2 from __future__
import unicode_literals
16 from ..compat
import (
17 compat_cookiejar_Cookie
,
18 compat_cookies_SimpleCookie
,
20 compat_etree_fromstring
,
27 compat_urllib_parse_unquote
,
28 compat_urllib_parse_urlencode
,
29 compat_urllib_request
,
31 compat_xml_parse_error
,
33 from ..downloader
import FileDownloader
34 from ..downloader
.f4m
import (
36 remove_encrypted_media
,
65 parse_m3u8_attributes
,
88 class InfoExtractor(object):
89 """Information Extractor class.
91 Information extractors are the classes that, given a URL, extract
92 information about the video (or videos) the URL refers to. This
93 information includes the real video URL, the video title, author and
94 others. The information is stored in a dictionary which is then
95 passed to the YoutubeDL. The YoutubeDL processes this
96 information possibly downloading the video to the file system, among
97 other possible outcomes.
99 The type field determines the type of the result.
100 By far the most common value (and the default if _type is missing) is
101 "video", which indicates a single video.
103 For a video, the dictionaries must include the following fields:
105 id: Video identifier.
106 title: Video title, unescaped.
108 Additionally, it must contain either a formats entry or a url one:
110 formats: A list of dictionaries for each format available, ordered
111 from worst to best quality.
114 * url The mandatory URL representing the media:
115 for plain file media - HTTP URL of this file,
117 for HLS - URL of the M3U8 media playlist,
118 for HDS - URL of the F4M manifest,
120 - HTTP URL to plain file media (in case of
122 - URL of the MPD manifest or base URL
123 representing the media if MPD manifest
124 is parsed from a string (in case of
126 for MSS - URL of the ISM manifest.
128 The URL of the manifest file in case of
130 for HLS - URL of the M3U8 master playlist,
131 for HDS - URL of the F4M manifest,
132 for DASH - URL of the MPD manifest,
133 for MSS - URL of the ISM manifest.
134 * ext Will be calculated from URL if missing
135 * format A human-readable description of the format
136 ("mp4 container with h264/opus").
137 Calculated from the format_id, width, height.
138 and format_note fields if missing.
139 * format_id A short description of the format
140 ("mp4_h264_opus" or "19").
141 Technically optional, but strongly recommended.
142 * format_note Additional info about the format
143 ("3D" or "DASH video")
144 * width Width of the video, if known
145 * height Height of the video, if known
146 * resolution Textual description of width and height
147 * tbr Average bitrate of audio and video in KBit/s
148 * abr Average audio bitrate in KBit/s
149 * acodec Name of the audio codec in use
150 * asr Audio sampling rate in Hertz
151 * vbr Average video bitrate in KBit/s
153 * vcodec Name of the video codec in use
154 * container Name of the container format
155 * filesize The number of bytes, if known in advance
156 * filesize_approx An estimate for the number of bytes
157 * player_url SWF Player URL (used for rtmpdump).
158 * protocol The protocol that will be used for the actual
159 download, lower-case.
160 "http", "https", "rtsp", "rtmp", "rtmp_ffmpeg", "rtmpe",
161 "m3u8", "m3u8_native" or "http_dash_segments".
163 Base URL for fragments. Each fragment's path
164 value (if present) will be relative to
166 * fragments A list of fragments of a fragmented media.
167 Each fragment entry must contain either an url
168 or a path. If an url is present it should be
169 considered by a client. Otherwise both path and
170 fragment_base_url must be present. Here is
171 the list of all potential fields:
172 * "url" - fragment's URL
173 * "path" - fragment's path relative to
175 * "duration" (optional, int or float)
176 * "filesize" (optional, int)
177 * preference Order number of this format. If this field is
178 present and not None, the formats get sorted
179 by this field, regardless of all other values.
180 -1 for default (order by other properties),
181 -2 or smaller for less than default.
182 < -1000 to hide the format (if there is
183 another one which is strictly better)
184 * language Language code, e.g. "de" or "en-US".
185 * language_preference Is this in the language mentioned in
187 10 if it's what the URL is about,
188 -1 for default (don't know),
189 -10 otherwise, other values reserved for now.
190 * quality Order number of the video quality of this
191 format, irrespective of the file format.
192 -1 for default (order by other properties),
193 -2 or smaller for less than default.
194 * source_preference Order number for this video source
195 (quality takes higher priority)
196 -1 for default (order by other properties),
197 -2 or smaller for less than default.
198 * http_headers A dictionary of additional HTTP headers
199 to add to the request.
200 * stretched_ratio If given and not 1, indicates that the
201 video's pixels are not square.
202 width : height ratio as float.
203 * no_resume The server does not support resuming the
204 (HTTP or RTMP) download. Boolean.
205 * downloader_options A dictionary of downloader options as
206 described in FileDownloader
207 RTMP formats can also have the additional fields: page_url,
208 app, play_path, tc_url, flash_version, rtmp_live, rtmp_conn,
209 rtmp_protocol, rtmp_real_time
211 url: Final video URL.
212 ext: Video filename extension.
213 format: The video format, defaults to ext (used for --get-format)
214 player_url: SWF Player URL (used for rtmpdump).
216 The following fields are optional:
218 alt_title: A secondary title of the video.
219 display_id An alternative identifier for the video, not necessarily
220 unique, but available before title. Typically, id is
221 something like "4234987", title "Dancing naked mole rats",
222 and display_id "dancing-naked-mole-rats"
223 thumbnails: A list of dictionaries, with the following entries:
224 * "id" (optional, string) - Thumbnail format ID
226 * "preference" (optional, int) - quality of the image
227 * "width" (optional, int)
228 * "height" (optional, int)
229 * "resolution" (optional, string "{width}x{height}",
231 * "filesize" (optional, int)
232 thumbnail: Full URL to a video thumbnail image.
233 description: Full video description.
234 uploader: Full name of the video uploader.
235 license: License name the video is licensed under.
236 creator: The creator of the video.
237 release_timestamp: UNIX timestamp of the moment the video was released.
238 release_date: The date (YYYYMMDD) when the video was released.
239 timestamp: UNIX timestamp of the moment the video was uploaded
240 upload_date: Video upload date (YYYYMMDD).
241 If not explicitly set, calculated from timestamp.
242 uploader_id: Nickname or id of the video uploader.
243 uploader_url: Full URL to a personal webpage of the video uploader.
244 channel: Full name of the channel the video is uploaded on.
245 Note that channel fields may or may not repeat uploader
246 fields. This depends on a particular extractor.
247 channel_id: Id of the channel.
248 channel_url: Full URL to a channel webpage.
249 location: Physical location where the video was filmed.
250 subtitles: The available subtitles as a dictionary in the format
251 {tag: subformats}. "tag" is usually a language code, and
252 "subformats" is a list sorted from lower to higher
253 preference, each element is a dictionary with the "ext"
255 * "data": The subtitles file contents
256 * "url": A URL pointing to the subtitles file
257 It can optionally also have:
258 * "name": Name or description of the subtitles
259 "ext" will be calculated from URL if missing
260 automatic_captions: Like 'subtitles'; contains automatically generated
261 captions instead of normal subtitles
262 duration: Length of the video in seconds, as an integer or float.
263 view_count: How many users have watched the video on the platform.
264 like_count: Number of positive ratings of the video
265 dislike_count: Number of negative ratings of the video
266 repost_count: Number of reposts of the video
267 average_rating: Average rating give by users, the scale used depends on the webpage
268 comment_count: Number of comments on the video
269 comments: A list of comments, each with one or more of the following
270 properties (all but one of text or html optional):
271 * "author" - human-readable name of the comment author
272 * "author_id" - user ID of the comment author
273 * "author_thumbnail" - The thumbnail of the comment author
275 * "html" - Comment as HTML
276 * "text" - Plain text of the comment
277 * "timestamp" - UNIX timestamp of comment
278 * "parent" - ID of the comment this one is replying to.
279 Set to "root" to indicate that this is a
280 comment to the original video.
281 * "like_count" - Number of positive ratings of the comment
282 * "dislike_count" - Number of negative ratings of the comment
283 * "is_favorited" - Whether the comment is marked as
284 favorite by the video uploader
285 * "author_is_uploader" - Whether the comment is made by
287 age_limit: Age restriction for the video, as an integer (years)
288 webpage_url: The URL to the video webpage, if given to yt-dlp it
289 should allow to get the same result again. (It will be set
290 by YoutubeDL if it's missing)
291 categories: A list of categories that the video falls in, for example
293 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
294 cast: A list of the video cast
295 is_live: True, False, or None (=unknown). Whether this video is a
296 live stream that goes on instead of a fixed-length video.
297 was_live: True, False, or None (=unknown). Whether this video was
298 originally a live stream.
299 start_time: Time in seconds where the reproduction should start, as
300 specified in the URL.
301 end_time: Time in seconds where the reproduction should end, as
302 specified in the URL.
303 chapters: A list of dictionaries, with the following entries:
304 * "start_time" - The start time of the chapter in seconds
305 * "end_time" - The end time of the chapter in seconds
306 * "title" (optional, string)
307 playable_in_embed: Whether this video is allowed to play in embedded
308 players on other sites. Can be True (=always allowed),
309 False (=never allowed), None (=unknown), or a string
310 specifying the criteria for embedability (Eg: 'whitelist')
311 availability: Under what condition the video is available. One of
312 'private', 'premium_only', 'subscriber_only', 'needs_auth',
313 'unlisted' or 'public'. Use 'InfoExtractor._availability'
315 __post_extractor: A function to be called just before the metadata is
316 written to either disk, logger or console. The function
317 must return a dict which will be added to the info_dict.
318 This is usefull for additional information that is
319 time-consuming to extract. Note that the fields thus
320 extracted will not be available to output template and
321 match_filter. So, only "comments" and "comment_count" are
322 currently allowed to be extracted via this method.
324 The following fields should only be used when the video belongs to some logical
327 chapter: Name or title of the chapter the video belongs to.
328 chapter_number: Number of the chapter the video belongs to, as an integer.
329 chapter_id: Id of the chapter the video belongs to, as a unicode string.
