preference else -100,
1972 'resolution': 'multiple',
1973 'format_note': 'Quality selection URL',
1976 def _report_ignoring_subs(self, name):
1977 self.report_warning(bug_reports_message(
1978 f'Ignoring subtitle tracks found in the {name} manifest; '
1979 'if any subtitle tracks are missing,'
1982 def _extract_m3u8_formats(self, *args, **kwargs):
1983 fmts, subs = self._extract_m3u8_formats_and_subtitles(*args, **kwargs)
1985 self._report_ignoring_subs('HLS')
1988 def _extract_m3u8_formats_and_subtitles(
1989 self, m3u8_url, video_id, ext=None, entry_protocol='m3u8_native',
1990 preference=None, quality=None, m3u8_id=None, note=None,
1991 errnote=None, fatal=True, live=False, data=None, headers={},
1994 res = self._download_webpage_handle(
1996 note='Downloading m3u8 information' if note is None else note,
1997 errnote='Failed to download m3u8 information' if errnote is None else errnote,
1998 fatal=fatal, data=data, headers=headers, query=query)
2003 m3u8_doc, urlh = res
2004 m3u8_url = urlh.geturl()
2006 return self._parse_m3u8_formats_and_subtitles(
2007 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
2008 preference=preference, quality=quality, m3u8_id=m3u8_id,
2009 note=note, errnote=errnote, fatal=fatal, live=live, data=data,
2010 headers=headers, query=query, video_id=video_id)
2012 def _parse_m3u8_formats_and_subtitles(
2013 self, m3u8_doc, m3u8_url, ext=None, entry_protocol='m3u8_native',
2014 preference=None, quality=None, m3u8_id=None, live=False, note=None,
2015 errnote=None, fatal=True, data=None, headers={}, query={},
2017 formats, subtitles = [], {}
2019 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
2020 return formats, subtitles
2022 has_drm = re.search(r'#EXT-X-(?:SESSION-)?KEY:.*?URI="skd://', m3u8_doc)
2024 def format_url(url):
2025 return url if re.match(r'^https?://', url) else compat_urlparse.urljoin(m3u8_url, url)
2027 if self.get_param('hls_split_discontinuity', False):
2028 def _extract_m3u8_playlist_indices(manifest_url=None, m3u8_doc=None):
2030 if not manifest_url:
2032 m3u8_doc = self._download_webpage(
2033 manifest_url, video_id, fatal=fatal, data=data, headers=headers,
2034 note=False, errnote='Failed to download m3u8 playlist information')
2035 if m3u8_doc is False:
2037 return range(1 + sum(line.startswith('#EXT-X-DISCONTINUITY') for line in m3u8_doc.splitlines()))
2040 def _extract_m3u8_playlist_indices(*args, **kwargs):
2044 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
2045 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
2046 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
2048 # We should try extracting formats only from master playlists [1, 4.3.4],
2049 # i.e. playlists that describe available qualities. On the other hand
2050 # media playlists [1, 4.3.3] should be returned as is since they contain
2051 # just the media without qualities renditions.
2052 # Fortunately, master playlist can be easily distinguished from media
2053 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
2054 # master playlist tags MUST NOT appear in a media playlist and vice versa.
2055 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
2056 # media playlist and MUST NOT appear in master playlist thus we can
2057 # clearly detect media playlist with this criterion.
2059 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
2061 'format_id': '-'.join(map(str, filter(None, [m3u8_id, idx]))),
2062 'format_index': idx,
2065 'protocol': entry_protocol,
2066 'preference': preference,
2069 } for idx in _extract_m3u8_playlist_indices(m3u8_doc=m3u8_doc)]
2071 return formats, subtitles
2074 last_stream_inf = {}
2076 def extract_media(x_media_line):
2077 media = parse_m3u8_attributes(x_media_line)
2078 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
2079 media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME')
2080 if not (media_type and group_id and name):
2082 groups.setdefault(group_id, []).append(media)
2083 # <https://tools.ietf.org/html/rfc8216#section-4.3.4.1>
2084 if media_type == 'SUBTITLES':
2085 # According to RFC 8216 §4.3.4.2.1, URI is REQUIRED in the
2086 # EXT-X-MEDIA tag if the media type is SUBTITLES.
2087 # However, lack of URI has been spotted in the wild.
2088 # e.g. NebulaIE; see https://github.com/yt-dlp/yt-dlp/issues/339
2089 if not media.get('URI'):
2091 url = format_url(media['URI'])
2094 'ext': determine_ext(url),
2096 if sub_info['ext'] == 'm3u8':
2097 # Per RFC 8216 §3.1, the only possible subtitle format m3u8
2098 # files may contain is WebVTT:
2099 # <https://tools.ietf.org/html/rfc8216#section-3.1>
2100 sub_info['ext'] = 'vtt'
2101 sub_info['protocol'] = 'm3u8_native'
2102 lang = media.get('LANGUAGE') or 'und'
2103 subtitles.setdefault(lang, []).append(sub_info)
2104 if media_type not in ('VIDEO', 'AUDIO'):
2106 media_url = media.get('URI')
2108 manifest_url = format_url(media_url)
2110 'format_id': '-'.join(map(str, filter(None, (m3u8_id, group_id, name, idx)))),
2111 'format_note': name,
2112 'format_index': idx,
2113 'url': manifest_url,
2114 'manifest_url': m3u8_url,
2115 'language': media.get('LANGUAGE'),
2117 'protocol': entry_protocol,
2118 'preference': preference,
2120 'vcodec': 'none' if media_type == 'AUDIO' else None,
2121 } for idx in _extract_m3u8_playlist_indices(manifest_url))
2123 def build_stream_name():
2124 # Despite specification does not mention NAME attribute for
2125 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
2126 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
2127 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
2128 stream_name = last_stream_inf.get('NAME')
2131 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
2132 # from corresponding rendition group
2133 stream_group_id = last_stream_inf.get('VIDEO')
2134 if not stream_group_id:
2136 stream_group = groups.get(stream_group_id)
2137 if not stream_group:
2138 return stream_group_id
2139 rendition = stream_group[0]
2140 return rendition.get('NAME') or stream_group_id
2142 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
2143 # chance to detect video only formats when EXT-X-STREAM-INF tags
2144 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
2145 for line in m3u8_doc.splitlines():
2146 if line.startswith('#EXT-X-MEDIA:'):
2149 for line in m3u8_doc.splitlines():
2150 if line.startswith('#EXT-X-STREAM-INF:'):
2151 last_stream_inf = parse_m3u8_attributes(line)
2152 elif line.startswith('#') or not line.strip():
2155 tbr = float_or_none(
2156 last_stream_inf.get('AVERAGE-BANDWIDTH')
2157 or last_stream_inf.get('BANDWIDTH'), scale=1000)
2158 manifest_url = format_url(line.strip())
2160 for idx in _extract_m3u8_playlist_indices(manifest_url):
2161 format_id = [m3u8_id, None, idx]
2162 # Bandwidth of live streams may differ over time thus making
2163 # format_id unpredictable. So it's better to keep provided
2166 stream_name = build_stream_name()
2167 format_id[1] = stream_name if stream_name else '%d' % (tbr if tbr else len(formats))
2169 'format_id': '-'.join(map(str, filter(None, format_id))),
2170 'format_index': idx,
2171 'url': manifest_url,
2172 'manifest_url': m3u8_url,
2175 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')),
2176 'protocol': entry_protocol,
2177 'preference': preference,
2180 resolution = last_stream_inf.get('RESOLUTION')
2182 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
2184 f['width'] = int(mobj.group('width'))
2185 f['height'] = int(mobj.group('height'))
2186 # Unified Streaming Platform
2188 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
2190 abr, vbr = mobj.groups()
2191 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
2196 codecs = parse_codecs(last_stream_inf.get('CODECS'))
2198 audio_group_id = last_stream_inf.get('AUDIO')
2199 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
2200 # references a rendition group MUST have a CODECS attribute.
