2 from __future__
import unicode_literals
17 from ..compat
import (
18 compat_cookiejar_Cookie
,
19 compat_cookies_SimpleCookie
,
21 compat_etree_fromstring
,
28 compat_urllib_parse_unquote
,
29 compat_urllib_parse_urlencode
,
30 compat_urllib_request
,
32 compat_xml_parse_error
,
34 from ..downloader
import FileDownloader
35 from ..downloader
.f4m
import (
37 remove_encrypted_media
,
67 parse_m3u8_attributes
,
91 class InfoExtractor(object):
92 """Information Extractor class.
94 Information extractors are the classes that, given a URL, extract
95 information about the video (or videos) the URL refers to. This
96 information includes the real video URL, the video title, author and
97 others. The information is stored in a dictionary which is then
98 passed to the YoutubeDL. The YoutubeDL processes this
99 information possibly downloading the video to the file system, among
100 other possible outcomes.
102 The type field determines the type of the result.
103 By far the most common value (and the default if _type is missing) is
104 "video", which indicates a single video.
106 For a video, the dictionaries must include the following fields:
108 id: Video identifier.
109 title: Video title, unescaped.
111 Additionally, it must contain either a formats entry or a url one:
113 formats: A list of dictionaries for each format available, ordered
114 from worst to best quality.
117 * url The mandatory URL representing the media:
118 for plain file media - HTTP URL of this file,
120 for HLS - URL of the M3U8 media playlist,
121 for HDS - URL of the F4M manifest,
123 - HTTP URL to plain file media (in case of
125 - URL of the MPD manifest or base URL
126 representing the media if MPD manifest
127 is parsed from a string (in case of
129 for MSS - URL of the ISM manifest.
131 The URL of the manifest file in case of
133 for HLS - URL of the M3U8 master playlist,
134 for HDS - URL of the F4M manifest,
135 for DASH - URL of the MPD manifest,
136 for MSS - URL of the ISM manifest.
137 * ext Will be calculated from URL if missing
138 * format A human-readable description of the format
139 ("mp4 container with h264/opus").
140 Calculated from the format_id, width, height.
141 and format_note fields if missing.
142 * format_id A short description of the format
143 ("mp4_h264_opus" or "19").
144 Technically optional, but strongly recommended.
145 * format_note Additional info about the format
146 ("3D" or "DASH video")
147 * width Width of the video, if known
148 * height Height of the video, if known
149 * resolution Textual description of width and height
150 * tbr Average bitrate of audio and video in KBit/s
151 * abr Average audio bitrate in KBit/s
152 * acodec Name of the audio codec in use
153 * asr Audio sampling rate in Hertz
154 * vbr Average video bitrate in KBit/s
156 * vcodec Name of the video codec in use
157 * container Name of the container format
158 * filesize The number of bytes, if known in advance
159 * filesize_approx An estimate for the number of bytes
160 * player_url SWF Player URL (used for rtmpdump).
161 * protocol The protocol that will be used for the actual
162 download, lower-case.
163 "http", "https", "rtsp", "rtmp", "rtmp_ffmpeg", "rtmpe",
164 "m3u8", "m3u8_native" or "http_dash_segments".
166 Base URL for fragments. Each fragment's path
167 value (if present) will be relative to
169 * fragments A list of fragments of a fragmented media.
170 Each fragment entry must contain either an url
171 or a path. If an url is present it should be
172 considered by a client. Otherwise both path and
173 fragment_base_url must be present. Here is
174 the list of all potential fields:
175 * "url" - fragment's URL
176 * "path" - fragment's path relative to
178 * "duration" (optional, int or float)
179 * "filesize" (optional, int)
180 * preference Order number of this format. If this field is
181 present and not None, the formats get sorted
182 by this field, regardless of all other values.
183 -1 for default (order by other properties),
184 -2 or smaller for less than default.
185 < -1000 to hide the format (if there is
186 another one which is strictly better)
187 * language Language code, e.g. "de" or "en-US".
188 * language_preference Is this in the language mentioned in
190 10 if it's what the URL is about,
191 -1 for default (don't know),
192 -10 otherwise, other values reserved for now.
193 * quality Order number of the video quality of this
194 format, irrespective of the file format.
195 -1 for default (order by other properties),
196 -2 or smaller for less than default.
197 * source_preference Order number for this video source
198 (quality takes higher priority)
199 -1 for default (order by other properties),
200 -2 or smaller for less than default.
201 * http_headers A dictionary of additional HTTP headers
202 to add to the request.
203 * stretched_ratio If given and not 1, indicates that the
204 video's pixels are not square.
205 width : height ratio as float.
206 * no_resume The server does not support resuming the
207 (HTTP or RTMP) download. Boolean.
208 * has_drm The format has DRM and cannot be downloaded. Boolean
209 * downloader_options A dictionary of downloader options as
210 described in FileDownloader
211 RTMP formats can also have the additional fields: page_url,
212 app, play_path, tc_url, flash_version, rtmp_live, rtmp_conn,
213 rtmp_protocol, rtmp_real_time
215 url: Final video URL.
216 ext: Video filename extension.
217 format: The video format, defaults to ext (used for --get-format)
218 player_url: SWF Player URL (used for rtmpdump).
220 The following fields are optional:
222 alt_title: A secondary title of the video.
223 display_id An alternative identifier for the video, not necessarily
224 unique, but available before title. Typically, id is
225 something like "4234987", title "Dancing naked mole rats",
226 and display_id "dancing-naked-mole-rats"
227 thumbnails: A list of dictionaries, with the following entries:
228 * "id" (optional, string) - Thumbnail format ID
230 * "preference" (optional, int) - quality of the image
231 * "width" (optional, int)
232 * "height" (optional, int)
233 * "resolution" (optional, string "{width}x{height}",
235 * "filesize" (optional, int)
236 thumbnail: Full URL to a video thumbnail image.
237 description: Full video description.
238 uploader: Full name of the video uploader.
239 license: License name the video is licensed under.
240 creator: The creator of the video.
241 release_timestamp: UNIX timestamp of the moment the video was released.
242 release_date: The date (YYYYMMDD) when the video was released.
243 timestamp: UNIX timestamp of the moment the video was uploaded
244 upload_date: Video upload date (YYYYMMDD).
245 If not explicitly set, calculated from timestamp.
246 uploader_id: Nickname or id of the video uploader.
247 uploader_url: Full URL to a personal webpage of the video uploader.
248 channel: Full name of the channel the video is uploaded on.
249 Note that channel fields may or may not repeat uploader
250 fields. This depends on a particular extractor.
251 channel_id: Id of the channel.
252 channel_url: Full URL to a channel webpage.
253 location: Physical location where the video was filmed.
254 subtitles: The available subtitles as a dictionary in the format
255 {tag: subformats}. "tag" is usually a language code, and
256 "subformats" is a list sorted from lower to higher
257 preference, each element is a dictionary with the "ext"
259 * "data": The subtitles file contents
260 * "url": A URL pointing to the subtitles file
261 It can optionally also have:
262 * "name": Name or description of the subtitles
263 "ext" will be calculated from URL if missing
264 automatic_captions: Like 'subtitles'; contains automatically generated
265 captions instead of normal subtitles
266 duration: Length of the video in seconds, as an integer or float.
267 view_count: How many users have watched the video on the platform.
268 like_count: Number of positive ratings of the video
269 dislike_count: Number of negative ratings of the video
270 repost_count: Number of reposts of the video
271 average_rating: Average rating give by users, the scale used depends on the webpage
272 comment_count: Number of comments on the video
273 comments: A list of comments, each with one or more of the following
274 properties (all but one of text or html optional):
275 * "author" - human-readable name of the comment author
276 * "author_id" - user ID of the comment author
277 * "author_thumbnail" - The thumbnail of the comment author
279 * "html" - Comment as HTML
280 * "text" - Plain text of the comment
281 * "timestamp" - UNIX timestamp of comment
282 * "parent" - ID of the comment this one is replying to.
283 Set to "root" to indicate that this is a
284 comment to the original video.
285 * "like_count" - Number of positive ratings of the comment
286 * "dislike_count" - Number of negative ratings of the comment
287 * "is_favorited" - Whether the comment is marked as
288 favorite by the video uploader
289 * "author_is_uploader" - Whether the comment is made by
291 age_limit: Age restriction for the video, as an integer (years)
292 webpage_url: The URL to the video webpage, if given to yt-dlp it
293 should allow to get the same result again. (It will be set
294 by YoutubeDL if it's missing)
295 categories: A list of categories that the video falls in, for example
297 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
298 cast: A list of the video cast
299 is_live: True, False, or None (=unknown). Whether this video is a
300 live stream that goes on instead of a fixed-length video.
301 was_live: True, False, or None (=unknown). Whether this video was
302 originally a live stream.
303 live_status: 'is_live', 'is_upcoming', 'was_live', 'not_live' or None (=unknown)
304 If absent, automatically set from is_live, was_live
305 start_time: Time in seconds where the reproduction should start, as
306 specified in the URL.
307 end_time: Time in seconds where the reproduction should end, as
308 specified in the URL.
309 chapters: A list of dictionaries, with the following entries:
310 * "start_time" - The start time of the chapter in seconds
311 * "end_time" - The end time of the chapter in seconds
312 * "title" (optional, string)
313 playable_in_embed: Whether this video is allowed to play in embedded
314 players on other sites. Can be True (=always allowed),
315 False (=never allowed), None (=unknown), or a string
316 specifying the criteria for embedability (Eg: 'whitelist')
317 availability: Under what condition the video is available. One of
318 'private', 'premium_only', 'subscriber_only', 'needs_auth',
319 'unlisted' or 'public'. Use 'InfoExtractor._availability'
321 __post_extractor: A function to be called just before the metadata is
322 written to either disk, logger or console. The function
323 must return a dict which will be added to the info_dict.
324 This is usefull for additional information that is
325 time-consuming to extract. Note that the fields thus
326 extracted will not be available to output template and
327 match_filter. So, only "comments" and "comment_count" are
328 currently allowed to be extracted via this method.
330 The following fields should only be used when the video belongs to some logical
333 chapter: Name or title of the chapter the video belongs to.
334 chapter_number: Number of the chapter the video belongs to, as an integer.
335 chapter_id: Id of the chapter the video belongs to, as a unicode string.
337 The following fields should only be used when the video is an episode of some
338 series, programme or podcast:
340 series: Title of the series or programme the video episode belongs to.
341 season: Title of the season the video episode belongs to.
342 season_number: Number of the season the video episode belongs to, as an integer.
343 season_id: Id of the season the video episode belongs to, as a unicode string.
344 episode: Title of the video episode. Unlike mandatory video title field,
345 this field should denote the exact title of the video episode
346 without any kind of decoration.
347 episode_number: Number of the video episode within a season, as an integer.
348 episode_id: Id of the video episode, as a unicode string.
350 The following fields should only be used when the media is a track or a part of
353 track: Title of the track.
354 track_number: Number of the track within an album or a disc, as an integer.
355 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
357 artist: Artist(s) of the track.
358 genre: Genre(s) of the track.
359 album: Title of the album the track belongs to.
360 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
361 album_artist: List of all artists appeared on the album (e.g.
362 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
364 disc_number: Number of the disc or other physical medium the track belongs to,
366 release_year: Year (YYYY) when the album was released.
368 Unless mentioned otherwise, the fields should be Unicode strings.
370 Unless mentioned otherwise, None is equivalent to absence of information.
373 _type "playlist" indicates multiple videos.
374 There must be a key "entries", which is a list, an iterable, or a PagedList
375 object, each element of which is a valid dictionary by this specification.
377 Additionally, playlists can have "id", "title", and any other relevent
378 attributes with the same semantics as videos (see above).
381 _type "multi_video" indicates that there are multiple videos that
382 form a single show, for examples multiple acts of an opera or TV episode.
383 It must have an entries key like a playlist and contain all the keys
384 required for a video at the same time.
387 _type "url" indicates that the video must be extracted from another
388 location, possibly by a different extractor. Its only required key is:
389 "url" - the next URL to extract.
390 The key "ie_key" can be set to the class name (minus the trailing "IE",
391 e.g. "Youtube") if the extractor class is known in advance.
392 Additionally, the dictionary may have any properties of the resolved entity
393 known in advance, for example "title" if the title of the referred video is
397 _type "url_transparent" entities have the same specification as "url", but
398 indicate that the given additional information is more precise than the one
399 associated with the resolved URL.
400 This is useful when a site employs a video service that hosts the video and
401 its technical metadata, but that video service does not embed a useful
402 title, description etc.
405 Subclasses of this one should re-define the _real_initialize() and
406 _real_extract() methods and define a _VALID_URL regexp.
407 Probably, they should also be added to the list of extractors.
409 Subclasses may also override suitable() if necessary, but ensure the function
410 signature is preserved and that this function imports everything it needs
411 (except other extractors), so that lazy_extractors works correctly
413 _GEO_BYPASS attribute may be set to False in order to disable
414 geo restriction bypass mechanisms for a particular extractor.
415 Though it won't disable explicit geo restriction bypass based on
416 country code provided with geo_bypass_country.
418 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
419 countries for this extractor. One of these countries will be used by
420 geo restriction bypass mechanism right away in order to bypass
421 geo restriction, of course, if the mechanism is not disabled.
423 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
424 IP blocks in CIDR notation for this extractor. One of these IP blocks
425 will be used by geo restriction bypass mechanism similarly
428 The _WORKING attribute should be set to False for broken IEs
429 in order to warn the users and skip the tests.
434 _x_forwarded_for_ip
= None
436 _GEO_COUNTRIES
= None
437 _GEO_IP_BLOCKS
= None
441 'any': 'Use --cookies, --username and --password or --netrc to provide account credentials',
443 'Use --cookies-from-browser or --cookies for the authentication. '
444 'See https://github.com/ytdl-org/youtube-dl#how-do-i-pass-cookies-to-youtube-dl for how to manually pass cookies'),
445 'password': 'Use --username and --password or --netrc to provide account credentials',
448 def __init__(self
, downloader
=None):
449 """Constructor. Receives an optional downloader."""
451 self
._x
_forwarded
_for
_ip
= None
452 self
._printed
_messages
= set()
453 self
.set_downloader(downloader
)
456 def _match_valid_url(cls
, url
):
457 # This does not use has/getattr intentionally - we want to know whether
458 # we have cached the regexp for *this* class, whereas getattr would also
459 # match the superclass
460 if '_VALID_URL_RE' not in cls
.__dict
__:
461 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
462 return cls
._VALID
_URL
_RE
.match(url
)
465 def suitable(cls
, url
):
466 """Receives a URL and returns True if suitable for this IE."""
467 # This function must import everything it needs (except other extractors),
468 # so that lazy_extractors works correctly
469 return cls
._match
_valid
_url
(url
) is not None
472 def _match_id(cls
, url
):
473 return cls
._match
_valid
_url
(url
).group('id')
476 def get_temp_id(cls
, url
):
478 return cls
._match
_id
(url
)
479 except (IndexError, AttributeError):
484 """Getter method for _WORKING."""
