2 from __future__
import unicode_literals
18 from ..compat
import (
19 compat_cookiejar_Cookie
,
20 compat_cookies_SimpleCookie
,
22 compat_etree_fromstring
,
29 compat_urllib_parse_unquote
,
30 compat_urllib_parse_urlencode
,
31 compat_urllib_request
,
33 compat_xml_parse_error
,
35 from ..downloader
import FileDownloader
36 from ..downloader
.f4m
import (
38 remove_encrypted_media
,
69 parse_m3u8_attributes
,
94 class InfoExtractor(object):
95 """Information Extractor class.
97 Information extractors are the classes that, given a URL, extract
98 information about the video (or videos) the URL refers to. This
99 information includes the real video URL, the video title, author and
100 others. The information is stored in a dictionary which is then
101 passed to the YoutubeDL. The YoutubeDL processes this
102 information possibly downloading the video to the file system, among
103 other possible outcomes.
105 The type field determines the type of the result.
106 By far the most common value (and the default if _type is missing) is
107 "video", which indicates a single video.
109 For a video, the dictionaries must include the following fields:
111 id: Video identifier.
112 title: Video title, unescaped.
114 Additionally, it must contain either a formats entry or a url one:
116 formats: A list of dictionaries for each format available, ordered
117 from worst to best quality.
120 * url The mandatory URL representing the media:
121 for plain file media - HTTP URL of this file,
123 for HLS - URL of the M3U8 media playlist,
124 for HDS - URL of the F4M manifest,
126 - HTTP URL to plain file media (in case of
128 - URL of the MPD manifest or base URL
129 representing the media if MPD manifest
130 is parsed from a string (in case of
132 for MSS - URL of the ISM manifest.
134 The URL of the manifest file in case of
136 for HLS - URL of the M3U8 master playlist,
137 for HDS - URL of the F4M manifest,
138 for DASH - URL of the MPD manifest,
139 for MSS - URL of the ISM manifest.
140 * ext Will be calculated from URL if missing
141 * format A human-readable description of the format
142 ("mp4 container with h264/opus").
143 Calculated from the format_id, width, height.
144 and format_note fields if missing.
145 * format_id A short description of the format
146 ("mp4_h264_opus" or "19").
147 Technically optional, but strongly recommended.
148 * format_note Additional info about the format
149 ("3D" or "DASH video")
150 * width Width of the video, if known
151 * height Height of the video, if known
152 * resolution Textual description of width and height
153 * dynamic_range The dynamic range of the video. One of:
154 "SDR" (None), "HDR10", "HDR10+, "HDR12", "HLG, "DV"
155 * tbr Average bitrate of audio and video in KBit/s
156 * abr Average audio bitrate in KBit/s
157 * acodec Name of the audio codec in use
158 * asr Audio sampling rate in Hertz
159 * vbr Average video bitrate in KBit/s
161 * vcodec Name of the video codec in use
162 * container Name of the container format
163 * filesize The number of bytes, if known in advance
164 * filesize_approx An estimate for the number of bytes
165 * player_url SWF Player URL (used for rtmpdump).
166 * protocol The protocol that will be used for the actual
167 download, lower-case.
168 "http", "https", "rtsp", "rtmp", "rtmp_ffmpeg", "rtmpe",
169 "m3u8", "m3u8_native" or "http_dash_segments".
171 Base URL for fragments. Each fragment's path
172 value (if present) will be relative to
174 * fragments A list of fragments of a fragmented media.
175 Each fragment entry must contain either an url
176 or a path. If an url is present it should be
177 considered by a client. Otherwise both path and
178 fragment_base_url must be present. Here is
179 the list of all potential fields:
180 * "url" - fragment's URL
181 * "path" - fragment's path relative to
183 * "duration" (optional, int or float)
184 * "filesize" (optional, int)
185 * preference Order number of this format. If this field is
186 present and not None, the formats get sorted
187 by this field, regardless of all other values.
188 -1 for default (order by other properties),
189 -2 or smaller for less than default.
190 < -1000 to hide the format (if there is
191 another one which is strictly better)
192 * language Language code, e.g. "de" or "en-US".
193 * language_preference Is this in the language mentioned in
195 10 if it's what the URL is about,
196 -1 for default (don't know),
197 -10 otherwise, other values reserved for now.
198 * quality Order number of the video quality of this
199 format, irrespective of the file format.
200 -1 for default (order by other properties),
201 -2 or smaller for less than default.
202 * source_preference Order number for this video source
203 (quality takes higher priority)
204 -1 for default (order by other properties),
205 -2 or smaller for less than default.
206 * http_headers A dictionary of additional HTTP headers
207 to add to the request.
208 * stretched_ratio If given and not 1, indicates that the
209 video's pixels are not square.
210 width : height ratio as float.
211 * no_resume The server does not support resuming the
212 (HTTP or RTMP) download. Boolean.
213 * has_drm The format has DRM and cannot be downloaded. Boolean
214 * downloader_options A dictionary of downloader options as
215 described in FileDownloader
216 RTMP formats can also have the additional fields: page_url,
217 app, play_path, tc_url, flash_version, rtmp_live, rtmp_conn,
218 rtmp_protocol, rtmp_real_time
220 url: Final video URL.
221 ext: Video filename extension.
222 format: The video format, defaults to ext (used for --get-format)
223 player_url: SWF Player URL (used for rtmpdump).
225 The following fields are optional:
227 alt_title: A secondary title of the video.
228 display_id An alternative identifier for the video, not necessarily
229 unique, but available before title. Typically, id is
230 something like "4234987", title "Dancing naked mole rats",
231 and display_id "dancing-naked-mole-rats"
232 thumbnails: A list of dictionaries, with the following entries:
233 * "id" (optional, string) - Thumbnail format ID
235 * "preference" (optional, int) - quality of the image
236 * "width" (optional, int)
237 * "height" (optional, int)
238 * "resolution" (optional, string "{width}x{height}",
240 * "filesize" (optional, int)
241 thumbnail: Full URL to a video thumbnail image.
242 description: Full video description.
243 uploader: Full name of the video uploader.
244 license: License name the video is licensed under.
245 creator: The creator of the video.
246 release_timestamp: UNIX timestamp of the moment the video was released.
247 release_date: The date (YYYYMMDD) when the video was released.
248 timestamp: UNIX timestamp of the moment the video was uploaded
249 upload_date: Video upload date (YYYYMMDD).
250 If not explicitly set, calculated from timestamp.
251 uploader_id: Nickname or id of the video uploader.
252 uploader_url: Full URL to a personal webpage of the video uploader.
253 channel: Full name of the channel the video is uploaded on.
254 Note that channel fields may or may not repeat uploader
255 fields. This depends on a particular extractor.
256 channel_id: Id of the channel.
257 channel_url: Full URL to a channel webpage.
258 location: Physical location where the video was filmed.
259 subtitles: The available subtitles as a dictionary in the format
260 {tag: subformats}. "tag" is usually a language code, and
261 "subformats" is a list sorted from lower to higher
262 preference, each element is a dictionary with the "ext"
264 * "data": The subtitles file contents
265 * "url": A URL pointing to the subtitles file
266 It can optionally also have:
267 * "name": Name or description of the subtitles
268 "ext" will be calculated from URL if missing
269 automatic_captions: Like 'subtitles'; contains automatically generated
270 captions instead of normal subtitles
271 duration: Length of the video in seconds, as an integer or float.
272 view_count: How many users have watched the video on the platform.
273 like_count: Number of positive ratings of the video
274 dislike_count: Number of negative ratings of the video
275 repost_count: Number of reposts of the video
276 average_rating: Average rating give by users, the scale used depends on the webpage
277 comment_count: Number of comments on the video
278 comments: A list of comments, each with one or more of the following
279 properties (all but one of text or html optional):
280 * "author" - human-readable name of the comment author
281 * "author_id" - user ID of the comment author
282 * "author_thumbnail" - The thumbnail of the comment author
284 * "html" - Comment as HTML
285 * "text" - Plain text of the comment
286 * "timestamp" - UNIX timestamp of comment
287 * "parent" - ID of the comment this one is replying to.
288 Set to "root" to indicate that this is a
289 comment to the original video.
290 * "like_count" - Number of positive ratings of the comment
291 * "dislike_count" - Number of negative ratings of the comment
292 * "is_favorited" - Whether the comment is marked as
293 favorite by the video uploader
294 * "author_is_uploader" - Whether the comment is made by
296 age_limit: Age restriction for the video, as an integer (years)
297 webpage_url: The URL to the video webpage, if given to yt-dlp it
298 should allow to get the same result again. (It will be set
299 by YoutubeDL if it's missing)
300 categories: A list of categories that the video falls in, for example
302 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
303 cast: A list of the video cast
304 is_live: True, False, or None (=unknown). Whether this video is a
305 live stream that goes on instead of a fixed-length video.
306 was_live: True, False, or None (=unknown). Whether this video was
307 originally a live stream.
308 live_status: 'is_live', 'is_upcoming', 'was_live', 'not_live' or None (=unknown)
309 If absent, automatically set from is_live, was_live
310 start_time: Time in seconds where the reproduction should start, as
311 specified in the URL.
312 end_time: Time in seconds where the reproduction should end, as
313 specified in the URL.
314 chapters: A list of dictionaries, with the following entries:
315 * "start_time" - The start time of the chapter in seconds
316 * "end_time" - The end time of the chapter in seconds
317 * "title" (optional, string)
318 playable_in_embed: Whether this video is allowed to play in embedded
319 players on other sites. Can be True (=always allowed),
320 False (=never allowed), None (=unknown), or a string
321 specifying the criteria for embedability (Eg: 'whitelist')
322 availability: Under what condition the video is available. One of
323 'private', 'premium_only', 'subscriber_only', 'needs_auth',
324 'unlisted' or 'public'. Use 'InfoExtractor._availability'
326 __post_extractor: A function to be called just before the metadata is
327 written to either disk, logger or console. The function
328 must return a dict which will be added to the info_dict.
329 This is usefull for additional information that is
330 time-consuming to extract. Note that the fields thus
331 extracted will not be available to output template and
332 match_filter. So, only "comments" and "comment_count" are
333 currently allowed to be extracted via this method.
335 The following fields should only be used when the video belongs to some logical
338 chapter: Name or title of the chapter the video belongs to.
339 chapter_number: Number of the chapter the video belongs to, as an integer.
340 chapter_id: Id of the chapter the video belongs to, as a unicode string.
342 The following fields should only be used when the video is an episode of some
343 series, programme or podcast:
345 series: Title of the series or programme the video episode belongs to.
346 series_id: Id of the series or programme the video episode belongs to, as a unicode string.
347 season: Title of the season the video episode belongs to.
348 season_number: Number of the season the video episode belongs to, as an integer.
349 season_id: Id of the season the video episode belongs to, as a unicode string.
350 episode: Title of the video episode. Unlike mandatory video title field,
351 this field should denote the exact title of the video episode
352 without any kind of decoration.
353 episode_number: Number of the video episode within a season, as an integer.
354 episode_id: Id of the video episode, as a unicode string.
356 The following fields should only be used when the media is a track or a part of
359 track: Title of the track.
360 track_number: Number of the track within an album or a disc, as an integer.
361 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
363 artist: Artist(s) of the track.
364 genre: Genre(s) of the track.
365 album: Title of the album the track belongs to.
366 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
367 album_artist: List of all artists appeared on the album (e.g.
368 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
370 disc_number: Number of the disc or other physical medium the track belongs to,
372 release_year: Year (YYYY) when the album was released.
374 Unless mentioned otherwise, the fields should be Unicode strings.
376 Unless mentioned otherwise, None is equivalent to absence of information.
379 _type "playlist" indicates multiple videos.
380 There must be a key "entries", which is a list, an iterable, or a PagedList
381 object, each element of which is a valid dictionary by this specification.
383 Additionally, playlists can have "id", "title", and any other relevent
384 attributes with the same semantics as videos (see above).
387 _type "multi_video" indicates that there are multiple videos that
388 form a single show, for examples multiple acts of an opera or TV episode.
389 It must have an entries key like a playlist and contain all the keys
390 required for a video at the same time.
393 _type "url" indicates that the video must be extracted from another
394 location, possibly by a different extractor. Its only required key is:
395 "url" - the next URL to extract.
396 The key "ie_key" can be set to the class name (minus the trailing "IE",
397 e.g. "Youtube") if the extractor class is known in advance.
398 Additionally, the dictionary may have any properties of the resolved entity
399 known in advance, for example "title" if the title of the referred video is
403 _type "url_transparent" entities have the same specification as "url", but
404 indicate that the given additional information is more precise than the one
405 associated with the resolved URL.
406 This is useful when a site employs a video service that hosts the video and
407 its technical metadata, but that video service does not embed a useful
408 title, description etc.
411 Subclasses of this one should re-define the _real_initialize() and
412 _real_extract() methods and define a _VALID_URL regexp.
413 Probably, they should also be added to the list of extractors.
415 Subclasses may also override suitable() if necessary, but ensure the function
416 signature is preserved and that this function imports everything it needs
417 (except other extractors), so that lazy_extractors works correctly
419 _GEO_BYPASS attribute may be set to False in order to disable
420 geo restriction bypass mechanisms for a particular extractor.
421 Though it won't disable explicit geo restriction bypass based on
422 country code provided with geo_bypass_country.
424 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
425 countries for this extractor. One of these countries will be used by
426 geo restriction bypass mechanism right away in order to bypass
427 geo restriction, of course, if the mechanism is not disabled.
429 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
430 IP blocks in CIDR notation for this extractor. One of these IP blocks
431 will be used by geo restriction bypass mechanism similarly
434 The _WORKING attribute should be set to False for broken IEs
435 in order to warn the users and skip the tests.
440 _x_forwarded_for_ip
= None
442 _GEO_COUNTRIES
= None
443 _GEO_IP_BLOCKS
= None
447 'any': 'Use --cookies, --username and --password, or --netrc to provide account credentials',
449 'Use --cookies-from-browser or --cookies for the authentication. '
450 'See https://github.com/ytdl-org/youtube-dl#how-do-i-pass-cookies-to-youtube-dl for how to manually pass cookies'),
451 'password': 'Use --username and --password, or --netrc to provide account credentials',
454 def __init__(self
, downloader
=None):
455 """Constructor. Receives an optional downloader (a YoutubeDL instance).
456 If a downloader is not passed during initialization,
457 it must be set using "set_downloader()" before "extract()" is called"""
459 self
._x
_forwarded
_for
_ip
= None
460 self
._printed
_messages
= set()
461 self
.set_downloader(downloader
)
464 def _match_valid_url(cls
, url
):
465 # This does not use has/getattr intentionally - we want to know whether
466 # we have cached the regexp for *this* class, whereas getattr would also
467 # match the superclass
468 if '_VALID_URL_RE' not in cls
.__dict
__:
469 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
470 return cls
._VALID
_URL
_RE
.match(url
)
473 def suitable(cls
, url
):
474 """Receives a URL and returns True if suitable for this IE."""
475 # This function must import everything it needs (except other extractors),
476 # so that lazy_extractors works correctly
477 return cls
._match
_valid
_url
(url
) is not None
480 def _match_id(cls
, url
):
481 return cls
._match
_valid
_url
(url
).group('id')
484 def get_temp_id(cls
, url
):
486 return cls
._match
_id
(url
)
487 except (IndexError, AttributeError):
492 """Getter method for _WORKING."""
495 def initialize(self
):
496 """Initializes an instance (authentication, etc)."""