2201 # However, this is not always respected, for example, [2]
2202 # contains EXT-X-STREAM-INF tag which references AUDIO
2203 # rendition group but does not have CODECS and despite
2204 # referencing an audio group it represents a complete
2205 # (with audio and video) format. So, for such cases we will
2206 # ignore references to rendition groups and treat them
2207 # as complete formats.
2208 if audio_group_id and codecs and f.get('vcodec') != 'none':
2209 audio_group = groups.get(audio_group_id)
2210 if audio_group and audio_group[0].get('URI'):
2211 # TODO: update acodec for audio only formats with
2213 f['acodec'] = 'none'
2214 if not f.get('ext'):
2215 f['ext'] = 'm4a' if f.get('vcodec') == 'none' else 'mp4'
2219 progressive_uri = last_stream_inf.get('PROGRESSIVE-URI')
2222 del http_f['manifest_url']
2224 'format_id': f['format_id'].replace('hls-', 'http-'),
2226 'url': progressive_uri,
2228 formats.append(http_f)
2230 last_stream_inf = {}
2231 return formats, subtitles
2233 def _extract_m3u8_vod_duration(
2234 self, m3u8_vod_url, video_id, note=None, errnote=None, data=None, headers={}, query={}):
2236 m3u8_vod = self._download_webpage(
2237 m3u8_vod_url, video_id,
2238 note='Downloading m3u8 VOD manifest' if note is None else note,
2239 errnote='Failed to download VOD manifest' if errnote is None else errnote,
2240 fatal=False, data=data, headers=headers, query=query)
2242 return self._parse_m3u8_vod_duration(m3u8_vod or '', video_id)
2244 def _parse_m3u8_vod_duration(self, m3u8_vod, video_id):
2245 if '#EXT-X-PLAYLIST-TYPE:VOD' not in m3u8_vod:
2249 float(line[len('#EXTINF:'):].split(',')[0])
2250 for line in m3u8_vod.splitlines() if line.startswith('#EXTINF:'))) or None
2253 def _xpath_ns(path, namespace=None):
2257 for c in path.split('/'):
2258 if not c or c == '.':
2261 out.append('{%s}%s' % (namespace, c))
2262 return '/'.join(out)
2264 def _extract_smil_formats_and_subtitles(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
2265 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
2271 namespace = self._parse_smil_namespace(smil)
2273 fmts = self._parse_smil_formats(
2274 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
2275 subs = self._parse_smil_subtitles(
2276 smil, namespace=namespace)
2280 def _extract_smil_formats(self, *args, **kwargs):
2281 fmts, subs = self._extract_smil_formats_and_subtitles(*args, **kwargs)
2283 self._report_ignoring_subs('SMIL')
2286 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
2287 smil = self._download_smil(smil_url, video_id, fatal=fatal)
2290 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
2292 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
2293 return self._download_xml(
2294 smil_url, video_id, 'Downloading SMIL file',
2295 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
2297 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
2298 namespace = self._parse_smil_namespace(smil)
2300 formats = self._parse_smil_formats(
2301 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
2302 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
2304 video_id = os.path.splitext(url_basename(smil_url))[0]
2308 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
2309 name = meta.attrib.get('name')
2310 content = meta.attrib.get('content')
2311 if not name or not content:
2313 if not title and name == 'title':
2315 elif not description and name in ('description', 'abstract'):
2316 description = content
2317 elif not upload_date and name == 'date':
2318 upload_date = unified_strdate(content)
2321 'id': image.get('type'),
2322 'url': image.get('src'),
2323 'width': int_or_none(image.get('width')),
2324 'height': int_or_none(image.get('height')),
2325 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
2329 'title': title or video_id,
2330 'description': description,
2331 'upload_date': upload_date,
2332 'thumbnails': thumbnails,
2334 'subtitles': subtitles,
2337 def _parse_smil_namespace(self, smil):
2338 return self._search_regex(
2339 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
2341 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
2343 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
2344 b = meta.get('base') or meta.get('httpBase')
2356 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
2357 for medium in media:
2358 src = medium.get('src')
2359 if not src or src in srcs:
2363 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
2364 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
2365 width = int_or_none(medium.get('width'))
2366 height = int_or_none(medium.get('height'))
2367 proto = medium.get('proto')
2368 ext = medium.get('ext')
2369 src_ext = determine_ext(src)
2370 streamer = medium.get('streamer') or base
2372 if proto == 'rtmp' or streamer.startswith('rtmp'):
2378 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
2380 'filesize': filesize,
2384 if transform_rtmp_url:
2385 streamer, src = transform_rtmp_url(streamer, src)
2386 formats[-1].update({
2392 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
2393 src_url = src_url.strip()
2395 if proto == 'm3u8' or src_ext == 'm3u8':
2396 m3u8_formats = self._extract_m3u8_formats(
2397 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
2398 if len(m3u8_formats) == 1:
2400 m3u8_formats[0].update({
2401 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
2406 formats.extend(m3u8_formats)
2407 elif src_ext == 'f4m':
2412 'plugin': 'flowplayer-3.2.0.1',
2414 f4m_url += '&' if '?' in f4m_url else '?'
