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
,
92 class InfoExtractor(object):
93 """Information Extractor class.
95 Information extractors are the classes that, given a URL, extract
96 information about the video (or videos) the URL refers to. This
97 information includes the real video URL, the video title, author and
98 others. The information is stored in a dictionary which is then
99 passed to the YoutubeDL. The YoutubeDL processes this
100 information possibly downloading the video to the file system, among
101 other possible outcomes.
103 The type field determines the type of the result.
104 By far the most common value (and the default if _type is missing) is
105 "video", which indicates a single video.
107 For a video, the dictionaries must include the following fields:
109 id: Video identifier.
110 title: Video title, unescaped.
112 Additionally, it must contain either a formats entry or a url one:
114 formats: A list of dictionaries for each format available, ordered
115 from worst to best quality.
118 * url The mandatory URL representing the media:
119 for plain file media - HTTP URL of this file,
121 for HLS - URL of the M3U8 media playlist,
122 for HDS - URL of the F4M manifest,
124 - HTTP URL to plain file media (in case of
126 - URL of the MPD manifest or base URL
127 representing the media if MPD manifest
128 is parsed from a string (in case of
130 for MSS - URL of the ISM manifest.
132 The URL of the manifest file in case of
134 for HLS - URL of the M3U8 master playlist,
135 for HDS - URL of the F4M manifest,
136 for DASH - URL of the MPD manifest,
137 for MSS - URL of the ISM manifest.
138 * ext Will be calculated from URL if missing
139 * format A human-readable description of the format
140 ("mp4 container with h264/opus").
141 Calculated from the format_id, width, height.
142 and format_note fields if missing.
143 * format_id A short description of the format
144 ("mp4_h264_opus" or "19").
145 Technically optional, but strongly recommended.
146 * format_note Additional info about the format
147 ("3D" or "DASH video")
148 * width Width of the video, if known
149 * height Height of the video, if known
150 * resolution Textual description of width and height
151 * dynamic_range The dynamic range of the video. One of:
152 "SDR" (None), "HDR10", "HDR10+, "HDR12", "HLG, "DV"
153 * tbr Average bitrate of audio and video in KBit/s
154 * abr Average audio bitrate in KBit/s
155 * acodec Name of the audio codec in use
156 * asr Audio sampling rate in Hertz
157 * vbr Average video bitrate in KBit/s
159 * vcodec Name of the video codec in use
160 * container Name of the container format
161 * filesize The number of bytes, if known in advance
162 * filesize_approx An estimate for the number of bytes
163 * player_url SWF Player URL (used for rtmpdump).
164 * protocol The protocol that will be used for the actual
165 download, lower-case.
166 "http", "https", "rtsp", "rtmp", "rtmp_ffmpeg", "rtmpe",
167 "m3u8", "m3u8_native" or "http_dash_segments".
169 Base URL for fragments. Each fragment's path
170 value (if present) will be relative to
172 * fragments A list of fragments of a fragmented media.
173 Each fragment entry must contain either an url
174 or a path. If an url is present it should be
175 considered by a client. Otherwise both path and
176 fragment_base_url must be present. Here is
177 the list of all potential fields:
178 * "url" - fragment's URL
179 * "path" - fragment's path relative to
181 * "duration" (optional, int or float)
182 * "filesize" (optional, int)
183 * preference Order number of this format. If this field is
184 present and not None, the formats get sorted
185 by this field, regardless of all other values.
186 -1 for default (order by other properties),
187 -2 or smaller for less than default.
188 < -1000 to hide the format (if there is
189 another one which is strictly better)
190 * language Language code, e.g. "de" or "en-US".
191 * language_preference Is this in the language mentioned in
193 10 if it's what the URL is about,
194 -1 for default (don't know),
195 -10 otherwise, other values reserved for now.
196 * quality Order number of the video quality of this
197 format, irrespective of the file format.
198 -1 for default (order by other properties),
199 -2 or smaller for less than default.
200 * source_preference Order number for this video source
201 (quality takes higher priority)
202 -1 for default (order by other properties),
203 -2 or smaller for less than default.
204 * http_headers A dictionary of additional HTTP headers
205 to add to the request.
206 * stretched_ratio If given and not 1, indicates that the
207 video's pixels are not square.
208 width : height ratio as float.
209 * no_resume The server does not support resuming the
210 (HTTP or RTMP) download. Boolean.
211 * has_drm The format has DRM and cannot be downloaded. Boolean
212 * downloader_options A dictionary of downloader options as
213 described in FileDownloader
214 RTMP formats can also have the additional fields: page_url,
215 app, play_path, tc_url, flash_version, rtmp_live, rtmp_conn,
216 rtmp_protocol, rtmp_real_time
218 url: Final video URL.
219 ext: Video filename extension.
220 format: The video format, defaults to ext (used for --get-format)
221 player_url: SWF Player URL (used for rtmpdump).
223 The following fields are optional:
225 alt_title: A secondary title of the video.
226 display_id An alternative identifier for the video, not necessarily
227 unique, but available before title. Typically, id is
228 something like "4234987", title "Dancing naked mole rats",
229 and display_id "dancing-naked-mole-rats"
230 thumbnails: A list of dictionaries, with the following entries:
231 * "id" (optional, string) - Thumbnail format ID
233 * "preference" (optional, int) - quality of the image
234 * "width" (optional, int)
235 * "height" (optional, int)
236 * "resolution" (optional, string "{width}x{height}",
238 * "filesize" (optional, int)
239 thumbnail: Full URL to a video thumbnail image.
240 description: Full video description.
241 uploader: Full name of the video uploader.
242 license: License name the video is licensed under.
243 creator: The creator of the video.
244 release_timestamp: UNIX timestamp of the moment the video was released.
245 release_date: The date (YYYYMMDD) when the video was released.
246 timestamp: UNIX timestamp of the moment the video was uploaded
247 upload_date: Video upload date (YYYYMMDD).
248 If not explicitly set, calculated from timestamp.
249 uploader_id: Nickname or id of the video uploader.
250 uploader_url: Full URL to a personal webpage of the video uploader.
251 channel: Full name of the channel the video is uploaded on.
252 Note that channel fields may or may not repeat uploader
253 fields. This depends on a particular extractor.
254 channel_id: Id of the channel.
255 channel_url: Full URL to a channel webpage.
256 location: Physical location where the video was filmed.
257 subtitles: The available subtitles as a dictionary in the format
258 {tag: subformats}. "tag" is usually a language code, and
259 "subformats" is a list sorted from lower to higher
260 preference, each element is a dictionary with the "ext"
262 * "data": The subtitles file contents
263 * "url": A URL pointing to the subtitles file
264 It can optionally also have:
265 * "name": Name or description of the subtitles
266 "ext" will be calculated from URL if missing
267 automatic_captions: Like 'subtitles'; contains automatically generated
268 captions instead of normal subtitles
269 duration: Length of the video in seconds, as an integer or float.
270 view_count: How many users have watched the video on the platform.
271 like_count: Number of positive ratings of the video
272 dislike_count: Number of negative ratings of the video
273 repost_count: Number of reposts of the video
274 average_rating: Average rating give by users, the scale used depends on the webpage
275 comment_count: Number of comments on the video
276 comments: A list of comments, each with one or more of the following
277 properties (all but one of text or html optional):
278 * "author" - human-readable name of the comment author
279 * "author_id" - user ID of the comment author
280 * "author_thumbnail" - The thumbnail of the comment author
282 * "html" - Comment as HTML
283 * "text" - Plain text of the comment
284 * "timestamp" - UNIX timestamp of comment
285 * "parent" - ID of the comment this one is replying to.
286 Set to "root" to indicate that this is a
287 comment to the original video.
288 * "like_count" - Number of positive ratings of the comment
289 * "dislike_count" - Number of negative ratings of the comment
290 * "is_favorited" - Whether the comment is marked as
291 favorite by the video uploader
292 * "author_is_uploader" - Whether the comment is made by
294 age_limit: Age restriction for the video, as an integer (years)
295 webpage_url: The URL to the video webpage, if given to yt-dlp it
296 should allow to get the same result again. (It will be set
297 by YoutubeDL if it's missing)
298 categories: A list of categories that the video falls in, for example
300 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
301 cast: A list of the video cast
302 is_live: True, False, or None (=unknown). Whether this video is a
303 live stream that goes on instead of a fixed-length video.
304 was_live: True, False, or None (=unknown). Whether this video was
305 originally a live stream.
306 live_status: 'is_live', 'is_upcoming', 'was_live', 'not_live' or None (=unknown)
307 If absent, automatically set from is_live, was_live
308 start_time: Time in seconds where the reproduction should start, as
309 specified in the URL.
310 end_time: Time in seconds where the reproduction should end, as
311 specified in the URL.
312 chapters: A list of dictionaries, with the following entries:
313 * "start_time" - The start time of the chapter in seconds
314 * "end_time" - The end time of the chapter in seconds
315 * "title" (optional, string)
316 playable_in_embed: Whether this video is allowed to play in embedded
317 players on other sites. Can be True (=always allowed),
318 False (=never allowed), None (=unknown), or a string
319 specifying the criteria for embedability (Eg: 'whitelist')
320 availability: Under what condition the video is available. One of
321 'private', 'premium_only', 'subscriber_only', 'needs_auth',
322 'unlisted' or 'public'. Use 'InfoExtractor._availability'
324 __post_extractor: A function to be called just before the metadata is
325 written to either disk, logger or console. The function
326 must return a dict which will be added to the info_dict.
327 This is usefull for additional information that is
328 time-consuming to extract. Note that the fields thus
329 extracted will not be available to output template and
330 match_filter. So, only "comments" and "comment_count" are
331 currently allowed to be extracted via this method.
333 The following fields should only be used when the video belongs to some logical
336 chapter: Name or title of the chapter the video belongs to.
337 chapter_number: Number of the chapter the video belongs to, as an integer.
338 chapter_id: Id of the chapter the video belongs to, as a unicode string.
340 The following fields should only be used when the video is an episode of some
341 series, programme or podcast:
343 series: Title of the series or programme the video episode belongs to.
344 season: Title of the season the video episode belongs to.
345 season_number: Number of the season the video episode belongs to, as an integer.
346 season_id: Id of the season the video episode belongs to, as a unicode string.
347 episode: Title of the video episode. Unlike mandatory video title field,
348 this field should denote the exact title of the video episode
349 without any kind of decoration.
350 episode_number: Number of the video episode within a season, as an integer.
351 episode_id: Id of the video episode, as a unicode string.
353 The following fields should only be used when the media is a track or a part of
356 track: Title of the track.
357 track_number: Number of the track within an album or a disc, as an integer.
358 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
360 artist: Artist(s) of the track.
361 genre: Genre(s) of the track.
362 album: Title of the album the track belongs to.
363 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
364 album_artist: List of all artists appeared on the album (e.g.
365 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
367 disc_number: Number of the disc or other physical medium the track belongs to,
369 release_year: Year (YYYY) when the album was released.
371 Unless mentioned otherwise, the fields should be Unicode strings.
373 Unless mentioned otherwise, None is equivalent to absence of information.
376 _type "playlist" indicates multiple videos.
377 There must be a key "entries", which is a list, an iterable, or a PagedList
378 object, each element of which is a valid dictionary by this specification.
380 Additionally, playlists can have "id", "title", and any other relevent
381 attributes with the same semantics as videos (see above).
384 _type "multi_video" indicates that there are multiple videos that
385 form a single show, for examples multiple acts of an opera or TV episode.
386 It must have an entries key like a playlist and contain all the keys
387 required for a video at the same time.
390 _type "url" indicates that the video must be extracted from another
391 location, possibly by a different extractor. Its only required key is:
392 "url" - the next URL to extract.
393 The key "ie_key" can be set to the class name (minus the trailing "IE",
394 e.g. "Youtube") if the extractor class is known in advance.
395 Additionally, the dictionary may have any properties of the resolved entity
396 known in advance, for example "title" if the title of the referred video is
400 _type "url_transparent" entities have the same specification as "url", but
401 indicate that the given additional information is more precise than the one
402 associated with the resolved URL.
403 This is useful when a site employs a video service that hosts the video and
404 its technical metadata, but that video service does not embed a useful
405 title, description etc.
408 Subclasses of this one should re-define the _real_initialize() and
409 _real_extract() methods and define a _VALID_URL regexp.
410 Probably, they should also be added to the list of extractors.
412 Subclasses may also override suitable() if necessary, but ensure the function
413 signature is preserved and that this function imports everything it needs
414 (except other extractors), so that lazy_extractors works correctly
416 _GEO_BYPASS attribute may be set to False in order to disable
417 geo restriction bypass mechanisms for a particular extractor.
418 Though it won't disable explicit geo restriction bypass based on
419 country code provided with geo_bypass_country.
421 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
422 countries for this extractor. One of these countries will be used by
423 geo restriction bypass mechanism right away in order to bypass
424 geo restriction, of course, if the mechanism is not disabled.
426 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
427 IP blocks in CIDR notation for this extractor. One of these IP blocks
428 will be used by geo restriction bypass mechanism similarly
431 The _WORKING attribute should be set to False for broken IEs
432 in order to warn the users and skip the tests.
437 _x_forwarded_for_ip
= None
439 _GEO_COUNTRIES
= None
440 _GEO_IP_BLOCKS
= None
444 'any': 'Use --cookies, --username and --password, or --netrc to provide account credentials',
446 'Use --cookies-from-browser or --cookies for the authentication. '
447 'See https://github.com/ytdl-org/youtube-dl#how-do-i-pass-cookies-to-youtube-dl for how to manually pass cookies'),
448 'password': 'Use --username and --password, or --netrc to provide account credentials',
451 def __init__(self
, downloader
=None):
452 """Constructor. Receives an optional downloader (a YoutubeDL instance).
453 If a downloader is not passed during initialization,
454 it must be set using "set_downloader()" before "extract()" is called"""
456 self
._x
_forwarded
_for
_ip
= None
457 self
._printed
_messages
= set()
458 self
.set_downloader(downloader
)
461 def _match_valid_url(cls
, url
):
462 # This does not use has/getattr intentionally - we want to know whether
463 # we have cached the regexp for *this* class, whereas getattr would also
464 # match the superclass
465 if '_VALID_URL_RE' not in cls
.__dict
__:
466 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
467 return cls
._VALID
_URL
_RE
.match(url
)
470 def suitable(cls
, url
):
471 """Receives a URL and returns True if suitable for this IE."""
472 # This function must import everything it needs (except other extractors),
473 # so that lazy_extractors works correctly
474 return cls
._match
_valid
_url
(url
) is not None
477 def _match_id(cls
, url
):
478 return cls
._match
_valid
_url
(url
).group('id')
481 def get_temp_id(cls
, url
):
483 return cls
._match
_id
(url
)
484 except (IndexError, AttributeError):
489 """Getter method for _WORKING."""
492 def initialize(self
):
493 """Initializes an instance (authentication, etc)."""