331 The following fields should only be used when the video is an episode of some
332 series, programme or podcast:
334 series: Title of the series or programme the video episode belongs to.
335 season: Title of the season the video episode belongs to.
336 season_number: Number of the season the video episode belongs to, as an integer.
337 season_id: Id of the season the video episode belongs to, as a unicode string.
338 episode: Title of the video episode. Unlike mandatory video title field,
339 this field should denote the exact title of the video episode
340 without any kind of decoration.
341 episode_number: Number of the video episode within a season, as an integer.
342 episode_id: Id of the video episode, as a unicode string.
344 The following fields should only be used when the media is a track or a part of
347 track: Title of the track.
348 track_number: Number of the track within an album or a disc, as an integer.
349 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
351 artist: Artist(s) of the track.
352 genre: Genre(s) of the track.
353 album: Title of the album the track belongs to.
354 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
355 album_artist: List of all artists appeared on the album (e.g.
356 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
358 disc_number: Number of the disc or other physical medium the track belongs to,
360 release_year: Year (YYYY) when the album was released.
362 Unless mentioned otherwise, the fields should be Unicode strings.
364 Unless mentioned otherwise, None is equivalent to absence of information.
367 _type "playlist" indicates multiple videos.
368 There must be a key "entries", which is a list, an iterable, or a PagedList
369 object, each element of which is a valid dictionary by this specification.
371 Additionally, playlists can have "id", "title", and any other relevent
372 attributes with the same semantics as videos (see above).
375 _type "multi_video" indicates that there are multiple videos that
376 form a single show, for examples multiple acts of an opera or TV episode.
377 It must have an entries key like a playlist and contain all the keys
378 required for a video at the same time.
381 _type "url" indicates that the video must be extracted from another
382 location, possibly by a different extractor. Its only required key is:
383 "url" - the next URL to extract.
384 The key "ie_key" can be set to the class name (minus the trailing "IE",
385 e.g. "Youtube") if the extractor class is known in advance.
386 Additionally, the dictionary may have any properties of the resolved entity
387 known in advance, for example "title" if the title of the referred video is
391 _type "url_transparent" entities have the same specification as "url", but
392 indicate that the given additional information is more precise than the one
393 associated with the resolved URL.
394 This is useful when a site employs a video service that hosts the video and
395 its technical metadata, but that video service does not embed a useful
396 title, description etc.
399 Subclasses of this one should re-define the _real_initialize() and
400 _real_extract() methods and define a _VALID_URL regexp.
401 Probably, they should also be added to the list of extractors.
403 _GEO_BYPASS attribute may be set to False in order to disable
404 geo restriction bypass mechanisms for a particular extractor.
405 Though it won't disable explicit geo restriction bypass based on
406 country code provided with geo_bypass_country.
408 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
409 countries for this extractor. One of these countries will be used by
410 geo restriction bypass mechanism right away in order to bypass
411 geo restriction, of course, if the mechanism is not disabled.
413 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
414 IP blocks in CIDR notation for this extractor. One of these IP blocks
415 will be used by geo restriction bypass mechanism similarly
418 Finally, the _WORKING attribute should be set to False for broken IEs
419 in order to warn the users and skip the tests.
424 _x_forwarded_for_ip
= None
426 _GEO_COUNTRIES
= None
427 _GEO_IP_BLOCKS
= None
431 'any': 'Use --cookies, --username and --password or --netrc to provide account credentials',
433 'Use --cookies for the authentication. '
434 'See https://github.com/ytdl-org/youtube-dl#how-do-i-pass-cookies-to-youtube-dl for how to pass cookies'),
435 'password': 'Use --username and --password or --netrc to provide account credentials',
438 def __init__(self
, downloader
=None):
439 """Constructor. Receives an optional downloader."""
441 self
._x
_forwarded
_for
_ip
= None
442 self
.set_downloader(downloader
)
445 def suitable(cls
, url
):
446 """Receives a URL and returns True if suitable for this IE."""
448 # This does not use has/getattr intentionally - we want to know whether
449 # we have cached the regexp for *this* class, whereas getattr would also
450 # match the superclass
451 if '_VALID_URL_RE' not in cls
.__dict
__:
452 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
453 return cls
._VALID
_URL
_RE
.match(url
) is not None
456 def _match_id(cls
, url
):
457 if '_VALID_URL_RE' not in cls
.__dict
__:
458 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
459 m
= cls
._VALID
_URL
_RE
.match(url
)
461 return compat_str(m
.group('id'))
465 """Getter method for _WORKING."""
468 def initialize(self
):
469 """Initializes an instance (authentication, etc)."""
470 self
._initialize
_geo
_bypass
({
471 'countries': self
._GEO
_COUNTRIES
,
472 'ip_blocks': self
._GEO
_IP
_BLOCKS
,
475 self
._real
_initialize
()
478 def _initialize_geo_bypass(self
, geo_bypass_context
):
480 Initialize geo restriction bypass mechanism.
482 This method is used to initialize geo bypass mechanism based on faking
483 X-Forwarded-For HTTP header. A random country from provided country list
484 is selected and a random IP belonging to this country is generated. This
485 IP will be passed as X-Forwarded-For HTTP header in all subsequent
488 This method will be used for initial geo bypass mechanism initialization
489 during the instance initialization with _GEO_COUNTRIES and
492 You may also manually call it from extractor's code if geo bypass
493 information is not available beforehand (e.g. obtained during
494 extraction) or due to some other reason. In this case you should pass
495 this information in geo bypass context passed as first argument. It may
496 contain following fields:
498 countries: List of geo unrestricted countries (similar
500 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
501 (similar to _GEO_IP_BLOCKS)
504 if not self
._x
_forwarded
_for
_ip
:
506 # Geo bypass mechanism is explicitly disabled by user
507 if not self
.get_param('geo_bypass', True):
510 if not geo_bypass_context
:
511 geo_bypass_context
= {}
513 # Backward compatibility: previously _initialize_geo_bypass
514 # expected a list of countries, some 3rd party code may still use
516 if isinstance(geo_bypass_context
, (list, tuple)):
517 geo_bypass_context
= {
518 'countries': geo_bypass_context
,
521 # The whole point of geo bypass mechanism is to fake IP
522 # as X-Forwarded-For HTTP header based on some IP block or
525 # Path 1: bypassing based on IP block in CIDR notation
527 # Explicit IP block specified by user, use it right away
528 # regardless of whether extractor is geo bypassable or not
529 ip_block
= self
.get_param('geo_bypass_ip_block', None)
531 # Otherwise use random IP block from geo bypass context but only
532 # if extractor is known as geo bypassable
534 ip_blocks
= geo_bypass_context
.get('ip_blocks')
535 if self
._GEO
_BYPASS
and ip_blocks
:
536 ip_block
= random
.choice(ip_blocks
)
539 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(ip_block
)
540 self
._downloader
.write_debug(
541 '[debug] Using fake IP %s as X-Forwarded-For' % self
._x
_forwarded
_for
_ip
)
544 # Path 2: bypassing based on country code
546 # Explicit country code specified by user, use it right away
547 # regardless of whether extractor is geo bypassable or not
548 country
= self
.get_param('geo_bypass_country', None)
550 # Otherwise use random country code from geo bypass context but
551 # only if extractor is known as geo bypassable
553 countries
= geo_bypass_context
.get('countries')
554 if self
._GEO
_BYPASS
and countries
:
555 country
= random
.choice(countries
)
558 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country
)
559 self
._downloader
.write_debug(
560 'Using fake IP %s (%s) as X-Forwarded-For' % (self
._x
_forwarded
_for
_ip
, country
.upper()))
562 def extract(self
, url
):
563 """Extracts URL information and returns it in list of dicts."""
568 self
.write_debug('Extracting URL: %s' % url
)
569 ie_result
= self
._real
_extract
(url
)
570 if ie_result
is None:
572 if self
._x
_forwarded
_for
_ip
:
573 ie_result
['__x_forwarded_for_ip'] = self
._x
_forwarded
_for
_ip
574 subtitles
= ie_result
.get('subtitles')
575 if (subtitles
and 'live_chat' in subtitles
576 and 'no-live-chat' in self
.get_param('compat_opts', [])):
577 del subtitles
['live_chat']
579 except GeoRestrictedError
as e
:
580 if self
.__maybe
_fake
_ip
_and
_retry
(e
.countries
):
583 except ExtractorError
:
585 except compat_http_client
.IncompleteRead
as e
:
586 raise ExtractorError('A network error has occurred.', cause
=e
, expected
=True)
587 except (KeyError, StopIteration) as e
:
588 raise ExtractorError('An extractor error has occurred.', cause
=e
)
590 def __maybe_fake_ip_and_retry(self
, countries
):
591 if (not self
.get_param('geo_bypass_country', None)
593 and self
.get_param('geo_bypass', True)
594 and not self
._x
_forwarded
_for
_ip
596 country_code
= random
.choice(countries
)
597 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country_code
)
598 if self
._x
_forwarded
_for
_ip
:
600 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
601 % (self
._x
_forwarded
_for
_ip
, country_code
.upper()))
605 def set_downloader(self
, downloader
):
606 """Sets the downloader for this IE."""
607 self
._downloader
= downloader
609 def _real_initialize(self
):
610 """Real initialization process. Redefine in subclasses."""
613 def _real_extract(self
, url
):
614 """Real extraction process. Redefine in subclasses."""
