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
,
68 parse_m3u8_attributes
,
93 class InfoExtractor(object):
94 """Information Extractor class.
96 Information extractors are the classes that, given a URL, extract
97 information about the video (or videos) the URL refers to. This
98 information includes the real video URL, the video title, author and
99 others. The information is stored in a dictionary which is then
100 passed to the YoutubeDL. The YoutubeDL processes this
101 information possibly downloading the video to the file system, among
102 other possible outcomes.
104 The type field determines the type of the result.
105 By far the most common value (and the default if _type is missing) is
106 "video", which indicates a single video.
108 For a video, the dictionaries must include the following fields:
110 id: Video identifier.
111 title: Video title, unescaped.
113 Additionally, it must contain either a formats entry or a url one:
115 formats: A list of dictionaries for each format available, ordered
116 from worst to best quality.
119 * url The mandatory URL representing the media:
120 for plain file media - HTTP URL of this file,
122 for HLS - URL of the M3U8 media playlist,
123 for HDS - URL of the F4M manifest,
125 - HTTP URL to plain file media (in case of
127 - URL of the MPD manifest or base URL
128 representing the media if MPD manifest
129 is parsed from a string (in case of
131 for MSS - URL of the ISM manifest.
133 The URL of the manifest file in case of
135 for HLS - URL of the M3U8 master playlist,
136 for HDS - URL of the F4M manifest,
137 for DASH - URL of the MPD manifest,
138 for MSS - URL of the ISM manifest.
139 * ext Will be calculated from URL if missing
140 * format A human-readable description of the format
141 ("mp4 container with h264/opus").
142 Calculated from the format_id, width, height.
143 and format_note fields if missing.
144 * format_id A short description of the format
145 ("mp4_h264_opus" or "19").
146 Technically optional, but strongly recommended.
147 * format_note Additional info about the format
148 ("3D" or "DASH video")
149 * width Width of the video, if known
150 * height Height of the video, if known
151 * resolution Textual description of width and height
152 * dynamic_range The dynamic range of the video. One of:
153 "SDR" (None), "HDR10", "HDR10+, "HDR12", "HLG, "DV"
154 * tbr Average bitrate of audio and video in KBit/s
155 * abr Average audio bitrate in KBit/s
156 * acodec Name of the audio codec in use
157 * asr Audio sampling rate in Hertz
158 * vbr Average video bitrate in KBit/s
160 * vcodec Name of the video codec in use
161 * container Name of the container format
162 * filesize The number of bytes, if known in advance
163 * filesize_approx An estimate for the number of bytes
164 * player_url SWF Player URL (used for rtmpdump).
165 * protocol The protocol that will be used for the actual
166 download, lower-case.
167 "http", "https", "rtsp", "rtmp", "rtmp_ffmpeg", "rtmpe",
168 "m3u8", "m3u8_native" or "http_dash_segments".
170 Base URL for fragments. Each fragment's path
171 value (if present) will be relative to
173 * fragments A list of fragments of a fragmented media.
174 Each fragment entry must contain either an url
175 or a path. If an url is present it should be
176 considered by a client. Otherwise both path and
177 fragment_base_url must be present. Here is
178 the list of all potential fields:
179 * "url" - fragment's URL
180 * "path" - fragment's path relative to
182 * "duration" (optional, int or float)
183 * "filesize" (optional, int)
184 * preference Order number of this format. If this field is
185 present and not None, the formats get sorted
186 by this field, regardless of all other values.
187 -1 for default (order by other properties),
188 -2 or smaller for less than default.
189 < -1000 to hide the format (if there is
190 another one which is strictly better)
191 * language Language code, e.g. "de" or "en-US".
192 * language_preference Is this in the language mentioned in
194 10 if it's what the URL is about,
195 -1 for default (don't know),
196 -10 otherwise, other values reserved for now.
197 * quality Order number of the video quality of this
198 format, irrespective of the file format.
199 -1 for default (order by other properties),
200 -2 or smaller for less than default.
201 * source_preference Order number for this video source
202 (quality takes higher priority)
203 -1 for default (order by other properties),
204 -2 or smaller for less than default.
205 * http_headers A dictionary of additional HTTP headers
206 to add to the request.
207 * stretched_ratio If given and not 1, indicates that the
208 video's pixels are not square.
209 width : height ratio as float.
210 * no_resume The server does not support resuming the
211 (HTTP or RTMP) download. Boolean.
212 * has_drm The format has DRM and cannot be downloaded. Boolean
213 * downloader_options A dictionary of downloader options as
214 described in FileDownloader
215 RTMP formats can also have the additional fields: page_url,
216 app, play_path, tc_url, flash_version, rtmp_live, rtmp_conn,
217 rtmp_protocol, rtmp_real_time
219 url: Final video URL.
220 ext: Video filename extension.
221 format: The video format, defaults to ext (used for --get-format)
222 player_url: SWF Player URL (used for rtmpdump).
224 The following fields are optional:
226 alt_title: A secondary title of the video.
227 display_id An alternative identifier for the video, not necessarily
228 unique, but available before title. Typically, id is
229 something like "4234987", title "Dancing naked mole rats",
230 and display_id "dancing-naked-mole-rats"
231 thumbnails: A list of dictionaries, with the following entries:
232 * "id" (optional, string) - Thumbnail format ID
234 * "preference" (optional, int) - quality of the image
235 * "width" (optional, int)
236 * "height" (optional, int)
237 * "resolution" (optional, string "{width}x{height}",
239 * "filesize" (optional, int)
240 thumbnail: Full URL to a video thumbnail image.
241 description: Full video description.
242 uploader: Full name of the video uploader.
243 license: License name the video is licensed under.
244 creator: The creator of the video.
245 release_timestamp: UNIX timestamp of the moment the video was released.
246 release_date: The date (YYYYMMDD) when the video was released.
247 timestamp: UNIX timestamp of the moment the video was uploaded
248 upload_date: Video upload date (YYYYMMDD).
249 If not explicitly set, calculated from timestamp.
250 uploader_id: Nickname or id of the video uploader.
251 uploader_url: Full URL to a personal webpage of the video uploader.
252 channel: Full name of the channel the video is uploaded on.
253 Note that channel fields may or may not repeat uploader
254 fields. This depends on a particular extractor.
255 channel_id: Id of the channel.
256 channel_url: Full URL to a channel webpage.
257 location: Physical location where the video was filmed.
258 subtitles: The available subtitles as a dictionary in the format
259 {tag: subformats}. "tag" is usually a language code, and
260 "subformats" is a list sorted from lower to higher
261 preference, each element is a dictionary with the "ext"
263 * "data": The subtitles file contents
264 * "url": A URL pointing to the subtitles file
265 It can optionally also have:
266 * "name": Name or description of the subtitles
267 "ext" will be calculated from URL if missing
268 automatic_captions: Like 'subtitles'; contains automatically generated
269 captions instead of normal subtitles
270 duration: Length of the video in seconds, as an integer or float.
271 view_count: How many users have watched the video on the platform.
272 like_count: Number of positive ratings of the video
273 dislike_count: Number of negative ratings of the video
274 repost_count: Number of reposts of the video
275 average_rating: Average rating give by users, the scale used depends on the webpage
276 comment_count: Number of comments on the video
277 comments: A list of comments, each with one or more of the following
278 properties (all but one of text or html optional):
279 * "author" - human-readable name of the comment author
280 * "author_id" - user ID of the comment author
281 * "author_thumbnail" - The thumbnail of the comment author
283 * "html" - Comment as HTML
284 * "text" - Plain text of the comment
285 * "timestamp" - UNIX timestamp of comment
286 * "parent" - ID of the comment this one is replying to.
287 Set to "root" to indicate that this is a
288 comment to the original video.
289 * "like_count" - Number of positive ratings of the comment
290 * "dislike_count" - Number of negative ratings of the comment
291 * "is_favorited" - Whether the comment is marked as
292 favorite by the video uploader
293 * "author_is_uploader" - Whether the comment is made by
295 age_limit: Age restriction for the video, as an integer (years)
296 webpage_url: The URL to the video webpage, if given to yt-dlp it
297 should allow to get the same result again. (It will be set
298 by YoutubeDL if it's missing)
299 categories: A list of categories that the video falls in, for example
301 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
302 cast: A list of the video cast
303 is_live: True, False, or None (=unknown). Whether this video is a
304 live stream that goes on instead of a fixed-length video.
305 was_live: True, False, or None (=unknown). Whether this video was
306 originally a live stream.
307 live_status: 'is_live', 'is_upcoming', 'was_live', 'not_live' or None (=unknown)
308 If absent, automatically set from is_live, was_live
309 start_time: Time in seconds where the reproduction should start, as
310 specified in the URL.
311 end_time: Time in seconds where the reproduction should end, as
312 specified in the URL.
313 chapters: A list of dictionaries, with the following entries:
314 * "start_time" - The start time of the chapter in seconds
315 * "end_time" - The end time of the chapter in seconds
316 * "title" (optional, string)
317 playable_in_embed: Whether this video is allowed to play in embedded
318 players on other sites. Can be True (=always allowed),
319 False (=never allowed), None (=unknown), or a string
320 specifying the criteria for embedability (Eg: 'whitelist')
321 availability: Under what condition the video is available. One of
322 'private', 'premium_only', 'subscriber_only', 'needs_auth',
323 'unlisted' or 'public'. Use 'InfoExtractor._availability'
325 __post_extractor: A function to be called just before the metadata is
326 written to either disk, logger or console. The function
327 must return a dict which will be added to the info_dict.
328 This is usefull for additional information that is
329 time-consuming to extract. Note that the fields thus
330 extracted will not be available to output template and
331 match_filter. So, only "comments" and "comment_count" are
332 currently allowed to be extracted via this method.
334 The following fields should only be used when the video belongs to some logical
337 chapter: Name or title of the chapter the video belongs to.
338 chapter_number: Number of the chapter the video belongs to, as an integer.
339 chapter_id: Id of the chapter the video belongs to, as a unicode string.
341 The following fields should only be used when the video is an episode of some
342 series, programme or podcast:
344 series: Title of the series or programme the video episode belongs to.
345 season: Title of the season the video episode belongs to.
346 season_number: Number of the season the video episode belongs to, as an integer.
347 season_id: Id of the season the video episode belongs to, as a unicode string.
348 episode: Title of the video episode. Unlike mandatory video title field,
349 this field should denote the exact title of the video episode
350 without any kind of decoration.
351 episode_number: Number of the video episode within a season, as an integer.
352 episode_id: Id of the video episode, as a unicode string.
354 The following fields should only be used when the media is a track or a part of
357 track: Title of the track.
358 track_number: Number of the track within an album or a disc, as an integer.
359 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
361 artist: Artist(s) of the track.
362 genre: Genre(s) of the track.
363 album: Title of the album the track belongs to.
364 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
365 album_artist: List of all artists appeared on the album (e.g.
366 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
368 disc_number: Number of the disc or other physical medium the track belongs to,
370 release_year: Year (YYYY) when the album was released.
372 Unless mentioned otherwise, the fields should be Unicode strings.
374 Unless mentioned otherwise, None is equivalent to absence of information.
377 _type "playlist" indicates multiple videos.
378 There must be a key "entries", which is a list, an iterable, or a PagedList
379 object, each element of which is a valid dictionary by this specification.
381 Additionally, playlists can have "id", "title", and any other relevent
382 attributes with the same semantics as videos (see above).
385 _type "multi_video" indicates that there are multiple videos that
386 form a single show, for examples multiple acts of an opera or TV episode.
387 It must have an entries key like a playlist and contain all the keys
388 required for a video at the same time.
391 _type "url" indicates that the video must be extracted from another
392 location, possibly by a different extractor. Its only required key is:
393 "url" - the next URL to extract.
394 The key "ie_key" can be set to the class name (minus the trailing "IE",
395 e.g. "Youtube") if the extractor class is known in advance.
396 Additionally, the dictionary may have any properties of the resolved entity
397 known in advance, for example "title" if the title of the referred video is
401 _type "url_transparent" entities have the same specification as "url", but
402 indicate that the given additional information is more precise than the one
403 associated with the resolved URL.
404 This is useful when a site employs a video service that hosts the video and
405 its technical metadata, but that video service does not embed a useful
406 title, description etc.
409 Subclasses of this one should re-define the _real_initialize() and
410 _real_extract() methods and define a _VALID_URL regexp.
411 Probably, they should also be added to the list of extractors.
413 Subclasses may also override suitable() if necessary, but ensure the function
414 signature is preserved and that this function imports everything it needs
415 (except other extractors), so that lazy_extractors works correctly
417 _GEO_BYPASS attribute may be set to False in order to disable
418 geo restriction bypass mechanisms for a particular extractor.
419 Though it won't disable explicit geo restriction bypass based on
420 country code provided with geo_bypass_country.
422 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
423 countries for this extractor. One of these countries will be used by
424 geo restriction bypass mechanism right away in order to bypass
425 geo restriction, of course, if the mechanism is not disabled.
427 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
428 IP blocks in CIDR notation for this extractor. One of these IP blocks
429 will be used by geo restriction bypass mechanism similarly
432 The _WORKING attribute should be set to False for broken IEs
433 in order to warn the users and skip the tests.
438 _x_forwarded_for_ip
= None
440 _GEO_COUNTRIES
= None
441 _GEO_IP_BLOCKS
= None
445 'any': 'Use --cookies, --username and --password, or --netrc to provide account credentials',
447 'Use --cookies-from-browser or --cookies for the authentication. '
448 'See https://github.com/ytdl-org/youtube-dl#how-do-i-pass-cookies-to-youtube-dl for how to manually pass cookies'),
449 'password': 'Use --username and --password, or --netrc to provide account credentials',
452 def __init__(self
, downloader
=None):
453 """Constructor. Receives an optional downloader (a YoutubeDL instance).
454 If a downloader is not passed during initialization,
455 it must be set using "set_downloader()" before "extract()" is called"""
457 self
._x
_forwarded
_for
_ip
= None
458 self
._printed
_messages
= set()
459 self
.set_downloader(downloader
)
462 def _match_valid_url(cls
, url
):
463 # This does not use has/getattr intentionally - we want to know whether
464 # we have cached the regexp for *this* class, whereas getattr would also
465 # match the superclass
466 if '_VALID_URL_RE' not in cls
.__dict
__:
467 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
468 return cls
._VALID
_URL
_RE
.match(url
)
471 def suitable(cls
, url
):
472 """Receives a URL and returns True if suitable for this IE."""
473 # This function must import everything it needs (except other extractors),
474 # so that lazy_extractors works correctly
475 return cls
._match
_valid
_url
(url
) is not None
478 def _match_id(cls
, url
):
479 return cls
._match
_valid
_url
(url
).group('id')
482 def get_temp_id(cls
, url
):
484 return cls
._match
_id
(url
)
485 except (IndexError, AttributeError):
490 """Getter method for _WORKING."""
493 def initialize(self
):
494 """Initializes an instance (authentication, etc)."""