494 self
._printed
_messages
= set()
495 self
._initialize
_geo
_bypass
({
496 'countries': self
._GEO
_COUNTRIES
,
497 'ip_blocks': self
._GEO
_IP
_BLOCKS
,
500 self
._real
_initialize
()
503 def _initialize_geo_bypass(self
, geo_bypass_context
):
505 Initialize geo restriction bypass mechanism.
507 This method is used to initialize geo bypass mechanism based on faking
508 X-Forwarded-For HTTP header. A random country from provided country list
509 is selected and a random IP belonging to this country is generated. This
510 IP will be passed as X-Forwarded-For HTTP header in all subsequent
513 This method will be used for initial geo bypass mechanism initialization
514 during the instance initialization with _GEO_COUNTRIES and
517 You may also manually call it from extractor's code if geo bypass
518 information is not available beforehand (e.g. obtained during
519 extraction) or due to some other reason. In this case you should pass
520 this information in geo bypass context passed as first argument. It may
521 contain following fields:
523 countries: List of geo unrestricted countries (similar
525 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
526 (similar to _GEO_IP_BLOCKS)
529 if not self
._x
_forwarded
_for
_ip
:
531 # Geo bypass mechanism is explicitly disabled by user
532 if not self
.get_param('geo_bypass', True):
535 if not geo_bypass_context
:
536 geo_bypass_context
= {}
538 # Backward compatibility: previously _initialize_geo_bypass
539 # expected a list of countries, some 3rd party code may still use
541 if isinstance(geo_bypass_context
, (list, tuple)):
542 geo_bypass_context
= {
543 'countries': geo_bypass_context
,
546 # The whole point of geo bypass mechanism is to fake IP
547 # as X-Forwarded-For HTTP header based on some IP block or
550 # Path 1: bypassing based on IP block in CIDR notation
552 # Explicit IP block specified by user, use it right away
553 # regardless of whether extractor is geo bypassable or not
554 ip_block
= self
.get_param('geo_bypass_ip_block', None)
556 # Otherwise use random IP block from geo bypass context but only
557 # if extractor is known as geo bypassable
559 ip_blocks
= geo_bypass_context
.get('ip_blocks')
560 if self
._GEO
_BYPASS
and ip_blocks
:
561 ip_block
= random
.choice(ip_blocks
)
564 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(ip_block
)
565 self
._downloader
.write_debug(
566 '[debug] Using fake IP %s as X-Forwarded-For' % self
._x
_forwarded
_for
_ip
)
569 # Path 2: bypassing based on country code
571 # Explicit country code specified by user, use it right away
572 # regardless of whether extractor is geo bypassable or not
573 country
= self
.get_param('geo_bypass_country', None)
575 # Otherwise use random country code from geo bypass context but
576 # only if extractor is known as geo bypassable
578 countries
= geo_bypass_context
.get('countries')
579 if self
._GEO
_BYPASS
and countries
:
580 country
= random
.choice(countries
)
583 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country
)
584 self
._downloader
.write_debug(
585 'Using fake IP %s (%s) as X-Forwarded-For' % (self
._x
_forwarded
_for
_ip
, country
.upper()))
587 def extract(self
, url
):
588 """Extracts URL information and returns it in list of dicts."""
593 self
.write_debug('Extracting URL: %s' % url
)
594 ie_result
= self
._real
_extract
(url
)
595 if ie_result
is None:
597 if self
._x
_forwarded
_for
_ip
:
598 ie_result
['__x_forwarded_for_ip'] = self
._x
_forwarded
_for
_ip
599 subtitles
= ie_result
.get('subtitles')
600 if (subtitles
and 'live_chat' in subtitles
601 and 'no-live-chat' in self
.get_param('compat_opts', [])):
602 del subtitles
['live_chat']
604 except GeoRestrictedError
as e
:
605 if self
.__maybe
_fake
_ip
_and
_retry
(e
.countries
):
608 except UnsupportedError
:
610 except ExtractorError
as e
:
612 'video_id': e
.video_id
or self
.get_temp_id(url
),
615 'expected': e
.expected
,
618 if hasattr(e
, 'countries'):
619 kwargs
['countries'] = e
.countries
620 raise type(e
)(e
.msg
, **kwargs
)
621 except compat_http_client
.IncompleteRead
as e
:
622 raise ExtractorError('A network error has occurred.', cause
=e
, expected
=True, video_id
=self
.get_temp_id(url
))
623 except (KeyError, StopIteration) as e
:
624 raise ExtractorError('An extractor error has occurred.', cause
=e
, video_id
=self
.get_temp_id(url
))
626 def __maybe_fake_ip_and_retry(self
, countries
):
627 if (not self
.get_param('geo_bypass_country', None)
629 and self
.get_param('geo_bypass', True)
630 and not self
._x
_forwarded
_for
_ip
632 country_code
= random
.choice(countries
)
633 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country_code
)
634 if self
._x
_forwarded
_for
_ip
:
636 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
637 % (self
._x
_forwarded
_for
_ip
, country_code
.upper()))
641 def set_downloader(self
, downloader
):
642 """Sets the downloader for this IE."""
643 self
._downloader
= downloader
645 def _real_initialize(self
):
646 """Real initialization process. Redefine in subclasses."""
649 def _real_extract(self
, url
):
650 """Real extraction process. Redefine in subclasses."""
655 """A string for getting the InfoExtractor with get_info_extractor"""
656 return cls
.__name
__[:-2]
660 return compat_str(type(self
).__name
__[:-2])
663 def __can_accept_status_code(err
, expected_status
):
664 assert isinstance(err
, compat_urllib_error
.HTTPError
)
665 if expected_status
is None:
667 elif callable(expected_status
):
668 return expected_status(err
.code
) is True
670 return err
.code
in variadic(expected_status
)
672 def _request_webpage(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
, expected_status
=None):
674 Return the response handle.
676 See _download_webpage docstring for arguments specification.
678 if not self
._downloader
._first
_webpage
_request
:
679 sleep_interval
= self
.get_param('sleep_interval_requests') or 0
680 if sleep_interval
> 0:
681 self
.to_screen('Sleeping %s seconds ...' % sleep_interval
)
682 time
.sleep(sleep_interval
)
684 self
._downloader
._first
_webpage
_request
= False
687 self
.report_download_webpage(video_id
)
688 elif note
is not False:
690 self
.to_screen('%s' % (note
,))
692 self
.to_screen('%s: %s' % (video_id
, note
))
694 # Some sites check X-Forwarded-For HTTP header in order to figure out
695 # the origin of the client behind proxy. This allows bypassing geo
696 # restriction by faking this header's value to IP that belongs to some
697 # geo unrestricted country. We will do so once we encounter any
698 # geo restriction error.
699 if self
._x
_forwarded
_for
_ip
:
700 if 'X-Forwarded-For' not in headers
:
701 headers
['X-Forwarded-For'] = self
._x
_forwarded
_for
_ip
703 if isinstance(url_or_request
, compat_urllib_request
.Request
):
704 url_or_request
= update_Request(
705 url_or_request
, data
=data
, headers
=headers
, query
=query
)
708 url_or_request
= update_url_query(url_or_request
, query
)
709 if data
is not None or headers
:
710 url_or_request
= sanitized_Request(url_or_request
, data
, headers
)
712 return self
._downloader
.urlopen(url_or_request
)
713 except network_exceptions
as err
:
714 if isinstance(err
, compat_urllib_error
.HTTPError
):
715 if self
.__can
_accept
_status
_code
(err
, expected_status
):
716 # Retain reference to error to prevent file object from
717 # being closed before it can be read. Works around the
718 # effects of <https://bugs.python.org/issue15002>
719 # introduced in Python 3.4.1.
726 errnote
= 'Unable to download webpage'
728 errmsg
= '%s: %s' % (errnote
, error_to_compat_str(err
))
730 raise ExtractorError(errmsg
, sys
.exc_info()[2], cause
=err
)
732 self
.report_warning(errmsg
)
735 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):
737 Return a tuple (page content as string, URL handle).
739 See _download_webpage docstring for arguments specification.
741 # Strip hashes from the URL (#1038)
742 if isinstance(url_or_request
, (compat_str
, str)):
743 url_or_request
= url_or_request
.partition('#')[0]
745 urlh
= self
._request
_webpage
(url_or_request
, video_id
, note
, errnote
, fatal
, data
=data
, headers
=headers
, query
=query
, expected_status
=expected_status
)
749 content
= self
._webpage
_read
_content
(urlh
, url_or_request
, video_id
, note
, errnote
, fatal
, encoding
=encoding
)
750 return (content
, urlh
)
753 def _guess_encoding_from_content(content_type
, webpage_bytes
):
754 m
= re
.match(r
'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type
)
756 encoding
= m
.group(1)
758 m
= re
.search(br
'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
759 webpage_bytes[:1024])
761 encoding = m.group(1).decode('ascii')
762 elif webpage_bytes.startswith(b'\xff\xfe'):
769 def __check_blocked(self, content):
770 first_block = content[:512]
771 if ('<title>Access to this site is blocked</title>' in content
772 and 'Websense' in first_block):
773 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
774 blocked_iframe = self._html_search_regex(
775 r'<iframe src="([^
"]+)"', content,
776 'Websense information URL
', default=None)
778 msg += ' Visit
%s for more details
' % blocked_iframe
779 raise ExtractorError(msg, expected=True)
780 if '<title
>The URL you requested has been blocked
</title
>' in first_block:
782 'Access to this webpage has been blocked by Indian censorship
. '
783 'Use a VPN
or proxy
server (with --proxy
) to route around it
.')
784 block_msg = self._html_search_regex(
785 r'</h1
><p
>(.*?
)</p
>',
786 content, 'block message
', default=None)
788 msg += ' (Message
: "%s")' % block_msg.replace('\n', ' ')
789 raise ExtractorError(msg, expected=True)
790 if ('<title
>TTK
:: Доступ к ресурсу ограничен
</title
>' in content
791 and 'blocklist
.rkn
.gov
.ru
' in content):
792 raise ExtractorError(
793 'Access to this webpage has been blocked by decision of the Russian government
. '
794 'Visit http
://blocklist
.rkn
.gov
.ru
/ for a block reason
.',
797 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
798 content_type = urlh.headers.get('Content
-Type
', '')
799 webpage_bytes = urlh.read()
800 if prefix is not None:
801 webpage_bytes = prefix + webpage_bytes
803 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
804 if self.get_param('dump_intermediate_pages
', False):
805 self.to_screen('Dumping request to
' + urlh.geturl())
806 dump = base64.b64encode(webpage_bytes).decode('ascii
')
807 self._downloader.to_screen(dump)
808 if self.get_param('write_pages
', False):
809 basen = '%s_%s' % (video_id, urlh.geturl())
810 trim_length = self.get_param('trim_file_name
') or 240
811 if len(basen) > trim_length:
812 h = '___
' + hashlib.md5(basen.encode('utf
-8')).hexdigest()
813 basen = basen[:trim_length - len(h)] + h
814 raw_filename = basen + '.dump
'
815 filename = sanitize_filename(raw_filename, restricted=True)
816 self.to_screen('Saving request to
' + filename)
817 # Working around MAX_PATH limitation on Windows (see
818 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
819 if compat_os_name == 'nt
':
820 absfilepath = os.path.abspath(filename)
821 if len(absfilepath) > 259:
822 filename = '\\\\?
\\' + absfilepath
823 with open(filename, 'wb
') as outf:
824 outf.write(webpage_bytes)
827 content = webpage_bytes.decode(encoding, 'replace
')
829 content = webpage_bytes.decode('utf
-8', 'replace
')
831 self.__check_blocked(content)
835 def _download_webpage(
836 self, url_or_request, video_id, note=None, errnote=None,
837 fatal=True, tries=1, timeout=5, encoding=None, data=None,
838 headers={}, query={}, expected_status=None):
840 Return the data of the page as a string.
843 url_or_request -- plain text URL as a string or
844 a compat_urllib_request.Requestobject
845 video_id -- Video/playlist/item identifier (string)
848 note -- note printed before downloading (string)
849 errnote -- note printed in case of an error (string)
850 fatal -- flag denoting whether error should be considered fatal,
851 i.e. whether it should cause ExtractionError to be raised,
852 otherwise a warning will be reported and extraction continued
853 tries -- number of tries
854 timeout -- sleep interval between tries
855 encoding -- encoding for a page content decoding, guessed automatically
856 when not explicitly specified
857 data -- POST data (bytes)
858 headers -- HTTP headers (dict)
859 query -- URL query (dict)
860 expected_status -- allows to accept failed HTTP requests (non 2xx
861 status code) by explicitly specifying a set of accepted status
862 codes. Can be any of the following entities:
863 - an integer type specifying an exact failed status code to
865 - a list or a tuple of integer types specifying a list of
866 failed status codes to accept
867 - a callable accepting an actual failed status code and
868 returning True if it should be accepted
869 Note that this argument does not affect success status codes (2xx)
870 which are always accepted.
875 while success is False:
877 res = self._download_webpage_handle(
878 url_or_request, video_id, note, errnote, fatal,
879 encoding=encoding, data=data, headers=headers, query=query,
880 expected_status=expected_status)
882 except compat_http_client.IncompleteRead as e:
884 if try_count >= tries:
886 self._sleep(timeout, video_id)
893 def _download_xml_handle(
894 self, url_or_request, video_id, note='Downloading XML
',
895 errnote='Unable to download XML
', transform_source=None,
896 fatal=True, encoding=None, data=None, headers={}, query={},
897 expected_status=None):
899 Return a tuple (xml as an compat_etree_Element, URL handle).
901 See _download_webpage docstring for arguments specification.
903 res = self._download_webpage_handle(
904 url_or_request, video_id, note, errnote, fatal=fatal,
905 encoding=encoding, data=data, headers=headers, query=query,
906 expected_status=expected_status)
909 xml_string, urlh = res
910 return self._parse_xml(
911 xml_string, video_id, transform_source=transform_source,
915 self, url_or_request, video_id,
916 note='Downloading XML
', errnote='Unable to download XML
',
917 transform_source=None, fatal=True, encoding=None,
918 data=None, headers={}, query={}, expected_status=None):
920 Return the xml as an compat_etree_Element.
922 See _download_webpage docstring for arguments specification.
924 res = self._download_xml_handle(
925 url_or_request, video_id, note=note, errnote=errnote,
926 transform_source=transform_source, fatal=fatal, encoding=encoding,
927 data=data, headers=headers, query=query,
928 expected_status=expected_status)
929 return res if res is False else res[0]
931 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
933 xml_string = transform_source(xml_string)
935 return compat_etree_fromstring(xml_string.encode('utf
-8'))
936 except compat_xml_parse_error as ve:
937 errmsg = '%s: Failed to parse XML
' % video_id
939 raise ExtractorError(errmsg, cause=ve)
941 self.report_warning(errmsg + str(ve))
943 def _download_json_handle(
944 self, url_or_request, video_id, note='Downloading JSON metadata
',
945 errnote='Unable to download JSON metadata
', transform_source=None,
946 fatal=True, encoding=None, data=None, headers={}, query={},
947 expected_status=None):