619 """A string for getting the InfoExtractor with get_info_extractor"""
620 return compat_str(cls
.__name
__[:-2])
624 return compat_str(type(self
).__name
__[:-2])
627 def __can_accept_status_code(err
, expected_status
):
628 assert isinstance(err
, compat_urllib_error
.HTTPError
)
629 if expected_status
is None:
631 if isinstance(expected_status
, compat_integer_types
):
632 return err
.code
== expected_status
633 elif isinstance(expected_status
, (list, tuple)):
634 return err
.code
in expected_status
635 elif callable(expected_status
):
636 return expected_status(err
.code
) is True
640 def _request_webpage(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
, expected_status
=None):
642 Return the response handle.
644 See _download_webpage docstring for arguments specification.
646 if not self
._downloader
._first
_webpage
_request
:
647 sleep_interval
= float_or_none(self
.get_param('sleep_interval_requests')) or 0
648 if sleep_interval
> 0:
649 self
.to_screen('Sleeping %s seconds ...' % sleep_interval
)
650 time
.sleep(sleep_interval
)
652 self
._downloader
._first
_webpage
_request
= False
655 self
.report_download_webpage(video_id
)
656 elif note
is not False:
658 self
.to_screen('%s' % (note
,))
660 self
.to_screen('%s: %s' % (video_id
, note
))
662 # Some sites check X-Forwarded-For HTTP header in order to figure out
663 # the origin of the client behind proxy. This allows bypassing geo
664 # restriction by faking this header's value to IP that belongs to some
665 # geo unrestricted country. We will do so once we encounter any
666 # geo restriction error.
667 if self
._x
_forwarded
_for
_ip
:
668 if 'X-Forwarded-For' not in headers
:
669 headers
['X-Forwarded-For'] = self
._x
_forwarded
_for
_ip
671 if isinstance(url_or_request
, compat_urllib_request
.Request
):
672 url_or_request
= update_Request(
673 url_or_request
, data
=data
, headers
=headers
, query
=query
)
676 url_or_request
= update_url_query(url_or_request
, query
)
677 if data
is not None or headers
:
678 url_or_request
= sanitized_Request(url_or_request
, data
, headers
)
680 return self
._downloader
.urlopen(url_or_request
)
681 except network_exceptions
as err
:
682 if isinstance(err
, compat_urllib_error
.HTTPError
):
683 if self
.__can
_accept
_status
_code
(err
, expected_status
):
684 # Retain reference to error to prevent file object from
685 # being closed before it can be read. Works around the
686 # effects of <https://bugs.python.org/issue15002>
687 # introduced in Python 3.4.1.
694 errnote
= 'Unable to download webpage'
696 errmsg
= '%s: %s' % (errnote
, error_to_compat_str(err
))
698 raise ExtractorError(errmsg
, sys
.exc_info()[2], cause
=err
)
700 self
.report_warning(errmsg
)
703 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):
705 Return a tuple (page content as string, URL handle).
707 See _download_webpage docstring for arguments specification.
709 # Strip hashes from the URL (#1038)
710 if isinstance(url_or_request
, (compat_str
, str)):
711 url_or_request
= url_or_request
.partition('#')[0]
713 urlh
= self
._request
_webpage
(url_or_request
, video_id
, note
, errnote
, fatal
, data
=data
, headers
=headers
, query
=query
, expected_status
=expected_status
)
717 content
= self
._webpage
_read
_content
(urlh
, url_or_request
, video_id
, note
, errnote
, fatal
, encoding
=encoding
)
718 return (content
, urlh
)
721 def _guess_encoding_from_content(content_type
, webpage_bytes
):
722 m
= re
.match(r
'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type
)
724 encoding
= m
.group(1)
726 m
= re
.search(br
'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
727 webpage_bytes[:1024])
729 encoding = m.group(1).decode('ascii')
730 elif webpage_bytes.startswith(b'\xff\xfe'):
737 def __check_blocked(self, content):
738 first_block = content[:512]
739 if ('<title>Access to this site is blocked</title>' in content
740 and 'Websense' in first_block):
741 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
742 blocked_iframe = self._html_search_regex(
743 r'<iframe src="([^
"]+)"', content,
744 'Websense information URL
', default=None)
746 msg += ' Visit
%s for more details
' % blocked_iframe
747 raise ExtractorError(msg, expected=True)
748 if '<title
>The URL you requested has been blocked
</title
>' in first_block:
750 'Access to this webpage has been blocked by Indian censorship
. '
751 'Use a VPN
or proxy
server (with --proxy
) to route around it
.')
752 block_msg = self._html_search_regex(
753 r'</h1
><p
>(.*?
)</p
>',
754 content, 'block message
', default=None)
756 msg += ' (Message
: "%s")' % block_msg.replace('\n', ' ')
757 raise ExtractorError(msg, expected=True)
758 if ('<title
>TTK
:: Доступ к ресурсу ограничен
</title
>' in content
759 and 'blocklist
.rkn
.gov
.ru
' in content):
760 raise ExtractorError(
761 'Access to this webpage has been blocked by decision of the Russian government
. '
762 'Visit http
://blocklist
.rkn
.gov
.ru
/ for a block reason
.',
765 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
766 content_type = urlh.headers.get('Content
-Type
', '')
767 webpage_bytes = urlh.read()
768 if prefix is not None:
769 webpage_bytes = prefix + webpage_bytes
771 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
772 if self.get_param('dump_intermediate_pages
', False):
773 self.to_screen('Dumping request to
' + urlh.geturl())
774 dump = base64.b64encode(webpage_bytes).decode('ascii
')
775 self._downloader.to_screen(dump)
776 if self.get_param('write_pages
', False):
777 basen = '%s_%s' % (video_id, urlh.geturl())
779 h = '___
' + hashlib.md5(basen.encode('utf
-8')).hexdigest()
780 basen = basen[:240 - len(h)] + h
781 raw_filename = basen + '.dump
'
782 filename = sanitize_filename(raw_filename, restricted=True)
783 self.to_screen('Saving request to
' + filename)
784 # Working around MAX_PATH limitation on Windows (see
785 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
786 if compat_os_name == 'nt
':
787 absfilepath = os.path.abspath(filename)
788 if len(absfilepath) > 259:
789 filename = '\\\\?
\\' + absfilepath
790 with open(filename, 'wb
') as outf:
791 outf.write(webpage_bytes)
794 content = webpage_bytes.decode(encoding, 'replace
')
796 content = webpage_bytes.decode('utf
-8', 'replace
')
798 self.__check_blocked(content)
802 def _download_webpage(
803 self, url_or_request, video_id, note=None, errnote=None,
804 fatal=True, tries=1, timeout=5, encoding=None, data=None,
805 headers={}, query={}, expected_status=None):
807 Return the data of the page as a string.
810 url_or_request -- plain text URL as a string or
811 a compat_urllib_request.Requestobject
812 video_id -- Video/playlist/item identifier (string)
815 note -- note printed before downloading (string)
816 errnote -- note printed in case of an error (string)
817 fatal -- flag denoting whether error should be considered fatal,
818 i.e. whether it should cause ExtractionError to be raised,
819 otherwise a warning will be reported and extraction continued
820 tries -- number of tries
821 timeout -- sleep interval between tries
822 encoding -- encoding for a page content decoding, guessed automatically
823 when not explicitly specified
824 data -- POST data (bytes)
825 headers -- HTTP headers (dict)
826 query -- URL query (dict)
827 expected_status -- allows to accept failed HTTP requests (non 2xx
828 status code) by explicitly specifying a set of accepted status
829 codes. Can be any of the following entities:
830 - an integer type specifying an exact failed status code to
832 - a list or a tuple of integer types specifying a list of
833 failed status codes to accept
834 - a callable accepting an actual failed status code and
835 returning True if it should be accepted
836 Note that this argument does not affect success status codes (2xx)
837 which are always accepted.
842 while success is False:
844 res = self._download_webpage_handle(
845 url_or_request, video_id, note, errnote, fatal,
846 encoding=encoding, data=data, headers=headers, query=query,
847 expected_status=expected_status)
849 except compat_http_client.IncompleteRead as e:
851 if try_count >= tries:
853 self._sleep(timeout, video_id)
860 def _download_xml_handle(
861 self, url_or_request, video_id, note='Downloading XML
',
862 errnote='Unable to download XML
', transform_source=None,
863 fatal=True, encoding=None, data=None, headers={}, query={},
864 expected_status=None):
866 Return a tuple (xml as an compat_etree_Element, URL handle).
868 See _download_webpage docstring for arguments specification.
870 res = self._download_webpage_handle(
871 url_or_request, video_id, note, errnote, fatal=fatal,
872 encoding=encoding, data=data, headers=headers, query=query,
873 expected_status=expected_status)
876 xml_string, urlh = res
877 return self._parse_xml(
878 xml_string, video_id, transform_source=transform_source,
882 self, url_or_request, video_id,
883 note='Downloading XML
', errnote='Unable to download XML
',
884 transform_source=None, fatal=True, encoding=None,
885 data=None, headers={}, query={}, expected_status=None):
887 Return the xml as an compat_etree_Element.
889 See _download_webpage docstring for arguments specification.
891 res = self._download_xml_handle(
892 url_or_request, video_id, note=note, errnote=errnote,
893 transform_source=transform_source, fatal=fatal, encoding=encoding,
894 data=data, headers=headers, query=query,
895 expected_status=expected_status)
896 return res if res is False else res[0]
898 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
900 xml_string = transform_source(xml_string)
902 return compat_etree_fromstring(xml_string.encode('utf
-8'))
903 except compat_xml_parse_error as ve:
904 errmsg = '%s: Failed to parse XML
' % video_id
906 raise ExtractorError(errmsg, cause=ve)
908 self.report_warning(errmsg + str(ve))
910 def _download_json_handle(
911 self, url_or_request, video_id, note='Downloading JSON metadata
',
912 errnote='Unable to download JSON metadata
', transform_source=None,
913 fatal=True, encoding=None, data=None, headers={}, query={},
914 expected_status=None):
916 Return a tuple (JSON object, URL handle).
918 See _download_webpage docstring for arguments specification.
920 res = self._download_webpage_handle(
921 url_or_request, video_id, note, errnote, fatal=fatal,
922 encoding=encoding, data=data, headers=headers, query=query,
923 expected_status=expected_status)
926 json_string, urlh = res
927 return self._parse_json(
928 json_string, video_id, transform_source=transform_source,
932 self, url_or_request, video_id, note='Downloading JSON metadata
',
933 errnote='Unable to download JSON metadata
', transform_source=None,
934 fatal=True, encoding=None, data=None, headers={}, query={},
935 expected_status=None):
