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 (
1083 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1084 self.report_warning(msg)
1085 if method is not None:
1086 msg = '%s. %s' % (msg, self._LOGIN_HINTS[method])
1087 raise ExtractorError(msg, expected=True)
1089 def raise_geo_restricted(
1090 self, msg='This video
is not available
from your location due to geo restriction
',
1091 countries=None, metadata_available=False):
1092 if metadata_available and (
1093 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1094 self.report_warning(msg)
1096 raise GeoRestrictedError(msg, countries=countries)
1098 def raise_no_formats(self, msg, expected=False, video_id=None):
1100 self.get_param('ignore_no_formats_error
') or self.get_param('wait_for_video
')):
1101 self.report_warning(msg, video_id)
1102 elif isinstance(msg, ExtractorError):
1105 raise ExtractorError(msg, expected=expected, video_id=video_id)
1107 # Methods for following #608
1109 def url_result(url, ie=None, video_id=None, video_title=None, **kwargs):
1110 """Returns a URL that points to a page that should be processed"""
1111 # TODO: ie should be the class used for getting the info
1112 video_info = {'_type
': 'url
',
1115 video_info.update(kwargs)
1116 if video_id is not None:
1117 video_info['id'] = video_id
1118 if video_title is not None:
1119 video_info['title
'] = video_title
1122 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None):
1124 self.url_result(self._proto_relative_url(getter(m) if getter else m), ie)
1126 return self.playlist_result(
1127 urls, playlist_id=playlist_id, playlist_title=playlist_title)
1130 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None, **kwargs):
1131 """Returns a playlist"""
1132 video_info = {'_type
': 'playlist
',
1134 video_info.update(kwargs)
1136 video_info['id'] = playlist_id
1138 video_info['title
'] = playlist_title
1139 if playlist_description is not None:
1140 video_info['description
'] = playlist_description
1143 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1145 Perform a regex search on the given string, using a single or a list of
1146 patterns returning the first matching group.
1147 In case of failure return a default value or raise a WARNING or a
1148 RegexNotFoundError, depending on fatal, specifying the field name.
1150 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
1151 mobj = re.search(pattern, string, flags)
1154 mobj = re.search(p, string, flags)
1158 _name = self._downloader._format_err(name, self._downloader.Styles.EMPHASIS)
1162 # return the first matching group
1163 return next(g for g in mobj.groups() if g is not None)
1164 elif isinstance(group, (list, tuple)):
1165 return tuple(mobj.group(g) for g in group)
1167 return mobj.group(group)
1168 elif default is not NO_DEFAULT:
1171 raise RegexNotFoundError('Unable to extract
%s' % _name)
1173 self.report_warning('unable to extract
%s' % _name + bug_reports_message())
1176 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1178 Like _search_regex, but strips HTML tags and unescapes entities.
1180 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1182 return clean_html(res).strip()
1186 def _get_netrc_login_info(self, netrc_machine=None):
1189 netrc_machine = netrc_machine or self._NETRC_MACHINE
1191 if self.get_param('usenetrc
', False):
1193 netrc_file = compat_expanduser(self.get_param('netrc_location
') or '~
')
1194 if os.path.isdir(netrc_file):
1195 netrc_file = os.path.join(netrc_file, '.netrc
')
1196 info = netrc.netrc(file=netrc_file).authenticators(netrc_machine)
1197 if info is not None:
1201 raise netrc.NetrcParseError(
1202 'No authenticators
for %s' % netrc_machine)
1203 except (IOError, netrc.NetrcParseError) as err:
1204 self.report_warning(
1205 'parsing
.netrc
: %s' % error_to_compat_str(err))
1207 return username, password
1209 def _get_login_info(self, username_option='username
', password_option='password
', netrc_machine=None):
1211 Get the login info as (username, password)
1212 First look for the manually specified credentials using username_option
1213 and password_option as keys in params dictionary. If no such credentials
1214 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1216 If there's no info available
, return (None, None)
1219 # Attempt to use provided username and password or .netrc data
1220 username = self.get_param(username_option)
1221 if username is not None:
1222 password = self.get_param(password_option)
1224 username, password = self._get_netrc_login_info(netrc_machine)
1226 return username, password
1228 def _get_tfa_info(self, note='two-factor verification code'):
1230 Get the two
-factor authentication info
1231 TODO
- asking the user will be required
for sms
/phone verify
1232 currently just uses the command line option
1233 If there
's no info available, return None
1236 tfa = self.get_param('twofactor
')
1240 return compat_getpass('Type
%s and press
[Return
]: ' % note)
1242 # Helper functions for extracting OpenGraph info
1244 def _og_regexes(prop):
1245 content_re = r'content
=(?
:"([^"]+?
)"|\'([^\']+?)\'|\s*([^\s"\'=<>`
]+?
))'
1246 property_re = (r'(?
:name|
property)=(?
:\'og
[:-]%(prop)s\'|
"og[:-]%(prop)s"|\s
*og
[:-]%(prop)s\b)'
1247 % {'prop': re.escape(prop)})
1248 template = r'<meta
[^
>]+?
%s[^
>]+?
%s'
1250 template % (property_re, content_re),
1251 template % (content_re, property_re),
1255 def _meta_regex(prop):
1256 return r'''(?isx)<meta
1257 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1258 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1260 def _og_search_property(self, prop, html, name=None, **kargs):
1261 prop = variadic(prop)
1263 name = 'OpenGraph
%s' % prop[0]
1266 og_regexes.extend(self._og_regexes(p))
1267 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1270 return unescapeHTML(escaped)
1272 def _og_search_thumbnail(self, html, **kargs):
1273 return self._og_search_property('image
', html, 'thumbnail URL
', fatal=False, **kargs)
1275 def _og_search_description(self, html, **kargs):
1276 return self._og_search_property('description
', html, fatal=False, **kargs)
1278 def _og_search_title(self, html, **kargs):
1279 return self._og_search_property('title
', html, **kargs)
1281 def _og_search_video_url(self, html, name='video url
', secure=True, **kargs):
1282 regexes = self._og_regexes('video
') + self._og_regexes('video
:url
')
1284 regexes = self._og_regexes('video
:secure_url
') + regexes
1285 return self._html_search_regex(regexes, html, name, **kargs)
1287 def _og_search_url(self, html, **kargs):
1288 return self._og_search_property('url
', html, **kargs)
1290 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1291 name = variadic(name)
1292 if display_name is None:
1293 display_name = name[0]
1294 return self._html_search_regex(
1295 [self._meta_regex(n) for n in name],
1296 html, display_name, fatal=fatal, group='content
', **kwargs)
1298 def _dc_search_uploader(self, html):
1299 return self._html_search_meta('dc
.creator
', html, 'uploader
')
1301 def _rta_search(self, html):
1302 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1303 if re.search(r'(?ix
)<meta\s
+name
="rating"\s
+'
1304 r' content
="RTA-5042-1996-1400-1577-RTA"',
1309 def _media_rating_search(self, html):
1310 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1311 rating = self._html_search_meta('rating
', html)
1323 return RATING_TABLE.get(rating.lower())
1325 def _family_friendly_search(self, html):
1326 # See http://schema.org/VideoObject
1327 family_friendly = self._html_search_meta(
1328 'isFamilyFriendly
', html, default=None)
1330 if not family_friendly:
1339 return RATING_TABLE.get(family_friendly.lower())
1341 def _twitter_search_player(self, html):
1342 return self._html_search_meta('twitter
:player
', html,
1343 'twitter card player
')
1345 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
1346 json_ld_list = list(re.finditer(JSON_LD_RE, html))
1347 default = kwargs.get('default
', NO_DEFAULT)
1348 # JSON-LD may be malformed and thus `fatal` should be respected.
1349 # At the same time `default` may be passed that assumes `fatal=False`
1350 # for _search_regex. Let's simulate the same behavior here
as well
.
1351 fatal
= kwargs
.get('fatal', True) if default
is NO_DEFAULT
else False
1353 for mobj
in json_ld_list
:
1354 json_ld_item
= self
._parse
_json
(
1355 mobj
.group('json_ld'), video_id
, fatal
=fatal
)
1356 if not json_ld_item
:
1358 if isinstance(json_ld_item
, dict):
1359 json_ld
.append(json_ld_item
)
1360 elif isinstance(json_ld_item
, (list, tuple)):
1361 json_ld
.extend(json_ld_item
)
1363 json_ld
= self
._json
_ld
(json_ld
, video_id
, fatal
=fatal
, expected_type
=expected_type
)
1366 if default
is not NO_DEFAULT
:
1369 raise RegexNotFoundError('Unable to extract JSON-LD')
1371 self
.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1374 def _json_ld(self
, json_ld
, video_id
, fatal
=True, expected_type
=None):
1375 if isinstance(json_ld
, compat_str
):
1376 json_ld
= self
._parse
_json
(json_ld
, video_id
, fatal
=fatal
)
1380 if not isinstance(json_ld
, (list, tuple, dict)):
1382 if isinstance(json_ld
, dict):
1385 INTERACTION_TYPE_MAP
= {
1386 'CommentAction': 'comment',
1387 'AgreeAction': 'like',
1388 'DisagreeAction': 'dislike',
1389 'LikeAction': 'like',
1390 'DislikeAction': 'dislike',
1391 'ListenAction': 'view',
1392 'WatchAction': 'view',
1393 'ViewAction': 'view',
1396 def extract_interaction_type(e
):
1397 interaction_type
= e
.get('interactionType')
1398 if isinstance(interaction_type
, dict):
1399 interaction_type
= interaction_type
.get('@type')
1400 return str_or_none(interaction_type
)
1402 def extract_interaction_statistic(e
):
1403 interaction_statistic
= e
.get('interactionStatistic')
1404 if isinstance(interaction_statistic
, dict):
1405 interaction_statistic
= [interaction_statistic
]
1406 if not isinstance(interaction_statistic
, list):
1408 for is_e
in interaction_statistic
:
1409 if not isinstance(is_e
, dict):
1411 if is_e
.get('@type') != 'InteractionCounter':
1413 interaction_type
= extract_interaction_type(is_e
)
1414 if not interaction_type
:
1416 # For interaction count some sites provide string instead of
1417 # an integer (as per spec) with non digit characters (e.g. ",")
1418 # so extracting count with more relaxed str_to_int
1419 interaction_count
= str_to_int(is_e
.get('userInteractionCount'))
1420 if interaction_count
is None:
1422 count_kind
= INTERACTION_TYPE_MAP
.get(interaction_type
.split('/')[-1])
1425 count_key
= '%s_count' % count_kind
1426 if info
.get(count_key
) is not None:
1428 info
[count_key
] = interaction_count
1430 def extract_video_object(e
):
1431 assert e
['@type'] == 'VideoObject'
1432 author
= e
.get('author')
1434 'url': url_or_none(e
.get('contentUrl')),
1435 'title': unescapeHTML(e
.get('name')),
1436 'description': unescapeHTML(e
.get('description')),
1437 'thumbnail': url_or_none(e
.get('thumbnailUrl') or e
.get('thumbnailURL')),
1438 'duration': parse_duration(e
.get('duration')),
1439 'timestamp': unified_timestamp(e
.get('uploadDate')),
1440 # author can be an instance of 'Organization' or 'Person' types.
1441 # both types can have 'name' property(inherited from 'Thing' type). [1]
1442 # however some websites are using 'Text' type instead.