497 self
._printed
_messages
= set()
498 self
._initialize
_geo
_bypass
({
499 'countries': self
._GEO
_COUNTRIES
,
500 'ip_blocks': self
._GEO
_IP
_BLOCKS
,
503 self
._real
_initialize
()
506 def _initialize_geo_bypass(self
, geo_bypass_context
):
508 Initialize geo restriction bypass mechanism.
510 This method is used to initialize geo bypass mechanism based on faking
511 X-Forwarded-For HTTP header. A random country from provided country list
512 is selected and a random IP belonging to this country is generated. This
513 IP will be passed as X-Forwarded-For HTTP header in all subsequent
516 This method will be used for initial geo bypass mechanism initialization
517 during the instance initialization with _GEO_COUNTRIES and
520 You may also manually call it from extractor's code if geo bypass
521 information is not available beforehand (e.g. obtained during
522 extraction) or due to some other reason. In this case you should pass
523 this information in geo bypass context passed as first argument. It may
524 contain following fields:
526 countries: List of geo unrestricted countries (similar
528 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
529 (similar to _GEO_IP_BLOCKS)
532 if not self
._x
_forwarded
_for
_ip
:
534 # Geo bypass mechanism is explicitly disabled by user
535 if not self
.get_param('geo_bypass', True):
538 if not geo_bypass_context
:
539 geo_bypass_context
= {}
541 # Backward compatibility: previously _initialize_geo_bypass
542 # expected a list of countries, some 3rd party code may still use
544 if isinstance(geo_bypass_context
, (list, tuple)):
545 geo_bypass_context
= {
546 'countries': geo_bypass_context
,
549 # The whole point of geo bypass mechanism is to fake IP
550 # as X-Forwarded-For HTTP header based on some IP block or
553 # Path 1: bypassing based on IP block in CIDR notation
555 # Explicit IP block specified by user, use it right away
556 # regardless of whether extractor is geo bypassable or not
557 ip_block
= self
.get_param('geo_bypass_ip_block', None)
559 # Otherwise use random IP block from geo bypass context but only
560 # if extractor is known as geo bypassable
562 ip_blocks
= geo_bypass_context
.get('ip_blocks')
563 if self
._GEO
_BYPASS
and ip_blocks
:
564 ip_block
= random
.choice(ip_blocks
)
567 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(ip_block
)
568 self
._downloader
.write_debug(
569 '[debug] Using fake IP %s as X-Forwarded-For' % self
._x
_forwarded
_for
_ip
)
572 # Path 2: bypassing based on country code
574 # Explicit country code specified by user, use it right away
575 # regardless of whether extractor is geo bypassable or not
576 country
= self
.get_param('geo_bypass_country', None)
578 # Otherwise use random country code from geo bypass context but
579 # only if extractor is known as geo bypassable
581 countries
= geo_bypass_context
.get('countries')
582 if self
._GEO
_BYPASS
and countries
:
583 country
= random
.choice(countries
)
586 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country
)
587 self
._downloader
.write_debug(
588 'Using fake IP %s (%s) as X-Forwarded-For' % (self
._x
_forwarded
_for
_ip
, country
.upper()))
590 def extract(self
, url
):
591 """Extracts URL information and returns it in list of dicts."""
596 self
.write_debug('Extracting URL: %s' % url
)
597 ie_result
= self
._real
_extract
(url
)
598 if ie_result
is None:
600 if self
._x
_forwarded
_for
_ip
:
601 ie_result
['__x_forwarded_for_ip'] = self
._x
_forwarded
_for
_ip
602 subtitles
= ie_result
.get('subtitles')
603 if (subtitles
and 'live_chat' in subtitles
604 and 'no-live-chat' in self
.get_param('compat_opts', [])):
605 del subtitles
['live_chat']
607 except GeoRestrictedError
as e
:
608 if self
.__maybe
_fake
_ip
_and
_retry
(e
.countries
):
611 except UnsupportedError
:
613 except ExtractorError
as e
:
615 'video_id': e
.video_id
or self
.get_temp_id(url
),
618 'expected': e
.expected
,
621 if hasattr(e
, 'countries'):
622 kwargs
['countries'] = e
.countries
623 raise type(e
)(e
.msg
, **kwargs
)
624 except compat_http_client
.IncompleteRead
as e
:
625 raise ExtractorError('A network error has occurred.', cause
=e
, expected
=True, video_id
=self
.get_temp_id(url
))
626 except (KeyError, StopIteration) as e
:
627 raise ExtractorError('An extractor error has occurred.', cause
=e
, video_id
=self
.get_temp_id(url
))
629 def __maybe_fake_ip_and_retry(self
, countries
):
630 if (not self
.get_param('geo_bypass_country', None)
632 and self
.get_param('geo_bypass', True)
633 and not self
._x
_forwarded
_for
_ip
635 country_code
= random
.choice(countries
)
636 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country_code
)
637 if self
._x
_forwarded
_for
_ip
:
639 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
640 % (self
._x
_forwarded
_for
_ip
, country_code
.upper()))
644 def set_downloader(self
, downloader
):
645 """Sets the downloader for this IE."""
646 self
._downloader
= downloader
648 def _real_initialize(self
):
649 """Real initialization process. Redefine in subclasses."""
652 def _real_extract(self
, url
):
653 """Real extraction process. Redefine in subclasses."""
658 """A string for getting the InfoExtractor with get_info_extractor"""
659 return cls
.__name
__[:-2]
663 return compat_str(type(self
).__name
__[:-2])
666 def __can_accept_status_code(err
, expected_status
):
667 assert isinstance(err
, compat_urllib_error
.HTTPError
)
668 if expected_status
is None:
670 elif callable(expected_status
):
671 return expected_status(err
.code
) is True
673 return err
.code
in variadic(expected_status
)
675 def _request_webpage(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
, expected_status
=None):
677 Return the response handle.
679 See _download_webpage docstring for arguments specification.
681 if not self
._downloader
._first
_webpage
_request
:
682 sleep_interval
= self
.get_param('sleep_interval_requests') or 0
683 if sleep_interval
> 0:
684 self
.to_screen('Sleeping %s seconds ...' % sleep_interval
)
685 time
.sleep(sleep_interval
)
687 self
._downloader
._first
_webpage
_request
= False
690 self
.report_download_webpage(video_id
)
691 elif note
is not False:
693 self
.to_screen('%s' % (note
,))
695 self
.to_screen('%s: %s' % (video_id
, note
))
697 # Some sites check X-Forwarded-For HTTP header in order to figure out
698 # the origin of the client behind proxy. This allows bypassing geo
699 # restriction by faking this header's value to IP that belongs to some
700 # geo unrestricted country. We will do so once we encounter any
701 # geo restriction error.
702 if self
._x
_forwarded
_for
_ip
:
703 if 'X-Forwarded-For' not in headers
:
704 headers
['X-Forwarded-For'] = self
._x
_forwarded
_for
_ip
706 if isinstance(url_or_request
, compat_urllib_request
.Request
):
707 url_or_request
= update_Request(
708 url_or_request
, data
=data
, headers
=headers
, query
=query
)
711 url_or_request
= update_url_query(url_or_request
, query
)
712 if data
is not None or headers
:
713 url_or_request
= sanitized_Request(url_or_request
, data
, headers
)
715 return self
._downloader
.urlopen(url_or_request
)
716 except network_exceptions
as err
:
717 if isinstance(err
, compat_urllib_error
.HTTPError
):
718 if self
.__can
_accept
_status
_code
(err
, expected_status
):
719 # Retain reference to error to prevent file object from
720 # being closed before it can be read. Works around the
721 # effects of <https://bugs.python.org/issue15002>
722 # introduced in Python 3.4.1.
729 errnote
= 'Unable to download webpage'
731 errmsg
= '%s: %s' % (errnote
, error_to_compat_str(err
))
733 raise ExtractorError(errmsg
, sys
.exc_info()[2], cause
=err
)
735 self
.report_warning(errmsg
)
738 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):
740 Return a tuple (page content as string, URL handle).
742 See _download_webpage docstring for arguments specification.
744 # Strip hashes from the URL (#1038)
745 if isinstance(url_or_request
, (compat_str
, str)):
746 url_or_request
= url_or_request
.partition('#')[0]
748 urlh
= self
._request
_webpage
(url_or_request
, video_id
, note
, errnote
, fatal
, data
=data
, headers
=headers
, query
=query
, expected_status
=expected_status
)
752 content
= self
._webpage
_read
_content
(urlh
, url_or_request
, video_id
, note
, errnote
, fatal
, encoding
=encoding
)
753 return (content
, urlh
)
756 def _guess_encoding_from_content(content_type
, webpage_bytes
):
757 m
= re
.match(r
'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type
)
759 encoding
= m
.group(1)
761 m
= re
.search(br
'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
762 webpage_bytes[:1024])
764 encoding = m.group(1).decode('ascii')
765 elif webpage_bytes.startswith(b'\xff\xfe'):
772 def __check_blocked(self, content):
773 first_block = content[:512]
774 if ('<title>Access to this site is blocked</title>' in content
775 and 'Websense' in first_block):
776 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
777 blocked_iframe = self._html_search_regex(
778 r'<iframe src="([^
"]+)"', content,
779 'Websense information URL
', default=None)
781 msg += ' Visit
%s for more details
' % blocked_iframe
782 raise ExtractorError(msg, expected=True)
783 if '<title
>The URL you requested has been blocked
</title
>' in first_block:
785 'Access to this webpage has been blocked by Indian censorship
. '
786 'Use a VPN
or proxy
server (with --proxy
) to route around it
.')
787 block_msg = self._html_search_regex(
788 r'</h1
><p
>(.*?
)</p
>',
789 content, 'block message
', default=None)
791 msg += ' (Message
: "%s")' % block_msg.replace('\n', ' ')
792 raise ExtractorError(msg, expected=True)
793 if ('<title
>TTK
:: Доступ к ресурсу ограничен
</title
>' in content
794 and 'blocklist
.rkn
.gov
.ru
' in content):
795 raise ExtractorError(
796 'Access to this webpage has been blocked by decision of the Russian government
. '
797 'Visit http
://blocklist
.rkn
.gov
.ru
/ for a block reason
.',
800 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
801 content_type = urlh.headers.get('Content
-Type
', '')
802 webpage_bytes = urlh.read()
803 if prefix is not None:
804 webpage_bytes = prefix + webpage_bytes
806 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
807 if self.get_param('dump_intermediate_pages
', False):
808 self.to_screen('Dumping request to
' + urlh.geturl())
809 dump = base64.b64encode(webpage_bytes).decode('ascii
')
810 self._downloader.to_screen(dump)
811 if self.get_param('write_pages
', False):
812 basen = '%s_%s' % (video_id, urlh.geturl())
813 trim_length = self.get_param('trim_file_name
') or 240
814 if len(basen) > trim_length:
815 h = '___
' + hashlib.md5(basen.encode('utf
-8')).hexdigest()
816 basen = basen[:trim_length - len(h)] + h
817 raw_filename = basen + '.dump
'
818 filename = sanitize_filename(raw_filename, restricted=True)
819 self.to_screen('Saving request to
' + filename)
820 # Working around MAX_PATH limitation on Windows (see
821 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
822 if compat_os_name == 'nt
':
823 absfilepath = os.path.abspath(filename)
824 if len(absfilepath) > 259:
825 filename = '\\\\?
\\' + absfilepath
826 with open(filename, 'wb
') as outf:
827 outf.write(webpage_bytes)
830 content = webpage_bytes.decode(encoding, 'replace
')
832 content = webpage_bytes.decode('utf
-8', 'replace
')
834 self.__check_blocked(content)
838 def _download_webpage(
839 self, url_or_request, video_id, note=None, errnote=None,
840 fatal=True, tries=1, timeout=5, encoding=None, data=None,
841 headers={}, query={}, expected_status=None):
843 Return the data of the page as a string.
846 url_or_request -- plain text URL as a string or
847 a compat_urllib_request.Requestobject
848 video_id -- Video/playlist/item identifier (string)
851 note -- note printed before downloading (string)
852 errnote -- note printed in case of an error (string)
853 fatal -- flag denoting whether error should be considered fatal,
854 i.e. whether it should cause ExtractionError to be raised,
855 otherwise a warning will be reported and extraction continued
856 tries -- number of tries
857 timeout -- sleep interval between tries
858 encoding -- encoding for a page content decoding, guessed automatically
859 when not explicitly specified
860 data -- POST data (bytes)
861 headers -- HTTP headers (dict)
862 query -- URL query (dict)
863 expected_status -- allows to accept failed HTTP requests (non 2xx
864 status code) by explicitly specifying a set of accepted status
865 codes. Can be any of the following entities:
866 - an integer type specifying an exact failed status code to
868 - a list or a tuple of integer types specifying a list of
869 failed status codes to accept
870 - a callable accepting an actual failed status code and
871 returning True if it should be accepted
872 Note that this argument does not affect success status codes (2xx)
873 which are always accepted.
878 while success is False:
880 res = self._download_webpage_handle(
881 url_or_request, video_id, note, errnote, fatal,
882 encoding=encoding, data=data, headers=headers, query=query,
883 expected_status=expected_status)
885 except compat_http_client.IncompleteRead as e:
887 if try_count >= tries:
889 self._sleep(timeout, video_id)
896 def _download_xml_handle(
897 self, url_or_request, video_id, note='Downloading XML
',
898 errnote='Unable to download XML
', transform_source=None,
899 fatal=True, encoding=None, data=None, headers={}, query={},
900 expected_status=None):
902 Return a tuple (xml as an compat_etree_Element, URL handle).
904 See _download_webpage docstring for arguments specification.
906 res = self._download_webpage_handle(
907 url_or_request, video_id, note, errnote, fatal=fatal,
908 encoding=encoding, data=data, headers=headers, query=query,
909 expected_status=expected_status)
912 xml_string, urlh = res
913 return self._parse_xml(
914 xml_string, video_id, transform_source=transform_source,
918 self, url_or_request, video_id,
919 note='Downloading XML
', errnote='Unable to download XML
',
920 transform_source=None, fatal=True, encoding=None,
921 data=None, headers={}, query={}, expected_status=None):
923 Return the xml as an compat_etree_Element.
925 See _download_webpage docstring for arguments specification.
927 res = self._download_xml_handle(
928 url_or_request, video_id, note=note, errnote=errnote,
929 transform_source=transform_source, fatal=fatal, encoding=encoding,
930 data=data, headers=headers, query=query,
931 expected_status=expected_status)
932 return res if res is False else res[0]
934 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
936 xml_string = transform_source(xml_string)
938 return compat_etree_fromstring(xml_string.encode('utf
-8'))
939 except compat_xml_parse_error as ve:
940 errmsg = '%s: Failed to parse XML
' % video_id
942 raise ExtractorError(errmsg, cause=ve)
944 self.report_warning(errmsg + str(ve))
946 def _download_json_handle(
947 self, url_or_request, video_id, note='Downloading JSON metadata
',
948 errnote='Unable to download JSON metadata
', transform_source=None,
949 fatal=True, encoding=None, data=None, headers={}, query={},
950 expected_status=None):
952 Return a tuple (JSON object, URL handle).
954 See _download_webpage docstring for arguments specification.
956 res = self._download_webpage_handle(
957 url_or_request, video_id, note, errnote, fatal=fatal,
958 encoding=encoding, data=data, headers=headers, query=query,
959 expected_status=expected_status)
962 json_string, urlh = res
963 return self._parse_json(
964 json_string, video_id, transform_source=transform_source,
968 self, url_or_request, video_id, note='Downloading JSON metadata
',
969 errnote='Unable to download JSON metadata
', transform_source=None,
970 fatal=True, encoding=None, data=None, headers={}, query={},
971 expected_status=None):
973 Return the JSON object as a dict.
975 See _download_webpage docstring for arguments specification.
977 res = self._download_json_handle(
978 url_or_request, video_id, note=note, errnote=errnote,
979 transform_source=transform_source, fatal=fatal, encoding=encoding,
980 data=data, headers=headers, query=query,
981 expected_status=expected_status)
982 return res if res is False else res[0]
984 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
986 json_string = transform_source(json_string)
988 return json.loads(json_string)
989 except ValueError as ve:
990 errmsg = '%s: Failed to parse JSON
' % video_id
992 raise ExtractorError(errmsg, cause=ve)
994 self.report_warning(errmsg + str(ve))
996 def _parse_socket_response_as_json(self, data, video_id, transform_source=None, fatal=True):
997 return self._parse_json(
998 data[data.find('{'):data.rfind('}
') + 1],
999 video_id, transform_source, fatal)
1001 def _download_socket_json_handle(
1002 self, url_or_request, video_id, note='Polling socket
',
1003 errnote='Unable to poll socket
', transform_source=None,
1004 fatal=True, encoding=None, data=None, headers={}, query={},
1005 expected_status=None):
1007 Return a tuple (JSON object, URL handle).
1009 See _download_webpage docstring for arguments specification.
1011 res = self._download_webpage_handle(
1012 url_or_request, video_id, note, errnote, fatal=fatal,
1013 encoding=encoding, data=data, headers=headers, query=query,
1014 expected_status=expected_status)
1018 return self._parse_socket_response_as_json(
1019 webpage, video_id, transform_source=transform_source,
1022 def _download_socket_json(
1023 self, url_or_request, video_id, note='Polling socket
',
1024 errnote='Unable to poll socket
', transform_source=None,
1025 fatal=True, encoding=None, data=None, headers={}, query={},
1026 expected_status=None):
1028 Return the JSON object as a dict.
1030 See _download_webpage docstring for arguments specification.
1032 res = self._download_socket_json_handle(
1033 url_or_request, video_id, note=note, errnote=errnote,
1034 transform_source=transform_source, fatal=fatal, encoding=encoding,
1035 data=data, headers=headers, query=query,
1036 expected_status=expected_status)
1037 return res if res is False else res[0]
1039 def report_warning(self, msg, video_id=None, *args, only_once=False, **kwargs):
1040 idstr = format_field(video_id, template='%s: ')
1041 msg = f'[{self.IE_NAME}
] {idstr}{msg}
'
1043 if f'WARNING
: {msg}
' in self._printed_messages:
1045 self._printed_messages.add(f'WARNING
: {msg}
')
1046 self._downloader.report_warning(msg, *args, **kwargs)
1048 def to_screen(self, msg, *args, **kwargs):
1049 """Print msg to screen, prefixing it with '[ie_name
]'"""
1050 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg), *args, **kwargs)
1052 def write_debug(self, msg, *args, **kwargs):
1053 self._downloader.write_debug('[%s] %s' % (self.IE_NAME, msg), *args, **kwargs)
1055 def get_param(self, name, default=None, *args, **kwargs):
1056 if self._downloader:
1057 return self._downloader.params.get(name, default, *args, **kwargs)
1060 def report_drm(self, video_id, partial=False):
1061 self.raise_no_formats('This video
is DRM protected
', expected=True, video_id=video_id)
1063 def report_extraction(self, id_or_name):
1064 """Report information extraction."""