487 def initialize(self
):
488 """Initializes an instance (authentication, etc)."""
489 self
._printed
_messages
= set()
490 self
._initialize
_geo
_bypass
({
491 'countries': self
._GEO
_COUNTRIES
,
492 'ip_blocks': self
._GEO
_IP
_BLOCKS
,
495 self
._real
_initialize
()
498 def _initialize_geo_bypass(self
, geo_bypass_context
):
500 Initialize geo restriction bypass mechanism.
502 This method is used to initialize geo bypass mechanism based on faking
503 X-Forwarded-For HTTP header. A random country from provided country list
504 is selected and a random IP belonging to this country is generated. This
505 IP will be passed as X-Forwarded-For HTTP header in all subsequent
508 This method will be used for initial geo bypass mechanism initialization
509 during the instance initialization with _GEO_COUNTRIES and
512 You may also manually call it from extractor's code if geo bypass
513 information is not available beforehand (e.g. obtained during
514 extraction) or due to some other reason. In this case you should pass
515 this information in geo bypass context passed as first argument. It may
516 contain following fields:
518 countries: List of geo unrestricted countries (similar
520 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
521 (similar to _GEO_IP_BLOCKS)
524 if not self
._x
_forwarded
_for
_ip
:
526 # Geo bypass mechanism is explicitly disabled by user
527 if not self
.get_param('geo_bypass', True):
530 if not geo_bypass_context
:
531 geo_bypass_context
= {}
533 # Backward compatibility: previously _initialize_geo_bypass
534 # expected a list of countries, some 3rd party code may still use
536 if isinstance(geo_bypass_context
, (list, tuple)):
537 geo_bypass_context
= {
538 'countries': geo_bypass_context
,
541 # The whole point of geo bypass mechanism is to fake IP
542 # as X-Forwarded-For HTTP header based on some IP block or
545 # Path 1: bypassing based on IP block in CIDR notation
547 # Explicit IP block specified by user, use it right away
548 # regardless of whether extractor is geo bypassable or not
549 ip_block
= self
.get_param('geo_bypass_ip_block', None)
551 # Otherwise use random IP block from geo bypass context but only
552 # if extractor is known as geo bypassable
554 ip_blocks
= geo_bypass_context
.get('ip_blocks')
555 if self
._GEO
_BYPASS
and ip_blocks
:
556 ip_block
= random
.choice(ip_blocks
)
559 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(ip_block
)
560 self
._downloader
.write_debug(
561 '[debug] Using fake IP %s as X-Forwarded-For' % self
._x
_forwarded
_for
_ip
)
564 # Path 2: bypassing based on country code
566 # Explicit country code specified by user, use it right away
567 # regardless of whether extractor is geo bypassable or not
568 country
= self
.get_param('geo_bypass_country', None)
570 # Otherwise use random country code from geo bypass context but
571 # only if extractor is known as geo bypassable
573 countries
= geo_bypass_context
.get('countries')
574 if self
._GEO
_BYPASS
and countries
:
575 country
= random
.choice(countries
)
578 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country
)
579 self
._downloader
.write_debug(
580 'Using fake IP %s (%s) as X-Forwarded-For' % (self
._x
_forwarded
_for
_ip
, country
.upper()))
582 def extract(self
, url
):
583 """Extracts URL information and returns it in list of dicts."""
588 self
.write_debug('Extracting URL: %s' % url
)
589 ie_result
= self
._real
_extract
(url
)
590 if ie_result
is None:
592 if self
._x
_forwarded
_for
_ip
:
593 ie_result
['__x_forwarded_for_ip'] = self
._x
_forwarded
_for
_ip
594 subtitles
= ie_result
.get('subtitles')
595 if (subtitles
and 'live_chat' in subtitles
596 and 'no-live-chat' in self
.get_param('compat_opts', [])):
597 del subtitles
['live_chat']
599 except GeoRestrictedError
as e
:
600 if self
.__maybe
_fake
_ip
_and
_retry
(e
.countries
):
603 except ExtractorError
as e
:
604 video_id
= e
.video_id
or self
.get_temp_id(url
)
605 raise ExtractorError(
606 e
.msg
, video_id
=video_id
, ie
=self
.IE_NAME
, tb
=e
.traceback
, expected
=e
.expected
, cause
=e
.cause
)
607 except compat_http_client
.IncompleteRead
as e
:
608 raise ExtractorError('A network error has occurred.', cause
=e
, expected
=True, video_id
=self
.get_temp_id(url
))
609 except (KeyError, StopIteration) as e
:
610 raise ExtractorError('An extractor error has occurred.', cause
=e
, video_id
=self
.get_temp_id(url
))
612 def __maybe_fake_ip_and_retry(self
, countries
):
613 if (not self
.get_param('geo_bypass_country', None)
615 and self
.get_param('geo_bypass', True)
616 and not self
._x
_forwarded
_for
_ip
618 country_code
= random
.choice(countries
)
619 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country_code
)
620 if self
._x
_forwarded
_for
_ip
:
622 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
623 % (self
._x
_forwarded
_for
_ip
, country_code
.upper()))
627 def set_downloader(self
, downloader
):
628 """Sets the downloader for this IE."""
629 self
._downloader
= downloader
631 def _real_initialize(self
):
632 """Real initialization process. Redefine in subclasses."""
635 def _real_extract(self
, url
):
636 """Real extraction process. Redefine in subclasses."""
641 """A string for getting the InfoExtractor with get_info_extractor"""
642 return cls
.__name
__[:-2]
646 return compat_str(type(self
).__name
__[:-2])
649 def __can_accept_status_code(err
, expected_status
):
650 assert isinstance(err
, compat_urllib_error
.HTTPError
)
651 if expected_status
is None:
653 elif callable(expected_status
):
654 return expected_status(err
.code
) is True
656 return err
.code
in variadic(expected_status
)
658 def _request_webpage(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
, expected_status
=None):
660 Return the response handle.
662 See _download_webpage docstring for arguments specification.
664 if not self
._downloader
._first
_webpage
_request
:
665 sleep_interval
= float_or_none(self
.get_param('sleep_interval_requests')) or 0
666 if sleep_interval
> 0:
667 self
.to_screen('Sleeping %s seconds ...' % sleep_interval
)
668 time
.sleep(sleep_interval
)
670 self
._downloader
._first
_webpage
_request
= False
673 self
.report_download_webpage(video_id
)
674 elif note
is not False:
676 self
.to_screen('%s' % (note
,))
678 self
.to_screen('%s: %s' % (video_id
, note
))
680 # Some sites check X-Forwarded-For HTTP header in order to figure out
681 # the origin of the client behind proxy. This allows bypassing geo
682 # restriction by faking this header's value to IP that belongs to some
683 # geo unrestricted country. We will do so once we encounter any
684 # geo restriction error.
685 if self
._x
_forwarded
_for
_ip
:
686 if 'X-Forwarded-For' not in headers
:
687 headers
['X-Forwarded-For'] = self
._x
_forwarded
_for
_ip
689 if isinstance(url_or_request
, compat_urllib_request
.Request
):
690 url_or_request
= update_Request(
691 url_or_request
, data
=data
, headers
=headers
, query
=query
)
694 url_or_request
= update_url_query(url_or_request
, query
)
695 if data
is not None or headers
:
696 url_or_request
= sanitized_Request(url_or_request
, data
, headers
)
698 return self
._downloader
.urlopen(url_or_request
)
699 except network_exceptions
as err
:
700 if isinstance(err
, compat_urllib_error
.HTTPError
):
701 if self
.__can
_accept
_status
_code
(err
, expected_status
):
702 # Retain reference to error to prevent file object from
703 # being closed before it can be read. Works around the
704 # effects of <https://bugs.python.org/issue15002>
705 # introduced in Python 3.4.1.
712 errnote
= 'Unable to download webpage'
714 errmsg
= '%s: %s' % (errnote
, error_to_compat_str(err
))
716 raise ExtractorError(errmsg
, sys
.exc_info()[2], cause
=err
)
718 self
.report_warning(errmsg
)
721 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):
723 Return a tuple (page content as string, URL handle).
725 See _download_webpage docstring for arguments specification.
727 # Strip hashes from the URL (#1038)
728 if isinstance(url_or_request
, (compat_str
, str)):
729 url_or_request
= url_or_request
.partition('#')[0]
731 urlh
= self
._request
_webpage
(url_or_request
, video_id
, note
, errnote
, fatal
, data
=data
, headers
=headers
, query
=query
, expected_status
=expected_status
)
735 content
= self
._webpage
_read
_content
(urlh
, url_or_request
, video_id
, note
, errnote
, fatal
, encoding
=encoding
)
736 return (content
, urlh
)
739 def _guess_encoding_from_content(content_type
, webpage_bytes
):
740 m
= re
.match(r
'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type
)
742 encoding
= m
.group(1)
744 m
= re
.search(br
'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
745 webpage_bytes[:1024])
747 encoding = m.group(1).decode('ascii')
748 elif webpage_bytes.startswith(b'\xff\xfe'):
755 def __check_blocked(self, content):
756 first_block = content[:512]
757 if ('<title>Access to this site is blocked</title>' in content
758 and 'Websense' in first_block):
759 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
760 blocked_iframe = self._html_search_regex(
761 r'<iframe src="([^
"]+)"', content,
762 'Websense information URL
', default=None)
764 msg += ' Visit
%s for more details
' % blocked_iframe
765 raise ExtractorError(msg, expected=True)
766 if '<title
>The URL you requested has been blocked
</title
>' in first_block:
768 'Access to this webpage has been blocked by Indian censorship
. '
769 'Use a VPN
or proxy
server (with --proxy
) to route around it
.')
770 block_msg = self._html_search_regex(
771 r'</h1
><p
>(.*?
)</p
>',
772 content, 'block message
', default=None)
774 msg += ' (Message
: "%s")' % block_msg.replace('\n', ' ')
775 raise ExtractorError(msg, expected=True)
776 if ('<title
>TTK
:: Доступ к ресурсу ограничен
</title
>' in content
777 and 'blocklist
.rkn
.gov
.ru
' in content):
778 raise ExtractorError(
779 'Access to this webpage has been blocked by decision of the Russian government
. '
780 'Visit http
://blocklist
.rkn
.gov
.ru
/ for a block reason
.',
783 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
784 content_type = urlh.headers.get('Content
-Type
', '')
785 webpage_bytes = urlh.read()
786 if prefix is not None:
787 webpage_bytes = prefix + webpage_bytes
789 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
790 if self.get_param('dump_intermediate_pages
', False):
791 self.to_screen('Dumping request to
' + urlh.geturl())
792 dump = base64.b64encode(webpage_bytes).decode('ascii
')
793 self._downloader.to_screen(dump)
794 if self.get_param('write_pages
', False):
795 basen = '%s_%s' % (video_id, urlh.geturl())
796 trim_length = self.get_param('trim_file_name
') or 240
797 if len(basen) > trim_length:
798 h = '___
' + hashlib.md5(basen.encode('utf
-8')).hexdigest()
799 basen = basen[:trim_length - len(h)] + h
800 raw_filename = basen + '.dump
'
801 filename = sanitize_filename(raw_filename, restricted=True)
802 self.to_screen('Saving request to
' + filename)
803 # Working around MAX_PATH limitation on Windows (see
804 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
805 if compat_os_name == 'nt
':
806 absfilepath = os.path.abspath(filename)
807 if len(absfilepath) > 259:
808 filename = '\\\\?
\\' + absfilepath
809 with open(filename, 'wb
') as outf:
810 outf.write(webpage_bytes)
813 content = webpage_bytes.decode(encoding, 'replace
')
815 content = webpage_bytes.decode('utf
-8', 'replace
')
817 self.__check_blocked(content)
821 def _download_webpage(
822 self, url_or_request, video_id, note=None, errnote=None,
823 fatal=True, tries=1, timeout=5, encoding=None, data=None,
824 headers={}, query={}, expected_status=None):
826 Return the data of the page as a string.
829 url_or_request -- plain text URL as a string or
830 a compat_urllib_request.Requestobject
831 video_id -- Video/playlist/item identifier (string)
834 note -- note printed before downloading (string)
835 errnote -- note printed in case of an error (string)
836 fatal -- flag denoting whether error should be considered fatal,
837 i.e. whether it should cause ExtractionError to be raised,
838 otherwise a warning will be reported and extraction continued
839 tries -- number of tries
840 timeout -- sleep interval between tries
841 encoding -- encoding for a page content decoding, guessed automatically
842 when not explicitly specified
843 data -- POST data (bytes)
844 headers -- HTTP headers (dict)
845 query -- URL query (dict)
846 expected_status -- allows to accept failed HTTP requests (non 2xx
847 status code) by explicitly specifying a set of accepted status
848 codes. Can be any of the following entities:
849 - an integer type specifying an exact failed status code to
851 - a list or a tuple of integer types specifying a list of
852 failed status codes to accept
853 - a callable accepting an actual failed status code and
854 returning True if it should be accepted
855 Note that this argument does not affect success status codes (2xx)
856 which are always accepted.