495 self
._printed
_messages
= set()
496 self
._initialize
_geo
_bypass
({
497 'countries': self
._GEO
_COUNTRIES
,
498 'ip_blocks': self
._GEO
_IP
_BLOCKS
,
501 self
._real
_initialize
()
504 def _initialize_geo_bypass(self
, geo_bypass_context
):
506 Initialize geo restriction bypass mechanism.
508 This method is used to initialize geo bypass mechanism based on faking
509 X-Forwarded-For HTTP header. A random country from provided country list
510 is selected and a random IP belonging to this country is generated. This
511 IP will be passed as X-Forwarded-For HTTP header in all subsequent
514 This method will be used for initial geo bypass mechanism initialization
515 during the instance initialization with _GEO_COUNTRIES and
518 You may also manually call it from extractor's code if geo bypass
519 information is not available beforehand (e.g. obtained during
520 extraction) or due to some other reason. In this case you should pass
521 this information in geo bypass context passed as first argument. It may
522 contain following fields:
524 countries: List of geo unrestricted countries (similar
526 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
527 (similar to _GEO_IP_BLOCKS)
530 if not self
._x
_forwarded
_for
_ip
:
532 # Geo bypass mechanism is explicitly disabled by user
533 if not self
.get_param('geo_bypass', True):
536 if not geo_bypass_context
:
537 geo_bypass_context
= {}
539 # Backward compatibility: previously _initialize_geo_bypass
540 # expected a list of countries, some 3rd party code may still use
542 if isinstance(geo_bypass_context
, (list, tuple)):
543 geo_bypass_context
= {
544 'countries': geo_bypass_context
,
547 # The whole point of geo bypass mechanism is to fake IP
548 # as X-Forwarded-For HTTP header based on some IP block or
551 # Path 1: bypassing based on IP block in CIDR notation
553 # Explicit IP block specified by user, use it right away
554 # regardless of whether extractor is geo bypassable or not
555 ip_block
= self
.get_param('geo_bypass_ip_block', None)
557 # Otherwise use random IP block from geo bypass context but only
558 # if extractor is known as geo bypassable
560 ip_blocks
= geo_bypass_context
.get('ip_blocks')
561 if self
._GEO
_BYPASS
and ip_blocks
:
562 ip_block
= random
.choice(ip_blocks
)
565 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(ip_block
)
566 self
._downloader
.write_debug(
567 '[debug] Using fake IP %s as X-Forwarded-For' % self
._x
_forwarded
_for
_ip
)
570 # Path 2: bypassing based on country code
572 # Explicit country code specified by user, use it right away
573 # regardless of whether extractor is geo bypassable or not
574 country
= self
.get_param('geo_bypass_country', None)
576 # Otherwise use random country code from geo bypass context but
577 # only if extractor is known as geo bypassable
579 countries
= geo_bypass_context
.get('countries')
580 if self
._GEO
_BYPASS
and countries
:
581 country
= random
.choice(countries
)
584 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country
)
585 self
._downloader
.write_debug(
586 'Using fake IP %s (%s) as X-Forwarded-For' % (self
._x
_forwarded
_for
_ip
, country
.upper()))
588 def extract(self
, url
):
589 """Extracts URL information and returns it in list of dicts."""
594 self
.write_debug('Extracting URL: %s' % url
)
595 ie_result
= self
._real
_extract
(url
)
596 if ie_result
is None:
598 if self
._x
_forwarded
_for
_ip
:
599 ie_result
['__x_forwarded_for_ip'] = self
._x
_forwarded
_for
_ip
600 subtitles
= ie_result
.get('subtitles')
601 if (subtitles
and 'live_chat' in subtitles
602 and 'no-live-chat' in self
.get_param('compat_opts', [])):
603 del subtitles
['live_chat']
605 except GeoRestrictedError
as e
:
606 if self
.__maybe
_fake
_ip
_and
_retry
(e
.countries
):
609 except UnsupportedError
:
611 except ExtractorError
as e
:
613 'video_id': e
.video_id
or self
.get_temp_id(url
),
616 'expected': e
.expected
,
619 if hasattr(e
, 'countries'):
620 kwargs
['countries'] = e
.countries
621 raise type(e
)(e
.msg
, **kwargs
)
622 except compat_http_client
.IncompleteRead
as e
:
623 raise ExtractorError('A network error has occurred.', cause
=e
, expected
=True, video_id
=self
.get_temp_id(url
))
624 except (KeyError, StopIteration) as e
:
625 raise ExtractorError('An extractor error has occurred.', cause
=e
, video_id
=self
.get_temp_id(url
))
627 def __maybe_fake_ip_and_retry(self
, countries
):
628 if (not self
.get_param('geo_bypass_country', None)
630 and self
.get_param('geo_bypass', True)
631 and not self
._x
_forwarded
_for
_ip
633 country_code
= random
.choice(countries
)
634 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country_code
)
635 if self
._x
_forwarded
_for
_ip
:
637 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
638 % (self
._x
_forwarded
_for
_ip
, country_code
.upper()))
642 def set_downloader(self
, downloader
):
643 """Sets the downloader for this IE."""
644 self
._downloader
= downloader
646 def _real_initialize(self
):
647 """Real initialization process. Redefine in subclasses."""
650 def _real_extract(self
, url
):
651 """Real extraction process. Redefine in subclasses."""
656 """A string for getting the InfoExtractor with get_info_extractor"""
657 return cls
.__name
__[:-2]
661 return compat_str(type(self
).__name
__[:-2])
664 def __can_accept_status_code(err
, expected_status
):
665 assert isinstance(err
, compat_urllib_error
.HTTPError
)
666 if expected_status
is None:
668 elif callable(expected_status
):
669 return expected_status(err
.code
) is True
671 return err
.code
in variadic(expected_status
)
673 def _request_webpage(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
, expected_status
=None):
675 Return the response handle.
677 See _download_webpage docstring for arguments specification.
679 if not self
._downloader
._first
_webpage
_request
:
680 sleep_interval
= self
.get_param('sleep_interval_requests') or 0
681 if sleep_interval
> 0:
682 self
.to_screen('Sleeping %s seconds ...' % sleep_interval
)
683 time
.sleep(sleep_interval
)
685 self
._downloader
._first
_webpage
_request
= False
688 self
.report_download_webpage(video_id
)
689 elif note
is not False:
691 self
.to_screen('%s' % (note
,))
693 self
.to_screen('%s: %s' % (video_id
, note
))
695 # Some sites check X-Forwarded-For HTTP header in order to figure out
696 # the origin of the client behind proxy. This allows bypassing geo
697 # restriction by faking this header's value to IP that belongs to some
698 # geo unrestricted country. We will do so once we encounter any
699 # geo restriction error.
700 if self
._x
_forwarded
_for
_ip
:
701 if 'X-Forwarded-For' not in headers
:
702 headers
['X-Forwarded-For'] = self
._x
_forwarded
_for
_ip
704 if isinstance(url_or_request
, compat_urllib_request
.Request
):
705 url_or_request
= update_Request(
706 url_or_request
, data
=data
, headers
=headers
, query
=query
)
709 url_or_request
= update_url_query(url_or_request
, query
)
710 if data
is not None or headers
:
711 url_or_request
= sanitized_Request(url_or_request
, data
, headers
)
713 return self
._downloader
.urlopen(url_or_request
)
714 except network_exceptions
as err
:
715 if isinstance(err
, compat_urllib_error
.HTTPError
):
716 if self
.__can
_accept
_status
_code
(err
, expected_status
):
717 # Retain reference to error to prevent file object from
718 # being closed before it can be read. Works around the
719 # effects of <https://bugs.python.org/issue15002>
720 # introduced in Python 3.4.1.
727 errnote
= 'Unable to download webpage'
729 errmsg
= '%s: %s' % (errnote
, error_to_compat_str(err
))
731 raise ExtractorError(errmsg
, sys
.exc_info()[2], cause
=err
)
733 self
.report_warning(errmsg
)
736 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):
738 Return a tuple (page content as string, URL handle).
740 See _download_webpage docstring for arguments specification.
742 # Strip hashes from the URL (#1038)
743 if isinstance(url_or_request
, (compat_str
, str)):
744 url_or_request
= url_or_request
.partition('#')[0]
746 urlh
= self
._request
_webpage
(url_or_request
, video_id
, note
, errnote
, fatal
, data
=data
, headers
=headers
, query
=query
, expected_status
=expected_status
)
750 content
= self
._webpage
_read
_content
(urlh
, url_or_request
, video_id
, note
, errnote
, fatal
, encoding
=encoding
)
751 return (content
, urlh
)
754 def _guess_encoding_from_content(content_type
, webpage_bytes
):
755 m
= re
.match(r
'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type
)
757 encoding
= m
.group(1)
759 m
= re
.search(br
'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
760 webpage_bytes[:1024])
762 encoding = m.group(1).decode('ascii')
763 elif webpage_bytes.startswith(b'\xff\xfe'):
770 def __check_blocked(self, content):
771 first_block = content[:512]
772 if ('<title>Access to this site is blocked</title>' in content
773 and 'Websense' in first_block):
774 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
775 blocked_iframe = self._html_search_regex(
776 r'<iframe src="([^
"]+)"', content,
777 'Websense information URL
', default=None)
779 msg += ' Visit
%s for more details
' % blocked_iframe
780 raise ExtractorError(msg, expected=True)
781 if '<title
>The URL you requested has been blocked
</title
>' in first_block:
783 'Access to this webpage has been blocked by Indian censorship
. '
784 'Use a VPN
or proxy
server (with --proxy
) to route around it
.')
785 block_msg = self._html_search_regex(
786 r'</h1
><p
>(.*?
)</p
>',
787 content, 'block message
', default=None)
789 msg += ' (Message
: "%s")' % block_msg.replace('\n', ' ')
790 raise ExtractorError(msg, expected=True)
791 if ('<title
>TTK
:: Доступ к ресурсу ограничен
</title
>' in content
792 and 'blocklist
.rkn
.gov
.ru
' in content):
793 raise ExtractorError(
794 'Access to this webpage has been blocked by decision of the Russian government
. '
795 'Visit http
://blocklist
.rkn
.gov
.ru
/ for a block reason
.',
798 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
799 content_type = urlh.headers.get('Content
-Type
', '')
800 webpage_bytes = urlh.read()
801 if prefix is not None:
802 webpage_bytes = prefix + webpage_bytes
804 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
805 if self.get_param('dump_intermediate_pages
', False):
806 self.to_screen('Dumping request to
' + urlh.geturl())
807 dump = base64.b64encode(webpage_bytes).decode('ascii
')
808 self._downloader.to_screen(dump)
809 if self.get_param('write_pages
', False):
810 basen = '%s_%s' % (video_id, urlh.geturl())
811 trim_length = self.get_param('trim_file_name
') or 240
812 if len(basen) > trim_length:
813 h = '___
' + hashlib.md5(basen.encode('utf
-8')).hexdigest()
814 basen = basen[:trim_length - len(h)] + h
815 raw_filename = basen + '.dump
'
816 filename = sanitize_filename(raw_filename, restricted=True)
817 self.to_screen('Saving request to
' + filename)
818 # Working around MAX_PATH limitation on Windows (see
819 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
820 if compat_os_name == 'nt
':
821 absfilepath = os.path.abspath(filename)
822 if len(absfilepath) > 259:
823 filename = '\\\\?
\\' + absfilepath
824 with open(filename, 'wb
') as outf:
825 outf.write(webpage_bytes)
828 content = webpage_bytes.decode(encoding, 'replace
')
830 content = webpage_bytes.decode('utf
-8', 'replace
')
832 self.__check_blocked(content)
836 def _download_webpage(
837 self, url_or_request, video_id, note=None, errnote=None,
838 fatal=True, tries=1, timeout=5, encoding=None, data=None,
839 headers={}, query={}, expected_status=None):
841 Return the data of the page as a string.
844 url_or_request -- plain text URL as a string or
845 a compat_urllib_request.Requestobject
846 video_id -- Video/playlist/item identifier (string)
849 note -- note printed before downloading (string)
850 errnote -- note printed in case of an error (string)
851 fatal -- flag denoting whether error should be considered fatal,
852 i.e. whether it should cause ExtractionError to be raised,
853 otherwise a warning will be reported and extraction continued
854 tries -- number of tries
855 timeout -- sleep interval between tries
856 encoding -- encoding for a page content decoding, guessed automatically
857 when not explicitly specified
858 data -- POST data (bytes)
859 headers -- HTTP headers (dict)
860 query -- URL query (dict)
861 expected_status -- allows to accept failed HTTP requests (non 2xx
862 status code) by explicitly specifying a set of accepted status
863 codes. Can be any of the following entities:
864 - an integer type specifying an exact failed status code to
866 - a list or a tuple of integer types specifying a list of
867 failed status codes to accept
868 - a callable accepting an actual failed status code and
869 returning True if it should be accepted
870 Note that this argument does not affect success status codes (2xx)
871 which are always accepted.
876 while success is False:
878 res = self._download_webpage_handle(
879 url_or_request, video_id, note, errnote, fatal,
880 encoding=encoding, data=data, headers=headers, query=query,
881 expected_status=expected_status)
883 except compat_http_client.IncompleteRead as e:
885 if try_count >= tries:
887 self._sleep(timeout, video_id)
894 def _download_xml_handle(
895 self, url_or_request, video_id, note='Downloading XML
',
896 errnote='Unable to download XML
', transform_source=None,
897 fatal=True, encoding=None, data=None, headers={}, query={},
898 expected_status=None):
900 Return a tuple (xml as an compat_etree_Element, URL handle).
902 See _download_webpage docstring for arguments specification.
904 res = self._download_webpage_handle(
905 url_or_request, video_id, note, errnote, fatal=fatal,
906 encoding=encoding, data=data, headers=headers, query=query,
907 expected_status=expected_status)
910 xml_string, urlh = res
911 return self._parse_xml(
912 xml_string, video_id, transform_source=transform_source,
916 self, url_or_request, video_id,
917 note='Downloading XML
', errnote='Unable to download XML
',
918 transform_source=None, fatal=True, encoding=None,
919 data=None, headers={}, query={}, expected_status=None):
921 Return the xml as an compat_etree_Element.
923 See _download_webpage docstring for arguments specification.
925 res = self._download_xml_handle(
926 url_or_request, video_id, note=note, errnote=errnote,
927 transform_source=transform_source, fatal=fatal, encoding=encoding,
928 data=data, headers=headers, query=query,
929 expected_status=expected_status)
930 return res if res is False else res[0]
932 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
934 xml_string = transform_source(xml_string)
936 return compat_etree_fromstring(xml_string.encode('utf
-8'))
937 except compat_xml_parse_error as ve:
938 errmsg = '%s: Failed to parse XML
' % video_id
940 raise ExtractorError(errmsg, cause=ve)
942 self.report_warning(errmsg + str(ve))
944 def _download_json_handle(
945 self, url_or_request, video_id, note='Downloading JSON metadata
',
946 errnote='Unable to download JSON metadata
', transform_source=None,
947 fatal=True, encoding=None, data=None, headers={}, query={},
948 expected_status=None):
950 Return a tuple (JSON object, URL handle).
952 See _download_webpage docstring for arguments specification.
954 res = self._download_webpage_handle(
955 url_or_request, video_id, note, errnote, fatal=fatal,
956 encoding=encoding, data=data, headers=headers, query=query,
957 expected_status=expected_status)
960 json_string, urlh = res
961 return self._parse_json(
962 json_string, video_id, transform_source=transform_source,
966 self, url_or_request, video_id, note='Downloading JSON metadata
',
967 errnote='Unable to download JSON metadata
', transform_source=None,
968 fatal=True, encoding=None, data=None, headers={}, query={},
969 expected_status=None):
971 Return the JSON object as a dict.
973 See _download_webpage docstring for arguments specification.
975 res = self._download_json_handle(
976 url_or_request, video_id, note=note, errnote=errnote,
977 transform_source=transform_source, fatal=fatal, encoding=encoding,
978 data=data, headers=headers, query=query,
979 expected_status=expected_status)
980 return res if res is False else res[0]
982 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
984 json_string = transform_source(json_string)
986 return json.loads(json_string)
987 except ValueError as ve:
988 errmsg = '%s: Failed to parse JSON
' % video_id
990 raise ExtractorError(errmsg, cause=ve)
992 self.report_warning(errmsg + str(ve))
994 def _parse_socket_response_as_json(self, data, video_id, transform_source=None, fatal=True):
995 return self._parse_json(
996 data[data.find('{'):data.rfind('}
') + 1],
997 video_id, transform_source, fatal)
999 def _download_socket_json_handle(
1000 self, url_or_request, video_id, note='Polling socket
',
1001 errnote='Unable to poll socket
', transform_source=None,
1002 fatal=True, encoding=None, data=None, headers={}, query={},
1003 expected_status=None):
1005 Return a tuple (JSON object, URL handle).
1007 See _download_webpage docstring for arguments specification.
1009 res = self._download_webpage_handle(
1010 url_or_request, video_id, note, errnote, fatal=fatal,
1011 encoding=encoding, data=data, headers=headers, query=query,
1012 expected_status=expected_status)
1016 return self._parse_socket_response_as_json(
1017 webpage, video_id, transform_source=transform_source,
1020 def _download_socket_json(
1021 self, url_or_request, video_id, note='Polling socket
',
1022 errnote='Unable to poll socket
', transform_source=None,
1023 fatal=True, encoding=None, data=None, headers={}, query={},
1024 expected_status=None):
1026 Return the JSON object as a dict.
1028 See _download_webpage docstring for arguments specification.
1030 res = self._download_socket_json_handle(
1031 url_or_request, video_id, note=note, errnote=errnote,
1032 transform_source=transform_source, fatal=fatal, encoding=encoding,
1033 data=data, headers=headers, query=query,
1034 expected_status=expected_status)
1035 return res if res is False else res[0]
1037 def report_warning(self, msg, video_id=None, *args, only_once=False, **kwargs):
1038 idstr = format_field(video_id, template='%s: ')
1039 msg = f'[{self.IE_NAME}
] {idstr}{msg}
'
1041 if f'WARNING
: {msg}
' in self._printed_messages:
1043 self._printed_messages.add(f'WARNING
: {msg}
')
1044 self._downloader.report_warning(msg, *args, **kwargs)
1046 def to_screen(self, msg, *args, **kwargs):
1047 """Print msg to screen, prefixing it with '[ie_name
]'"""
1048 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg), *args, **kwargs)
1050 def write_debug(self, msg, *args, **kwargs):
1051 self._downloader.write_debug('[%s] %s' % (self.IE_NAME, msg), *args, **kwargs)
1053 def get_param(self, name, default=None, *args, **kwargs):
1054 if self._downloader:
1055 return self._downloader.params.get(name, default, *args, **kwargs)
1058 def report_drm(self, video_id, partial=False):
1059 self.raise_no_formats('This video
is DRM protected
', expected=True, video_id=video_id)
1061 def report_extraction(self, id_or_name):
1062 """Report information extraction."""