937 Return the JSON object as a dict.
939 See _download_webpage docstring for arguments specification.
941 res = self._download_json_handle(
942 url_or_request, video_id, note=note, errnote=errnote,
943 transform_source=transform_source, fatal=fatal, encoding=encoding,
944 data=data, headers=headers, query=query,
945 expected_status=expected_status)
946 return res if res is False else res[0]
948 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
950 json_string = transform_source(json_string)
952 return json.loads(json_string)
953 except ValueError as ve:
954 errmsg = '%s: Failed to parse JSON
' % video_id
956 raise ExtractorError(errmsg, cause=ve)
958 self.report_warning(errmsg + str(ve))
960 def _parse_socket_response_as_json(self, data, video_id, transform_source=None, fatal=True):
961 return self._parse_json(
962 data[data.find('{'):data.rfind('}
') + 1],
963 video_id, transform_source, fatal)
965 def _download_socket_json_handle(
966 self, url_or_request, video_id, note='Polling socket
',
967 errnote='Unable to poll socket
', transform_source=None,
968 fatal=True, encoding=None, data=None, headers={}, query={},
969 expected_status=None):
971 Return a tuple (JSON object, URL handle).
973 See _download_webpage docstring for arguments specification.
975 res = self._download_webpage_handle(
976 url_or_request, video_id, note, errnote, fatal=fatal,
977 encoding=encoding, data=data, headers=headers, query=query,
978 expected_status=expected_status)
982 return self._parse_socket_response_as_json(
983 webpage, video_id, transform_source=transform_source,
986 def _download_socket_json(
987 self, url_or_request, video_id, note='Polling socket
',
988 errnote='Unable to poll socket
', transform_source=None,
989 fatal=True, encoding=None, data=None, headers={}, query={},
990 expected_status=None):
992 Return the JSON object as a dict.
994 See _download_webpage docstring for arguments specification.
996 res = self._download_socket_json_handle(
997 url_or_request, video_id, note=note, errnote=errnote,
998 transform_source=transform_source, fatal=fatal, encoding=encoding,
999 data=data, headers=headers, query=query,
1000 expected_status=expected_status)
1001 return res if res is False else res[0]
1003 def report_warning(self, msg, video_id=None, *args, **kwargs):
1004 idstr = '' if video_id is None else '%s: ' % video_id
1005 self._downloader.report_warning(
1006 '[%s] %s%s' % (self.IE_NAME, idstr, msg), *args, **kwargs)
1008 def to_screen(self, msg, *args, **kwargs):
1009 """Print msg to screen, prefixing it with '[ie_name
]'"""
1010 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg), *args, **kwargs)
1012 def write_debug(self, msg, *args, **kwargs):
1013 self._downloader.write_debug('[%s] %s' % (self.IE_NAME, msg), *args, **kwargs)
1015 def get_param(self, name, default=None, *args, **kwargs):
1016 if self._downloader:
1017 return self._downloader.params.get(name, default, *args, **kwargs)
1020 def report_extraction(self, id_or_name):
1021 """Report information extraction."""
1022 self.to_screen('%s: Extracting information
' % id_or_name)
1024 def report_download_webpage(self, video_id):
1025 """Report webpage download."""
1026 self.to_screen('%s: Downloading webpage
' % video_id)
1028 def report_age_confirmation(self):
1029 """Report attempt to confirm age."""
1030 self.to_screen('Confirming age
')
1032 def report_login(self):
1033 """Report attempt to log in."""
1034 self.to_screen('Logging
in')
1036 def raise_login_required(
1037 self, msg='This video
is only available
for registered users
',
1038 metadata_available=False, method='any
'):
1039 if metadata_available and self.get_param('ignore_no_formats_error
'):
1040 self.report_warning(msg)
1041 raise ExtractorError('%s. %s' % (msg, self._LOGIN_HINTS[method]), expected=True)
1043 def raise_geo_restricted(
1044 self, msg='This video
is not available
from your location due to geo restriction
',
1045 countries=None, metadata_available=False):
1046 if metadata_available and self.get_param('ignore_no_formats_error
'):
1047 self.report_warning(msg)
1049 raise GeoRestrictedError(msg, countries=countries)
1051 def raise_no_formats(self, msg, expected=False, video_id=None):
1052 if expected and self.get_param('ignore_no_formats_error
'):
1053 self.report_warning(msg, video_id)
1055 raise ExtractorError(msg, expected=expected, video_id=video_id)
1057 # Methods for following #608
1059 def url_result(url, ie=None, video_id=None, video_title=None):
1060 """Returns a URL that points to a page that should be processed"""
1061 # TODO: ie should be the class used for getting the info
1062 video_info = {'_type
': 'url
',
1065 if video_id is not None:
1066 video_info['id'] = video_id
1067 if video_title is not None:
1068 video_info['title
'] = video_title
1071 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None):
1073 self.url_result(self._proto_relative_url(getter(m) if getter else m), ie)
1075 return self.playlist_result(
1076 urls, playlist_id=playlist_id, playlist_title=playlist_title)
1079 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None, **kwargs):
1080 """Returns a playlist"""
1081 video_info = {'_type
': 'playlist
',
1083 video_info.update(kwargs)
1085 video_info['id'] = playlist_id
1087 video_info['title
'] = playlist_title
1088 if playlist_description is not None:
1089 video_info['description
'] = playlist_description
1092 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1094 Perform a regex search on the given string, using a single or a list of
1095 patterns returning the first matching group.
1096 In case of failure return a default value or raise a WARNING or a
1097 RegexNotFoundError, depending on fatal, specifying the field name.
1099 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
1100 mobj = re.search(pattern, string, flags)
1103 mobj = re.search(p, string, flags)
1107 if not self.get_param('no_color
') and compat_os_name != 'nt
' and sys.stderr.isatty():
1108 _name = '\033[0;34m
%s\033[0m
' % name
1114 # return the first matching group
1115 return next(g for g in mobj.groups() if g is not None)
1117 return mobj.group(group)
1118 elif default is not NO_DEFAULT:
1121 raise RegexNotFoundError('Unable to extract
%s' % _name)
1123 self.report_warning('unable to extract
%s' % _name + bug_reports_message())
1126 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1128 Like _search_regex, but strips HTML tags and unescapes entities.
1130 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1132 return clean_html(res).strip()
1136 def _get_netrc_login_info(self, netrc_machine=None):
1139 netrc_machine = netrc_machine or self._NETRC_MACHINE
1141 if self.get_param('usenetrc
', False):
1143 info = netrc.netrc().authenticators(netrc_machine)
1144 if info is not None:
1148 raise netrc.NetrcParseError(
1149 'No authenticators
for %s' % netrc_machine)
1150 except (IOError, netrc.NetrcParseError) as err:
1151 self.report_warning(
1152 'parsing
.netrc
: %s' % error_to_compat_str(err))
1154 return username, password
1156 def _get_login_info(self, username_option='username
', password_option='password
', netrc_machine=None):
1158 Get the login info as (username, password)
1159 First look for the manually specified credentials using username_option
1160 and password_option as keys in params dictionary. If no such credentials
1161 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1163 If there's no info available
, return (None, None)
1166 # Attempt to use provided username and password or .netrc data
1167 username = self.get_param(username_option)
1168 if username is not None:
1169 password = self.get_param(password_option)
1171 username, password = self._get_netrc_login_info(netrc_machine)
1173 return username, password
1175 def _get_tfa_info(self, note='two-factor verification code'):
1177 Get the two
-factor authentication info
1178 TODO
- asking the user will be required
for sms
/phone verify
1179 currently just uses the command line option
1180 If there
's no info available, return None
1183 tfa = self.get_param('twofactor
')
1187 return compat_getpass('Type
%s and press
[Return
]: ' % note)
1189 # Helper functions for extracting OpenGraph info
1191 def _og_regexes(prop):
1192 content_re = r'content
=(?
:"([^"]+?
)"|\'([^\']+?)\'|\s*([^\s"\'=<>`
]+?
))'
1193 property_re = (r'(?
:name|
property)=(?
:\'og
[:-]%(prop)s\'|
"og[:-]%(prop)s"|\s
*og
[:-]%(prop)s\b)'
1194 % {'prop': re.escape(prop)})
1195 template = r'<meta
[^
>]+?
%s[^
>]+?
%s'
1197 template % (property_re, content_re),
1198 template % (content_re, property_re),
1202 def _meta_regex(prop):
1203 return r'''(?isx)<meta
1204 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1205 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1207 def _og_search_property(self, prop, html, name=None, **kargs):
1208 if not isinstance(prop, (list, tuple)):
1211 name = 'OpenGraph
%s' % prop[0]
1214 og_regexes.extend(self._og_regexes(p))
1215 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1218 return unescapeHTML(escaped)
1220 def _og_search_thumbnail(self, html, **kargs):
1221 return self._og_search_property('image
', html, 'thumbnail URL
', fatal=False, **kargs)
1223 def _og_search_description(self, html, **kargs):
1224 return self._og_search_property('description
', html, fatal=False, **kargs)
1226 def _og_search_title(self, html, **kargs):
1227 return self._og_search_property('title
', html, **kargs)
1229 def _og_search_video_url(self, html, name='video url
', secure=True, **kargs):
1230 regexes = self._og_regexes('video
') + self._og_regexes('video
:url
')
1232 regexes = self._og_regexes('video
:secure_url
') + regexes
1233 return self._html_search_regex(regexes, html, name, **kargs)
1235 def _og_search_url(self, html, **kargs):
1236 return self._og_search_property('url
', html, **kargs)
1238 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1239 if not isinstance(name, (list, tuple)):
1241 if display_name is None:
1242 display_name = name[0]
1243 return self._html_search_regex(
1244 [self._meta_regex(n) for n in name],
1245 html, display_name, fatal=fatal, group='content
', **kwargs)
1247 def _dc_search_uploader(self, html):
1248 return self._html_search_meta('dc
.creator
', html, 'uploader
')
1250 def _rta_search(self, html):
1251 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1252 if re.search(r'(?ix
)<meta\s
+name
="rating"\s
+'
1253 r' content
="RTA-5042-1996-1400-1577-RTA"',
1258 def _media_rating_search(self, html):
1259 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1260 rating = self._html_search_meta('rating
', html)
1272 return RATING_TABLE.get(rating.lower())
1274 def _family_friendly_search(self, html):
1275 # See http://schema.org/VideoObject
1276 family_friendly = self._html_search_meta(
1277 'isFamilyFriendly
', html, default=None)
1279 if not family_friendly:
1288 return RATING_TABLE.get(family_friendly.lower())
1290 def _twitter_search_player(self, html):
1291 return self._html_search_meta('twitter
:player
', html,
1292 'twitter card player
')
1294 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
1295 json_ld_list = list(re.finditer(JSON_LD_RE, html))
1296 default = kwargs.get('default
', NO_DEFAULT)
1297 # JSON-LD may be malformed and thus `fatal` should be respected.