861 while success is False:
863 res = self._download_webpage_handle(
864 url_or_request, video_id, note, errnote, fatal,
865 encoding=encoding, data=data, headers=headers, query=query,
866 expected_status=expected_status)
868 except compat_http_client.IncompleteRead as e:
870 if try_count >= tries:
872 self._sleep(timeout, video_id)
879 def _download_xml_handle(
880 self, url_or_request, video_id, note='Downloading XML
',
881 errnote='Unable to download XML
', transform_source=None,
882 fatal=True, encoding=None, data=None, headers={}, query={},
883 expected_status=None):
885 Return a tuple (xml as an compat_etree_Element, URL handle).
887 See _download_webpage docstring for arguments specification.
889 res = self._download_webpage_handle(
890 url_or_request, video_id, note, errnote, fatal=fatal,
891 encoding=encoding, data=data, headers=headers, query=query,
892 expected_status=expected_status)
895 xml_string, urlh = res
896 return self._parse_xml(
897 xml_string, video_id, transform_source=transform_source,
901 self, url_or_request, video_id,
902 note='Downloading XML
', errnote='Unable to download XML
',
903 transform_source=None, fatal=True, encoding=None,
904 data=None, headers={}, query={}, expected_status=None):
906 Return the xml as an compat_etree_Element.
908 See _download_webpage docstring for arguments specification.
910 res = self._download_xml_handle(
911 url_or_request, video_id, note=note, errnote=errnote,
912 transform_source=transform_source, fatal=fatal, encoding=encoding,
913 data=data, headers=headers, query=query,
914 expected_status=expected_status)
915 return res if res is False else res[0]
917 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
919 xml_string = transform_source(xml_string)
921 return compat_etree_fromstring(xml_string.encode('utf
-8'))
922 except compat_xml_parse_error as ve:
923 errmsg = '%s: Failed to parse XML
' % video_id
925 raise ExtractorError(errmsg, cause=ve)
927 self.report_warning(errmsg + str(ve))
929 def _download_json_handle(
930 self, url_or_request, video_id, note='Downloading JSON metadata
',
931 errnote='Unable to download JSON metadata
', transform_source=None,
932 fatal=True, encoding=None, data=None, headers={}, query={},
933 expected_status=None):
935 Return a tuple (JSON object, URL handle).
937 See _download_webpage docstring for arguments specification.
939 res = self._download_webpage_handle(
940 url_or_request, video_id, note, errnote, fatal=fatal,
941 encoding=encoding, data=data, headers=headers, query=query,
942 expected_status=expected_status)
945 json_string, urlh = res
946 return self._parse_json(
947 json_string, video_id, transform_source=transform_source,
951 self, url_or_request, video_id, note='Downloading JSON metadata
',
952 errnote='Unable to download JSON metadata
', transform_source=None,
953 fatal=True, encoding=None, data=None, headers={}, query={},
954 expected_status=None):
956 Return the JSON object as a dict.
958 See _download_webpage docstring for arguments specification.
960 res = self._download_json_handle(
961 url_or_request, video_id, note=note, errnote=errnote,
962 transform_source=transform_source, fatal=fatal, encoding=encoding,
963 data=data, headers=headers, query=query,
964 expected_status=expected_status)
965 return res if res is False else res[0]
967 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
969 json_string = transform_source(json_string)
971 return json.loads(json_string)
972 except ValueError as ve:
973 errmsg = '%s: Failed to parse JSON
' % video_id
975 raise ExtractorError(errmsg, cause=ve)
977 self.report_warning(errmsg + str(ve))
979 def _parse_socket_response_as_json(self, data, video_id, transform_source=None, fatal=True):
980 return self._parse_json(
981 data[data.find('{'):data.rfind('}
') + 1],
982 video_id, transform_source, fatal)
984 def _download_socket_json_handle(
985 self, url_or_request, video_id, note='Polling socket
',
986 errnote='Unable to poll socket
', transform_source=None,
987 fatal=True, encoding=None, data=None, headers={}, query={},
988 expected_status=None):
990 Return a tuple (JSON object, URL handle).
992 See _download_webpage docstring for arguments specification.
994 res = self._download_webpage_handle(
995 url_or_request, video_id, note, errnote, fatal=fatal,
996 encoding=encoding, data=data, headers=headers, query=query,
997 expected_status=expected_status)
1001 return self._parse_socket_response_as_json(
1002 webpage, video_id, transform_source=transform_source,
1005 def _download_socket_json(
1006 self, url_or_request, video_id, note='Polling socket
',
1007 errnote='Unable to poll socket
', transform_source=None,
1008 fatal=True, encoding=None, data=None, headers={}, query={},
1009 expected_status=None):
1011 Return the JSON object as a dict.
1013 See _download_webpage docstring for arguments specification.
1015 res = self._download_socket_json_handle(
1016 url_or_request, video_id, note=note, errnote=errnote,
1017 transform_source=transform_source, fatal=fatal, encoding=encoding,
1018 data=data, headers=headers, query=query,
1019 expected_status=expected_status)
1020 return res if res is False else res[0]
1022 def report_warning(self, msg, video_id=None, *args, only_once=False, **kwargs):
1023 idstr = format_field(video_id, template='%s: ')
1024 msg = f'[{self.IE_NAME}
] {idstr}{msg}
'
1026 if f'WARNING
: {msg}
' in self._printed_messages:
1028 self._printed_messages.add(f'WARNING
: {msg}
')
1029 self._downloader.report_warning(msg, *args, **kwargs)
1031 def to_screen(self, msg, *args, **kwargs):
1032 """Print msg to screen, prefixing it with '[ie_name
]'"""
1033 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg), *args, **kwargs)
1035 def write_debug(self, msg, *args, **kwargs):
1036 self._downloader.write_debug('[%s] %s' % (self.IE_NAME, msg), *args, **kwargs)
1038 def get_param(self, name, default=None, *args, **kwargs):
1039 if self._downloader:
1040 return self._downloader.params.get(name, default, *args, **kwargs)
1043 def report_drm(self, video_id, partial=False):
1044 self.raise_no_formats('This video
is DRM protected
', expected=True, video_id=video_id)
1046 def report_extraction(self, id_or_name):
1047 """Report information extraction."""
1048 self.to_screen('%s: Extracting information
' % id_or_name)
1050 def report_download_webpage(self, video_id):
1051 """Report webpage download."""
1052 self.to_screen('%s: Downloading webpage
' % video_id)
1054 def report_age_confirmation(self):
1055 """Report attempt to confirm age."""
1056 self.to_screen('Confirming age
')
1058 def report_login(self):
1059 """Report attempt to log in."""
1060 self.to_screen('Logging
in')
1062 def raise_login_required(
1063 self, msg='This video
is only available
for registered users
',
1064 metadata_available=False, method='any
'):
1065 if metadata_available and self.get_param('ignore_no_formats_error
'):
1066 self.report_warning(msg)
1067 if method is not None:
1068 msg = '%s. %s' % (msg, self._LOGIN_HINTS[method])
1069 raise ExtractorError(msg, expected=True)
1071 def raise_geo_restricted(
1072 self, msg='This video
is not available
from your location due to geo restriction
',
1073 countries=None, metadata_available=False):
1074 if metadata_available and self.get_param('ignore_no_formats_error
'):
1075 self.report_warning(msg)
1077 raise GeoRestrictedError(msg, countries=countries)
1079 def raise_no_formats(self, msg, expected=False, video_id=None):
1080 if expected and self.get_param('ignore_no_formats_error
'):
1081 self.report_warning(msg, video_id)
1082 elif isinstance(msg, ExtractorError):
1085 raise ExtractorError(msg, expected=expected, video_id=video_id)
1087 # Methods for following #608
1089 def url_result(url, ie=None, video_id=None, video_title=None, **kwargs):
1090 """Returns a URL that points to a page that should be processed"""
1091 # TODO: ie should be the class used for getting the info
1092 video_info = {'_type
': 'url
',
1095 video_info.update(kwargs)
1096 if video_id is not None:
1097 video_info['id'] = video_id
1098 if video_title is not None:
1099 video_info['title
'] = video_title
1102 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None):
1104 self.url_result(self._proto_relative_url(getter(m) if getter else m), ie)
1106 return self.playlist_result(
1107 urls, playlist_id=playlist_id, playlist_title=playlist_title)
1110 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None, **kwargs):
1111 """Returns a playlist"""
1112 video_info = {'_type
': 'playlist
',
1114 video_info.update(kwargs)
1116 video_info['id'] = playlist_id
1118 video_info['title
'] = playlist_title
1119 if playlist_description is not None:
1120 video_info['description
'] = playlist_description
1123 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1125 Perform a regex search on the given string, using a single or a list of
1126 patterns returning the first matching group.
1127 In case of failure return a default value or raise a WARNING or a
1128 RegexNotFoundError, depending on fatal, specifying the field name.
1130 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
1131 mobj = re.search(pattern, string, flags)
1134 mobj = re.search(p, string, flags)
1138 _name = self._downloader._color_text(name, 'blue
')
1142 # return the first matching group
1143 return next(g for g in mobj.groups() if g is not None)
1144 elif isinstance(group, (list, tuple)):
1145 return tuple(mobj.group(g) for g in group)
1147 return mobj.group(group)
1148 elif default is not NO_DEFAULT:
1151 raise RegexNotFoundError('Unable to extract
%s' % _name)
1153 self.report_warning('unable to extract
%s' % _name + bug_reports_message())
1156 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1158 Like _search_regex, but strips HTML tags and unescapes entities.
1160 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1162 return clean_html(res).strip()
1166 def _get_netrc_login_info(self, netrc_machine=None):
1169 netrc_machine = netrc_machine or self._NETRC_MACHINE
1171 if self.get_param('usenetrc
', False):
1173 netrc_file = compat_expanduser(self.get_param('netrc_location
') or '~
')
1174 if os.path.isdir(netrc_file):
1175 netrc_file = os.path.join(netrc_file, '.netrc
')
1176 info = netrc.netrc(file=netrc_file).authenticators(netrc_machine)
1177 if info is not None:
1181 raise netrc.NetrcParseError(
1182 'No authenticators
for %s' % netrc_machine)
1183 except (IOError, netrc.NetrcParseError) as err:
1184 self.report_warning(
1185 'parsing
.netrc
: %s' % error_to_compat_str(err))
1187 return username, password
1189 def _get_login_info(self, username_option='username
', password_option='password
', netrc_machine=None):
1191 Get the login info as (username, password)
1192 First look for the manually specified credentials using username_option
1193 and password_option as keys in params dictionary. If no such credentials
1194 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1196 If there's no info available
, return (None, None)
1199 # Attempt to use provided username and password or .netrc data
1200 username = self.get_param(username_option)
1201 if username is not None:
1202 password = self.get_param(password_option)
1204 username, password = self._get_netrc_login_info(netrc_machine)
1206 return username, password
1208 def _get_tfa_info(self, note='two-factor verification code'):
1210 Get the two
-factor authentication info
1211 TODO
- asking the user will be required
for sms
/phone verify
1212 currently just uses the command line option
1213 If there
's no info available, return None
1216 tfa = self.get_param('twofactor
')
1220 return compat_getpass('Type
%s and press
[Return
]: ' % note)
1222 # Helper functions for extracting OpenGraph info
1224 def _og_regexes(prop):
1225 content_re = r'content
=(?
:"([^"]+?
)"|\'([^\']+?)\'|\s*([^\s"\'=<>`
]+?
))'
1226 property_re = (r'(?
:name|
property)=(?
:\'og
[:-]%(prop)s\'|
"og[:-]%(prop)s"|\s
*og
[:-]%(prop)s\b)'
1227 % {'prop': re.escape(prop)})
1228 template = r'<meta
[^
>]+?
%s[^
>]+?
%s'
1230 template % (property_re, content_re),
1231 template % (content_re, property_re),
1235 def _meta_regex(prop):
1236 return r'''(?isx)<meta
1237 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1238 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1240 def _og_search_property(self, prop, html, name=None, **kargs):
1241 prop = variadic(prop)
1243 name = 'OpenGraph
%s' % prop[0]
1246 og_regexes.extend(self._og_regexes(p))
1247 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1250 return unescapeHTML(escaped)
1252 def _og_search_thumbnail(self, html, **kargs):
1253 return self._og_search_property('image
', html, 'thumbnail URL
', fatal=False, **kargs)
1255 def _og_search_description(self, html, **kargs):
1256 return self._og_search_property('description
', html, fatal=False, **kargs)
1258 def _og_search_title(self, html, **kargs):
1259 return self._og_search_property('title
', html, **kargs)
1261 def _og_search_video_url(self, html, name='video url
', secure=True, **kargs):
1262 regexes = self._og_regexes('video
') + self._og_regexes('video
:url
')
1264 regexes = self._og_regexes('video
:secure_url
') + regexes
1265 return self._html_search_regex(regexes, html, name, **kargs)
1267 def _og_search_url(self, html, **kargs):
1268 return self._og_search_property('url
', html, **kargs)
1270 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1271 name = variadic(name)
1272 if display_name is None:
1273 display_name = name[0]
1274 return self._html_search_regex(
1275 [self._meta_regex(n) for n in name],
1276 html, display_name, fatal=fatal, group='content
', **kwargs)
1278 def _dc_search_uploader(self, html):
1279 return self._html_search_meta('dc
.creator
', html, 'uploader
')
1281 def _rta_search(self, html):
1282 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1283 if re.search(r'(?ix
)<meta\s
+name
="rating"\s
+'
1284 r' content
="RTA-5042-1996-1400-1577-RTA"',
1289 def _media_rating_search(self, html):
1290 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1291 rating = self._html_search_meta('rating
', html)
1303 return RATING_TABLE.get(rating.lower())
1305 def _family_friendly_search(self, html):
1306 # See http://schema.org/VideoObject
1307 family_friendly = self._html_search_meta(
1308 'isFamilyFriendly
', html, default=None)
1310 if not family_friendly:
1319 return RATING_TABLE.get(family_friendly.lower())
1321 def _twitter_search_player(self, html):
1322 return self._html_search_meta('twitter
:player
', html,
1323 'twitter card player
')
1325 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
1326 json_ld_list = list(re.finditer(JSON_LD_RE, html))
1327 default = kwargs.get('default
', NO_DEFAULT)
1328 # JSON-LD may be malformed and thus `fatal` should be respected.
1329 # At the same time `default` may be passed that assumes `fatal=False`
1330 # for _search_regex. Let's simulate the same behavior here
as well
.
1331 fatal
= kwargs
.get('fatal', True) if default
is NO_DEFAULT
else False
1333 for mobj
in json_ld_list
:
1334 json_ld_item
= self
._parse
_json
(
1335 mobj
.group('json_ld'), video_id
, fatal
=fatal
)
1336 if not json_ld_item
:
1338 if isinstance(json_ld_item
, dict):
1339 json_ld
.append(json_ld_item
)
1340 elif isinstance(json_ld_item
, (list, tuple)):
1341 json_ld
.extend(json_ld_item
)
1343 json_ld
= self
._json
_ld
(json_ld
, video_id
, fatal
=fatal
, expected_type
=expected_type
)
1346 if default
is not NO_DEFAULT
:
1349 raise RegexNotFoundError('Unable to extract JSON-LD')
1351 self
.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1354 def _json_ld(self
, json_ld
, video_id
, fatal
=True, expected_type
=None):
1355 if isinstance(json_ld
, compat_str
):
1356 json_ld
= self
._parse
_json
(json_ld
, video_id
, fatal
=fatal
)
1360 if not isinstance(json_ld
, (list, tuple, dict)):
1362 if isinstance(json_ld
, dict):
1365 INTERACTION_TYPE_MAP
= {
1366 'CommentAction': 'comment',
1367 'AgreeAction': 'like',
1368 'DisagreeAction': 'dislike',
1369 'LikeAction': 'like',
1370 'DislikeAction': 'dislike',
1371 'ListenAction': 'view',
1372 'WatchAction': 'view',
1373 'ViewAction': 'view',
1376 def extract_interaction_type(e
):
1377 interaction_type
= e
.get('interactionType')
1378 if isinstance(interaction_type
, dict):
1379 interaction_type
= interaction_type
.get('@type')
1380 return str_or_none(interaction_type
)
1382 def extract_interaction_statistic(e
):
1383 interaction_statistic
= e
.get('interactionStatistic')
1384 if isinstance(interaction_statistic
, dict):
1385 interaction_statistic
= [interaction_statistic
]
1386 if not isinstance(interaction_statistic
, list):
1388 for is_e
in interaction_statistic
:
1389 if not isinstance(is_e
, dict):
1391 if is_e
.get('@type') != 'InteractionCounter':
1393 interaction_type
= extract_interaction_type(is_e
)
1394 if not interaction_type
:
1396 # For interaction count some sites provide string instead of
1397 # an integer (as per spec) with non digit characters (e.g. ",")
1398 # so extracting count with more relaxed str_to_int
1399 interaction_count
= str_to_int(is_e
.get('userInteractionCount'))
1400 if interaction_count
is None:
1402 count_kind
= INTERACTION_TYPE_MAP
.get(interaction_type
.split('/')[-1])
1405 count_key
= '%s_count' % count_kind
1406 if info
.get(count_key
) is not None:
1408 info
[count_key
] = interaction_count
1410 def extract_video_object(e
):
1411 assert e
['@type'] == 'VideoObject'
1412 author
= e
.get('author')
1414 'url': url_or_none(e
.get('contentUrl')),
1415 'title': unescapeHTML(e
.get('name')),
1416 'description': unescapeHTML(e
.get('description')),
1417 'thumbnail': url_or_none(e
.get('thumbnailUrl') or e
.get('thumbnailURL')),
1418 'duration': parse_duration(e
.get('duration')),
1419 'timestamp': unified_timestamp(e
.get('uploadDate')),
1420 # author can be an instance of 'Organization' or 'Person' types.