1443 # 1. https://schema.org/VideoObject
1444 'uploader': author
.get('name') if isinstance(author
, dict) else author
if isinstance(author
, compat_str
) else None,
1445 'filesize': float_or_none(e
.get('contentSize')),
1446 'tbr': int_or_none(e
.get('bitrate')),
1447 'width': int_or_none(e
.get('width')),
1448 'height': int_or_none(e
.get('height')),
1449 'view_count': int_or_none(e
.get('interactionCount')),
1451 extract_interaction_statistic(e
)
1455 item_type
= e
.get('@type')
1456 if expected_type
is not None and expected_type
!= item_type
:
1458 rating
= traverse_obj(e
, ('aggregateRating', 'ratingValue'), expected_type
=float_or_none
)
1459 if rating
is not None:
1460 info
['average_rating'] = rating
1461 if item_type
in ('TVEpisode', 'Episode'):
1462 episode_name
= unescapeHTML(e
.get('name'))
1464 'episode': episode_name
,
1465 'episode_number': int_or_none(e
.get('episodeNumber')),
1466 'description': unescapeHTML(e
.get('description')),
1468 if not info
.get('title') and episode_name
:
1469 info
['title'] = episode_name
1470 part_of_season
= e
.get('partOfSeason')
1471 if isinstance(part_of_season
, dict) and part_of_season
.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1473 'season': unescapeHTML(part_of_season
.get('name')),
1474 'season_number': int_or_none(part_of_season
.get('seasonNumber')),
1476 part_of_series
= e
.get('partOfSeries') or e
.get('partOfTVSeries')
1477 if isinstance(part_of_series
, dict) and part_of_series
.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1478 info
['series'] = unescapeHTML(part_of_series
.get('name'))
1479 elif item_type
== 'Movie':
1481 'title': unescapeHTML(e
.get('name')),
1482 'description': unescapeHTML(e
.get('description')),
1483 'duration': parse_duration(e
.get('duration')),
1484 'timestamp': unified_timestamp(e
.get('dateCreated')),
1486 elif item_type
in ('Article', 'NewsArticle'):
1488 'timestamp': parse_iso8601(e
.get('datePublished')),
1489 'title': unescapeHTML(e
.get('headline')),
1490 'description': unescapeHTML(e
.get('articleBody')),
1492 elif item_type
== 'VideoObject':
1493 extract_video_object(e
)
1494 if expected_type
is None:
1498 video
= e
.get('video')
1499 if isinstance(video
, dict) and video
.get('@type') == 'VideoObject':
1500 extract_video_object(video
)
1501 if expected_type
is None:
1505 return dict((k
, v
) for k
, v
in info
.items() if v
is not None)
1507 def _search_nextjs_data(self
, webpage
, video_id
, **kw
):
1508 return self
._parse
_json
(
1510 r
'(?s)<script[^>]+id=[\'"]__NEXT_DATA__[\'"][^
>]*>([^
<]+)</script
>',
1511 webpage, 'next
.js data
', **kw),
1515 def _hidden_inputs(html):
1516 html = re.sub(r'<!--(?
:(?
!<!--).)*-->', '', html)
1518 for input in re.findall(r'(?i
)(<input[^
>]+>)', html):
1519 attrs = extract_attributes(input)
1522 if attrs.get('type') not in ('hidden
', 'submit
'):
1524 name = attrs.get('name
') or attrs.get('id')
1525 value = attrs.get('value
')
1526 if name and value is not None:
1527 hidden_inputs[name] = value
1528 return hidden_inputs
1530 def _form_hidden_inputs(self, form_id, html):
1531 form = self._search_regex(
1532 r'(?
is)<form
[^
>]+?
id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1533 html, '%s form' % form_id, group='form')
1534 return self._hidden_inputs(form)
1537 regex = r' *((?P<reverse>\+)?(?P<field>[a-zA-Z0-9_]+)((?P<separator>[~:])(?P<limit>.*?))?)? *$'
1539 default = ('hidden', 'aud_or_vid', 'hasvid', 'ie_pref', 'lang', 'quality',
1540 'res', 'fps', 'hdr:12', 'codec:vp9.2', 'size', 'br', 'asr',
1541 'proto', 'ext', 'hasaud', 'source', 'id') # These must not be aliases
1542 ytdl_default = ('hasaud', 'lang', 'quality', 'tbr', 'filesize', 'vbr',
1543 'height', 'width', 'proto', 'vext', 'abr', 'aext',
1544 'fps', 'fs_approx', 'source', 'id')
1547 'vcodec': {'type': 'ordered', 'regex': True,
1548 'order': ['av0?1', 'vp0?9.2', 'vp0?9', '[hx]265|he?vc?', '[hx]264|avc', 'vp0?8', 'mp4v|h263', 'theora', '', None, 'none']},
1549 'acodec': {'type': 'ordered', 'regex': True,
1550 'order': ['opus', 'vorbis', 'aac', 'mp?4a?', 'mp3', 'e-?a?c-?3', 'ac-?3', 'dts', '', None, 'none']},
1551 'hdr': {'type': 'ordered', 'regex': True, 'field': 'dynamic_range',
1552 'order': ['dv', '(hdr)?12', r'(hdr)?10\+', '(hdr)?10', 'hlg', '', 'sdr', None]},
1553 'proto': {'type': 'ordered', 'regex': True, 'field': 'protocol',
1554 'order': ['(ht|f)tps', '(ht|f)tp$', 'm3u8.*', '.*dash', 'websocket_frag', 'rtmpe?', '', 'mms|rtsp', 'ws|websocket', 'f4']},
1555 'vext': {'type': 'ordered', 'field': 'video_ext',
1556 'order': ('mp4', 'webm', 'flv', '', 'none'),
1557 'order_free': ('webm', 'mp4', 'flv', '', 'none')},
1558 'aext': {'type': 'ordered', 'field': 'audio_ext',
1559 'order': ('m4a', 'aac', 'mp3', 'ogg', 'opus', 'webm', '', 'none'),
1560 'order_free': ('opus', 'ogg', 'webm', 'm4a', 'mp3', 'aac', '', 'none')},
1561 'hidden': {'visible': False, 'forced': True, 'type': 'extractor', 'max': -1000},
1562 'aud_or_vid': {'visible': False, 'forced': True, 'type': 'multiple',
1563 'field': ('vcodec', 'acodec'),
1564 'function': lambda it: int(any(v != 'none' for v in it))},
1565 'ie_pref': {'priority': True, 'type': 'extractor'},
1566 'hasvid': {'priority': True, 'field': 'vcodec', 'type': 'boolean', 'not_in_list': ('none',)},
1567 'hasaud': {'field': 'acodec', 'type': 'boolean', 'not_in_list': ('none',)},
1568 'lang': {'convert': 'float', 'field': 'language_preference', 'default': -1},
1569 'quality': {'convert': 'float', 'default': -1},
1570 'filesize': {'convert': 'bytes'},
1571 'fs_approx': {'convert': 'bytes', 'field': 'filesize_approx'},
1572 'id': {'convert': 'string', 'field': 'format_id'},
1573 'height': {'convert': 'float_none'},
1574 'width': {'convert': 'float_none'},
1575 'fps': {'convert': 'float_none'},
1576 'tbr': {'convert': 'float_none'},
1577 'vbr': {'convert': 'float_none'},
1578 'abr': {'convert': 'float_none'},
1579 'asr': {'convert': 'float_none'},
1580 'source': {'convert': 'float', 'field': 'source_preference', 'default': -1},
1582 'codec': {'type': 'combined', 'field': ('vcodec', 'acodec')},
1583 'br': {'type': 'combined', 'field': ('tbr', 'vbr', 'abr'), 'same_limit': True},
1584 'size': {'type': 'combined', 'same_limit': True, 'field': ('filesize', 'fs_approx')},
1585 'ext': {'type': 'combined', 'field': ('vext', 'aext')},
1586 'res': {'type': 'multiple', 'field': ('height', 'width'),
1587 'function': lambda it: (lambda l: min(l) if l else 0)(tuple(filter(None, it)))},
1590 'dimension': {'type': 'alias', 'field': 'res'},
1591 'resolution': {'type': 'alias', 'field': 'res'},
1592 'extension': {'type': 'alias', 'field': 'ext'},
1593 'bitrate': {'type': 'alias', 'field': 'br'},
1594 'total_bitrate': {'type': 'alias', 'field': 'tbr'},
1595 'video_bitrate': {'type': 'alias', 'field': 'vbr'},
1596 'audio_bitrate': {'type': 'alias', 'field': 'abr'},
1597 'framerate': {'type': 'alias', 'field': 'fps'},
1598 'language_preference': {'type': 'alias', 'field': 'lang'},
1599 'protocol': {'type': 'alias', 'field': 'proto'},
1600 'source_preference': {'type': 'alias', 'field': 'source'},
1601 'filesize_approx': {'type': 'alias', 'field': 'fs_approx'},
1602 'filesize_estimate': {'type': 'alias', 'field': 'size'},
1603 'samplerate': {'type': 'alias', 'field': 'asr'},
1604 'video_ext': {'type': 'alias', 'field': 'vext'},
1605 'audio_ext': {'type': 'alias', 'field': 'aext'},
1606 'video_codec': {'type': 'alias', 'field': 'vcodec'},
1607 'audio_codec': {'type': 'alias', 'field': 'acodec'},
1608 'video': {'type': 'alias', 'field': 'hasvid'},
1609 'has_video': {'type': 'alias', 'field': 'hasvid'},
1610 'audio': {'type': 'alias', 'field': 'hasaud'},
1611 'has_audio': {'type': 'alias', 'field': 'hasaud'},
1612 'extractor': {'type': 'alias', 'field': 'ie_pref'},
1613 'preference': {'type': 'alias', 'field': 'ie_pref'},
1614 'extractor_preference': {'type': 'alias', 'field': 'ie_pref'},
1615 'format_id': {'type': 'alias', 'field': 'id'},
1618 def __init__(self, ie, field_preference):
1620 self.ydl = ie._downloader
1621 self.evaluate_params(self.ydl.params, field_preference)
1622 if ie.get_param('verbose'):
1623 self.print_verbose_info(self.ydl.write_debug)
1625 def _get_field_setting(self, field, key):
1626 if field not in self.settings:
1627 if key in ('forced', 'priority'):
1629 self.ydl.deprecation_warning(
1630 f'Using arbitrary fields ({field}) for format sorting is deprecated '
1631 'and may be removed in a future version')
1632 self.settings[field] = {}
1633 propObj = self.settings[field]
1634 if key not in propObj:
1635 type = propObj.get('type')
1637 default = 'preference' if type == 'extractor' else (field,) if type in ('combined', 'multiple') else field
1638 elif key == 'convert':
1639 default = 'order' if type == 'ordered' else 'float_string' if field else 'ignore'
1641 default = {'type': 'field', 'visible': True, 'order': [], 'not_in_list': (None,)}.get(key, None)
1642 propObj[key] = default
1645 def _resolve_field_value(self, field, value, convertNone=False):
1650 value = value.lower()
1651 conversion = self._get_field_setting(field, 'convert')
1652 if conversion == 'ignore':
1654 if conversion == 'string':
1656 elif conversion == 'float_none':
1657 return float_or_none(value)
1658 elif conversion == 'bytes':
1659 return FileDownloader.parse_bytes(value)
1660 elif conversion == 'order':
1661 order_list = (self._use_free_order and self._get_field_setting(field, 'order_free')) or self._get_field_setting(field, 'order')
1662 use_regex = self._get_field_setting(field, 'regex')
1663 list_length = len(order_list)
1664 empty_pos = order_list.index('') if '' in order_list else list_length + 1
1665 if use_regex and value is not None:
1666 for i, regex in enumerate(order_list):
1667 if regex and re.match(regex, value):
1668 return list_length - i
1669 return list_length - empty_pos # not in list
1670 else: # not regex or value = None
1671 return list_length - (order_list.index(value) if value in order_list else empty_pos)
1673 if value.isnumeric():
1676 self.settings[field]['convert'] = 'string'
1679 def evaluate_params(self, params, sort_extractor):
1680 self._use_free_order = params.get('prefer_free_formats', False)
1681 self._sort_user = params.get('format_sort', [])
1682 self._sort_extractor = sort_extractor
1684 def add_item(field, reverse, closest, limit_text):
1685 field = field.lower()
1686 if field in self._order:
1688 self._order.append(field)
1689 limit = self._resolve_field_value(field, limit_text)
1692 'closest': False if limit is None else closest,
1693 'limit_text': limit_text,
1695 if field in self.settings:
1696 self.settings[field].update(data)
1698 self.settings[field] = data
1701 tuple(field for field in self.default if self._get_field_setting(field, 'forced'))
1702 + (tuple() if params.get('format_sort_force', False)
1703 else tuple(field for field in self.default if self._get_field_setting(field, 'priority')))
1704 + tuple(self._sort_user) + tuple(sort_extractor) + self.default)
1706 for item in sort_list:
1707 match = re.match(self.regex, item)
1709 raise ExtractorError('Invalid format sort string "%s" given by extractor' % item)
1710 field = match.group('field')
1713 if self._get_field_setting(field, 'type') == 'alias':
1714 alias, field = field, self._get_field_setting(field, 'field')
1715 self.ydl.deprecation_warning(
1716 f'Format sorting alias {alias} is deprecated '
1717 f'and may be removed in a future version. Please use {field} instead')
1718 reverse = match.group('reverse') is not None
1719 closest = match.group('separator') == '~'
1720 limit_text = match.group('limit')
1722 has_limit = limit_text is not None
1723 has_multiple_fields = self._get_field_setting(field, 'type') == 'combined'
1724 has_multiple_limits = has_limit and has_multiple_fields and not self._get_field_setting(field, 'same_limit')