1298 # At the same time `default` may be passed that assumes `fatal=False`
1299 # for _search_regex. Let's simulate the same behavior here
as well
.
1300 fatal
= kwargs
.get('fatal', True) if default
== NO_DEFAULT
else False
1302 for mobj
in json_ld_list
:
1303 json_ld_item
= self
._parse
_json
(
1304 mobj
.group('json_ld'), video_id
, fatal
=fatal
)
1305 if not json_ld_item
:
1307 if isinstance(json_ld_item
, dict):
1308 json_ld
.append(json_ld_item
)
1309 elif isinstance(json_ld_item
, (list, tuple)):
1310 json_ld
.extend(json_ld_item
)
1312 json_ld
= self
._json
_ld
(json_ld
, video_id
, fatal
=fatal
, expected_type
=expected_type
)
1315 if default
is not NO_DEFAULT
:
1318 raise RegexNotFoundError('Unable to extract JSON-LD')
1320 self
.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1323 def _json_ld(self
, json_ld
, video_id
, fatal
=True, expected_type
=None):
1324 if isinstance(json_ld
, compat_str
):
1325 json_ld
= self
._parse
_json
(json_ld
, video_id
, fatal
=fatal
)
1329 if not isinstance(json_ld
, (list, tuple, dict)):
1331 if isinstance(json_ld
, dict):
1334 INTERACTION_TYPE_MAP
= {
1335 'CommentAction': 'comment',
1336 'AgreeAction': 'like',
1337 'DisagreeAction': 'dislike',
1338 'LikeAction': 'like',
1339 'DislikeAction': 'dislike',
1340 'ListenAction': 'view',
1341 'WatchAction': 'view',
1342 'ViewAction': 'view',
1345 def extract_interaction_type(e
):
1346 interaction_type
= e
.get('interactionType')
1347 if isinstance(interaction_type
, dict):
1348 interaction_type
= interaction_type
.get('@type')
1349 return str_or_none(interaction_type
)
1351 def extract_interaction_statistic(e
):
1352 interaction_statistic
= e
.get('interactionStatistic')
1353 if isinstance(interaction_statistic
, dict):
1354 interaction_statistic
= [interaction_statistic
]
1355 if not isinstance(interaction_statistic
, list):
1357 for is_e
in interaction_statistic
:
1358 if not isinstance(is_e
, dict):
1360 if is_e
.get('@type') != 'InteractionCounter':
1362 interaction_type
= extract_interaction_type(is_e
)
1363 if not interaction_type
:
1365 # For interaction count some sites provide string instead of
1366 # an integer (as per spec) with non digit characters (e.g. ",")
1367 # so extracting count with more relaxed str_to_int
1368 interaction_count
= str_to_int(is_e
.get('userInteractionCount'))
1369 if interaction_count
is None:
1371 count_kind
= INTERACTION_TYPE_MAP
.get(interaction_type
.split('/')[-1])
1374 count_key
= '%s_count' % count_kind
1375 if info
.get(count_key
) is not None:
1377 info
[count_key
] = interaction_count
1379 def extract_video_object(e
):
1380 assert e
['@type'] == 'VideoObject'
1381 author
= e
.get('author')
1383 'url': url_or_none(e
.get('contentUrl')),
1384 'title': unescapeHTML(e
.get('name')),
1385 'description': unescapeHTML(e
.get('description')),
1386 'thumbnail': url_or_none(e
.get('thumbnailUrl') or e
.get('thumbnailURL')),
1387 'duration': parse_duration(e
.get('duration')),
1388 'timestamp': unified_timestamp(e
.get('uploadDate')),
1389 # author can be an instance of 'Organization' or 'Person' types.
1390 # both types can have 'name' property(inherited from 'Thing' type). [1]
1391 # however some websites are using 'Text' type instead.
1392 # 1. https://schema.org/VideoObject
1393 'uploader': author
.get('name') if isinstance(author
, dict) else author
if isinstance(author
, compat_str
) else None,
1394 'filesize': float_or_none(e
.get('contentSize')),
1395 'tbr': int_or_none(e
.get('bitrate')),
1396 'width': int_or_none(e
.get('width')),
1397 'height': int_or_none(e
.get('height')),
1398 'view_count': int_or_none(e
.get('interactionCount')),
1400 extract_interaction_statistic(e
)
1404 item_type
= e
.get('@type')
1405 if expected_type
is not None and expected_type
!= item_type
:
1407 if item_type
in ('TVEpisode', 'Episode'):
1408 episode_name
= unescapeHTML(e
.get('name'))
1410 'episode': episode_name
,
1411 'episode_number': int_or_none(e
.get('episodeNumber')),
1412 'description': unescapeHTML(e
.get('description')),
1414 if not info
.get('title') and episode_name
:
1415 info
['title'] = episode_name
1416 part_of_season
= e
.get('partOfSeason')
1417 if isinstance(part_of_season
, dict) and part_of_season
.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1419 'season': unescapeHTML(part_of_season
.get('name')),
1420 'season_number': int_or_none(part_of_season
.get('seasonNumber')),
1422 part_of_series
= e
.get('partOfSeries') or e
.get('partOfTVSeries')
1423 if isinstance(part_of_series
, dict) and part_of_series
.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1424 info
['series'] = unescapeHTML(part_of_series
.get('name'))
1425 elif item_type
== 'Movie':
1427 'title': unescapeHTML(e
.get('name')),
1428 'description': unescapeHTML(e
.get('description')),
1429 'duration': parse_duration(e
.get('duration')),
1430 'timestamp': unified_timestamp(e
.get('dateCreated')),
1432 elif item_type
in ('Article', 'NewsArticle'):
1434 'timestamp': parse_iso8601(e
.get('datePublished')),
1435 'title': unescapeHTML(e
.get('headline')),
1436 'description': unescapeHTML(e
.get('articleBody')),
1438 elif item_type
== 'VideoObject':
1439 extract_video_object(e
)
1440 if expected_type
is None:
1444 video
= e
.get('video')
1445 if isinstance(video
, dict) and video
.get('@type') == 'VideoObject':
1446 extract_video_object(video
)
1447 if expected_type
is None:
1451 return dict((k
, v
) for k
, v
in info
.items() if v
is not None)
1454 def _hidden_inputs(html
):
1455 html
= re
.sub(r
'<!--(?:(?!<!--).)*-->', '', html
)
1457 for input in re
.findall(r
'(?i)(<input[^>]+>)', html
):
1458 attrs
= extract_attributes(input)
1461 if attrs
.get('type') not in ('hidden', 'submit'):
1463 name
= attrs
.get('name') or attrs
.get('id')
1464 value
= attrs
.get('value')
1465 if name
and value
is not None:
1466 hidden_inputs
[name
] = value
1467 return hidden_inputs
1469 def _form_hidden_inputs(self
, form_id
, html
):
1470 form
= self
._search
_regex
(
1471 r
'(?is)<form[^>]+?id=(["\'])%s\
1[^
>]*>(?P
<form
>.+?
)</form
>' % form_id,
1472 html, '%s form
' % form_id, group='form
')
1473 return self._hidden_inputs(form)
1476 regex = r' *((?P
<reverse
>\
+)?
(?P
<field
>[a
-zA
-Z0
-9_]+)((?P
<separator
>[~
:])(?P
<limit
>.*?
))?
)?
*$
'
1478 default = ('hidden
', 'aud_or_vid
', 'hasvid
', 'ie_pref
', 'lang
', 'quality
',
1479 'res
', 'fps
', 'codec
:vp9
.2
', 'size
', 'br
', 'asr
',
1480 'proto
', 'ext
', 'hasaud
', 'source
', 'format_id
') # These must not be aliases
1481 ytdl_default = ('hasaud
', 'quality
', 'tbr
', 'filesize
', 'vbr
',
1482 'height
', 'width
', 'proto
', 'vext
', 'abr
', 'aext
',
1483 'fps
', 'fs_approx
', 'source
', 'format_id
')
1486 'vcodec
': {'type': 'ordered
', 'regex
': True,
1487 'order
': ['av0?
1', 'vp0?
9.2', 'vp0?
9', '[hx
]265|he?vc?
', '[hx
]264|avc
', 'vp0?
8', 'mp4v|h263
', 'theora
', '', None, 'none
']},
1488 'acodec
': {'type': 'ordered
', 'regex
': True,
1489 'order
': ['opus
', 'vorbis
', 'aac
', 'mp?
4a?
', 'mp3
', 'e?a?c
-?