949 Return a tuple (JSON object, URL handle).
951 See _download_webpage docstring for arguments specification.
953 res = self._download_webpage_handle(
954 url_or_request, video_id, note, errnote, fatal=fatal,
955 encoding=encoding, data=data, headers=headers, query=query,
956 expected_status=expected_status)
959 json_string, urlh = res
960 return self._parse_json(
961 json_string, video_id, transform_source=transform_source,
965 self, url_or_request, video_id, note='Downloading JSON metadata
',
966 errnote='Unable to download JSON metadata
', transform_source=None,
967 fatal=True, encoding=None, data=None, headers={}, query={},
968 expected_status=None):
970 Return the JSON object as a dict.
972 See _download_webpage docstring for arguments specification.
974 res = self._download_json_handle(
975 url_or_request, video_id, note=note, errnote=errnote,
976 transform_source=transform_source, fatal=fatal, encoding=encoding,
977 data=data, headers=headers, query=query,
978 expected_status=expected_status)
979 return res if res is False else res[0]
981 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
983 json_string = transform_source(json_string)
985 return json.loads(json_string)
986 except ValueError as ve:
987 errmsg = '%s: Failed to parse JSON
' % video_id
989 raise ExtractorError(errmsg, cause=ve)
991 self.report_warning(errmsg + str(ve))
993 def _parse_socket_response_as_json(self, data, video_id, transform_source=None, fatal=True):
994 return self._parse_json(
995 data[data.find('{'):data.rfind('}
') + 1],
996 video_id, transform_source, fatal)
998 def _download_socket_json_handle(
999 self, url_or_request, video_id, note='Polling socket
',
1000 errnote='Unable to poll socket
', transform_source=None,
1001 fatal=True, encoding=None, data=None, headers={}, query={},
1002 expected_status=None):
1004 Return a tuple (JSON object, URL handle).
1006 See _download_webpage docstring for arguments specification.
1008 res = self._download_webpage_handle(
1009 url_or_request, video_id, note, errnote, fatal=fatal,
1010 encoding=encoding, data=data, headers=headers, query=query,
1011 expected_status=expected_status)
1015 return self._parse_socket_response_as_json(
1016 webpage, video_id, transform_source=transform_source,
1019 def _download_socket_json(
1020 self, url_or_request, video_id, note='Polling socket
',
1021 errnote='Unable to poll socket
', transform_source=None,
1022 fatal=True, encoding=None, data=None, headers={}, query={},
1023 expected_status=None):
1025 Return the JSON object as a dict.
1027 See _download_webpage docstring for arguments specification.
1029 res = self._download_socket_json_handle(
1030 url_or_request, video_id, note=note, errnote=errnote,
1031 transform_source=transform_source, fatal=fatal, encoding=encoding,
1032 data=data, headers=headers, query=query,
1033 expected_status=expected_status)
1034 return res if res is False else res[0]
1036 def report_warning(self, msg, video_id=None, *args, only_once=False, **kwargs):
1037 idstr = format_field(video_id, template='%s: ')
1038 msg = f'[{self.IE_NAME}
] {idstr}{msg}
'
1040 if f'WARNING
: {msg}
' in self._printed_messages:
1042 self._printed_messages.add(f'WARNING
: {msg}
')
1043 self._downloader.report_warning(msg, *args, **kwargs)
1045 def to_screen(self, msg, *args, **kwargs):
1046 """Print msg to screen, prefixing it with '[ie_name
]'"""
1047 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg), *args, **kwargs)
1049 def write_debug(self, msg, *args, **kwargs):
1050 self._downloader.write_debug('[%s] %s' % (self.IE_NAME, msg), *args, **kwargs)
1052 def get_param(self, name, default=None, *args, **kwargs):
1053 if self._downloader:
1054 return self._downloader.params.get(name, default, *args, **kwargs)
1057 def report_drm(self, video_id, partial=False):
1058 self.raise_no_formats('This video
is DRM protected
', expected=True, video_id=video_id)
1060 def report_extraction(self, id_or_name):
1061 """Report information extraction."""
1062 self.to_screen('%s: Extracting information
' % id_or_name)
1064 def report_download_webpage(self, video_id):
1065 """Report webpage download."""
1066 self.to_screen('%s: Downloading webpage
' % video_id)
1068 def report_age_confirmation(self):
1069 """Report attempt to confirm age."""
1070 self.to_screen('Confirming age
')
1072 def report_login(self):
1073 """Report attempt to log in."""
1074 self.to_screen('Logging
in')
1076 def raise_login_required(
1077 self, msg='This video
is only available
for registered users
',
1078 metadata_available=False, method='any
'):
1079 if metadata_available and self.get_param('ignore_no_formats_error
'):
1080 self.report_warning(msg)
1081 if method is not None:
1082 msg = '%s. %s' % (msg, self._LOGIN_HINTS[method])
1083 raise ExtractorError(msg, expected=True)
1085 def raise_geo_restricted(
1086 self, msg='This video
is not available
from your location due to geo restriction
',
1087 countries=None, metadata_available=False):
1088 if metadata_available and self.get_param('ignore_no_formats_error
'):
1089 self.report_warning(msg)
1091 raise GeoRestrictedError(msg, countries=countries)
1093 def raise_no_formats(self, msg, expected=False, video_id=None):
1094 if expected and self.get_param('ignore_no_formats_error
'):
1095 self.report_warning(msg, video_id)
1096 elif isinstance(msg, ExtractorError):
1099 raise ExtractorError(msg, expected=expected, video_id=video_id)
1101 # Methods for following #608
1103 def url_result(url, ie=None, video_id=None, video_title=None, **kwargs):
1104 """Returns a URL that points to a page that should be processed"""
1105 # TODO: ie should be the class used for getting the info
1106 video_info = {'_type
': 'url
',
1109 video_info.update(kwargs)
1110 if video_id is not None:
1111 video_info['id'] = video_id
1112 if video_title is not None:
1113 video_info['title
'] = video_title
1116 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None):
1118 self.url_result(self._proto_relative_url(getter(m) if getter else m), ie)
1120 return self.playlist_result(
1121 urls, playlist_id=playlist_id, playlist_title=playlist_title)
1124 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None, **kwargs):
1125 """Returns a playlist"""
1126 video_info = {'_type
': 'playlist
',
1128 video_info.update(kwargs)
1130 video_info['id'] = playlist_id
1132 video_info['title
'] = playlist_title
1133 if playlist_description is not None:
1134 video_info['description
'] = playlist_description
1137 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1139 Perform a regex search on the given string, using a single or a list of
1140 patterns returning the first matching group.
1141 In case of failure return a default value or raise a WARNING or a
1142 RegexNotFoundError, depending on fatal, specifying the field name.
1144 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
1145 mobj = re.search(pattern, string, flags)
1148 mobj = re.search(p, string, flags)
1152 _name = self._downloader._format_err(name, self._downloader.Styles.EMPHASIS)
1156 # return the first matching group
1157 return next(g for g in mobj.groups() if g is not None)
1158 elif isinstance(group, (list, tuple)):
1159 return tuple(mobj.group(g) for g in group)
1161 return mobj.group(group)
1162 elif default is not NO_DEFAULT:
1165 raise RegexNotFoundError('Unable to extract
%s' % _name)
1167 self.report_warning('unable to extract
%s' % _name + bug_reports_message())
1170 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1172 Like _search_regex, but strips HTML tags and unescapes entities.
1174 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1176 return clean_html(res).strip()
1180 def _get_netrc_login_info(self, netrc_machine=None):
1183 netrc_machine = netrc_machine or self._NETRC_MACHINE
1185 if self.get_param('usenetrc
', False):
1187 netrc_file = compat_expanduser(self.get_param('netrc_location
') or '~
')
1188 if os.path.isdir(netrc_file):
1189 netrc_file = os.path.join(netrc_file, '.netrc
')
1190 info = netrc.netrc(file=netrc_file).authenticators(netrc_machine)
1191 if info is not None:
1195 raise netrc.NetrcParseError(
1196 'No authenticators
for %s' % netrc_machine)
1197 except (IOError, netrc.NetrcParseError) as err:
1198 self.report_warning(
1199 'parsing
.netrc
: %s' % error_to_compat_str(err))
1201 return username, password
1203 def _get_login_info(self, username_option='username
', password_option='password
', netrc_machine=None):
1205 Get the login info as (username, password)
1206 First look for the manually specified credentials using username_option
1207 and password_option as keys in params dictionary. If no such credentials
1208 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1210 If there's no info available
, return (None, None)
1213 # Attempt to use provided username and password or .netrc data
1214 username = self.get_param(username_option)
1215 if username is not None:
1216 password = self.get_param(password_option)
1218 username, password = self._get_netrc_login_info(netrc_machine)
1220 return username, password
1222 def _get_tfa_info(self, note='two-factor verification code'):
1224 Get the two
-factor authentication info
1225 TODO
- asking the user will be required
for sms
/phone verify
1226 currently just uses the command line option
1227 If there
's no info available, return None
1230 tfa = self.get_param('twofactor
')
1234 return compat_getpass('Type
%s and press
[Return
]: ' % note)
1236 # Helper functions for extracting OpenGraph info
1238 def _og_regexes(prop):
1239 content_re = r'content
=(?
:"([^"]+?
)"|\'([^\']+?)\'|\s*([^\s"\'=<>`
]+?
))'
1240 property_re = (r'(?
:name|
property)=(?
:\'og
[:-]%(prop)s\'|
"og[:-]%(prop)s"|\s
*og
[:-]%(prop)s\b)'
1241 % {'prop': re.escape(prop)})
1242 template = r'<meta
[^
>]+?
%s[^
>]+?
%s'
1244 template % (property_re, content_re),
1245 template % (content_re, property_re),
1249 def _meta_regex(prop):
1250 return r'''(?isx)<meta
1251 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1252 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1254 def _og_search_property(self, prop, html, name=None, **kargs):
1255 prop = variadic(prop)
1257 name = 'OpenGraph
%s' % prop[0]
1260 og_regexes.extend(self._og_regexes(p))
1261 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1264 return unescapeHTML(escaped)
1266 def _og_search_thumbnail(self, html, **kargs):
1267 return self._og_search_property('image
', html, 'thumbnail URL
', fatal=False, **kargs)
1269 def _og_search_description(self, html, **kargs):
1270 return self._og_search_property('description
', html, fatal=False, **kargs)
1272 def _og_search_title(self, html, **kargs):
1273 return self._og_search_property('title
', html, **kargs)
1275 def _og_search_video_url(self, html, name='video url
', secure=True, **kargs):
1276 regexes = self._og_regexes('video
') + self._og_regexes('video
:url
')
1278 regexes = self._og_regexes('video
:secure_url
') + regexes
1279 return self._html_search_regex(regexes, html, name, **kargs)
1281 def _og_search_url(self, html, **kargs):
1282 return self._og_search_property('url
', html, **kargs)
1284 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1285 name = variadic(name)
1286 if display_name is None:
1287 display_name = name[0]
1288 return self._html_search_regex(
1289 [self._meta_regex(n) for n in name],
1290 html, display_name, fatal=fatal, group='content
', **kwargs)
1292 def _dc_search_uploader(self, html):
1293 return self._html_search_meta('dc
.creator
', html, 'uploader
')
1295 def _rta_search(self, html):
1296 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1297 if re.search(r'(?ix
)<meta\s
+name
="rating"\s
+'
1298 r' content
="RTA-5042-1996-1400-1577-RTA"',
1303 def _media_rating_search(self, html):
1304 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1305 rating = self._html_search_meta('rating
', html)
1317 return RATING_TABLE.get(rating.lower())
1319 def _family_friendly_search(self, html):
1320 # See http://schema.org/VideoObject
1321 family_friendly = self._html_search_meta(
1322 'isFamilyFriendly
', html, default=None)
1324 if not family_friendly:
1333 return RATING_TABLE.get(family_friendly.lower())
1335 def _twitter_search_player(self, html):
1336 return self._html_search_meta('twitter
:player
', html,
1337 'twitter card player
')
1339 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
1340 json_ld_list = list(re.finditer(JSON_LD_RE, html))
1341 default = kwargs.get('default
', NO_DEFAULT)
1342 # JSON-LD may be malformed and thus `fatal` should be respected.
1343 # At the same time `default` may be passed that assumes `fatal=False`
1344 # for _search_regex. Let's simulate the same behavior here
as well
.
1345 fatal
= kwargs
.get('fatal', True) if default
is NO_DEFAULT
else False
1347 for mobj
in json_ld_list
:
1348 json_ld_item
= self
._parse
_json
(
1349 mobj
.group('json_ld'), video_id
, fatal
=fatal
)
1350 if not json_ld_item
:
1352 if isinstance(json_ld_item
, dict):
1353 json_ld
.append(json_ld_item
)
1354 elif isinstance(json_ld_item
, (list, tuple)):
1355 json_ld
.extend(json_ld_item
)
1357 json_ld
= self
._json
_ld
(json_ld
, video_id
, fatal
=fatal
, expected_type
=expected_type
)
1360 if default
is not NO_DEFAULT
:
1363 raise RegexNotFoundError('Unable to extract JSON-LD')
1365 self
.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1368 def _json_ld(self
, json_ld
, video_id
, fatal
=True, expected_type
=None):
1369 if isinstance(json_ld
, compat_str
):
1370 json_ld
= self
._parse
_json
(json_ld
, video_id
, fatal
=fatal
)
1374 if not isinstance(json_ld
, (list, tuple, dict)):
1376 if isinstance(json_ld
, dict):
1379 INTERACTION_TYPE_MAP
= {
1380 'CommentAction': 'comment',
1381 'AgreeAction': 'like',
1382 'DisagreeAction': 'dislike',
1383 'LikeAction': 'like',
1384 'DislikeAction': 'dislike',
1385 'ListenAction': 'view',
1386 'WatchAction': 'view',
1387 'ViewAction': 'view',
1390 def extract_interaction_type(e
):
1391 interaction_type
= e
.get('interactionType')
1392 if isinstance(interaction_type
, dict):
1393 interaction_type
= interaction_type
.get('@type')
1394 return str_or_none(interaction_type
)
1396 def extract_interaction_statistic(e
):
1397 interaction_statistic
= e
.get('interactionStatistic')
1398 if isinstance(interaction_statistic
, dict):
1399 interaction_statistic
= [interaction_statistic
]
1400 if not isinstance(interaction_statistic
, list):
1402 for is_e
in interaction_statistic
:
1403 if not isinstance(is_e
, dict):
1405 if is_e
.get('@type') != 'InteractionCounter':
1407 interaction_type
= extract_interaction_type(is_e
)
1408 if not interaction_type
:
1410 # For interaction count some sites provide string instead of
1411 # an integer (as per spec) with non digit characters (e.g. ",")
1412 # so extracting count with more relaxed str_to_int
1413 interaction_count
= str_to_int(is_e
.get('userInteractionCount'))
1414 if interaction_count
is None:
1416 count_kind
= INTERACTION_TYPE_MAP
.get(interaction_type
.split('/')[-1])
1419 count_key
= '%s_count' % count_kind
1420 if info
.get(count_key
) is not None:
1422 info
[count_key
] = interaction_count
1424 def extract_video_object(e
):
1425 assert e
['@type'] == 'VideoObject'
1426 author
= e
.get('author')
1428 'url': url_or_none(e
.get('contentUrl')),
1429 'title': unescapeHTML(e
.get('name')),
1430 'description': unescapeHTML(e
.get('description')),
1431 'thumbnail': url_or_none(e
.get('thumbnailUrl') or e
.get('thumbnailURL')),
1432 'duration': parse_duration(e
.get('duration')),
1433 'timestamp': unified_timestamp(e
.get('uploadDate')),
1434 # author can be an instance of 'Organization' or 'Person' types.