1726 fields = self._get_field_setting(field, 'field') if has_multiple_fields else (field,)
1727 limits = limit_text.split(':') if has_multiple_limits else (limit_text,) if has_limit else tuple()
1728 limit_count = len(limits)
1729 for (i, f) in enumerate(fields):
1730 add_item(f, reverse, closest,
1731 limits[i] if i < limit_count
1732 else limits[0] if has_limit and not has_multiple_limits
1735 def print_verbose_info(self, write_debug):
1737 write_debug('Sort order given by user: %s' % ', '.join(self._sort_user))
1738 if self._sort_extractor:
1739 write_debug('Sort order given by extractor: %s' % ', '.join(self._sort_extractor))
1740 write_debug('Formats sorted by: %s' % ', '.join(['%s%s%s' % (
1741 '+' if self._get_field_setting(field, 'reverse') else '', field,
1742 '%s%s(%s)' % ('~' if self._get_field_setting(field, 'closest') else ':',
1743 self._get_field_setting(field, 'limit_text'),
1744 self._get_field_setting(field, 'limit'))
1745 if self._get_field_setting(field, 'limit_text') is not None else '')
1746 for field in self._order if self._get_field_setting(field, 'visible')]))
1748 def _calculate_field_preference_from_value(self, format, field, type, value):
1749 reverse = self._get_field_setting(field, 'reverse')
1750 closest = self._get_field_setting(field, 'closest')
1751 limit = self._get_field_setting(field, 'limit')
1753 if type == 'extractor':
1754 maximum = self._get_field_setting(field, 'max')
1755 if value is None or (maximum is not None and value >= maximum):
1757 elif type == 'boolean':
1758 in_list = self._get_field_setting(field, 'in_list')
1759 not_in_list = self._get_field_setting(field, 'not_in_list')
1760 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
1761 elif type == 'ordered':
1762 value = self._resolve_field_value(field, value, True)
1764 # try to convert to number
1765 val_num = float_or_none(value, default=self._get_field_setting(field, 'default'))
1766 is_num = self._get_field_setting(field, 'convert') != 'string' and val_num is not None
1770 return ((-10, 0) if value is None
1771 else (1, value, 0) if not is_num # if a field has mixed strings and numbers, strings are sorted higher
1772 else (0, -abs(value - limit), value - limit if reverse else limit - value) if closest
1773 else (0, value, 0) if not reverse and (limit is None or value <= limit)
1774 else (0, -value, 0) if limit is None or (reverse and value == limit) or value > limit
1775 else (-1, value, 0))
1777 def _calculate_field_preference(self, format, field):
1778 type = self._get_field_setting(field, 'type') # extractor, boolean, ordered, field, multiple
1779 get_value = lambda f: format.get(self._get_field_setting(f, 'field'))
1780 if type == 'multiple':
1781 type = 'field' # Only 'field' is allowed in multiple for now
1782 actual_fields = self._get_field_setting(field, 'field')
1784 value = self._get_field_setting(field, 'function')(get_value(f) for f in actual_fields)
1786 value = get_value(field)
1787 return self._calculate_field_preference_from_value(format, field, type, value)
1789 def calculate_preference(self, format):
1790 # Determine missing protocol
1791 if not format.get('protocol'):
1792 format['protocol'] = determine_protocol(format)
1794 # Determine missing ext
1795 if not format.get('ext') and 'url' in format:
1796 format['ext'] = determine_ext(format['url'])
1797 if format.get('vcodec') == 'none':
1798 format['audio_ext'] = format['ext'] if format.get('acodec') != 'none' else 'none'
1799 format['video_ext'] = 'none'
1801 format['video_ext'] = format['ext']
1802 format['audio_ext'] = 'none'
1803 # if format.get('preference') is None and format.get('ext') in ('f4f', 'f4m'): # Not supported?
1804 # format['preference'] = -1000
1806 # Determine missing bitrates
1807 if format.get('tbr') is None:
1808 if format.get('vbr') is not None and format.get('abr') is not None:
1809 format['tbr'] = format.get('vbr', 0) + format.get('abr', 0)
1811 if format.get('vcodec') != 'none' and format.get('vbr') is None:
1812 format['vbr'] = format.get('tbr') - format.get('abr', 0)
1813 if format.get('acodec') != 'none' and format.get('abr') is None:
1814 format['abr'] = format.get('tbr') - format.get('vbr', 0)
1816 return tuple(self._calculate_field_preference(format, field) for field in self._order)
1818 def _sort_formats(self, formats, field_preference=[]):
1821 format_sort = self.FormatSort(self, field_preference)
1822 formats.sort(key=lambda f: format_sort.calculate_preference(f))
1824 def _check_formats(self, formats, video_id):
1826 formats[:] = filter(
1827 lambda f: self._is_valid_url(
1829 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1833 def _remove_duplicate_formats(formats):
1837 if f['url'] not in format_urls:
1838 format_urls.add(f['url'])
1839 unique_formats.append(f)
1840 formats[:] = unique_formats
1842 def _is_valid_url(self, url, video_id, item='video', headers={}):
1843 url = self._proto_relative_url(url, scheme='http:')
1844 # For now assume non HTTP(S) URLs always valid
1845 if not (url.startswith('http://') or url.startswith('https://')):
1848 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1850 except ExtractorError as e:
1852 '%s: %s URL is invalid, skipping: %s'
1853 % (video_id, item, error_to_compat_str(e.cause)))
1856 def http_scheme(self):
1857 """ Either "http
:" or "https
:", depending on the user's preferences """
1860 if self.get_param('prefer_insecure', False)
1863 def _proto_relative_url(self, url, scheme=None):
1866 if url.startswith('//'):
1868 scheme = self.http_scheme()
1873 def _sleep(self, timeout, video_id, msg_template=None):
1874 if msg_template is None:
1875 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1876 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1880 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1881 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1882 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
1883 manifest = self._download_xml(
1884 manifest_url, video_id, 'Downloading f4m manifest',
1885 'Unable to download f4m manifest',
1886 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1887 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
1888 transform_source=transform_source,
1889 fatal=fatal, data=data, headers=headers, query=query)
1891 if manifest is False:
1894 return self._parse_f4m_formats(
1895 manifest, manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1896 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1898 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1899 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1900 fatal=True, m3u8_id=None):
1901 if not isinstance(manifest, compat_etree_Element) and not fatal:
1904 # currently yt-dlp cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1905 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1906 if akamai_pv is not None and ';' in akamai_pv.text:
1907 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1908 if playerVerificationChallenge.strip() != '':
1912 manifest_version = '1.0'
1913 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1915 manifest_version = '2.0'
1916 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1917 # Remove unsupported DRM protected media from final formats
1918 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
1919 media_nodes = remove_encrypted_media(media_nodes)
1923 manifest_base_url = get_base_url(manifest)
1925 bootstrap_info = xpath_element(
1926 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1927 'bootstrap info', default=None)
1930 mime_type = xpath_text(
1931 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1932 'base URL', default=None)
1933 if mime_type and mime_type.startswith('audio/'):
1936 for i, media_el in enumerate(media_nodes):
1937 tbr = int_or_none(media_el.attrib.get('bitrate'))
1938 width = int_or_none(media_el.attrib.get('width'))
1939 height = int_or_none(media_el.attrib.get('height'))
1940 format_id = join_nonempty(f4m_id, tbr or i)
1941 # If <bootstrapInfo> is present, the specified f4m is a
1942 # stream-level manifest, and only set-level manifests may refer to
1943 # external resources. See section 11.4 and section 4 of F4M spec
1944 if bootstrap_info is None:
1946 # @href is introduced in 2.0, see section 11.6 of F4M spec
1947 if manifest_version == '2.0':
1948 media_url = media_el.attrib.get('href')
1949 if media_url is None:
1950 media_url = media_el.attrib.get('url')
1954 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1955 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1956 # If media_url is itself a f4m manifest do the recursive extraction
1957 # since bitrates in parent manifest (this one) and media_url manifest
1958 # may differ leading to inability to resolve the format by requested
1959 # bitrate in f4m downloader
1960 ext = determine_ext(manifest_url)
1962 f4m_formats = self._extract_f4m_formats(
1963 manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1964 transform_source=transform_source, fatal=fatal)
1965 # Sometimes stream-level manifest contains single media entry that
1966 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1967 # At the same time parent's media entry in set-level manifest may
1968 # contain it. We will copy it from parent in such cases.
1969 if len(f4m_formats) == 1:
1972 'tbr': f.get('tbr') or tbr,
1973 'width': f.get('width') or width,
1974 'height': f.get('height') or height,
1975 'format_id': f.get('format_id') if not tbr else format_id,
1978 formats.extend(f4m_formats)
1981 formats.extend(self._extract_m3u8_formats(
1982 manifest_url, video_id, 'mp4', preference=preference,
1983 quality=quality, m3u8_id=m3u8_id, fatal=fatal))
1986 'format_id': format_id,
1987 'url': manifest_url,
1988 'manifest_url': manifest_url,
1989 'ext': 'flv' if bootstrap_info is not None else None,
1995 'preference': preference,
2000 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, quality=None, m3u8_id=None):
2002 'format_id': join_nonempty(m3u8_id, 'meta'),
2006 'preference': preference - 100 if preference else -100,
2008 'resolution': 'multiple',
2009 'format_note': 'Quality selection URL',
2012 def _report_ignoring_subs(self, name):
2013 self.report_warning(bug_reports_message(
2014 f'Ignoring subtitle tracks found in the {name} manifest; '
2015 'if any subtitle tracks are missing,'
2018 def _extract_m3u8_formats(self, *args, **kwargs):
2019 fmts, subs = self._extract_m3u8_formats_and_subtitles(*args, **kwargs)
2021 self._report_ignoring_subs('HLS')
2024 def _extract_m3u8_formats_and_subtitles(
2025 self, m3u8_url, video_id, ext=None, entry_protocol='m3u8_native',
2026 preference=None, quality=None, m3u8_id=None, note=None,
2027 errnote=None, fatal=True, live=False, data=None, headers={},
2030 res = self._download_webpage_handle(
2032 note='Downloading m3u8 information' if note is None else note,
2033 errnote='Failed to download m3u8 information' if errnote is None else errnote,
2034 fatal=fatal, data=data, headers=headers, query=query)
2039 m3u8_doc, urlh = res
2040 m3u8_url = urlh.geturl()
2042 return self._parse_m3u8_formats_and_subtitles(
2043 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
2044 preference=preference, quality=quality, m3u8_id=m3u8_id,
2045 note=note, errnote=errnote, fatal=fatal, live=live, data=data,
2046 headers=headers, query=query, video_id=video_id)
2048 def _parse_m3u8_formats_and_subtitles(
2049 self, m3u8_doc, m3u8_url, ext=None, entry_protocol='m3u8_native',
2050 preference=None, quality=None, m3u8_id=None, live=False, note=None,
2051 errnote=None, fatal=True, data=None, headers={}, query={},
2053 formats, subtitles = [], {}
2055 has_drm = re.search('|'.join([
2056 r'#EXT-X-FAXS-CM:', # Adobe Flash Access
2057 r'#EXT-X-(?:SESSION-)?KEY:.*?URI="skd
://', # Apple FairPlay
2060 def format_url(url):
2061 return url if re.match(r'^https?