1421 # both types can have 'name' property(inherited from 'Thing' type). [1]
1422 # however some websites are using 'Text' type instead.
1423 # 1. https://schema.org/VideoObject
1424 'uploader': author
.get('name') if isinstance(author
, dict) else author
if isinstance(author
, compat_str
) else None,
1425 'filesize': float_or_none(e
.get('contentSize')),
1426 'tbr': int_or_none(e
.get('bitrate')),
1427 'width': int_or_none(e
.get('width')),
1428 'height': int_or_none(e
.get('height')),
1429 'view_count': int_or_none(e
.get('interactionCount')),
1431 extract_interaction_statistic(e
)
1435 item_type
= e
.get('@type')
1436 if expected_type
is not None and expected_type
!= item_type
:
1438 if item_type
in ('TVEpisode', 'Episode'):
1439 episode_name
= unescapeHTML(e
.get('name'))
1441 'episode': episode_name
,
1442 'episode_number': int_or_none(e
.get('episodeNumber')),
1443 'description': unescapeHTML(e
.get('description')),
1445 if not info
.get('title') and episode_name
:
1446 info
['title'] = episode_name
1447 part_of_season
= e
.get('partOfSeason')
1448 if isinstance(part_of_season
, dict) and part_of_season
.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1450 'season': unescapeHTML(part_of_season
.get('name')),
1451 'season_number': int_or_none(part_of_season
.get('seasonNumber')),
1453 part_of_series
= e
.get('partOfSeries') or e
.get('partOfTVSeries')
1454 if isinstance(part_of_series
, dict) and part_of_series
.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1455 info
['series'] = unescapeHTML(part_of_series
.get('name'))
1456 elif item_type
== 'Movie':
1458 'title': unescapeHTML(e
.get('name')),
1459 'description': unescapeHTML(e
.get('description')),
1460 'duration': parse_duration(e
.get('duration')),
1461 'timestamp': unified_timestamp(e
.get('dateCreated')),
1463 elif item_type
in ('Article', 'NewsArticle'):
1465 'timestamp': parse_iso8601(e
.get('datePublished')),
1466 'title': unescapeHTML(e
.get('headline')),
1467 'description': unescapeHTML(e
.get('articleBody')),
1469 elif item_type
== 'VideoObject':
1470 extract_video_object(e
)
1471 if expected_type
is None:
1475 video
= e
.get('video')
1476 if isinstance(video
, dict) and video
.get('@type') == 'VideoObject':
1477 extract_video_object(video
)
1478 if expected_type
is None:
1482 return dict((k
, v
) for k
, v
in info
.items() if v
is not None)
1485 def _hidden_inputs(html
):
1486 html
= re
.sub(r
'<!--(?:(?!<!--).)*-->', '', html
)
1488 for input in re
.findall(r
'(?i)(<input[^>]+>)', html
):
1489 attrs
= extract_attributes(input)
1492 if attrs
.get('type') not in ('hidden', 'submit'):
1494 name
= attrs
.get('name') or attrs
.get('id')
1495 value
= attrs
.get('value')
1496 if name
and value
is not None:
1497 hidden_inputs
[name
] = value
1498 return hidden_inputs
1500 def _form_hidden_inputs(self
, form_id
, html
):
1501 form
= self
._search
_regex
(
1502 r
'(?is)<form[^>]+?id=(["\'])%s\
1[^
>]*>(?P
<form
>.+?
)</form
>' % form_id,
1503 html, '%s form
' % form_id, group='form
')
1504 return self._hidden_inputs(form)
1507 regex = r' *((?P
<reverse
>\
+)?
(?P
<field
>[a
-zA
-Z0
-9_]+)((?P
<separator
>[~
:])(?P
<limit
>.*?
))?
)?
*$
'
1509 default = ('hidden
', 'aud_or_vid
', 'hasvid
', 'ie_pref
', 'lang
', 'quality
',
1510 'res
', 'fps
', 'codec
:vp9
.2
', 'size
', 'br
', 'asr
',
1511 'proto
', 'ext
', 'hasaud
', 'source
', 'format_id
') # These must not be aliases
1512 ytdl_default = ('hasaud
', 'lang
', 'quality
', 'tbr
', 'filesize
', 'vbr
',
1513 'height
', 'width
', 'proto
', 'vext
', 'abr
', 'aext
',
1514 'fps
', 'fs_approx
', 'source
', 'format_id
')
1517 'vcodec
': {'type': 'ordered
', 'regex
': True,
1518 'order
': ['av0?
1', 'vp0?
9.2', 'vp0?
9', '[hx
]265|he?vc?
', '[hx
]264|avc
', 'vp0?
8', 'mp4v|h263
', 'theora
', '', None, 'none
']},
1519 'acodec
': {'type': 'ordered
', 'regex
': True,
1520 'order
': ['opus
', 'vorbis
', 'aac
', 'mp?
4a?
', 'mp3
', 'e?a?c
-?
3', 'dts
', '', None, 'none
']},
1521 'proto
': {'type': 'ordered
', 'regex
': True, 'field
': 'protocol
',
1522 'order
': ['(ht|f
)tps
', '(ht|f
)tp$
', 'm3u8
.+', '.*dash
', 'ws|websocket
', '', 'mms|rtsp
', 'none
', 'f4
']},
1523 'vext
': {'type': 'ordered
', 'field
': 'video_ext
',
1524 'order
': ('mp4
', 'webm
', 'flv
', '', 'none
'),
1525 'order_free
': ('webm
', 'mp4
', 'flv
', '', 'none
')},
1526 'aext
': {'type': 'ordered
', 'field
': 'audio_ext
',
1527 'order
': ('m4a
', 'aac
', 'mp3
', 'ogg
', 'opus
', 'webm
', '', 'none
'),
1528 'order_free
': ('opus
', 'ogg
', 'webm
', 'm4a
', 'mp3
', 'aac
', '', 'none
')},
1529 'hidden
': {'visible': False, 'forced': True, 'type': 'extractor', 'max': -1000},
1530 'aud_or_vid
': {'visible
': False, 'forced
': True, 'type': 'multiple
',
1531 'field
': ('vcodec
', 'acodec
'),
1532 'function
': lambda it: int(any(v != 'none
' for v in it))},
1533 'ie_pref
': {'priority': True, 'type': 'extractor'},
1534 'hasvid
': {'priority': True, 'field': 'vcodec', 'type': 'boolean', 'not_in_list': ('none',)},
1535 'hasaud
': {'field': 'acodec', 'type': 'boolean', 'not_in_list': ('none',)},
1536 'lang
': {'convert': 'ignore', 'field': 'language_preference'},
1537 'quality
': {'convert': 'float_none', 'default': -1},
1538 'filesize
': {'convert': 'bytes'},
1539 'fs_approx
': {'convert': 'bytes', 'field': 'filesize_approx'},
1540 'id': {'convert': 'string', 'field': 'format_id'},
1541 'height
': {'convert': 'float_none'},
1542 'width
': {'convert': 'float_none'},
1543 'fps
': {'convert': 'float_none'},
1544 'tbr
': {'convert': 'float_none'},
1545 'vbr
': {'convert': 'float_none'},
1546 'abr
': {'convert': 'float_none'},
1547 'asr
': {'convert': 'float_none'},
1548 'source
': {'convert': 'ignore', 'field': 'source_preference'},
1550 'codec
': {'type': 'combined', 'field': ('vcodec', 'acodec')},
1551 'br
': {'type': 'combined', 'field': ('tbr', 'vbr', 'abr'), 'same_limit': True},
1552 'size
': {'type': 'combined', 'same_limit': True, 'field': ('filesize', 'fs_approx')},
1553 'ext
': {'type': 'combined', 'field': ('vext', 'aext')},
1554 'res
': {'type': 'multiple
', 'field
': ('height
', 'width
'),
1555 'function
': lambda it: (lambda l: min(l) if l else 0)(tuple(filter(None, it)))},
1557 # Most of these exist only for compatibility reasons
1558 'dimension
': {'type': 'alias', 'field': 'res'},
1559 'resolution
': {'type': 'alias', 'field': 'res'},
1560 'extension
': {'type': 'alias', 'field': 'ext'},
1561 'bitrate
': {'type': 'alias', 'field': 'br'},
1562 'total_bitrate
': {'type': 'alias', 'field': 'tbr'},
1563 'video_bitrate
': {'type': 'alias', 'field': 'vbr'},
1564 'audio_bitrate
': {'type': 'alias', 'field': 'abr'},
1565 'framerate
': {'type': 'alias', 'field': 'fps'},
1566 'language_preference
': {'type': 'alias', 'field': 'lang'}, # not named as 'language
' because such a field exists
1567 'protocol
': {'type': 'alias', 'field': 'proto'},
1568 'source_preference
': {'type': 'alias', 'field': 'source'},
1569 'filesize_approx
': {'type': 'alias', 'field': 'fs_approx'},
1570 'filesize_estimate
': {'type': 'alias', 'field': 'size'},
1571 'samplerate
': {'type': 'alias', 'field': 'asr'},
1572 'video_ext
': {'type': 'alias', 'field': 'vext'},
1573 'audio_ext
': {'type': 'alias', 'field': 'aext'},
1574 'video_codec
': {'type': 'alias', 'field': 'vcodec'},
1575 'audio_codec
': {'type': 'alias', 'field': 'acodec'},
1576 'video
': {'type': 'alias', 'field': 'hasvid'},
1577 'has_video
': {'type': 'alias', 'field': 'hasvid'},
1578 'audio
': {'type': 'alias', 'field': 'hasaud'},
1579 'has_audio
': {'type': 'alias', 'field': 'hasaud'},
1580 'extractor
': {'type': 'alias', 'field': 'ie_pref'},
1581 'preference
': {'type': 'alias', 'field': 'ie_pref'},
1582 'extractor_preference
': {'type': 'alias', 'field': 'ie_pref'},
1583 'format_id
': {'type': 'alias', 'field': 'id'},
1588 def _get_field_setting(self, field, key):
1589 if field not in self.settings:
1590 self.settings[field] = {}
1591 propObj = self.settings[field]
1592 if key not in propObj:
1593 type = propObj.get('type')
1595 default = 'preference
' if type == 'extractor
' else (field,) if type in ('combined
', 'multiple
') else field
1596 elif key == 'convert
':
1597 default = 'order
' if type == 'ordered
' else 'float_string
' if field else 'ignore
'
1599 default = {'type': 'field', 'visible': True, 'order': [], 'not_in_list': (None,)}.get(key, None)
1600 propObj[key] = default
1603 def _resolve_field_value(self, field, value, convertNone=False):
1608 value = value.lower()
1609 conversion = self._get_field_setting(field, 'convert
')
1610 if conversion == 'ignore
':
1612 if conversion == 'string
':
1614 elif conversion == 'float_none
':
1615 return float_or_none(value)
1616 elif conversion == 'bytes':
1617 return FileDownloader.parse_bytes(value)
1618 elif conversion == 'order
':
1619 order_list = (self._use_free_order and self._get_field_setting(field, 'order_free
')) or self._get_field_setting(field, 'order
')
1620 use_regex = self._get_field_setting(field, 'regex
')
1621 list_length = len(order_list)
1622 empty_pos = order_list.index('') if '' in order_list else list_length + 1
1623 if use_regex and value is not None:
1624 for i, regex in enumerate(order_list):
1625 if regex and re.match(regex, value):
1626 return list_length - i
1627 return list_length - empty_pos # not in list
1628 else: # not regex or value = None
1629 return list_length - (order_list.index(value) if value in order_list else empty_pos)
1631 if value.isnumeric():
1634 self.settings[field]['convert
'] = 'string
'
1637 def evaluate_params(self, params, sort_extractor):
1638 self._use_free_order = params.get('prefer_free_formats
', False)
1639 self._sort_user = params.get('format_sort
', [])
1640 self._sort_extractor = sort_extractor
1642 def add_item(field, reverse, closest, limit_text):
1643 field = field.lower()
1644 if field in self._order:
1646 self._order.append(field)
1647 limit = self._resolve_field_value(field, limit_text)
1650 'closest
': False if limit is None else closest,
1651 'limit_text
': limit_text,
1653 if field in self.settings:
1654 self.settings[field].update(data)
1656 self.settings[field] = data
1659 tuple(field for field in self.default if self._get_field_setting(field, 'forced
'))
1660 + (tuple() if params.get('format_sort_force
', False)
1661 else tuple(field for field in self.default if self._get_field_setting(field, 'priority
')))
1662 + tuple(self._sort_user) + tuple(sort_extractor) + self.default)
1664 for item in sort_list:
1665 match = re.match(self.regex, item)
1667 raise ExtractorError('Invalid format sort string
"%s" given by extractor
' % item)
1668 field = match.group('field
')
1671 if self._get_field_setting(field, 'type') == 'alias
':
1672 field = self._get_field_setting(field, 'field
')
1673 reverse = match.group('reverse
') is not None
1674 closest = match.group('separator
') == '~
'
1675 limit_text = match.group('limit
')
1677 has_limit = limit_text is not None
1678 has_multiple_fields = self._get_field_setting(field, 'type') == 'combined
'
1679 has_multiple_limits = has_limit and has_multiple_fields and not self._get_field_setting(field, 'same_limit
')
1681 fields = self._get_field_setting(field, 'field
') if has_multiple_fields else (field,)
1682 limits = limit_text.split(':') if has_multiple_limits else (limit_text,) if has_limit else tuple()
1683 limit_count = len(limits)
1684 for (i, f) in enumerate(fields):
1685 add_item(f, reverse, closest,
1686 limits[i] if i < limit_count
1687 else limits[0] if has_limit and not has_multiple_limits
1690 def print_verbose_info(self, write_debug):
1692 write_debug('Sort order given by user
: %s' % ', '.join(self._sort_user))
1693 if self._sort_extractor:
1694 write_debug('Sort order given by extractor
: %s' % ', '.join(self._sort_extractor))
1695 write_debug('Formats
sorted by
: %s' % ', '.join(['%s%s%s' % (
1696 '+' if self._get_field_setting(field, 'reverse
') else '', field,
1697 '%s%s(%s)' % ('~
' if self._get_field_setting(field, 'closest
') else ':',
1698 self._get_field_setting(field, 'limit_text
'),
1699 self._get_field_setting(field, 'limit
'))
1700 if self._get_field_setting(field, 'limit_text
') is not None else '')
1701 for field in self._order if self._get_field_setting(field, 'visible
')]))
1703 def _calculate_field_preference_from_value(self, format, field, type, value):
1704 reverse = self._get_field_setting(field, 'reverse
')
1705 closest = self._get_field_setting(field, 'closest
')
1706 limit = self._get_field_setting(field, 'limit
')
1708 if type == 'extractor
':
1709 maximum = self._get_field_setting(field, 'max')
1710 if value is None or (maximum is not None and value >= maximum):
1712 elif type == 'boolean
':
1713 in_list = self._get_field_setting(field, 'in_list
')
1714 not_in_list = self._get_field_setting(field, 'not_in_list
')
1715 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
1716 elif type == 'ordered
':
1717 value = self._resolve_field_value(field, value, True)
1719 # try to convert to number
1720 val_num = float_or_none(value, default=self._get_field_setting(field, 'default
'))
1721 is_num = self._get_field_setting(field, 'convert
') != 'string
' and val_num is not None
1725 return ((-10, 0) if value is None
1726 else (1, value, 0) if not is_num # if a field has mixed strings and numbers, strings are sorted higher
1727 else (0, -abs(value - limit), value - limit if reverse else limit - value) if closest
1728 else (0, value, 0) if not reverse and (limit is None or value <= limit)
1729 else (0, -value, 0) if limit is None or (reverse and value == limit) or value > limit
1730 else (-1, value, 0))
1732 def _calculate_field_preference(self, format, field):
1733 type = self._get_field_setting(field, 'type') # extractor, boolean, ordered, field, multiple
1734 get_value = lambda f: format.get(self._get_field_setting(f, 'field
'))
1735 if type == 'multiple
':
1736 type = 'field
' # Only 'field
' is allowed in multiple for now
1737 actual_fields = self._get_field_setting(field, 'field
')
1739 value = self._get_field_setting(field, 'function
')(get_value(f) for f in actual_fields)
1741 value = get_value(field)
1742 return self._calculate_field_preference_from_value(format, field, type, value)
1744 def calculate_preference(self, format):
1745 # Determine missing protocol
1746 if not format.get('protocol
'):
1747 format['protocol
'] = determine_protocol(format)
1749 # Determine missing ext
1750 if not format.get('ext
') and 'url
' in format:
1751 format['ext
'] = determine_ext(format['url
'])
1752 if format.get('vcodec
') == 'none
':
1753 format['audio_ext
'] = format['ext
'] if format.get('acodec
') != 'none
' else 'none
'
1754 format['video_ext
'] = 'none
'
1756 format['video_ext
'] = format['ext
']
1757 format['audio_ext
'] = 'none
'
1758 # if format.get('preference
') is None and format.get('ext
') in ('f4f
', 'f4m
'): # Not supported?