1435 # both types can have 'name' property(inherited from 'Thing' type). [1]
1436 # however some websites are using 'Text' type instead.
1437 # 1. https://schema.org/VideoObject
1438 'uploader': author
.get('name') if isinstance(author
, dict) else author
if isinstance(author
, compat_str
) else None,
1439 'filesize': float_or_none(e
.get('contentSize')),
1440 'tbr': int_or_none(e
.get('bitrate')),
1441 'width': int_or_none(e
.get('width')),
1442 'height': int_or_none(e
.get('height')),
1443 'view_count': int_or_none(e
.get('interactionCount')),
1445 extract_interaction_statistic(e
)
1449 item_type
= e
.get('@type')
1450 if expected_type
is not None and expected_type
!= item_type
:
1452 if item_type
in ('TVEpisode', 'Episode'):
1453 episode_name
= unescapeHTML(e
.get('name'))
1455 'episode': episode_name
,
1456 'episode_number': int_or_none(e
.get('episodeNumber')),
1457 'description': unescapeHTML(e
.get('description')),
1459 if not info
.get('title') and episode_name
:
1460 info
['title'] = episode_name
1461 part_of_season
= e
.get('partOfSeason')
1462 if isinstance(part_of_season
, dict) and part_of_season
.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1464 'season': unescapeHTML(part_of_season
.get('name')),
1465 'season_number': int_or_none(part_of_season
.get('seasonNumber')),
1467 part_of_series
= e
.get('partOfSeries') or e
.get('partOfTVSeries')
1468 if isinstance(part_of_series
, dict) and part_of_series
.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1469 info
['series'] = unescapeHTML(part_of_series
.get('name'))
1470 elif item_type
== 'Movie':
1472 'title': unescapeHTML(e
.get('name')),
1473 'description': unescapeHTML(e
.get('description')),
1474 'duration': parse_duration(e
.get('duration')),
1475 'timestamp': unified_timestamp(e
.get('dateCreated')),
1477 elif item_type
in ('Article', 'NewsArticle'):
1479 'timestamp': parse_iso8601(e
.get('datePublished')),
1480 'title': unescapeHTML(e
.get('headline')),
1481 'description': unescapeHTML(e
.get('articleBody')),
1483 elif item_type
== 'VideoObject':
1484 extract_video_object(e
)
1485 if expected_type
is None:
1489 video
= e
.get('video')
1490 if isinstance(video
, dict) and video
.get('@type') == 'VideoObject':
1491 extract_video_object(video
)
1492 if expected_type
is None:
1496 return dict((k
, v
) for k
, v
in info
.items() if v
is not None)
1499 def _hidden_inputs(html
):
1500 html
= re
.sub(r
'<!--(?:(?!<!--).)*-->', '', html
)
1502 for input in re
.findall(r
'(?i)(<input[^>]+>)', html
):
1503 attrs
= extract_attributes(input)
1506 if attrs
.get('type') not in ('hidden', 'submit'):
1508 name
= attrs
.get('name') or attrs
.get('id')
1509 value
= attrs
.get('value')
1510 if name
and value
is not None:
1511 hidden_inputs
[name
] = value
1512 return hidden_inputs
1514 def _form_hidden_inputs(self
, form_id
, html
):
1515 form
= self
._search
_regex
(
1516 r
'(?is)<form[^>]+?id=(["\'])%s\
1[^
>]*>(?P
<form
>.+?
)</form
>' % form_id,
1517 html, '%s form
' % form_id, group='form
')
1518 return self._hidden_inputs(form)
1521 regex = r' *((?P
<reverse
>\
+)?
(?P
<field
>[a
-zA
-Z0
-9_]+)((?P
<separator
>[~
:])(?P
<limit
>.*?
))?
)?
*$
'
1523 default = ('hidden
', 'aud_or_vid
', 'hasvid
', 'ie_pref
', 'lang
', 'quality
',
1524 'res
', 'fps
', 'hdr
:12', 'codec
:vp9
.2
', 'size
', 'br
', 'asr
',
1525 'proto
', 'ext
', 'hasaud
', 'source
', 'format_id
') # These must not be aliases
1526 ytdl_default = ('hasaud
', 'lang
', 'quality
', 'tbr
', 'filesize
', 'vbr
',
1527 'height
', 'width
', 'proto
', 'vext
', 'abr
', 'aext
',
1528 'fps
', 'fs_approx
', 'source
', 'format_id
')
1531 'vcodec
': {'type': 'ordered
', 'regex
': True,
1532 'order
': ['av0?
1', 'vp0?
9.2', 'vp0?
9', '[hx
]265|he?vc?
', '[hx
]264|avc
', 'vp0?
8', 'mp4v|h263
', 'theora
', '', None, 'none
']},
1533 'acodec
': {'type': 'ordered
', 'regex
': True,
1534 'order
': ['opus
', 'vorbis
', 'aac
', 'mp?
4a?
', 'mp3
', 'e
-?a?c
-?
3', 'ac
-?
3', 'dts
', '', None, 'none
']},
1535 'hdr
': {'type': 'ordered
', 'regex
': True, 'field
': 'dynamic_range
',
1536 'order
': ['dv
', '(hdr
)?
12', r'(hdr
)?
10\
+', '(hdr
)?
10', 'hlg
', '', 'sdr
', None]},
1537 'proto
': {'type': 'ordered
', 'regex
': True, 'field
': 'protocol
',
1538 'order
': ['(ht|f
)tps
', '(ht|f
)tp$
', 'm3u8
.+', '.*dash
', 'ws|websocket
', '', 'mms|rtsp
', 'none
', 'f4
']},
1539 'vext
': {'type': 'ordered
', 'field
': 'video_ext
',
1540 'order
': ('mp4
', 'webm
', 'flv
', '', 'none
'),
1541 'order_free
': ('webm
', 'mp4
', 'flv
', '', 'none
')},
1542 'aext
': {'type': 'ordered
', 'field
': 'audio_ext
',
1543 'order
': ('m4a
', 'aac
', 'mp3
', 'ogg
', 'opus
', 'webm
', '', 'none
'),
1544 'order_free
': ('opus
', 'ogg
', 'webm
', 'm4a
', 'mp3
', 'aac
', '', 'none
')},
1545 'hidden
': {'visible': False, 'forced': True, 'type': 'extractor', 'max': -1000},
1546 'aud_or_vid
': {'visible
': False, 'forced
': True, 'type': 'multiple
',
1547 'field
': ('vcodec
', 'acodec
'),
1548 'function
': lambda it: int(any(v != 'none
' for v in it))},
1549 'ie_pref
': {'priority': True, 'type': 'extractor'},
1550 'hasvid
': {'priority': True, 'field': 'vcodec', 'type': 'boolean', 'not_in_list': ('none',)},
1551 'hasaud
': {'field': 'acodec', 'type': 'boolean', 'not_in_list': ('none',)},
1552 'lang
': {'convert': 'float', 'field': 'language_preference', 'default': -1},
1553 'quality
': {'convert': 'float', 'default': -1},
1554 'filesize
': {'convert': 'bytes'},
1555 'fs_approx
': {'convert': 'bytes', 'field': 'filesize_approx'},
1556 'id': {'convert': 'string', 'field': 'format_id'},
1557 'height
': {'convert': 'float_none'},
1558 'width
': {'convert': 'float_none'},
1559 'fps
': {'convert': 'float_none'},
1560 'tbr
': {'convert': 'float_none'},
1561 'vbr
': {'convert': 'float_none'},
1562 'abr
': {'convert': 'float_none'},
1563 'asr
': {'convert': 'float_none'},
1564 'source
': {'convert': 'float', 'field': 'source_preference', 'default': -1},
1566 'codec
': {'type': 'combined', 'field': ('vcodec', 'acodec')},
1567 'br
': {'type': 'combined', 'field': ('tbr', 'vbr', 'abr'), 'same_limit': True},
1568 'size
': {'type': 'combined', 'same_limit': True, 'field': ('filesize', 'fs_approx')},
1569 'ext
': {'type': 'combined', 'field': ('vext', 'aext')},
1570 'res
': {'type': 'multiple
', 'field
': ('height
', 'width
'),
1571 'function
': lambda it: (lambda l: min(l) if l else 0)(tuple(filter(None, it)))},
1573 # Most of these exist only for compatibility reasons
1574 'dimension
': {'type': 'alias', 'field': 'res'},
1575 'resolution
': {'type': 'alias', 'field': 'res'},
1576 'extension
': {'type': 'alias', 'field': 'ext'},
1577 'bitrate
': {'type': 'alias', 'field': 'br'},
1578 'total_bitrate
': {'type': 'alias', 'field': 'tbr'},
1579 'video_bitrate
': {'type': 'alias', 'field': 'vbr'},
1580 'audio_bitrate
': {'type': 'alias', 'field': 'abr'},
1581 'framerate
': {'type': 'alias', 'field': 'fps'},
1582 'language_preference
': {'type': 'alias', 'field': 'lang'}, # not named as 'language
' because such a field exists
1583 'protocol
': {'type': 'alias', 'field': 'proto'},
1584 'source_preference
': {'type': 'alias', 'field': 'source'},
1585 'filesize_approx
': {'type': 'alias', 'field': 'fs_approx'},
1586 'filesize_estimate
': {'type': 'alias', 'field': 'size'},
1587 'samplerate
': {'type': 'alias', 'field': 'asr'},
1588 'video_ext
': {'type': 'alias', 'field': 'vext'},
1589 'audio_ext
': {'type': 'alias', 'field': 'aext'},
1590 'video_codec
': {'type': 'alias', 'field': 'vcodec'},
1591 'audio_codec
': {'type': 'alias', 'field': 'acodec'},
1592 'video
': {'type': 'alias', 'field': 'hasvid'},
1593 'has_video
': {'type': 'alias', 'field': 'hasvid'},
1594 'audio
': {'type': 'alias', 'field': 'hasaud'},
1595 'has_audio
': {'type': 'alias', 'field': 'hasaud'},
1596 'extractor
': {'type': 'alias', 'field': 'ie_pref'},
1597 'preference
': {'type': 'alias', 'field': 'ie_pref'},
1598 'extractor_preference
': {'type': 'alias', 'field': 'ie_pref'},
1599 'format_id
': {'type': 'alias', 'field': 'id'},
1604 def _get_field_setting(self, field, key):
1605 if field not in self.settings:
1606 self.settings[field] = {}
1607 propObj = self.settings[field]
1608 if key not in propObj:
1609 type = propObj.get('type')
1611 default = 'preference
' if type == 'extractor
' else (field,) if type in ('combined
', 'multiple
') else field
1612 elif key == 'convert
':
1613 default = 'order
' if type == 'ordered
' else 'float_string
' if field else 'ignore
'
1615 default = {'type': 'field', 'visible': True, 'order': [], 'not_in_list': (None,)}.get(key, None)
1616 propObj[key] = default
1619 def _resolve_field_value(self, field, value, convertNone=False):
1624 value = value.lower()
1625 conversion = self._get_field_setting(field, 'convert
')
1626 if conversion == 'ignore
':
1628 if conversion == 'string
':
1630 elif conversion == 'float_none
':
1631 return float_or_none(value)
1632 elif conversion == 'bytes':
1633 return FileDownloader.parse_bytes(value)
1634 elif conversion == 'order
':
1635 order_list = (self._use_free_order and self._get_field_setting(field, 'order_free
')) or self._get_field_setting(field, 'order
')
1636 use_regex = self._get_field_setting(field, 'regex
')
1637 list_length = len(order_list)
1638 empty_pos = order_list.index('') if '' in order_list else list_length + 1
1639 if use_regex and value is not None:
1640 for i, regex in enumerate(order_list):
1641 if regex and re.match(regex, value):
1642 return list_length - i
1643 return list_length - empty_pos # not in list
1644 else: # not regex or value = None
1645 return list_length - (order_list.index(value) if value in order_list else empty_pos)
1647 if value.isnumeric():
1650 self.settings[field]['convert
'] = 'string
'
1653 def evaluate_params(self, params, sort_extractor):
1654 self._use_free_order = params.get('prefer_free_formats
', False)
1655 self._sort_user = params.get('format_sort
', [])
1656 self._sort_extractor = sort_extractor
1658 def add_item(field, reverse, closest, limit_text):
1659 field = field.lower()
1660 if field in self._order:
1662 self._order.append(field)
1663 limit = self._resolve_field_value(field, limit_text)
1666 'closest
': False if limit is None else closest,
1667 'limit_text
': limit_text,
1669 if field in self.settings:
1670 self.settings[field].update(data)
1672 self.settings[field] = data
1675 tuple(field for field in self.default if self._get_field_setting(field, 'forced
'))
1676 + (tuple() if params.get('format_sort_force
', False)
1677 else tuple(field for field in self.default if self._get_field_setting(field, 'priority
')))
1678 + tuple(self._sort_user) + tuple(sort_extractor) + self.default)
1680 for item in sort_list:
1681 match = re.match(self.regex, item)
1683 raise ExtractorError('Invalid format sort string
"%s" given by extractor
' % item)
1684 field = match.group('field
')
1687 if self._get_field_setting(field, 'type') == 'alias
':
1688 field = self._get_field_setting(field, 'field
')
1689 reverse = match.group('reverse
') is not None
1690 closest = match.group('separator
') == '~
'
1691 limit_text = match.group('limit
')
1693 has_limit = limit_text is not None
1694 has_multiple_fields = self._get_field_setting(field, 'type') == 'combined
'
1695 has_multiple_limits = has_limit and has_multiple_fields and not self._get_field_setting(field, 'same_limit
')
1697 fields = self._get_field_setting(field, 'field
') if has_multiple_fields else (field,)
1698 limits = limit_text.split(':') if has_multiple_limits else (limit_text,) if has_limit else tuple()
1699 limit_count = len(limits)
1700 for (i, f) in enumerate(fields):
1701 add_item(f, reverse, closest,
1702 limits[i] if i < limit_count
1703 else limits[0] if has_limit and not has_multiple_limits
1706 def print_verbose_info(self, write_debug):
1708 write_debug('Sort order given by user
: %s' % ', '.join(self._sort_user))
1709 if self._sort_extractor:
1710 write_debug('Sort order given by extractor
: %s' % ', '.join(self._sort_extractor))
1711 write_debug('Formats
sorted by
: %s' % ', '.join(['%s%s%s' % (
1712 '+' if self._get_field_setting(field, 'reverse
') else '', field,
1713 '%s%s(%s)' % ('~
' if self._get_field_setting(field, 'closest
') else ':',
1714 self._get_field_setting(field, 'limit_text
'),
1715 self._get_field_setting(field, 'limit
'))
1716 if self._get_field_setting(field, 'limit_text
') is not None else '')
1717 for field in self._order if self._get_field_setting(field, 'visible
')]))
1719 def _calculate_field_preference_from_value(self, format, field, type, value):
1720 reverse = self._get_field_setting(field, 'reverse
')
1721 closest = self._get_field_setting(field, 'closest
')
1722 limit = self._get_field_setting(field, 'limit
')
1724 if type == 'extractor
':
1725 maximum = self._get_field_setting(field, 'max')
1726 if value is None or (maximum is not None and value >= maximum):
1728 elif type == 'boolean
':
1729 in_list = self._get_field_setting(field, 'in_list
')
1730 not_in_list = self._get_field_setting(field, 'not_in_list
')
1731 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
1732 elif type == 'ordered
':
1733 value = self._resolve_field_value(field, value, True)
1735 # try to convert to number
1736 val_num = float_or_none(value, default=self._get_field_setting(field, 'default
'))
1737 is_num = self._get_field_setting(field, 'convert
') != 'string
' and val_num is not None
1741 return ((-10, 0) if value is None
1742 else (1, value, 0) if not is_num # if a field has mixed strings and numbers, strings are sorted higher
1743 else (0, -abs(value - limit), value - limit if reverse else limit - value) if closest
1744 else (0, value, 0) if not reverse and (limit is None or value <= limit)
1745 else (0, -value, 0) if limit is None or (reverse and value == limit) or value > limit
1746 else (-1, value, 0))
1748 def _calculate_field_preference(self, format, field):
1749 type = self._get_field_setting(field, 'type') # extractor, boolean, ordered, field, multiple
1750 get_value = lambda f: format.get(self._get_field_setting(f, 'field
'))
1751 if type == 'multiple
':
1752 type = 'field
' # Only 'field
' is allowed in multiple for now
1753 actual_fields = self._get_field_setting(field, 'field
')
1755 value = self._get_field_setting(field, 'function
')(get_value(f) for f in actual_fields)
1757 value = get_value(field)
1758 return self._calculate_field_preference_from_value(format, field, type, value)
1760 def calculate_preference(self, format):
1761 # Determine missing protocol
1762 if not format.get('protocol
'):
1763 format['protocol
'] = determine_protocol(format)
1765 # Determine missing ext
1766 if not format.get('ext
') and 'url
' in format:
1767 format['ext
'] = determine_ext(format['url
'])
1768 if format.get('vcodec
') == 'none
':
1769 format['audio_ext
'] = format['ext
'] if format.get('acodec
') != 'none
' else 'none
'
1770 format['video_ext
'] = 'none
'
1772 format['video_ext
'] = format['ext
']
1773 format['audio_ext
'] = 'none
'
1774 # if format.get('preference
') is None and format.get('ext
') in ('f4f
', 'f4m
'): # Not supported?