3', 'dts
', '', None, 'none
']},
1490 'proto
': {'type': 'ordered
', 'regex
': True, 'field
': 'protocol
',
1491 'order
': ['(ht|f
)tps
', '(ht|f
)tp$
', 'm3u8
.+', '.*dash
', 'ws|websocket
', '', 'mms|rtsp
', 'none
', 'f4
']},
1492 'vext
': {'type': 'ordered
', 'field
': 'video_ext
',
1493 'order
': ('mp4
', 'webm
', 'flv
', '', 'none
'),
1494 'order_free
': ('webm
', 'mp4
', 'flv
', '', 'none
')},
1495 'aext
': {'type': 'ordered
', 'field
': 'audio_ext
',
1496 'order
': ('m4a
', 'aac
', 'mp3
', 'ogg
', 'opus
', 'webm
', '', 'none
'),
1497 'order_free
': ('opus
', 'ogg
', 'webm
', 'm4a
', 'mp3
', 'aac
', '', 'none
')},
1498 'hidden
': {'visible': False, 'forced': True, 'type': 'extractor', 'max': -1000},
1499 'aud_or_vid
': {'visible
': False, 'forced
': True, 'type': 'multiple
', 'default
': 1,
1500 'field
': ('vcodec
', 'acodec
'),
1501 'function
': lambda it: int(any(v != 'none
' for v in it))},
1502 'ie_pref
': {'priority': True, 'type': 'extractor'},
1503 'hasvid
': {'priority': True, 'field': 'vcodec', 'type': 'boolean', 'not_in_list': ('none',)},
1504 'hasaud
': {'field': 'acodec', 'type': 'boolean', 'not_in_list': ('none',)},
1505 'lang
': {'priority': True, 'convert': 'ignore', 'field': 'language_preference'},
1506 'quality
': {'convert': 'float_none', 'default': -1},
1507 'filesize
': {'convert': 'bytes'},
1508 'fs_approx
': {'convert': 'bytes', 'field': 'filesize_approx'},
1509 'id': {'convert': 'string', 'field': 'format_id'},
1510 'height
': {'convert': 'float_none'},
1511 'width
': {'convert': 'float_none'},
1512 'fps
': {'convert': 'float_none'},
1513 'tbr
': {'convert': 'float_none'},
1514 'vbr
': {'convert': 'float_none'},
1515 'abr
': {'convert': 'float_none'},
1516 'asr
': {'convert': 'float_none'},
1517 'source
': {'convert': 'ignore', 'field': 'source_preference'},
1519 'codec
': {'type': 'combined', 'field': ('vcodec', 'acodec')},
1520 'br
': {'type': 'combined', 'field': ('tbr', 'vbr', 'abr'), 'same_limit': True},
1521 'size
': {'type': 'combined', 'same_limit': True, 'field': ('filesize', 'fs_approx')},
1522 'ext
': {'type': 'combined', 'field': ('vext', 'aext')},
1523 'res
': {'type': 'multiple', 'field': ('height', 'width'), 'function': min},
1525 # Most of these exist only for compatibility reasons
1526 'dimension
': {'type': 'alias', 'field': 'res'},
1527 'resolution
': {'type': 'alias', 'field': 'res'},
1528 'extension
': {'type': 'alias', 'field': 'ext'},
1529 'bitrate
': {'type': 'alias', 'field': 'br'},
1530 'total_bitrate
': {'type': 'alias', 'field': 'tbr'},
1531 'video_bitrate
': {'type': 'alias', 'field': 'vbr'},
1532 'audio_bitrate
': {'type': 'alias', 'field': 'abr'},
1533 'framerate
': {'type': 'alias', 'field': 'fps'},
1534 'language_preference
': {'type': 'alias', 'field': 'lang'}, # not named as 'language
' because such a field exists
1535 'protocol
': {'type': 'alias', 'field': 'proto'},
1536 'source_preference
': {'type': 'alias', 'field': 'source'},
1537 'filesize_approx
': {'type': 'alias', 'field': 'fs_approx'},
1538 'filesize_estimate
': {'type': 'alias', 'field': 'size'},
1539 'samplerate
': {'type': 'alias', 'field': 'asr'},
1540 'video_ext
': {'type': 'alias', 'field': 'vext'},
1541 'audio_ext
': {'type': 'alias', 'field': 'aext'},
1542 'video_codec
': {'type': 'alias', 'field': 'vcodec'},
1543 'audio_codec
': {'type': 'alias', 'field': 'acodec'},
1544 'video
': {'type': 'alias', 'field': 'hasvid'},
1545 'has_video
': {'type': 'alias', 'field': 'hasvid'},
1546 'audio
': {'type': 'alias', 'field': 'hasaud'},
1547 'has_audio
': {'type': 'alias', 'field': 'hasaud'},
1548 'extractor
': {'type': 'alias', 'field': 'ie_pref'},
1549 'preference
': {'type': 'alias', 'field': 'ie_pref'},
1550 'extractor_preference
': {'type': 'alias', 'field': 'ie_pref'},
1551 'format_id
': {'type': 'alias', 'field': 'id'},
1556 def _get_field_setting(self, field, key):
1557 if field not in self.settings:
1558 self.settings[field] = {}
1559 propObj = self.settings[field]
1560 if key not in propObj:
1561 type = propObj.get('type')
1563 default = 'preference
' if type == 'extractor
' else (field,) if type in ('combined
', 'multiple
') else field
1564 elif key == 'convert
':
1565 default = 'order
' if type == 'ordered
' else 'float_string
' if field else 'ignore
'
1567 default = {'type': 'field', 'visible': True, 'order': [], 'not_in_list': (None,), 'function': max}.get(key, None)
1568 propObj[key] = default
1571 def _resolve_field_value(self, field, value, convertNone=False):
1576 value = value.lower()
1577 conversion = self._get_field_setting(field, 'convert
')
1578 if conversion == 'ignore
':
1580 if conversion == 'string
':
1582 elif conversion == 'float_none
':
1583 return float_or_none(value)
1584 elif conversion == 'bytes':
1585 return FileDownloader.parse_bytes(value)
1586 elif conversion == 'order
':
1587 order_list = (self._use_free_order and self._get_field_setting(field, 'order_free
')) or self._get_field_setting(field, 'order
')
1588 use_regex = self._get_field_setting(field, 'regex
')
1589 list_length = len(order_list)
1590 empty_pos = order_list.index('') if '' in order_list else list_length + 1
1591 if use_regex and value is not None:
1592 for i, regex in enumerate(order_list):
1593 if regex and re.match(regex, value):
1594 return list_length - i
1595 return list_length - empty_pos # not in list
1596 else: # not regex or value = None
1597 return list_length - (order_list.index(value) if value in order_list else empty_pos)
1599 if value.isnumeric():
1602 self.settings[field]['convert
'] = 'string
'
1605 def evaluate_params(self, params, sort_extractor):
1606 self._use_free_order = params.get('prefer_free_formats
', False)
1607 self._sort_user = params.get('format_sort
', [])
1608 self._sort_extractor = sort_extractor
1610 def add_item(field, reverse, closest, limit_text):
1611 field = field.lower()
1612 if field in self._order:
1614 self._order.append(field)
1615 limit = self._resolve_field_value(field, limit_text)
1618 'closest
': False if limit is None else closest,
1619 'limit_text
': limit_text,
1621 if field in self.settings:
1622 self.settings[field].update(data)
1624 self.settings[field] = data
1627 tuple(field for field in self.default if self._get_field_setting(field, 'forced
'))
1628 + (tuple() if params.get('format_sort_force
', False)
1629 else tuple(field for field in self.default if self._get_field_setting(field, 'priority
')))
1630 + tuple(self._sort_user) + tuple(sort_extractor) + self.default)
1632 for item in sort_list:
1633 match = re.match(self.regex, item)
1635 raise ExtractorError('Invalid format sort string
"%s" given by extractor
' % item)
1636 field = match.group('field
')
1639 if self._get_field_setting(field, 'type') == 'alias
':
1640 field = self._get_field_setting(field, 'field
')
1641 reverse = match.group('reverse
') is not None
1642 closest = match.group('separator
') == '~
'
1643 limit_text = match.group('limit
')
1645 has_limit = limit_text is not None
1646 has_multiple_fields = self._get_field_setting(field, 'type') == 'combined
'
1647 has_multiple_limits = has_limit and has_multiple_fields and not self._get_field_setting(field, 'same_limit
')
1649 fields = self._get_field_setting(field, 'field
') if has_multiple_fields else (field,)
1650 limits = limit_text.split(":") if has_multiple_limits else (limit_text,) if has_limit else tuple()
1651 limit_count = len(limits)
1652 for (i, f) in enumerate(fields):
1653 add_item(f, reverse, closest,
1654 limits[i] if i < limit_count
1655 else limits[0] if has_limit and not has_multiple_limits
1658 def print_verbose_info(self, write_debug):
1660 write_debug('Sort order given by user
: %s' % ', '.join(self._sort_user))
1661 if self._sort_extractor:
1662 write_debug('Sort order given by extractor
: %s' % ', '.join(self._sort_extractor))
1663 write_debug('Formats
sorted by
: %s' % ', '.join(['%s%s%s' % (
1664 '+' if self._get_field_setting(field, 'reverse
') else '', field,
1665 '%s%s(%s)' % ('~
' if self._get_field_setting(field, 'closest
') else ':',
1666 self._get_field_setting(field, 'limit_text
'),
1667 self._get_field_setting(field, 'limit
'))
1668 if self._get_field_setting(field, 'limit_text
') is not None else '')
1669 for field in self._order if self._get_field_setting(field, 'visible
')]))
1671 def _calculate_field_preference_from_value(self, format, field, type, value):
1672 reverse = self._get_field_setting(field, 'reverse
')
1673 closest = self._get_field_setting(field, 'closest
')
1674 limit = self._get_field_setting(field, 'limit
')
1676 if type == 'extractor
':
1677 maximum = self._get_field_setting(field, 'max')
1678 if value is None or (maximum is not None and value >= maximum):
1680 elif type == 'boolean
':
1681 in_list = self._get_field_setting(field, 'in_list
')
1682 not_in_list = self._get_field_setting(field, 'not_in_list
')
1683 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
1684 elif type == 'ordered
':
1685 value = self._resolve_field_value(field, value, True)
1687 # try to convert to number
1688 val_num = float_or_none(value, default=self._get_field_setting(field, 'default
'))
1689 is_num = self._get_field_setting(field, 'convert
') != 'string
' and val_num is not None
1693 return ((-10, 0) if value is None
1694 else (1, value, 0) if not is_num # if a field has mixed strings and numbers, strings are sorted higher
1695 else (0, -abs(value - limit), value - limit if reverse else limit - value) if closest
1696 else (0, value, 0) if not reverse and (limit is None or value <= limit)
1697 else (0, -value, 0) if limit is None or (reverse and value == limit) or value > limit
1698 else (-1, value, 0))
1700 def _calculate_field_preference(self, format, field):
1701 type = self._get_field_setting(field, 'type') # extractor, boolean, ordered, field, multiple
1702 get_value = lambda f: format.get(self._get_field_setting(f, 'field
'))
1703 if type == 'multiple
':
1704 type = 'field
' # Only 'field
' is allowed in multiple for now
1705 actual_fields = self._get_field_setting(field, 'field
')
1707 def wrapped_function(values):
1708 values = tuple(filter(lambda x: x is not None, values))
1709 return self._get_field_setting(field, 'function
')(values) if values else None
1711 value = wrapped_function((get_value(f) for f in actual_fields))
1713 value = get_value(field)
1714 return self._calculate_field_preference_from_value(format, field, type, value)
1716 def calculate_preference(self, format):
1717 # Determine missing protocol
1718 if not format.get('protocol
'):
1719 format['protocol
'] = determine_protocol(format)
1721 # Determine missing ext
1722 if not format.get('ext
') and 'url
' in format:
1723 format['ext
'] = determine_ext(format['url
'])
1724 if format.get('vcodec
') == 'none
':
1725 format['audio_ext
'] = format['ext
'] if format.get('acodec
') != 'none
' else 'none
'
1726 format['video_ext
'] = 'none
'
1728 format['video_ext
'] = format['ext
']
1729 format['audio_ext
'] = 'none
'
1730 # if format.get('preference
') is None and format.get('ext
') in ('f4f
', 'f4m
'): # Not supported?