1759 # format['preference
'] = -1000
1761 # Determine missing bitrates
1762 if format.get('tbr
') is None:
1763 if format.get('vbr
') is not None and format.get('abr
') is not None:
1764 format['tbr
'] = format.get('vbr
', 0) + format.get('abr
', 0)
1766 if format.get('vcodec
') != 'none
' and format.get('vbr
') is None:
1767 format['vbr
'] = format.get('tbr
') - format.get('abr
', 0)
1768 if format.get('acodec
') != 'none
' and format.get('abr
') is None:
1769 format['abr
'] = format.get('tbr
') - format.get('vbr
', 0)
1771 return tuple(self._calculate_field_preference(format, field) for field in self._order)
1773 def _sort_formats(self, formats, field_preference=[]):
1776 format_sort = self.FormatSort() # params and to_screen are taken from the downloader
1777 format_sort.evaluate_params(self._downloader.params, field_preference)
1778 if self.get_param('verbose
', False):
1779 format_sort.print_verbose_info(self._downloader.write_debug)
1780 formats.sort(key=lambda f: format_sort.calculate_preference(f))
1782 def _check_formats(self, formats, video_id):
1784 formats[:] = filter(
1785 lambda f: self._is_valid_url(
1787 item='%s video format
' % f.get('format_id
') if f.get('format_id
') else 'video
'),
1791 def _remove_duplicate_formats(formats):
1795 if f['url
'] not in format_urls:
1796 format_urls.add(f['url
'])
1797 unique_formats.append(f)
1798 formats[:] = unique_formats
1800 def _is_valid_url(self, url, video_id, item='video
', headers={}):
1801 url = self._proto_relative_url(url, scheme='http
:')
1802 # For now assume non HTTP(S) URLs always valid
1803 if not (url.startswith('http
://') or url.startswith('https
://')):
1806 self._request_webpage(url, video_id, 'Checking
%s URL
' % item, headers=headers)
1808 except ExtractorError as e:
1810 '%s: %s URL
is invalid
, skipping
: %s'
1811 % (video_id, item, error_to_compat_str(e.cause)))
1814 def http_scheme(self):
1815 """ Either "http:" or "https:", depending on the user's preferences
"""
1818 if self.get_param('prefer_insecure', False)
1821 def _proto_relative_url(self, url, scheme=None):
1824 if url.startswith('//'):
1826 scheme = self.http_scheme()
1831 def _sleep(self, timeout, video_id, msg_template=None):
1832 if msg_template is None:
1833 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1834 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1838 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1839 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1840 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
1841 manifest = self._download_xml(
1842 manifest_url, video_id, 'Downloading f4m manifest',
1843 'Unable to download f4m manifest',
1844 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1845 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
1846 transform_source=transform_source,
1847 fatal=fatal, data=data, headers=headers, query=query)
1849 if manifest is False:
1852 return self._parse_f4m_formats(
1853 manifest, manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1854 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1856 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1857 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1858 fatal=True, m3u8_id=None):
1859 if not isinstance(manifest, compat_etree_Element) and not fatal:
1862 # currently yt-dlp cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1863 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1864 if akamai_pv is not None and ';' in akamai_pv.text:
1865 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1866 if playerVerificationChallenge.strip() != '':
1870 manifest_version = '1.0'
1871 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1873 manifest_version = '2.0'
1874 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1875 # Remove unsupported DRM protected media from final formats
1876 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
1877 media_nodes = remove_encrypted_media(media_nodes)
1881 manifest_base_url = get_base_url(manifest)
1883 bootstrap_info = xpath_element(
1884 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1885 'bootstrap info', default=None)
1888 mime_type = xpath_text(
1889 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1890 'base URL', default=None)
1891 if mime_type and mime_type.startswith('audio/'):
1894 for i, media_el in enumerate(media_nodes):
1895 tbr = int_or_none(media_el.attrib.get('bitrate'))
1896 width = int_or_none(media_el.attrib.get('width'))
1897 height = int_or_none(media_el.attrib.get('height'))
1898 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1899 # If <bootstrapInfo> is present, the specified f4m is a
1900 # stream-level manifest, and only set-level manifests may refer to
1901 # external resources. See section 11.4 and section 4 of F4M spec
1902 if bootstrap_info is None:
1904 # @href is introduced in 2.0, see section 11.6 of F4M spec
1905 if manifest_version == '2.0':
1906 media_url = media_el.attrib.get('href')
1907 if media_url is None:
1908 media_url = media_el.attrib.get('url')
1912 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1913 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1914 # If media_url is itself a f4m manifest do the recursive extraction
1915 # since bitrates in parent manifest (this one) and media_url manifest
1916 # may differ leading to inability to resolve the format by requested
1917 # bitrate in f4m downloader
1918 ext = determine_ext(manifest_url)
1920 f4m_formats = self._extract_f4m_formats(
1921 manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1922 transform_source=transform_source, fatal=fatal)
1923 # Sometimes stream-level manifest contains single media entry that
1924 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1925 # At the same time parent's media entry in set-level manifest may
1926 # contain it. We will copy it from parent in such cases.
1927 if len(f4m_formats) == 1:
1930 'tbr': f.get('tbr') or tbr,
1931 'width': f.get('width') or width,
1932 'height': f.get('height') or height,
1933 'format_id': f.get('format_id') if not tbr else format_id,
1936 formats.extend(f4m_formats)
1939 formats.extend(self._extract_m3u8_formats(
1940 manifest_url, video_id, 'mp4', preference=preference,
1941 quality=quality, m3u8_id=m3u8_id, fatal=fatal))
1944 'format_id': format_id,
1945 'url': manifest_url,
1946 'manifest_url': manifest_url,
1947 'ext': 'flv' if bootstrap_info is not None else None,
1953 'preference': preference,
1958 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, quality=None, m3u8_id=None):
1960 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1964 'preference': preference - 100 if preference else -100,
1966 'resolution': 'multiple',
1967 'format_note': 'Quality selection URL',
1970 def _report_ignoring_subs(self, name):
1971 self.report_warning(bug_reports_message(
1972 f'Ignoring subtitle tracks found in the {name} manifest; '
1973 'if any subtitle tracks are missing,'
1976 def _extract_m3u8_formats(self, *args, **kwargs):
1977 fmts, subs = self._extract_m3u8_formats_and_subtitles(*args, **kwargs)
1979 self._report_ignoring_subs('HLS')
1982 def _extract_m3u8_formats_and_subtitles(
1983 self, m3u8_url, video_id, ext=None, entry_protocol='m3u8_native',
1984 preference=None, quality=None, m3u8_id=None, note=None,
1985 errnote=None, fatal=True, live=False, data=None, headers={},
1988 res = self._download_webpage_handle(
1990 note='Downloading m3u8 information' if note is None else note,
1991 errnote='Failed to download m3u8 information' if errnote is None else errnote,
1992 fatal=fatal, data=data, headers=headers, query=query)
1997 m3u8_doc, urlh = res
1998 m3u8_url = urlh.geturl()
2000 return self._parse_m3u8_formats_and_subtitles(
2001 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
2002 preference=preference, quality=quality, m3u8_id=m3u8_id,
2003 note=note, errnote=errnote, fatal=fatal, live=live, data=data,
2004 headers=headers, query=query, video_id=video_id)
2006 def _parse_m3u8_formats_and_subtitles(
2007 self, m3u8_doc, m3u8_url, ext=None, entry_protocol='m3u8_native',
2008 preference=None, quality=None, m3u8_id=None, live=False, note=None,
2009 errnote=None, fatal=True, data=None, headers={}, query={},
2011 formats, subtitles = [], {}
2013 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
2014 return formats, subtitles
2016 has_drm = re.search(r'#EXT-X-(?:SESSION-)?KEY:.*?URI="skd://', m3u8_doc)
2018 def format_url(url):
2019 return url if re.match(r'^https?://', url) else compat_urlparse.urljoin(m3u8_url, url)
2021 if self.get_param('hls_split_discontinuity', False):
2022 def _extract_m3u8_playlist_indices(manifest_url=None, m3u8_doc=None):
2024 if not manifest_url:
2026 m3u8_doc = self._download_webpage(
2027 manifest_url, video_id, fatal=fatal, data=data, headers=headers,
2028 note=False, errnote='Failed to download m3u8 playlist information')
2029 if m3u8_doc is False:
2031 return range(1 + sum(line.startswith('#EXT-X-DISCONTINUITY') for line in m3u8_doc.splitlines()))
2034 def _extract_m3u8_playlist_indices(*args, **kwargs):
2038 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
2039 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
2040 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
2042 # We should try extracting formats only from master playlists [1, 4.3.4],
2043 # i.e. playlists that describe available qualities. On the other hand
2044 # media playlists [1, 4.3.3] should be returned as is since they contain
2045 # just the media without qualities renditions.
2046 # Fortunately, master playlist can be easily distinguished from media
2047 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
2048 # master playlist tags MUST NOT appear in a media playlist and vice versa.
2049 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
2050 # media playlist and MUST NOT appear in master playlist thus we can
2051 # clearly detect media playlist with this criterion.
2053 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
2055 'format_id': '-'.join(map(str, filter(None, [m3u8_id, idx]))),
2056 'format_index': idx,
2059 'protocol': entry_protocol,
2060 'preference': preference,
2063 } for idx in _extract_m3u8_playlist_indices(m3u8_doc=m3u8_doc)]
2065 return formats, subtitles
2068 last_stream_inf = {}
2070 def extract_media(x_media_line):
2071 media = parse_m3u8_attributes(x_media_line)
2072 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
2073 media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME')
2074 if not (media_type and group_id and name):
2076 groups.setdefault(group_id, []).append(media)
2077 # <https://tools.ietf.org/html/rfc8216#section-4.3.4.1>
2078 if media_type == 'SUBTITLES':
2079 # According to RFC 8216 §4.3.4.2.1, URI is REQUIRED in the
2080 # EXT-X-MEDIA tag if the media type is SUBTITLES.
2081 # However, lack of URI has been spotted in the wild.
2082 # e.g. NebulaIE; see https://github.com/yt-dlp/yt-dlp/issues/339
2083 if not media.get('URI'):
2085 url = format_url(media['URI'])
2088 'ext': determine_ext(url),
2090 if sub_info['ext'] == 'm3u8':
2091 # Per RFC 8216 §3.1, the only possible subtitle format m3u8
2092 # files may contain is WebVTT:
2093 # <https://tools.ietf.org/html/rfc8216#section-3.1>
2094 sub_info['ext'] = 'vtt'
2095 sub_info['protocol'] = 'm3u8_native'
2096 lang = media.get('LANGUAGE') or 'und'
2097 subtitles.setdefault(lang, []).append(sub_info)
2098 if media_type not in ('VIDEO', 'AUDIO'):
2100 media_url = media.get('URI')
2102 manifest_url = format_url(media_url)
2104 'format_id': '-'.join(map(str, filter(None, (m3u8_id, group_id, name, idx)))),
2105 'format_note': name,
2106 'format_index': idx,
2107 'url': manifest_url,
2108 'manifest_url': m3u8_url,
2109 'language': media.get('LANGUAGE'),
2111 'protocol': entry_protocol,
2112 'preference': preference,
2114 'vcodec': 'none' if media_type == 'AUDIO' else None,
2115 } for idx in _extract_m3u8_playlist_indices(manifest_url))
2117 def build_stream_name():
2118 # Despite specification does not mention NAME attribute for
2119 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
2120 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
2121 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
2122 stream_name = last_stream_inf.get('NAME')
2125 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
2126 # from corresponding rendition group
2127 stream_group_id = last_stream_inf.get('VIDEO')
2128 if not stream_group_id:
2130 stream_group = groups.get(stream_group_id)
2131 if not stream_group:
2132 return stream_group_id
2133 rendition = stream_group[0]
2134 return rendition.get('NAME') or stream_group_id
2136 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
2137 # chance to detect video only formats when EXT-X-STREAM-INF tags
2138 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
2139 for line in m3u8_doc.splitlines():
2140 if line.startswith('#EXT-X-MEDIA:'):
2143 for line in m3u8_doc.splitlines():
2144 if line.startswith('#EXT-X-STREAM-INF:'):
2145 last_stream_inf = parse_m3u8_attributes(line)
2146 elif line.startswith('#') or not line.strip():
2149 tbr = float_or_none(
2150 last_stream_inf.get('AVERAGE-BANDWIDTH')
2151 or last_stream_inf.get('BANDWIDTH'), scale=1000)
2152 manifest_url = format_url(line.strip())
2154 for idx in _extract_m3u8_playlist_indices(manifest_url):
2155 format_id = [m3u8_id, None, idx]
2156 # Bandwidth of live streams may differ over time thus making
2157 # format_id unpredictable. So it's better to keep provided
2160 stream_name = build_stream_name()
2161 format_id[1] = stream_name if stream_name else '%d' % (tbr if tbr else len(formats))
2163 'format_id': '-'.join(map(str, filter(None, format_id))),
2164 'format_index': idx,
2165 'url': manifest_url,
2166 'manifest_url': m3u8_url,
2169 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')),
2170 'protocol': entry_protocol,
2171 'preference': preference,
2174 resolution = last_stream_inf.get('RESOLUTION')
2176 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
2178 f['width'] = int(mobj.group('width'))
2179 f['height'] = int(mobj.group('height'))
2180 # Unified Streaming Platform
2182 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
2184 abr, vbr = mobj.groups()
2185 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
2190 codecs = parse_codecs(last_stream_inf.get('CODECS'))
2192 audio_group_id = last_stream_inf.get('AUDIO')
2193 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
2194 # references a rendition group MUST have a CODECS attribute.