1775 # format['preference
'] = -1000
1777 # Determine missing bitrates
1778 if format.get('tbr
') is None:
1779 if format.get('vbr
') is not None and format.get('abr
') is not None:
1780 format['tbr
'] = format.get('vbr
', 0) + format.get('abr
', 0)
1782 if format.get('vcodec
') != 'none
' and format.get('vbr
') is None:
1783 format['vbr
'] = format.get('tbr
') - format.get('abr
', 0)
1784 if format.get('acodec
') != 'none
' and format.get('abr
') is None:
1785 format['abr
'] = format.get('tbr
') - format.get('vbr
', 0)
1787 return tuple(self._calculate_field_preference(format, field) for field in self._order)
1789 def _sort_formats(self, formats, field_preference=[]):
1792 format_sort = self.FormatSort() # params and to_screen are taken from the downloader
1793 format_sort.evaluate_params(self._downloader.params, field_preference)
1794 if self.get_param('verbose
', False):
1795 format_sort.print_verbose_info(self._downloader.write_debug)
1796 formats.sort(key=lambda f: format_sort.calculate_preference(f))
1798 def _check_formats(self, formats, video_id):
1800 formats[:] = filter(
1801 lambda f: self._is_valid_url(
1803 item='%s video format
' % f.get('format_id
') if f.get('format_id
') else 'video
'),
1807 def _remove_duplicate_formats(formats):
1811 if f['url
'] not in format_urls:
1812 format_urls.add(f['url
'])
1813 unique_formats.append(f)
1814 formats[:] = unique_formats
1816 def _is_valid_url(self, url, video_id, item='video
', headers={}):
1817 url = self._proto_relative_url(url, scheme='http
:')
1818 # For now assume non HTTP(S) URLs always valid
1819 if not (url.startswith('http
://') or url.startswith('https
://')):
1822 self._request_webpage(url, video_id, 'Checking
%s URL
' % item, headers=headers)
1824 except ExtractorError as e:
1826 '%s: %s URL
is invalid
, skipping
: %s'
1827 % (video_id, item, error_to_compat_str(e.cause)))
1830 def http_scheme(self):
1831 """ Either "http:" or "https:", depending on the user's preferences
"""
1834 if self.get_param('prefer_insecure', False)
1837 def _proto_relative_url(self, url, scheme=None):
1840 if url.startswith('//'):
1842 scheme = self.http_scheme()
1847 def _sleep(self, timeout, video_id, msg_template=None):
1848 if msg_template is None:
1849 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1850 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1854 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1855 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1856 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
1857 manifest = self._download_xml(
1858 manifest_url, video_id, 'Downloading f4m manifest',
1859 'Unable to download f4m manifest',
1860 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1861 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
1862 transform_source=transform_source,
1863 fatal=fatal, data=data, headers=headers, query=query)
1865 if manifest is False:
1868 return self._parse_f4m_formats(
1869 manifest, manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1870 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1872 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1873 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1874 fatal=True, m3u8_id=None):
1875 if not isinstance(manifest, compat_etree_Element) and not fatal:
1878 # currently yt-dlp cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1879 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1880 if akamai_pv is not None and ';' in akamai_pv.text:
1881 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1882 if playerVerificationChallenge.strip() != '':
1886 manifest_version = '1.0'
1887 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1889 manifest_version = '2.0'
1890 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1891 # Remove unsupported DRM protected media from final formats
1892 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
1893 media_nodes = remove_encrypted_media(media_nodes)
1897 manifest_base_url = get_base_url(manifest)
1899 bootstrap_info = xpath_element(
1900 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1901 'bootstrap info', default=None)
1904 mime_type = xpath_text(
1905 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1906 'base URL', default=None)
1907 if mime_type and mime_type.startswith('audio/'):
1910 for i, media_el in enumerate(media_nodes):
1911 tbr = int_or_none(media_el.attrib.get('bitrate'))
1912 width = int_or_none(media_el.attrib.get('width'))
1913 height = int_or_none(media_el.attrib.get('height'))
1914 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1915 # If <bootstrapInfo> is present, the specified f4m is a
1916 # stream-level manifest, and only set-level manifests may refer to
1917 # external resources. See section 11.4 and section 4 of F4M spec
1918 if bootstrap_info is None:
1920 # @href is introduced in 2.0, see section 11.6 of F4M spec
1921 if manifest_version == '2.0':
1922 media_url = media_el.attrib.get('href')
1923 if media_url is None:
1924 media_url = media_el.attrib.get('url')
1928 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1929 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1930 # If media_url is itself a f4m manifest do the recursive extraction
1931 # since bitrates in parent manifest (this one) and media_url manifest
1932 # may differ leading to inability to resolve the format by requested
1933 # bitrate in f4m downloader
1934 ext = determine_ext(manifest_url)
1936 f4m_formats = self._extract_f4m_formats(
1937 manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1938 transform_source=transform_source, fatal=fatal)
1939 # Sometimes stream-level manifest contains single media entry that
1940 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1941 # At the same time parent's media entry in set-level manifest may
1942 # contain it. We will copy it from parent in such cases.
1943 if len(f4m_formats) == 1:
1946 'tbr': f.get('tbr') or tbr,
1947 'width': f.get('width') or width,
1948 'height': f.get('height') or height,
1949 'format_id': f.get('format_id') if not tbr else format_id,
1952 formats.extend(f4m_formats)
1955 formats.extend(self._extract_m3u8_formats(
1956 manifest_url, video_id, 'mp4', preference=preference,
1957 quality=quality, m3u8_id=m3u8_id, fatal=fatal))
1960 'format_id': format_id,
1961 'url': manifest_url,
1962 'manifest_url': manifest_url,
1963 'ext': 'flv' if bootstrap_info is not None else None,
1969 'preference': preference,
1974 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, quality=None, m3u8_id=None):
1976 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1980 'preference': preference - 100 if preference else -100,
1982 'resolution': 'multiple',
1983 'format_note': 'Quality selection URL',
1986 def _report_ignoring_subs(self, name):
1987 self.report_warning(bug_reports_message(
1988 f'Ignoring subtitle tracks found in the {name} manifest; '
1989 'if any subtitle tracks are missing,'
1992 def _extract_m3u8_formats(self, *args, **kwargs):
1993 fmts, subs = self._extract_m3u8_formats_and_subtitles(*args, **kwargs)
1995 self._report_ignoring_subs('HLS')
1998 def _extract_m3u8_formats_and_subtitles(
1999 self, m3u8_url, video_id, ext=None, entry_protocol='m3u8_native',
2000 preference=None, quality=None, m3u8_id=None, note=None,
2001 errnote=None, fatal=True, live=False, data=None, headers={},
2004 res = self._download_webpage_handle(
2006 note='Downloading m3u8 information' if note is None else note,
2007 errnote='Failed to download m3u8 information' if errnote is None else errnote,
2008 fatal=fatal, data=data, headers=headers, query=query)
2013 m3u8_doc, urlh = res
2014 m3u8_url = urlh.geturl()
2016 return self._parse_m3u8_formats_and_subtitles(
2017 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
2018 preference=preference, quality=quality, m3u8_id=m3u8_id,
2019 note=note, errnote=errnote, fatal=fatal, live=live, data=data,
2020 headers=headers, query=query, video_id=video_id)
2022 def _parse_m3u8_formats_and_subtitles(
2023 self, m3u8_doc, m3u8_url, ext=None, entry_protocol='m3u8_native',
2024 preference=None, quality=None, m3u8_id=None, live=False, note=None,
2025 errnote=None, fatal=True, data=None, headers={}, query={},
2027 formats, subtitles = [], {}
2029 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
2030 return formats, subtitles
2032 has_drm = re.search(r'#EXT-X-(?:SESSION-)?KEY:.*?URI="skd://', m3u8_doc)
2034 def format_url(url):
2035 return url if re.match(r'^https?://', url) else compat_urlparse.urljoin(m3u8_url, url)
2037 if self.get_param('hls_split_discontinuity', False):
2038 def _extract_m3u8_playlist_indices(manifest_url=None, m3u8_doc=None):
2040 if not manifest_url:
2042 m3u8_doc = self._download_webpage(
2043 manifest_url, video_id, fatal=fatal, data=data, headers=headers,
2044 note=False, errnote='Failed to download m3u8 playlist information')
2045 if m3u8_doc is False:
2047 return range(1 + sum(line.startswith('#EXT-X-DISCONTINUITY') for line in m3u8_doc.splitlines()))
2050 def _extract_m3u8_playlist_indices(*args, **kwargs):
2054 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
2055 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
2056 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
2058 # We should try extracting formats only from master playlists [1, 4.3.4],
2059 # i.e. playlists that describe available qualities. On the other hand
2060 # media playlists [1, 4.3.3] should be returned as is since they contain
2061 # just the media without qualities renditions.
2062 # Fortunately, master playlist can be easily distinguished from media
2063 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
2064 # master playlist tags MUST NOT appear in a media playlist and vice versa.
2065 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
2066 # media playlist and MUST NOT appear in master playlist thus we can
2067 # clearly detect media playlist with this criterion.
2069 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
2071 'format_id': '-'.join(map(str, filter(None, [m3u8_id, idx]))),
2072 'format_index': idx,
2075 'protocol': entry_protocol,
2076 'preference': preference,
2079 } for idx in _extract_m3u8_playlist_indices(m3u8_doc=m3u8_doc)]
2081 return formats, subtitles
2084 last_stream_inf = {}
2086 def extract_media(x_media_line):
2087 media = parse_m3u8_attributes(x_media_line)
2088 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
2089 media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME')
2090 if not (media_type and group_id and name):
2092 groups.setdefault(group_id, []).append(media)
2093 # <https://tools.ietf.org/html/rfc8216#section-4.3.4.1>
2094 if media_type == 'SUBTITLES':
2095 # According to RFC 8216 §4.3.4.2.1, URI is REQUIRED in the
2096 # EXT-X-MEDIA tag if the media type is SUBTITLES.
2097 # However, lack of URI has been spotted in the wild.
2098 # e.g. NebulaIE; see https://github.com/yt-dlp/yt-dlp/issues/339
2099 if not media.get('URI'):
2101 url = format_url(media['URI'])
2104 'ext': determine_ext(url),
2106 if sub_info['ext'] == 'm3u8':
2107 # Per RFC 8216 §3.1, the only possible subtitle format m3u8
2108 # files may contain is WebVTT:
2109 # <https://tools.ietf.org/html/rfc8216#section-3.1>
2110 sub_info['ext'] = 'vtt'
2111 sub_info['protocol'] = 'm3u8_native'
2112 lang = media.get('LANGUAGE') or 'und'
2113 subtitles.setdefault(lang, []).append(sub_info)
2114 if media_type not in ('VIDEO', 'AUDIO'):
2116 media_url = media.get('URI')
2118 manifest_url = format_url(media_url)
2120 'format_id': '-'.join(map(str, filter(None, (m3u8_id, group_id, name, idx)))),
2121 'format_note': name,
2122 'format_index': idx,
2123 'url': manifest_url,
2124 'manifest_url': m3u8_url,
2125 'language': media.get('LANGUAGE'),
2127 'protocol': entry_protocol,
2128 'preference': preference,
2130 'vcodec': 'none' if media_type == 'AUDIO' else None,
2131 } for idx in _extract_m3u8_playlist_indices(manifest_url))
2133 def build_stream_name():
2134 # Despite specification does not mention NAME attribute for
2135 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
2136 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
2137 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
2138 stream_name = last_stream_inf.get('NAME')
2141 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
2142 # from corresponding rendition group
2143 stream_group_id = last_stream_inf.get('VIDEO')
2144 if not stream_group_id:
2146 stream_group = groups.get(stream_group_id)
2147 if not stream_group:
2148 return stream_group_id
2149 rendition = stream_group[0]
2150 return rendition.get('NAME') or stream_group_id
2152 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
2153 # chance to detect video only formats when EXT-X-STREAM-INF tags
2154 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
2155 for line in m3u8_doc.splitlines():
2156 if line.startswith('#EXT-X-MEDIA:'):
2159 for line in m3u8_doc.splitlines():
2160 if line.startswith('#EXT-X-STREAM-INF:'):
2161 last_stream_inf = parse_m3u8_attributes(line)
2162 elif line.startswith('#') or not line.strip():
2165 tbr = float_or_none(
2166 last_stream_inf.get('AVERAGE-BANDWIDTH')
2167 or last_stream_inf.get('BANDWIDTH'), scale=1000)
2168 manifest_url = format_url(line.strip())
2170 for idx in _extract_m3u8_playlist_indices(manifest_url):
2171 format_id = [m3u8_id, None, idx]
2172 # Bandwidth of live streams may differ over time thus making
2173 # format_id unpredictable. So it's better to keep provided
2176 stream_name = build_stream_name()
2177 format_id[1] = stream_name if stream_name else '%d' % (tbr if tbr else len(formats))
2179 'format_id': '-'.join(map(str, filter(None, format_id))),
2180 'format_index': idx,
2181 'url': manifest_url,
2182 'manifest_url': m3u8_url,
2185 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')),
2186 'protocol': entry_protocol,
2187 'preference': preference,
2190 resolution = last_stream_inf.get('RESOLUTION')
2192 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
2194 f['width'] = int(mobj.group('width'))
2195 f['height'] = int(mobj.group('height'))
2196 # Unified Streaming Platform
2198 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
2200 abr, vbr = mobj.groups()
2201 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
2206 codecs = parse_codecs(last_stream_inf.get('CODECS'))
2208 audio_group_id = last_stream_inf.get('AUDIO')
2209 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
2210 # references a rendition group MUST have a CODECS attribute.