1063 self.to_screen('%s: Extracting information
' % id_or_name)
1065 def report_download_webpage(self, video_id):
1066 """Report webpage download."""
1067 self.to_screen('%s: Downloading webpage
' % video_id)
1069 def report_age_confirmation(self):
1070 """Report attempt to confirm age."""
1071 self.to_screen('Confirming age
')
1073 def report_login(self):
1074 """Report attempt to log in."""
1075 self.to_screen('Logging
in')
1077 def raise_login_required(
1078 self, msg='This video
is only available
for registered users
',
1079 metadata_available=False, method='any
'):
1080 if metadata_available and self.get_param('ignore_no_formats_error
'):
1081 self.report_warning(msg)
1082 if method is not None:
1083 msg = '%s. %s' % (msg, self._LOGIN_HINTS[method])
1084 raise ExtractorError(msg, expected=True)
1086 def raise_geo_restricted(
1087 self, msg='This video
is not available
from your location due to geo restriction
',
1088 countries=None, metadata_available=False):
1089 if metadata_available and self.get_param('ignore_no_formats_error
'):
1090 self.report_warning(msg)
1092 raise GeoRestrictedError(msg, countries=countries)
1094 def raise_no_formats(self, msg, expected=False, video_id=None):
1095 if expected and self.get_param('ignore_no_formats_error
'):
1096 self.report_warning(msg, video_id)
1097 elif isinstance(msg, ExtractorError):
1100 raise ExtractorError(msg, expected=expected, video_id=video_id)
1102 # Methods for following #608
1104 def url_result(url, ie=None, video_id=None, video_title=None, **kwargs):
1105 """Returns a URL that points to a page that should be processed"""
1106 # TODO: ie should be the class used for getting the info
1107 video_info = {'_type
': 'url
',
1110 video_info.update(kwargs)
1111 if video_id is not None:
1112 video_info['id'] = video_id
1113 if video_title is not None:
1114 video_info['title
'] = video_title
1117 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None):
1119 self.url_result(self._proto_relative_url(getter(m) if getter else m), ie)
1121 return self.playlist_result(
1122 urls, playlist_id=playlist_id, playlist_title=playlist_title)
1125 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None, **kwargs):
1126 """Returns a playlist"""
1127 video_info = {'_type
': 'playlist
',
1129 video_info.update(kwargs)
1131 video_info['id'] = playlist_id
1133 video_info['title
'] = playlist_title
1134 if playlist_description is not None:
1135 video_info['description
'] = playlist_description
1138 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1140 Perform a regex search on the given string, using a single or a list of
1141 patterns returning the first matching group.
1142 In case of failure return a default value or raise a WARNING or a
1143 RegexNotFoundError, depending on fatal, specifying the field name.
1145 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
1146 mobj = re.search(pattern, string, flags)
1149 mobj = re.search(p, string, flags)
1153 _name = self._downloader._format_err(name, self._downloader.Styles.EMPHASIS)
1157 # return the first matching group
1158 return next(g for g in mobj.groups() if g is not None)
1159 elif isinstance(group, (list, tuple)):
1160 return tuple(mobj.group(g) for g in group)
1162 return mobj.group(group)
1163 elif default is not NO_DEFAULT:
1166 raise RegexNotFoundError('Unable to extract
%s' % _name)
1168 self.report_warning('unable to extract
%s' % _name + bug_reports_message())
1171 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1173 Like _search_regex, but strips HTML tags and unescapes entities.
1175 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1177 return clean_html(res).strip()
1181 def _get_netrc_login_info(self, netrc_machine=None):
1184 netrc_machine = netrc_machine or self._NETRC_MACHINE
1186 if self.get_param('usenetrc
', False):
1188 netrc_file = compat_expanduser(self.get_param('netrc_location
') or '~
')
1189 if os.path.isdir(netrc_file):
1190 netrc_file = os.path.join(netrc_file, '.netrc
')
1191 info = netrc.netrc(file=netrc_file).authenticators(netrc_machine)
1192 if info is not None:
1196 raise netrc.NetrcParseError(
1197 'No authenticators
for %s' % netrc_machine)
1198 except (IOError, netrc.NetrcParseError) as err:
1199 self.report_warning(
1200 'parsing
.netrc
: %s' % error_to_compat_str(err))
1202 return username, password
1204 def _get_login_info(self, username_option='username
', password_option='password
', netrc_machine=None):
1206 Get the login info as (username, password)
1207 First look for the manually specified credentials using username_option
1208 and password_option as keys in params dictionary. If no such credentials
1209 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1211 If there's no info available
, return (None, None)
1214 # Attempt to use provided username and password or .netrc data
1215 username = self.get_param(username_option)
1216 if username is not None:
1217 password = self.get_param(password_option)
1219 username, password = self._get_netrc_login_info(netrc_machine)
1221 return username, password
1223 def _get_tfa_info(self, note='two-factor verification code'):
1225 Get the two
-factor authentication info
1226 TODO
- asking the user will be required
for sms
/phone verify
1227 currently just uses the command line option
1228 If there
's no info available, return None
1231 tfa = self.get_param('twofactor
')
1235 return compat_getpass('Type
%s and press
[Return
]: ' % note)
1237 # Helper functions for extracting OpenGraph info
1239 def _og_regexes(prop):
1240 content_re = r'content
=(?
:"([^"]+?
)"|\'([^\']+?)\'|\s*([^\s"\'=<>`
]+?
))'
1241 property_re = (r'(?
:name|
property)=(?
:\'og
[:-]%(prop)s\'|
"og[:-]%(prop)s"|\s
*og
[:-]%(prop)s\b)'
1242 % {'prop': re.escape(prop)})
1243 template = r'<meta
[^
>]+?
%s[^
>]+?
%s'
1245 template % (property_re, content_re),
1246 template % (content_re, property_re),
1250 def _meta_regex(prop):
1251 return r'''(?isx)<meta
1252 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1253 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1255 def _og_search_property(self, prop, html, name=None, **kargs):
1256 prop = variadic(prop)
1258 name = 'OpenGraph
%s' % prop[0]
1261 og_regexes.extend(self._og_regexes(p))
1262 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1265 return unescapeHTML(escaped)
1267 def _og_search_thumbnail(self, html, **kargs):
1268 return self._og_search_property('image
', html, 'thumbnail URL
', fatal=False, **kargs)
1270 def _og_search_description(self, html, **kargs):
1271 return self._og_search_property('description
', html, fatal=False, **kargs)
1273 def _og_search_title(self, html, **kargs):
1274 return self._og_search_property('title
', html, **kargs)
1276 def _og_search_video_url(self, html, name='video url
', secure=True, **kargs):
1277 regexes = self._og_regexes('video
') + self._og_regexes('video
:url
')
1279 regexes = self._og_regexes('video
:secure_url
') + regexes
1280 return self._html_search_regex(regexes, html, name, **kargs)
1282 def _og_search_url(self, html, **kargs):
1283 return self._og_search_property('url
', html, **kargs)
1285 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1286 name = variadic(name)
1287 if display_name is None:
1288 display_name = name[0]
1289 return self._html_search_regex(
1290 [self._meta_regex(n) for n in name],
1291 html, display_name, fatal=fatal, group='content
', **kwargs)
1293 def _dc_search_uploader(self, html):
1294 return self._html_search_meta('dc
.creator
', html, 'uploader
')
1296 def _rta_search(self, html):
1297 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1298 if re.search(r'(?ix
)<meta\s
+name
="rating"\s
+'
1299 r' content
="RTA-5042-1996-1400-1577-RTA"',
1304 def _media_rating_search(self, html):
1305 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1306 rating = self._html_search_meta('rating
', html)
1318 return RATING_TABLE.get(rating.lower())
1320 def _family_friendly_search(self, html):
1321 # See http://schema.org/VideoObject
1322 family_friendly = self._html_search_meta(
1323 'isFamilyFriendly
', html, default=None)
1325 if not family_friendly:
1334 return RATING_TABLE.get(family_friendly.lower())
1336 def _twitter_search_player(self, html):
1337 return self._html_search_meta('twitter
:player
', html,
1338 'twitter card player
')
1340 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
1341 json_ld_list = list(re.finditer(JSON_LD_RE, html))
1342 default = kwargs.get('default
', NO_DEFAULT)
1343 # JSON-LD may be malformed and thus `fatal` should be respected.
1344 # At the same time `default` may be passed that assumes `fatal=False`
1345 # for _search_regex. Let's simulate the same behavior here
as well
.
1346 fatal
= kwargs
.get('fatal', True) if default
is NO_DEFAULT
else False
1348 for mobj
in json_ld_list
:
1349 json_ld_item
= self
._parse
_json
(
1350 mobj
.group('json_ld'), video_id
, fatal
=fatal
)
1351 if not json_ld_item
:
1353 if isinstance(json_ld_item
, dict):
1354 json_ld
.append(json_ld_item
)
1355 elif isinstance(json_ld_item
, (list, tuple)):
1356 json_ld
.extend(json_ld_item
)
1358 json_ld
= self
._json
_ld
(json_ld
, video_id
, fatal
=fatal
, expected_type
=expected_type
)
1361 if default
is not NO_DEFAULT
:
1364 raise RegexNotFoundError('Unable to extract JSON-LD')
1366 self
.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1369 def _json_ld(self
, json_ld
, video_id
, fatal
=True, expected_type
=None):
1370 if isinstance(json_ld
, compat_str
):
1371 json_ld
= self
._parse
_json
(json_ld
, video_id
, fatal
=fatal
)
1375 if not isinstance(json_ld
, (list, tuple, dict)):
1377 if isinstance(json_ld
, dict):
1380 INTERACTION_TYPE_MAP
= {
1381 'CommentAction': 'comment',
1382 'AgreeAction': 'like',
1383 'DisagreeAction': 'dislike',
1384 'LikeAction': 'like',
1385 'DislikeAction': 'dislike',
1386 'ListenAction': 'view',
1387 'WatchAction': 'view',
1388 'ViewAction': 'view',
1391 def extract_interaction_type(e
):
1392 interaction_type
= e
.get('interactionType')
1393 if isinstance(interaction_type
, dict):
1394 interaction_type
= interaction_type
.get('@type')
1395 return str_or_none(interaction_type
)
1397 def extract_interaction_statistic(e
):
1398 interaction_statistic
= e
.get('interactionStatistic')
1399 if isinstance(interaction_statistic
, dict):
1400 interaction_statistic
= [interaction_statistic
]
1401 if not isinstance(interaction_statistic
, list):
1403 for is_e
in interaction_statistic
:
1404 if not isinstance(is_e
, dict):
1406 if is_e
.get('@type') != 'InteractionCounter':
1408 interaction_type
= extract_interaction_type(is_e
)
1409 if not interaction_type
:
1411 # For interaction count some sites provide string instead of
1412 # an integer (as per spec) with non digit characters (e.g. ",")
1413 # so extracting count with more relaxed str_to_int
1414 interaction_count
= str_to_int(is_e
.get('userInteractionCount'))
1415 if interaction_count
is None:
1417 count_kind
= INTERACTION_TYPE_MAP
.get(interaction_type
.split('/')[-1])
1420 count_key
= '%s_count' % count_kind
1421 if info
.get(count_key
) is not None:
1423 info
[count_key
] = interaction_count
1425 def extract_video_object(e
):
1426 assert e
['@type'] == 'VideoObject'
1427 author
= e
.get('author')
1429 'url': url_or_none(e
.get('contentUrl')),
1430 'title': unescapeHTML(e
.get('name')),
1431 'description': unescapeHTML(e
.get('description')),
1432 'thumbnail': url_or_none(e
.get('thumbnailUrl') or e
.get('thumbnailURL')),
1433 'duration': parse_duration(e
.get('duration')),
1434 'timestamp': unified_timestamp(e
.get('uploadDate')),
1435 # author can be an instance of 'Organization' or 'Person' types.
1436 # both types can have 'name' property(inherited from 'Thing' type). [1]
1437 # however some websites are using 'Text' type instead.
1438 # 1. https://schema.org/VideoObject
1439 'uploader': author
.get('name') if isinstance(author
, dict) else author
if isinstance(author
, compat_str
) else None,
1440 'filesize': float_or_none(e
.get('contentSize')),
1441 'tbr': int_or_none(e
.get('bitrate')),
1442 'width': int_or_none(e
.get('width')),
1443 'height': int_or_none(e
.get('height')),
1444 'view_count': int_or_none(e
.get('interactionCount')),
1446 extract_interaction_statistic(e
)
1450 item_type
= e
.get('@type')
1451 if expected_type
is not None and expected_type
!= item_type
:
1453 if item_type
in ('TVEpisode', 'Episode'):
1454 episode_name
= unescapeHTML(e
.get('name'))
1456 'episode': episode_name
,
1457 'episode_number': int_or_none(e
.get('episodeNumber')),
1458 'description': unescapeHTML(e
.get('description')),
1460 if not info
.get('title') and episode_name
:
1461 info
['title'] = episode_name
1462 part_of_season
= e
.get('partOfSeason')
1463 if isinstance(part_of_season
, dict) and part_of_season
.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1465 'season': unescapeHTML(part_of_season
.get('name')),
1466 'season_number': int_or_none(part_of_season
.get('seasonNumber')),
1468 part_of_series
= e
.get('partOfSeries') or e
.get('partOfTVSeries')
1469 if isinstance(part_of_series
, dict) and part_of_series
.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1470 info
['series'] = unescapeHTML(part_of_series
.get('name'))
1471 elif item_type
== 'Movie':
1473 'title': unescapeHTML(e
.get('name')),
1474 'description': unescapeHTML(e
.get('description')),
1475 'duration': parse_duration(e
.get('duration')),
1476 'timestamp': unified_timestamp(e
.get('dateCreated')),
1478 elif item_type
in ('Article', 'NewsArticle'):
1480 'timestamp': parse_iso8601(e
.get('datePublished')),
1481 'title': unescapeHTML(e
.get('headline')),
1482 'description': unescapeHTML(e
.get('articleBody')),
1484 elif item_type
== 'VideoObject':
1485 extract_video_object(e
)
1486 if expected_type
is None:
1490 video
= e
.get('video')
1491 if isinstance(video
, dict) and video
.get('@type') == 'VideoObject':
1492 extract_video_object(video
)
1493 if expected_type
is None:
1497 return dict((k
, v
) for k
, v
in info
.items() if v
is not None)
1499 def _search_nextjs_data(self
, webpage
, video_id
, **kw
):
1500 return self
._parse
_json
(
1502 r
'(?s)<script[^>]+id=[\'"]__NEXT_DATA__[\'"][^
>]*>([^
<]+)</script
>',
1503 webpage, 'next
.js data
', **kw),
1507 def _hidden_inputs(html):
1508 html = re.sub(r'<!--(?
:(?