://', url) else compat_urlparse.urljoin(m3u8_url, url)
2063 if self.get_param('hls_split_discontinuity
', False):
2064 def _extract_m3u8_playlist_indices(manifest_url=None, m3u8_doc=None):
2066 if not manifest_url:
2068 m3u8_doc = self._download_webpage(
2069 manifest_url, video_id, fatal=fatal, data=data, headers=headers,
2070 note=False, errnote='Failed to download m3u8 playlist information
')
2071 if m3u8_doc is False:
2073 return range(1 + sum(line.startswith('#EXT-X-DISCONTINUITY') for line in m3u8_doc.splitlines()))
2076 def _extract_m3u8_playlist_indices(*args
, **kwargs
):
2080 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
2081 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
2082 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
2084 # We should try extracting formats only from master playlists [1, 4.3.4],
2085 # i.e. playlists that describe available qualities. On the other hand
2086 # media playlists [1, 4.3.3] should be returned as is since they contain
2087 # just the media without qualities renditions.
2088 # Fortunately, master playlist can be easily distinguished from media
2089 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
2090 # master playlist tags MUST NOT appear in a media playlist and vice versa.
2091 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
2092 # media playlist and MUST NOT appear in master playlist thus we can
2093 # clearly detect media playlist with this criterion.
2095 if '#EXT-X-TARGETDURATION' in m3u8_doc
: # media playlist, return as is
2097 'format_id': join_nonempty(m3u8_id
, idx
),
2098 'format_index': idx
,
2101 'protocol': entry_protocol
,
2102 'preference': preference
,
2105 } for idx
in _extract_m3u8_playlist_indices(m3u8_doc
=m3u8_doc
)]
2107 return formats
, subtitles
2110 last_stream_inf
= {}
2112 def extract_media(x_media_line
):
2113 media
= parse_m3u8_attributes(x_media_line
)
2114 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
2115 media_type
, group_id
, name
= media
.get('TYPE'), media
.get('GROUP-ID'), media
.get('NAME')
2116 if not (media_type
and group_id
and name
):
2118 groups
.setdefault(group_id
, []).append(media
)
2119 # <https://tools.ietf.org/html/rfc8216#section-4.3.4.1>
2120 if media_type
== 'SUBTITLES':
2121 # According to RFC 8216 §4.3.4.2.1, URI is REQUIRED in the
2122 # EXT-X-MEDIA tag if the media type is SUBTITLES.
2123 # However, lack of URI has been spotted in the wild.
2124 # e.g. NebulaIE; see https://github.com/yt-dlp/yt-dlp/issues/339
2125 if not media
.get('URI'):
2127 url
= format_url(media
['URI'])
2130 'ext': determine_ext(url
),
2132 if sub_info
['ext'] == 'm3u8':
2133 # Per RFC 8216 §3.1, the only possible subtitle format m3u8
2134 # files may contain is WebVTT:
2135 # <https://tools.ietf.org/html/rfc8216#section-3.1>
2136 sub_info
['ext'] = 'vtt'
2137 sub_info
['protocol'] = 'm3u8_native'
2138 lang
= media
.get('LANGUAGE') or 'und'
2139 subtitles
.setdefault(lang
, []).append(sub_info
)
2140 if media_type
not in ('VIDEO', 'AUDIO'):
2142 media_url
= media
.get('URI')
2144 manifest_url
= format_url(media_url
)
2146 'format_id': join_nonempty(m3u8_id
, group_id
, name
, idx
),
2147 'format_note': name
,
2148 'format_index': idx
,
2149 'url': manifest_url
,
2150 'manifest_url': m3u8_url
,
2151 'language': media
.get('LANGUAGE'),
2153 'protocol': entry_protocol
,
2154 'preference': preference
,
2156 'vcodec': 'none' if media_type
== 'AUDIO' else None,
2157 } for idx
in _extract_m3u8_playlist_indices(manifest_url
))
2159 def build_stream_name():
2160 # Despite specification does not mention NAME attribute for
2161 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
2162 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
2163 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
2164 stream_name
= last_stream_inf
.get('NAME')
2167 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
2168 # from corresponding rendition group
2169 stream_group_id
= last_stream_inf
.get('VIDEO')
2170 if not stream_group_id
:
2172 stream_group
= groups
.get(stream_group_id
)
2173 if not stream_group
:
2174 return stream_group_id
2175 rendition
= stream_group
[0]
2176 return rendition
.get('NAME') or stream_group_id
2178 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
2179 # chance to detect video only formats when EXT-X-STREAM-INF tags
2180 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
2181 for line
in m3u8_doc
.splitlines():
2182 if line
.startswith('#EXT-X-MEDIA:'):
2185 for line
in m3u8_doc
.splitlines():
2186 if line
.startswith('#EXT-X-STREAM-INF:'):
2187 last_stream_inf
= parse_m3u8_attributes(line
)
2188 elif line
.startswith('#') or not line
.strip():
2191 tbr
= float_or_none(
2192 last_stream_inf
.get('AVERAGE-BANDWIDTH')
2193 or last_stream_inf
.get('BANDWIDTH'), scale
=1000)
2194 manifest_url
= format_url(line
.strip())
2196 for idx
in _extract_m3u8_playlist_indices(manifest_url
):
2197 format_id
= [m3u8_id
, None, idx
]
2198 # Bandwidth of live streams may differ over time thus making
2199 # format_id unpredictable. So it's better to keep provided
2202 stream_name
= build_stream_name()
2203 format_id
[1] = stream_name
or '%d' % (tbr
or len(formats
))
2205 'format_id': join_nonempty(*format_id
),
2206 'format_index': idx
,
2207 'url': manifest_url
,
2208 'manifest_url': m3u8_url
,
2211 'fps': float_or_none(last_stream_inf
.get('FRAME-RATE')),
2212 'protocol': entry_protocol
,
2213 'preference': preference
,
2216 resolution
= last_stream_inf
.get('RESOLUTION')
2218 mobj
= re
.search(r
'(?P<width>\d+)[xX](?P<height>\d+)', resolution
)
2220 f
['width'] = int(mobj
.group('width'))
2221 f
['height'] = int(mobj
.group('height'))
2222 # Unified Streaming Platform
2224 r
'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f
['url'])
2226 abr
, vbr
= mobj
.groups()
2227 abr
, vbr
= float_or_none(abr
, 1000), float_or_none(vbr
, 1000)
2232 codecs
= parse_codecs(last_stream_inf
.get('CODECS'))
2234 audio_group_id
= last_stream_inf
.get('AUDIO')
2235 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
2236 # references a rendition group MUST have a CODECS attribute.
2237 # However, this is not always respected, for example, [2]
2238 # contains EXT-X-STREAM-INF tag which references AUDIO
2239 # rendition group but does not have CODECS and despite
2240 # referencing an audio group it represents a complete
2241 # (with audio and video) format. So, for such cases we will
2242 # ignore references to rendition groups and treat them
2243 # as complete formats.
2244 if audio_group_id
and codecs
and f
.get('vcodec') != 'none':
2245 audio_group
= groups
.get(audio_group_id
)
2246 if audio_group
and audio_group
[0].get('URI'):
2247 # TODO: update acodec for audio only formats with
2249 f
['acodec'] = 'none'
2250 if not f
.get('ext'):
2251 f
['ext'] = 'm4a' if f
.get('vcodec') == 'none' else 'mp4'
2255 progressive_uri
= last_stream_inf
.get('PROGRESSIVE-URI')
2258 del http_f
['manifest_url']
2260 'format_id': f
['format_id'].replace('hls-', 'http-'),
2262 'url': progressive_uri
,
2264 formats
.append(http_f
)
2266 last_stream_inf
= {}
2267 return formats
, subtitles
2269 def _extract_m3u8_vod_duration(
2270 self
, m3u8_vod_url
, video_id
, note
=None, errnote
=None, data
=None, headers
={}, query={}
):
2272 m3u8_vod
= self
._download
_webpage
(
2273 m3u8_vod_url
, video_id
,
2274 note
='Downloading m3u8 VOD manifest' if note
is None else note
,
2275 errnote
='Failed to download VOD manifest' if errnote
is None else errnote
,
2276 fatal
=False, data
=data
, headers
=headers
, query
=query
)
2278 return self
._parse
_m
3u8_vod
_duration
(m3u8_vod
or '', video_id
)
2280 def _parse_m3u8_vod_duration(self
, m3u8_vod
, video_id
):
2281 if '#EXT-X-PLAYLIST-TYPE:VOD' not in m3u8_vod
:
2285 float(line
[len('#EXTINF:'):].split(',')[0])
2286 for line
in m3u8_vod
.splitlines() if line
.startswith('#EXTINF:'))) or None
2289 def _xpath_ns(path
, namespace
=None):
2293 for c
in path
.split('/'):
2294 if not c
or c
== '.':
2297 out
.append('{%s}%s' % (namespace
, c
))
2298 return '/'.join(out
)
2300 def _extract_smil_formats_and_subtitles(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None, transform_source
=None):
2301 smil
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
, transform_source
=transform_source
)
2307 namespace
= self
._parse
_smil
_namespace
(smil
)
2309 fmts
= self
._parse
_smil
_formats
(
2310 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2311 subs
= self
._parse
_smil
_subtitles
(
2312 smil
, namespace
=namespace
)
2316 def _extract_smil_formats(self
, *args
, **kwargs
):
2317 fmts
, subs
= self
._extract
_smil
_formats
_and
_subtitles
(*args
, **kwargs
)
2319 self
._report
_ignoring
_subs
('SMIL')
2322 def _extract_smil_info(self
, smil_url
, video_id
, fatal
=True, f4m_params
=None):
2323 smil
= self
._download
_smil
(smil_url
, video_id
, fatal
=fatal
)
2326 return self
._parse
_smil
(smil
, smil_url
, video_id
, f4m_params
=f4m_params
)
2328 def _download_smil(self
, smil_url
, video_id
, fatal
=True, transform_source
=None):
2329 return self
._download
_xml
(
2330 smil_url
, video_id
, 'Downloading SMIL file',
2331 'Unable to download SMIL file', fatal
=fatal
, transform_source
=transform_source
)
2333 def _parse_smil(self
, smil
, smil_url
, video_id
, f4m_params
=None):
2334 namespace
= self
._parse
_smil
_namespace
(smil
)
2336 formats
= self
._parse
_smil
_formats
(
2337 smil
, smil_url
, video_id
, namespace
=namespace
, f4m_params
=f4m_params
)
2338 subtitles
= self
._parse
_smil
_subtitles
(smil
, namespace
=namespace
)
2340 video_id
= os
.path
.splitext(url_basename(smil_url
))[0]
2344 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2345 name
= meta
.attrib
.get('name')
2346 content
= meta
.attrib
.get('content')
2347 if not name
or not content
:
2349 if not title
and name
== 'title':
2351 elif not description
and name
in ('description', 'abstract'):
2352 description
= content
2353 elif not upload_date
and name
== 'date':
2354 upload_date
= unified_strdate(content
)
2357 'id': image
.get('type'),
2358 'url': image
.get('src'),
2359 'width': int_or_none(image
.get('width')),
2360 'height': int_or_none(image
.get('height')),
2361 } for image
in smil
.findall(self
._xpath
_ns
('.//image', namespace
)) if image
.get('src')]
2365 'title': title
or video_id
,
2366 'description': description
,
2367 'upload_date': upload_date
,
2368 'thumbnails': thumbnails
,
2370 'subtitles': subtitles
,
2373 def _parse_smil_namespace(self
, smil
):
2374 return self
._search
_regex
(
2375 r
'(?i)^{([^}]+)?}smil$', smil
.tag
, 'namespace', default
=None)
2377 def _parse_smil_formats(self
, smil
, smil_url
, video_id
, namespace
=None, f4m_params
=None, transform_rtmp_url
=None):
2379 for meta
in smil
.findall(self
._xpath
_ns
('./head/meta', namespace
)):
2380 b
= meta
.get('base') or meta
.get('httpBase')
2392 media
= smil
.findall(self
._xpath
_ns
('.//video', namespace
)) + smil
.findall(self
._xpath
_ns
('.//audio', namespace
))
2393 for medium
in media
:
2394 src
= medium
.get('src')
2395 if not src
or src
in srcs
:
2399 bitrate
= float_or_none(medium
.get('system-bitrate') or medium
.get('systemBitrate'), 1000)
2400 filesize
= int_or_none(medium
.get('size') or medium
.get('fileSize'))
2401 width
= int_or_none(medium
.get('width'))
2402 height
= int_or_none(medium
.get('height'))
2403 proto
= medium
.get('proto')
2404 ext
= medium
.get('ext')
2405 src_ext
= determine_ext(src
)
2406 streamer
= medium
.get('streamer') or base
2408 if proto
== 'rtmp' or streamer
.startswith('rtmp'):
2414 'format_id': 'rtmp-%d' % (rtmp_count
if bitrate
is None else bitrate
),
2416 'filesize': filesize
,
2420 if transform_rtmp_url
:
2421 streamer
, src
= transform_rtmp_url(streamer
, src
)
2422 formats
[-1].update({
2428 src_url
= src
if src
.startswith('http') else compat_urlparse
.urljoin(base
, src
)
2429 src_url
= src_url
.strip()
2431 if proto
== 'm3u8' or src_ext
== 'm3u8':
2432 m3u8_formats
= self
._extract
_m
3u8_formats
(
2433 src_url
, video_id
, ext
or 'mp4', m3u8_id
='hls', fatal
=False)
2434 if len(m3u8_formats
) == 1:
2436 m3u8_formats
[0].update({
2437 'format_id': 'hls-%d' % (m3u8_count
if bitrate
is None else bitrate
),
2442 formats
.extend(m3u8_formats
)
2443 elif src_ext
== 'f4m':
2448 'plugin': 'flowplayer-3.2.0.1',
2450 f4m_url
+= '&' if '?' in f4m_url
else '?'