2211 # However, this is not always respected, for example, [2]
2212 # contains EXT-X-STREAM-INF tag which references AUDIO
2213 # rendition group but does not have CODECS and despite
2214 # referencing an audio group it represents a complete
2215 # (with audio and video) format. So, for such cases we will
2216 # ignore references to rendition groups and treat them
2217 # as complete formats.
2218 if audio_group_id and codecs and f.get('vcodec') != 'none':
2219 audio_group = groups.get(audio_group_id)
2220 if audio_group and audio_group[0].get('URI'):
2221 # TODO: update acodec for audio only formats with
2223 f['acodec'] = 'none'
2224 if not f.get('ext'):
2225 f['ext'] = 'm4a' if f.get('vcodec') == 'none' else 'mp4'
2229 progressive_uri = last_stream_inf.get('PROGRESSIVE-URI')
2232 del http_f['manifest_url']
2234 'format_id': f['format_id'].replace('hls-', 'http-'),
2236 'url': progressive_uri,
2238 formats.append(http_f)
2240 last_stream_inf = {}
2241 return formats, subtitles
2243 def _extract_m3u8_vod_duration(
2244 self, m3u8_vod_url, video_id, note=None, errnote=None, data=None, headers={}, query={}):
2246 m3u8_vod = self._download_webpage(
2247 m3u8_vod_url, video_id,
2248 note='Downloading m3u8 VOD manifest' if note is None else note,
2249 errnote='Failed to download VOD manifest' if errnote is None else errnote,
2250 fatal=False, data=data, headers=headers, query=query)
2252 return self._parse_m3u8_vod_duration(m3u8_vod or '', video_id)
2254 def _parse_m3u8_vod_duration(self, m3u8_vod, video_id):
2255 if '#EXT-X-PLAYLIST-TYPE:VOD' not in m3u8_vod:
2259 float(line[len('#EXTINF:'):].split(',')[0])
2260 for line in m3u8_vod.splitlines() if line.startswith('#EXTINF:'))) or None
2263 def _xpath_ns(path, namespace=None):
2267 for c in path.split('/'):
2268 if not c or c == '.':
2271 out.append('{%s}%s' % (namespace, c))
2272 return '/'.join(out)
2274 def _extract_smil_formats_and_subtitles(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
2275 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
2281 namespace = self._parse_smil_namespace(smil)
2283 fmts = self._parse_smil_formats(
2284 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
2285 subs = self._parse_smil_subtitles(
2286 smil, namespace=namespace)
2290 def _extract_smil_formats(self, *args, **kwargs):
2291 fmts, subs = self._extract_smil_formats_and_subtitles(*args, **kwargs)
2293 self._report_ignoring_subs('SMIL')
2296 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
2297 smil = self._download_smil(smil_url, video_id, fatal=fatal)
2300 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
2302 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
2303 return self._download_xml(
2304 smil_url, video_id, 'Downloading SMIL file',
2305 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
2307 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
2308 namespace = self._parse_smil_namespace(smil)
2310 formats = self._parse_smil_formats(
2311 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
2312 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
2314 video_id = os.path.splitext(url_basename(smil_url))[0]
2318 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
2319 name = meta.attrib.get('name')
2320 content = meta.attrib.get('content')
2321 if not name or not content:
2323 if not title and name == 'title':
2325 elif not description and name in ('description', 'abstract'):
2326 description = content
2327 elif not upload_date and name == 'date':
2328 upload_date = unified_strdate(content)
2331 'id': image.get('type'),
2332 'url': image.get('src'),
2333 'width': int_or_none(image.get('width')),
2334 'height': int_or_none(image.get('height')),
2335 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
2339 'title': title or video_id,
2340 'description': description,
2341 'upload_date': upload_date,
2342 'thumbnails': thumbnails,
2344 'subtitles': subtitles,
2347 def _parse_smil_namespace(self, smil):
2348 return self._search_regex(
2349 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
2351 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
2353 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
2354 b = meta.get('base') or meta.get('httpBase')
2366 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
2367 for medium in media:
2368 src = medium.get('src')
2369 if not src or src in srcs:
2373 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
2374 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
2375 width = int_or_none(medium.get('width'))
2376 height = int_or_none(medium.get('height'))
2377 proto = medium.get('proto')
2378 ext = medium.get('ext')
2379 src_ext = determine_ext(src)
2380 streamer = medium.get('streamer') or base
2382 if proto == 'rtmp' or streamer.startswith('rtmp'):
2388 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
2390 'filesize': filesize,
2394 if transform_rtmp_url:
2395 streamer, src = transform_rtmp_url(streamer, src)
2396 formats[-1].update({
2402 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
2403 src_url = src_url.strip()
2405 if proto == 'm3u8' or src_ext == 'm3u8':
2406 m3u8_formats = self._extract_m3u8_formats(
2407 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
2408 if len(m3u8_formats) == 1:
2410 m3u8_formats[0].update({
2411 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
2416 formats.extend(m3u8_formats)
2417 elif src_ext == 'f4m':
2422 'plugin': 'flowplayer-3.2.0.1',
2424 f4m_url += '&' if '?' in f4m_url else '?'
2425 f4m_url += compat_urllib_parse_urlencode(f4m_params)
2426 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
2427 elif src_ext == 'mpd':
2428 formats.extend(self._extract_mpd_formats(
2429 src_url, video_id, mpd_id='dash', fatal=False))
2430 elif re.search(r'\.ism/[Mm]anifest', src_url):
2431 formats.extend(self._extract_ism_formats(
2432 src_url, video_id, ism_id='mss', fatal=False))
2433 elif src_url.startswith('http') and self._is_valid_url(src, video_id):
2437 'ext': ext or src_ext or 'flv',
2438 'format_id': 'http-%d' % (bitrate or http_count),
2440 'filesize': filesize,
2445 for medium in smil.findall(self._xpath_ns('.//imagestream', namespace)):
2446 src = medium.get('src')
2447 if not src or src in srcs:
2453 'format_id': 'imagestream-%d' % (imgs_count),
2455 'ext': mimetype2ext(medium.get('type')),
2458 'width': int_or_none(medium.get('width')),
2459 'height': int_or_none(medium.get('height')),
2460 'format_note': 'SMIL storyboards',
2465 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
2468 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
2469 src = textstream.get('src')
2470 if not src or src in urls:
2473 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
2474 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
2475 subtitles.setdefault(lang, []).append({
2481 def _extract_xspf_playlist(self, xspf_url, playlist_id, fatal=True):
2482 xspf = self._download_xml(
2483 xspf_url, playlist_id, 'Downloading xpsf playlist',
2484 'Unable to download xspf manifest', fatal=fatal)
2487 return self._parse_xspf(
2488 xspf, playlist_id, xspf_url=xspf_url,
2489 xspf_base_url=base_url(xspf_url))
2491 def _parse_xspf(self, xspf_doc, playlist_id, xspf_url=None, xspf_base_url=None):
2493 'xspf': 'http://xspf.org/ns/0/',
2494 's1': 'http://static.streamone.nl/player/ns/0',
2498 for track in xspf_doc.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
2500 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
2501 description = xpath_text(
2502 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
2503 thumbnail = xpath_text(
2504 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
2505 duration = float_or_none(
2506 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
2509 for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP)):
2510 format_url = urljoin(xspf_base_url, location.text)
2515 'manifest_url': xspf_url,
2516 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
2517 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
2518 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
2520 self._sort_formats(formats)
2525 'description': description,
2526 'thumbnail': thumbnail,
2527 'duration': duration,
2532 def _extract_mpd_formats(self, *args, **kwargs):
2533 fmts, subs = self._extract_mpd_formats_and_subtitles(*args, **kwargs)
2535 self._report_ignoring_subs('DASH')
2538 def _extract_mpd_formats_and_subtitles(
2539 self, mpd_url, video_id, mpd_id=None, note=None, errnote=None,
2540 fatal=True, data=None, headers={}, query={}):
2541 res = self._download_xml_handle(
2543 note='Downloading MPD manifest' if note is None else note,
2544 errnote='Failed to download MPD manifest' if errnote is None else errnote,
2545 fatal=fatal, data=data, headers=headers, query=query)
2551 mpd_base_url = base_url(urlh.geturl())
2553 return self._parse_mpd_formats_and_subtitles(
2554 mpd_doc, mpd_id, mpd_base_url, mpd_url)
2556 def _parse_mpd_formats(self, *args, **kwargs):
2557 fmts, subs = self._parse_mpd_formats_and_subtitles(*args, **kwargs)
2559 self._report_ignoring_subs('DASH')
2562 def _parse_mpd_formats_and_subtitles(
2563 self, mpd_doc, mpd_id=None, mpd_base_url='', mpd_url=None):
2565 Parse formats
from MPD manifest
.
2567 1. MPEG
-DASH Standard
, ISO
/IEC
23009-1:2014(E
),
2568 http
://standards
.iso
.org
/ittf
/PubliclyAvailableStandards
/c065274_ISO_IEC_23009
-1_2014.zip
2569 2. https
://en
.wikipedia
.org
/wiki
/Dynamic_Adaptive_Streaming_over_HTTP
2571 if not self.get_param('dynamic_mpd', True):
2572 if mpd_doc.get('type') == 'dynamic':
2575 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
2578 return self._xpath_ns(path, namespace)
2580 def is_drm_protected(element):
2581 return element.find(_add_ns('ContentProtection')) is not None
2583 def extract_multisegment_info(element, ms_parent_info):
2584 ms_info = ms_parent_info.copy()
2586 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2587 # common attributes and elements. We will only extract relevant
2589 def extract_common(source):
2590 segment_timeline = source.find(_add_ns('SegmentTimeline'))
2591 if segment_timeline is not None:
2592 s_e = segment_timeline.findall(_add_ns('S'))
2594 ms_info['total_number'] = 0
2597 r = int(s.get('r', 0))
2598 ms_info['total_number'] += 1 + r
2599 ms_info['s'].append({
2600 't': int(s.get('t', 0)),
2601 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2602 'd': int(s.attrib['d']),
2605 start_number = source.get('startNumber')
2607 ms_info['start_number'] = int(start_number)
2608 timescale = source.get('timescale')
2610 ms_info['timescale'] = int(timescale)
2611 segment_duration = source.get('duration')
2612 if segment_duration:
2613 ms_info['segment_duration'] = float(segment_duration)
2615 def extract_Initialization(source):
2616 initialization = source.find(_add_ns('Initialization'))
2617 if initialization is not None:
2618 ms_info['initialization_url'] = initialization.attrib['sourceURL']
2620 segment_list = element.find(_add_ns('SegmentList'))
2621 if segment_list is not None:
2622 extract_common(segment_list)
2623 extract_Initialization(segment_list)
2624 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
2626 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
2628 segment_template = element.find(_add_ns('SegmentTemplate'))
2629 if segment_template is not None:
2630 extract_common(segment_template)
2631 media = segment_template.get('media')
2633 ms_info['media'] = media
2634 initialization = segment_template.get('initialization')
2636 ms_info['initialization'] = initialization
2638 extract_Initialization(segment_template)
2641 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
2642 formats, subtitles = [], {}
2643 stream_numbers = {'audio': 0, 'video': 0}
2644 for period in mpd_doc.findall(_add_ns('Period')):
2645 period_duration = parse_duration(period.get('duration')) or mpd_duration
2646 period_ms_info = extract_multisegment_info(period, {
2650 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
2651 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
2652 for representation in adaptation_set.findall(_add_ns('Representation')):
2653 representation_attrib = adaptation_set.attrib.copy()
2654 representation_attrib.update(representation.attrib)
2655 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2656 mime_type = representation_attrib['mimeType']
2657 content_type = representation_attrib.get('contentType', mime_type.split('/')[0])
2659 codecs = representation_attrib.get('codecs', '')
2660 if content_type not in ('video', 'audio', 'text'):
2661 if mime_type == 'image/jpeg':
2662 content_type = mime_type
2663 elif codecs.split('.')[0] == 'stpp':
2664 content_type = 'text'
2665 elif mimetype2ext(mime_type) in ('tt', 'dfxp', 'ttml', 'xml', 'json'):
2666 content_type = 'text'
2668 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
2672 for element in (representation, adaptation_set, period, mpd_doc):
2673 base_url_e = element.find(_add_ns('BaseURL'))
2674 if base_url_e is not None:
2675 base_url = base_url_e.text + base_url
2676 if re.match(r'^https?://', base_url):
2678 if mpd_base_url and base_url.startswith('/'):
2679 base_url = compat_urlparse.urljoin(mpd_base_url, base_url)
2680 elif mpd_base_url and not re.match(r'^https?://', base_url):
2681 if not mpd_base_url.endswith('/'):
2683 base_url = mpd_base_url + base_url
2684 representation_id = representation_attrib.get('id')
2685 lang = representation_attrib.get('lang')
2686 url_el = representation.find(_add_ns('BaseURL'))
2687 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
2688 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
2689 if representation_id is not None:
2690 format_id = representation_id
2692 format_id = content_type
2694 format_id = mpd_id + '-' + format_id
2695 if content_type in ('video', 'audio'):
2697 'format_id': format_id,
2698 'manifest_url': mpd_url,
2699 'ext': mimetype2ext(mime_type),
2700 'width': int_or_none(representation_attrib.get('width')),
2701 'height': int_or_none(representation_attrib.get('height')),
2702 'tbr': float_or_none(bandwidth, 1000),
2703 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
2704 'fps': int_or_none(representation_attrib.get('frameRate')),
2705 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
2706 'format_note': 'DASH %s' % content_type,
2707 'filesize': filesize,
2708 'container': mimetype2ext(mime_type) + '_dash',
2709 'manifest_stream_number': stream_numbers[content_type]
2711 f.update(parse_codecs(codecs))
2712 stream_numbers[content_type] += 1
2713 elif content_type == 'text':
2715 'ext': mimetype2ext(mime_type),
2716 'manifest_url': mpd_url,
2717 'filesize': filesize,
2719 elif content_type == 'image/jpeg':
2720 # See test case in VikiIE
2721 # https://www.viki.com/videos/1175236v-choosing-spouse-by-lottery-episode-1
2723 'format_id': format_id,
2725 'manifest_url': mpd_url,
2726 'format_note': 'DASH storyboards (jpeg)',
2730 if is_drm_protected(adaptation_set) or is_drm_protected(representation):
2732 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
2734 def prepare_template(template_name, identifiers):
2735 tmpl = representation_ms_info[template_name]
2736 # First of, % characters outside $...$ templates
2737 # must be escaped by doubling for proper processing
2738 # by % operator string formatting used further (see
2739 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2745 in_template = not in_template
2746 elif c == '%' and not in_template:
2748 # Next, $...$ templates are translated to their
2749 # %(...) counterparts to be used with % operator
2750 if representation_id is not None:
2751 t = t.replace('$RepresentationID$', representation_id)
2752 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
2753 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
2754 t.replace('$$', '$')
2757 # @initialization is a regular template like @media one
2758 # so it should be handled just the same way (see
2759 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2760 if 'initialization' in representation_ms_info:
2761 initialization_template = prepare_template(
2763 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2764 # $Time$ shall not be included for @initialization thus
2765 # only $Bandwidth$ remains
2767 representation_ms_info['initialization_url'] = initialization_template % {
2768 'Bandwidth': bandwidth,
2771 def location_key(location):
2772 return 'url' if re.match(r'^https?://', location) else 'path'
2774 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
2776 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2777 media_location_key = location_key(media_template)
2779 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2780 # can't be used at the same time
2781 if '%(Number' in media_template and 's' not in representation_ms_info:
2782 segment_duration = None
2783 if 'total_number' not in representation_ms_info and 'segment_duration' in representation_ms_info:
2784 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
2785 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
2786 representation_ms_info['fragments'] = [{
2787 media_location_key: media_template % {
2788 'Number': segment_number,
2789 'Bandwidth': bandwidth,
2791 'duration': segment_duration,
2792 } for segment_number in range(
2793 representation_ms_info['start_number'],
2794 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
2796 # $Number*$ or $Time$ in media template with S list available
2797 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2798 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2799 representation_ms_info['fragments'] = []
2802 segment_number = representation_ms_info['start_number']
2804 def add_segment_url():
2805 segment_url = media_template % {
2806 'Time': segment_time,
2807 'Bandwidth': bandwidth,
2808 'Number': segment_number,
2810 representation_ms_info['fragments'].append({
2811 media_location_key: segment_url,
2812 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
2815 for num, s in enumerate(representation_ms_info['s']):
2816 segment_time = s.get('t') or segment_time
2820 for r in range(s.get('r', 0)):
2821 segment_time += segment_d
2824 segment_time += segment_d
2825 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
2827 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2828 # or any YouTube dashsegments video
2831 timescale = representation_ms_info['timescale']
2832 for s in representation_ms_info['s']:
2833 duration = float_or_none(s['d'], timescale)
2834 for r in range(s.get('r', 0) + 1):
2835 segment_uri = representation_ms_info['segment_urls'][segment_index]
2837 location_key(segment_uri): segment_uri,
2838 'duration': duration,
2841 representation_ms_info['fragments'] = fragments
2842 elif 'segment_urls' in representation_ms_info:
2843 # Segment URLs with no SegmentTimeline
2844 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2845 # https://github.com/ytdl-org/youtube-dl/pull/14844
2847 segment_duration = float_or_none(
2848 representation_ms_info['segment_duration'],
2849 representation_ms_info['timescale']) if 'segment_duration' in representation_ms_info else None
2850 for segment_url in representation_ms_info['segment_urls']:
2852 location_key(segment_url): segment_url,
2854 if segment_duration:
2855 fragment['duration'] = segment_duration
2856 fragments.append(fragment)
2857 representation_ms_info['fragments'] = fragments
2858 # If there is a fragments key available then we correctly recognized fragmented media.