2415 f4m_url += compat_urllib_parse_urlencode(f4m_params)
2416 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
2417 elif src_ext == 'mpd':
2418 formats.extend(self._extract_mpd_formats(
2419 src_url, video_id, mpd_id='dash', fatal=False))
2420 elif re.search(r'\.ism/[Mm]anifest', src_url):
2421 formats.extend(self._extract_ism_formats(
2422 src_url, video_id, ism_id='mss', fatal=False))
2423 elif src_url.startswith('http') and self._is_valid_url(src, video_id):
2427 'ext': ext or src_ext or 'flv',
2428 'format_id': 'http-%d' % (bitrate or http_count),
2430 'filesize': filesize,
2435 for medium in smil.findall(self._xpath_ns('.//imagestream', namespace)):
2436 src = medium.get('src')
2437 if not src or src in srcs:
2443 'format_id': 'imagestream-%d' % (imgs_count),
2445 'ext': mimetype2ext(medium.get('type')),
2448 'width': int_or_none(medium.get('width')),
2449 'height': int_or_none(medium.get('height')),
2450 'format_note': 'SMIL storyboards',
2455 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
2458 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
2459 src = textstream.get('src')
2460 if not src or src in urls:
2463 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
2464 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
2465 subtitles.setdefault(lang, []).append({
2471 def _extract_xspf_playlist(self, xspf_url, playlist_id, fatal=True):
2472 xspf = self._download_xml(
2473 xspf_url, playlist_id, 'Downloading xpsf playlist',
2474 'Unable to download xspf manifest', fatal=fatal)
2477 return self._parse_xspf(
2478 xspf, playlist_id, xspf_url=xspf_url,
2479 xspf_base_url=base_url(xspf_url))
2481 def _parse_xspf(self, xspf_doc, playlist_id, xspf_url=None, xspf_base_url=None):
2483 'xspf': 'http://xspf.org/ns/0/',
2484 's1': 'http://static.streamone.nl/player/ns/0',
2488 for track in xspf_doc.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
2490 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
2491 description = xpath_text(
2492 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
2493 thumbnail = xpath_text(
2494 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
2495 duration = float_or_none(
2496 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
2499 for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP)):
2500 format_url = urljoin(xspf_base_url, location.text)
2505 'manifest_url': xspf_url,
2506 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
2507 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
2508 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
2510 self._sort_formats(formats)
2515 'description': description,
2516 'thumbnail': thumbnail,
2517 'duration': duration,
2522 def _extract_mpd_formats(self, *args, **kwargs):
2523 fmts, subs = self._extract_mpd_formats_and_subtitles(*args, **kwargs)
2525 self._report_ignoring_subs('DASH')
2528 def _extract_mpd_formats_and_subtitles(
2529 self, mpd_url, video_id, mpd_id=None, note=None, errnote=None,
2530 fatal=True, data=None, headers={}, query={}):
2531 res = self._download_xml_handle(
2533 note='Downloading MPD manifest' if note is None else note,
2534 errnote='Failed to download MPD manifest' if errnote is None else errnote,
2535 fatal=fatal, data=data, headers=headers, query=query)
2541 mpd_base_url = base_url(urlh.geturl())
2543 return self._parse_mpd_formats_and_subtitles(
2544 mpd_doc, mpd_id, mpd_base_url, mpd_url)
2546 def _parse_mpd_formats(self, *args, **kwargs):
2547 fmts, subs = self._parse_mpd_formats_and_subtitles(*args, **kwargs)
2549 self._report_ignoring_subs('DASH')
2552 def _parse_mpd_formats_and_subtitles(
2553 self, mpd_doc, mpd_id=None, mpd_base_url='', mpd_url=None):
2555 Parse formats
from MPD manifest
.
2557 1. MPEG
-DASH Standard
, ISO
/IEC
23009-1:2014(E
),
2558 http
://standards
.iso
.org
/ittf
/PubliclyAvailableStandards
/c065274_ISO_IEC_23009
-1_2014.zip
2559 2. https
://en
.wikipedia
.org
/wiki
/Dynamic_Adaptive_Streaming_over_HTTP
2561 if not self.get_param('dynamic_mpd', True):
2562 if mpd_doc.get('type') == 'dynamic':
2565 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
2568 return self._xpath_ns(path, namespace)
2570 def is_drm_protected(element):
2571 return element.find(_add_ns('ContentProtection')) is not None
2573 def extract_multisegment_info(element, ms_parent_info):
2574 ms_info = ms_parent_info.copy()
2576 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2577 # common attributes and elements. We will only extract relevant
2579 def extract_common(source):
2580 segment_timeline = source.find(_add_ns('SegmentTimeline'))
2581 if segment_timeline is not None:
2582 s_e = segment_timeline.findall(_add_ns('S'))
2584 ms_info['total_number'] = 0
2587 r = int(s.get('r', 0))
2588 ms_info['total_number'] += 1 + r
2589 ms_info['s'].append({
2590 't': int(s.get('t', 0)),
2591 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2592 'd': int(s.attrib['d']),
2595 start_number = source.get('startNumber')
2597 ms_info['start_number'] = int(start_number)
2598 timescale = source.get('timescale')
2600 ms_info['timescale'] = int(timescale)
2601 segment_duration = source.get('duration')
2602 if segment_duration:
2603 ms_info['segment_duration'] = float(segment_duration)
2605 def extract_Initialization(source):
2606 initialization = source.find(_add_ns('Initialization'))
2607 if initialization is not None:
2608 ms_info['initialization_url'] = initialization.attrib['sourceURL']
2610 segment_list = element.find(_add_ns('SegmentList'))
2611 if segment_list is not None:
2612 extract_common(segment_list)
2613 extract_Initialization(segment_list)
2614 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
2616 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
2618 segment_template = element.find(_add_ns('SegmentTemplate'))
2619 if segment_template is not None:
2620 extract_common(segment_template)
2621 media = segment_template.get('media')
2623 ms_info['media'] = media
2624 initialization = segment_template.get('initialization')
2626 ms_info['initialization'] = initialization
2628 extract_Initialization(segment_template)
2631 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
2632 formats, subtitles = [], {}
2633 stream_numbers = {'audio': 0, 'video': 0}
2634 for period in mpd_doc.findall(_add_ns('Period')):
2635 period_duration = parse_duration(period.get('duration')) or mpd_duration
2636 period_ms_info = extract_multisegment_info(period, {
2640 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
2641 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
2642 for representation in adaptation_set.findall(_add_ns('Representation')):
2643 representation_attrib = adaptation_set.attrib.copy()
2644 representation_attrib.update(representation.attrib)
2645 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2646 mime_type = representation_attrib['mimeType']
2647 content_type = representation_attrib.get('contentType', mime_type.split('/')[0])
2649 codecs = representation_attrib.get('codecs', '')