!<!--).)*-->', '', html)
1510 for input in re.findall(r'(?i
)(<input[^
>]+>)', html):
1511 attrs = extract_attributes(input)
1514 if attrs.get('type') not in ('hidden
', 'submit
'):
1516 name = attrs.get('name
') or attrs.get('id')
1517 value = attrs.get('value
')
1518 if name and value is not None:
1519 hidden_inputs[name] = value
1520 return hidden_inputs
1522 def _form_hidden_inputs(self, form_id, html):
1523 form = self._search_regex(
1524 r'(?
is)<form
[^
>]+?
id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1525 html, '%s form' % form_id, group='form')
1526 return self._hidden_inputs(form)
1529 regex = r' *((?P<reverse>\+)?(?P<field>[a-zA-Z0-9_]+)((?P<separator>[~:])(?P<limit>.*?))?)? *$'
1531 default = ('hidden', 'aud_or_vid', 'hasvid', 'ie_pref', 'lang', 'quality',
1532 'res', 'fps', 'hdr:12', 'codec:vp9.2', 'size', 'br', 'asr',
1533 'proto', 'ext', 'hasaud', 'source', 'format_id') # These must not be aliases
1534 ytdl_default = ('hasaud', 'lang', 'quality', 'tbr', 'filesize', 'vbr',
1535 'height', 'width', 'proto', 'vext', 'abr', 'aext',
1536 'fps', 'fs_approx', 'source', 'format_id')
1539 'vcodec': {'type': 'ordered', 'regex': True,
1540 'order': ['av0?1', 'vp0?9.2', 'vp0?9', '[hx]265|he?vc?', '[hx]264|avc', 'vp0?8', 'mp4v|h263', 'theora', '', None, 'none']},
1541 'acodec': {'type': 'ordered', 'regex': True,
1542 'order': ['opus', 'vorbis', 'aac', 'mp?4a?', 'mp3', 'e-?a?c-?3', 'ac-?3', 'dts', '', None, 'none']},
1543 'hdr': {'type': 'ordered', 'regex': True, 'field': 'dynamic_range',
1544 'order': ['dv', '(hdr)?12', r'(hdr)?10\+', '(hdr)?10', 'hlg', '', 'sdr', None]},
1545 'proto': {'type': 'ordered', 'regex': True, 'field': 'protocol',
1546 'order': ['(ht|f)tps', '(ht|f)tp$', 'm3u8.+', '.*dash', 'ws|websocket', '', 'mms|rtsp', 'none', 'f4']},
1547 'vext': {'type': 'ordered', 'field': 'video_ext',
1548 'order': ('mp4', 'webm', 'flv', '', 'none'),
1549 'order_free': ('webm', 'mp4', 'flv', '', 'none')},
1550 'aext': {'type': 'ordered', 'field': 'audio_ext',
1551 'order': ('m4a', 'aac', 'mp3', 'ogg', 'opus', 'webm', '', 'none'),
1552 'order_free': ('opus', 'ogg', 'webm', 'm4a', 'mp3', 'aac', '', 'none')},
1553 'hidden': {'visible': False, 'forced': True, 'type': 'extractor', 'max': -1000},
1554 'aud_or_vid': {'visible': False, 'forced': True, 'type': 'multiple',
1555 'field': ('vcodec', 'acodec'),
1556 'function': lambda it: int(any(v != 'none' for v in it))},
1557 'ie_pref': {'priority': True, 'type': 'extractor'},
1558 'hasvid': {'priority': True, 'field': 'vcodec', 'type': 'boolean', 'not_in_list': ('none',)},
1559 'hasaud': {'field': 'acodec', 'type': 'boolean', 'not_in_list': ('none',)},
1560 'lang': {'convert': 'float', 'field': 'language_preference', 'default': -1},
1561 'quality': {'convert': 'float', 'default': -1},
1562 'filesize': {'convert': 'bytes'},
1563 'fs_approx': {'convert': 'bytes', 'field': 'filesize_approx'},
1564 'id': {'convert': 'string', 'field': 'format_id'},
1565 'height': {'convert': 'float_none'},
1566 'width': {'convert': 'float_none'},
1567 'fps': {'convert': 'float_none'},
1568 'tbr': {'convert': 'float_none'},
1569 'vbr': {'convert': 'float_none'},
1570 'abr': {'convert': 'float_none'},
1571 'asr': {'convert': 'float_none'},
1572 'source': {'convert': 'float', 'field': 'source_preference', 'default': -1},
1574 'codec': {'type': 'combined', 'field': ('vcodec', 'acodec')},
1575 'br': {'type': 'combined', 'field': ('tbr', 'vbr', 'abr'), 'same_limit': True},
1576 'size': {'type': 'combined', 'same_limit': True, 'field': ('filesize', 'fs_approx')},
1577 'ext': {'type': 'combined', 'field': ('vext', 'aext')},
1578 'res': {'type': 'multiple', 'field': ('height', 'width'),
1579 'function': lambda it: (lambda l: min(l) if l else 0)(tuple(filter(None, it)))},
1581 # Most of these exist only for compatibility reasons
1582 'dimension': {'type': 'alias', 'field': 'res'},
1583 'resolution': {'type': 'alias', 'field': 'res'},
1584 'extension': {'type': 'alias', 'field': 'ext'},
1585 'bitrate': {'type': 'alias', 'field': 'br'},
1586 'total_bitrate': {'type': 'alias', 'field': 'tbr'},
1587 'video_bitrate': {'type': 'alias', 'field': 'vbr'},
1588 'audio_bitrate': {'type': 'alias', 'field': 'abr'},
1589 'framerate': {'type': 'alias', 'field': 'fps'},
1590 'language_preference': {'type': 'alias', 'field': 'lang'}, # not named as 'language' because such a field exists
1591 'protocol': {'type': 'alias', 'field': 'proto'},
1592 'source_preference': {'type': 'alias', 'field': 'source'},
1593 'filesize_approx': {'type': 'alias', 'field': 'fs_approx'},
1594 'filesize_estimate': {'type': 'alias', 'field': 'size'},
1595 'samplerate': {'type': 'alias', 'field': 'asr'},
1596 'video_ext': {'type': 'alias', 'field': 'vext'},
1597 'audio_ext': {'type': 'alias', 'field': 'aext'},
1598 'video_codec': {'type': 'alias', 'field': 'vcodec'},
1599 'audio_codec': {'type': 'alias', 'field': 'acodec'},
1600 'video': {'type': 'alias', 'field': 'hasvid'},
1601 'has_video': {'type': 'alias', 'field': 'hasvid'},
1602 'audio': {'type': 'alias', 'field': 'hasaud'},
1603 'has_audio': {'type': 'alias', 'field': 'hasaud'},
1604 'extractor': {'type': 'alias', 'field': 'ie_pref'},
1605 'preference': {'type': 'alias', 'field': 'ie_pref'},
1606 'extractor_preference': {'type': 'alias', 'field': 'ie_pref'},
1607 'format_id': {'type': 'alias', 'field': 'id'},
1612 def _get_field_setting(self, field, key):
1613 if field not in self.settings:
1614 self.settings[field] = {}
1615 propObj = self.settings[field]
1616 if key not in propObj:
1617 type = propObj.get('type')
1619 default = 'preference' if type == 'extractor' else (field,) if type in ('combined', 'multiple') else field
1620 elif key == 'convert':
1621 default = 'order' if type == 'ordered' else 'float_string' if field else 'ignore'
1623 default = {'type': 'field', 'visible': True, 'order': [], 'not_in_list': (None,)}.get(key, None)
1624 propObj[key] = default
1627 def _resolve_field_value(self, field, value, convertNone=False):
1632 value = value.lower()
1633 conversion = self._get_field_setting(field, 'convert')
1634 if conversion == 'ignore':
1636 if conversion == 'string':
1638 elif conversion == 'float_none':
1639 return float_or_none(value)
1640 elif conversion == 'bytes':
1641 return FileDownloader.parse_bytes(value)
1642 elif conversion == 'order':
1643 order_list = (self._use_free_order and self._get_field_setting(field, 'order_free')) or self._get_field_setting(field, 'order')
1644 use_regex = self._get_field_setting(field, 'regex')
1645 list_length = len(order_list)
1646 empty_pos = order_list.index('') if '' in order_list else list_length + 1
1647 if use_regex and value is not None:
1648 for i, regex in enumerate(order_list):
1649 if regex and re.match(regex, value):
1650 return list_length - i
1651 return list_length - empty_pos # not in list
1652 else: # not regex or value = None
1653 return list_length - (order_list.index(value) if value in order_list else empty_pos)
1655 if value.isnumeric():
1658 self.settings[field]['convert'] = 'string'
1661 def evaluate_params(self, params, sort_extractor):
1662 self._use_free_order = params.get('prefer_free_formats', False)
1663 self._sort_user = params.get('format_sort', [])
1664 self._sort_extractor = sort_extractor
1666 def add_item(field, reverse, closest, limit_text):
1667 field = field.lower()
1668 if field in self._order:
1670 self._order.append(field)
1671 limit = self._resolve_field_value(field, limit_text)
1674 'closest': False if limit is None else closest,
1675 'limit_text': limit_text,
1677 if field in self.settings:
1678 self.settings[field].update(data)
1680 self.settings[field] = data
1683 tuple(field for field in self.default if self._get_field_setting(field, 'forced'))
1684 + (tuple() if params.get('format_sort_force', False)
1685 else tuple(field for field in self.default if self._get_field_setting(field, 'priority')))
1686 + tuple(self._sort_user) + tuple(sort_extractor) + self.default)
1688 for item in sort_list:
1689 match = re.match(self.regex, item)
1691 raise ExtractorError('Invalid format sort string "%s" given by extractor' % item)
1692 field = match.group('field')
1695 if self._get_field_setting(field, 'type') == 'alias':
1696 field = self._get_field_setting(field, 'field')
1697 reverse = match.group('reverse') is not None
1698 closest = match.group('separator') == '~'
1699 limit_text = match.group('limit')
1701 has_limit = limit_text is not None
1702 has_multiple_fields = self._get_field_setting(field, 'type') == 'combined'
1703 has_multiple_limits = has_limit and has_multiple_fields and not self._get_field_setting(field, 'same_limit')
1705 fields = self._get_field_setting(field, 'field') if has_multiple_fields else (field,)
1706 limits = limit_text.split(':') if has_multiple_limits else (limit_text,) if has_limit else tuple()
1707 limit_count = len(limits)
1708 for (i, f) in enumerate(fields):
1709 add_item(f, reverse, closest,
1710 limits[i] if i < limit_count
1711 else limits[0] if has_limit and not has_multiple_limits
1714 def print_verbose_info(self, write_debug):
1716 write_debug('Sort order given by user: %s' % ', '.join(self._sort_user))
1717 if self._sort_extractor:
1718 write_debug('Sort order given by extractor: %s' % ', '.join(self._sort_extractor))
1719 write_debug('Formats sorted by: %s' % ', '.join(['%s%s%s' % (
1720 '+' if self._get_field_setting(field, 'reverse') else '', field,
1721 '%s%s(%s)' % ('~' if self._get_field_setting(field, 'closest') else ':',
1722 self._get_field_setting(field, 'limit_text'),
1723 self._get_field_setting(field, 'limit'))
1724 if self._get_field_setting(field, 'limit_text') is not None else '')
1725 for field in self._order if self._get_field_setting(field, 'visible')]))
1727 def _calculate_field_preference_from_value(self, format, field, type, value):
1728 reverse = self._get_field_setting(field, 'reverse')
1729 closest = self._get_field_setting(field, 'closest')
1730 limit = self._get_field_setting(field, 'limit')
1732 if type == 'extractor':
1733 maximum = self._get_field_setting(field, 'max')
1734 if value is None or (maximum is not None and value >= maximum):
1736 elif type == 'boolean':
1737 in_list = self._get_field_setting(field, 'in_list')
1738 not_in_list = self._get_field_setting(field, 'not_in_list')
1739 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
1740 elif type == 'ordered':
1741 value = self._resolve_field_value(field, value, True)
1743 # try to convert to number
1744 val_num = float_or_none(value, default=self._get_field_setting(field, 'default'))
1745 is_num = self._get_field_setting(field, 'convert') != 'string' and val_num is not None
1749 return ((-10, 0) if value is None
1750 else (1, value, 0) if not is_num # if a field has mixed strings and numbers, strings are sorted higher
1751 else (0, -abs(value - limit), value - limit if reverse else limit - value) if closest
1752 else (0, value, 0) if not reverse and (limit is None or value <= limit)
1753 else (0, -value, 0) if limit is None or (reverse and value == limit) or value > limit
1754 else (-1, value, 0))
1756 def _calculate_field_preference(self, format, field):
1757 type = self._get_field_setting(field, 'type') # extractor, boolean, ordered, field, multiple
1758 get_value = lambda f: format.get(self._get_field_setting(f, 'field'))
1759 if type == 'multiple':
1760 type = 'field' # Only 'field' is allowed in multiple for now
1761 actual_fields = self._get_field_setting(field, 'field')
1763 value = self._get_field_setting(field, 'function')(get_value(f) for f in actual_fields)
1765 value = get_value(field)
1766 return self._calculate_field_preference_from_value(format, field, type, value)
1768 def calculate_preference(self, format):
1769 # Determine missing protocol
1770 if not format.get('protocol'):
1771 format['protocol'] = determine_protocol(format)
1773 # Determine missing ext
1774 if not format.get('ext') and 'url' in format:
1775 format['ext'] = determine_ext(format['url'])
1776 if format.get('vcodec') == 'none':
1777 format['audio_ext'] = format['ext'] if format.get('acodec') != 'none' else 'none'
1778 format['video_ext'] = 'none'
1780 format['video_ext'] = format['ext']
1781 format['audio_ext'] = 'none'
1782 # if format.get('preference') is None and format.get('ext') in ('f4f', 'f4m'): # Not supported?
1783 # format['preference'] = -1000
1785 # Determine missing bitrates
1786 if format.get('tbr') is None:
1787 if format.get('vbr') is not None and format.get('abr') is not None:
1788 format['tbr'] = format.get('vbr', 0) + format.get('abr', 0)
1790 if format.get('vcodec') != 'none' and format.get('vbr') is None:
1791 format['vbr'] = format.get('tbr') - format.get('abr', 0)
1792 if format.get('acodec') != 'none' and format.get('abr') is None:
1793 format['abr'] = format.get('tbr') - format.get('vbr', 0)
1795 return tuple(self._calculate_field_preference(format, field) for field in self._order)
1797 def _sort_formats(self, formats, field_preference=[]):
1800 format_sort = self.FormatSort() # params and to_screen are taken from the downloader
1801 format_sort.evaluate_params(self._downloader.params, field_preference)
1802 if self.get_param('verbose', False):
1803 format_sort.print_verbose_info(self._downloader.write_debug)
1804 formats.sort(key=lambda f: format_sort.calculate_preference(f))
1806 def _check_formats(self, formats, video_id):
1808 formats[:] = filter(
1809 lambda f: self._is_valid_url(
1811 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1815 def _remove_duplicate_formats(formats):
1819 if f['url'] not in format_urls:
1820 format_urls.add(f['url'])
1821 unique_formats.append(f)
1822 formats[:] = unique_formats
1824 def _is_valid_url(self, url, video_id, item='video', headers={}):
1825 url = self._proto_relative_url(url, scheme='http:')
1826 # For now assume non HTTP(S) URLs always valid
1827 if not (url.startswith('http://') or url.startswith('https://')):
1830 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1832 except ExtractorError as e:
1834 '%s: %s URL is invalid, skipping: %s'
1835 % (video_id, item, error_to_compat_str(e.cause)))
1838 def http_scheme(self):
1839 """ Either "http
:" or "https
:", depending on the user's preferences """
1842 if self.get_param('prefer_insecure', False)
1845 def _proto_relative_url(self, url, scheme=None):
1848 if url.startswith('//'):
1850 scheme = self.http_scheme()
1855 def _sleep(self, timeout, video_id, msg_template=None):
1856 if msg_template is None:
1857 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1858 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1862 def _extract_f4m_formats(self, 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, data=None, headers={}, query={}):
1865 manifest = self._download_xml(
1866 manifest_url, video_id, 'Downloading f4m manifest',
1867 'Unable to download f4m manifest',
1868 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1869 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
1870 transform_source=transform_source,
1871 fatal=fatal, data=data, headers=headers, query=query)
1873 if manifest is False:
1876 return self._parse_f4m_formats(
1877 manifest, manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1878 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1880 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1881 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1882 fatal=True, m3u8_id=None):
1883 if not isinstance(manifest, compat_etree_Element) and not fatal:
1886 # currently yt-dlp cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1887 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1888 if akamai_pv is not None and ';' in akamai_pv.text:
1889 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1890 if playerVerificationChallenge.strip() != '':
1894 manifest_version = '1.0'
1895 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1897 manifest_version = '2.0'
1898 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1899 # Remove unsupported DRM protected media from final formats
1900 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
1901 media_nodes = remove_encrypted_media(media_nodes)
1905 manifest_base_url = get_base_url(manifest)
1907 bootstrap_info = xpath_element(
1908 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1909 'bootstrap info', default=None)
1912 mime_type = xpath_text(
1913 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1914 'base URL', default=None)
1915 if mime_type and mime_type.startswith('audio/'):
1918 for i, media_el in enumerate(media_nodes):
1919 tbr = int_or_none(media_el.attrib.get('bitrate'))
1920 width = int_or_none(media_el.attrib.get('width'))
1921 height = int_or_none(media_el.attrib.get('height'))
1922 format_id = join_nonempty(f4m_id, tbr or i)
1923 # If <bootstrapInfo> is present, the specified f4m is a
1924 # stream-level manifest, and only set-level manifests may refer to
1925 # external resources. See section 11.4 and section 4 of F4M spec
1926 if bootstrap_info is None:
1928 # @href is introduced in 2.0, see section 11.6 of F4M spec
1929 if manifest_version == '2.0':
1930 media_url = media_el.attrib.get('href')
1931 if media_url is None:
1932 media_url = media_el.attrib.get('url')
1936 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1937 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1938 # If media_url is itself a f4m manifest do the recursive extraction
1939 # since bitrates in parent manifest (this one) and media_url manifest
1940 # may differ leading to inability to resolve the format by requested
1941 # bitrate in f4m downloader
1942 ext = determine_ext(manifest_url)
1944 f4m_formats = self._extract_f4m_formats(
1945 manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1946 transform_source=transform_source, fatal=fatal)
1947 # Sometimes stream-level manifest contains single media entry that
1948 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1949 # At the same time parent's media entry in set-level manifest may
1950 # contain it. We will copy it from parent in such cases.