2451 f4m_url
+= compat_urllib_parse_urlencode(f4m_params
)
2452 formats
.extend(self
._extract
_f
4m
_formats
(f4m_url
, video_id
, f4m_id
='hds', fatal
=False))
2453 elif src_ext
== 'mpd':
2454 formats
.extend(self
._extract
_mpd
_formats
(
2455 src_url
, video_id
, mpd_id
='dash', fatal
=False))
2456 elif re
.search(r
'\.ism/[Mm]anifest', src_url
):
2457 formats
.extend(self
._extract
_ism
_formats
(
2458 src_url
, video_id
, ism_id
='mss', fatal
=False))
2459 elif src_url
.startswith('http') and self
._is
_valid
_url
(src
, video_id
):
2463 'ext': ext
or src_ext
or 'flv',
2464 'format_id': 'http-%d' % (bitrate
or http_count
),
2466 'filesize': filesize
,
2471 for medium
in smil
.findall(self
._xpath
_ns
('.//imagestream', namespace
)):
2472 src
= medium
.get('src')
2473 if not src
or src
in srcs
:
2479 'format_id': 'imagestream-%d' % (imgs_count
),
2481 'ext': mimetype2ext(medium
.get('type')),
2484 'width': int_or_none(medium
.get('width')),
2485 'height': int_or_none(medium
.get('height')),
2486 'format_note': 'SMIL storyboards',
2491 def _parse_smil_subtitles(self
, smil
, namespace
=None, subtitles_lang
='en'):
2494 for num
, textstream
in enumerate(smil
.findall(self
._xpath
_ns
('.//textstream', namespace
))):
2495 src
= textstream
.get('src')
2496 if not src
or src
in urls
:
2499 ext
= textstream
.get('ext') or mimetype2ext(textstream
.get('type')) or determine_ext(src
)
2500 lang
= textstream
.get('systemLanguage') or textstream
.get('systemLanguageName') or textstream
.get('lang') or subtitles_lang
2501 subtitles
.setdefault(lang
, []).append({
2507 def _extract_xspf_playlist(self
, xspf_url
, playlist_id
, fatal
=True):
2508 xspf
= self
._download
_xml
(
2509 xspf_url
, playlist_id
, 'Downloading xpsf playlist',
2510 'Unable to download xspf manifest', fatal
=fatal
)
2513 return self
._parse
_xspf
(
2514 xspf
, playlist_id
, xspf_url
=xspf_url
,
2515 xspf_base_url
=base_url(xspf_url
))
2517 def _parse_xspf(self
, xspf_doc
, playlist_id
, xspf_url
=None, xspf_base_url
=None):
2519 'xspf': 'http://xspf.org/ns/0/',
2520 's1': 'http://static.streamone.nl/player/ns/0',
2524 for track
in xspf_doc
.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP
)):
2526 track
, xpath_with_ns('./xspf:title', NS_MAP
), 'title', default
=playlist_id
)
2527 description
= xpath_text(
2528 track
, xpath_with_ns('./xspf:annotation', NS_MAP
), 'description')
2529 thumbnail
= xpath_text(
2530 track
, xpath_with_ns('./xspf:image', NS_MAP
), 'thumbnail')
2531 duration
= float_or_none(
2532 xpath_text(track
, xpath_with_ns('./xspf:duration', NS_MAP
), 'duration'), 1000)
2535 for location
in track
.findall(xpath_with_ns('./xspf:location', NS_MAP
)):
2536 format_url
= urljoin(xspf_base_url
, location
.text
)
2541 'manifest_url': xspf_url
,
2542 'format_id': location
.get(xpath_with_ns('s1:label', NS_MAP
)),
2543 'width': int_or_none(location
.get(xpath_with_ns('s1:width', NS_MAP
))),
2544 'height': int_or_none(location
.get(xpath_with_ns('s1:height', NS_MAP
))),
2546 self
._sort
_formats
(formats
)
2551 'description': description
,
2552 'thumbnail': thumbnail
,
2553 'duration': duration
,
2558 def _extract_mpd_formats(self
, *args
, **kwargs
):
2559 fmts
, subs
= self
._extract
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2561 self
._report
_ignoring
_subs
('DASH')
2564 def _extract_mpd_formats_and_subtitles(
2565 self
, mpd_url
, video_id
, mpd_id
=None, note
=None, errnote
=None,
2566 fatal
=True, data
=None, headers
={}, query={}
):
2567 res
= self
._download
_xml
_handle
(
2569 note
='Downloading MPD manifest' if note
is None else note
,
2570 errnote
='Failed to download MPD manifest' if errnote
is None else errnote
,
2571 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
2577 mpd_base_url
= base_url(urlh
.geturl())
2579 return self
._parse
_mpd
_formats
_and
_subtitles
(
2580 mpd_doc
, mpd_id
, mpd_base_url
, mpd_url
)
2582 def _parse_mpd_formats(self
, *args
, **kwargs
):
2583 fmts
, subs
= self
._parse
_mpd
_formats
_and
_subtitles
(*args
, **kwargs
)
2585 self
._report
_ignoring
_subs
('DASH')
2588 def _parse_mpd_formats_and_subtitles(
2589 self
, mpd_doc
, mpd_id
=None, mpd_base_url
='', mpd_url
=None):
2591 Parse formats from MPD manifest.
2593 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
2594 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
2595 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
2597 if not self
.get_param('dynamic_mpd', True):
2598 if mpd_doc
.get('type') == 'dynamic':
2601 namespace
= self
._search
_regex
(r
'(?i)^{([^}]+)?}MPD$', mpd_doc
.tag
, 'namespace', default
=None)
2604 return self
._xpath
_ns
(path
, namespace
)
2606 def is_drm_protected(element
):
2607 return element
.find(_add_ns('ContentProtection')) is not None
2609 def extract_multisegment_info(element
, ms_parent_info
):
2610 ms_info
= ms_parent_info
.copy()
2612 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2613 # common attributes and elements. We will only extract relevant
2615 def extract_common(source
):
2616 segment_timeline
= source
.find(_add_ns('SegmentTimeline'))
2617 if segment_timeline
is not None:
2618 s_e
= segment_timeline
.findall(_add_ns('S'))
2620 ms_info
['total_number'] = 0
2623 r
= int(s
.get('r', 0))
2624 ms_info
['total_number'] += 1 + r
2625 ms_info
['s'].append({
2626 't': int(s
.get('t', 0)),
2627 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2628 'd': int(s
.attrib
['d']),
2631 start_number
= source
.get('startNumber')
2633 ms_info
['start_number'] = int(start_number
)
2634 timescale
= source
.get('timescale')
2636 ms_info
['timescale'] = int(timescale
)
2637 segment_duration
= source
.get('duration')
2638 if segment_duration
:
2639 ms_info
['segment_duration'] = float(segment_duration
)
2641 def extract_Initialization(source
):
2642 initialization
= source
.find(_add_ns('Initialization'))
2643 if initialization
is not None:
2644 ms_info
['initialization_url'] = initialization
.attrib
['sourceURL']
2646 segment_list
= element
.find(_add_ns('SegmentList'))
2647 if segment_list
is not None:
2648 extract_common(segment_list
)
2649 extract_Initialization(segment_list
)
2650 segment_urls_e
= segment_list
.findall(_add_ns('SegmentURL'))
2652 ms_info
['segment_urls'] = [segment
.attrib
['media'] for segment
in segment_urls_e
]
2654 segment_template
= element
.find(_add_ns('SegmentTemplate'))
2655 if segment_template
is not None:
2656 extract_common(segment_template
)
2657 media
= segment_template
.get('media')
2659 ms_info
['media'] = media
2660 initialization
= segment_template
.get('initialization')
2662 ms_info
['initialization'] = initialization
2664 extract_Initialization(segment_template
)
2667 mpd_duration
= parse_duration(mpd_doc
.get('mediaPresentationDuration'))
2668 formats
, subtitles
= [], {}
2669 stream_numbers
= collections
.defaultdict(int)
2670 for period
in mpd_doc
.findall(_add_ns('Period')):
2671 period_duration
= parse_duration(period
.get('duration')) or mpd_duration
2672 period_ms_info
= extract_multisegment_info(period
, {
2676 for adaptation_set
in period
.findall(_add_ns('AdaptationSet')):
2677 adaption_set_ms_info
= extract_multisegment_info(adaptation_set
, period_ms_info
)
2678 for representation
in adaptation_set
.findall(_add_ns('Representation')):
2679 representation_attrib
= adaptation_set
.attrib
.copy()
2680 representation_attrib
.update(representation
.attrib
)
2681 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2682 mime_type
= representation_attrib
['mimeType']
2683 content_type
= representation_attrib
.get('contentType', mime_type
.split('/')[0])
2685 codecs
= representation_attrib
.get('codecs', '')
2686 if content_type
not in ('video', 'audio', 'text'):
2687 if mime_type
== 'image/jpeg':
2688 content_type
= mime_type
2689 elif codecs
.split('.')[0] == 'stpp':
2690 content_type
= 'text'
2691 elif mimetype2ext(mime_type
) in ('tt', 'dfxp', 'ttml', 'xml', 'json'):
2692 content_type
= 'text'
2694 self
.report_warning('Unknown MIME type %s in DASH manifest' % mime_type
)
2698 for element
in (representation
, adaptation_set
, period
, mpd_doc
):
2699 base_url_e
= element
.find(_add_ns('BaseURL'))
2700 if base_url_e
is not None:
2701 base_url
= base_url_e
.text
+ base_url
2702 if re
.match(r
'^https?://', base_url
):
2704 if mpd_base_url
and base_url
.startswith('/'):
2705 base_url
= compat_urlparse
.urljoin(mpd_base_url
, base_url
)
2706 elif mpd_base_url
and not re
.match(r
'^https?://', base_url
):
2707 if not mpd_base_url
.endswith('/'):
2709 base_url
= mpd_base_url
+ base_url
2710 representation_id
= representation_attrib
.get('id')
2711 lang
= representation_attrib
.get('lang')
2712 url_el
= representation
.find(_add_ns('BaseURL'))
2713 filesize
= int_or_none(url_el
.attrib
.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el
is not None else None)
2714 bandwidth
= int_or_none(representation_attrib
.get('bandwidth'))
2715 if representation_id
is not None:
2716 format_id
= representation_id
2718 format_id
= content_type
2720 format_id
= mpd_id
+ '-' + format_id
2721 if content_type
in ('video', 'audio'):
2723 'format_id': format_id
,
2724 'manifest_url': mpd_url
,
2725 'ext': mimetype2ext(mime_type
),
2726 'width': int_or_none(representation_attrib
.get('width')),
2727 'height': int_or_none(representation_attrib
.get('height')),
2728 'tbr': float_or_none(bandwidth
, 1000),
2729 'asr': int_or_none(representation_attrib
.get('audioSamplingRate')),
2730 'fps': int_or_none(representation_attrib
.get('frameRate')),
2731 'language': lang
if lang
not in ('mul', 'und', 'zxx', 'mis') else None,
2732 'format_note': 'DASH %s' % content_type
,
2733 'filesize': filesize
,
2734 'container': mimetype2ext(mime_type
) + '_dash',
2736 f
.update(parse_codecs(codecs
))
2737 elif content_type
== 'text':
2739 'ext': mimetype2ext(mime_type
),
2740 'manifest_url': mpd_url
,
2741 'filesize': filesize
,
2743 elif content_type
== 'image/jpeg':
2744 # See test case in VikiIE
2745 # https://www.viki.com/videos/1175236v-choosing-spouse-by-lottery-episode-1
2747 'format_id': format_id
,
2749 'manifest_url': mpd_url
,
2750 'format_note': 'DASH storyboards (jpeg)',
2754 if is_drm_protected(adaptation_set
) or is_drm_protected(representation
):
2756 representation_ms_info
= extract_multisegment_info(representation
, adaption_set_ms_info
)
2758 def prepare_template(template_name
, identifiers
):
2759 tmpl
= representation_ms_info
[template_name
]
2760 # First of, % characters outside $...$ templates
2761 # must be escaped by doubling for proper processing
2762 # by % operator string formatting used further (see
2763 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2769 in_template
= not in_template
2770 elif c
== '%' and not in_template
:
2772 # Next, $...$ templates are translated to their
2773 # %(...) counterparts to be used with % operator
2774 if representation_id
is not None:
2775 t
= t
.replace('$RepresentationID$', representation_id
)
2776 t
= re
.sub(r
'\$(%s)\$' % '|'.join(identifiers
), r
'%(\1)d', t
)
2777 t
= re
.sub(r
'\$(%s)%%([^$]+)\$' % '|'.join(identifiers
), r
'%(\1)\2', t
)
2778 t
.replace('$$', '$')
2781 # @initialization is a regular template like @media one
2782 # so it should be handled just the same way (see