2650 if content_type not in ('video', 'audio', 'text'):
2651 if mime_type == 'image/jpeg':
2652 content_type = mime_type
2653 elif codecs.split('.')[0] == 'stpp':
2654 content_type = 'text'
2655 elif mimetype2ext(mime_type) in ('tt', 'dfxp', 'ttml', 'xml', 'json'):
2656 content_type = 'text'
2658 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
2662 for element in (representation, adaptation_set, period, mpd_doc):
2663 base_url_e = element.find(_add_ns('BaseURL'))
2664 if base_url_e is not None:
2665 base_url = base_url_e.text + base_url
2666 if re.match(r'^https?://', base_url):
2668 if mpd_base_url and base_url.startswith('/'):
2669 base_url = compat_urlparse.urljoin(mpd_base_url, base_url)
2670 elif mpd_base_url and not re.match(r'^https?://', base_url):
2671 if not mpd_base_url.endswith('/'):
2673 base_url = mpd_base_url + base_url
2674 representation_id = representation_attrib.get('id')
2675 lang = representation_attrib.get('lang')
2676 url_el = representation.find(_add_ns('BaseURL'))
2677 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
2678 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
2679 if representation_id is not None:
2680 format_id = representation_id
2682 format_id = content_type
2684 format_id = mpd_id + '-' + format_id
2685 if content_type in ('video', 'audio'):
2687 'format_id': format_id,
2688 'manifest_url': mpd_url,
2689 'ext': mimetype2ext(mime_type),
2690 'width': int_or_none(representation_attrib.get('width')),
2691 'height': int_or_none(representation_attrib.get('height')),
2692 'tbr': float_or_none(bandwidth, 1000),
2693 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
2694 'fps': int_or_none(representation_attrib.get('frameRate')),
2695 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
2696 'format_note': 'DASH %s' % content_type,
2697 'filesize': filesize,
2698 'container': mimetype2ext(mime_type) + '_dash',
2699 'manifest_stream_number': stream_numbers[content_type]
2701 f.update(parse_codecs(codecs))
2702 stream_numbers[content_type] += 1
2703 elif content_type == 'text':
2705 'ext': mimetype2ext(mime_type),
2706 'manifest_url': mpd_url,
2707 'filesize': filesize,
2709 elif content_type == 'image/jpeg':
2710 # See test case in VikiIE
2711 # https://www.viki.com/videos/1175236v-choosing-spouse-by-lottery-episode-1
2713 'format_id': format_id,
2715 'manifest_url': mpd_url,
2716 'format_note': 'DASH storyboards (jpeg)',
2720 if is_drm_protected(adaptation_set) or is_drm_protected(representation):
2722 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
2724 def prepare_template(template_name, identifiers):
2725 tmpl = representation_ms_info[template_name]
2726 # First of, % characters outside $...$ templates
2727 # must be escaped by doubling for proper processing
2728 # by % operator string formatting used further (see
2729 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2735 in_template = not in_template
2736 elif c == '%' and not in_template:
2738 # Next, $...$ templates are translated to their
2739 # %(...) counterparts to be used with % operator
2740 if representation_id is not None:
2741 t = t.replace('$RepresentationID$', representation_id)
2742 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
2743 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
2744 t.replace('$$', '$')
2747 # @initialization is a regular template like @media one
2748 # so it should be handled just the same way (see
2749 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2750 if 'initialization' in representation_ms_info:
2751 initialization_template = prepare_template(
2753 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2754 # $Time$ shall not be included for @initialization thus
2755 # only $Bandwidth$ remains
2757 representation_ms_info['initialization_url'] = initialization_template % {
2758 'Bandwidth': bandwidth,
2761 def location_key(location):
2762 return 'url' if re.match(r'^https?://', location) else 'path'
2764 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
2766 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2767 media_location_key = location_key(media_template)
2769 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2770 # can't be used at the same time
2771 if '%(Number' in media_template and 's' not in representation_ms_info:
2772 segment_duration = None
2773 if 'total_number' not in representation_ms_info and 'segment_duration' in representation_ms_info:
2774 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
2775 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
2776 representation_ms_info['fragments'] = [{
2777 media_location_key: media_template % {
2778 'Number': segment_number,
2779 'Bandwidth': bandwidth,
2781 'duration': segment_duration,
2782 } for segment_number in range(
2783 representation_ms_info['start_number'],
2784 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
2786 # $Number*$ or $Time$ in media template with S list available
2787 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2788 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2789 representation_ms_info['fragments'] = []
2792 segment_number = representation_ms_info['start_number']
2794 def add_segment_url():
2795 segment_url = media_template % {
2796 'Time': segment_time,
2797 'Bandwidth': bandwidth,
2798 'Number': segment_number,
2800 representation_ms_info['fragments'].append({
2801 media_location_key: segment_url,
2802 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
2805 for num, s in enumerate(representation_ms_info['s']):
2806 segment_time = s.get('t') or segment_time
2810 for r in range(s.get('r', 0)):
2811 segment_time += segment_d
2814 segment_time += segment_d
2815 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
2817 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2818 # or any YouTube dashsegments video
2821 timescale = representation_ms_info['timescale']
2822 for s in representation_ms_info['s']:
2823 duration = float_or_none(s['d'], timescale)
2824 for r in range(s.get('r', 0) + 1):
2825 segment_uri = representation_ms_info['segment_urls'][segment_index]
2827 location_key(segment_uri): segment_uri,
2828 'duration': duration,
2831 representation_ms_info['fragments'] = fragments
2832 elif 'segment_urls' in representation_ms_info:
2833 # Segment URLs with no SegmentTimeline
2834 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2835 # https://github.com/ytdl-org/youtube-dl/pull/14844
2837 segment_duration = float_or_none(
2838 representation_ms_info['segment_duration'],
2839 representation_ms_info['timescale']) if 'segment_duration' in representation_ms_info else None
2840 for segment_url in representation_ms_info['segment_urls']:
2842 location_key(segment_url): segment_url,
2844 if segment_duration:
2845 fragment['duration'] = segment_duration
2846 fragments.append(fragment)
2847 representation_ms_info['fragments'] = fragments
2848 # If there is a fragments key available then we correctly recognized fragmented media.