1065 self.to_screen('%s: Extracting information
' % id_or_name)
1067 def report_download_webpage(self, video_id):
1068 """Report webpage download."""
1069 self.to_screen('%s: Downloading webpage
' % video_id)
1071 def report_age_confirmation(self):
1072 """Report attempt to confirm age."""
1073 self.to_screen('Confirming age
')
1075 def report_login(self):
1076 """Report attempt to log in."""
1077 self.to_screen('Logging
in')
1079 def raise_login_required(
1080 self, msg='This video
is only available
for registered users
',
1081 metadata_available=False, method='any
'):
1082 if metadata_available and self.get_param('ignore_no_formats_error
'):
1083 self.report_warning(msg)
1084 if method is not None:
1085 msg = '%s. %s' % (msg, self._LOGIN_HINTS[method])
1086 raise ExtractorError(msg, expected=True)
1088 def raise_geo_restricted(
1089 self, msg='This video
is not available
from your location due to geo restriction
',
1090 countries=None, metadata_available=False):
1091 if metadata_available and self.get_param('ignore_no_formats_error
'):
1092 self.report_warning(msg)
1094 raise GeoRestrictedError(msg, countries=countries)
1096 def raise_no_formats(self, msg, expected=False, video_id=None):
1097 if expected and self.get_param('ignore_no_formats_error
'):
1098 self.report_warning(msg, video_id)
1099 elif isinstance(msg, ExtractorError):
1102 raise ExtractorError(msg, expected=expected, video_id=video_id)
1104 # Methods for following #608
1106 def url_result(url, ie=None, video_id=None, video_title=None, **kwargs):
1107 """Returns a URL that points to a page that should be processed"""
1108 # TODO: ie should be the class used for getting the info
1109 video_info = {'_type
': 'url
',
1112 video_info.update(kwargs)
1113 if video_id is not None:
1114 video_info['id'] = video_id
1115 if video_title is not None:
1116 video_info['title
'] = video_title
1119 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None):
1121 self.url_result(self._proto_relative_url(getter(m) if getter else m), ie)
1123 return self.playlist_result(
1124 urls, playlist_id=playlist_id, playlist_title=playlist_title)
1127 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None, **kwargs):
1128 """Returns a playlist"""
1129 video_info = {'_type
': 'playlist
',
1131 video_info.update(kwargs)
1133 video_info['id'] = playlist_id
1135 video_info['title
'] = playlist_title
1136 if playlist_description is not None:
1137 video_info['description
'] = playlist_description
1140 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1142 Perform a regex search on the given string, using a single or a list of
1143 patterns returning the first matching group.
1144 In case of failure return a default value or raise a WARNING or a
1145 RegexNotFoundError, depending on fatal, specifying the field name.
1147 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
1148 mobj = re.search(pattern, string, flags)
1151 mobj = re.search(p, string, flags)
1155 _name = self._downloader._format_err(name, self._downloader.Styles.EMPHASIS)
1159 # return the first matching group
1160 return next(g for g in mobj.groups() if g is not None)
1161 elif isinstance(group, (list, tuple)):
1162 return tuple(mobj.group(g) for g in group)
1164 return mobj.group(group)
1165 elif default is not NO_DEFAULT:
1168 raise RegexNotFoundError('Unable to extract
%s' % _name)
1170 self.report_warning('unable to extract
%s' % _name + bug_reports_message())
1173 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1175 Like _search_regex, but strips HTML tags and unescapes entities.
1177 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1179 return clean_html(res).strip()
1183 def _get_netrc_login_info(self, netrc_machine=None):
1186 netrc_machine = netrc_machine or self._NETRC_MACHINE
1188 if self.get_param('usenetrc
', False):
1190 netrc_file = compat_expanduser(self.get_param('netrc_location
') or '~
')
1191 if os.path.isdir(netrc_file):
1192 netrc_file = os.path.join(netrc_file, '.netrc
')
1193 info = netrc.netrc(file=netrc_file).authenticators(netrc_machine)
1194 if info is not None:
1198 raise netrc.NetrcParseError(
1199 'No authenticators
for %s' % netrc_machine)
1200 except (IOError, netrc.NetrcParseError) as err:
1201 self.report_warning(
1202 'parsing
.netrc
: %s' % error_to_compat_str(err))
1204 return username, password
1206 def _get_login_info(self, username_option='username
', password_option='password
', netrc_machine=None):
1208 Get the login info as (username, password)
1209 First look for the manually specified credentials using username_option
1210 and password_option as keys in params dictionary. If no such credentials
1211 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1213 If there's no info available
, return (None, None)
1216 # Attempt to use provided username and password or .netrc data
1217 username = self.get_param(username_option)
1218 if username is not None:
1219 password = self.get_param(password_option)
1221 username, password = self._get_netrc_login_info(netrc_machine)
1223 return username, password
1225 def _get_tfa_info(self, note='two-factor verification code'):
1227 Get the two
-factor authentication info
1228 TODO
- asking the user will be required
for sms
/phone verify
1229 currently just uses the command line option
1230 If there
's no info available, return None
1233 tfa = self.get_param('twofactor
')
1237 return compat_getpass('Type
%s and press
[Return
]: ' % note)
1239 # Helper functions for extracting OpenGraph info
1241 def _og_regexes(prop):
1242 content_re = r'content
=(?
:"([^"]+?
)"|\'([^\']+?)\'|\s*([^\s"\'=<>`
]+?
))'
1243 property_re = (r'(?
:name|
property)=(?
:\'og
[:-]%(prop)s\'|
"og[:-]%(prop)s"|\s
*og
[:-]%(prop)s\b)'
1244 % {'prop': re.escape(prop)})
1245 template = r'<meta
[^
>]+?
%s[^
>]+?
%s'
1247 template % (property_re, content_re),
1248 template % (content_re, property_re),
1252 def _meta_regex(prop):
1253 return r'''(?isx)<meta
1254 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1255 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1257 def _og_search_property(self, prop, html, name=None, **kargs):
1258 prop = variadic(prop)
1260 name = 'OpenGraph
%s' % prop[0]
1263 og_regexes.extend(self._og_regexes(p))
1264 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1267 return unescapeHTML(escaped)
1269 def _og_search_thumbnail(self, html, **kargs):
1270 return self._og_search_property('image
', html, 'thumbnail URL
', fatal=False, **kargs)
1272 def _og_search_description(self, html, **kargs):
1273 return self._og_search_property('description
', html, fatal=False, **kargs)
1275 def _og_search_title(self, html, **kargs):
1276 return self._og_search_property('title
', html, **kargs)
1278 def _og_search_video_url(self, html, name='video url
', secure=True, **kargs):
1279 regexes = self._og_regexes('video
') + self._og_regexes('video
:url
')
1281 regexes = self._og_regexes('video
:secure_url
') + regexes
1282 return self._html_search_regex(regexes, html, name, **kargs)
1284 def _og_search_url(self, html, **kargs):
1285 return self._og_search_property('url
', html, **kargs)
1287 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1288 name = variadic(name)
1289 if display_name is None:
1290 display_name = name[0]
1291 return self._html_search_regex(
1292 [self._meta_regex(n) for n in name],
1293 html, display_name, fatal=fatal, group='content
', **kwargs)
1295 def _dc_search_uploader(self, html):
1296 return self._html_search_meta('dc
.creator
', html, 'uploader
')
1298 def _rta_search(self, html):
1299 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1300 if re.search(r'(?ix
)<meta\s
+name
="rating"\s
+'
1301 r' content
="RTA-5042-1996-1400-1577-RTA"',
1306 def _media_rating_search(self, html):
1307 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1308 rating = self._html_search_meta('rating
', html)
1320 return RATING_TABLE.get(rating.lower())
1322 def _family_friendly_search(self, html):
1323 # See http://schema.org/VideoObject
1324 family_friendly = self._html_search_meta(
1325 'isFamilyFriendly
', html, default=None)
1327 if not family_friendly:
1336 return RATING_TABLE.get(family_friendly.lower())
1338 def _twitter_search_player(self, html):
1339 return self._html_search_meta('twitter
:player
', html,
1340 'twitter card player
')
1342 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
1343 json_ld_list = list(re.finditer(JSON_LD_RE, html))
1344 default = kwargs.get('default
', NO_DEFAULT)
1345 # JSON-LD may be malformed and thus `fatal` should be respected.
1346 # At the same time `default` may be passed that assumes `fatal=False`
1347 # for _search_regex. Let's simulate the same behavior here
as well
.
1348 fatal
= kwargs
.get('fatal', True) if default
is NO_DEFAULT
else False
1350 for mobj
in json_ld_list
:
1351 json_ld_item
= self
._parse
_json
(
1352 mobj
.group('json_ld'), video_id
, fatal
=fatal
)
1353 if not json_ld_item
:
1355 if isinstance(json_ld_item
, dict):
1356 json_ld
.append(json_ld_item
)
1357 elif isinstance(json_ld_item
, (list, tuple)):
1358 json_ld
.extend(json_ld_item
)
1360 json_ld
= self
._json
_ld
(json_ld
, video_id
, fatal
=fatal
, expected_type
=expected_type
)
1363 if default
is not NO_DEFAULT
:
1366 raise RegexNotFoundError('Unable to extract JSON-LD')
1368 self
.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1371 def _json_ld(self
, json_ld
, video_id
, fatal
=True, expected_type
=None):
1372 if isinstance(json_ld
, compat_str
):
1373 json_ld
= self
._parse
_json
(json_ld
, video_id
, fatal
=fatal
)
1377 if not isinstance(json_ld
, (list, tuple, dict)):
1379 if isinstance(json_ld
, dict):
1382 INTERACTION_TYPE_MAP
= {
1383 'CommentAction': 'comment',
1384 'AgreeAction': 'like',
1385 'DisagreeAction': 'dislike',
1386 'LikeAction': 'like',
1387 'DislikeAction': 'dislike',
1388 'ListenAction': 'view',
1389 'WatchAction': 'view',
1390 'ViewAction': 'view',
1393 def extract_interaction_type(e
):
1394 interaction_type
= e
.get('interactionType')
1395 if isinstance(interaction_type
, dict):
1396 interaction_type
= interaction_type
.get('@type')
1397 return str_or_none(interaction_type
)
1399 def extract_interaction_statistic(e
):
1400 interaction_statistic
= e
.get('interactionStatistic')
1401 if isinstance(interaction_statistic
, dict):
1402 interaction_statistic
= [interaction_statistic
]
1403 if not isinstance(interaction_statistic
, list):
1405 for is_e
in interaction_statistic
:
1406 if not isinstance(is_e
, dict):
1408 if is_e
.get('@type') != 'InteractionCounter':
1410 interaction_type
= extract_interaction_type(is_e
)
1411 if not interaction_type
:
1413 # For interaction count some sites provide string instead of
1414 # an integer (as per spec) with non digit characters (e.g. ",")
1415 # so extracting count with more relaxed str_to_int
1416 interaction_count
= str_to_int(is_e
.get('userInteractionCount'))
1417 if interaction_count
is None:
1419 count_kind
= INTERACTION_TYPE_MAP
.get(interaction_type
.split('/')[-1])
1422 count_key
= '%s_count' % count_kind
1423 if info
.get(count_key
) is not None:
1425 info
[count_key
] = interaction_count
1427 def extract_video_object(e
):
1428 assert e
['@type'] == 'VideoObject'
1429 author
= e
.get('author')
1431 'url': url_or_none(e
.get('contentUrl')),
1432 'title': unescapeHTML(e
.get('name')),
1433 'description': unescapeHTML(e
.get('description')),
1434 'thumbnail': url_or_none(e
.get('thumbnailUrl') or e
.get('thumbnailURL')),
1435 'duration': parse_duration(e
.get('duration')),
1436 'timestamp': unified_timestamp(e
.get('uploadDate')),
1437 # author can be an instance of 'Organization' or 'Person' types.
1438 # both types can have 'name' property(inherited from 'Thing' type). [1]
1439 # however some websites are using 'Text' type instead.
1440 # 1. https://schema.org/VideoObject
1441 'uploader': author
.get('name') if isinstance(author
, dict) else author
if isinstance(author
, compat_str
) else None,
1442 'filesize': float_or_none(e
.get('contentSize')),
1443 'tbr': int_or_none(e
.get('bitrate')),
1444 'width': int_or_none(e
.get('width')),
1445 'height': int_or_none(e
.get('height')),
1446 'view_count': int_or_none(e
.get('interactionCount')),
1448 extract_interaction_statistic(e
)
1452 item_type
= e
.get('@type')
1453 if expected_type
is not None and expected_type
!= item_type
:
1455 rating
= traverse_obj(e
, ('aggregateRating', 'ratingValue'), expected_type
=float_or_none
)
1456 if rating
is not None:
1457 info
['average_rating'] = rating
1458 if item_type
in ('TVEpisode', 'Episode'):
1459 episode_name
= unescapeHTML(e
.get('name'))
1461 'episode': episode_name
,
1462 'episode_number': int_or_none(e
.get('episodeNumber')),
1463 'description': unescapeHTML(e
.get('description')),
1465 if not info
.get('title') and episode_name
:
1466 info
['title'] = episode_name
1467 part_of_season
= e
.get('partOfSeason')
1468 if isinstance(part_of_season
, dict) and part_of_season
.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1470 'season': unescapeHTML(part_of_season
.get('name')),
1471 'season_number': int_or_none(part_of_season
.get('seasonNumber')),
1473 part_of_series
= e
.get('partOfSeries') or e
.get('partOfTVSeries')
1474 if isinstance(part_of_series
, dict) and part_of_series
.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1475 info
['series'] = unescapeHTML(part_of_series
.get('name'))
1476 elif item_type
== 'Movie':
1478 'title': unescapeHTML(e
.get('name')),
1479 'description': unescapeHTML(e
.get('description')),
1480 'duration': parse_duration(e
.get('duration')),
1481 'timestamp': unified_timestamp(e
.get('dateCreated')),
1483 elif item_type
in ('Article', 'NewsArticle'):
1485 'timestamp': parse_iso8601(e
.get('datePublished')),
1486 'title': unescapeHTML(e
.get('headline')),
1487 'description': unescapeHTML(e
.get('articleBody')),
1489 elif item_type
== 'VideoObject':
1490 extract_video_object(e
)
1491 if expected_type
is None:
1495 video
= e
.get('video')
1496 if isinstance(video
, dict) and video
.get('@type') == 'VideoObject':
1497 extract_video_object(video
)
1498 if expected_type
is None:
1502 return dict((k
, v
) for k
, v
in info
.items() if v
is not None)
1504 def _search_nextjs_data(self
, webpage
, video_id
, **kw
):
1505 return self
._parse
_json
(
1507 r
'(?s)<script[^>]+id=[\'"]__NEXT_DATA__[\'"][^
>]*>([^
<]+)</script
>',
1508 webpage, 'next
.js data
', **kw),
1512 def _hidden_inputs(html):
1513 html = re.sub(r'<!--(?