2783 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2784 if 'initialization' in representation_ms_info
:
2785 initialization_template
= prepare_template(
2787 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2788 # $Time$ shall not be included for @initialization thus
2789 # only $Bandwidth$ remains
2791 representation_ms_info
['initialization_url'] = initialization_template
% {
2792 'Bandwidth': bandwidth
,
2795 def location_key(location
):
2796 return 'url' if re
.match(r
'^https?://', location
) else 'path'
2798 if 'segment_urls' not in representation_ms_info
and 'media' in representation_ms_info
:
2800 media_template
= prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2801 media_location_key
= location_key(media_template
)
2803 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2804 # can't be used at the same time
2805 if '%(Number' in media_template
and 's' not in representation_ms_info
:
2806 segment_duration
= None
2807 if 'total_number' not in representation_ms_info
and 'segment_duration' in representation_ms_info
:
2808 segment_duration
= float_or_none(representation_ms_info
['segment_duration'], representation_ms_info
['timescale'])
2809 representation_ms_info
['total_number'] = int(math
.ceil(float(period_duration
) / segment_duration
))
2810 representation_ms_info
['fragments'] = [{
2811 media_location_key
: media_template
% {
2812 'Number': segment_number
,
2813 'Bandwidth': bandwidth
,
2815 'duration': segment_duration
,
2816 } for segment_number
in range(
2817 representation_ms_info
['start_number'],
2818 representation_ms_info
['total_number'] + representation_ms_info
['start_number'])]
2820 # $Number*$ or $Time$ in media template with S list available
2821 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2822 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2823 representation_ms_info
['fragments'] = []
2826 segment_number
= representation_ms_info
['start_number']
2828 def add_segment_url():
2829 segment_url
= media_template
% {
2830 'Time': segment_time
,
2831 'Bandwidth': bandwidth
,
2832 'Number': segment_number
,
2834 representation_ms_info
['fragments'].append({
2835 media_location_key
: segment_url
,
2836 'duration': float_or_none(segment_d
, representation_ms_info
['timescale']),
2839 for num
, s
in enumerate(representation_ms_info
['s']):
2840 segment_time
= s
.get('t') or segment_time
2844 for r
in range(s
.get('r', 0)):
2845 segment_time
+= segment_d
2848 segment_time
+= segment_d
2849 elif 'segment_urls' in representation_ms_info
and 's' in representation_ms_info
:
2851 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2852 # or any YouTube dashsegments video
2855 timescale
= representation_ms_info
['timescale']
2856 for s
in representation_ms_info
['s']:
2857 duration
= float_or_none(s
['d'], timescale
)
2858 for r
in range(s
.get('r', 0) + 1):
2859 segment_uri
= representation_ms_info
['segment_urls'][segment_index
]
2861 location_key(segment_uri
): segment_uri
,
2862 'duration': duration
,
2865 representation_ms_info
['fragments'] = fragments
2866 elif 'segment_urls' in representation_ms_info
:
2867 # Segment URLs with no SegmentTimeline
2868 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2869 # https://github.com/ytdl-org/youtube-dl/pull/14844
2871 segment_duration
= float_or_none(
2872 representation_ms_info
['segment_duration'],
2873 representation_ms_info
['timescale']) if 'segment_duration' in representation_ms_info
else None
2874 for segment_url
in representation_ms_info
['segment_urls']:
2876 location_key(segment_url
): segment_url
,
2878 if segment_duration
:
2879 fragment
['duration'] = segment_duration
2880 fragments
.append(fragment
)
2881 representation_ms_info
['fragments'] = fragments
2882 # If there is a fragments key available then we correctly recognized fragmented media.
2883 # Otherwise we will assume unfragmented media with direct access. Technically, such
2884 # assumption is not necessarily correct since we may simply have no support for
2885 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2886 if 'fragments' in representation_ms_info
:
2888 # NB: mpd_url may be empty when MPD manifest is parsed from a string
2889 'url': mpd_url
or base_url
,
2890 'fragment_base_url': base_url
,
2892 'protocol': 'http_dash_segments' if mime_type
!= 'image/jpeg' else 'mhtml',
2894 if 'initialization_url' in representation_ms_info
:
2895 initialization_url
= representation_ms_info
['initialization_url']
2896 if not f
.get('url'):
2897 f
['url'] = initialization_url
2898 f
['fragments'].append({location_key(initialization_url): initialization_url}
)
2899 f
['fragments'].extend(representation_ms_info
['fragments'])
2901 # Assuming direct URL to unfragmented media.
2903 if content_type
in ('video', 'audio', 'image/jpeg'):
2904 f
['manifest_stream_number'] = stream_numbers
[f
['url']]
2905 stream_numbers
[f
['url']] += 1
2907 elif content_type
== 'text':
2908 subtitles
.setdefault(lang
or 'und', []).append(f
)
2910 return formats
, subtitles
2912 def _extract_ism_formats(self
, *args
, **kwargs
):
2913 fmts
, subs
= self
._extract
_ism
_formats
_and
_subtitles
(*args
, **kwargs
)
2915 self
._report
_ignoring
_subs
('ISM')
2918 def _extract_ism_formats_and_subtitles(self
, ism_url
, video_id
, ism_id
=None, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
):
2919 res
= self
._download
_xml
_handle
(
2921 note
='Downloading ISM manifest' if note
is None else note
,
2922 errnote
='Failed to download ISM manifest' if errnote
is None else errnote
,
2923 fatal
=fatal
, data
=data
, headers
=headers
, query
=query
)
2930 return self
._parse
_ism
_formats
_and
_subtitles
(ism_doc
, urlh
.geturl(), ism_id
)
2932 def _parse_ism_formats_and_subtitles(self
, ism_doc
, ism_url
, ism_id
=None):
2934 Parse formats from ISM manifest.
2936 1. [MS-SSTR]: Smooth Streaming Protocol,
2937 https://msdn.microsoft.com/en-us/library/ff469518.aspx
2939 if ism_doc
.get('IsLive') == 'TRUE':
2942 duration
= int(ism_doc
.attrib
['Duration'])
2943 timescale
= int_or_none(ism_doc
.get('TimeScale')) or 10000000
2947 for stream
in ism_doc
.findall('StreamIndex'):
2948 stream_type
= stream
.get('Type')
2949 if stream_type
not in ('video', 'audio', 'text'):
2951 url_pattern
= stream
.attrib
['Url']
2952 stream_timescale
= int_or_none(stream
.get('TimeScale')) or timescale
2953 stream_name
= stream
.get('Name')
2954 stream_language
= stream
.get('Language', 'und')
2955 for track
in stream
.findall('QualityLevel'):
2956 fourcc
= track
.get('FourCC') or ('AACL' if track
.get('AudioTag') == '255' else None)
2957 # TODO: add support for WVC1 and WMAP
2958 if fourcc
not in ('H264', 'AVC1', 'AACL', 'TTML'):
2959 self
.report_warning('%s is not a supported codec' % fourcc
)
2961 tbr
= int(track
.attrib
['Bitrate']) // 1000
2962 # [1] does not mention Width and Height attributes. However,
2963 # they're often present while MaxWidth and MaxHeight are
2964 # missing, so should be used as fallbacks
2965 width
= int_or_none(track
.get('MaxWidth') or track
.get('Width'))
2966 height
= int_or_none(track
.get('MaxHeight') or track
.get('Height'))
2967 sampling_rate
= int_or_none(track
.get('SamplingRate'))
2969 track_url_pattern
= re
.sub(r
'{[Bb]itrate}', track
.attrib
['Bitrate'], url_pattern
)
2970 track_url_pattern
= compat_urlparse
.urljoin(ism_url
, track_url_pattern
)
2976 stream_fragments
= stream
.findall('c')
2977 for stream_fragment_index
, stream_fragment
in enumerate(stream_fragments
):
2978 fragment_ctx
['time'] = int_or_none(stream_fragment
.get('t')) or fragment_ctx
['time']
2979 fragment_repeat
= int_or_none(stream_fragment
.get('r')) or 1
2980 fragment_ctx
['duration'] = int_or_none(stream_fragment
.get('d'))
2981 if not fragment_ctx
['duration']:
2983 next_fragment_time
= int(stream_fragment
[stream_fragment_index
+ 1].attrib
['t'])
2985 next_fragment_time
= duration
2986 fragment_ctx
['duration'] = (next_fragment_time
- fragment_ctx
['time']) / fragment_repeat
2987 for _
in range(fragment_repeat
):
2989 'url': re
.sub(r
'{start[ _]time}', compat_str(fragment_ctx
['time']), track_url_pattern
),
2990 'duration': fragment_ctx
['duration'] / stream_timescale
,
2992 fragment_ctx
['time'] += fragment_ctx
['duration']
2994 if stream_type
== 'text':
2995 subtitles
.setdefault(stream_language
, []).append({
2999 'manifest_url': ism_url
,
3000 'fragments': fragments
,
3001 '_download_params': {
3002 'stream_type': stream_type
,
3003 'duration': duration
,
3004 'timescale': stream_timescale
,
3006 'language': stream_language
,
3007 'codec_private_data': track
.get('CodecPrivateData'),
3010 elif stream_type
in ('video', 'audio'):
3012 'format_id': join_nonempty(ism_id
, stream_name
, tbr
),
3014 'manifest_url': ism_url
,
3015 'ext': 'ismv' if stream_type
== 'video' else 'isma',
3019 'asr': sampling_rate
,
3020 'vcodec': 'none' if stream_type
== 'audio' else fourcc
,
3021 'acodec': 'none' if stream_type
== 'video' else fourcc
,
3023 'fragments': fragments
,
3024 'has_drm': ism_doc
.find('Protection') is not None,
3025 '_download_params': {
3026 'stream_type': stream_type
,
3027 'duration': duration
,
3028 'timescale': stream_timescale
,
3029 'width': width
or 0,
3030 'height': height
or 0,
3032 'language': stream_language
,
3033 'codec_private_data': track
.get('CodecPrivateData'),
3034 'sampling_rate': sampling_rate
,
3035 'channels': int_or_none(track
.get('Channels', 2)),
3036 'bits_per_sample': int_or_none(track
.get('BitsPerSample', 16)),
3037 'nal_unit_length_field': int_or_none(track
.get('NALUnitLengthField', 4)),
3040 return formats
, subtitles
3042 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):
3043 def absolute_url(item_url
):
3044 return urljoin(base_url
, item_url
)
3046 def parse_content_type(content_type
):
3047 if not content_type
:
3049 ctr
= re
.search(r
'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type
)
3051 mimetype
, codecs
= ctr
.groups()
3052 f
= parse_codecs(codecs
)
3053 f
['ext'] = mimetype2ext(mimetype
)
3057 def _media_formats(src
, cur_media_type
, type_info
={}):
3058 full_url
= absolute_url(src
)
3059 ext
= type_info
.get('ext') or determine_ext(full_url
)
3061 is_plain_url
= False
3062 formats
= self
._extract
_m
3u8_formats
(
3063 full_url
, video_id
, ext
='mp4',
3064 entry_protocol
=m3u8_entry_protocol
, m3u8_id
=m3u8_id
,
3065 preference
=preference
, quality
=quality
, fatal
=False)
3067 is_plain_url
= False
3068 formats
= self
._extract
_mpd
_formats
(
3069 full_url
, video_id
, mpd_id
=mpd_id
, fatal
=False)
3074 'vcodec': 'none' if cur_media_type
== 'audio' else None,
3076 return is_plain_url
, formats
3079 # amp-video and amp-audio are very similar to their HTML5 counterparts
3080 # so we wll include them right here (see
3081 # https://www.ampproject.org/docs/reference/components/amp-video)
3082 # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/
3083 _MEDIA_TAG_NAME_RE
= r
'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)'
3084 media_tags
= [(media_tag
, media_tag_name
, media_type
, '')
3085 for media_tag
, media_tag_name
, media_type
3086 in re
.findall(r
'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE
, webpage
)]
3087 media_tags
.extend(re
.findall(
3088 # We only allow video|audio followed by a whitespace or '>'.