1951 if len(f4m_formats) == 1:
1954 'tbr': f.get('tbr') or tbr,
1955 'width': f.get('width') or width,
1956 'height': f.get('height') or height,
1957 'format_id': f.get('format_id') if not tbr else format_id,
1960 formats.extend(f4m_formats)
1963 formats.extend(self._extract_m3u8_formats(
1964 manifest_url, video_id, 'mp4', preference=preference,
1965 quality=quality, m3u8_id=m3u8_id, fatal=fatal))
1968 'format_id': format_id,
1969 'url': manifest_url,
1970 'manifest_url': manifest_url,
1971 'ext': 'flv' if bootstrap_info is not None else None,
1977 'preference': preference,
1982 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, quality=None, m3u8_id=None):
1984 'format_id': join_nonempty(m3u8_id, 'meta'),
1988 'preference': preference - 100 if preference else -100,
1990 'resolution': 'multiple',
1991 'format_note': 'Quality selection URL',
1994 def _report_ignoring_subs(self, name):
1995 self.report_warning(bug_reports_message(
1996 f'Ignoring subtitle tracks found in the {name} manifest; '
1997 'if any subtitle tracks are missing,'
2000 def _extract_m3u8_formats(self, *args, **kwargs):
2001 fmts, subs = self._extract_m3u8_formats_and_subtitles(*args, **kwargs)
2003 self._report_ignoring_subs('HLS')
2006 def _extract_m3u8_formats_and_subtitles(
2007 self, m3u8_url, video_id, ext=None, entry_protocol='m3u8_native',
2008 preference=None, quality=None, m3u8_id=None, note=None,
2009 errnote=None, fatal=True, live=False, data=None, headers={},
2012 res = self._download_webpage_handle(
2014 note='Downloading m3u8 information' if note is None else note,
2015 errnote='Failed to download m3u8 information' if errnote is None else errnote,
2016 fatal=fatal, data=data, headers=headers, query=query)
2021 m3u8_doc, urlh = res
2022 m3u8_url = urlh.geturl()
2024 return self._parse_m3u8_formats_and_subtitles(
2025 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
2026 preference=preference, quality=quality, m3u8_id=m3u8_id,
2027 note=note, errnote=errnote, fatal=fatal, live=live, data=data,
2028 headers=headers, query=query, video_id=video_id)
2030 def _parse_m3u8_formats_and_subtitles(
2031 self, m3u8_doc, m3u8_url, ext=None, entry_protocol='m3u8_native',
2032 preference=None, quality=None, m3u8_id=None, live=False, note=None,
2033 errnote=None, fatal=True, data=None, headers={}, query={},
2035 formats, subtitles = [], {}
2037 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
2038 return formats, subtitles
2040 has_drm = re.search(r'#EXT-X-(?:SESSION-)?KEY:.*?URI="skd
://', m3u8_doc)
2042 def format_url(url):
2043 return url if re.match(r'^https?
://', url) else compat_urlparse.urljoin(m3u8_url, url)
2045 if self.get_param('hls_split_discontinuity
', False):
2046 def _extract_m3u8_playlist_indices(manifest_url=None, m3u8_doc=None):
2048 if not manifest_url:
2050 m3u8_doc = self._download_webpage(
2051 manifest_url, video_id, fatal=fatal, data=data, headers=headers,
2052 note=False, errnote='Failed to download m3u8 playlist information
')
2053 if m3u8_doc is False:
2055 return range(1 + sum(line.startswith('#EXT-X-DISCONTINUITY') for line in m3u8_doc.splitlines()))
2058 def _extract_m3u8_playlist_indices(*args
, **kwargs
):
2062 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
2063 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
2064 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
2066 # We should try extracting formats only from master playlists [1, 4.3.4],
2067 # i.e. playlists that describe available qualities. On the other hand
2068 # media playlists [1, 4.3.3] should be returned as is since they contain
2069 # just the media without qualities renditions.
2070 # Fortunately, master playlist can be easily distinguished from media
2071 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
2072 # master playlist tags MUST NOT appear in a media playlist and vice versa.
2073 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
2074 # media playlist and MUST NOT appear in master playlist thus we can
2075 # clearly detect media playlist with this criterion.
2077 if '#EXT-X-TARGETDURATION' in m3u8_doc
: # media playlist, return as is
2079 'format_id': join_nonempty(m3u8_id
, idx
),
2080 'format_index': idx
,
2083 'protocol': entry_protocol
,
2084 'preference': preference
,
2087 } for idx
in _extract_m3u8_playlist_indices(m3u8_doc
=m3u8_doc
)]
2089 return formats
, subtitles
2092 last_stream_inf
= {}
2094 def extract_media(x_media_line
):
2095 media
= parse_m3u8_attributes(x_media_line
)
2096 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
2097 media_type
, group_id
, name
= media
.get('TYPE'), media
.get('GROUP-ID'), media
.get('NAME')
2098 if not (media_type
and group_id
and name
):
2100 groups
.setdefault(group_id
, []).append(media
)
2101 # <https://tools.ietf.org/html/rfc8216#section-4.3.4.1>
2102 if media_type
== 'SUBTITLES':
2103 # According to RFC 8216 §4.3.4.2.1, URI is REQUIRED in the
2104 # EXT-X-MEDIA tag if the media type is SUBTITLES.
2105 # However, lack of URI has been spotted in the wild.
2106 # e.g. NebulaIE; see https://github.com/yt-dlp/yt-dlp/issues/339
2107 if not media
.get('URI'):
2109 url
= format_url(media
['URI'])
2112 'ext': determine_ext(url
),
2114 if sub_info
['ext'] == 'm3u8':
2115 # Per RFC 8216 §3.1, the only possible subtitle format m3u8
2116 # files may contain is WebVTT:
2117 # <https://tools.ietf.org/html/rfc8216#section-3.1>
2118 sub_info
['ext'] = 'vtt'
2119 sub_info
['protocol'] = 'm3u8_native'
2120 lang
= media
.get('LANGUAGE') or 'und'
2121 subtitles
.setdefault(lang
, []).append(sub_info
)
2122 if media_type
not in ('VIDEO', 'AUDIO'):
2124 media_url
= media
.get('URI')
2126 manifest_url
= format_url(media_url
)
2128 'format_id': join_nonempty(m3u8_id
, group_id
, name
, idx
),
2129 'format_note': name
,
2130 'format_index': idx
,
2131 'url': manifest_url
,
2132 'manifest_url': m3u8_url
,
2133 'language': media
.get('LANGUAGE'),
2135 'protocol': entry_protocol
,
2136 'preference': preference
,
2138 'vcodec': 'none' if media_type
== 'AUDIO' else None,
2139 } for idx
in _extract_m3u8_playlist_indices(manifest_url
))
2141 def build_stream_name():
2142 # Despite specification does not mention NAME attribute for
2143 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
2144 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
2145 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
2146 stream_name
= last_stream_inf
.get('NAME')
2149 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
2150 # from corresponding rendition group
2151 stream_group_id
= last_stream_inf
.get('VIDEO')
2152 if not stream_group_id
:
2154 stream_group
= groups
.get(stream_group_id
)
2155 if not stream_group
:
2156 return stream_group_id
2157 rendition
= stream_group
[0]
2158 return rendition
.get('NAME') or stream_group_id
2160 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
2161 # chance to detect video only formats when EXT-X-STREAM-INF tags
2162 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
2163 for line
in m3u8_doc
.splitlines():
2164 if line
.startswith('#EXT-X-MEDIA:'):
2167 for line
in m3u8_doc
.splitlines():
2168 if line
.startswith('#EXT-X-STREAM-INF:'):
2169 last_stream_inf
= parse_m3u8_attributes(line
)
2170 elif line
.startswith('#') or not line
.strip():
2173 tbr
= float_or_none(
2174 last_stream_inf
.get('AVERAGE-BANDWIDTH')
2175 or last_stream_inf
.get('BANDWIDTH'), scale
=1000)
2176 manifest_url
= format_url(line
.strip())
2178 for idx
in _extract_m3u8_playlist_indices(manifest_url
):
2179 format_id
= [m3u8_id
, None, idx
]
2180 # Bandwidth of live streams may differ over time thus making
2181 # format_id unpredictable. So it's better to keep provided
2184 stream_name
= build_stream_name()
2185 format_id
[1] = stream_name
or '%d' % (tbr
or len(formats
))
2187 'format_id': join_nonempty(*format_id
),
2188 'format_index': idx
,
2189 'url': manifest_url
,
2190 'manifest_url': m3u8_url
,
2193 'fps': float_or_none(last_stream_inf
.get('FRAME-RATE')),
2194 'protocol': entry_protocol
,
2195 'preference': preference
,
2198 resolution
= last_stream_inf
.get('RESOLUTION')
2200 mobj
= re
.search(r
'(?P<width>\d+)[xX](?P<height>\d+)', resolution
)
2202 f
['width'] = int(mobj
.group('width'))
2203 f
['height'] = int(mobj
.group('height'))
2204 # Unified Streaming Platform
2206 r
'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f
['url'])
2208 abr
, vbr
= mobj
.groups()
2209 abr
, vbr
= float_or_none(abr
, 1000), float_or_none(vbr
, 1000)
2214 codecs
= parse_codecs(last_stream_inf
.get('CODECS'))
2216 audio_group_id
= last_stream_inf
.get('AUDIO')
2217 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
2218 # references a rendition group MUST have a CODECS attribute.
2219 # However, this is not always respected, for example, [2]
2220 # contains EXT-X-STREAM-INF tag which references AUDIO
2221 # rendition group but does not have CODECS and despite
2222 # referencing an audio group it represents a complete
2223 # (with audio and video) format. So, for such cases we will
2224 # ignore references to rendition groups and treat them
2225 # as complete formats.
2226 if audio_group_id
and codecs
and f
.get('vcodec') != 'none':
2227 audio_group
= groups
.get(audio_group_id
)
2228 if audio_group
and audio_group
[0].get('URI'):
2229 # TODO: update acodec for audio only formats with
2231 f
['acodec'] = 'none'
2232 if not f
.get('ext'):
2233 f
['ext'] = 'm4a' if f
.get('vcodec') == 'none' else 'mp4'
2237 progressive_uri
= last_stream_inf
.get('PROGRESSIVE-URI')
2240 del http_f
['manifest_url']
2242 'format_id': f
['format_id'].replace('hls-', 'http-'),
2244 'url': progressive_uri
,
2246 formats
.append(http_f
)
2248 last_stream_inf
= {}
2249 return formats
, subtitles
2251 def _extract_m3u8_vod_duration(
2252 self
, m3u8_vod_url
, video_id
, note
=None, errnote
=None, data
=None, headers
={}, query={}
):
2254 m3u8_vod
= self
._download
_webpage
(
2255 m3u8_vod_url
, video_id
,
2256 note
='Downloading m3u8 VOD manifest' if note
is None else note
,
2257 errnote
='Failed to download VOD manifest' if errnote
is None else errnote
,
2258 fatal
=False, data
=data
, headers
=headers
, query
=query
)
2260 return self
._parse
_m
3u8_vod
_duration
(m3u8_vod
or '', video_id
)
2262 def _parse_m3u8_vod_duration(self
, m3u8_vod
, video_id
):
2263 if '#EXT-X-PLAYLIST-TYPE:VOD' not in m3u8_vod
:
2267 float(line
[len('#EXTINF:'):].split(',')[0])
2268 for line
in m3u8_vod
.splitlines() if line
.startswith('#EXTINF:'))) or None
2271 def _xpath_ns(path
, namespace
=None):
2275 for c
in path
.split('/'):
2276 if not c
or c
== '.':
2279 out
.append('{%s}%s' % (namespace
, c
))
2280 return '/'.join(out
)
2282 def _extract_smil_formats_and_subtitles(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None, transform_source
=None):
2283 smil
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
, transform_source
=transform_source
)
2289 namespace
= self
._parse
_smil
_namespace
(smil
)
2291 fmts
= self
._parse
_smil
_formats
(
2292 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2293 subs
= self
._parse
_smil
_subtitles
(
2294 smil
, namespace
=namespace
)
2298 def _extract_smil_formats(self
, *args
, **kwargs
):
2299 fmts
, subs
= self
._extract
_smil
_formats
_and
_subtitles
(*args
, **kwargs
)
2301 self
._report
_ignoring
_subs
('SMIL')
2304 def _extract_smil_info(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None):
2305 smil
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
)
2308 return self
._parse
_smil
(smil
, smil_url
, video_id
, f4m_params
=f4m_params
)
2310 def _download_smil(self
, smil_url
, video_id
, fatal
=True, transform_source
=None):
2311 return self
._download
_xml
(
2312 smil_url
, video_id
, 'Downloading SMIL file',
2313 'Unable to download SMIL file', fatal
=fatal
, transform_source
=transform_source
)
2315 def _parse_smil(self
, smil
, smil_url
, video_id
, f4m_params
=None):
2316 namespace
= self
._parse
_smil
_namespace
(smil
)
2318 formats
= self
._parse
_smil
_formats
(
2319 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2320 subtitles
= self
._parse
_smil
_subtitles
(smil
, namespace
=namespace
)
2322 video_id
= os
.path
.splitext(url_basename(smil_url
))[0]
2326 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2327 name
= meta
.attrib
.get('name')
2328 content
= meta
.attrib
.get('content')
2329 if not name
or not content
:
2331 if not title
and name
== 'title':
2333 elif not description
and name
in ('description', 'abstract'):
2334 description
= content
2335 elif not upload_date
and name
== 'date':
2336 upload_date
= unified_strdate(content
)
2339 'id': image
.get('type'),
2340 'url': image
.get('src'),
2341 'width': int_or_none(image
.get('width')),
2342 'height': int_or_none(image
.get('height')),
2343 } for image
in smil
.findall(self
._xpath
_ns
('.//image', namespace
)) if image
.get('src')]
2347 'title': title
or video_id
,
2348 'description': description
,
2349 'upload_date': upload_date
,
2350 'thumbnails': thumbnails
,
2352 'subtitles': subtitles
,
2355 def _parse_smil_namespace(self
, smil
):
2356 return self
._search
_regex
(
2357 r
'(?i)^{([^}]+)?}smil$', smil
.tag
, 'namespace', default
=None)
2359 def _parse_smil_formats(self
, smil
, smil_url
, video_id
, namespace
=None, f4m_params
=None, transform_rtmp_url
=None):
2361 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2362 b
= meta
.get('base') or meta
.get('httpBase')
2374 media
= smil
.findall(self
._xpath
_ns
('.//video', namespace
)) + smil
.findall(self
._xpath
_ns
('.//audio', namespace
))
2375 for medium
in media
:
2376 src
= medium
.get('src')
2377 if not src
or src
in srcs
:
2381 bitrate
= float_or_none(medium
.get('system-bitrate') or medium
.get('systemBitrate'), 1000)
2382 filesize
= int_or_none(medium
.get('size') or medium
.get('fileSize'))
2383 width
= int_or_none(medium
.get('width'))
2384 height
= int_or_none(medium
.get('height'))
2385 proto
= medium
.get('proto')
2386 ext
= medium
.get('ext')
2387 src_ext
= determine_ext(src
)
2388 streamer
= medium
.get('streamer') or base
2390 if proto
== 'rtmp' or streamer
.startswith('rtmp'):
2396 'format_id': 'rtmp-%d' % (rtmp_count
if bitrate
is None else bitrate
),
2398 'filesize': filesize
,
2402 if transform_rtmp_url
:
2403 streamer
, src
= transform_rtmp_url(streamer
, src
)
2404 formats
[-1].update({
2410 src_url
= src
if src
.startswith('http') else compat_urlparse
.urljoin(base
, src
)
2411 src_url
= src_url
.strip()
2413 if proto
== 'm3u8' or src_ext
== 'm3u8':
2414 m3u8_formats
= self
._extract
_m
3u8_formats
(
2415 src_url
, video_id
, ext
or 'mp4', m3u8_id
='hls', fatal
=False)
2416 if len(m3u8_formats
) == 1:
2418 m3u8_formats
[0].update({
2419 'format_id': 'hls-%d' % (m3u8_count
if bitrate
is None else bitrate
),
2424 formats
.extend(m3u8_formats
)
2425 elif src_ext
== 'f4m':
2430 'plugin': 'flowplayer-3.2.0.1',
2432 f4m_url
+= '&' if '?' in f4m_url
else '?'