:(?
!<!--).)*-->', '', html)
1515 for input in re.findall(r'(?i
)(<input[^
>]+>)', html):
1516 attrs = extract_attributes(input)
1519 if attrs.get('type') not in ('hidden
', 'submit
'):
1521 name = attrs.get('name
') or attrs.get('id')
1522 value = attrs.get('value
')
1523 if name and value is not None:
1524 hidden_inputs[name] = value
1525 return hidden_inputs
1527 def _form_hidden_inputs(self, form_id, html):
1528 form = self._search_regex(
1529 r'(?
is)<form
[^
>]+?
id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1530 html, '%s form' % form_id, group='form')
1531 return self._hidden_inputs(form)
1534 regex = r' *((?P<reverse>\+)?(?P<field>[a-zA-Z0-9_]+)((?P<separator>[~:])(?P<limit>.*?))?)? *$'
1536 default = ('hidden', 'aud_or_vid', 'hasvid', 'ie_pref', 'lang', 'quality',
1537 'res', 'fps', 'hdr:12', 'codec:vp9.2', 'size', 'br', 'asr',
1538 'proto', 'ext', 'hasaud', 'source', 'format_id') # These must not be aliases
1539 ytdl_default = ('hasaud', 'lang', 'quality', 'tbr', 'filesize', 'vbr',
1540 'height', 'width', 'proto', 'vext', 'abr', 'aext',
1541 'fps', 'fs_approx', 'source', 'format_id')
1544 'vcodec': {'type': 'ordered', 'regex': True,
1545 'order': ['av0?1', 'vp0?9.2', 'vp0?9', '[hx]265|he?vc?', '[hx]264|avc', 'vp0?8', 'mp4v|h263', 'theora', '', None, 'none']},
1546 'acodec': {'type': 'ordered', 'regex': True,
1547 'order': ['opus', 'vorbis', 'aac', 'mp?4a?', 'mp3', 'e-?a?c-?3', 'ac-?3', 'dts', '', None, 'none']},
1548 'hdr': {'type': 'ordered', 'regex': True, 'field': 'dynamic_range',
1549 'order': ['dv', '(hdr)?12', r'(hdr)?10\+', '(hdr)?10', 'hlg', '', 'sdr', None]},
1550 'proto': {'type': 'ordered', 'regex': True, 'field': 'protocol',
1551 'order': ['(ht|f)tps', '(ht|f)tp$', 'm3u8.+', '.*dash', 'ws|websocket', '', 'mms|rtsp', 'none', 'f4']},
1552 'vext': {'type': 'ordered', 'field': 'video_ext',
1553 'order': ('mp4', 'webm', 'flv', '', 'none'),
1554 'order_free': ('webm', 'mp4', 'flv', '', 'none')},
1555 'aext': {'type': 'ordered', 'field': 'audio_ext',
1556 'order': ('m4a', 'aac', 'mp3', 'ogg', 'opus', 'webm', '', 'none'),
1557 'order_free': ('opus', 'ogg', 'webm', 'm4a', 'mp3', 'aac', '', 'none')},
1558 'hidden': {'visible': False, 'forced': True, 'type': 'extractor', 'max': -1000},
1559 'aud_or_vid': {'visible': False, 'forced': True, 'type': 'multiple',
1560 'field': ('vcodec', 'acodec'),
1561 'function': lambda it: int(any(v != 'none' for v in it))},
1562 'ie_pref': {'priority': True, 'type': 'extractor'},
1563 'hasvid': {'priority': True, 'field': 'vcodec', 'type': 'boolean', 'not_in_list': ('none',)},
1564 'hasaud': {'field': 'acodec', 'type': 'boolean', 'not_in_list': ('none',)},
1565 'lang': {'convert': 'float', 'field': 'language_preference', 'default': -1},
1566 'quality': {'convert': 'float', 'default': -1},
1567 'filesize': {'convert': 'bytes'},
1568 'fs_approx': {'convert': 'bytes', 'field': 'filesize_approx'},
1569 'id': {'convert': 'string', 'field': 'format_id'},
1570 'height': {'convert': 'float_none'},
1571 'width': {'convert': 'float_none'},
1572 'fps': {'convert': 'float_none'},
1573 'tbr': {'convert': 'float_none'},
1574 'vbr': {'convert': 'float_none'},
1575 'abr': {'convert': 'float_none'},
1576 'asr': {'convert': 'float_none'},
1577 'source': {'convert': 'float', 'field': 'source_preference', 'default': -1},
1579 'codec': {'type': 'combined', 'field': ('vcodec', 'acodec')},
1580 'br': {'type': 'combined', 'field': ('tbr', 'vbr', 'abr'), 'same_limit': True},
1581 'size': {'type': 'combined', 'same_limit': True, 'field': ('filesize', 'fs_approx')},
1582 'ext': {'type': 'combined', 'field': ('vext', 'aext')},
1583 'res': {'type': 'multiple', 'field': ('height', 'width'),
1584 'function': lambda it: (lambda l: min(l) if l else 0)(tuple(filter(None, it)))},
1586 # Most of these exist only for compatibility reasons
1587 'dimension': {'type': 'alias', 'field': 'res'},
1588 'resolution': {'type': 'alias', 'field': 'res'},
1589 'extension': {'type': 'alias', 'field': 'ext'},
1590 'bitrate': {'type': 'alias', 'field': 'br'},
1591 'total_bitrate': {'type': 'alias', 'field': 'tbr'},
1592 'video_bitrate': {'type': 'alias', 'field': 'vbr'},
1593 'audio_bitrate': {'type': 'alias', 'field': 'abr'},
1594 'framerate': {'type': 'alias', 'field': 'fps'},
1595 'language_preference': {'type': 'alias', 'field': 'lang'}, # not named as 'language' because such a field exists
1596 'protocol': {'type': 'alias', 'field': 'proto'},
1597 'source_preference': {'type': 'alias', 'field': 'source'},
1598 'filesize_approx': {'type': 'alias', 'field': 'fs_approx'},
1599 'filesize_estimate': {'type': 'alias', 'field': 'size'},
1600 'samplerate': {'type': 'alias', 'field': 'asr'},
1601 'video_ext': {'type': 'alias', 'field': 'vext'},
1602 'audio_ext': {'type': 'alias', 'field': 'aext'},
1603 'video_codec': {'type': 'alias', 'field': 'vcodec'},
1604 'audio_codec': {'type': 'alias', 'field': 'acodec'},
1605 'video': {'type': 'alias', 'field': 'hasvid'},
1606 'has_video': {'type': 'alias', 'field': 'hasvid'},
1607 'audio': {'type': 'alias', 'field': 'hasaud'},
1608 'has_audio': {'type': 'alias', 'field': 'hasaud'},
1609 'extractor': {'type': 'alias', 'field': 'ie_pref'},
1610 'preference': {'type': 'alias', 'field': 'ie_pref'},
1611 'extractor_preference': {'type': 'alias', 'field': 'ie_pref'},
1612 'format_id': {'type': 'alias', 'field': 'id'},
1617 def _get_field_setting(self, field, key):
1618 if field not in self.settings:
1619 self.settings[field] = {}
1620 propObj = self.settings[field]
1621 if key not in propObj:
1622 type = propObj.get('type')
1624 default = 'preference' if type == 'extractor' else (field,) if type in ('combined', 'multiple') else field
1625 elif key == 'convert':
1626 default = 'order' if type == 'ordered' else 'float_string' if field else 'ignore'
1628 default = {'type': 'field', 'visible': True, 'order': [], 'not_in_list': (None,)}.get(key, None)
1629 propObj[key] = default
1632 def _resolve_field_value(self, field, value, convertNone=False):
1637 value = value.lower()
1638 conversion = self._get_field_setting(field, 'convert')
1639 if conversion == 'ignore':
1641 if conversion == 'string':
1643 elif conversion == 'float_none':
1644 return float_or_none(value)
1645 elif conversion == 'bytes':
1646 return FileDownloader.parse_bytes(value)
1647 elif conversion == 'order':
1648 order_list = (self._use_free_order and self._get_field_setting(field, 'order_free')) or self._get_field_setting(field, 'order')
1649 use_regex = self._get_field_setting(field, 'regex')
1650 list_length = len(order_list)
1651 empty_pos = order_list.index('') if '' in order_list else list_length + 1
1652 if use_regex and value is not None:
1653 for i, regex in enumerate(order_list):
1654 if regex and re.match(regex, value):
1655 return list_length - i
1656 return list_length - empty_pos # not in list
1657 else: # not regex or value = None
1658 return list_length - (order_list.index(value) if value in order_list else empty_pos)
1660 if value.isnumeric():
1663 self.settings[field]['convert'] = 'string'
1666 def evaluate_params(self, params, sort_extractor):
1667 self._use_free_order = params.get('prefer_free_formats', False)
1668 self._sort_user = params.get('format_sort', [])
1669 self._sort_extractor = sort_extractor
1671 def add_item(field, reverse, closest, limit_text):
1672 field = field.lower()
1673 if field in self._order:
1675 self._order.append(field)
1676 limit = self._resolve_field_value(field, limit_text)
1679 'closest': False if limit is None else closest,
1680 'limit_text': limit_text,
1682 if field in self.settings:
1683 self.settings[field].update(data)
1685 self.settings[field] = data
1688 tuple(field for field in self.default if self._get_field_setting(field, 'forced'))
1689 + (tuple() if params.get('format_sort_force', False)
1690 else tuple(field for field in self.default if self._get_field_setting(field, 'priority')))
1691 + tuple(self._sort_user) + tuple(sort_extractor) + self.default)
1693 for item in sort_list:
1694 match = re.match(self.regex, item)
1696 raise ExtractorError('Invalid format sort string "%s" given by extractor' % item)
1697 field = match.group('field')
1700 if self._get_field_setting(field, 'type') == 'alias':
1701 field = self._get_field_setting(field, 'field')
1702 reverse = match.group('reverse') is not None
1703 closest = match.group('separator') == '~'
1704 limit_text = match.group('limit')
1706 has_limit = limit_text is not None
1707 has_multiple_fields = self._get_field_setting(field, 'type') == 'combined'
1708 has_multiple_limits = has_limit and has_multiple_fields and not self._get_field_setting(field, 'same_limit')
1710 fields = self._get_field_setting(field, 'field') if has_multiple_fields else (field,)
1711 limits = limit_text.split(':') if has_multiple_limits else (limit_text,) if has_limit else tuple()
1712 limit_count = len(limits)
1713 for (i, f) in enumerate(fields):
1714 add_item(f, reverse, closest,
1715 limits[i] if i < limit_count
1716 else limits[0] if has_limit and not has_multiple_limits
1719 def print_verbose_info(self, write_debug):
1721 write_debug('Sort order given by user: %s' % ', '.join(self._sort_user))
1722 if self._sort_extractor:
1723 write_debug('Sort order given by extractor: %s' % ', '.join(self._sort_extractor))
1724 write_debug('Formats sorted by: %s' % ', '.join(['%s%s%s' % (
1725 '+' if self._get_field_setting(field, 'reverse') else '', field,
1726 '%s%s(%s)' % ('~' if self._get_field_setting(field, 'closest') else ':',
1727 self._get_field_setting(field, 'limit_text'),
1728 self._get_field_setting(field, 'limit'))
1729 if self._get_field_setting(field, 'limit_text') is not None else '')
1730 for field in self._order if self._get_field_setting(field, 'visible')]))
1732 def _calculate_field_preference_from_value(self, format, field, type, value):
1733 reverse = self._get_field_setting(field, 'reverse')
1734 closest = self._get_field_setting(field, 'closest')
1735 limit = self._get_field_setting(field, 'limit')
1737 if type == 'extractor':
1738 maximum = self._get_field_setting(field, 'max')
1739 if value is None or (maximum is not None and value >= maximum):
1741 elif type == 'boolean':
1742 in_list = self._get_field_setting(field, 'in_list')
1743 not_in_list = self._get_field_setting(field, 'not_in_list')
1744 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
1745 elif type == 'ordered':
1746 value = self._resolve_field_value(field, value, True)
1748 # try to convert to number
1749 val_num = float_or_none(value, default=self._get_field_setting(field, 'default'))
1750 is_num = self._get_field_setting(field, 'convert') != 'string' and val_num is not None
1754 return ((-10, 0) if value is None
1755 else (1, value, 0) if not is_num # if a field has mixed strings and numbers, strings are sorted higher
1756 else (0, -abs(value - limit), value - limit if reverse else limit - value) if closest
1757 else (0, value, 0) if not reverse and (limit is None or value <= limit)
1758 else (0, -value, 0) if limit is None or (reverse and value == limit) or value > limit
1759 else (-1, value, 0))
1761 def _calculate_field_preference(self, format, field):
1762 type = self._get_field_setting(field, 'type') # extractor, boolean, ordered, field, multiple
1763 get_value = lambda f: format.get(self._get_field_setting(f, 'field'))
1764 if type == 'multiple':
1765 type = 'field' # Only 'field' is allowed in multiple for now
1766 actual_fields = self._get_field_setting(field, 'field')
1768 value = self._get_field_setting(field, 'function')(get_value(f) for f in actual_fields)
1770 value = get_value(field)
1771 return self._calculate_field_preference_from_value(format, field, type, value)
1773 def calculate_preference(self, format):
1774 # Determine missing protocol
1775 if not format.get('protocol'):
1776 format['protocol'] = determine_protocol(format)
1778 # Determine missing ext
1779 if not format.get('ext') and 'url' in format:
1780 format['ext'] = determine_ext(format['url'])
1781 if format.get('vcodec') == 'none':
1782 format['audio_ext'] = format['ext'] if format.get('acodec') != 'none' else 'none'
1783 format['video_ext'] = 'none'
1785 format['video_ext'] = format['ext']
1786 format['audio_ext'] = 'none'
1787 # if format.get('preference') is None and format.get('ext') in ('f4f', 'f4m'): # Not supported?