2849 # Otherwise we will assume unfragmented media with direct access. Technically, such
2850 # assumption is not necessarily correct since we may simply have no support for
2851 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2852 if 'fragments' in representation_ms_info:
2854 # NB: mpd_url may be empty when MPD manifest is parsed from a string
2855 'url': mpd_url or base_url,
2856 'fragment_base_url': base_url,
2858 'protocol': 'http_dash_segments' if mime_type != 'image/jpeg' else 'mhtml',
2860 if 'initialization_url' in representation_ms_info:
2861 initialization_url = representation_ms_info['initialization_url']
2862 if not f.get('url'):
2863 f['url'] = initialization_url
2864 f['fragments'].append({location_key(initialization_url): initialization_url})
2865 f['fragments'].extend(representation_ms_info['fragments'])
2867 # Assuming direct URL to unfragmented media.
2869 if content_type in ('video', 'audio') or mime_type == 'image/jpeg':
2871 elif content_type == 'text':
2872 subtitles.setdefault(lang or 'und', []).append(f)
2874 return formats, subtitles
2876 def _extract_ism_formats(self, *args, **kwargs):
2877 fmts, subs = self._extract_ism_formats_and_subtitles(*args, **kwargs)
2879 self._report_ignoring_subs('ISM')
2882 def _extract_ism_formats_and_subtitles(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
2883 res = self._download_xml_handle(
2885 note='Downloading ISM manifest' if note is None else note,
2886 errnote='Failed to download ISM manifest' if errnote is None else errnote,
2887 fatal=fatal, data=data, headers=headers, query=query)
2894 return self._parse_ism_formats_and_subtitles(ism_doc, urlh.geturl(), ism_id)
2896 def _parse_ism_formats_and_subtitles(self, ism_doc, ism_url, ism_id=None):
2898 Parse formats
from ISM manifest
.
2900 1. [MS
-SSTR
]: Smooth Streaming Protocol
,
2901 https
://msdn
.microsoft
.com
/en
-us
/library
/ff469518
.aspx
2903 if ism_doc.get('IsLive') == 'TRUE':
2906 duration = int(ism_doc.attrib['Duration'])
2907 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
2911 for stream in ism_doc.findall('StreamIndex'):
2912 stream_type = stream.get('Type')
2913 if stream_type not in ('video', 'audio', 'text'):
2915 url_pattern = stream.attrib['Url']
2916 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
2917 stream_name = stream.get('Name')
2918 stream_language = stream.get('Language', 'und')
2919 for track in stream.findall('QualityLevel'):
2920 fourcc = track.get('FourCC') or ('AACL' if track.get('AudioTag') == '255' else None)
2921 # TODO: add support for WVC1 and WMAP
2922 if fourcc not in ('H264', 'AVC1', 'AACL', 'TTML'):
2923 self.report_warning('%s is not a supported codec' % fourcc)
2925 tbr = int(track.attrib['Bitrate']) // 1000
2926 # [1] does not mention Width and Height attributes. However,
2927 # they're often present while MaxWidth and MaxHeight are
2928 # missing, so should be used as fallbacks
2929 width = int_or_none(track.get('MaxWidth') or track.get('Width'))
2930 height = int_or_none(track.get('MaxHeight') or track.get('Height'))
2931 sampling_rate = int_or_none(track.get('SamplingRate'))
2933 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
2934 track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
2940 stream_fragments = stream.findall('c')
2941 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
2942 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
2943 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
2944 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
2945 if not fragment_ctx['duration']:
2947 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
2949 next_fragment_time = duration
2950 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
2951 for _ in range(fragment_repeat):
2953 'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
2954 'duration': fragment_ctx['duration'] / stream_timescale,
2956 fragment_ctx['time'] += fragment_ctx['duration']
2960 format_id.append(ism_id)
2962 format_id.append(stream_name)
2963 format_id.append(compat_str(tbr))
2965 if stream_type == 'text':
2966 subtitles.setdefault(stream_language, []).append({
2970 'manifest_url': ism_url,
2971 'fragments': fragments,
2972 '_download_params': {
2973 'stream_type': stream_type,
2974 'duration': duration,
2975 'timescale': stream_timescale,
2977 'language': stream_language,
2978 'codec_private_data': track.get('CodecPrivateData'),
2981 elif stream_type in ('video', 'audio'):
2983 'format_id': '-'.join(format_id),
2985 'manifest_url': ism_url,
2986 'ext': 'ismv' if stream_type == 'video' else 'isma',
2990 'asr': sampling_rate,
2991 'vcodec': 'none' if stream_type == 'audio' else fourcc,
2992 'acodec': 'none' if stream_type == 'video' else fourcc,
2994 'fragments': fragments,
2995 'has_drm': ism_doc.find('Protection') is not None,
2996 '_download_params': {
2997 'stream_type': stream_type,
2998 'duration': duration,
2999 'timescale': stream_timescale,
3000 'width': width or 0,
3001 'height': height or 0,
3003 'language': stream_language,
3004 'codec_private_data': track.get('CodecPrivateData'),
3005 'sampling_rate': sampling_rate,
3006 'channels': int_or_none(track.get('Channels', 2)),
3007 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
3008 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
3011 return formats, subtitles
3013 def _parse_html5_media_entries(self, base_url, webpage, video_id, m3u8_id=None, m3u8_entry_protocol='m3u8', mpd_id=None, preference=None, quality=None):
3014 def absolute_url(item_url):
3015 return urljoin(base_url, item_url)
3017 def parse_content_type(content_type):
3018 if not content_type:
3020 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
3022 mimetype, codecs = ctr.groups()
3023 f = parse_codecs(codecs)
3024 f['ext'] = mimetype2ext(mimetype)
3028 def _media_formats(src, cur_media_type, type_info={}):
3029 full_url = absolute_url(src)
3030 ext = type_info.get('ext') or determine_ext(full_url)
3032 is_plain_url = False
3033 formats = self._extract_m3u8_formats(
3034 full_url, video_id, ext='mp4',
3035 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
3036 preference=preference, quality=quality, fatal=False)
3038 is_plain_url = False
3039 formats = self._extract_mpd_formats(
3040 full_url, video_id, mpd_id=mpd_id, fatal=False)
3045 'vcodec': 'none' if cur_media_type == 'audio' else None,
3047 return is_plain_url, formats
3050 # amp-video and amp-audio are very similar to their HTML5 counterparts
3051 # so we wll include them right here (see
3052 # https://www.ampproject.org/docs/reference/components/amp-video)
3053 # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/
3054 _MEDIA_TAG_NAME_RE = r'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)'
3055 media_tags = [(media_tag, media_tag_name, media_type, '')
3056 for media_tag, media_tag_name, media_type
3057 in re.findall(r'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE, webpage)]
3058 media_tags.extend(re.findall(
3059 # We only allow video|audio followed by a whitespace or '>'.