3089 # Allowing more characters may end up in significant slow down (see
3090 # https://github.com/ytdl-org/youtube-dl/issues/11979, example URL:
3091 # http://www.porntrex.com/maps/videositemap.xml).
3092 r
'(?s)(<(?P<tag>%s)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>' % _MEDIA_TAG_NAME_RE
, webpage
))
3093 for media_tag
, _
, media_type
, media_content
in media_tags
:
3098 media_attributes
= extract_attributes(media_tag
)
3099 src
= strip_or_none(media_attributes
.get('src'))
3101 _
, formats
= _media_formats(src
, media_type
)
3102 media_info
['formats'].extend(formats
)
3103 media_info
['thumbnail'] = absolute_url(media_attributes
.get('poster'))
3105 for source_tag
in re
.findall(r
'<source[^>]+>', media_content
):
3106 s_attr
= extract_attributes(source_tag
)
3107 # data-video-src and data-src are non standard but seen
3108 # several times in the wild
3109 src
= strip_or_none(dict_get(s_attr
, ('src', 'data-video-src', 'data-src')))
3112 f
= parse_content_type(s_attr
.get('type'))
3113 is_plain_url
, formats
= _media_formats(src
, media_type
, f
)
3115 # width, height, res, label and title attributes are
3116 # all not standard but seen several times in the wild
3119 for lbl
in ('label', 'title')
3120 if str_or_none(s_attr
.get(lbl
))
3122 width
= int_or_none(s_attr
.get('width'))
3123 height
= (int_or_none(s_attr
.get('height'))
3124 or int_or_none(s_attr
.get('res')))
3125 if not width
or not height
:
3127 resolution
= parse_resolution(lbl
)
3130 width
= width
or resolution
.get('width')
3131 height
= height
or resolution
.get('height')
3133 tbr
= parse_bitrate(lbl
)
3142 'format_id': s_attr
.get('label') or s_attr
.get('title'),
3144 f
.update(formats
[0])
3145 media_info
['formats'].append(f
)
3147 media_info
['formats'].extend(formats
)
3148 for track_tag
in re
.findall(r
'<track[^>]+>', media_content
):
3149 track_attributes
= extract_attributes(track_tag
)
3150 kind
= track_attributes
.get('kind')
3151 if not kind
or kind
in ('subtitles', 'captions'):
3152 src
= strip_or_none(track_attributes
.get('src'))
3155 lang
= track_attributes
.get('srclang') or track_attributes
.get('lang') or track_attributes
.get('label')
3156 media_info
['subtitles'].setdefault(lang
, []).append({
3157 'url': absolute_url(src
),
3159 for f
in media_info
['formats']:
3160 f
.setdefault('http_headers', {})['Referer'] = base_url
3161 if media_info
['formats'] or media_info
['subtitles']:
3162 entries
.append(media_info
)
3165 def _extract_akamai_formats(self
, *args
, **kwargs
):
3166 fmts
, subs
= self
._extract
_akamai
_formats
_and
_subtitles
(*args
, **kwargs
)
3168 self
._report
_ignoring
_subs
('akamai')
3171 def _extract_akamai_formats_and_subtitles(self
, manifest_url
, video_id
, hosts
={}):
3172 signed
= 'hdnea=' in manifest_url
3174 # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
3175 manifest_url
= re
.sub(
3176 r
'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
3177 '', manifest_url
).strip('?')
3182 hdcore_sign
= 'hdcore=3.7.0'
3183 f4m_url
= re
.sub(r
'(https?://[^/]+)/i/', r
'\1/z/', manifest_url
).replace('/master.m3u8', '/manifest.f4m')
3184 hds_host
= hosts
.get('hds')
3186 f4m_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hds_host
, f4m_url
)
3187 if 'hdcore=' not in f4m_url
:
3188 f4m_url
+= ('&' if '?' in f4m_url
else '?') + hdcore_sign
3189 f4m_formats
= self
._extract
_f
4m
_formats
(
3190 f4m_url
, video_id
, f4m_id
='hds', fatal
=False)
3191 for entry
in f4m_formats
:
3192 entry
.update({'extra_param_to_segment_url': hdcore_sign}
)
3193 formats
.extend(f4m_formats
)
3195 m3u8_url
= re
.sub(r
'(https?://[^/]+)/z/', r
'\1/i/', manifest_url
).replace('/manifest.f4m', '/master.m3u8')
3196 hls_host
= hosts
.get('hls')
3198 m3u8_url
= re
.sub(r
'(https?://)[^/]+', r
'\1' + hls_host
, m3u8_url
)
3199 m3u8_formats
, m3u8_subtitles
= self
._extract
_m
3u8_formats
_and
_subtitles
(
3200 m3u8_url
, video_id
, 'mp4', 'm3u8_native',
3201 m3u8_id
='hls', fatal
=False)
3202 formats
.extend(m3u8_formats
)
3203 subtitles
= self
._merge
_subtitles
(subtitles
, m3u8_subtitles
)
3205 http_host
= hosts
.get('http')
3206 if http_host
and m3u8_formats
and not signed
:
3207 REPL_REGEX
= r
'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
3208 qualities
= re
.match(REPL_REGEX
, m3u8_url
).group(2).split(',')
3209 qualities_length
= len(qualities
)
3210 if len(m3u8_formats
) in (qualities_length
, qualities_length
+ 1):
3212 for f
in m3u8_formats
:
3213 if f
['vcodec'] != 'none':
3214 for protocol
in ('http', 'https'):
3216 del http_f
['manifest_url']
3218 REPL_REGEX
, protocol
+ r
'://%s/\g<1>%s\3' % (http_host
, qualities
[i
]), f
['url'])
3220 'format_id': http_f
['format_id'].replace('hls-', protocol
+ '-'),
3222 'protocol': protocol
,
3224 formats
.append(http_f
)
3227 return formats
, subtitles
3229 def _extract_wowza_formats(self
, url
, video_id
, m3u8_entry_protocol
='m3u8_native', skip_protocols
=[]):
3230 query
= compat_urlparse
.urlparse(url
).query
3231 url
= re
.sub(r
'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url
)
3233 r
'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url
)
3234 url_base
= mobj
.group('url')
3235 http_base_url
= '%s%s:%s' % ('http', mobj
.group('s') or '', url_base
)
3238 def manifest_url(manifest
):
3239 m_url
= '%s/%s' % (http_base_url
, manifest
)
3241 m_url
+= '?%s' % query
3244 if 'm3u8' not in skip_protocols
:
3245 formats
.extend(self
._extract
_m
3u8_formats
(
3246 manifest_url('playlist.m3u8'), video_id
, 'mp4',
3247 m3u8_entry_protocol
, m3u8_id
='hls', fatal
=False))
3248 if 'f4m' not in skip_protocols
:
3249 formats
.extend(self
._extract
_f
4m
_formats
(
3250 manifest_url('manifest.f4m'),
3251 video_id
, f4m_id
='hds', fatal
=False))
3252 if 'dash' not in skip_protocols
:
3253 formats
.extend(self
._extract
_mpd
_formats
(
3254 manifest_url('manifest.mpd'),
3255 video_id
, mpd_id
='dash', fatal
=False))
3256 if re
.search(r
'(?:/smil:|\.smil)', url_base
):
3257 if 'smil' not in skip_protocols
:
3258 rtmp_formats
= self
._extract
_smil
_formats
(
3259 manifest_url('jwplayer.smil'),
3260 video_id
, fatal
=False)
3261 for rtmp_format
in rtmp_formats
:
3262 rtsp_format
= rtmp_format
.copy()
3263 rtsp_format
['url'] = '%s/%s' % (rtmp_format
['url'], rtmp_format
['play_path'])
3264 del rtsp_format
['play_path']
3265 del rtsp_format
['ext']
3266 rtsp_format
.update({
3267 'url': rtsp_format
['url'].replace('rtmp://', 'rtsp://'),
3268 'format_id': rtmp_format
['format_id'].replace('rtmp', 'rtsp'),
3271 formats
.extend([rtmp_format
, rtsp_format
])
3273 for protocol
in ('rtmp', 'rtsp'):
3274 if protocol
not in skip_protocols
:
3276 'url': '%s:%s' % (protocol
, url_base
),
3277 'format_id': protocol
,
3278 'protocol': protocol
,
3282 def _find_jwplayer_data(self
, webpage
, video_id
=None, transform_source
=js_to_json
):
3284 r
'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P
=quote
)\
)(?
!</script
>).*?\
.setup\s
*\
((?P
<options
>[^
)]+)\
)',
3288 jwplayer_data = self._parse_json(mobj.group('options
'),
3290 transform_source=transform_source)
3291 except ExtractorError:
3294 if isinstance(jwplayer_data, dict):
3295 return jwplayer_data
3297 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
3298 jwplayer_data = self._find_jwplayer_data(
3299 webpage, video_id, transform_source=js_to_json)
3300 return self._parse_jwplayer_data(
3301 jwplayer_data, video_id, *args, **kwargs)
3303 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
3304 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3305 # JWPlayer backward compatibility: flattened playlists
3306 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
3307 if 'playlist
' not in jwplayer_data:
3308 jwplayer_data = {'playlist': [jwplayer_data]}
3312 # JWPlayer backward compatibility: single playlist item
3313 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
3314 if not isinstance(jwplayer_data['playlist
'], list):
3315 jwplayer_data['playlist
'] = [jwplayer_data['playlist
']]
3317 for video_data in jwplayer_data['playlist
']:
3318 # JWPlayer backward compatibility: flattened sources
3319 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
3320 if 'sources
' not in video_data:
3321 video_data['sources
'] = [video_data]
3323 this_video_id = video_id or video_data['mediaid
']
3325 formats = self._parse_jwplayer_formats(
3326 video_data['sources
'], video_id=this_video_id, m3u8_id=m3u8_id,
3327 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
3330 tracks = video_data.get('tracks
')
3331 if tracks and isinstance(tracks, list):
3332 for track in tracks:
3333 if not isinstance(track, dict):
3335 track_kind = track.get('kind
')
3336 if not track_kind or not isinstance(track_kind, compat_str):
3338 if track_kind.lower() not in ('captions
', 'subtitles
'):
3340 track_url = urljoin(base_url, track.get('file'))
3343 subtitles.setdefault(track.get('label
') or 'en
', []).append({
3344 'url
': self._proto_relative_url(track_url)
3348 'id': this_video_id,
3349 'title
': unescapeHTML(video_data['title
'] if require_title else video_data.get('title
')),
3350 'description
': clean_html(video_data.get('description
')),
3351 'thumbnail
': urljoin(base_url, self._proto_relative_url(video_data.get('image
'))),
3352 'timestamp
': int_or_none(video_data.get('pubdate
')),
3353 'duration
': float_or_none(jwplayer_data.get('duration
') or video_data.get('duration
')),
3354 'subtitles
': subtitles,
3356 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
3357 if len(formats) == 1 and re.search(r'^
(?