1788 # format['preference'] = -1000
1790 # Determine missing bitrates
1791 if format.get('tbr') is None:
1792 if format.get('vbr') is not None and format.get('abr') is not None:
1793 format['tbr'] = format.get('vbr', 0) + format.get('abr', 0)
1795 if format.get('vcodec') != 'none' and format.get('vbr') is None:
1796 format['vbr'] = format.get('tbr') - format.get('abr', 0)
1797 if format.get('acodec') != 'none' and format.get('abr') is None:
1798 format['abr'] = format.get('tbr') - format.get('vbr', 0)
1800 return tuple(self._calculate_field_preference(format, field) for field in self._order)
1802 def _sort_formats(self, formats, field_preference=[]):
1805 format_sort = self.FormatSort() # params and to_screen are taken from the downloader
1806 format_sort.evaluate_params(self._downloader.params, field_preference)
1807 if self.get_param('verbose', False):
1808 format_sort.print_verbose_info(self._downloader.write_debug)
1809 formats.sort(key=lambda f: format_sort.calculate_preference(f))
1811 def _check_formats(self, formats, video_id):
1813 formats[:] = filter(
1814 lambda f: self._is_valid_url(
1816 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1820 def _remove_duplicate_formats(formats):
1824 if f['url'] not in format_urls:
1825 format_urls.add(f['url'])
1826 unique_formats.append(f)
1827 formats[:] = unique_formats
1829 def _is_valid_url(self, url, video_id, item='video', headers={}):
1830 url = self._proto_relative_url(url, scheme='http:')
1831 # For now assume non HTTP(S) URLs always valid
1832 if not (url.startswith('http://') or url.startswith('https://')):
1835 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1837 except ExtractorError as e:
1839 '%s: %s URL is invalid, skipping: %s'
1840 % (video_id, item, error_to_compat_str(e.cause)))
1843 def http_scheme(self):
1844 """ Either "http
:" or "https
:", depending on the user's preferences """
1847 if self.get_param('prefer_insecure', False)
1850 def _proto_relative_url(self, url, scheme=None):
1853 if url.startswith('//'):
1855 scheme = self.http_scheme()
1860 def _sleep(self, timeout, video_id, msg_template=None):
1861 if msg_template is None:
1862 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1863 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1867 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1868 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1869 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
1870 manifest = self._download_xml(
1871 manifest_url, video_id, 'Downloading f4m manifest',
1872 'Unable to download f4m manifest',
1873 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1874 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
1875 transform_source=transform_source,
1876 fatal=fatal, data=data, headers=headers, query=query)
1878 if manifest is False:
1881 return self._parse_f4m_formats(
1882 manifest, manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1883 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1885 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1886 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1887 fatal=True, m3u8_id=None):
1888 if not isinstance(manifest, compat_etree_Element) and not fatal:
1891 # currently yt-dlp cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1892 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1893 if akamai_pv is not None and ';' in akamai_pv.text:
1894 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1895 if playerVerificationChallenge.strip() != '':
1899 manifest_version = '1.0'
1900 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1902 manifest_version = '2.0'
1903 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1904 # Remove unsupported DRM protected media from final formats
1905 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
1906 media_nodes = remove_encrypted_media(media_nodes)
1910 manifest_base_url = get_base_url(manifest)
1912 bootstrap_info = xpath_element(
1913 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1914 'bootstrap info', default=None)
1917 mime_type = xpath_text(
1918 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1919 'base URL', default=None)
1920 if mime_type and mime_type.startswith('audio/'):
1923 for i, media_el in enumerate(media_nodes):
1924 tbr = int_or_none(media_el.attrib.get('bitrate'))
1925 width = int_or_none(media_el.attrib.get('width'))
1926 height = int_or_none(media_el.attrib.get('height'))
1927 format_id = join_nonempty(f4m_id, tbr or i)
1928 # If <bootstrapInfo> is present, the specified f4m is a
1929 # stream-level manifest, and only set-level manifests may refer to
1930 # external resources. See section 11.4 and section 4 of F4M spec
1931 if bootstrap_info is None:
1933 # @href is introduced in 2.0, see section 11.6 of F4M spec
1934 if manifest_version == '2.0':
1935 media_url = media_el.attrib.get('href')
1936 if media_url is None:
1937 media_url = media_el.attrib.get('url')
1941 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1942 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1943 # If media_url is itself a f4m manifest do the recursive extraction
1944 # since bitrates in parent manifest (this one) and media_url manifest
1945 # may differ leading to inability to resolve the format by requested
1946 # bitrate in f4m downloader
1947 ext = determine_ext(manifest_url)
1949 f4m_formats = self._extract_f4m_formats(
1950 manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1951 transform_source=transform_source, fatal=fatal)
1952 # Sometimes stream-level manifest contains single media entry that
1953 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1954 # At the same time parent's media entry in set-level manifest may
1955 # contain it. We will copy it from parent in such cases.
1956 if len(f4m_formats) == 1:
1959 'tbr': f.get('tbr') or tbr,
1960 'width': f.get('width') or width,
1961 'height': f.get('height') or height,
1962 'format_id': f.get('format_id') if not tbr else format_id,
1965 formats.extend(f4m_formats)
1968 formats.extend(self._extract_m3u8_formats(
1969 manifest_url, video_id, 'mp4', preference=preference,
1970 quality=quality, m3u8_id=m3u8_id, fatal=fatal))
1973 'format_id': format_id,
1974 'url': manifest_url,
1975 'manifest_url': manifest_url,
1976 'ext': 'flv' if bootstrap_info is not None else None,
1982 'preference': preference,
1987 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, quality=None, m3u8_id=None):
1989 'format_id': join_nonempty(m3u8_id, 'meta'),
1993 'preference': preference - 100 if preference else -100,
1995 'resolution': 'multiple',
1996 'format_note': 'Quality selection URL',
1999 def _report_ignoring_subs(self, name):
2000 self.report_warning(bug_reports_message(
2001 f'Ignoring subtitle tracks found in the {name} manifest; '
2002 'if any subtitle tracks are missing,'
2005 def _extract_m3u8_formats(self, *args, **kwargs):
2006 fmts, subs = self._extract_m3u8_formats_and_subtitles(*args, **kwargs)
2008 self._report_ignoring_subs('HLS')
2011 def _extract_m3u8_formats_and_subtitles(
2012 self, m3u8_url, video_id, ext=None, entry_protocol='m3u8_native',
2013 preference=None, quality=None, m3u8_id=None, note=None,
2014 errnote=None, fatal=True, live=False, data=None, headers={},
2017 res = self._download_webpage_handle(
2019 note='Downloading m3u8 information' if note is None else note,
2020 errnote='Failed to download m3u8 information' if errnote is None else errnote,
2021 fatal=fatal, data=data, headers=headers, query=query)
2026 m3u8_doc, urlh = res
2027 m3u8_url = urlh.geturl()
2029 return self._parse_m3u8_formats_and_subtitles(
2030 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
2031 preference=preference, quality=quality, m3u8_id=m3u8_id,
2032 note=note, errnote=errnote, fatal=fatal, live=live, data=data,
2033 headers=headers, query=query, video_id=video_id)
2035 def _parse_m3u8_formats_and_subtitles(
2036 self, m3u8_doc, m3u8_url, ext=None, entry_protocol='m3u8_native',
2037 preference=None, quality=None, m3u8_id=None, live=False, note=None,
2038 errnote=None, fatal=True, data=None, headers={}, query={},
2040 formats, subtitles = [], {}
2042 has_drm = re.search('|'.join([
2043 r'#EXT-X-FAXS-CM:', # Adobe Flash Access
2044 r'#EXT-X-(?:SESSION-)?KEY:.*?URI="skd
://', # Apple FairPlay
2047 def format_url(url):
2048 return url if re.match(r'^https?
://', url) else compat_urlparse.urljoin(m3u8_url, url)
2050 if self.get_param('hls_split_discontinuity
', False):
2051 def _extract_m3u8_playlist_indices(manifest_url=None, m3u8_doc=None):
2053 if not manifest_url:
2055 m3u8_doc = self._download_webpage(
2056 manifest_url, video_id, fatal=fatal, data=data, headers=headers,
2057 note=False, errnote='Failed to download m3u8 playlist information
')
2058 if m3u8_doc is False:
2060 return range(1 + sum(line.startswith('#EXT-X-DISCONTINUITY') for line in m3u8_doc.splitlines()))
2063 def _extract_m3u8_playlist_indices(*args
, **kwargs
):
2067 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
2068 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
2069 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
2071 # We should try extracting formats only from master playlists [1, 4.3.4],
2072 # i.e. playlists that describe available qualities. On the other hand
2073 # media playlists [1, 4.3.3] should be returned as is since they contain
2074 # just the media without qualities renditions.
2075 # Fortunately, master playlist can be easily distinguished from media
2076 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
2077 # master playlist tags MUST NOT appear in a media playlist and vice versa.
2078 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
2079 # media playlist and MUST NOT appear in master playlist thus we can
2080 # clearly detect media playlist with this criterion.
2082 if '#EXT-X-TARGETDURATION' in m3u8_doc
: # media playlist, return as is
2084 'format_id': join_nonempty(m3u8_id
, idx
),
2085 'format_index': idx
,
2088 'protocol': entry_protocol
,
2089 'preference': preference
,
2092 } for idx
in _extract_m3u8_playlist_indices(m3u8_doc
=m3u8_doc
)]
2094 return formats
, subtitles
2097 last_stream_inf
= {}
2099 def extract_media(x_media_line
):
2100 media
= parse_m3u8_attributes(x_media_line
)
2101 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
2102 media_type
, group_id
, name
= media
.get('TYPE'), media
.get('GROUP-ID'), media
.get('NAME')
2103 if not (media_type
and group_id
and name
):
2105 groups
.setdefault(group_id
, []).append(media
)
2106 # <https://tools.ietf.org/html/rfc8216#section-4.3.4.1>
2107 if media_type
== 'SUBTITLES':
2108 # According to RFC 8216 §4.3.4.2.1, URI is REQUIRED in the
2109 # EXT-X-MEDIA tag if the media type is SUBTITLES.
2110 # However, lack of URI has been spotted in the wild.
2111 # e.g. NebulaIE; see https://github.com/yt-dlp/yt-dlp/issues/339
2112 if not media
.get('URI'):
2114 url
= format_url(media
['URI'])
2117 'ext': determine_ext(url
),
2119 if sub_info
['ext'] == 'm3u8':
2120 # Per RFC 8216 §3.1, the only possible subtitle format m3u8
2121 # files may contain is WebVTT:
2122 # <https://tools.ietf.org/html/rfc8216#section-3.1>
2123 sub_info
['ext'] = 'vtt'
2124 sub_info
['protocol'] = 'm3u8_native'
2125 lang
= media
.get('LANGUAGE') or 'und'
2126 subtitles
.setdefault(lang
, []).append(sub_info
)
2127 if media_type
not in ('VIDEO', 'AUDIO'):
2129 media_url
= media
.get('URI')
2131 manifest_url
= format_url(media_url
)
2133 'format_id': join_nonempty(m3u8_id
, group_id
, name
, idx
),
2134 'format_note': name
,
2135 'format_index': idx
,
2136 'url': manifest_url
,
2137 'manifest_url': m3u8_url
,
2138 'language': media
.get('LANGUAGE'),
2140 'protocol': entry_protocol
,
2141 'preference': preference
,
2143 'vcodec': 'none' if media_type
== 'AUDIO' else None,
2144 } for idx
in _extract_m3u8_playlist_indices(manifest_url
))
2146 def build_stream_name():
2147 # Despite specification does not mention NAME attribute for
2148 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
2149 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
2150 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
2151 stream_name
= last_stream_inf
.get('NAME')
2154 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
2155 # from corresponding rendition group
2156 stream_group_id
= last_stream_inf
.get('VIDEO')
2157 if not stream_group_id
:
2159 stream_group
= groups
.get(stream_group_id
)
2160 if not stream_group
:
2161 return stream_group_id
2162 rendition
= stream_group
[0]
2163 return rendition
.get('NAME') or stream_group_id
2165 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
2166 # chance to detect video only formats when EXT-X-STREAM-INF tags
2167 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
2168 for line
in m3u8_doc
.splitlines():
2169 if line
.startswith('#EXT-X-MEDIA:'):
2172 for line
in m3u8_doc
.splitlines():
2173 if line
.startswith('#EXT-X-STREAM-INF:'):
2174 last_stream_inf
= parse_m3u8_attributes(line
)
2175 elif line
.startswith('#') or not line
.strip():
2178 tbr
= float_or_none(
2179 last_stream_inf
.get('AVERAGE-BANDWIDTH')
2180 or last_stream_inf
.get('BANDWIDTH'), scale
=1000)
2181 manifest_url
= format_url(line
.strip())
2183 for idx
in _extract_m3u8_playlist_indices(manifest_url
):
2184 format_id
= [m3u8_id
, None, idx
]
2185 # Bandwidth of live streams may differ over time thus making
2186 # format_id unpredictable. So it's better to keep provided
2189 stream_name
= build_stream_name()
2190 format_id
[1] = stream_name
or '%d' % (tbr
or len(formats
))
2192 'format_id': join_nonempty(*format_id
),
2193 'format_index': idx
,
2194 'url': manifest_url
,
2195 'manifest_url': m3u8_url
,
2198 'fps': float_or_none(last_stream_inf
.get('FRAME-RATE')),
2199 'protocol': entry_protocol
,
2200 'preference': preference
,
2203 resolution
= last_stream_inf
.get('RESOLUTION')
2205 mobj
= re
.search(r
'(?P<width>\d+)[xX](?P<height>\d+)', resolution
)
2207 f
['width'] = int(mobj
.group('width'))
2208 f
['height'] = int(mobj
.group('height'))
2209 # Unified Streaming Platform
2211 r
'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f
['url'])
2213 abr
, vbr
= mobj
.groups()
2214 abr
, vbr
= float_or_none(abr
, 1000), float_or_none(vbr
, 1000)
2219 codecs
= parse_codecs(last_stream_inf
.get('CODECS'))
2221 audio_group_id
= last_stream_inf
.get('AUDIO')
2222 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
2223 # references a rendition group MUST have a CODECS attribute.
2224 # However, this is not always respected, for example, [2]
2225 # contains EXT-X-STREAM-INF tag which references AUDIO
2226 # rendition group but does not have CODECS and despite
2227 # referencing an audio group it represents a complete
2228 # (with audio and video) format. So, for such cases we will
2229 # ignore references to rendition groups and treat them
2230 # as complete formats.
2231 if audio_group_id
and codecs
and f
.get('vcodec') != 'none':
2232 audio_group
= groups
.get(audio_group_id
)
2233 if audio_group
and audio_group
[0].get('URI'):
2234 # TODO: update acodec for audio only formats with
2236 f
['acodec'] = 'none'
2237 if not f
.get('ext'):
2238 f
['ext'] = 'm4a' if f
.get('vcodec') == 'none' else 'mp4'
2242 progressive_uri
= last_stream_inf
.get('PROGRESSIVE-URI')
2245 del http_f
['manifest_url']
2247 'format_id': f
['format_id'].replace('hls-', 'http-'),
2249 'url': progressive_uri
,
2251 formats
.append(http_f
)
2253 last_stream_inf
= {}
2254 return formats
, subtitles
2256 def _extract_m3u8_vod_duration(
2257 self
, m3u8_vod_url
, video_id
, note
=None, errnote
=None, data
=None, headers
={}, query={}
):
2259 m3u8_vod
= self
._download
_webpage
(
2260 m3u8_vod_url
, video_id
,
2261 note
='Downloading m3u8 VOD manifest' if note
is None else note
,
2262 errnote
='Failed to download VOD manifest' if errnote
is None else errnote
,
2263 fatal
=False, data
=data
, headers
=headers
, query
=query
)
2265 return self
._parse
_m
3u8_vod
_duration
(m3u8_vod
or '', video_id
)
2267 def _parse_m3u8_vod_duration(self
, m3u8_vod
, video_id
):
2268 if '#EXT-X-PLAYLIST-TYPE:VOD' not in m3u8_vod
:
2272 float(line
[len('#EXTINF:'):].split(',')[0])
2273 for line
in m3u8_vod
.splitlines() if line
.startswith('#EXTINF:'))) or None
2276 def _xpath_ns(path
, namespace
=None):
2280 for c
in path
.split('/'):
2281 if not c
or c
== '.':
2284 out
.append('{%s}%s' % (namespace
, c
))
2285 return '/'.join(out
)
2287 def _extract_smil_formats_and_subtitles(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None, transform_source
=None):
2288 smil
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
, transform_source
=transform_source
)
2294 namespace
= self
._parse
_smil
_namespace
(smil
)
2296 fmts
= self
._parse
_smil
_formats
(
2297 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2298 subs
= self
._parse
_smil
_subtitles
(
2299 smil
, namespace
=namespace
)
2303 def _extract_smil_formats(self
, *args
, **kwargs
):
2304 fmts
, subs
= self
._extract
_smil
_formats
_and
_subtitles
(*args
, **kwargs
)
2306 self
._report
_ignoring
_subs
('SMIL')
2309 def _extract_smil_info(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None):
2310 smil
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
)
2313 return self
._parse
_smil
(smil
, smil_url
, video_id
, f4m_params
=f4m_params
)
2315 def _download_smil(self
, smil_url
, video_id
, fatal
=True, transform_source
=None):
2316 return self
._download
_xml
(
2317 smil_url
, video_id
, 'Downloading SMIL file',
2318 'Unable to download SMIL file', fatal
=fatal
, transform_source
=transform_source
)
2320 def _parse_smil(self
, smil
, smil_url
, video_id
, f4m_params
=None):
2321 namespace
= self
._parse
_smil
_namespace
(smil
)
2323 formats
= self
._parse
_smil
_formats
(
2324 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2325 subtitles
= self
._parse
_smil
_subtitles
(smil
, namespace
=namespace
)
2327 video_id
= os
.path
.splitext(url_basename(smil_url
))[0]
2331 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2332 name
= meta
.attrib
.get('name')
2333 content
= meta
.attrib
.get('content')
2334 if not name
or not content
:
2336 if not title
and name
== 'title':
2338 elif not description
and name
in ('description', 'abstract'):
2339 description
= content
2340 elif not upload_date
and name
== 'date':
2341 upload_date
= unified_strdate(content
)
2344 'id': image
.get('type'),
2345 'url': image
.get('src'),
2346 'width': int_or_none(image
.get('width')),
2347 'height': int_or_none(image
.get('height')),
2348 } for image
in smil
.findall(self
._xpath
_ns
('.//image', namespace
)) if image
.get('src')]
2352 'title': title
or video_id
,
2353 'description': description
,
2354 'upload_date': upload_date
,
2355 'thumbnails': thumbnails
,
2357 'subtitles': subtitles
,
2360 def _parse_smil_namespace(self
, smil
):
2361 return self
._search
_regex
(
2362 r
'(?i)^{([^}]+)?}smil$', smil
.tag
, 'namespace', default
=None)
2364 def _parse_smil_formats(self
, smil
, smil_url
, video_id
, namespace
=None, f4m_params
=None, transform_rtmp_url
=None):
2366 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2367 b
= meta
.get('base') or meta
.get('httpBase')
2379 media
= smil
.findall(self
._xpath
_ns
('.//video', namespace
)) + smil
.findall(self
._xpath
_ns
('.//audio', namespace
))
2380 for medium
in media
:
2381 src
= medium
.get('src')
2382 if not src
or src
in srcs
:
2386 bitrate
= float_or_none(medium
.get('system-bitrate') or medium
.get('systemBitrate'), 1000)
2387 filesize
= int_or_none(medium
.get('size') or medium
.get('fileSize'))
2388 width
= int_or_none(medium
.get('width'))
2389 height
= int_or_none(medium
.get('height'))
2390 proto
= medium
.get('proto')
2391 ext
= medium
.get('ext')
2392 src_ext
= determine_ext(src
)
2393 streamer
= medium
.get('streamer') or base
2395 if proto
== 'rtmp' or streamer
.startswith('rtmp'):
2401 'format_id': 'rtmp-%d' % (rtmp_count
if bitrate
is None else bitrate
),
2403 'filesize': filesize
,
2407 if transform_rtmp_url
:
2408 streamer
, src
= transform_rtmp_url(streamer
, src
)
2409 formats
[-1].update({
2415 src_url
= src
if src
.startswith('http') else compat_urlparse
.urljoin(base
, src
)
2416 src_url
= src_url
.strip()
2418 if proto
== 'm3u8' or src_ext
== 'm3u8':
2419 m3u8_formats
= self
._extract
_m
3u8_formats
(
2420 src_url
, video_id
, ext
or 'mp4', m3u8_id
='hls', fatal
=False)
2421 if len(m3u8_formats
) == 1:
2423 m3u8_formats
[0].update({
2424 'format_id': 'hls-%d' % (m3u8_count
if bitrate
is None else bitrate
),
2429 formats
.extend(m3u8_formats
)
2430 elif src_ext
== 'f4m':
2435 'plugin': 'flowplayer-3.2.0.1',
2437 f4m_url
+= '&' if '?' in f4m_url
else '?'