2195 # However, this is not always respected, for example, [2]
2196 # contains EXT-X-STREAM-INF tag which references AUDIO
2197 # rendition group but does not have CODECS and despite
2198 # referencing an audio group it represents a complete
2199 # (with audio and video) format. So, for such cases we will
2200 # ignore references to rendition groups and treat them
2201 # as complete formats.
2202 if audio_group_id and codecs and f.get('vcodec') != 'none':
2203 audio_group = groups.get(audio_group_id)
2204 if audio_group and audio_group[0].get('URI'):
2205 # TODO: update acodec for audio only formats with
2207 f['acodec'] = 'none'
2208 if not f.get('ext'):
2209 f['ext'] = 'm4a' if f.get('vcodec') == 'none' else 'mp4'
2213 progressive_uri = last_stream_inf.get('PROGRESSIVE-URI')
2216 del http_f['manifest_url']
2218 'format_id': f['format_id'].replace('hls-', 'http-'),
2220 'url': progressive_uri,
2222 formats.append(http_f)
2224 last_stream_inf = {}
2225 return formats, subtitles
2227 def _extract_m3u8_vod_duration(
2228 self, m3u8_vod_url, video_id, note=None, errnote=None, data=None, headers={}, query={}):
2230 m3u8_vod = self._download_webpage(
2231 m3u8_vod_url, video_id,
2232 note='Downloading m3u8 VOD manifest' if note is None else note,
2233 errnote='Failed to download VOD manifest' if errnote is None else errnote,
2234 fatal=False, data=data, headers=headers, query=query)
2236 return self._parse_m3u8_vod_duration(m3u8_vod or '', video_id)
2238 def _parse_m3u8_vod_duration(self, m3u8_vod, video_id):
2239 if '#EXT-X-PLAYLIST-TYPE:VOD' not in m3u8_vod:
2243 float(line[len('#EXTINF:'):].split(',')[0])
2244 for line in m3u8_vod.splitlines() if line.startswith('#EXTINF:'))) or None
2247 def _xpath_ns(path, namespace=None):
2251 for c in path.split('/'):
2252 if not c or c == '.':
2255 out.append('{%s}%s' % (namespace, c))
2256 return '/'.join(out)
2258 def _extract_smil_formats_and_subtitles(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
2259 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
2265 namespace = self._parse_smil_namespace(smil)
2267 fmts = self._parse_smil_formats(
2268 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
2269 subs = self._parse_smil_subtitles(
2270 smil, namespace=namespace)
2274 def _extract_smil_formats(self, *args, **kwargs):
2275 fmts, subs = self._extract_smil_formats_and_subtitles(*args, **kwargs)
2277 self._report_ignoring_subs('SMIL')
2280 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
2281 smil = self._download_smil(smil_url, video_id, fatal=fatal)
2284 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
2286 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
2287 return self._download_xml(
2288 smil_url, video_id, 'Downloading SMIL file',
2289 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
2291 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
2292 namespace = self._parse_smil_namespace(smil)
2294 formats = self._parse_smil_formats(
2295 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
2296 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
2298 video_id = os.path.splitext(url_basename(smil_url))[0]
2302 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
2303 name = meta.attrib.get('name')
2304 content = meta.attrib.get('content')
2305 if not name or not content:
2307 if not title and name == 'title':
2309 elif not description and name in ('description', 'abstract'):
2310 description = content
2311 elif not upload_date and name == 'date':
2312 upload_date = unified_strdate(content)
2315 'id': image.get('type'),
2316 'url': image.get('src'),
2317 'width': int_or_none(image.get('width')),
2318 'height': int_or_none(image.get('height')),
2319 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
2323 'title': title or video_id,
2324 'description': description,
2325 'upload_date': upload_date,
2326 'thumbnails': thumbnails,
2328 'subtitles': subtitles,
2331 def _parse_smil_namespace(self, smil):
2332 return self._search_regex(
2333 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
2335 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
2337 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
2338 b = meta.get('base') or meta.get('httpBase')
2350 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
2351 for medium in media:
2352 src = medium.get('src')
2353 if not src or src in srcs:
2357 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
2358 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
2359 width = int_or_none(medium.get('width'))
2360 height = int_or_none(medium.get('height'))
2361 proto = medium.get('proto')
2362 ext = medium.get('ext')
2363 src_ext = determine_ext(src)
2364 streamer = medium.get('streamer') or base
2366 if proto == 'rtmp' or streamer.startswith('rtmp'):
2372 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
2374 'filesize': filesize,
2378 if transform_rtmp_url:
2379 streamer, src = transform_rtmp_url(streamer, src)
2380 formats[-1].update({
2386 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
2387 src_url = src_url.strip()
2389 if proto == 'm3u8' or src_ext == 'm3u8':
2390 m3u8_formats = self._extract_m3u8_formats(
2391 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
2392 if len(m3u8_formats) == 1:
2394 m3u8_formats[0].update({
2395 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
2400 formats.extend(m3u8_formats)
2401 elif src_ext == 'f4m':
2406 'plugin': 'flowplayer-3.2.0.1',
2408 f4m_url += '&' if '?' in f4m_url else '?'
2409 f4m_url += compat_urllib_parse_urlencode(f4m_params)
2410 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
2411 elif src_ext == 'mpd':
2412 formats.extend(self._extract_mpd_formats(
2413 src_url, video_id, mpd_id='dash', fatal=False))
2414 elif re.search(r'\.ism/[Mm]anifest', src_url):
2415 formats.extend(self._extract_ism_formats(
2416 src_url, video_id, ism_id='mss', fatal=False))
2417 elif src_url.startswith('http') and self._is_valid_url(src, video_id):
2421 'ext': ext or src_ext or 'flv',
2422 'format_id': 'http-%d' % (bitrate or http_count),
2424 'filesize': filesize,
2429 for medium in smil.findall(self._xpath_ns('.//imagestream', namespace)):
2430 src = medium.get('src')
2431 if not src or src in srcs:
2437 'format_id': 'imagestream-%d' % (imgs_count),
2439 'ext': mimetype2ext(medium.get('type')),
2442 'width': int_or_none(medium.get('width')),
2443 'height': int_or_none(medium.get('height')),
2444 'format_note': 'SMIL storyboards',
2449 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
2452 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
2453 src = textstream.get('src')
2454 if not src or src in urls:
2457 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
2458 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
2459 subtitles.setdefault(lang, []).append({
2465 def _extract_xspf_playlist(self, xspf_url, playlist_id, fatal=True):
2466 xspf = self._download_xml(
2467 xspf_url, playlist_id, 'Downloading xpsf playlist',
2468 'Unable to download xspf manifest', fatal=fatal)
2471 return self._parse_xspf(
2472 xspf, playlist_id, xspf_url=xspf_url,
2473 xspf_base_url=base_url(xspf_url))
2475 def _parse_xspf(self, xspf_doc, playlist_id, xspf_url=None, xspf_base_url=None):
2477 'xspf': 'http://xspf.org/ns/0/',
2478 's1': 'http://static.streamone.nl/player/ns/0',
2482 for track in xspf_doc.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
2484 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
2485 description = xpath_text(
2486 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
2487 thumbnail = xpath_text(
2488 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
2489 duration = float_or_none(
2490 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
2493 for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP)):
2494 format_url = urljoin(xspf_base_url, location.text)
2499 'manifest_url': xspf_url,
2500 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
2501 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
2502 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
2504 self._sort_formats(formats)
2509 'description': description,
2510 'thumbnail': thumbnail,
2511 'duration': duration,
2516 def _extract_mpd_formats(self, *args, **kwargs):
2517 fmts, subs = self._extract_mpd_formats_and_subtitles(*args, **kwargs)
2519 self._report_ignoring_subs('DASH')
2522 def _extract_mpd_formats_and_subtitles(
2523 self, mpd_url, video_id, mpd_id=None, note=None, errnote=None,
2524 fatal=True, data=None, headers={}, query={}):
2525 res = self._download_xml_handle(
2527 note='Downloading MPD manifest' if note is None else note,
2528 errnote='Failed to download MPD manifest' if errnote is None else errnote,
2529 fatal=fatal, data=data, headers=headers, query=query)
2535 mpd_base_url = base_url(urlh.geturl())
2537 return self._parse_mpd_formats_and_subtitles(
2538 mpd_doc, mpd_id, mpd_base_url, mpd_url)
2540 def _parse_mpd_formats(self, *args, **kwargs):
2541 fmts, subs = self._parse_mpd_formats_and_subtitles(*args, **kwargs)
2543 self._report_ignoring_subs('DASH')
2546 def _parse_mpd_formats_and_subtitles(
2547 self, mpd_doc, mpd_id=None, mpd_base_url='', mpd_url=None):
2549 Parse formats
from MPD manifest
.
2551 1. MPEG
-DASH Standard
, ISO
/IEC
23009-1:2014(E
),
2552 http
://standards
.iso
.org
/ittf
/PubliclyAvailableStandards
/c065274_ISO_IEC_23009
-1_2014.zip
2553 2. https
://en
.wikipedia
.org
/wiki
/Dynamic_Adaptive_Streaming_over_HTTP
2555 if not self.get_param('dynamic_mpd', True):
2556 if mpd_doc.get('type') == 'dynamic':
2559 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
2562 return self._xpath_ns(path, namespace)
2564 def is_drm_protected(element):
2565 return element.find(_add_ns('ContentProtection')) is not None
2567 def extract_multisegment_info(element, ms_parent_info):
2568 ms_info = ms_parent_info.copy()
2570 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2571 # common attributes and elements. We will only extract relevant
2573 def extract_common(source):
2574 segment_timeline = source.find(_add_ns('SegmentTimeline'))
2575 if segment_timeline is not None:
2576 s_e = segment_timeline.findall(_add_ns('S'))
2578 ms_info['total_number'] = 0
2581 r = int(s.get('r', 0))
2582 ms_info['total_number'] += 1 + r
2583 ms_info['s'].append({
2584 't': int(s.get('t', 0)),
2585 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2586 'd': int(s.attrib['d']),
2589 start_number = source.get('startNumber')
2591 ms_info['start_number'] = int(start_number)
2592 timescale = source.get('timescale')
2594 ms_info['timescale'] = int(timescale)
2595 segment_duration = source.get('duration')
2596 if segment_duration:
2597 ms_info['segment_duration'] = float(segment_duration)
2599 def extract_Initialization(source):
2600 initialization = source.find(_add_ns('Initialization'))
2601 if initialization is not None:
2602 ms_info['initialization_url'] = initialization.attrib['sourceURL']
2604 segment_list = element.find(_add_ns('SegmentList'))
2605 if segment_list is not None:
2606 extract_common(segment_list)
2607 extract_Initialization(segment_list)
2608 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
2610 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
2612 segment_template = element.find(_add_ns('SegmentTemplate'))
2613 if segment_template is not None:
2614 extract_common(segment_template)
2615 media = segment_template.get('media')
2617 ms_info['media'] = media
2618 initialization = segment_template.get('initialization')
2620 ms_info['initialization'] = initialization
2622 extract_Initialization(segment_template)
2625 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
2626 formats, subtitles = [], {}
2627 stream_numbers = {'audio': 0, 'video': 0}
2628 for period in mpd_doc.findall(_add_ns('Period')):
2629 period_duration = parse_duration(period.get('duration')) or mpd_duration
2630 period_ms_info = extract_multisegment_info(period, {
2634 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
2635 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
2636 for representation in adaptation_set.findall(_add_ns('Representation')):
2637 representation_attrib = adaptation_set.attrib.copy()
2638 representation_attrib.update(representation.attrib)
2639 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2640 mime_type = representation_attrib['mimeType']
2641 content_type = representation_attrib.get('contentType', mime_type.split('/')[0])
2643 codecs = representation_attrib.get('codecs', '')
2644 if content_type not in ('video', 'audio', 'text'):
2645 if mime_type == 'image/jpeg':
2646 content_type = mime_type
2647 elif codecs.split('.')[0] == 'stpp':
2648 content_type = 'text'
2649 elif mimetype2ext(mime_type) in ('tt', 'dfxp', 'ttml', 'xml', 'json'):
2650 content_type = 'text'
2652 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
2656 for element in (representation, adaptation_set, period, mpd_doc):
2657 base_url_e = element.find(_add_ns('BaseURL'))
2658 if base_url_e is not None:
2659 base_url = base_url_e.text + base_url
2660 if re.match(r'^https?://', base_url):
2662 if mpd_base_url and base_url.startswith('/'):
2663 base_url = compat_urlparse.urljoin(mpd_base_url, base_url)
2664 elif mpd_base_url and not re.match(r'^https?://', base_url):
2665 if not mpd_base_url.endswith('/'):
2667 base_url = mpd_base_url + base_url
2668 representation_id = representation_attrib.get('id')
2669 lang = representation_attrib.get('lang')
2670 url_el = representation.find(_add_ns('BaseURL'))
2671 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
2672 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
2673 if representation_id is not None:
2674 format_id = representation_id
2676 format_id = content_type
2678 format_id = mpd_id + '-' + format_id
2679 if content_type in ('video', 'audio'):
2681 'format_id': format_id,
2682 'manifest_url': mpd_url,
2683 'ext': mimetype2ext(mime_type),
2684 'width': int_or_none(representation_attrib.get('width')),
2685 'height': int_or_none(representation_attrib.get('height')),
2686 'tbr': float_or_none(bandwidth, 1000),
2687 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
2688 'fps': int_or_none(representation_attrib.get('frameRate')),
2689 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
2690 'format_note': 'DASH %s' % content_type,
2691 'filesize': filesize,
2692 'container': mimetype2ext(mime_type) + '_dash',
2693 'manifest_stream_number': stream_numbers[content_type]
2695 f.update(parse_codecs(codecs))
2696 stream_numbers[content_type] += 1
2697 elif content_type == 'text':
2699 'ext': mimetype2ext(mime_type),
2700 'manifest_url': mpd_url,
2701 'filesize': filesize,
2703 elif content_type == 'image/jpeg':
2704 # See test case in VikiIE
2705 # https://www.viki.com/videos/1175236v-choosing-spouse-by-lottery-episode-1
2707 'format_id': format_id,
2709 'manifest_url': mpd_url,
2710 'format_note': 'DASH storyboards (jpeg)',
2714 if is_drm_protected(adaptation_set) or is_drm_protected(representation):
2716 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
2718 def prepare_template(template_name, identifiers):
2719 tmpl = representation_ms_info[template_name]
2720 # First of, % characters outside $...$ templates
2721 # must be escaped by doubling for proper processing
2722 # by % operator string formatting used further (see
2723 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2729 in_template = not in_template
2730 elif c == '%' and not in_template:
2732 # Next, $...$ templates are translated to their
2733 # %(...) counterparts to be used with % operator
2734 if representation_id is not None:
2735 t = t.replace('$RepresentationID$', representation_id)
2736 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
2737 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
2738 t.replace('$$', '$')
2741 # @initialization is a regular template like @media one
2742 # so it should be handled just the same way (see
2743 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2744 if 'initialization' in representation_ms_info:
2745 initialization_template = prepare_template(
2747 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2748 # $Time$ shall not be included for @initialization thus
2749 # only $Bandwidth$ remains
2751 representation_ms_info['initialization_url'] = initialization_template % {
2752 'Bandwidth': bandwidth,
2755 def location_key(location):
2756 return 'url' if re.match(r'^https?://', location) else 'path'
2758 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
2760 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2761 media_location_key = location_key(media_template)
2763 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2764 # can't be used at the same time
2765 if '%(Number' in media_template and 's' not in representation_ms_info:
2766 segment_duration = None
2767 if 'total_number' not in representation_ms_info and 'segment_duration' in representation_ms_info:
2768 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
2769 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
2770 representation_ms_info['fragments'] = [{
2771 media_location_key: media_template % {
2772 'Number': segment_number,
2773 'Bandwidth': bandwidth,
2775 'duration': segment_duration,
2776 } for segment_number in range(
2777 representation_ms_info['start_number'],
2778 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
2780 # $Number*$ or $Time$ in media template with S list available
2781 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2782 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2783 representation_ms_info['fragments'] = []
2786 segment_number = representation_ms_info['start_number']
2788 def add_segment_url():
2789 segment_url = media_template % {
2790 'Time': segment_time,
2791 'Bandwidth': bandwidth,
2792 'Number': segment_number,
2794 representation_ms_info['fragments'].append({
2795 media_location_key: segment_url,
2796 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
2799 for num, s in enumerate(representation_ms_info['s']):
2800 segment_time = s.get('t') or segment_time
2804 for r in range(s.get('r', 0)):
2805 segment_time += segment_d
2808 segment_time += segment_d
2809 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
2811 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2812 # or any YouTube dashsegments video
2815 timescale = representation_ms_info['timescale']
2816 for s in representation_ms_info['s']:
2817 duration = float_or_none(s['d'], timescale)
2818 for r in range(s.get('r', 0) + 1):
2819 segment_uri = representation_ms_info['segment_urls'][segment_index]
2821 location_key(segment_uri): segment_uri,
2822 'duration': duration,
2825 representation_ms_info['fragments'] = fragments
2826 elif 'segment_urls' in representation_ms_info:
2827 # Segment URLs with no SegmentTimeline
2828 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2829 # https://github.com/ytdl-org/youtube-dl/pull/14844
2831 segment_duration = float_or_none(
2832 representation_ms_info['segment_duration'],
2833 representation_ms_info['timescale']) if 'segment_duration' in representation_ms_info else None
2834 for segment_url in representation_ms_info['segment_urls']:
2836 location_key(segment_url): segment_url,
2838 if segment_duration:
2839 fragment['duration'] = segment_duration
2840 fragments.append(fragment)
2841 representation_ms_info['fragments'] = fragments
2842 # If there is a fragments key available then we correctly recognized fragmented media.