2859 # Otherwise we will assume unfragmented media with direct access. Technically, such
2860 # assumption is not necessarily correct since we may simply have no support for
2861 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2862 if 'fragments' in representation_ms_info:
2864 # NB: mpd_url may be empty when MPD manifest is parsed from a string
2865 'url': mpd_url or base_url,
2866 'fragment_base_url': base_url,
2868 'protocol': 'http_dash_segments' if mime_type != 'image/jpeg' else 'mhtml',
2870 if 'initialization_url' in representation_ms_info:
2871 initialization_url = representation_ms_info['initialization_url']
2872 if not f.get('url'):
2873 f['url'] = initialization_url
2874 f['fragments'].append({location_key(initialization_url): initialization_url})
2875 f['fragments'].extend(representation_ms_info['fragments'])
2877 # Assuming direct URL to unfragmented media.
2879 if content_type in ('video', 'audio') or mime_type == 'image/jpeg':
2881 elif content_type == 'text':
2882 subtitles.setdefault(lang or 'und', []).append(f)
2884 return formats, subtitles
2886 def _extract_ism_formats(self, *args, **kwargs):
2887 fmts, subs = self._extract_ism_formats_and_subtitles(*args, **kwargs)
2889 self._report_ignoring_subs('ISM')
2892 def _extract_ism_formats_and_subtitles(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
2893 res = self._download_xml_handle(
2895 note='Downloading ISM manifest' if note is None else note,
2896 errnote='Failed to download ISM manifest' if errnote is None else errnote,
2897 fatal=fatal, data=data, headers=headers, query=query)
2904 return self._parse_ism_formats_and_subtitles(ism_doc, urlh.geturl(), ism_id)
2906 def _parse_ism_formats_and_subtitles(self, ism_doc, ism_url, ism_id=None):
2908 Parse formats
from ISM manifest
.
2910 1. [MS
-SSTR
]: Smooth Streaming Protocol
,
2911 https
://msdn
.microsoft
.com
/en
-us
/library
/ff469518
.aspx
2913 if ism_doc.get('IsLive') == 'TRUE':
2916 duration = int(ism_doc.attrib['Duration'])
2917 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
2921 for stream in ism_doc.findall('StreamIndex'):
2922 stream_type = stream.get('Type')
2923 if stream_type not in ('video', 'audio', 'text'):
2925 url_pattern = stream.attrib['Url']
2926 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
2927 stream_name = stream.get('Name')
2928 stream_language = stream.get('Language', 'und')
2929 for track in stream.findall('QualityLevel'):
2930 fourcc = track.get('FourCC') or ('AACL' if track.get('AudioTag') == '255' else None)
2931 # TODO: add support for WVC1 and WMAP
2932 if fourcc not in ('H264', 'AVC1', 'AACL', 'TTML'):
2933 self.report_warning('%s is not a supported codec' % fourcc)
2935 tbr = int(track.attrib['Bitrate']) // 1000
2936 # [1] does not mention Width and Height attributes. However,
2937 # they're often present while MaxWidth and MaxHeight are
2938 # missing, so should be used as fallbacks
2939 width = int_or_none(track.get('MaxWidth') or track.get('Width'))
2940 height = int_or_none(track.get('MaxHeight') or track.get('Height'))
2941 sampling_rate = int_or_none(track.get('SamplingRate'))
2943 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
2944 track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
2950 stream_fragments = stream.findall('c')
2951 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
2952 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
2953 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
2954 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
2955 if not fragment_ctx['duration']:
2957 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
2959 next_fragment_time = duration
2960 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
2961 for _ in range(fragment_repeat):
2963 'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
2964 'duration': fragment_ctx['duration'] / stream_timescale,
2966 fragment_ctx['time'] += fragment_ctx['duration']
2970 format_id.append(ism_id)
2972 format_id.append(stream_name)
2973 format_id.append(compat_str(tbr))
2975 if stream_type == 'text':
2976 subtitles.setdefault(stream_language, []).append({
2980 'manifest_url': ism_url,
2981 'fragments': fragments,
2982 '_download_params': {
2983 'stream_type': stream_type,
2984 'duration': duration,
2985 'timescale': stream_timescale,
2987 'language': stream_language,
2988 'codec_private_data': track.get('CodecPrivateData'),
2991 elif stream_type in ('video', 'audio'):
2993 'format_id': '-'.join(format_id),
2995 'manifest_url': ism_url,
2996 'ext': 'ismv' if stream_type == 'video' else 'isma',
3000 'asr': sampling_rate,
3001 'vcodec': 'none' if stream_type == 'audio' else fourcc,
3002 'acodec': 'none' if stream_type == 'video' else fourcc,
3004 'fragments': fragments,
3005 'has_drm': ism_doc.find('Protection') is not None,
3006 '_download_params': {
3007 'stream_type': stream_type,
3008 'duration': duration,
3009 'timescale': stream_timescale,
3010 'width': width or 0,
3011 'height': height or 0,
3013 'language': stream_language,
3014 'codec_private_data': track.get('CodecPrivateData'),
3015 'sampling_rate': sampling_rate,
3016 'channels': int_or_none(track.get('Channels', 2)),
3017 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
3018 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
3021 return formats, subtitles
3023 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):
3024 def absolute_url(item_url):
3025 return urljoin(base_url, item_url)
3027 def parse_content_type(content_type):
3028 if not content_type:
3030 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
3032 mimetype, codecs = ctr.groups()
3033 f = parse_codecs(codecs)
3034 f['ext'] = mimetype2ext(mimetype)
3038 def _media_formats(src, cur_media_type, type_info={}):
3039 full_url = absolute_url(src)
3040 ext = type_info.get('ext') or determine_ext(full_url)
3042 is_plain_url = False
3043 formats = self._extract_m3u8_formats(
3044 full_url, video_id, ext='mp4',
3045 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
3046 preference=preference, quality=quality, fatal=False)
3048 is_plain_url = False
3049 formats = self._extract_mpd_formats(
3050 full_url, video_id, mpd_id=mpd_id, fatal=False)
3055 'vcodec': 'none' if cur_media_type == 'audio' else None,
3057 return is_plain_url, formats
3060 # amp-video and amp-audio are very similar to their HTML5 counterparts
3061 # so we wll include them right here (see
3062 # https://www.ampproject.org/docs/reference/components/amp-video)
3063 # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/
3064 _MEDIA_TAG_NAME_RE = r'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)'
3065 media_tags = [(media_tag, media_tag_name, media_type, '')
3066 for media_tag, media_tag_name, media_type
3067 in re.findall(r'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE, webpage)]
3068 media_tags.extend(re.findall(
3069 # We only allow video|audio followed by a whitespace or '>'.
3070 # Allowing more characters may end up in significant slow down (see
3071 # https://github.com/ytdl-org/youtube-dl/issues/11979, example URL:
3072 # http://www.porntrex.com/maps/videositemap.xml).
3073 r'(?s)(<(?P<tag>%s)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>' % _MEDIA_TAG_NAME_RE, webpage))
3074 for media_tag, _, media_type, media_content in media_tags:
3079 media_attributes = extract_attributes(media_tag)
3080 src = strip_or_none(media_attributes.get('src'))
3082 _, formats = _media_formats(src, media_type)
3083 media_info['formats'].extend(formats)
3084 media_info['thumbnail'] = absolute_url(media_attributes.get('poster'))
3086 for source_tag in re.findall(r'<source[^>]+>', media_content):
3087 s_attr = extract_attributes(source_tag)
3088 # data-video-src and data-src are non standard but seen
3089 # several times in the wild
3090 src = strip_or_none(dict_get(s_attr, ('src', 'data-video-src', 'data-src')))
3093 f = parse_content_type(s_attr.get('type'))
3094 is_plain_url, formats = _media_formats(src, media_type, f)
3096 # width, height, res, label and title attributes are
3097 # all not standard but seen several times in the wild
3100 for lbl in ('label', 'title')
3101 if str_or_none(s_attr.get(lbl))
3103 width = int_or_none(s_attr.get('width'))
3104 height = (int_or_none(s_attr.get('height'))
3105 or int_or_none(s_attr.get('res')))
3106 if not width or not height:
3108 resolution = parse_resolution(lbl)
3111 width = width or resolution.get('width')
3112 height = height or resolution.get('height')
3114 tbr = parse_bitrate(lbl)
3123 'format_id': s_attr.get('label') or s_attr.get('title'),
3125 f.update(formats[0])
3126 media_info['formats'].append(f)
3128 media_info['formats'].extend(formats)
3129 for track_tag in re.findall(r'<track[^>]+>', media_content):
3130 track_attributes = extract_attributes(track_tag)
3131 kind = track_attributes.get('kind')
3132 if not kind or kind in ('subtitles', 'captions'):
3133 src = strip_or_none(track_attributes.get('src'))
3136 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
3137 media_info['subtitles'].setdefault(lang, []).append({
3138 'url': absolute_url(src),
3140 for f in media_info['formats']:
3141 f.setdefault('http_headers', {})['Referer'] = base_url
3142 if media_info['formats'] or media_info['subtitles']:
3143 entries.append(media_info)
3146 def _extract_akamai_formats(self, *args, **kwargs):
3147 fmts, subs = self._extract_akamai_formats_and_subtitles(*args, **kwargs)
3149 self._report_ignoring_subs('akamai')
3152 def _extract_akamai_formats_and_subtitles(self, manifest_url, video_id, hosts={}):
3153 signed = 'hdnea=' in manifest_url
3155 # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
3156 manifest_url = re.sub(
3157 r'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
3158 '', manifest_url).strip('?')