3060 # Allowing more characters may end up in significant slow down (see
3061 # https://github.com/ytdl-org/youtube-dl/issues/11979, example URL:
3062 # http://www.porntrex.com/maps/videositemap.xml).
3063 r'(?s)(<(?P<tag>%s)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>' % _MEDIA_TAG_NAME_RE, webpage))
3064 for media_tag, _, media_type, media_content in media_tags:
3069 media_attributes = extract_attributes(media_tag)
3070 src = strip_or_none(media_attributes.get('src'))
3072 _, formats = _media_formats(src, media_type)
3073 media_info['formats'].extend(formats)
3074 media_info['thumbnail'] = absolute_url(media_attributes.get('poster'))
3076 for source_tag in re.findall(r'<source[^>]+>', media_content):
3077 s_attr = extract_attributes(source_tag)
3078 # data-video-src and data-src are non standard but seen
3079 # several times in the wild
3080 src = strip_or_none(dict_get(s_attr, ('src', 'data-video-src', 'data-src')))
3083 f = parse_content_type(s_attr.get('type'))
3084 is_plain_url, formats = _media_formats(src, media_type, f)
3086 # width, height, res, label and title attributes are
3087 # all not standard but seen several times in the wild
3090 for lbl in ('label', 'title')
3091 if str_or_none(s_attr.get(lbl))
3093 width = int_or_none(s_attr.get('width'))
3094 height = (int_or_none(s_attr.get('height'))
3095 or int_or_none(s_attr.get('res')))
3096 if not width or not height:
3098 resolution = parse_resolution(lbl)
3101 width = width or resolution.get('width')
3102 height = height or resolution.get('height')
3104 tbr = parse_bitrate(lbl)
3113 'format_id': s_attr.get('label') or s_attr.get('title'),
3115 f.update(formats[0])
3116 media_info['formats'].append(f)
3118 media_info['formats'].extend(formats)
3119 for track_tag in re.findall(r'<track[^>]+>', media_content):
3120 track_attributes = extract_attributes(track_tag)
3121 kind = track_attributes.get('kind')
3122 if not kind or kind in ('subtitles', 'captions'):
3123 src = strip_or_none(track_attributes.get('src'))
3126 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
3127 media_info['subtitles'].setdefault(lang, []).append({
3128 'url': absolute_url(src),
3130 for f in media_info['formats']:
3131 f.setdefault('http_headers', {})['Referer'] = base_url
3132 if media_info['formats'] or media_info['subtitles']:
3133 entries.append(media_info)
3136 def _extract_akamai_formats(self, *args, **kwargs):
3137 fmts, subs = self._extract_akamai_formats_and_subtitles(*args, **kwargs)
3139 self._report_ignoring_subs('akamai')
3142 def _extract_akamai_formats_and_subtitles(self, manifest_url, video_id, hosts={}):
3143 signed = 'hdnea=' in manifest_url
3145 # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
3146 manifest_url = re.sub(
3147 r'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
3148 '', manifest_url).strip('?')
3153 hdcore_sign = 'hdcore=3.7.0'
3154 f4m_url = re.sub(r'(https?://[^/]+)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
3155 hds_host = hosts.get('hds')
3157 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
3158 if 'hdcore=' not in f4m_url:
3159 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
3160 f4m_formats = self._extract_f4m_formats(
3161 f4m_url, video_id, f4m_id='hds', fatal=False)
3162 for entry in f4m_formats:
3163 entry.update({'extra_param_to_segment_url': hdcore_sign})
3164 formats.extend(f4m_formats)
3166 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
3167 hls_host = hosts.get('hls')
3169 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
3170 m3u8_formats, m3u8_subtitles = self._extract_m3u8_formats_and_subtitles(
3171 m3u8_url, video_id, 'mp4', 'm3u8_native',
3172 m3u8_id='hls', fatal=False)
3173 formats.extend(m3u8_formats)
3174 subtitles = self._merge_subtitles(subtitles, m3u8_subtitles)
3176 http_host = hosts.get('http')
3177 if http_host and m3u8_formats and not signed:
3178 REPL_REGEX = r'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
3179 qualities = re.match(REPL_REGEX, m3u8_url).group(2).split(',')
3180 qualities_length = len(qualities)
3181 if len(m3u8_formats) in (qualities_length, qualities_length + 1):
3183 for f in m3u8_formats:
3184 if f['vcodec'] != 'none':
3185 for protocol in ('http', 'https'):
3187 del http_f['manifest_url']
3189 REPL_REGEX, protocol + r'://%s/\g<1>%s\3' % (http_host, qualities[i]), f['url'])
3191 'format_id': http_f['format_id'].replace('hls-', protocol + '-'),
3193 'protocol': protocol,
3195 formats.append(http_f)
3198 return formats, subtitles
3200 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
3201 query = compat_urlparse.urlparse(url).query
3202 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
3204 r'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url)
3205 url_base = mobj.group('url')
3206 http_base_url = '%s%s:%s' % ('http', mobj.group('s') or '', url_base)
3209 def manifest_url(manifest):
3210 m_url = '%s/%s' % (http_base_url, manifest)
3212 m_url += '?%s' % query
3215 if 'm3u8' not in skip_protocols:
3216 formats.extend(self._extract_m3u8_formats(
3217 manifest_url('playlist.m3u8'), video_id, 'mp4',
3218 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
3219 if 'f4m' not in skip_protocols:
3220 formats.extend(self._extract_f4m_formats(
3221 manifest_url('manifest.f4m'),
3222 video_id, f4m_id='hds', fatal=False))
3223 if 'dash' not in skip_protocols:
3224 formats.extend(self._extract_mpd_formats(
3225 manifest_url('manifest.mpd'),
3226 video_id, mpd_id='dash', fatal=False))
3227 if re.search(r'(?:/smil:|\.smil)', url_base):
3228 if 'smil' not in skip_protocols:
3229 rtmp_formats = self._extract_smil_formats(
3230 manifest_url('jwplayer.smil'),
3231 video_id, fatal=False)
3232 for rtmp_format in rtmp_formats:
3233 rtsp_format = rtmp_format.copy()
3234 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
3235 del rtsp_format['play_path']
3236 del rtsp_format['ext']
3237 rtsp_format.update({
3238 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
3239 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
3242 formats.extend([rtmp_format, rtsp_format])
3244 for protocol in ('rtmp', 'rtsp'):
3245 if protocol not in skip_protocols:
3247 'url': '%s:%s' % (protocol, url_base),
3248 'format_id': protocol,
3249 'protocol': protocol,
3253 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
3255 r'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)(?!</script>).*?\.setup\s*\((?P<options>[^)]+)\)',
3259 jwplayer_data = self._parse_json(mobj.group('options'),