2843 # Otherwise we will assume unfragmented media with direct access. Technically, such
2844 # assumption is not necessarily correct since we may simply have no support for
2845 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2846 if 'fragments' in representation_ms_info:
2848 # NB: mpd_url may be empty when MPD manifest is parsed from a string
2849 'url': mpd_url or base_url,
2850 'fragment_base_url': base_url,
2852 'protocol': 'http_dash_segments' if mime_type != 'image/jpeg' else 'mhtml',
2854 if 'initialization_url' in representation_ms_info:
2855 initialization_url = representation_ms_info['initialization_url']
2856 if not f.get('url'):
2857 f['url'] = initialization_url
2858 f['fragments'].append({location_key(initialization_url): initialization_url})
2859 f['fragments'].extend(representation_ms_info['fragments'])
2861 # Assuming direct URL to unfragmented media.
2863 if content_type in ('video', 'audio') or mime_type == 'image/jpeg':
2865 elif content_type == 'text':
2866 subtitles.setdefault(lang or 'und', []).append(f)
2868 return formats, subtitles
2870 def _extract_ism_formats(self, *args, **kwargs):
2871 fmts, subs = self._extract_ism_formats_and_subtitles(*args, **kwargs)
2873 self._report_ignoring_subs('ISM')
2876 def _extract_ism_formats_and_subtitles(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
2877 res = self._download_xml_handle(
2879 note='Downloading ISM manifest' if note is None else note,
2880 errnote='Failed to download ISM manifest' if errnote is None else errnote,
2881 fatal=fatal, data=data, headers=headers, query=query)
2888 return self._parse_ism_formats_and_subtitles(ism_doc, urlh.geturl(), ism_id)
2890 def _parse_ism_formats_and_subtitles(self, ism_doc, ism_url, ism_id=None):
2892 Parse formats
from ISM manifest
.
2894 1. [MS
-SSTR
]: Smooth Streaming Protocol
,
2895 https
://msdn
.microsoft
.com
/en
-us
/library
/ff469518
.aspx
2897 if ism_doc.get('IsLive') == 'TRUE':
2900 duration = int(ism_doc.attrib['Duration'])
2901 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
2905 for stream in ism_doc.findall('StreamIndex'):
2906 stream_type = stream.get('Type')
2907 if stream_type not in ('video', 'audio', 'text'):
2909 url_pattern = stream.attrib['Url']
2910 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
2911 stream_name = stream.get('Name')
2912 stream_language = stream.get('Language', 'und')
2913 for track in stream.findall('QualityLevel'):
2914 fourcc = track.get('FourCC') or ('AACL' if track.get('AudioTag') == '255' else None)
2915 # TODO: add support for WVC1 and WMAP
2916 if fourcc not in ('H264', 'AVC1', 'AACL', 'TTML'):
2917 self.report_warning('%s is not a supported codec' % fourcc)
2919 tbr = int(track.attrib['Bitrate']) // 1000
2920 # [1] does not mention Width and Height attributes. However,
2921 # they're often present while MaxWidth and MaxHeight are
2922 # missing, so should be used as fallbacks
2923 width = int_or_none(track.get('MaxWidth') or track.get('Width'))
2924 height = int_or_none(track.get('MaxHeight') or track.get('Height'))
2925 sampling_rate = int_or_none(track.get('SamplingRate'))
2927 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
2928 track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
2934 stream_fragments = stream.findall('c')
2935 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
2936 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
2937 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
2938 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
2939 if not fragment_ctx['duration']:
2941 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
2943 next_fragment_time = duration
2944 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
2945 for _ in range(fragment_repeat):
2947 'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
2948 'duration': fragment_ctx['duration'] / stream_timescale,
2950 fragment_ctx['time'] += fragment_ctx['duration']
2954 format_id.append(ism_id)
2956 format_id.append(stream_name)
2957 format_id.append(compat_str(tbr))
2959 if stream_type == 'text':
2960 subtitles.setdefault(stream_language, []).append({
2964 'manifest_url': ism_url,
2965 'fragments': fragments,
2966 '_download_params': {
2967 'stream_type': stream_type,
2968 'duration': duration,
2969 'timescale': stream_timescale,
2971 'language': stream_language,
2972 'codec_private_data': track.get('CodecPrivateData'),
2975 elif stream_type in ('video', 'audio'):
2977 'format_id': '-'.join(format_id),
2979 'manifest_url': ism_url,
2980 'ext': 'ismv' if stream_type == 'video' else 'isma',
2984 'asr': sampling_rate,
2985 'vcodec': 'none' if stream_type == 'audio' else fourcc,
2986 'acodec': 'none' if stream_type == 'video' else fourcc,
2988 'fragments': fragments,
2989 'has_drm': ism_doc.find('Protection') is not None,
2990 '_download_params': {
2991 'stream_type': stream_type,
2992 'duration': duration,
2993 'timescale': stream_timescale,
2994 'width': width or 0,
2995 'height': height or 0,
2997 'language': stream_language,
2998 'codec_private_data': track.get('CodecPrivateData'),
2999 'sampling_rate': sampling_rate,
3000 'channels': int_or_none(track.get('Channels', 2)),
3001 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
3002 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
3005 return formats, subtitles
3007 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):
3008 def absolute_url(item_url):
3009 return urljoin(base_url, item_url)
3011 def parse_content_type(content_type):
3012 if not content_type:
3014 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
3016 mimetype, codecs = ctr.groups()
3017 f = parse_codecs(codecs)
3018 f['ext'] = mimetype2ext(mimetype)
3022 def _media_formats(src, cur_media_type, type_info={}):
3023 full_url = absolute_url(src)
3024 ext = type_info.get('ext') or determine_ext(full_url)
3026 is_plain_url = False
3027 formats = self._extract_m3u8_formats(
3028 full_url, video_id, ext='mp4',
3029 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
3030 preference=preference, quality=quality, fatal=False)
3032 is_plain_url = False
3033 formats = self._extract_mpd_formats(
3034 full_url, video_id, mpd_id=mpd_id, fatal=False)
3039 'vcodec': 'none' if cur_media_type == 'audio' else None,
3041 return is_plain_url, formats
3044 # amp-video and amp-audio are very similar to their HTML5 counterparts
3045 # so we wll include them right here (see
3046 # https://www.ampproject.org/docs/reference/components/amp-video)
3047 # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/
3048 _MEDIA_TAG_NAME_RE = r'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)'
3049 media_tags = [(media_tag, media_tag_name, media_type, '')
3050 for media_tag, media_tag_name, media_type
3051 in re.findall(r'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE, webpage)]
3052 media_tags.extend(re.findall(
3053 # We only allow video|audio followed by a whitespace or '>'.
3054 # Allowing more characters may end up in significant slow down (see
3055 # https://github.com/ytdl-org/youtube-dl/issues/11979, example URL:
3056 # http://www.porntrex.com/maps/videositemap.xml).
3057 r'(?s)(<(?P<tag>%s)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>' % _MEDIA_TAG_NAME_RE, webpage))
3058 for media_tag, _, media_type, media_content in media_tags:
3063 media_attributes = extract_attributes(media_tag)
3064 src = strip_or_none(media_attributes.get('src'))
3066 _, formats = _media_formats(src, media_type)
3067 media_info['formats'].extend(formats)
3068 media_info['thumbnail'] = absolute_url(media_attributes.get('poster'))
3070 for source_tag in re.findall(r'<source[^>]+>', media_content):
3071 s_attr = extract_attributes(source_tag)
3072 # data-video-src and data-src are non standard but seen
3073 # several times in the wild
3074 src = strip_or_none(dict_get(s_attr, ('src', 'data-video-src', 'data-src')))
3077 f = parse_content_type(s_attr.get('type'))
3078 is_plain_url, formats = _media_formats(src, media_type, f)
3080 # width, height, res, label and title attributes are
3081 # all not standard but seen several times in the wild
3084 for lbl in ('label', 'title')
3085 if str_or_none(s_attr.get(lbl))
3087 width = int_or_none(s_attr.get('width'))
3088 height = (int_or_none(s_attr.get('height'))
3089 or int_or_none(s_attr.get('res')))
3090 if not width or not height:
3092 resolution = parse_resolution(lbl)
3095 width = width or resolution.get('width')
3096 height = height or resolution.get('height')
3098 tbr = parse_bitrate(lbl)
3107 'format_id': s_attr.get('label') or s_attr.get('title'),
3109 f.update(formats[0])
3110 media_info['formats'].append(f)
3112 media_info['formats'].extend(formats)
3113 for track_tag in re.findall(r'<track[^>]+>', media_content):
3114 track_attributes = extract_attributes(track_tag)
3115 kind = track_attributes.get('kind')
3116 if not kind or kind in ('subtitles', 'captions'):
3117 src = strip_or_none(track_attributes.get('src'))
3120 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
3121 media_info['subtitles'].setdefault(lang, []).append({
3122 'url': absolute_url(src),
3124 for f in media_info['formats']:
3125 f.setdefault('http_headers', {})['Referer'] = base_url
3126 if media_info['formats'] or media_info['subtitles']:
3127 entries.append(media_info)
3130 def _extract_akamai_formats(self, *args, **kwargs):
3131 fmts, subs = self._extract_akamai_formats_and_subtitles(*args, **kwargs)
3133 self._report_ignoring_subs('akamai')
3136 def _extract_akamai_formats_and_subtitles(self, manifest_url, video_id, hosts={}):
3137 signed = 'hdnea=' in manifest_url
3139 # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
3140 manifest_url = re.sub(
3141 r'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
3142 '', manifest_url).strip('?')