2438 f4m_url
+= compat_urllib_parse_urlencode(f4m_params
)
2439 formats
.extend(self
._extract
_f
4m
_formats
(f4m_url
, video_id
, f4m_id
='hds', fatal
=False))
2440 elif src_ext
== 'mpd':
2441 formats
.extend(self
._extract
_mpd
_formats
(
2442 src_url
, video_id
, mpd_id
='dash', fatal
=False))
2443 elif re
.search(r
'\.ism/[Mm]anifest', src_url
):
2444 formats
.extend(self
._extract
_ism
_formats
(
2445 src_url
, video_id
, ism_id
='mss', fatal
=False))
2446 elif src_url
.startswith('http') and self
._is
_valid
_url
(src
, video_id
):
2450 'ext': ext
or src_ext
or 'flv',
2451 'format_id': 'http-%d' % (bitrate
or http_count
),
2453 'filesize': filesize
,
2458 for medium
in smil
.findall(self
._xpath
_ns
('.//imagestream', namespace
)):
2459 src
= medium
.get('src')
2460 if not src
or src
in srcs
:
2466 'format_id': 'imagestream-%d' % (imgs_count
),
2468 'ext': mimetype2ext(medium
.get('type')),
2471 'width': int_or_none(medium
.get('width')),
2472 'height': int_or_none(medium
.get('height')),
2473 'format_note': 'SMIL storyboards',
2478 def _parse_smil_subtitles(self
, smil
, namespace
=None, subtitles_lang
='en'):
2481 for num
, textstream
in enumerate(smil
.findall(self
._xpath
_ns
('.//textstream', namespace
))):
2482 src
= textstream
.get('src')
2483 if not src
or src
in urls
:
2486 ext
= textstream
.get('ext') or mimetype2ext(textstream
.get('type')) or determine_ext(src
)
2487 lang
= textstream
.get('systemLanguage') or textstream
.get('systemLanguageName') or textstream
.get('lang') or subtitles_lang
2488 subtitles
.setdefault(lang
, []).append({
2494 def _extract_xspf_playlist(self
, xspf_url
, playlist_id
, fatal
=True):
2495 xspf
= self
._download
_xml
(
2496 xspf_url
, playlist_id
, 'Downloading xpsf playlist',
2497 'Unable to download xspf manifest', fatal
=fatal
)
2500 return self
._parse
_xspf
(
2501 xspf
, playlist_id
, xspf_url
=xspf_url
,
2502 xspf_base_url
=base_url(xspf_url
))
2504 def _parse_xspf(self
, xspf_doc
, playlist_id
, xspf_url
=None, xspf_base_url
=None):
2506 'xspf': 'http://xspf.org/ns/0/',
2507 's1': 'http://static.streamone.nl/player/ns/0',
2511 for track
in xspf_doc
.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP
)):
2513 track
, xpath_with_ns('./xspf:title', NS_MAP
), 'title', default
=playlist_id
)
2514 description
= xpath_text(
2515 track
, xpath_with_ns('./xspf:annotation', NS_MAP
), 'description')
2516 thumbnail
= xpath_text(
2517 track
, xpath_with_ns('./xspf:image', NS_MAP
), 'thumbnail')
2518 duration
= float_or_none(
2519 xpath_text(track
, xpath_with_ns('./xspf:duration', NS_MAP
), 'duration'), 1000)
2522 for location
in track
.findall(xpath_with_ns('./xspf:location', NS_MAP
)):
2523 format_url
= urljoin(xspf_base_url
, location
.text
)
2528 'manifest_url': xspf_url
,
2529 'format_id': location
.get(xpath_with_ns('s1:label', NS_MAP
)),
2530 'width': int_or_none(location
.get(xpath_with_ns('s1:width', NS_MAP
))),
2531 'height': int_or_none(location
.get(xpath_with_ns('s1:height', NS_MAP
))),
2533 self
._sort
_formats
(formats
)
2538 'description': description
,
2539 'thumbnail': thumbnail
,
2540 'duration': duration
,
2545 def _extract_mpd_formats(self
, *args
, **kwargs
):
2546 fmts
, subs
= self
._extract
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2548 self
._report
_ignoring
_subs
('DASH')
2551 def _extract_mpd_formats_and_subtitles(
2552 self
, mpd_url
, video_id
, mpd_id
=None, note
=None, errnote
=None,
2553 fatal
=True, data
=None, headers
={}, query={}
):
2554 res
= self
._download
_xml
_handle
(
2556 note
='Downloading MPD manifest' if note
is None else note
,
2557 errnote
='Failed to download MPD manifest' if errnote
is None else errnote
,
2558 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
2564 mpd_base_url
= base_url(urlh
.geturl())
2566 return self
._parse
_mpd
_formats
_and
_subtitles
(
2567 mpd_doc
, mpd_id
, mpd_base_url
, mpd_url
)
2569 def _parse_mpd_formats(self
, *args
, **kwargs
):
2570 fmts
, subs
= self
._parse
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2572 self
._report
_ignoring
_subs
('DASH')
2575 def _parse_mpd_formats_and_subtitles(
2576 self
, mpd_doc
, mpd_id
=None, mpd_base_url
='', mpd_url
=None):
2578 Parse formats from MPD manifest.
2580 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
2581 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
2582 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
2584 if not self
.get_param('dynamic_mpd', True):
2585 if mpd_doc
.get('type') == 'dynamic':
2588 namespace
= self
._search
_regex
(r
'(?i)^{([^}]+)?}MPD$', mpd_doc
.tag
, 'namespace', default
=None)
2591 return self
._xpath
_ns
(path
, namespace
)
2593 def is_drm_protected(element
):
2594 return element
.find(_add_ns('ContentProtection')) is not None
2596 def extract_multisegment_info(element
, ms_parent_info
):
2597 ms_info
= ms_parent_info
.copy()
2599 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2600 # common attributes and elements. We will only extract relevant
2602 def extract_common(source
):
2603 segment_timeline
= source
.find(_add_ns('SegmentTimeline'))
2604 if segment_timeline
is not None:
2605 s_e
= segment_timeline
.findall(_add_ns('S'))
2607 ms_info
['total_number'] = 0
2610 r
= int(s
.get('r', 0))
2611 ms_info
['total_number'] += 1 + r
2612 ms_info
['s'].append({
2613 't': int(s
.get('t', 0)),
2614 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2615 'd': int(s
.attrib
['d']),
2618 start_number
= source
.get('startNumber')
2620 ms_info
['start_number'] = int(start_number
)
2621 timescale
= source
.get('timescale')
2623 ms_info
['timescale'] = int(timescale
)
2624 segment_duration
= source
.get('duration')
2625 if segment_duration
:
2626 ms_info
['segment_duration'] = float(segment_duration
)
2628 def extract_Initialization(source
):
2629 initialization
= source
.find(_add_ns('Initialization'))
2630 if initialization
is not None:
2631 ms_info
['initialization_url'] = initialization
.attrib
['sourceURL']
2633 segment_list
= element
.find(_add_ns('SegmentList'))
2634 if segment_list
is not None:
2635 extract_common(segment_list
)
2636 extract_Initialization(segment_list
)
2637 segment_urls_e
= segment_list
.findall(_add_ns('SegmentURL'))
2639 ms_info
['segment_urls'] = [segment
.attrib
['media'] for segment
in segment_urls_e
]
2641 segment_template
= element
.find(_add_ns('SegmentTemplate'))
2642 if segment_template
is not None:
2643 extract_common(segment_template
)
2644 media
= segment_template
.get('media')
2646 ms_info
['media'] = media
2647 initialization
= segment_template
.get('initialization')
2649 ms_info
['initialization'] = initialization
2651 extract_Initialization(segment_template
)
2654 mpd_duration
= parse_duration(mpd_doc
.get('mediaPresentationDuration'))
2655 formats
, subtitles
= [], {}
2656 stream_numbers
= collections
.defaultdict(int)
2657 for period
in mpd_doc
.findall(_add_ns('Period')):
2658 period_duration
= parse_duration(period
.get('duration')) or mpd_duration
2659 period_ms_info
= extract_multisegment_info(period
, {
2663 for adaptation_set
in period
.findall(_add_ns('AdaptationSet')):
2664 adaption_set_ms_info
= extract_multisegment_info(adaptation_set
, period_ms_info
)
2665 for representation
in adaptation_set
.findall(_add_ns('Representation')):
2666 representation_attrib
= adaptation_set
.attrib
.copy()
2667 representation_attrib
.update(representation
.attrib
)
2668 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2669 mime_type
= representation_attrib
['mimeType']
2670 content_type
= representation_attrib
.get('contentType', mime_type
.split('/')[0])
2672 codecs
= representation_attrib
.get('codecs', '')
2673 if content_type
not in ('video', 'audio', 'text'):
2674 if mime_type
== 'image/jpeg':
2675 content_type
= mime_type
2676 elif codecs
.split('.')[0] == 'stpp':
2677 content_type
= 'text'
2678 elif mimetype2ext(mime_type
) in ('tt', 'dfxp', 'ttml', 'xml', 'json'):
2679 content_type
= 'text'
2681 self
.report_warning('Unknown MIME type %s in DASH manifest' % mime_type
)
2685 for element
in (representation
, adaptation_set
, period
, mpd_doc
):
2686 base_url_e
= element
.find(_add_ns('BaseURL'))
2687 if base_url_e
is not None:
2688 base_url
= base_url_e
.text
+ base_url
2689 if re
.match(r
'^https?://', base_url
):
2691 if mpd_base_url
and base_url
.startswith('/'):
2692 base_url
= compat_urlparse
.urljoin(mpd_base_url
, base_url
)
2693 elif mpd_base_url
and not re
.match(r
'^https?://', base_url
):
2694 if not mpd_base_url
.endswith('/'):
2696 base_url
= mpd_base_url
+ base_url
2697 representation_id
= representation_attrib
.get('id')
2698 lang
= representation_attrib
.get('lang')
2699 url_el
= representation
.find(_add_ns('BaseURL'))
2700 filesize
= int_or_none(url_el
.attrib
.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el
is not None else None)
2701 bandwidth
= int_or_none(representation_attrib
.get('bandwidth'))
2702 if representation_id
is not None:
2703 format_id
= representation_id
2705 format_id
= content_type
2707 format_id
= mpd_id
+ '-' + format_id
2708 if content_type
in ('video', 'audio'):
2710 'format_id': format_id
,
2711 'manifest_url': mpd_url
,
2712 'ext': mimetype2ext(mime_type
),
2713 'width': int_or_none(representation_attrib
.get('width')),
2714 'height': int_or_none(representation_attrib
.get('height')),
2715 'tbr': float_or_none(bandwidth
, 1000),
2716 'asr': int_or_none(representation_attrib
.get('audioSamplingRate')),
2717 'fps': int_or_none(representation_attrib
.get('frameRate')),
2718 'language': lang
if lang
not in ('mul', 'und', 'zxx', 'mis') else None,
2719 'format_note': 'DASH %s' % content_type
,
2720 'filesize': filesize
,
2721 'container': mimetype2ext(mime_type
) + '_dash',
2723 f
.update(parse_codecs(codecs
))
2724 elif content_type
== 'text':
2726 'ext': mimetype2ext(mime_type
),
2727 'manifest_url': mpd_url
,
2728 'filesize': filesize
,
2730 elif content_type
== 'image/jpeg':
2731 # See test case in VikiIE
2732 # https://www.viki.com/videos/1175236v-choosing-spouse-by-lottery-episode-1
2734 'format_id': format_id
,
2736 'manifest_url': mpd_url
,
2737 'format_note': 'DASH storyboards (jpeg)',
2741 if is_drm_protected(adaptation_set
) or is_drm_protected(representation
):
2743 representation_ms_info
= extract_multisegment_info(representation
, adaption_set_ms_info
)
2745 def prepare_template(template_name
, identifiers
):
2746 tmpl
= representation_ms_info
[template_name
]
2747 # First of, % characters outside $...$ templates
2748 # must be escaped by doubling for proper processing
2749 # by % operator string formatting used further (see
2750 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2756 in_template
= not in_template
2757 elif c
== '%' and not in_template
:
2759 # Next, $...$ templates are translated to their
2760 # %(...) counterparts to be used with % operator
2761 if representation_id
is not None:
2762 t
= t
.replace('$RepresentationID$', representation_id
)
2763 t
= re
.sub(r
'\$(%s)\$' % '|'.join(identifiers
), r
'%(\1)d', t
)
2764 t
= re
.sub(r
'\$(%s)%%([^$]+)\$' % '|'.join(identifiers
), r
'%(\1)\2', t
)
2765 t
.replace('$$', '$')
2768 # @initialization is a regular template like @media one
2769 # so it should be handled just the same way (see
2770 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2771 if 'initialization' in representation_ms_info
:
2772 initialization_template
= prepare_template(
2774 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2775 # $Time$ shall not be included for @initialization thus
2776 # only $Bandwidth$ remains
2778 representation_ms_info
['initialization_url'] = initialization_template
% {
2779 'Bandwidth': bandwidth
,
2782 def location_key(location
):
2783 return 'url' if re
.match(r
'^https?://', location
) else 'path'
2785 if 'segment_urls' not in representation_ms_info
and 'media' in representation_ms_info
:
2787 media_template
= prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2788 media_location_key
= location_key(media_template
)
2790 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2791 # can't be used at the same time
2792 if '%(Number' in media_template
and 's' not in representation_ms_info
:
2793 segment_duration
= None
2794 if 'total_number' not in representation_ms_info
and 'segment_duration' in representation_ms_info
:
2795 segment_duration
= float_or_none(representation_ms_info
['segment_duration'], representation_ms_info
['timescale'])
2796 representation_ms_info
['total_number'] = int(math
.ceil(float(period_duration
) / segment_duration
))
2797 representation_ms_info
['fragments'] = [{
2798 media_location_key
: media_template
% {
2799 'Number': segment_number
,
2800 'Bandwidth': bandwidth
,
2802 'duration': segment_duration
,
2803 } for segment_number
in range(
2804 representation_ms_info
['start_number'],
2805 representation_ms_info
['total_number'] + representation_ms_info
['start_number'])]
2807 # $Number*$ or $Time$ in media template with S list available
2808 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2809 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2810 representation_ms_info
['fragments'] = []
2813 segment_number
= representation_ms_info
['start_number']
2815 def add_segment_url():
2816 segment_url
= media_template
% {
2817 'Time': segment_time
,
2818 'Bandwidth': bandwidth
,
2819 'Number': segment_number
,
2821 representation_ms_info
['fragments'].append({
2822 media_location_key
: segment_url
,
2823 'duration': float_or_none(segment_d
, representation_ms_info
['timescale']),
2826 for num
, s
in enumerate(representation_ms_info
['s']):
2827 segment_time
= s
.get('t') or segment_time
2831 for r
in range(s
.get('r', 0)):
2832 segment_time
+= segment_d
2835 segment_time
+= segment_d
2836 elif 'segment_urls' in representation_ms_info
and 's' in representation_ms_info
:
2838 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2839 # or any YouTube dashsegments video
2842 timescale
= representation_ms_info
['timescale']
2843 for s
in representation_ms_info
['s']:
2844 duration
= float_or_none(s
['d'], timescale
)
2845 for r
in range(s
.get('r', 0) + 1):
2846 segment_uri
= representation_ms_info
['segment_urls'][segment_index
]
2848 location_key(segment_uri
): segment_uri
,
2849 'duration': duration
,
2852 representation_ms_info
['fragments'] = fragments
2853 elif 'segment_urls' in representation_ms_info
:
2854 # Segment URLs with no SegmentTimeline
2855 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2856 # https://github.com/ytdl-org/youtube-dl/pull/14844
2858 segment_duration
= float_or_none(
2859 representation_ms_info
['segment_duration'],
2860 representation_ms_info
['timescale']) if 'segment_duration' in representation_ms_info
else None
2861 for segment_url
in representation_ms_info
['segment_urls']:
2863 location_key(segment_url
): segment_url
,
2865 if segment_duration
:
2866 fragment
['duration'] = segment_duration
2867 fragments
.append(fragment
)
2868 representation_ms_info
['fragments'] = fragments
2869 # If there is a fragments key available then we correctly recognized fragmented media.