3261 transform_source=transform_source)
3262 except ExtractorError:
3265 if isinstance(jwplayer_data, dict):
3266 return jwplayer_data
3268 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
3269 jwplayer_data = self._find_jwplayer_data(
3270 webpage, video_id, transform_source=js_to_json)
3271 return self._parse_jwplayer_data(
3272 jwplayer_data, video_id, *args, **kwargs)
3274 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
3275 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3276 # JWPlayer backward compatibility: flattened playlists
3277 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
3278 if 'playlist' not in jwplayer_data:
3279 jwplayer_data = {'playlist': [jwplayer_data]}
3283 # JWPlayer backward compatibility: single playlist item
3284 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
3285 if not isinstance(jwplayer_data['playlist'], list):
3286 jwplayer_data['playlist'] = [jwplayer_data['playlist']]
3288 for video_data in jwplayer_data['playlist']:
3289 # JWPlayer backward compatibility: flattened sources
3290 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
3291 if 'sources' not in video_data:
3292 video_data['sources'] = [video_data]
3294 this_video_id = video_id or video_data['mediaid']
3296 formats = self._parse_jwplayer_formats(
3297 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
3298 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
3301 tracks = video_data.get('tracks')
3302 if tracks and isinstance(tracks, list):
3303 for track in tracks:
3304 if not isinstance(track, dict):
3306 track_kind = track.get('kind')
3307 if not track_kind or not isinstance(track_kind, compat_str):
3309 if track_kind.lower() not in ('captions', 'subtitles'):
3311 track_url = urljoin(base_url, track.get('file'))
3314 subtitles.setdefault(track.get('label') or 'en', []).append({
3315 'url': self._proto_relative_url(track_url)
3319 'id': this_video_id,
3320 'title': unescapeHTML(video_data['title'] if require_title else video_data.get('title')),
3321 'description': clean_html(video_data.get('description')),
3322 'thumbnail': urljoin(base_url, self._proto_relative_url(video_data.get('image'))),
3323 'timestamp': int_or_none(video_data.get('pubdate')),
3324 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
3325 'subtitles': subtitles,
3327 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
3328 if len(formats) == 1 and re.search(r'^(?:http|//).*(?:youtube\.com|youtu\.be)/.+', formats[0]['url']):
3330 '_type': 'url_transparent',
3331 'url': formats[0]['url'],
3334 self._sort_formats(formats)
3335 entry['formats'] = formats
3336 entries.append(entry)
3337 if len(entries) == 1:
3340 return self.playlist_result(entries)
3342 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
3343 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3346 for source in jwplayer_sources_data:
3347 if not isinstance(source, dict):
3349 source_url = urljoin(
3350 base_url, self._proto_relative_url(source.get('file')))
3351 if not source_url or source_url in urls:
3353 urls.append(source_url)
3354 source_type = source.get('type') or ''
3355 ext = mimetype2ext(source_type) or determine_ext(source_url)
3356 if source_type == 'hls' or ext == 'm3u8':
3357 formats.extend(self._extract_m3u8_formats(
3358 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
3359 m3u8_id=m3u8_id, fatal=False))
3360 elif source_type == 'dash' or ext == 'mpd':
3361 formats.extend(self._extract_mpd_formats(
3362 source_url, video_id, mpd_id=mpd_id, fatal=False))
3364 formats.extend(self._extract_smil_formats(
3365 source_url, video_id, fatal=False))
3366 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3367 elif source_type.startswith('audio') or ext in (
3368 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
3375 height = int_or_none(source.get('height'))
3377 # Often no height is provided but there is a label in
3378 # format like "1080p", "720p SD", or 1080.
3379 height = int_or_none(self._search_regex(
3380 r'^(\d{3,4})[pP]?(?:\b|$)', compat_str(source.get('label') or ''),
3381 'height', default=None))
3384 'width': int_or_none(source.get('width')),
3386 'tbr': int_or_none(source.get('bitrate')),
3389 if source_url.startswith('rtmp'):
3390 a_format['ext'] = 'flv'
3391 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3392 # of jwplayer.flash.swf
3393 rtmp_url_parts = re.split(
3394 r'((?:mp4|mp3|flv):)', source_url, 1)
3395 if len(rtmp_url_parts) == 3:
3396 rtmp_url, prefix, play_path = rtmp_url_parts
3399 'play_path': prefix + play_path,
3402 a_format.update(rtmp_params)
3403 formats.append(a_format)
3406 def _live_title(self, name):
3407 """ Generate the title
for a live video
"""
3408 now = datetime.datetime.now()
3409 now_str = now.strftime('%Y-%m-%d %H:%M')
3410 return name + ' ' + now_str
3412 def _int(self, v, name, fatal=False, **kwargs):
3413 res = int_or_none(v, **kwargs)
3414 if 'get_attr' in kwargs:
3415 print(getattr(v, kwargs['get_attr']))
3417 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
3419 raise ExtractorError(msg)
3421 self.report_warning(msg)
3424 def _float(self, v, name, fatal=False, **kwargs):
3425 res = float_or_none(v, **kwargs)
3427 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
3429 raise ExtractorError(msg)
3431 self.report_warning(msg)
3434 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
3435 path='/', secure=False, discard=False, rest={}, **kwargs):
3436 cookie = compat_cookiejar_Cookie(
3437 0, name, value, port, port is not None, domain, True,
3438 domain.startswith('.'), path, True, secure, expire_time,
3439 discard, None, None, rest)
3440 self._downloader.cookiejar.set_cookie(cookie)
3442 def _get_cookies(self, url):
3443 """ Return a compat_cookies_SimpleCookie
with the cookies
for the url
"""
3444 req = sanitized_Request(url)
3445 self._downloader.cookiejar.add_cookie_header(req)
3446 return compat_cookies_SimpleCookie(req.get_header('Cookie'))
3448 def _apply_first_set_cookie_header(self, url_handle, cookie):
3450 Apply first Set
-Cookie header instead of the last
. Experimental
.
3452 Some
sites (e
.g
. [1-3]) may serve two cookies under the same name
3453 in Set
-Cookie header
and expect the
first (old
) one to be
set rather
3454 than
second (new
). However
, as of RFC6265 the newer one cookie
3455 should be
set into cookie store what actually happens
.
3456 We will workaround this issue by resetting the cookie to
3457 the first one manually
.