2433 f4m_url
+= compat_urllib_parse_urlencode(f4m_params
)
2434 formats
.extend(self
._extract
_f
4m
_formats
(f4m_url
, video_id
, f4m_id
='hds', fatal
=False))
2435 elif src_ext
== 'mpd':
2436 formats
.extend(self
._extract
_mpd
_formats
(
2437 src_url
, video_id
, mpd_id
='dash', fatal
=False))
2438 elif re
.search(r
'\.ism/[Mm]anifest', src_url
):
2439 formats
.extend(self
._extract
_ism
_formats
(
2440 src_url
, video_id
, ism_id
='mss', fatal
=False))
2441 elif src_url
.startswith('http') and self
._is
_valid
_url
(src
, video_id
):
2445 'ext': ext
or src_ext
or 'flv',
2446 'format_id': 'http-%d' % (bitrate
or http_count
),
2448 'filesize': filesize
,
2453 for medium
in smil
.findall(self
._xpath
_ns
('.//imagestream', namespace
)):
2454 src
= medium
.get('src')
2455 if not src
or src
in srcs
:
2461 'format_id': 'imagestream-%d' % (imgs_count
),
2463 'ext': mimetype2ext(medium
.get('type')),
2466 'width': int_or_none(medium
.get('width')),
2467 'height': int_or_none(medium
.get('height')),
2468 'format_note': 'SMIL storyboards',
2473 def _parse_smil_subtitles(self
, smil
, namespace
=None, subtitles_lang
='en'):
2476 for num
, textstream
in enumerate(smil
.findall(self
._xpath
_ns
('.//textstream', namespace
))):
2477 src
= textstream
.get('src')
2478 if not src
or src
in urls
:
2481 ext
= textstream
.get('ext') or mimetype2ext(textstream
.get('type')) or determine_ext(src
)
2482 lang
= textstream
.get('systemLanguage') or textstream
.get('systemLanguageName') or textstream
.get('lang') or subtitles_lang
2483 subtitles
.setdefault(lang
, []).append({
2489 def _extract_xspf_playlist(self
, xspf_url
, playlist_id
, fatal
=True):
2490 xspf
= self
._download
_xml
(
2491 xspf_url
, playlist_id
, 'Downloading xpsf playlist',
2492 'Unable to download xspf manifest', fatal
=fatal
)
2495 return self
._parse
_xspf
(
2496 xspf
, playlist_id
, xspf_url
=xspf_url
,
2497 xspf_base_url
=base_url(xspf_url
))
2499 def _parse_xspf(self
, xspf_doc
, playlist_id
, xspf_url
=None, xspf_base_url
=None):
2501 'xspf': 'http://xspf.org/ns/0/',
2502 's1': 'http://static.streamone.nl/player/ns/0',
2506 for track
in xspf_doc
.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP
)):
2508 track
, xpath_with_ns('./xspf:title', NS_MAP
), 'title', default
=playlist_id
)
2509 description
= xpath_text(
2510 track
, xpath_with_ns('./xspf:annotation', NS_MAP
), 'description')
2511 thumbnail
= xpath_text(
2512 track
, xpath_with_ns('./xspf:image', NS_MAP
), 'thumbnail')
2513 duration
= float_or_none(
2514 xpath_text(track
, xpath_with_ns('./xspf:duration', NS_MAP
), 'duration'), 1000)
2517 for location
in track
.findall(xpath_with_ns('./xspf:location', NS_MAP
)):
2518 format_url
= urljoin(xspf_base_url
, location
.text
)
2523 'manifest_url': xspf_url
,
2524 'format_id': location
.get(xpath_with_ns('s1:label', NS_MAP
)),
2525 'width': int_or_none(location
.get(xpath_with_ns('s1:width', NS_MAP
))),
2526 'height': int_or_none(location
.get(xpath_with_ns('s1:height', NS_MAP
))),
2528 self
._sort
_formats
(formats
)
2533 'description': description
,
2534 'thumbnail': thumbnail
,
2535 'duration': duration
,
2540 def _extract_mpd_formats(self
, *args
, **kwargs
):
2541 fmts
, subs
= self
._extract
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2543 self
._report
_ignoring
_subs
('DASH')
2546 def _extract_mpd_formats_and_subtitles(
2547 self
, mpd_url
, video_id
, mpd_id
=None, note
=None, errnote
=None,
2548 fatal
=True, data
=None, headers
={}, query={}
):
2549 res
= self
._download
_xml
_handle
(
2551 note
='Downloading MPD manifest' if note
is None else note
,
2552 errnote
='Failed to download MPD manifest' if errnote
is None else errnote
,
2553 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
2559 mpd_base_url
= base_url(urlh
.geturl())
2561 return self
._parse
_mpd
_formats
_and
_subtitles
(
2562 mpd_doc
, mpd_id
, mpd_base_url
, mpd_url
)
2564 def _parse_mpd_formats(self
, *args
, **kwargs
):
2565 fmts
, subs
= self
._parse
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2567 self
._report
_ignoring
_subs
('DASH')
2570 def _parse_mpd_formats_and_subtitles(
2571 self
, mpd_doc
, mpd_id
=None, mpd_base_url
='', mpd_url
=None):
2573 Parse formats from MPD manifest.
2575 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
2576 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
2577 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
2579 if not self
.get_param('dynamic_mpd', True):
2580 if mpd_doc
.get('type') == 'dynamic':
2583 namespace
= self
._search
_regex
(r
'(?i)^{([^}]+)?}MPD$', mpd_doc
.tag
, 'namespace', default
=None)
2586 return self
._xpath
_ns
(path
, namespace
)
2588 def is_drm_protected(element
):
2589 return element
.find(_add_ns('ContentProtection')) is not None
2591 def extract_multisegment_info(element
, ms_parent_info
):
2592 ms_info
= ms_parent_info
.copy()
2594 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2595 # common attributes and elements. We will only extract relevant
2597 def extract_common(source
):
2598 segment_timeline
= source
.find(_add_ns('SegmentTimeline'))
2599 if segment_timeline
is not None:
2600 s_e
= segment_timeline
.findall(_add_ns('S'))
2602 ms_info
['total_number'] = 0
2605 r
= int(s
.get('r', 0))
2606 ms_info
['total_number'] += 1 + r
2607 ms_info
['s'].append({
2608 't': int(s
.get('t', 0)),
2609 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2610 'd': int(s
.attrib
['d']),
2613 start_number
= source
.get('startNumber')
2615 ms_info
['start_number'] = int(start_number
)
2616 timescale
= source
.get('timescale')
2618 ms_info
['timescale'] = int(timescale
)
2619 segment_duration
= source
.get('duration')
2620 if segment_duration
:
2621 ms_info
['segment_duration'] = float(segment_duration
)
2623 def extract_Initialization(source
):
2624 initialization
= source
.find(_add_ns('Initialization'))
2625 if initialization
is not None:
2626 ms_info
['initialization_url'] = initialization
.attrib
['sourceURL']
2628 segment_list
= element
.find(_add_ns('SegmentList'))
2629 if segment_list
is not None:
2630 extract_common(segment_list
)
2631 extract_Initialization(segment_list
)
2632 segment_urls_e
= segment_list
.findall(_add_ns('SegmentURL'))
2634 ms_info
['segment_urls'] = [segment
.attrib
['media'] for segment
in segment_urls_e
]
2636 segment_template
= element
.find(_add_ns('SegmentTemplate'))
2637 if segment_template
is not None:
2638 extract_common(segment_template
)
2639 media
= segment_template
.get('media')
2641 ms_info
['media'] = media
2642 initialization
= segment_template
.get('initialization')
2644 ms_info
['initialization'] = initialization
2646 extract_Initialization(segment_template
)
2649 mpd_duration
= parse_duration(mpd_doc
.get('mediaPresentationDuration'))
2650 formats
, subtitles
= [], {}
2651 stream_numbers
= {'audio': 0, 'video': 0}
2652 for period
in mpd_doc
.findall(_add_ns('Period')):
2653 period_duration
= parse_duration(period
.get('duration')) or mpd_duration
2654 period_ms_info
= extract_multisegment_info(period
, {
2658 for adaptation_set
in period
.findall(_add_ns('AdaptationSet')):
2659 adaption_set_ms_info
= extract_multisegment_info(adaptation_set
, period_ms_info
)
2660 for representation
in adaptation_set
.findall(_add_ns('Representation')):
2661 representation_attrib
= adaptation_set
.attrib
.copy()
2662 representation_attrib
.update(representation
.attrib
)
2663 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2664 mime_type
= representation_attrib
['mimeType']
2665 content_type
= representation_attrib
.get('contentType', mime_type
.split('/')[0])
2667 codecs
= representation_attrib
.get('codecs', '')
2668 if content_type
not in ('video', 'audio', 'text'):
2669 if mime_type
== 'image/jpeg':
2670 content_type
= mime_type
2671 elif codecs
.split('.')[0] == 'stpp':
2672 content_type
= 'text'
2673 elif mimetype2ext(mime_type
) in ('tt', 'dfxp', 'ttml', 'xml', 'json'):
2674 content_type
= 'text'
2676 self
.report_warning('Unknown MIME type %s in DASH manifest' % mime_type
)
2680 for element
in (representation
, adaptation_set
, period
, mpd_doc
):
2681 base_url_e
= element
.find(_add_ns('BaseURL'))
2682 if base_url_e
is not None:
2683 base_url
= base_url_e
.text
+ base_url
2684 if re
.match(r
'^https?://', base_url
):
2686 if mpd_base_url
and base_url
.startswith('/'):
2687 base_url
= compat_urlparse
.urljoin(mpd_base_url
, base_url
)
2688 elif mpd_base_url
and not re
.match(r
'^https?://', base_url
):
2689 if not mpd_base_url
.endswith('/'):
2691 base_url
= mpd_base_url
+ base_url
2692 representation_id
= representation_attrib
.get('id')
2693 lang
= representation_attrib
.get('lang')
2694 url_el
= representation
.find(_add_ns('BaseURL'))
2695 filesize
= int_or_none(url_el
.attrib
.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el
is not None else None)
2696 bandwidth
= int_or_none(representation_attrib
.get('bandwidth'))
2697 if representation_id
is not None:
2698 format_id
= representation_id
2700 format_id
= content_type
2702 format_id
= mpd_id
+ '-' + format_id
2703 if content_type
in ('video', 'audio'):
2705 'format_id': format_id
,
2706 'manifest_url': mpd_url
,
2707 'ext': mimetype2ext(mime_type
),
2708 'width': int_or_none(representation_attrib
.get('width')),
2709 'height': int_or_none(representation_attrib
.get('height')),
2710 'tbr': float_or_none(bandwidth
, 1000),
2711 'asr': int_or_none(representation_attrib
.get('audioSamplingRate')),
2712 'fps': int_or_none(representation_attrib
.get('frameRate')),
2713 'language': lang
if lang
not in ('mul', 'und', 'zxx', 'mis') else None,
2714 'format_note': 'DASH %s' % content_type
,
2715 'filesize': filesize
,
2716 'container': mimetype2ext(mime_type
) + '_dash',
2717 'manifest_stream_number': stream_numbers
[content_type
]
2719 f
.update(parse_codecs(codecs
))
2720 stream_numbers
[content_type
] += 1
2721 elif content_type
== 'text':
2723 'ext': mimetype2ext(mime_type
),
2724 'manifest_url': mpd_url
,
2725 'filesize': filesize
,
2727 elif content_type
== 'image/jpeg':
2728 # See test case in VikiIE
2729 # https://www.viki.com/videos/1175236v-choosing-spouse-by-lottery-episode-1
2731 'format_id': format_id
,
2733 'manifest_url': mpd_url
,
2734 'format_note': 'DASH storyboards (jpeg)',
2738 if is_drm_protected(adaptation_set
) or is_drm_protected(representation
):
2740 representation_ms_info
= extract_multisegment_info(representation
, adaption_set_ms_info
)
2742 def prepare_template(template_name
, identifiers
):
2743 tmpl
= representation_ms_info
[template_name
]
2744 # First of, % characters outside $...$ templates
2745 # must be escaped by doubling for proper processing
2746 # by % operator string formatting used further (see
2747 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2753 in_template
= not in_template
2754 elif c
== '%' and not in_template
:
2756 # Next, $...$ templates are translated to their
2757 # %(...) counterparts to be used with % operator
2758 if representation_id
is not None:
2759 t
= t
.replace('$RepresentationID$', representation_id
)
2760 t
= re
.sub(r
'\$(%s)\$' % '|'.join(identifiers
), r
'%(\1)d', t
)
2761 t
= re
.sub(r
'\$(%s)%%([^$]+)\$' % '|'.join(identifiers
), r
'%(\1)\2', t
)
2762 t
.replace('$$', '$')
2765 # @initialization is a regular template like @media one
2766 # so it should be handled just the same way (see
2767 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2768 if 'initialization' in representation_ms_info
:
2769 initialization_template
= prepare_template(
2771 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2772 # $Time$ shall not be included for @initialization thus
2773 # only $Bandwidth$ remains
2775 representation_ms_info
['initialization_url'] = initialization_template
% {
2776 'Bandwidth': bandwidth
,
2779 def location_key(location
):
2780 return 'url' if re
.match(r
'^https?://', location
) else 'path'
2782 if 'segment_urls' not in representation_ms_info
and 'media' in representation_ms_info
:
2784 media_template
= prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2785 media_location_key
= location_key(media_template
)
2787 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2788 # can't be used at the same time
2789 if '%(Number' in media_template
and 's' not in representation_ms_info
:
2790 segment_duration
= None
2791 if 'total_number' not in representation_ms_info
and 'segment_duration' in representation_ms_info
:
2792 segment_duration
= float_or_none(representation_ms_info
['segment_duration'], representation_ms_info
['timescale'])
2793 representation_ms_info
['total_number'] = int(math
.ceil(float(period_duration
) / segment_duration
))
2794 representation_ms_info
['fragments'] = [{
2795 media_location_key
: media_template
% {
2796 'Number': segment_number
,
2797 'Bandwidth': bandwidth
,
2799 'duration': segment_duration
,
2800 } for segment_number
in range(
2801 representation_ms_info
['start_number'],
2802 representation_ms_info
['total_number'] + representation_ms_info
['start_number'])]
2804 # $Number*$ or $Time$ in media template with S list available
2805 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2806 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2807 representation_ms_info
['fragments'] = []
2810 segment_number
= representation_ms_info
['start_number']
2812 def add_segment_url():
2813 segment_url
= media_template
% {
2814 'Time': segment_time
,
2815 'Bandwidth': bandwidth
,
2816 'Number': segment_number
,
2818 representation_ms_info
['fragments'].append({
2819 media_location_key
: segment_url
,
2820 'duration': float_or_none(segment_d
, representation_ms_info
['timescale']),
2823 for num
, s
in enumerate(representation_ms_info
['s']):
2824 segment_time
= s
.get('t') or segment_time
2828 for r
in range(s
.get('r', 0)):
2829 segment_time
+= segment_d
2832 segment_time
+= segment_d
2833 elif 'segment_urls' in representation_ms_info
and 's' in representation_ms_info
:
2835 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2836 # or any YouTube dashsegments video
2839 timescale
= representation_ms_info
['timescale']
2840 for s
in representation_ms_info
['s']:
2841 duration
= float_or_none(s
['d'], timescale
)
2842 for r
in range(s
.get('r', 0) + 1):
2843 segment_uri
= representation_ms_info
['segment_urls'][segment_index
]
2845 location_key(segment_uri
): segment_uri
,
2846 'duration': duration
,
2849 representation_ms_info
['fragments'] = fragments
2850 elif 'segment_urls' in representation_ms_info
:
2851 # Segment URLs with no SegmentTimeline
2852 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2853 # https://github.com/ytdl-org/youtube-dl/pull/14844
2855 segment_duration
= float_or_none(
2856 representation_ms_info
['segment_duration'],
2857 representation_ms_info
['timescale']) if 'segment_duration' in representation_ms_info
else None
2858 for segment_url
in representation_ms_info
['segment_urls']:
2860 location_key(segment_url
): segment_url
,
2862 if segment_duration
:
2863 fragment
['duration'] = segment_duration
2864 fragments
.append(fragment
)
2865 representation_ms_info
['fragments'] = fragments
2866 # If there is a fragments key available then we correctly recognized fragmented media.