3163 hdcore_sign = 'hdcore=3.7.0'
3164 f4m_url = re.sub(r'(https?://[^/]+)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
3165 hds_host = hosts.get('hds')
3167 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
3168 if 'hdcore=' not in f4m_url:
3169 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
3170 f4m_formats = self._extract_f4m_formats(
3171 f4m_url, video_id, f4m_id='hds', fatal=False)
3172 for entry in f4m_formats:
3173 entry.update({'extra_param_to_segment_url': hdcore_sign})
3174 formats.extend(f4m_formats)
3176 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
3177 hls_host = hosts.get('hls')
3179 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
3180 m3u8_formats, m3u8_subtitles = self._extract_m3u8_formats_and_subtitles(
3181 m3u8_url, video_id, 'mp4', 'm3u8_native',
3182 m3u8_id='hls', fatal=False)
3183 formats.extend(m3u8_formats)
3184 subtitles = self._merge_subtitles(subtitles, m3u8_subtitles)
3186 http_host = hosts.get('http')
3187 if http_host and m3u8_formats and not signed:
3188 REPL_REGEX = r'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
3189 qualities = re.match(REPL_REGEX, m3u8_url).group(2).split(',')
3190 qualities_length = len(qualities)
3191 if len(m3u8_formats) in (qualities_length, qualities_length + 1):
3193 for f in m3u8_formats:
3194 if f['vcodec'] != 'none':
3195 for protocol in ('http', 'https'):
3197 del http_f['manifest_url']
3199 REPL_REGEX, protocol + r'://%s/\g<1>%s\3' % (http_host, qualities[i]), f['url'])
3201 'format_id': http_f['format_id'].replace('hls-', protocol + '-'),
3203 'protocol': protocol,
3205 formats.append(http_f)
3208 return formats, subtitles
3210 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
3211 query = compat_urlparse.urlparse(url).query
3212 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
3214 r'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url)
3215 url_base = mobj.group('url')
3216 http_base_url = '%s%s:%s' % ('http', mobj.group('s') or '', url_base)
3219 def manifest_url(manifest):
3220 m_url = '%s/%s' % (http_base_url, manifest)
3222 m_url += '?%s' % query
3225 if 'm3u8' not in skip_protocols:
3226 formats.extend(self._extract_m3u8_formats(
3227 manifest_url('playlist.m3u8'), video_id, 'mp4',
3228 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
3229 if 'f4m' not in skip_protocols:
3230 formats.extend(self._extract_f4m_formats(
3231 manifest_url('manifest.f4m'),
3232 video_id, f4m_id='hds', fatal=False))
3233 if 'dash' not in skip_protocols:
3234 formats.extend(self._extract_mpd_formats(
3235 manifest_url('manifest.mpd'),
3236 video_id, mpd_id='dash', fatal=False))
3237 if re.search(r'(?:/smil:|\.smil)', url_base):
3238 if 'smil' not in skip_protocols:
3239 rtmp_formats = self._extract_smil_formats(
3240 manifest_url('jwplayer.smil'),
3241 video_id, fatal=False)
3242 for rtmp_format in rtmp_formats:
3243 rtsp_format = rtmp_format.copy()
3244 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
3245 del rtsp_format['play_path']
3246 del rtsp_format['ext']
3247 rtsp_format.update({
3248 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
3249 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
3252 formats.extend([rtmp_format, rtsp_format])
3254 for protocol in ('rtmp', 'rtsp'):
3255 if protocol not in skip_protocols:
3257 'url': '%s:%s' % (protocol, url_base),
3258 'format_id': protocol,
3259 'protocol': protocol,
3263 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
3265 r'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)(?!</script>).*?\.setup\s*\((?P<options>[^)]+)\)',
3269 jwplayer_data = self._parse_json(mobj.group('options'),
3271 transform_source=transform_source)
3272 except ExtractorError:
3275 if isinstance(jwplayer_data, dict):
3276 return jwplayer_data
3278 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
3279 jwplayer_data = self._find_jwplayer_data(
3280 webpage, video_id, transform_source=js_to_json)
3281 return self._parse_jwplayer_data(
3282 jwplayer_data, video_id, *args, **kwargs)
3284 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
3285 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3286 # JWPlayer backward compatibility: flattened playlists
3287 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
3288 if 'playlist' not in jwplayer_data:
3289 jwplayer_data = {'playlist': [jwplayer_data]}
3293 # JWPlayer backward compatibility: single playlist item
3294 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
3295 if not isinstance(jwplayer_data['playlist'], list):
3296 jwplayer_data['playlist'] = [jwplayer_data['playlist']]
3298 for video_data in jwplayer_data['playlist']:
3299 # JWPlayer backward compatibility: flattened sources
3300 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
3301 if 'sources' not in video_data:
3302 video_data['sources'] = [video_data]
3304 this_video_id = video_id or video_data['mediaid']
3306 formats = self._parse_jwplayer_formats(
3307 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
3308 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
3311 tracks = video_data.get('tracks')
3312 if tracks and isinstance(tracks, list):
3313 for track in tracks:
3314 if not isinstance(track, dict):
3316 track_kind = track.get('kind')
3317 if not track_kind or not isinstance(track_kind, compat_str):
3319 if track_kind.lower() not in ('captions', 'subtitles'):
3321 track_url = urljoin(base_url, track.get('file'))
3324 subtitles.setdefault(track.get('label') or 'en', []).append({
3325 'url': self._proto_relative_url(track_url)
3329 'id': this_video_id,
3330 'title': unescapeHTML(video_data['title'] if require_title else video_data.get('title')),
3331 'description': clean_html(video_data.get('description')),
3332 'thumbnail': urljoin(base_url, self._proto_relative_url(video_data.get('image'))),
3333 'timestamp': int_or_none(video_data.get('pubdate')),
3334 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
3335 'subtitles': subtitles,
3337 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
3338 if len(formats) == 1 and re.search(r'^(?:http|//).*(?:youtube\.com|youtu\.be)/.+', formats[0]['url']):
3340 '_type': 'url_transparent',
3341 'url': formats[0]['url'],
3344 self._sort_formats(formats)
3345 entry['formats'] = formats
3346 entries.append(entry)
3347 if len(entries) == 1:
3350 return self.playlist_result(entries)
3352 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
3353 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3356 for source in jwplayer_sources_data:
3357 if not isinstance(source, dict):
3359 source_url = urljoin(
3360 base_url, self._proto_relative_url(source.get('file')))
3361 if not source_url or source_url in urls:
3363 urls.append(source_url)
3364 source_type = source.get('type') or ''
3365 ext = mimetype2ext(source_type) or determine_ext(source_url)
3366 if source_type == 'hls' or ext == 'm3u8':
3367 formats.extend(self._extract_m3u8_formats(
3368 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
3369 m3u8_id=m3u8_id, fatal=False))
3370 elif source_type == 'dash' or ext == 'mpd':
3371 formats.extend(self._extract_mpd_formats(
3372 source_url, video_id, mpd_id=mpd_id, fatal=False))
3374 formats.extend(self._extract_smil_formats(
3375 source_url, video_id, fatal=False))
3376 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3377 elif source_type.startswith('audio') or ext in (
3378 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
3385 height = int_or_none(source.get('height'))
3387 # Often no height is provided but there is a label in
3388 # format like "1080p", "720p SD", or 1080.
3389 height = int_or_none(self._search_regex(
3390 r'^(\d{3,4})[pP]?(?:\b|$)', compat_str(source.get('label') or ''),
3391 'height', default=None))
3394 'width': int_or_none(source.get('width')),
3396 'tbr': int_or_none(source.get('bitrate')),
3399 if source_url.startswith('rtmp'):
3400 a_format['ext'] = 'flv'
3401 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3402 # of jwplayer.flash.swf
3403 rtmp_url_parts = re.split(
3404 r'((?:mp4|mp3|flv):)', source_url, 1)
3405 if len(rtmp_url_parts) == 3:
3406 rtmp_url, prefix, play_path = rtmp_url_parts
3409 'play_path': prefix + play_path,
3412 a_format.update(rtmp_params)
3413 formats.append(a_format)
3416 def _live_title(self, name):
3417 """ Generate the title
for a live video
"""
3418 now = datetime.datetime.now()
3419 now_str = now.strftime('%Y-%m-%d %H:%M')
3420 return name + ' ' + now_str
3422 def _int(self, v, name, fatal=False, **kwargs):
3423 res = int_or_none(v, **kwargs)
3424 if 'get_attr' in kwargs:
3425 print(getattr(v, kwargs['get_attr']))
3427 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
3429 raise ExtractorError(msg)
3431 self.report_warning(msg)
3434 def _float(self, v, name, fatal=False, **kwargs):
3435 res = float_or_none(v, **kwargs)
3437 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
3439 raise ExtractorError(msg)
3441 self.report_warning(msg)
3444 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
3445 path='/', secure=False, discard=False, rest={}, **kwargs):
3446 cookie = compat_cookiejar_Cookie(
3447 0, name, value, port, port is not None, domain, True,
3448 domain.startswith('.'), path, True, secure, expire_time,
3449 discard, None, None, rest)
3450 self._downloader.cookiejar.set_cookie(cookie)
3452 def _get_cookies(self, url):
3453 """ Return a compat_cookies_SimpleCookie
with the cookies
for the url
"""
3454 req = sanitized_Request(url)
3455 self._downloader.cookiejar.add_cookie_header(req)
3456 return compat_cookies_SimpleCookie(req.get_header('Cookie'))
3458 def _apply_first_set_cookie_header(self, url_handle, cookie):
3460 Apply first Set
-Cookie header instead of the last
. Experimental
.
3462 Some
sites (e
.g
. [1-3]) may serve two cookies under the same name
3463 in Set
-Cookie header
and expect the
first (old
) one to be
set rather
3464 than
second (new
). However
, as of RFC6265 the newer one cookie
3465 should be
set into cookie store what actually happens
.
3466 We will workaround this issue by resetting the cookie to
3467 the first one manually
.
3468 1. https
://new
.vk
.com
/
3469 2. https
://github
.com
/ytdl
-org
/youtube
-dl
/issues
/9841#issuecomment-227871201
3470 3. https
://learning
.oreilly
.com
/
3472 for header, cookies in url_handle.headers.items():
3473 if header.lower() != 'set-cookie':
3475 if sys.version_info[0] >= 3:
3476 cookies = cookies.encode('iso-8859-1')
3477 cookies = cookies.decode('utf-8')
3478 cookie_value = re.search(
3479 r'%s=(.+?);.*?\b[Dd]omain=(.+?)(?:[,;]|$)' % cookie, cookies)
3481 value, domain = cookie_value.groups()
3482 self._set_cookie(domain, cookie, value)
3485 def get_testcases(self, include_onlymatching=False):
3486 t = getattr(self, '_TEST', None)
3488 assert not hasattr(self, '_TESTS'), \
3489 '%s has _TEST and _TESTS' % type(self).__name__
3492 tests = getattr(self, '_TESTS', [])
3494 if not include_onlymatching and t.get('only_matching', False):
3496 t['name'] = type(self).__name__[:-len('IE')]
3499 def is_suitable(self, age_limit):
3500 """ Test whether the extractor
is generally suitable
for the given
3501 age
limit (i
.e
. pornographic sites are
not, all others usually are
) """
3503 any_restricted = False
3504 for tc in self.get_testcases(include_onlymatching=False):
3505 if tc.get('playlist', []):
3506 tc = tc['playlist'][0]
3507 is_restricted = age_restricted(
3508 tc.get('info_dict', {}).get('age_limit'), age_limit)
3509 if not is_restricted:
3511 any_restricted = any_restricted or is_restricted
3512 return not any_restricted
3514 def extract_subtitles(self, *args, **kwargs):
3515 if (self.get_param('writesubtitles', False)
3516 or self.get_param('listsubtitles')):
3517 return self._get_subtitles(*args, **kwargs)
3520 def _get_subtitles(self, *args, **kwargs):
3521 raise NotImplementedError('This method must be implemented by subclasses')
3523 def extract_comments(self, *args, **kwargs):
3524 if not self.get_param('getcomments'):
3526 generator = self._get_comments(*args, **kwargs)
3532 comments.append(next(generator))
3533 except KeyboardInterrupt:
3535 self.to_screen('Interrupted by user')
3536 except StopIteration:
3538 comment_count = len(comments)
3539 self.to_screen(f'Extracted {comment_count} comments')
3541 'comments': comments,
3542 'comment_count': None if interrupted else comment_count
3546 def _get_comments(self, *args, **kwargs):
3547 raise NotImplementedError('This method must be implemented by subclasses')
3550 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
3551 """ Merge subtitle items
for one language
. Items
with duplicated URLs
3552 will be dropped
. """
3553 list1_urls = set([item['url'] for item in subtitle_list1])
3554 ret = list(subtitle_list1)
3555 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
3559 def _merge_subtitles(cls, *dicts, target=None):
3560 """ Merge subtitle dictionaries
, language by language
. """
3564 for lang, subs in d.items():
3565 target[lang] = cls._merge_subtitle_items(target.get(lang, []), subs)
3568 def extract_automatic_captions(self, *args, **kwargs):
3569 if (self.get_param('writeautomaticsub', False)
3570 or self.get_param('listsubtitles')):
3571 return self._get_automatic_captions(*args, **kwargs)
3574 def _get_automatic_captions(self, *args, **kwargs):
3575 raise NotImplementedError('This method must be implemented by subclasses')
3577 def mark_watched(self, *args, **kwargs):
3578 if not self.get_param('mark_watched', False):
3580 if (self._get_login_info()[0] is not None
3581 or self.get_param('cookiefile')
3582 or self.get_param('cookiesfrombrowser')):
3583 self._mark_watched(*args, **kwargs)
3585 def _mark_watched(self, *args, **kwargs):
3586 raise NotImplementedError('This method must be implemented by subclasses')
3588 def geo_verification_headers(self):
3590 geo_verification_proxy = self.get_param('geo_verification_proxy')
3591 if geo_verification_proxy:
3592 headers['Ytdl-request-proxy'] = geo_verification_proxy
3595 def _generic_id(self, url):
3596 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
3598 def _generic_title(self, url):
3599 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
3602 def _availability(is_private=None, needs_premium=None, needs_subscription=None, needs_auth=None, is_unlisted=None):
3603 all_known = all(map(
3604 lambda x: x is not None,
3605 (is_private, needs_premium, needs_subscription, needs_auth, is_unlisted)))
3607 'private' if is_private
3608 else 'premium_only' if needs_premium
3609 else 'subscriber_only' if needs_subscription
3610 else 'needs_auth' if needs_auth
3611 else 'unlisted' if is_unlisted
3612 else 'public' if all_known
3615 def _configuration_arg(self, key, default=NO_DEFAULT, casesense=False):
3617 @returns A list of values for the extractor argument given by "key"
3618 or "default" if no such key is present
3619 @param default The default value to return when the key is not present (default: [])
3620 @param casesense When false, the values are converted to lower case
3623 self._downloader.params, ('extractor_args', self.ie_key().lower(), key))
3625 return [] if default is NO_DEFAULT else default
3626 return list(val) if casesense else [x.lower() for x in val]
3629 class SearchInfoExtractor(InfoExtractor):
3631 Base
class for paged search queries extractors
.
3632 They accept URLs
in the format
_SEARCH_KEY(|all|
[0-9]):{query}
3633 Instances should define _SEARCH_KEY
and optionally _MAX_RESULTS
3636 _MAX_RESULTS = float('inf')
3639 def _make_valid_url(cls):
3640 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
3643 def suitable(cls, url):
3644 return re.match(cls._make_valid_url(), url) is not None
3646 def _real_extract(self, query):
3647 mobj = re.match(self._make_valid_url(), query)
3649 raise ExtractorError('Invalid search query "%s"' % query)
3651 prefix = mobj.group('prefix')
3652 query = mobj.group('query')
3654 return self._get_n_results(query, 1)
3655 elif prefix == 'all':
3656 return self._get_n_results(query, self._MAX_RESULTS)
3660 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
3661 elif n > self._MAX_RESULTS:
3662 self.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3663 n = self._MAX_RESULTS
3664 return self._get_n_results(query, n)
3666 def _get_n_results(self, query, n):
3667 """Get a specified number of results
for a query
.
3668 Either this function
or _search_results must be overridden by subclasses
"""
3669 return self.playlist_result(
3670 itertools.islice(self._search_results(query), 0, None if n == float('inf') else n),
3673 def _search_results(self, query):
3674 """Returns an iterator of search results
"""
3675 raise NotImplementedError('This method must be implemented by subclasses')
3678 def SEARCH_KEY(self):
3679 return self._SEARCH_KEY