3458 1. https
://new
.vk
.com
/
3459 2. https
://github
.com
/ytdl
-org
/youtube
-dl
/issues
/9841#issuecomment-227871201
3460 3. https
://learning
.oreilly
.com
/
3462 for header, cookies in url_handle.headers.items():
3463 if header.lower() != 'set-cookie':
3465 if sys.version_info[0] >= 3:
3466 cookies = cookies.encode('iso-8859-1')
3467 cookies = cookies.decode('utf-8')
3468 cookie_value = re.search(
3469 r'%s=(.+?);.*?\b[Dd]omain=(.+?)(?:[,;]|$)' % cookie, cookies)
3471 value, domain = cookie_value.groups()
3472 self._set_cookie(domain, cookie, value)
3475 def get_testcases(self, include_onlymatching=False):
3476 t = getattr(self, '_TEST', None)
3478 assert not hasattr(self, '_TESTS'), \
3479 '%s has _TEST and _TESTS' % type(self).__name__
3482 tests = getattr(self, '_TESTS', [])
3484 if not include_onlymatching and t.get('only_matching', False):
3486 t['name'] = type(self).__name__[:-len('IE')]
3489 def is_suitable(self, age_limit):
3490 """ Test whether the extractor
is generally suitable
for the given
3491 age
limit (i
.e
. pornographic sites are
not, all others usually are
) """
3493 any_restricted = False
3494 for tc in self.get_testcases(include_onlymatching=False):
3495 if tc.get('playlist', []):
3496 tc = tc['playlist'][0]
3497 is_restricted = age_restricted(
3498 tc.get('info_dict', {}).get('age_limit'), age_limit)
3499 if not is_restricted:
3501 any_restricted = any_restricted or is_restricted
3502 return not any_restricted
3504 def extract_subtitles(self, *args, **kwargs):
3505 if (self.get_param('writesubtitles', False)
3506 or self.get_param('listsubtitles')):
3507 return self._get_subtitles(*args, **kwargs)
3510 def _get_subtitles(self, *args, **kwargs):
3511 raise NotImplementedError('This method must be implemented by subclasses')
3513 def extract_comments(self, *args, **kwargs):
3514 if not self.get_param('getcomments'):
3516 generator = self._get_comments(*args, **kwargs)
3522 comments.append(next(generator))
3523 except KeyboardInterrupt:
3525 self.to_screen('Interrupted by user')
3526 except StopIteration:
3528 comment_count = len(comments)
3529 self.to_screen(f'Extracted {comment_count} comments')
3531 'comments': comments,
3532 'comment_count': None if interrupted else comment_count
3536 def _get_comments(self, *args, **kwargs):
3537 raise NotImplementedError('This method must be implemented by subclasses')
3540 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
3541 """ Merge subtitle items
for one language
. Items
with duplicated URLs
3542 will be dropped
. """
3543 list1_urls = set([item['url'] for item in subtitle_list1])
3544 ret = list(subtitle_list1)
3545 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
3549 def _merge_subtitles(cls, *dicts, target=None):
3550 """ Merge subtitle dictionaries
, language by language
. """
3554 for lang, subs in d.items():
3555 target[lang] = cls._merge_subtitle_items(target.get(lang, []), subs)
3558 def extract_automatic_captions(self, *args, **kwargs):
3559 if (self.get_param('writeautomaticsub', False)
3560 or self.get_param('listsubtitles')):
3561 return self._get_automatic_captions(*args, **kwargs)
3564 def _get_automatic_captions(self, *args, **kwargs):
3565 raise NotImplementedError('This method must be implemented by subclasses')
3567 def mark_watched(self, *args, **kwargs):
3568 if not self.get_param('mark_watched', False):
3570 if (self._get_login_info()[0] is not None
3571 or self.get_param('cookiefile')
3572 or self.get_param('cookiesfrombrowser')):
3573 self._mark_watched(*args, **kwargs)
3575 def _mark_watched(self, *args, **kwargs):
3576 raise NotImplementedError('This method must be implemented by subclasses')
3578 def geo_verification_headers(self):
3580 geo_verification_proxy = self.get_param('geo_verification_proxy')
3581 if geo_verification_proxy:
3582 headers['Ytdl-request-proxy'] = geo_verification_proxy
3585 def _generic_id(self, url):
3586 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
3588 def _generic_title(self, url):
3589 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
3592 def _availability(is_private=None, needs_premium=None, needs_subscription=None, needs_auth=None, is_unlisted=None):
3593 all_known = all(map(
3594 lambda x: x is not None,
3595 (is_private, needs_premium, needs_subscription, needs_auth, is_unlisted)))
3597 'private' if is_private
3598 else 'premium_only' if needs_premium
3599 else 'subscriber_only' if needs_subscription
3600 else 'needs_auth' if needs_auth
3601 else 'unlisted' if is_unlisted
3602 else 'public' if all_known
3605 def _configuration_arg(self, key, default=NO_DEFAULT, casesense=False):
3607 @returns A list of values for the extractor argument given by "key"
3608 or "default" if no such key is present
3609 @param default The default value to return when the key is not present (default: [])
3610 @param casesense When false, the values are converted to lower case
3613 self._downloader.params, ('extractor_args', self.ie_key().lower(), key))
3615 return [] if default is NO_DEFAULT else default
3616 return list(val) if casesense else [x.lower() for x in val]
3619 class SearchInfoExtractor(InfoExtractor):
3621 Base
class for paged search queries extractors
.
3622 They accept URLs
in the format
_SEARCH_KEY(|all|
[0-9]):{query}
3623 Instances should define _SEARCH_KEY
and optionally _MAX_RESULTS
3626 _MAX_RESULTS = float('inf')
3629 def _make_valid_url(cls):
3630 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
3633 def suitable(cls, url):
3634 return re.match(cls._make_valid_url(), url) is not None
3636 def _real_extract(self, query):
3637 mobj = re.match(self._make_valid_url(), query)
3639 raise ExtractorError('Invalid search query "%s"' % query)
3641 prefix = mobj.group('prefix')
3642 query = mobj.group('query')
3644 return self._get_n_results(query, 1)
3645 elif prefix == 'all':
3646 return self._get_n_results(query, self._MAX_RESULTS)
3650 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
3651 elif n > self._MAX_RESULTS:
3652 self.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3653 n = self._MAX_RESULTS
3654 return self._get_n_results(query, n)
3656 def _get_n_results(self, query, n):
3657 """Get a specified number of results
for a query
.
3658 Either this function
or _search_results must be overridden by subclasses
"""
3659 return self.playlist_result(
3660 itertools.islice(self._search_results(query), 0, None if n == float('inf') else n),
3663 def _search_results(self, query):
3664 """Returns an iterator of search results
"""
3665 raise NotImplementedError('This method must be implemented by subclasses')
3668 def SEARCH_KEY(self):
3669 return self._SEARCH_KEY