2867 # Otherwise we will assume unfragmented media with direct access. Technically, such
2868 # assumption is not necessarily correct since we may simply have no support for
2869 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2870 if 'fragments' in representation_ms_info
:
2872 # NB: mpd_url may be empty when MPD manifest is parsed from a string
2873 'url': mpd_url
or base_url
,
2874 'fragment_base_url': base_url
,
2876 'protocol': 'http_dash_segments' if mime_type
!= 'image/jpeg' else 'mhtml',
2878 if 'initialization_url' in representation_ms_info
:
2879 initialization_url
= representation_ms_info
['initialization_url']
2880 if not f
.get('url'):
2881 f
['url'] = initialization_url
2882 f
['fragments'].append({location_key(initialization_url): initialization_url}
)
2883 f
['fragments'].extend(representation_ms_info
['fragments'])
2885 # Assuming direct URL to unfragmented media.
2887 if content_type
in ('video', 'audio') or mime_type
== 'image/jpeg':
2889 elif content_type
== 'text':
2890 subtitles
.setdefault(lang
or 'und', []).append(f
)
2892 return formats
, subtitles
2894 def _extract_ism_formats(self
, *args
, **kwargs
):
2895 fmts
, subs
= self
._extract
_ism
_formats
_and
_subtitles
(*args
, **kwargs
)
2897 self
._report
_ignoring
_subs
('ISM')
2900 def _extract_ism_formats_and_subtitles(self
, ism_url
, video_id
, ism_id
=None, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
):
2901 res
= self
._download
_xml
_handle
(
2903 note
='Downloading ISM manifest' if note
is None else note
,
2904 errnote
='Failed to download ISM manifest' if errnote
is None else errnote
,
2905 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
2912 return self
._parse
_ism
_formats
_and
_subtitles
(ism_doc
, urlh
.geturl(), ism_id
)
2914 def _parse_ism_formats_and_subtitles(self
, ism_doc
, ism_url
, ism_id
=None):
2916 Parse formats from ISM manifest.
2918 1. [MS-SSTR]: Smooth Streaming Protocol,
2919 https://msdn.microsoft.com/en-us/library/ff469518.aspx
2921 if ism_doc
.get('IsLive') == 'TRUE':
2924 duration
= int(ism_doc
.attrib
['Duration'])
2925 timescale
= int_or_none(ism_doc
.get('TimeScale')) or 10000000
2929 for stream
in ism_doc
.findall('StreamIndex'):
2930 stream_type
= stream
.get('Type')
2931 if stream_type
not in ('video', 'audio', 'text'):
2933 url_pattern
= stream
.attrib
['Url']
2934 stream_timescale
= int_or_none(stream
.get('TimeScale')) or timescale
2935 stream_name
= stream
.get('Name')
2936 stream_language
= stream
.get('Language', 'und')
2937 for track
in stream
.findall('QualityLevel'):
2938 fourcc
= track
.get('FourCC') or ('AACL' if track
.get('AudioTag') == '255' else None)
2939 # TODO: add support for WVC1 and WMAP
2940 if fourcc
not in ('H264', 'AVC1', 'AACL', 'TTML'):
2941 self
.report_warning('%s is not a supported codec' % fourcc
)
2943 tbr
= int(track
.attrib
['Bitrate']) // 1000
2944 # [1] does not mention Width and Height attributes. However,
2945 # they're often present while MaxWidth and MaxHeight are
2946 # missing, so should be used as fallbacks
2947 width
= int_or_none(track
.get('MaxWidth') or track
.get('Width'))
2948 height
= int_or_none(track
.get('MaxHeight') or track
.get('Height'))
2949 sampling_rate
= int_or_none(track
.get('SamplingRate'))
2951 track_url_pattern
= re
.sub(r
'{[Bb]itrate}', track
.attrib
['Bitrate'], url_pattern
)
2952 track_url_pattern
= compat_urlparse
.urljoin(ism_url
, track_url_pattern
)
2958 stream_fragments
= stream
.findall('c')
2959 for stream_fragment_index
, stream_fragment
in enumerate(stream_fragments
):
2960 fragment_ctx
['time'] = int_or_none(stream_fragment
.get('t')) or fragment_ctx
['time']
2961 fragment_repeat
= int_or_none(stream_fragment
.get('r')) or 1
2962 fragment_ctx
['duration'] = int_or_none(stream_fragment
.get('d'))
2963 if not fragment_ctx
['duration']:
2965 next_fragment_time
= int(stream_fragment
[stream_fragment_index
+ 1].attrib
['t'])
2967 next_fragment_time
= duration
2968 fragment_ctx
['duration'] = (next_fragment_time
- fragment_ctx
['time']) / fragment_repeat
2969 for _
in range(fragment_repeat
):
2971 'url': re
.sub(r
'{start[ _]time}', compat_str(fragment_ctx
['time']), track_url_pattern
),
2972 'duration': fragment_ctx
['duration'] / stream_timescale
,
2974 fragment_ctx
['time'] += fragment_ctx
['duration']
2976 if stream_type
== 'text':
2977 subtitles
.setdefault(stream_language
, []).append({
2981 'manifest_url': ism_url
,
2982 'fragments': fragments
,
2983 '_download_params': {
2984 'stream_type': stream_type
,
2985 'duration': duration
,
2986 'timescale': stream_timescale
,
2988 'language': stream_language
,
2989 'codec_private_data': track
.get('CodecPrivateData'),
2992 elif stream_type
in ('video', 'audio'):
2994 'format_id': join_nonempty(ism_id
, stream_name
, tbr
),
2996 'manifest_url': ism_url
,
2997 'ext': 'ismv' if stream_type
== 'video' else 'isma',
3001 'asr': sampling_rate
,
3002 'vcodec': 'none' if stream_type
== 'audio' else fourcc
,
3003 'acodec': 'none' if stream_type
== 'video' else fourcc
,
3005 'fragments': fragments
,
3006 'has_drm': ism_doc
.find('Protection') is not None,
3007 '_download_params': {
3008 'stream_type': stream_type
,
3009 'duration': duration
,
3010 'timescale': stream_timescale
,
3011 'width': width
or 0,
3012 'height': height
or 0,
3014 'language': stream_language
,
3015 'codec_private_data': track
.get('CodecPrivateData'),
3016 'sampling_rate': sampling_rate
,
3017 'channels': int_or_none(track
.get('Channels', 2)),
3018 'bits_per_sample': int_or_none(track
.get('BitsPerSample', 16)),
3019 'nal_unit_length_field': int_or_none(track
.get('NALUnitLengthField', 4)),
3022 return formats
, subtitles
3024 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):
3025 def absolute_url(item_url
):
3026 return urljoin(base_url
, item_url
)
3028 def parse_content_type(content_type
):
3029 if not content_type
:
3031 ctr
= re
.search(r
'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type
)
3033 mimetype
, codecs
= ctr
.groups()
3034 f
= parse_codecs(codecs
)
3035 f
['ext'] = mimetype2ext(mimetype
)
3039 def _media_formats(src
, cur_media_type
, type_info
={}):
3040 full_url
= absolute_url(src
)
3041 ext
= type_info
.get('ext') or determine_ext(full_url
)
3043 is_plain_url
= False
3044 formats
= self
._extract
_m
3u8_formats
(
3045 full_url
, video_id
, ext
='mp4',
3046 entry_protocol
=m3u8_entry_protocol
, m3u8_id
=m3u8_id
,
3047 preference
=preference
, quality
=quality
, fatal
=False)
3049 is_plain_url
= False
3050 formats
= self
._extract
_mpd
_formats
(
3051 full_url
, video_id
, mpd_id
=mpd_id
, fatal
=False)
3056 'vcodec': 'none' if cur_media_type
== 'audio' else None,
3058 return is_plain_url
, formats
3061 # amp-video and amp-audio are very similar to their HTML5 counterparts
3062 # so we wll include them right here (see
3063 # https://www.ampproject.org/docs/reference/components/amp-video)
3064 # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/
3065 _MEDIA_TAG_NAME_RE
= r
'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)'
3066 media_tags
= [(media_tag
, media_tag_name
, media_type
, '')
3067 for media_tag
, media_tag_name
, media_type
3068 in re
.findall(r
'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE
, webpage
)]
3069 media_tags
.extend(re
.findall(
3070 # We only allow video|audio followed by a whitespace or '>'.
3071 # Allowing more characters may end up in significant slow down (see
3072 # https://github.com/ytdl-org/youtube-dl/issues/11979, example URL:
3073 # http://www.porntrex.com/maps/videositemap.xml).
3074 r
'(?s)(<(?P<tag>%s)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>' % _MEDIA_TAG_NAME_RE
, webpage
))
3075 for media_tag
, _
, media_type
, media_content
in media_tags
:
3080 media_attributes
= extract_attributes(media_tag
)
3081 src
= strip_or_none(media_attributes
.get('src'))
3083 _
, formats
= _media_formats(src
, media_type
)
3084 media_info
['formats'].extend(formats
)
3085 media_info
['thumbnail'] = absolute_url(media_attributes
.get('poster'))
3087 for source_tag
in re
.findall(r
'<source[^>]+>', media_content
):
3088 s_attr
= extract_attributes(source_tag
)
3089 # data-video-src and data-src are non standard but seen
3090 # several times in the wild
3091 src
= strip_or_none(dict_get(s_attr
, ('src', 'data-video-src', 'data-src')))
3094 f
= parse_content_type(s_attr
.get('type'))
3095 is_plain_url
, formats
= _media_formats(src
, media_type
, f
)
3097 # width, height, res, label and title attributes are
3098 # all not standard but seen several times in the wild
3101 for lbl
in ('label', 'title')
3102 if str_or_none(s_attr
.get(lbl
))
3104 width
= int_or_none(s_attr
.get('width'))
3105 height
= (int_or_none(s_attr
.get('height'))
3106 or int_or_none(s_attr
.get('res')))
3107 if not width
or not height
:
3109 resolution
= parse_resolution(lbl
)
3112 width
= width
or resolution
.get('width')
3113 height
= height
or resolution
.get('height')
3115 tbr
= parse_bitrate(lbl
)
3124 'format_id': s_attr
.get('label') or s_attr
.get('title'),
3126 f
.update(formats
[0])
3127 media_info
['formats'].append(f
)
3129 media_info
['formats'].extend(formats
)
3130 for track_tag
in re
.findall(r
'<track[^>]+>', media_content
):
3131 track_attributes
= extract_attributes(track_tag
)
3132 kind
= track_attributes
.get('kind')
3133 if not kind
or kind
in ('subtitles', 'captions'):
3134 src
= strip_or_none(track_attributes
.get('src'))
3137 lang
= track_attributes
.get('srclang') or track_attributes
.get('lang') or track_attributes
.get('label')
3138 media_info
['subtitles'].setdefault(lang
, []).append({
3139 'url': absolute_url(src
),
3141 for f
in media_info
['formats']:
3142 f
.setdefault('http_headers', {})['Referer'] = base_url
3143 if media_info
['formats'] or media_info
['subtitles']:
3144 entries
.append(media_info
)
3147 def _extract_akamai_formats(self
, *args
, **kwargs
):
3148 fmts
, subs
= self
._extract
_akamai
_formats
_and
_subtitles
(*args
, **kwargs
)
3150 self
._report
_ignoring
_subs
('akamai')
3153 def _extract_akamai_formats_and_subtitles(self
, manifest_url
, video_id
, hosts
={}):
3154 signed
= 'hdnea=' in manifest_url
3156 # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
3157 manifest_url
= re
.sub(
3158 r
'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
3159 '', manifest_url
).strip('?')
3164 hdcore_sign
= 'hdcore=3.7.0'
3165 f4m_url
= re
.sub(r
'(https?://[^/]+)/i/', r
'\1/z/', manifest_url
).replace('/master.m3u8', '/manifest.f4m')
3166 hds_host
= hosts
.get('hds')
3168 f4m_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hds_host
, f4m_url
)
3169 if 'hdcore=' not in f4m_url
:
3170 f4m_url
+= ('&' if '?' in f4m_url
else '?') + hdcore_sign
3171 f4m_formats
= self
._extract
_f
4m
_formats
(
3172 f4m_url
, video_id
, f4m_id
='hds', fatal
=False)
3173 for entry
in f4m_formats
:
3174 entry
.update({'extra_param_to_segment_url': hdcore_sign}
)
3175 formats
.extend(f4m_formats
)
3177 m3u8_url
= re
.sub(r
'(https?://[^/]+)/z/', r
'\1/i/', manifest_url
).replace('/manifest.f4m', '/master.m3u8')
3178 hls_host
= hosts
.get('hls')
3180 m3u8_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hls_host
, m3u8_url
)
3181 m3u8_formats
, m3u8_subtitles
= self
._extract
_m
3u8_formats
_and
_subtitles
(
3182 m3u8_url
, video_id
, 'mp4', 'm3u8_native',
3183 m3u8_id
='hls', fatal
=False)
3184 formats
.extend(m3u8_formats
)
3185 subtitles
= self
._merge
_subtitles
(subtitles
, m3u8_subtitles
)
3187 http_host
= hosts
.get('http')
3188 if http_host
and m3u8_formats
and not signed
:
3189 REPL_REGEX
= r
'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
3190 qualities
= re
.match(REPL_REGEX
, m3u8_url
).group(2).split(',')
3191 qualities_length
= len(qualities
)
3192 if len(m3u8_formats
) in (qualities_length
, qualities_length
+ 1):
3194 for f
in m3u8_formats
:
3195 if f
['vcodec'] != 'none':
3196 for protocol
in ('http', 'https'):
3198 del http_f
['manifest_url']
3200 REPL_REGEX
, protocol
+ r
'://%s/\g<1>%s\3' % (http_host
, qualities
[i
]), f
['url'])
3202 'format_id': http_f
['format_id'].replace('hls-', protocol
+ '-'),
3204 'protocol': protocol
,
3206 formats
.append(http_f
)
3209 return formats
, subtitles
3211 def _extract_wowza_formats(self
, url
, video_id
, m3u8_entry_protocol
='m3u8_native', skip_protocols
=[]):
3212 query
= compat_urlparse
.urlparse(url
).query
3213 url
= re
.sub(r
'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url
)
3215 r
'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url
)
3216 url_base
= mobj
.group('url')
3217 http_base_url
= '%s%s:%s' % ('http', mobj
.group('s') or '', url_base
)
3220 def manifest_url(manifest
):
3221 m_url
= '%s/%s' % (http_base_url
, manifest
)
3223 m_url
+= '?%s' % query
3226 if 'm3u8' not in skip_protocols
:
3227 formats
.extend(self
._extract
_m
3u8_formats
(
3228 manifest_url('playlist.m3u8'), video_id
, 'mp4',
3229 m3u8_entry_protocol
, m3u8_id
='hls', fatal
=False))
3230 if 'f4m' not in skip_protocols
:
3231 formats
.extend(self
._extract
_f
4m
_formats
(
3232 manifest_url('manifest.f4m'),
3233 video_id
, f4m_id
='hds', fatal
=False))
3234 if 'dash' not in skip_protocols
:
3235 formats
.extend(self
._extract
_mpd
_formats
(
3236 manifest_url('manifest.mpd'),
3237 video_id
, mpd_id
='dash', fatal
=False))
3238 if re
.search(r
'(?:/smil:|\.smil)', url_base
):
3239 if 'smil' not in skip_protocols
:
3240 rtmp_formats
= self
._extract
_smil
_formats
(
3241 manifest_url('jwplayer.smil'),
3242 video_id
, fatal
=False)
3243 for rtmp_format
in rtmp_formats
:
3244 rtsp_format
= rtmp_format
.copy()
3245 rtsp_format
['url'] = '%s/%s' % (rtmp_format
['url'], rtmp_format
['play_path'])
3246 del rtsp_format
['play_path']
3247 del rtsp_format
['ext']
3248 rtsp_format
.update({
3249 'url': rtsp_format
['url'].replace('rtmp://', 'rtsp://'),
3250 'format_id': rtmp_format
['format_id'].replace('rtmp', 'rtsp'),
3253 formats
.extend([rtmp_format
, rtsp_format
])
3255 for protocol
in ('rtmp', 'rtsp'):
3256 if protocol
not in skip_protocols
:
3258 'url': '%s:%s' % (protocol
, url_base
),
3259 'format_id': protocol
,
3260 'protocol': protocol
,
3264 def _find_jwplayer_data(self
, webpage
, video_id
=None, transform_source
=js_to_json
):
3266 r
'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P
=quote
)\
)(?