3147 hdcore_sign = 'hdcore=3.7.0'
3148 f4m_url = re.sub(r'(https?://[^/]+)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
3149 hds_host = hosts.get('hds')
3151 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
3152 if 'hdcore=' not in f4m_url:
3153 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
3154 f4m_formats = self._extract_f4m_formats(
3155 f4m_url, video_id, f4m_id='hds', fatal=False)
3156 for entry in f4m_formats:
3157 entry.update({'extra_param_to_segment_url': hdcore_sign})
3158 formats.extend(f4m_formats)
3160 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
3161 hls_host = hosts.get('hls')
3163 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
3164 m3u8_formats, m3u8_subtitles = self._extract_m3u8_formats_and_subtitles(
3165 m3u8_url, video_id, 'mp4', 'm3u8_native',
3166 m3u8_id='hls', fatal=False)
3167 formats.extend(m3u8_formats)
3168 subtitles = self._merge_subtitles(subtitles, m3u8_subtitles)
3170 http_host = hosts.get('http')
3171 if http_host and m3u8_formats and not signed:
3172 REPL_REGEX = r'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
3173 qualities = re.match(REPL_REGEX, m3u8_url).group(2).split(',')
3174 qualities_length = len(qualities)
3175 if len(m3u8_formats) in (qualities_length, qualities_length + 1):
3177 for f in m3u8_formats:
3178 if f['vcodec'] != 'none':
3179 for protocol in ('http', 'https'):
3181 del http_f['manifest_url']
3183 REPL_REGEX, protocol + r'://%s/\g<1>%s\3' % (http_host, qualities[i]), f['url'])
3185 'format_id': http_f['format_id'].replace('hls-', protocol + '-'),
3187 'protocol': protocol,
3189 formats.append(http_f)
3192 return formats, subtitles
3194 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
3195 query = compat_urlparse.urlparse(url).query
3196 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
3198 r'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url)
3199 url_base = mobj.group('url')
3200 http_base_url = '%s%s:%s' % ('http', mobj.group('s') or '', url_base)
3203 def manifest_url(manifest):
3204 m_url = '%s/%s' % (http_base_url, manifest)
3206 m_url += '?%s' % query
3209 if 'm3u8' not in skip_protocols:
3210 formats.extend(self._extract_m3u8_formats(
3211 manifest_url('playlist.m3u8'), video_id, 'mp4',
3212 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
3213 if 'f4m' not in skip_protocols:
3214 formats.extend(self._extract_f4m_formats(
3215 manifest_url('manifest.f4m'),
3216 video_id, f4m_id='hds', fatal=False))
3217 if 'dash' not in skip_protocols:
3218 formats.extend(self._extract_mpd_formats(
3219 manifest_url('manifest.mpd'),
3220 video_id, mpd_id='dash', fatal=False))
3221 if re.search(r'(?:/smil:|\.smil)', url_base):
3222 if 'smil' not in skip_protocols:
3223 rtmp_formats = self._extract_smil_formats(
3224 manifest_url('jwplayer.smil'),
3225 video_id, fatal=False)
3226 for rtmp_format in rtmp_formats:
3227 rtsp_format = rtmp_format.copy()
3228 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
3229 del rtsp_format['play_path']
3230 del rtsp_format['ext']
3231 rtsp_format.update({
3232 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
3233 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
3236 formats.extend([rtmp_format, rtsp_format])
3238 for protocol in ('rtmp', 'rtsp'):
3239 if protocol not in skip_protocols:
3241 'url': '%s:%s' % (protocol, url_base),
3242 'format_id': protocol,
3243 'protocol': protocol,
3247 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
3249 r'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)(?!</script>).*?\.setup\s*\((?P<options>[^)]+)\)',
3253 jwplayer_data = self._parse_json(mobj.group('options'),
3255 transform_source=transform_source)
3256 except ExtractorError:
3259 if isinstance(jwplayer_data, dict):
3260 return jwplayer_data
3262 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
3263 jwplayer_data = self._find_jwplayer_data(
3264 webpage, video_id, transform_source=js_to_json)
3265 return self._parse_jwplayer_data(
3266 jwplayer_data, video_id, *args, **kwargs)
3268 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
3269 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3270 # JWPlayer backward compatibility: flattened playlists
3271 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
3272 if 'playlist' not in jwplayer_data:
3273 jwplayer_data = {'playlist': [jwplayer_data]}
3277 # JWPlayer backward compatibility: single playlist item
3278 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
3279 if not isinstance(jwplayer_data['playlist'], list):
3280 jwplayer_data['playlist'] = [jwplayer_data['playlist']]
3282 for video_data in jwplayer_data['playlist']:
3283 # JWPlayer backward compatibility: flattened sources
3284 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
3285 if 'sources' not in video_data:
3286 video_data['sources'] = [video_data]
3288 this_video_id = video_id or video_data['mediaid']
3290 formats = self._parse_jwplayer_formats(
3291 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
3292 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
3295 tracks = video_data.get('tracks')
3296 if tracks and isinstance(tracks, list):
3297 for track in tracks:
3298 if not isinstance(track, dict):
3300 track_kind = track.get('kind')
3301 if not track_kind or not isinstance(track_kind, compat_str):
3303 if track_kind.lower() not in ('captions', 'subtitles'):
3305 track_url = urljoin(base_url, track.get('file'))
3308 subtitles.setdefault(track.get('label') or 'en', []).append({
3309 'url': self._proto_relative_url(track_url)
3313 'id': this_video_id,
3314 'title': unescapeHTML(video_data['title'] if require_title else video_data.get('title')),
3315 'description': clean_html(video_data.get('description')),
3316 'thumbnail': urljoin(base_url, self._proto_relative_url(video_data.get('image'))),
3317 'timestamp': int_or_none(video_data.get('pubdate')),
3318 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
3319 'subtitles': subtitles,
3321 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
3322 if len(formats) == 1 and re.search(r'^(?:http|//).*(?:youtube\.com|youtu\.be)/.+', formats[0]['url']):
3324 '_type': 'url_transparent',
3325 'url': formats[0]['url'],
3328 self._sort_formats(formats)
3329 entry['formats'] = formats
3330 entries.append(entry)
3331 if len(entries) == 1:
3334 return self.playlist_result(entries)
3336 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
3337 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3340 for source in jwplayer_sources_data:
3341 if not isinstance(source, dict):
3343 source_url = urljoin(
3344 base_url, self._proto_relative_url(source.get('file')))
3345 if not source_url or source_url in urls:
3347 urls.append(source_url)
3348 source_type = source.get('type') or ''
3349 ext = mimetype2ext(source_type) or determine_ext(source_url)
3350 if source_type == 'hls' or ext == 'm3u8':
3351 formats.extend(self._extract_m3u8_formats(
3352 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
3353 m3u8_id=m3u8_id, fatal=False))
3354 elif source_type == 'dash' or ext == 'mpd':
3355 formats.extend(self._extract_mpd_formats(
3356 source_url, video_id, mpd_id=mpd_id, fatal=False))
3358 formats.extend(self._extract_smil_formats(
3359 source_url, video_id, fatal=False))
3360 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3361 elif source_type.startswith('audio') or ext in (
3362 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
3369 height = int_or_none(source.get('height'))
3371 # Often no height is provided but there is a label in
3372 # format like "1080p", "720p SD", or 1080.
3373 height = int_or_none(self._search_regex(
3374 r'^(\d{3,4})[pP]?(?:\b|$)', compat_str(source.get('label') or ''),
3375 'height', default=None))
3378 'width': int_or_none(source.get('width')),
3380 'tbr': int_or_none(source.get('bitrate')),
3383 if source_url.startswith('rtmp'):
3384 a_format['ext'] = 'flv'
3385 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3386 # of jwplayer.flash.swf
3387 rtmp_url_parts = re.split(
3388 r'((?:mp4|mp3|flv):)', source_url, 1)
3389 if len(rtmp_url_parts) == 3:
3390 rtmp_url, prefix, play_path = rtmp_url_parts
3393 'play_path': prefix + play_path,
3396 a_format.update(rtmp_params)
3397 formats.append(a_format)
3400 def _live_title(self, name):
3401 """ Generate the title
for a live video
"""
3402 now = datetime.datetime.now()
3403 now_str = now.strftime('%Y-%m-%d %H:%M')
3404 return name + ' ' + now_str
3406 def _int(self, v, name, fatal=False, **kwargs):
3407 res = int_or_none(v, **kwargs)
3408 if 'get_attr' in kwargs:
3409 print(getattr(v, kwargs['get_attr']))
3411 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
3413 raise ExtractorError(msg)
3415 self.report_warning(msg)
3418 def _float(self, v, name, fatal=False, **kwargs):
3419 res = float_or_none(v, **kwargs)
3421 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
3423 raise ExtractorError(msg)
3425 self.report_warning(msg)
3428 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
3429 path='/', secure=False, discard=False, rest={}, **kwargs):
3430 cookie = compat_cookiejar_Cookie(
3431 0, name, value, port, port is not None, domain, True,
3432 domain.startswith('.'), path, True, secure, expire_time,
3433 discard, None, None, rest)
3434 self._downloader.cookiejar.set_cookie(cookie)
3436 def _get_cookies(self, url):
3437 """ Return a compat_cookies_SimpleCookie
with the cookies
for the url
"""
3438 req = sanitized_Request(url)
3439 self._downloader.cookiejar.add_cookie_header(req)
3440 return compat_cookies_SimpleCookie(req.get_header('Cookie'))
3442 def _apply_first_set_cookie_header(self, url_handle, cookie):
3444 Apply first Set
-Cookie header instead of the last
. Experimental
.
3446 Some
sites (e
.g
. [1-3]) may serve two cookies under the same name
3447 in Set
-Cookie header
and expect the
first (old
) one to be
set rather
3448 than
second (new
). However
, as of RFC6265 the newer one cookie
3449 should be
set into cookie store what actually happens
.
3450 We will workaround this issue by resetting the cookie to
3451 the first one manually
.
3452 1. https
://new
.vk
.com
/
3453 2. https
://github
.com
/ytdl
-org
/youtube
-dl
/issues
/9841#issuecomment-227871201
3454 3. https
://learning
.oreilly
.com
/
3456 for header, cookies in url_handle.headers.items():
3457 if header.lower() != 'set-cookie':
3459 if sys.version_info[0] >= 3:
3460 cookies = cookies.encode('iso-8859-1')
3461 cookies = cookies.decode('utf-8')
3462 cookie_value = re.search(
3463 r'%s=(.+?);.*?\b[Dd]omain=(.+?)(?:[,;]|$)' % cookie, cookies)
3465 value, domain = cookie_value.groups()
3466 self._set_cookie(domain, cookie, value)
3469 def get_testcases(self, include_onlymatching=False):
3470 t = getattr(self, '_TEST', None)
3472 assert not hasattr(self, '_TESTS'), \
3473 '%s has _TEST and _TESTS' % type(self).__name__
3476 tests = getattr(self, '_TESTS', [])
3478 if not include_onlymatching and t.get('only_matching', False):
3480 t['name'] = type(self).__name__[:-len('IE')]
3483 def is_suitable(self, age_limit):
3484 """ Test whether the extractor
is generally suitable
for the given
3485 age
limit (i
.e
. pornographic sites are
not, all others usually are
) """
3487 any_restricted = False
3488 for tc in self.get_testcases(include_onlymatching=False):
3489 if tc.get('playlist', []):
3490 tc = tc['playlist'][0]
3491 is_restricted = age_restricted(
3492 tc.get('info_dict', {}).get('age_limit'), age_limit)
3493 if not is_restricted:
3495 any_restricted = any_restricted or is_restricted
3496 return not any_restricted
3498 def extract_subtitles(self, *args, **kwargs):
3499 if (self.get_param('writesubtitles', False)
3500 or self.get_param('listsubtitles')):
3501 return self._get_subtitles(*args, **kwargs)
3504 def _get_subtitles(self, *args, **kwargs):
3505 raise NotImplementedError('This method must be implemented by subclasses')
3507 def extract_comments(self, *args, **kwargs):
3508 if not self.get_param('getcomments'):
3510 generator = self._get_comments(*args, **kwargs)
3516 comments.append(next(generator))
3517 except KeyboardInterrupt:
3519 self.to_screen('Interrupted by user')
3520 except StopIteration:
3522 comment_count = len(comments)
3523 self.to_screen(f'Extracted {comment_count} comments')
3525 'comments': comments,
3526 'comment_count': None if interrupted else comment_count
3530 def _get_comments(self, *args, **kwargs):
3531 raise NotImplementedError('This method must be implemented by subclasses')
3534 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
3535 """ Merge subtitle items
for one language
. Items
with duplicated URLs
3536 will be dropped
. """
3537 list1_urls = set([item['url'] for item in subtitle_list1])
3538 ret = list(subtitle_list1)
3539 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
3543 def _merge_subtitles(cls, *dicts, target=None):
3544 """ Merge subtitle dictionaries
, language by language
. """
3548 for lang, subs in d.items():
3549 target[lang] = cls._merge_subtitle_items(target.get(lang, []), subs)
3552 def extract_automatic_captions(self, *args, **kwargs):
3553 if (self.get_param('writeautomaticsub', False)
3554 or self.get_param('listsubtitles')):
3555 return self._get_automatic_captions(*args, **kwargs)
3558 def _get_automatic_captions(self, *args, **kwargs):
3559 raise NotImplementedError('This method must be implemented by subclasses')
3561 def mark_watched(self, *args, **kwargs):
3562 if not self.get_param('mark_watched', False):
3564 if (self._get_login_info()[0] is not None
3565 or self.get_param('cookiefile')
3566 or self.get_param('cookiesfrombrowser')):
3567 self._mark_watched(*args, **kwargs)
3569 def _mark_watched(self, *args, **kwargs):
3570 raise NotImplementedError('This method must be implemented by subclasses')
3572 def geo_verification_headers(self):
3574 geo_verification_proxy = self.get_param('geo_verification_proxy')
3575 if geo_verification_proxy:
3576 headers['Ytdl-request-proxy'] = geo_verification_proxy
3579 def _generic_id(self, url):
3580 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
3582 def _generic_title(self, url):
3583 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
3586 def _availability(is_private=None, needs_premium=None, needs_subscription=None, needs_auth=None, is_unlisted=None):
3587 all_known = all(map(
3588 lambda x: x is not None,
3589 (is_private, needs_premium, needs_subscription, needs_auth, is_unlisted)))
3591 'private' if is_private
3592 else 'premium_only' if needs_premium
3593 else 'subscriber_only' if needs_subscription
3594 else 'needs_auth' if needs_auth
3595 else 'unlisted' if is_unlisted
3596 else 'public' if all_known
3599 def _configuration_arg(self, key, default=NO_DEFAULT, casesense=False):
3601 @returns A list of values for the extractor argument given by "key"
3602 or "default" if no such key is present
3603 @param default The default value to return when the key is not present (default: [])
3604 @param casesense When false, the values are converted to lower case
3607 self._downloader.params, ('extractor_args', self.ie_key().lower(), key))
3609 return [] if default is NO_DEFAULT else default
3610 return list(val) if casesense else [x.lower() for x in val]
3613 class SearchInfoExtractor(InfoExtractor):
3615 Base
class for paged search queries extractors
.
3616 They accept URLs
in the format
_SEARCH_KEY(|all|
[0-9]):{query}
3617 Instances should define _SEARCH_KEY
and _MAX_RESULTS
.
3621 def _make_valid_url(cls):
3622 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
3625 def suitable(cls, url):
3626 return re.match(cls._make_valid_url(), url) is not None
3628 def _real_extract(self, query):
3629 mobj = re.match(self._make_valid_url(), query)
3631 raise ExtractorError('Invalid search query "%s"' % query)
3633 prefix = mobj.group('prefix')
3634 query = mobj.group('query')
3636 return self._get_n_results(query, 1)
3637 elif prefix == 'all':
3638 return self._get_n_results(query, self._MAX_RESULTS)
3642 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
3643 elif n > self._MAX_RESULTS:
3644 self.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3645 n = self._MAX_RESULTS
3646 return self._get_n_results(query, n)
3648 def _get_n_results(self, query, n):
3649 """Get a specified number of results
for a query
.
3650 Either this function
or _search_results must be overridden by subclasses
"""
3651 return self.playlist_result(
3652 itertools.islice(self._search_results(query), 0, None if n == float('inf') else n),
3655 def _search_results(self, query):
3656 """Returns an iterator of search results
"""
3657 raise NotImplementedError('This method must be implemented by subclasses')
3660 def SEARCH_KEY(self):
3661 return self._SEARCH_KEY