:http|
//).*(?
:youtube\
.com|youtu\
.be
)/.+', formats[0]['url
']):
3359 '_type
': 'url_transparent
',
3360 'url
': formats[0]['url
'],
3363 self._sort_formats(formats)
3364 entry['formats
'] = formats
3365 entries.append(entry)
3366 if len(entries) == 1:
3369 return self.playlist_result(entries)
3371 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
3372 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3375 for source in jwplayer_sources_data:
3376 if not isinstance(source, dict):
3378 source_url = urljoin(
3379 base_url, self._proto_relative_url(source.get('file')))
3380 if not source_url or source_url in urls:
3382 urls.append(source_url)
3383 source_type = source.get('type') or ''
3384 ext = mimetype2ext(source_type) or determine_ext(source_url)
3385 if source_type == 'hls
' or ext == 'm3u8
':
3386 formats.extend(self._extract_m3u8_formats(
3387 source_url, video_id, 'mp4
', entry_protocol='m3u8_native
',
3388 m3u8_id=m3u8_id, fatal=False))
3389 elif source_type == 'dash
' or ext == 'mpd
':
3390 formats.extend(self._extract_mpd_formats(
3391 source_url, video_id, mpd_id=mpd_id, fatal=False))
3393 formats.extend(self._extract_smil_formats(
3394 source_url, video_id, fatal=False))
3395 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3396 elif source_type.startswith('audio
') or ext in (
3397 'oga
', 'aac
', 'mp3
', 'mpeg
', 'vorbis
'):
3404 height = int_or_none(source.get('height
'))
3406 # Often no height is provided but there is a label in
3407 # format like "1080p", "720p SD", or 1080.
3408 height = int_or_none(self._search_regex(
3409 r'^
(\d{3,4}
)[pP
]?
(?
:\b|$
)', compat_str(source.get('label
') or ''),
3410 'height
', default=None))
3413 'width
': int_or_none(source.get('width
')),
3415 'tbr
': int_or_none(source.get('bitrate
')),
3418 if source_url.startswith('rtmp
'):
3419 a_format['ext
'] = 'flv
'
3420 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3421 # of jwplayer.flash.swf
3422 rtmp_url_parts = re.split(
3423 r'((?
:mp4|mp3|flv
):)', source_url, 1)
3424 if len(rtmp_url_parts) == 3:
3425 rtmp_url, prefix, play_path = rtmp_url_parts
3428 'play_path
': prefix + play_path,
3431 a_format.update(rtmp_params)
3432 formats.append(a_format)
3435 def _live_title(self, name):
3436 """ Generate the title for a live video """
3437 now = datetime.datetime.now()
3438 now_str = now.strftime('%Y
-%m
-%d %H
:%M
')
3439 return name + ' ' + now_str
3441 def _int(self, v, name, fatal=False, **kwargs):
3442 res = int_or_none(v, **kwargs)
3443 if 'get_attr
' in kwargs:
3444 print(getattr(v, kwargs['get_attr
']))
3446 msg = 'Failed to extract
%s: Could
not parse value
%r' % (name, v)
3448 raise ExtractorError(msg)
3450 self.report_warning(msg)
3453 def _float(self, v, name, fatal=False, **kwargs):
3454 res = float_or_none(v, **kwargs)
3456 msg = 'Failed to extract
%s: Could
not parse value
%r' % (name, v)
3458 raise ExtractorError(msg)
3460 self.report_warning(msg)
3463 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
3464 path='/', secure=False, discard=False, rest={}, **kwargs):
3465 cookie = compat_cookiejar_Cookie(
3466 0, name, value, port, port is not None, domain, True,
3467 domain.startswith('.'), path, True, secure, expire_time,
3468 discard, None, None, rest)
3469 self._downloader.cookiejar.set_cookie(cookie)
3471 def _get_cookies(self, url):
3472 """ Return a compat_cookies_SimpleCookie with the cookies for the url """
3473 req = sanitized_Request(url)
3474 self._downloader.cookiejar.add_cookie_header(req)
3475 return compat_cookies_SimpleCookie(req.get_header('Cookie
'))
3477 def _apply_first_set_cookie_header(self, url_handle, cookie):
3479 Apply first Set-Cookie header instead of the last. Experimental.
3481 Some sites (e.g. [1-3]) may serve two cookies under the same name
3482 in Set-Cookie header and expect the first (old) one to be set rather
3483 than second (new). However, as of RFC6265 the newer one cookie
3484 should be set into cookie store what actually happens.
3485 We will workaround this issue by resetting the cookie to
3486 the first one manually.
3487 1. https://new.vk.com/
3488 2. https://github.com/ytdl-org/youtube-dl/issues/9841#issuecomment-227871201
3489 3. https://learning.oreilly.com/
3491 for header, cookies in url_handle.headers.items():
3492 if header.lower() != 'set-cookie
':
3494 if sys.version_info[0] >= 3:
3495 cookies = cookies.encode('iso
-8859-1')
3496 cookies = cookies.decode('utf
-8')
3497 cookie_value = re.search(
3498 r'%s=(.+?
);.*?
\b[Dd
]omain
=(.+?
)(?
:[,;]|$
)' % cookie, cookies)
3500 value, domain = cookie_value.groups()
3501 self._set_cookie(domain, cookie, value)
3504 def get_testcases(self, include_onlymatching=False):
3505 t = getattr(self, '_TEST
', None)
3507 assert not hasattr(self, '_TESTS
'), \
3508 '%s has _TEST
and _TESTS
' % type(self).__name__
3511 tests = getattr(self, '_TESTS
', [])
3513 if not include_onlymatching and t.get('only_matching
', False):
3515 t['name
'] = type(self).__name__[:-len('IE
')]
3518 def is_suitable(self, age_limit):
3519 """ Test whether the extractor is generally suitable for the given
3520 age limit (i.e. pornographic sites are not, all others usually are) """
3522 any_restricted = False
3523 for tc in self.get_testcases(include_onlymatching=False):
3524 if tc.get('playlist
', []):
3525 tc = tc['playlist
'][0]
3526 is_restricted = age_restricted(
3527 tc.get('info_dict
', {}).get('age_limit
'), age_limit)
3528 if not is_restricted:
3530 any_restricted = any_restricted or is_restricted
3531 return not any_restricted
3533 def extract_subtitles(self, *args, **kwargs):
3534 if (self.get_param('writesubtitles
', False)
3535 or self.get_param('listsubtitles
')):
3536 return self._get_subtitles(*args, **kwargs)
3539 def _get_subtitles(self, *args, **kwargs):
3540 raise NotImplementedError('This method must be implemented by subclasses
')
3542 def extract_comments(self, *args, **kwargs):
3543 if not self.get_param('getcomments
'):
3545 generator = self._get_comments(*args, **kwargs)
3551 comments.append(next(generator))
3552 except KeyboardInterrupt:
3554 self.to_screen('Interrupted by user
')
3555 except StopIteration:
3557 comment_count = len(comments)
3558 self.to_screen(f'Extracted {comment_count} comments
')
3560 'comments
': comments,
3561 'comment_count
': None if interrupted else comment_count
3565 def _get_comments(self, *args, **kwargs):
3566 raise NotImplementedError('This method must be implemented by subclasses
')
3569 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
3570 """ Merge subtitle items for one language. Items with duplicated URLs
3571 will be dropped. """
3572 list1_urls = set([item['url
'] for item in subtitle_list1])
3573 ret = list(subtitle_list1)
3574 ret.extend([item for item in subtitle_list2 if item['url
'] not in list1_urls])
3578 def _merge_subtitles(cls, *dicts, target=None):
3579 """ Merge subtitle dictionaries, language by language. """
3583 for lang, subs in d.items():
3584 target[lang] = cls._merge_subtitle_items(target.get(lang, []), subs)
3587 def extract_automatic_captions(self, *args, **kwargs):
3588 if (self.get_param('writeautomaticsub
', False)
3589 or self.get_param('listsubtitles
')):
3590 return self._get_automatic_captions(*args, **kwargs)
3593 def _get_automatic_captions(self, *args, **kwargs):
3594 raise NotImplementedError('This method must be implemented by subclasses
')
3596 def mark_watched(self, *args, **kwargs):
3597 if not self.get_param('mark_watched
', False):
3599 if (self._get_login_info()[0] is not None
3600 or self.get_param('cookiefile
')
3601 or self.get_param('cookiesfrombrowser
')):
3602 self._mark_watched(*args, **kwargs)
3604 def _mark_watched(self, *args, **kwargs):
3605 raise NotImplementedError('This method must be implemented by subclasses
')
3607 def geo_verification_headers(self):
3609 geo_verification_proxy = self.get_param('geo_verification_proxy
')
3610 if geo_verification_proxy:
3611 headers['Ytdl
-request
-proxy
'] = geo_verification_proxy
3614 def _generic_id(self, url):
3615 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
3617 def _generic_title(self, url):
3618 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
3621 def _availability(is_private=None, needs_premium=None, needs_subscription=None, needs_auth=None, is_unlisted=None):
3622 all_known = all(map(
3623 lambda x: x is not None,
3624 (is_private, needs_premium, needs_subscription, needs_auth, is_unlisted)))
3626 'private
' if is_private
3627 else 'premium_only
' if needs_premium
3628 else 'subscriber_only
' if needs_subscription
3629 else 'needs_auth
' if needs_auth
3630 else 'unlisted
' if is_unlisted
3631 else 'public
' if all_known
3634 def _configuration_arg(self, key, default=NO_DEFAULT, casesense=False):
3636 @returns A list of values for the extractor argument given by "key"
3637 or "default" if no such key is present
3638 @param default The default value to return when the key is not present (default: [])
3639 @param casesense When false, the values are converted to lower case
3642 self._downloader.params, ('extractor_args
', self.ie_key().lower(), key))
3644 return [] if default is NO_DEFAULT else default
3645 return list(val) if casesense else [x.lower() for x in val]
3648 class SearchInfoExtractor(InfoExtractor):
3650 Base class for paged search queries extractors.
3651 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
3652 Instances should define _SEARCH_KEY and optionally _MAX_RESULTS
3655 _MAX_RESULTS = float('inf
')
3658 def _make_valid_url(cls):
3659 return r'%s(?P
<prefix
>|
[1-9][0-9]*|all
):(?P
<query
>[\s\S
]+)' % cls._SEARCH_KEY
3662 def suitable(cls, url):
3663 return re.match(cls._make_valid_url(), url) is not None
3665 def _real_extract(self, query):
3666 mobj = re.match(self._make_valid_url(), query)
3668 raise ExtractorError('Invalid search query
"%s"' % query)
3670 prefix = mobj.group('prefix
')
3671 query = mobj.group('query
')
3673 return self._get_n_results(query, 1)
3674 elif prefix == 'all
':
3675 return self._get_n_results(query, self._MAX_RESULTS)
3679 raise ExtractorError('invalid download number
%s for query
"%s"' % (n, query))
3680 elif n > self._MAX_RESULTS:
3681 self.report_warning('%s returns
max %i results (you requested
%i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3682 n = self._MAX_RESULTS
3683 return self._get_n_results(query, n)
3685 def _get_n_results(self, query, n):
3686 """Get a specified number of results for a query.
3687 Either this function or _search_results must be overridden by subclasses """
3688 return self.playlist_result(
3689 itertools.islice(self._search_results(query), 0, None if n == float('inf
') else n),
3692 def _search_results(self, query):
3693 """Returns an iterator of search results"""
3694 raise NotImplementedError('This method must be implemented by subclasses
')
3697 def SEARCH_KEY(self):
3698 return self._SEARCH_KEY