!</script
>).*?\
.setup\s
*\
((?P
<options
>[^
)]+)\
)',
3270 jwplayer_data = self._parse_json(mobj.group('options
'),
3272 transform_source=transform_source)
3273 except ExtractorError:
3276 if isinstance(jwplayer_data, dict):
3277 return jwplayer_data
3279 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
3280 jwplayer_data = self._find_jwplayer_data(
3281 webpage, video_id, transform_source=js_to_json)
3282 return self._parse_jwplayer_data(
3283 jwplayer_data, video_id, *args, **kwargs)
3285 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
3286 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3287 # JWPlayer backward compatibility: flattened playlists
3288 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
3289 if 'playlist
' not in jwplayer_data:
3290 jwplayer_data = {'playlist': [jwplayer_data]}
3294 # JWPlayer backward compatibility: single playlist item
3295 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
3296 if not isinstance(jwplayer_data['playlist
'], list):
3297 jwplayer_data['playlist
'] = [jwplayer_data['playlist
']]
3299 for video_data in jwplayer_data['playlist
']:
3300 # JWPlayer backward compatibility: flattened sources
3301 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
3302 if 'sources
' not in video_data:
3303 video_data['sources
'] = [video_data]
3305 this_video_id = video_id or video_data['mediaid
']
3307 formats = self._parse_jwplayer_formats(
3308 video_data['sources
'], video_id=this_video_id, m3u8_id=m3u8_id,
3309 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
3312 tracks = video_data.get('tracks
')
3313 if tracks and isinstance(tracks, list):
3314 for track in tracks:
3315 if not isinstance(track, dict):
3317 track_kind = track.get('kind
')
3318 if not track_kind or not isinstance(track_kind, compat_str):
3320 if track_kind.lower() not in ('captions
', 'subtitles
'):
3322 track_url = urljoin(base_url, track.get('file'))
3325 subtitles.setdefault(track.get('label
') or 'en
', []).append({
3326 'url
': self._proto_relative_url(track_url)
3330 'id': this_video_id,
3331 'title
': unescapeHTML(video_data['title
'] if require_title else video_data.get('title
')),
3332 'description
': clean_html(video_data.get('description
')),
3333 'thumbnail
': urljoin(base_url, self._proto_relative_url(video_data.get('image
'))),
3334 'timestamp
': int_or_none(video_data.get('pubdate
')),
3335 'duration
': float_or_none(jwplayer_data.get('duration
') or video_data.get('duration
')),
3336 'subtitles
': subtitles,
3338 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
3339 if len(formats) == 1 and re.search(r'^
(?
:http|
//).*(?
:youtube\
.com|youtu\
.be
)/.+', formats[0]['url
']):
3341 '_type
': 'url_transparent
',
3342 'url
': formats[0]['url
'],
3345 self._sort_formats(formats)
3346 entry['formats
'] = formats
3347 entries.append(entry)
3348 if len(entries) == 1:
3351 return self.playlist_result(entries)
3353 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
3354 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3357 for source in jwplayer_sources_data:
3358 if not isinstance(source, dict):
3360 source_url = urljoin(
3361 base_url, self._proto_relative_url(source.get('file')))
3362 if not source_url or source_url in urls:
3364 urls.append(source_url)
3365 source_type = source.get('type') or ''
3366 ext = mimetype2ext(source_type) or determine_ext(source_url)
3367 if source_type == 'hls
' or ext == 'm3u8
':
3368 formats.extend(self._extract_m3u8_formats(
3369 source_url, video_id, 'mp4
', entry_protocol='m3u8_native
',
3370 m3u8_id=m3u8_id, fatal=False))
3371 elif source_type == 'dash
' or ext == 'mpd
':
3372 formats.extend(self._extract_mpd_formats(
3373 source_url, video_id, mpd_id=mpd_id, fatal=False))
3375 formats.extend(self._extract_smil_formats(
3376 source_url, video_id, fatal=False))
3377 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3378 elif source_type.startswith('audio
') or ext in (
3379 'oga
', 'aac
', 'mp3
', 'mpeg
', 'vorbis
'):
3386 height = int_or_none(source.get('height
'))
3388 # Often no height is provided but there is a label in
3389 # format like "1080p", "720p SD", or 1080.
3390 height = int_or_none(self._search_regex(
3391 r'^
(\d{3,4}
)[pP
]?
(?
:\b|$
)', compat_str(source.get('label
') or ''),
3392 'height
', default=None))
3395 'width
': int_or_none(source.get('width
')),
3397 'tbr
': int_or_none(source.get('bitrate
')),
3400 if source_url.startswith('rtmp
'):
3401 a_format['ext
'] = 'flv
'
3402 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3403 # of jwplayer.flash.swf
3404 rtmp_url_parts = re.split(
3405 r'((?
:mp4|mp3|flv
):)', source_url, 1)
3406 if len(rtmp_url_parts) == 3:
3407 rtmp_url, prefix, play_path = rtmp_url_parts
3410 'play_path
': prefix + play_path,
3413 a_format.update(rtmp_params)
3414 formats.append(a_format)
3417 def _live_title(self, name):
3418 """ Generate the title for a live video """
3419 now = datetime.datetime.now()
3420 now_str = now.strftime('%Y
-%m
-%d %H
:%M
')
3421 return name + ' ' + now_str
3423 def _int(self, v, name, fatal=False, **kwargs):
3424 res = int_or_none(v, **kwargs)
3425 if 'get_attr
' in kwargs:
3426 print(getattr(v, kwargs['get_attr
']))
3428 msg = 'Failed to extract
%s: Could
not parse value
%r' % (name, v)
3430 raise ExtractorError(msg)
3432 self.report_warning(msg)
3435 def _float(self, v, name, fatal=False, **kwargs):
3436 res = float_or_none(v, **kwargs)
3438 msg = 'Failed to extract
%s: Could
not parse value
%r' % (name, v)
3440 raise ExtractorError(msg)
3442 self.report_warning(msg)
3445 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
3446 path='/', secure=False, discard=False, rest={}, **kwargs):
3447 cookie = compat_cookiejar_Cookie(
3448 0, name, value, port, port is not None, domain, True,
3449 domain.startswith('.'), path, True, secure, expire_time,
3450 discard, None, None, rest)
3451 self._downloader.cookiejar.set_cookie(cookie)
3453 def _get_cookies(self, url):
3454 """ Return a compat_cookies_SimpleCookie with the cookies for the url """
3455 req = sanitized_Request(url)
3456 self._downloader.cookiejar.add_cookie_header(req)
3457 return compat_cookies_SimpleCookie(req.get_header('Cookie
'))
3459 def _apply_first_set_cookie_header(self, url_handle, cookie):
3461 Apply first Set-Cookie header instead of the last. Experimental.
3463 Some sites (e.g. [1-3]) may serve two cookies under the same name
3464 in Set-Cookie header and expect the first (old) one to be set rather
3465 than second (new). However, as of RFC6265 the newer one cookie
3466 should be set into cookie store what actually happens.
3467 We will workaround this issue by resetting the cookie to
3468 the first one manually.
3469 1. https://new.vk.com/
3470 2. https://github.com/ytdl-org/youtube-dl/issues/9841#issuecomment-227871201
3471 3. https://learning.oreilly.com/
3473 for header, cookies in url_handle.headers.items():
3474 if header.lower() != 'set-cookie
':
3476 if sys.version_info[0] >= 3:
3477 cookies = cookies.encode('iso
-8859-1')
3478 cookies = cookies.decode('utf
-8')
3479 cookie_value = re.search(
3480 r'%s=(.+?
);.*?
\b[Dd
]omain
=(.+?
)(?
:[,;]|$
)' % cookie, cookies)
3482 value, domain = cookie_value.groups()
3483 self._set_cookie(domain, cookie, value)
3486 def get_testcases(self, include_onlymatching=False):
3487 t = getattr(self, '_TEST
', None)
3489 assert not hasattr(self, '_TESTS
'), \
3490 '%s has _TEST
and _TESTS
' % type(self).__name__
3493 tests = getattr(self, '_TESTS
', [])
3495 if not include_onlymatching and t.get('only_matching
', False):
3497 t['name
'] = type(self).__name__[:-len('IE
')]
3500 def is_suitable(self, age_limit):
3501 """ Test whether the extractor is generally suitable for the given
3502 age limit (i.e. pornographic sites are not, all others usually are) """
3504 any_restricted = False
3505 for tc in self.get_testcases(include_onlymatching=False):
3506 if tc.get('playlist
', []):
3507 tc = tc['playlist
'][0]
3508 is_restricted = age_restricted(
3509 tc.get('info_dict
', {}).get('age_limit
'), age_limit)
3510 if not is_restricted:
3512 any_restricted = any_restricted or is_restricted
3513 return not any_restricted
3515 def extract_subtitles(self, *args, **kwargs):
3516 if (self.get_param('writesubtitles
', False)
3517 or self.get_param('listsubtitles
')):
3518 return self._get_subtitles(*args, **kwargs)
3521 def _get_subtitles(self, *args, **kwargs):
3522 raise NotImplementedError('This method must be implemented by subclasses
')
3524 def extract_comments(self, *args, **kwargs):
3525 if not self.get_param('getcomments
'):
3527 generator = self._get_comments(*args, **kwargs)
3533 comments.append(next(generator))
3534 except KeyboardInterrupt:
3536 self.to_screen('Interrupted by user
')
3537 except StopIteration:
3539 comment_count = len(comments)
3540 self.to_screen(f'Extracted {comment_count} comments
')
3542 'comments
': comments,
3543 'comment_count
': None if interrupted else comment_count
3547 def _get_comments(self, *args, **kwargs):
3548 raise NotImplementedError('This method must be implemented by subclasses
')
3551 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
3552 """ Merge subtitle items for one language. Items with duplicated URLs
3553 will be dropped. """
3554 list1_urls = set([item['url
'] for item in subtitle_list1])
3555 ret = list(subtitle_list1)
3556 ret.extend([item for item in subtitle_list2 if item['url
'] not in list1_urls])
3560 def _merge_subtitles(cls, *dicts, target=None):
3561 """ Merge subtitle dictionaries, language by language. """
3565 for lang, subs in d.items():
3566 target[lang] = cls._merge_subtitle_items(target.get(lang, []), subs)
3569 def extract_automatic_captions(self, *args, **kwargs):
3570 if (self.get_param('writeautomaticsub
', False)
3571 or self.get_param('listsubtitles
')):
3572 return self._get_automatic_captions(*args, **kwargs)
3575 def _get_automatic_captions(self, *args, **kwargs):
3576 raise NotImplementedError('This method must be implemented by subclasses
')
3578 def mark_watched(self, *args, **kwargs):
3579 if not self.get_param('mark_watched
', False):
3581 if (self._get_login_info()[0] is not None
3582 or self.get_param('cookiefile
')
3583 or self.get_param('cookiesfrombrowser
')):
3584 self._mark_watched(*args, **kwargs)
3586 def _mark_watched(self, *args, **kwargs):
3587 raise NotImplementedError('This method must be implemented by subclasses
')
3589 def geo_verification_headers(self):
3591 geo_verification_proxy = self.get_param('geo_verification_proxy
')
3592 if geo_verification_proxy:
3593 headers['Ytdl
-request
-proxy
'] = geo_verification_proxy
3596 def _generic_id(self, url):
3597 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
3599 def _generic_title(self, url):
3600 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
3603 def _availability(is_private=None, needs_premium=None, needs_subscription=None, needs_auth=None, is_unlisted=None):
3604 all_known = all(map(
3605 lambda x: x is not None,
3606 (is_private, needs_premium, needs_subscription, needs_auth, is_unlisted)))
3608 'private
' if is_private
3609 else 'premium_only
' if needs_premium
3610 else 'subscriber_only
' if needs_subscription
3611 else 'needs_auth
' if needs_auth
3612 else 'unlisted
' if is_unlisted
3613 else 'public
' if all_known
3616 def _configuration_arg(self, key, default=NO_DEFAULT, casesense=False):
3618 @returns A list of values for the extractor argument given by "key"
3619 or "default" if no such key is present
3620 @param default The default value to return when the key is not present (default: [])
3621 @param casesense When false, the values are converted to lower case
3624 self._downloader.params, ('extractor_args
', self.ie_key().lower(), key))
3626 return [] if default is NO_DEFAULT else default
3627 return list(val) if casesense else [x.lower() for x in val]
3630 class SearchInfoExtractor(InfoExtractor):
3632 Base class for paged search queries extractors.
3633 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
3634 Instances should define _SEARCH_KEY and optionally _MAX_RESULTS
3637 _MAX_RESULTS = float('inf
')
3640 def _make_valid_url(cls):
3641 return r'%s(?P
<prefix
>|
[1-9][0-9]*|all
):(?P
<query
>[\s\S
]+)' % cls._SEARCH_KEY
3644 def suitable(cls, url):
3645 return re.match(cls._make_valid_url(), url) is not None
3647 def _real_extract(self, query):
3648 mobj = re.match(self._make_valid_url(), query)
3650 raise ExtractorError('Invalid search query
"%s"' % query)
3652 prefix = mobj.group('prefix
')
3653 query = mobj.group('query
')
3655 return self._get_n_results(query, 1)
3656 elif prefix == 'all
':
3657 return self._get_n_results(query, self._MAX_RESULTS)
3661 raise ExtractorError('invalid download number
%s for query
"%s"' % (n, query))
3662 elif n > self._MAX_RESULTS:
3663 self.report_warning('%s returns
max %i results (you requested
%i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3664 n = self._MAX_RESULTS
3665 return self._get_n_results(query, n)
3667 def _get_n_results(self, query, n):
3668 """Get a specified number of results for a query.
3669 Either this function or _search_results must be overridden by subclasses """
3670 return self.playlist_result(
3671 itertools.islice(self._search_results(query), 0, None if n == float('inf
') else n),
3674 def _search_results(self, query):
3675 """Returns an iterator of search results"""
3676 raise NotImplementedError('This method must be implemented by subclasses
')
3679 def SEARCH_KEY(self):
3680 return self._SEARCH_KEY