2870 # Otherwise we will assume unfragmented media with direct access. Technically, such
2871 # assumption is not necessarily correct since we may simply have no support for
2872 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2873 if 'fragments' in representation_ms_info
:
2875 # NB: mpd_url may be empty when MPD manifest is parsed from a string
2876 'url': mpd_url
or base_url
,
2877 'fragment_base_url': base_url
,
2879 'protocol': 'http_dash_segments' if mime_type
!= 'image/jpeg' else 'mhtml',
2881 if 'initialization_url' in representation_ms_info
:
2882 initialization_url
= representation_ms_info
['initialization_url']
2883 if not f
.get('url'):
2884 f
['url'] = initialization_url
2885 f
['fragments'].append({location_key(initialization_url): initialization_url}
)
2886 f
['fragments'].extend(representation_ms_info
['fragments'])
2888 # Assuming direct URL to unfragmented media.
2890 if content_type
in ('video', 'audio', 'image/jpeg'):
2891 f
['manifest_stream_number'] = stream_numbers
[f
['url']]
2892 stream_numbers
[f
['url']] += 1
2894 elif content_type
== 'text':
2895 subtitles
.setdefault(lang
or 'und', []).append(f
)
2897 return formats
, subtitles
2899 def _extract_ism_formats(self
, *args
, **kwargs
):
2900 fmts
, subs
= self
._extract
_ism
_formats
_and
_subtitles
(*args
, **kwargs
)
2902 self
._report
_ignoring
_subs
('ISM')
2905 def _extract_ism_formats_and_subtitles(self
, ism_url
, video_id
, ism_id
=None, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
):
2906 res
= self
._download
_xml
_handle
(
2908 note
='Downloading ISM manifest' if note
is None else note
,
2909 errnote
='Failed to download ISM manifest' if errnote
is None else errnote
,
2910 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
2917 return self
._parse
_ism
_formats
_and
_subtitles
(ism_doc
, urlh
.geturl(), ism_id
)
2919 def _parse_ism_formats_and_subtitles(self
, ism_doc
, ism_url
, ism_id
=None):
2921 Parse formats from ISM manifest.
2923 1. [MS-SSTR]: Smooth Streaming Protocol,
2924 https://msdn.microsoft.com/en-us/library/ff469518.aspx
2926 if ism_doc
.get('IsLive') == 'TRUE':
2929 duration
= int(ism_doc
.attrib
['Duration'])
2930 timescale
= int_or_none(ism_doc
.get('TimeScale')) or 10000000
2934 for stream
in ism_doc
.findall('StreamIndex'):
2935 stream_type
= stream
.get('Type')
2936 if stream_type
not in ('video', 'audio', 'text'):
2938 url_pattern
= stream
.attrib
['Url']
2939 stream_timescale
= int_or_none(stream
.get('TimeScale')) or timescale
2940 stream_name
= stream
.get('Name')
2941 stream_language
= stream
.get('Language', 'und')
2942 for track
in stream
.findall('QualityLevel'):
2943 fourcc
= track
.get('FourCC') or ('AACL' if track
.get('AudioTag') == '255' else None)
2944 # TODO: add support for WVC1 and WMAP
2945 if fourcc
not in ('H264', 'AVC1', 'AACL', 'TTML'):
2946 self
.report_warning('%s is not a supported codec' % fourcc
)
2948 tbr
= int(track
.attrib
['Bitrate']) // 1000
2949 # [1] does not mention Width and Height attributes. However,
2950 # they're often present while MaxWidth and MaxHeight are
2951 # missing, so should be used as fallbacks
2952 width
= int_or_none(track
.get('MaxWidth') or track
.get('Width'))
2953 height
= int_or_none(track
.get('MaxHeight') or track
.get('Height'))
2954 sampling_rate
= int_or_none(track
.get('SamplingRate'))
2956 track_url_pattern
= re
.sub(r
'{[Bb]itrate}', track
.attrib
['Bitrate'], url_pattern
)
2957 track_url_pattern
= compat_urlparse
.urljoin(ism_url
, track_url_pattern
)
2963 stream_fragments
= stream
.findall('c')
2964 for stream_fragment_index
, stream_fragment
in enumerate(stream_fragments
):
2965 fragment_ctx
['time'] = int_or_none(stream_fragment
.get('t')) or fragment_ctx
['time']
2966 fragment_repeat
= int_or_none(stream_fragment
.get('r')) or 1
2967 fragment_ctx
['duration'] = int_or_none(stream_fragment
.get('d'))
2968 if not fragment_ctx
['duration']:
2970 next_fragment_time
= int(stream_fragment
[stream_fragment_index
+ 1].attrib
['t'])
2972 next_fragment_time
= duration
2973 fragment_ctx
['duration'] = (next_fragment_time
- fragment_ctx
['time']) / fragment_repeat
2974 for _
in range(fragment_repeat
):
2976 'url': re
.sub(r
'{start[ _]time}', compat_str(fragment_ctx
['time']), track_url_pattern
),
2977 'duration': fragment_ctx
['duration'] / stream_timescale
,
2979 fragment_ctx
['time'] += fragment_ctx
['duration']
2981 if stream_type
== 'text':
2982 subtitles
.setdefault(stream_language
, []).append({
2986 'manifest_url': ism_url
,
2987 'fragments': fragments
,
2988 '_download_params': {
2989 'stream_type': stream_type
,
2990 'duration': duration
,
2991 'timescale': stream_timescale
,
2993 'language': stream_language
,
2994 'codec_private_data': track
.get('CodecPrivateData'),
2997 elif stream_type
in ('video', 'audio'):
2999 'format_id': join_nonempty(ism_id
, stream_name
, tbr
),
3001 'manifest_url': ism_url
,
3002 'ext': 'ismv' if stream_type
== 'video' else 'isma',
3006 'asr': sampling_rate
,
3007 'vcodec': 'none' if stream_type
== 'audio' else fourcc
,
3008 'acodec': 'none' if stream_type
== 'video' else fourcc
,
3010 'fragments': fragments
,
3011 'has_drm': ism_doc
.find('Protection') is not None,
3012 '_download_params': {
3013 'stream_type': stream_type
,
3014 'duration': duration
,
3015 'timescale': stream_timescale
,
3016 'width': width
or 0,
3017 'height': height
or 0,
3019 'language': stream_language
,
3020 'codec_private_data': track
.get('CodecPrivateData'),
3021 'sampling_rate': sampling_rate
,
3022 'channels': int_or_none(track
.get('Channels', 2)),
3023 'bits_per_sample': int_or_none(track
.get('BitsPerSample', 16)),
3024 'nal_unit_length_field': int_or_none(track
.get('NALUnitLengthField', 4)),
3027 return formats
, subtitles
3029 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):
3030 def absolute_url(item_url
):
3031 return urljoin(base_url
, item_url
)
3033 def parse_content_type(content_type
):
3034 if not content_type
:
3036 ctr
= re
.search(r
'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type
)
3038 mimetype
, codecs
= ctr
.groups()
3039 f
= parse_codecs(codecs
)
3040 f
['ext'] = mimetype2ext(mimetype
)
3044 def _media_formats(src
, cur_media_type
, type_info
={}):
3045 full_url
= absolute_url(src
)
3046 ext
= type_info
.get('ext') or determine_ext(full_url
)
3048 is_plain_url
= False
3049 formats
= self
._extract
_m
3u8_formats
(
3050 full_url
, video_id
, ext
='mp4',
3051 entry_protocol
=m3u8_entry_protocol
, m3u8_id
=m3u8_id
,
3052 preference
=preference
, quality
=quality
, fatal
=False)
3054 is_plain_url
= False
3055 formats
= self
._extract
_mpd
_formats
(
3056 full_url
, video_id
, mpd_id
=mpd_id
, fatal
=False)
3061 'vcodec': 'none' if cur_media_type
== 'audio' else None,
3063 return is_plain_url
, formats
3066 # amp-video and amp-audio are very similar to their HTML5 counterparts
3067 # so we wll include them right here (see
3068 # https://www.ampproject.org/docs/reference/components/amp-video)
3069 # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/
3070 _MEDIA_TAG_NAME_RE
= r
'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)'
3071 media_tags
= [(media_tag
, media_tag_name
, media_type
, '')
3072 for media_tag
, media_tag_name
, media_type
3073 in re
.findall(r
'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE
, webpage
)]
3074 media_tags
.extend(re
.findall(
3075 # We only allow video|audio followed by a whitespace or '>'.
3076 # Allowing more characters may end up in significant slow down (see
3077 # https://github.com/ytdl-org/youtube-dl/issues/11979, example URL:
3078 # http://www.porntrex.com/maps/videositemap.xml).
3079 r
'(?s)(<(?P<tag>%s)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>' % _MEDIA_TAG_NAME_RE
, webpage
))
3080 for media_tag
, _
, media_type
, media_content
in media_tags
:
3085 media_attributes
= extract_attributes(media_tag
)
3086 src
= strip_or_none(media_attributes
.get('src'))
3088 _
, formats
= _media_formats(src
, media_type
)
3089 media_info
['formats'].extend(formats
)
3090 media_info
['thumbnail'] = absolute_url(media_attributes
.get('poster'))
3092 for source_tag
in re
.findall(r
'<source[^>]+>', media_content
):
3093 s_attr
= extract_attributes(source_tag
)
3094 # data-video-src and data-src are non standard but seen
3095 # several times in the wild
3096 src
= strip_or_none(dict_get(s_attr
, ('src', 'data-video-src', 'data-src')))
3099 f
= parse_content_type(s_attr
.get('type'))
3100 is_plain_url
, formats
= _media_formats(src
, media_type
, f
)
3102 # width, height, res, label and title attributes are
3103 # all not standard but seen several times in the wild
3106 for lbl
in ('label', 'title')
3107 if str_or_none(s_attr
.get(lbl
))
3109 width
= int_or_none(s_attr
.get('width'))
3110 height
= (int_or_none(s_attr
.get('height'))
3111 or int_or_none(s_attr
.get('res')))
3112 if not width
or not height
:
3114 resolution
= parse_resolution(lbl
)
3117 width
= width
or resolution
.get('width')
3118 height
= height
or resolution
.get('height')
3120 tbr
= parse_bitrate(lbl
)
3129 'format_id': s_attr
.get('label') or s_attr
.get('title'),
3131 f
.update(formats
[0])
3132 media_info
['formats'].append(f
)
3134 media_info
['formats'].extend(formats
)
3135 for track_tag
in re
.findall(r
'<track[^>]+>', media_content
):
3136 track_attributes
= extract_attributes(track_tag
)
3137 kind
= track_attributes
.get('kind')
3138 if not kind
or kind
in ('subtitles', 'captions'):
3139 src
= strip_or_none(track_attributes
.get('src'))
3142 lang
= track_attributes
.get('srclang') or track_attributes
.get('lang') or track_attributes
.get('label')
3143 media_info
['subtitles'].setdefault(lang
, []).append({
3144 'url': absolute_url(src
),
3146 for f
in media_info
['formats']:
3147 f
.setdefault('http_headers', {})['Referer'] = base_url
3148 if media_info
['formats'] or media_info
['subtitles']:
3149 entries
.append(media_info
)
3152 def _extract_akamai_formats(self
, *args
, **kwargs
):
3153 fmts
, subs
= self
._extract
_akamai
_formats
_and
_subtitles
(*args
, **kwargs
)
3155 self
._report
_ignoring
_subs
('akamai')
3158 def _extract_akamai_formats_and_subtitles(self
, manifest_url
, video_id
, hosts
={}):
3159 signed
= 'hdnea=' in manifest_url
3161 # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
3162 manifest_url
= re
.sub(
3163 r
'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
3164 '', manifest_url
).strip('?')
3169 hdcore_sign
= 'hdcore=3.7.0'
3170 f4m_url
= re
.sub(r
'(https?://[^/]+)/i/', r
'\1/z/', manifest_url
).replace('/master.m3u8', '/manifest.f4m')
3171 hds_host
= hosts
.get('hds')
3173 f4m_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hds_host
, f4m_url
)
3174 if 'hdcore=' not in f4m_url
:
3175 f4m_url
+= ('&' if '?' in f4m_url
else '?') + hdcore_sign
3176 f4m_formats
= self
._extract
_f
4m
_formats
(
3177 f4m_url
, video_id
, f4m_id
='hds', fatal
=False)
3178 for entry
in f4m_formats
:
3179 entry
.update({'extra_param_to_segment_url': hdcore_sign}
)
3180 formats
.extend(f4m_formats
)
3182 m3u8_url
= re
.sub(r
'(https?://[^/]+)/z/', r
'\1/i/', manifest_url
).replace('/manifest.f4m', '/master.m3u8')
3183 hls_host
= hosts
.get('hls')
3185 m3u8_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hls_host
, m3u8_url
)
3186 m3u8_formats
, m3u8_subtitles
= self
._extract
_m
3u8_formats
_and
_subtitles
(
3187 m3u8_url
, video_id
, 'mp4', 'm3u8_native',
3188 m3u8_id
='hls', fatal
=False)
3189 formats
.extend(m3u8_formats
)
3190 subtitles
= self
._merge
_subtitles
(subtitles
, m3u8_subtitles
)
3192 http_host
= hosts
.get('http')
3193 if http_host
and m3u8_formats
and not signed
:
3194 REPL_REGEX
= r
'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
3195 qualities
= re
.match(REPL_REGEX
, m3u8_url
).group(2).split(',')
3196 qualities_length
= len(qualities
)
3197 if len(m3u8_formats
) in (qualities_length
, qualities_length
+ 1):
3199 for f
in m3u8_formats
:
3200 if f
['vcodec'] != 'none':
3201 for protocol
in ('http', 'https'):
3203 del http_f
['manifest_url']
3205 REPL_REGEX
, protocol
+ r
'://%s/\g<1>%s\3' % (http_host
, qualities
[i
]), f
['url'])
3207 'format_id': http_f
['format_id'].replace('hls-', protocol
+ '-'),
3209 'protocol': protocol
,
3211 formats
.append(http_f
)
3214 return formats
, subtitles
3216 def _extract_wowza_formats(self
, url
, video_id
, m3u8_entry_protocol
='m3u8_native', skip_protocols
=[]):
3217 query
= compat_urlparse
.urlparse(url
).query
3218 url
= re
.sub(r
'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url
)
3220 r
'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url
)
3221 url_base
= mobj
.group('url')
3222 http_base_url
= '%s%s:%s' % ('http', mobj
.group('s') or '', url_base
)
3225 def manifest_url(manifest
):
3226 m_url
= '%s/%s' % (http_base_url
, manifest
)
3228 m_url
+= '?%s' % query
3231 if 'm3u8' not in skip_protocols
:
3232 formats
.extend(self
._extract
_m
3u8_formats
(
3233 manifest_url('playlist.m3u8'), video_id
, 'mp4',
3234 m3u8_entry_protocol
, m3u8_id
='hls', fatal
=False))
3235 if 'f4m' not in skip_protocols
:
3236 formats
.extend(self
._extract
_f
4m
_formats
(
3237 manifest_url('manifest.f4m'),
3238 video_id
, f4m_id
='hds', fatal
=False))
3239 if 'dash' not in skip_protocols
:
3240 formats
.extend(self
._extract
_mpd
_formats
(
3241 manifest_url('manifest.mpd'),
3242 video_id
, mpd_id
='dash', fatal
=False))
3243 if re
.search(r
'(?:/smil:|\.smil)', url_base
):
3244 if 'smil' not in skip_protocols
:
3245 rtmp_formats
= self
._extract
_smil
_formats
(
3246 manifest_url('jwplayer.smil'),
3247 video_id
, fatal
=False)
3248 for rtmp_format
in rtmp_formats
:
3249 rtsp_format
= rtmp_format
.copy()
3250 rtsp_format
['url'] = '%s/%s' % (rtmp_format
['url'], rtmp_format
['play_path'])
3251 del rtsp_format
['play_path']
3252 del rtsp_format
['ext']
3253 rtsp_format
.update({
3254 'url': rtsp_format
['url'].replace('rtmp://', 'rtsp://'),
3255 'format_id': rtmp_format
['format_id'].replace('rtmp', 'rtsp'),
3258 formats
.extend([rtmp_format
, rtsp_format
])
3260 for protocol
in ('rtmp', 'rtsp'):
3261 if protocol
not in skip_protocols
:
3263 'url': '%s:%s' % (protocol
, url_base
),
3264 'format_id': protocol
,
3265 'protocol': protocol
,
3269 def _find_jwplayer_data(self
, webpage
, video_id
=None, transform_source
=js_to_json
):
3271 r
'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P
=quote
)\
)(?