1731 # format['preference
'] = -1000
1733 # Determine missing bitrates
1734 if format.get('tbr
') is None:
1735 if format.get('vbr
') is not None and format.get('abr
') is not None:
1736 format['tbr
'] = format.get('vbr
', 0) + format.get('abr
', 0)
1738 if format.get('vcodec
') != "none" and format.get('vbr
') is None:
1739 format['vbr
'] = format.get('tbr
') - format.get('abr
', 0)
1740 if format.get('acodec
') != "none" and format.get('abr
') is None:
1741 format['abr
'] = format.get('tbr
') - format.get('vbr
', 0)
1743 return tuple(self._calculate_field_preference(format, field) for field in self._order)
1745 def _sort_formats(self, formats, field_preference=[]):
1747 if self.get_param('ignore_no_formats_error
'):
1749 raise ExtractorError('No video formats found
')
1750 format_sort = self.FormatSort() # params and to_screen are taken from the downloader
1751 format_sort.evaluate_params(self._downloader.params, field_preference)
1752 if self.get_param('verbose
', False):
1753 format_sort.print_verbose_info(self._downloader.write_debug)
1754 formats.sort(key=lambda f: format_sort.calculate_preference(f))
1756 def _check_formats(self, formats, video_id):
1758 formats[:] = filter(
1759 lambda f: self._is_valid_url(
1761 item='%s video format
' % f.get('format_id
') if f.get('format_id
') else 'video
'),
1765 def _remove_duplicate_formats(formats):
1769 if f['url
'] not in format_urls:
1770 format_urls.add(f['url
'])
1771 unique_formats.append(f)
1772 formats[:] = unique_formats
1774 def _is_valid_url(self, url, video_id, item='video
', headers={}):
1775 url = self._proto_relative_url(url, scheme='http
:')
1776 # For now assume non HTTP(S) URLs always valid
1777 if not (url.startswith('http
://') or url.startswith('https
://')):
1780 self._request_webpage(url, video_id, 'Checking
%s URL
' % item, headers=headers)
1782 except ExtractorError as e:
1784 '%s: %s URL
is invalid
, skipping
: %s'
1785 % (video_id, item, error_to_compat_str(e.cause)))
1788 def http_scheme(self):
1789 """ Either "http:" or "https:", depending on the user's preferences
"""
1792 if self.get_param('prefer_insecure', False)
1795 def _proto_relative_url(self, url, scheme=None):
1798 if url.startswith('//'):
1800 scheme = self.http_scheme()
1805 def _sleep(self, timeout, video_id, msg_template=None):
1806 if msg_template is None:
1807 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1808 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1812 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1813 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1814 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
1815 manifest = self._download_xml(
1816 manifest_url, video_id, 'Downloading f4m manifest',
1817 'Unable to download f4m manifest',
1818 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1819 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
1820 transform_source=transform_source,
1821 fatal=fatal, data=data, headers=headers, query=query)
1823 if manifest is False:
1826 return self._parse_f4m_formats(
1827 manifest, manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1828 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1830 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1831 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1832 fatal=True, m3u8_id=None):
1833 if not isinstance(manifest, compat_etree_Element) and not fatal:
1836 # currently yt-dlp cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1837 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1838 if akamai_pv is not None and ';' in akamai_pv.text:
1839 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1840 if playerVerificationChallenge.strip() != '':
1844 manifest_version = '1.0'
1845 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1847 manifest_version = '2.0'
1848 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1849 # Remove unsupported DRM protected media from final formats
1850 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
1851 media_nodes = remove_encrypted_media(media_nodes)
1855 manifest_base_url = get_base_url(manifest)
1857 bootstrap_info = xpath_element(
1858 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1859 'bootstrap info', default=None)
1862 mime_type = xpath_text(
1863 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1864 'base URL', default=None)
1865 if mime_type and mime_type.startswith('audio/'):
1868 for i, media_el in enumerate(media_nodes):
1869 tbr = int_or_none(media_el.attrib.get('bitrate'))
1870 width = int_or_none(media_el.attrib.get('width'))
1871 height = int_or_none(media_el.attrib.get('height'))
1872 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1873 # If <bootstrapInfo> is present, the specified f4m is a
1874 # stream-level manifest, and only set-level manifests may refer to
1875 # external resources. See section 11.4 and section 4 of F4M spec
1876 if bootstrap_info is None:
1878 # @href is introduced in 2.0, see section 11.6 of F4M spec
1879 if manifest_version == '2.0':
1880 media_url = media_el.attrib.get('href')
1881 if media_url is None:
1882 media_url = media_el.attrib.get('url')
1886 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1887 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1888 # If media_url is itself a f4m manifest do the recursive extraction
1889 # since bitrates in parent manifest (this one) and media_url manifest
1890 # may differ leading to inability to resolve the format by requested
1891 # bitrate in f4m downloader
1892 ext = determine_ext(manifest_url)
1894 f4m_formats = self._extract_f4m_formats(
1895 manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1896 transform_source=transform_source, fatal=fatal)
1897 # Sometimes stream-level manifest contains single media entry that
1898 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1899 # At the same time parent's media entry in set-level manifest may
1900 # contain it. We will copy it from parent in such cases.
1901 if len(f4m_formats) == 1:
1904 'tbr': f.get('tbr') or tbr,
1905 'width': f.get('width') or width,
1906 'height': f.get('height') or height,
1907 'format_id': f.get('format_id') if not tbr else format_id,
1910 formats.extend(f4m_formats)
1913 formats.extend(self._extract_m3u8_formats(
1914 manifest_url, video_id, 'mp4', preference=preference,
1915 quality=quality, m3u8_id=m3u8_id, fatal=fatal))
1918 'format_id': format_id,
1919 'url': manifest_url,
1920 'manifest_url': manifest_url,
1921 'ext': 'flv' if bootstrap_info is not None else None,
1927 'preference': preference,
1932 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, quality=None, m3u8_id=None):
1934 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1938 'preference': preference - 100 if preference else -100,
1940 'resolution': 'multiple',
1941 'format_note': 'Quality selection URL',
1944 def _extract_m3u8_formats(self, *args, **kwargs):
1945 fmts, subs = self._extract_m3u8_formats_and_subtitles(*args, **kwargs)
1947 self.report_warning(bug_reports_message(
1948 "Ignoring subtitle tracks found in the HLS manifest; "
1949 "if any subtitle tracks are missing,"
1953 def _extract_m3u8_formats_and_subtitles(
1954 self, m3u8_url, video_id, ext=None, entry_protocol='m3u8_native',
1955 preference=None, quality=None, m3u8_id=None, note=None,
1956 errnote=None, fatal=True, live=False, data=None, headers={},
1959 res = self._download_webpage_handle(
1961 note='Downloading m3u8 information' if note is None else note,
1962 errnote='Failed to download m3u8 information' if errnote is None else errnote,
1963 fatal=fatal, data=data, headers=headers, query=query)
1968 m3u8_doc, urlh = res
1969 m3u8_url = urlh.geturl()
1971 return self._parse_m3u8_formats_and_subtitles(
1972 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
1973 preference=preference, quality=quality, m3u8_id=m3u8_id,
1974 note=note, errnote=errnote, fatal=fatal, live=live, data=data,
1975 headers=headers, query=query, video_id=video_id)
1977 def _parse_m3u8_formats_and_subtitles(
1978 self, m3u8_doc, m3u8_url, ext=None, entry_protocol='m3u8_native',
1979 preference=None, quality=None, m3u8_id=None, live=False, note=None,
1980 errnote=None, fatal=True, data=None, headers={}, query={},
1983 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
1986 if (not self.get_param('allow_unplayable_formats')
1987 and re.search(r'#EXT-X-SESSION-KEY:.*?URI="skd://', m3u8_doc)): # Apple FairPlay
1994 format_url = lambda u: (
1996 if re.match(r'^https?://', u)
1997 else compat_urlparse.urljoin(m3u8_url, u))
1999 split_discontinuity = self.get_param('hls_split_discontinuity', False)
2002 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
2003 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
2004 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
2006 # We should try extracting formats only from master playlists [1, 4.3.4],
2007 # i.e. playlists that describe available qualities. On the other hand
2008 # media playlists [1, 4.3.3] should be returned as is since they contain
2009 # just the media without qualities renditions.
2010 # Fortunately, master playlist can be easily distinguished from media
2011 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
2012 # master playlist tags MUST NOT appear in a media playlist and vice versa.
2013 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
2014 # media playlist and MUST NOT appear in master playlist thus we can
2015 # clearly detect media playlist with this criterion.