!</script
>).*?\
.setup\s
*\
((?P
<options
>[^
)]+)\
)',
3275 jwplayer_data = self._parse_json(mobj.group('options
'),
3277 transform_source=transform_source)
3278 except ExtractorError:
3281 if isinstance(jwplayer_data, dict):
3282 return jwplayer_data
3284 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
3285 jwplayer_data = self._find_jwplayer_data(
3286 webpage, video_id, transform_source=js_to_json)
3287 return self._parse_jwplayer_data(
3288 jwplayer_data, video_id, *args, **kwargs)
3290 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
3291 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3292 # JWPlayer backward compatibility: flattened playlists
3293 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
3294 if 'playlist
' not in jwplayer_data:
3295 jwplayer_data = {'playlist': [jwplayer_data]}
3299 # JWPlayer backward compatibility: single playlist item
3300 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
3301 if not isinstance(jwplayer_data['playlist
'], list):
3302 jwplayer_data['playlist
'] = [jwplayer_data['playlist
']]
3304 for video_data in jwplayer_data['playlist
']:
3305 # JWPlayer backward compatibility: flattened sources
3306 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
3307 if 'sources
' not in video_data:
3308 video_data['sources
'] = [video_data]
3310 this_video_id = video_id or video_data['mediaid
']
3312 formats = self._parse_jwplayer_formats(
3313 video_data['sources
'], video_id=this_video_id, m3u8_id=m3u8_id,
3314 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
3317 tracks = video_data.get('tracks
')
3318 if tracks and isinstance(tracks, list):
3319 for track in tracks:
3320 if not isinstance(track, dict):
3322 track_kind = track.get('kind
')
3323 if not track_kind or not isinstance(track_kind, compat_str):
3325 if track_kind.lower() not in ('captions
', 'subtitles
'):
3327 track_url = urljoin(base_url, track.get('file'))
3330 subtitles.setdefault(track.get('label
') or 'en
', []).append({
3331 'url
': self._proto_relative_url(track_url)
3335 'id': this_video_id,
3336 'title
': unescapeHTML(video_data['title
'] if require_title else video_data.get('title
')),
3337 'description
': clean_html(video_data.get('description
')),
3338 'thumbnail
': urljoin(base_url, self._proto_relative_url(video_data.get('image
'))),
3339 'timestamp
': int_or_none(video_data.get('pubdate
')),
3340 'duration
': float_or_none(jwplayer_data.get('duration
') or video_data.get('duration
')),
3341 'subtitles
': subtitles,
3343 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
3344 if len(formats) == 1 and re.search(r'^
(?
:http|
//).*(?
:youtube\
.com|youtu\
.be
)/.+', formats[0]['url
']):
3346 '_type
': 'url_transparent
',
3347 'url
': formats[0]['url
'],
3350 self._sort_formats(formats)
3351 entry['formats
'] = formats
3352 entries.append(entry)
3353 if len(entries) == 1:
3356 return self.playlist_result(entries)
3358 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
3359 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3362 for source in jwplayer_sources_data:
3363 if not isinstance(source, dict):
3365 source_url = urljoin(
3366 base_url, self._proto_relative_url(source.get('file')))
3367 if not source_url or source_url in urls:
3369 urls.append(source_url)
3370 source_type = source.get('type') or ''
3371 ext = mimetype2ext(source_type) or determine_ext(source_url)
3372 if source_type == 'hls
' or ext == 'm3u8
':
3373 formats.extend(self._extract_m3u8_formats(
3374 source_url, video_id, 'mp4
', entry_protocol='m3u8_native
',
3375 m3u8_id=m3u8_id, fatal=False))
3376 elif source_type == 'dash
' or ext == 'mpd
':
3377 formats.extend(self._extract_mpd_formats(
3378 source_url, video_id, mpd_id=mpd_id, fatal=False))
3380 formats.extend(self._extract_smil_formats(
3381 source_url, video_id, fatal=False))
3382 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3383 elif source_type.startswith('audio
') or ext in (
3384 'oga
', 'aac
', 'mp3
', 'mpeg
', 'vorbis
'):
3391 height = int_or_none(source.get('height
'))
3393 # Often no height is provided but there is a label in
3394 # format like "1080p", "720p SD", or 1080.
3395 height = int_or_none(self._search_regex(
3396 r'^
(\d{3,4}
)[pP
]?
(?
:\b|$
)', compat_str(source.get('label
') or ''),
3397 'height
', default=None))
3400 'width
': int_or_none(source.get('width
')),
3402 'tbr
': int_or_none(source.get('bitrate
')),
3405 if source_url.startswith('rtmp
'):
3406 a_format['ext
'] = 'flv
'
3407 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3408 # of jwplayer.flash.swf
3409 rtmp_url_parts = re.split(
3410 r'((?
:mp4|mp3|flv
):)', source_url, 1)
3411 if len(rtmp_url_parts) == 3:
3412 rtmp_url, prefix, play_path = rtmp_url_parts
3415 'play_path
': prefix + play_path,
3418 a_format.update(rtmp_params)
3419 formats.append(a_format)
3422 def _live_title(self, name):
3423 """ Generate the title for a live video """
3424 now = datetime.datetime.now()
3425 now_str = now.strftime('%Y
-%m
-%d %H
:%M
')
3426 return name + ' ' + now_str
3428 def _int(self, v, name, fatal=False, **kwargs):
3429 res = int_or_none(v, **kwargs)
3430 if 'get_attr
' in kwargs:
3431 print(getattr(v, kwargs['get_attr
']))
3433 msg = 'Failed to extract
%s: Could
not parse value
%r' % (name, v)
3435 raise ExtractorError(msg)
3437 self.report_warning(msg)
3440 def _float(self, v, name, fatal=False, **kwargs):
3441 res = float_or_none(v, **kwargs)
3443 msg = 'Failed to extract
%s: Could
not parse value
%r' % (name, v)
3445 raise ExtractorError(msg)
3447 self.report_warning(msg)
3450 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
3451 path='/', secure=False, discard=False, rest={}, **kwargs):
3452 cookie = compat_cookiejar_Cookie(
3453 0, name, value, port, port is not None, domain, True,
3454 domain.startswith('.'), path, True, secure, expire_time,
3455 discard, None, None, rest)
3456 self._downloader.cookiejar.set_cookie(cookie)
3458 def _get_cookies(self, url):
3459 """ Return a compat_cookies_SimpleCookie with the cookies for the url """
3460 req = sanitized_Request(url)
3461 self._downloader.cookiejar.add_cookie_header(req)
3462 return compat_cookies_SimpleCookie(req.get_header('Cookie
'))
3464 def _apply_first_set_cookie_header(self, url_handle, cookie):
3466 Apply first Set-Cookie header instead of the last. Experimental.
3468 Some sites (e.g. [1-3]) may serve two cookies under the same name
3469 in Set-Cookie header and expect the first (old) one to be set rather
3470 than second (new). However, as of RFC6265 the newer one cookie
3471 should be set into cookie store what actually happens.
3472 We will workaround this issue by resetting the cookie to
3473 the first one manually.
3474 1. https://new.vk.com/
3475 2. https://github.com/ytdl-org/youtube-dl/issues/9841#issuecomment-227871201
3476 3. https://learning.oreilly.com/
3478 for header, cookies in url_handle.headers.items():
3479 if header.lower() != 'set-cookie
':
3481 if sys.version_info[0] >= 3:
3482 cookies = cookies.encode('iso
-8859-1')
3483 cookies = cookies.decode('utf
-8')
3484 cookie_value = re.search(
3485 r'%s=(.+?
);.*?
\b[Dd
]omain
=(.+?
)(?
:[,;]|$
)' % cookie, cookies)
3487 value, domain = cookie_value.groups()
3488 self._set_cookie(domain, cookie, value)
3491 def get_testcases(self, include_onlymatching=False):
3492 t = getattr(self, '_TEST
', None)
3494 assert not hasattr(self, '_TESTS
'), \
3495 '%s has _TEST
and _TESTS
' % type(self).__name__
3498 tests = getattr(self, '_TESTS
', [])
3500 if not include_onlymatching and t.get('only_matching
', False):
3502 t['name
'] = type(self).__name__[:-len('IE
')]
3505 def is_suitable(self, age_limit):
3506 """ Test whether the extractor is generally suitable for the given
3507 age limit (i.e. pornographic sites are not, all others usually are) """
3509 any_restricted = False
3510 for tc in self.get_testcases(include_onlymatching=False):
3511 if tc.get('playlist
', []):
3512 tc = tc['playlist
'][0]
3513 is_restricted = age_restricted(
3514 tc.get('info_dict
', {}).get('age_limit
'), age_limit)
3515 if not is_restricted:
3517 any_restricted = any_restricted or is_restricted
3518 return not any_restricted
3520 def extract_subtitles(self, *args, **kwargs):
3521 if (self.get_param('writesubtitles
', False)
3522 or self.get_param('listsubtitles
')):
3523 return self._get_subtitles(*args, **kwargs)
3526 def _get_subtitles(self, *args, **kwargs):
3527 raise NotImplementedError('This method must be implemented by subclasses
')
3529 def extract_comments(self, *args, **kwargs):
3530 if not self.get_param('getcomments
'):
3532 generator = self._get_comments(*args, **kwargs)
3538 comments.append(next(generator))
3539 except KeyboardInterrupt:
3541 self.to_screen('Interrupted by user
')
3542 except StopIteration:
3544 comment_count = len(comments)
3545 self.to_screen(f'Extracted {comment_count} comments
')
3547 'comments
': comments,
3548 'comment_count
': None if interrupted else comment_count
3552 def _get_comments(self, *args, **kwargs):
3553 raise NotImplementedError('This method must be implemented by subclasses
')
3556 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
3557 """ Merge subtitle items for one language. Items with duplicated URLs
3558 will be dropped. """
3559 list1_urls = set([item['url
'] for item in subtitle_list1])
3560 ret = list(subtitle_list1)
3561 ret.extend([item for item in subtitle_list2 if item['url
'] not in list1_urls])
3565 def _merge_subtitles(cls, *dicts, target=None):
3566 """ Merge subtitle dictionaries, language by language. """
3570 for lang, subs in d.items():
3571 target[lang] = cls._merge_subtitle_items(target.get(lang, []), subs)
3574 def extract_automatic_captions(self, *args, **kwargs):
3575 if (self.get_param('writeautomaticsub
', False)
3576 or self.get_param('listsubtitles
')):
3577 return self._get_automatic_captions(*args, **kwargs)
3580 def _get_automatic_captions(self, *args, **kwargs):
3581 raise NotImplementedError('This method must be implemented by subclasses
')
3583 def mark_watched(self, *args, **kwargs):
3584 if not self.get_param('mark_watched
', False):
3586 if (self._get_login_info()[0] is not None
3587 or self.get_param('cookiefile
')
3588 or self.get_param('cookiesfrombrowser
')):
3589 self._mark_watched(*args, **kwargs)
3591 def _mark_watched(self, *args, **kwargs):
3592 raise NotImplementedError('This method must be implemented by subclasses
')
3594 def geo_verification_headers(self):
3596 geo_verification_proxy = self.get_param('geo_verification_proxy
')
3597 if geo_verification_proxy:
3598 headers['Ytdl
-request
-proxy
'] = geo_verification_proxy
3601 def _generic_id(self, url):
3602 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
3604 def _generic_title(self, url):
3605 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
3608 def _availability(is_private=None, needs_premium=None, needs_subscription=None, needs_auth=None, is_unlisted=None):
3609 all_known = all(map(
3610 lambda x: x is not None,
3611 (is_private, needs_premium, needs_subscription, needs_auth, is_unlisted)))
3613 'private
' if is_private
3614 else 'premium_only
' if needs_premium
3615 else 'subscriber_only
' if needs_subscription
3616 else 'needs_auth
' if needs_auth
3617 else 'unlisted
' if is_unlisted
3618 else 'public
' if all_known
3621 def _configuration_arg(self, key, default=NO_DEFAULT, casesense=False):
3623 @returns A list of values for the extractor argument given by "key"
3624 or "default" if no such key is present
3625 @param default The default value to return when the key is not present (default: [])
3626 @param casesense When false, the values are converted to lower case
3629 self._downloader.params, ('extractor_args
', self.ie_key().lower(), key))
3631 return [] if default is NO_DEFAULT else default
3632 return list(val) if casesense else [x.lower() for x in val]
3635 class SearchInfoExtractor(InfoExtractor):
3637 Base class for paged search queries extractors.
3638 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
3639 Instances should define _SEARCH_KEY and optionally _MAX_RESULTS
3642 _MAX_RESULTS = float('inf
')
3645 def _make_valid_url(cls):
3646 return r'%s(?P
<prefix
>|
[1-9][0-9]*|all
):(?P
<query
>[\s\S
]+)' % cls._SEARCH_KEY
3649 def suitable(cls, url):
3650 return re.match(cls._make_valid_url(), url) is not None
3652 def _real_extract(self, query):
3653 mobj = re.match(self._make_valid_url(), query)
3655 raise ExtractorError('Invalid search query
"%s"' % query)
3657 prefix = mobj.group('prefix
')
3658 query = mobj.group('query
')
3660 return self._get_n_results(query, 1)
3661 elif prefix == 'all
':
3662 return self._get_n_results(query, self._MAX_RESULTS)
3666 raise ExtractorError('invalid download number
%s for query
"%s"' % (n, query))
3667 elif n > self._MAX_RESULTS:
3668 self.report_warning('%s returns
max %i results (you requested
%i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3669 n = self._MAX_RESULTS
3670 return self._get_n_results(query, n)
3672 def _get_n_results(self, query, n):
3673 """Get a specified number of results for a query.
3674 Either this function or _search_results must be overridden by subclasses """
3675 return self.playlist_result(
3676 itertools.islice(self._search_results(query), 0, None if n == float('inf
') else n),
3679 def _search_results(self, query):
3680 """Returns an iterator of search results"""
3681 raise NotImplementedError('This method must be implemented by subclasses
')
3684 def SEARCH_KEY(self):
3685 return self._SEARCH_KEY