2017 def _extract_m3u8_playlist_formats(format_url=None, m3u8_doc=None, video_id=None,
2018 fatal=True, data=None, headers={}):
2022 res = self._download_webpage_handle(
2023 format_url, video_id,
2025 errnote='Failed to download m3u8 playlist information',
2026 fatal=fatal, data=data, headers=headers)
2031 m3u8_doc, urlh = res
2032 format_url = urlh.geturl()
2034 playlist_formats = []
2037 if split_discontinuity
2044 for line in m3u8_doc.splitlines():
2045 if not line.startswith('#'):
2046 format_info['files'].append(line)
2047 elif split_discontinuity and line.startswith('#EXT-X-DISCONTINUITY'):
2049 playlist_formats.append(format_info)
2055 playlist_formats.append(format_info)
2056 return playlist_formats
2058 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
2060 playlist_formats = _extract_m3u8_playlist_formats(m3u8_doc=m3u8_doc)
2062 for format in playlist_formats:
2065 format_id.append(m3u8_id)
2066 format_index = format.get('index')
2068 format_id.append(str(format_index))
2070 'format_id': '-'.join(format_id),
2071 'format_index': format_index,
2074 'protocol': entry_protocol,
2075 'preference': preference,
2080 return formats, subtitles
2083 last_stream_inf = {}
2085 def extract_media(x_media_line):
2086 media = parse_m3u8_attributes(x_media_line)
2087 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
2088 media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME')
2089 if not (media_type and group_id and name):
2091 groups.setdefault(group_id, []).append(media)
2092 # <https://tools.ietf.org/html/rfc8216#section-4.3.4.1>
2093 if media_type == 'SUBTITLES':
2094 # According to RFC 8216 §4.3.4.2.1, URI is REQUIRED in the
2095 # EXT-X-MEDIA tag if the media type is SUBTITLES.
2096 # However, lack of URI has been spotted in the wild.
2097 # e.g. NebulaIE; see https://github.com/yt-dlp/yt-dlp/issues/339
2098 if not media.get('URI'):
2100 url = format_url(media['URI'])
2103 'ext': determine_ext(url),
2105 if sub_info['ext'] == 'm3u8':
2106 # Per RFC 8216 §3.1, the only possible subtitle format m3u8
2107 # files may contain is WebVTT:
2108 # <https://tools.ietf.org/html/rfc8216#section-3.1>
2109 sub_info['ext'] = 'vtt'
2110 sub_info['protocol'] = 'm3u8_native'
2111 lang = media.get('LANGUAGE') or 'und'
2112 subtitles.setdefault(lang, []).append(sub_info)
2113 if media_type not in ('VIDEO', 'AUDIO'):
2115 media_url = media.get('URI')
2117 manifest_url = format_url(media_url)
2119 playlist_formats = _extract_m3u8_playlist_formats(manifest_url, video_id=video_id,
2120 fatal=fatal, data=data, headers=headers)
2122 for format in playlist_formats:
2123 format_index = format.get('index')
2124 for v in (m3u8_id, group_id, name):
2128 format_id.append(str(format_index))
2130 'format_id': '-'.join(format_id),
2131 'format_note': name,
2132 'format_index': format_index,
2133 'url': manifest_url,
2134 'manifest_url': m3u8_url,
2135 'language': media.get('LANGUAGE'),
2137 'protocol': entry_protocol,
2138 'preference': preference,
2141 if media_type == 'AUDIO':
2142 f['vcodec'] = 'none'
2145 def build_stream_name():
2146 # Despite specification does not mention NAME attribute for
2147 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
2148 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
2149 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
2150 stream_name = last_stream_inf.get('NAME')
2153 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
2154 # from corresponding rendition group
2155 stream_group_id = last_stream_inf.get('VIDEO')
2156 if not stream_group_id:
2158 stream_group = groups.get(stream_group_id)
2159 if not stream_group:
2160 return stream_group_id
2161 rendition = stream_group[0]
2162 return rendition.get('NAME') or stream_group_id
2164 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
2165 # chance to detect video only formats when EXT-X-STREAM-INF tags
2166 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
2167 for line in m3u8_doc.splitlines():
2168 if line.startswith('#EXT-X-MEDIA:'):
2171 for line in m3u8_doc.splitlines():
2172 if line.startswith('#EXT-X-STREAM-INF:'):
2173 last_stream_inf = parse_m3u8_attributes(line)
2174 elif line.startswith('#') or not line.strip():
2177 tbr = float_or_none(
2178 last_stream_inf.get('AVERAGE-BANDWIDTH')
2179 or last_stream_inf.get('BANDWIDTH'), scale=1000)
2180 manifest_url = format_url(line.strip())
2182 playlist_formats = _extract_m3u8_playlist_formats(manifest_url, video_id=video_id,
2183 fatal=fatal, data=data, headers=headers)
2185 for frmt in playlist_formats:
2188 format_id.append(m3u8_id)
2189 format_index = frmt.get('index')
2190 stream_name = build_stream_name()
2191 # Bandwidth of live streams may differ over time thus making
2192 # format_id unpredictable. So it's better to keep provided
2195 format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
2197 format_id.append(str(format_index))
2199 'format_id': '-'.join(format_id),
2200 'format_index': format_index,
2201 'url': manifest_url,
2202 'manifest_url': m3u8_url,
2205 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')),
2206 'protocol': entry_protocol,
2207 'preference': preference,
2210 resolution = last_stream_inf.get('RESOLUTION')
2212 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
2214 f['width'] = int(mobj.group('width'))
2215 f['height'] = int(mobj.group('height'))
2216 # Unified Streaming Platform
2218 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
2220 abr, vbr = mobj.groups()
2221 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
2226 codecs = parse_codecs(last_stream_inf.get('CODECS'))
2228 audio_group_id = last_stream_inf.get('AUDIO')
2229 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
2230 # references a rendition group MUST have a CODECS attribute.
2231 # However, this is not always respected, for example, [2]
2232 # contains EXT-X-STREAM-INF tag which references AUDIO
2233 # rendition group but does not have CODECS and despite
2234 # referencing an audio group it represents a complete
2235 # (with audio and video) format. So, for such cases we will
2236 # ignore references to rendition groups and treat them
2237 # as complete formats.
2238 if audio_group_id and codecs and f.get('vcodec') != 'none':
2239 audio_group = groups.get(audio_group_id)
2240 if audio_group and audio_group[0].get('URI'):
2241 # TODO: update acodec for audio only formats with
2243 f['acodec'] = 'none'
2244 if not f.get('ext'):
2245 f['ext'] = 'm4a' if f.get('vcodec') == 'none' else 'mp4'
2249 progressive_uri = last_stream_inf.get('PROGRESSIVE-URI')
2252 del http_f['manifest_url']
2254 'format_id': f['format_id'].replace('hls-', 'http-'),
2256 'url': progressive_uri,
2258 formats.append(http_f)
2260 last_stream_inf = {}
2261 return formats, subtitles
2264 def _xpath_ns(path, namespace=None):
2268 for c in path.split('/'):
2269 if not c or c == '.':
2272 out.append('{%s}%s' % (namespace, c))
2273 return '/'.join(out)
2275 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
2276 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
2282 namespace = self._parse_smil_namespace(smil)
2284 return self._parse_smil_formats(
2285 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
2287 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
2288 smil = self._download_smil(smil_url, video_id, fatal=fatal)
2291 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
2293 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
2294 return self._download_xml(
2295 smil_url, video_id, 'Downloading SMIL file',
2296 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
2298 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
2299 namespace = self._parse_smil_namespace(smil)
2301 formats = self._parse_smil_formats(
2302 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
2303 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
2305 video_id = os.path.splitext(url_basename(smil_url))[0]
2309 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
2310 name = meta.attrib.get('name')
2311 content = meta.attrib.get('content')
2312 if not name or not content:
2314 if not title and name == 'title':
2316 elif not description and name in ('description', 'abstract'):
2317 description = content
2318 elif not upload_date and name == 'date':
2319 upload_date = unified_strdate(content)
2322 'id': image.get('type'),
2323 'url': image.get('src'),
2324 'width': int_or_none(image.get('width')),
2325 'height': int_or_none(image.get('height')),
2326 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
2330 'title': title or video_id,
2331 'description': description,
2332 'upload_date': upload_date,
2333 'thumbnails': thumbnails,
2335 'subtitles': subtitles,
2338 def _parse_smil_namespace(self, smil):
2339 return self._search_regex(
2340 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
2342 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
2344 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
2345 b = meta.get('base') or meta.get('httpBase')
2356 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
2357 for medium in media:
2358 src = medium.get('src')
2359 if not src or src in srcs:
2363 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
2364 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
2365 width = int_or_none(medium.get('width'))
2366 height = int_or_none(medium.get('height'))
2367 proto = medium.get('proto')
2368 ext = medium.get('ext')
2369 src_ext = determine_ext(src)
2370 streamer = medium.get('streamer') or base
2372 if proto == 'rtmp' or streamer.startswith('rtmp'):
2378 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
2380 'filesize': filesize,
2384 if transform_rtmp_url:
2385 streamer, src = transform_rtmp_url(streamer, src)
2386 formats[-1].update({
2392 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
2393 src_url = src_url.strip()
2395 if proto == 'm3u8' or src_ext == 'm3u8':
2396 m3u8_formats = self._extract_m3u8_formats(
2397 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
2398 if len(m3u8_formats) == 1:
2400 m3u8_formats[0].update({
2401 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
2406 formats.extend(m3u8_formats)
2407 elif src_ext == 'f4m':
2412 'plugin': 'flowplayer-3.2.0.1',
2414 f4m_url += '&' if '?' in f4m_url else '?'
2415 f4m_url += compat_urllib_parse_urlencode(f4m_params)
2416 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
2417 elif src_ext == 'mpd':
2418 formats.extend(self._extract_mpd_formats(
2419 src_url, video_id, mpd_id='dash', fatal=False))
2420 elif re.search(r'\.ism/[Mm]anifest', src_url):
2421 formats.extend(self._extract_ism_formats(
2422 src_url, video_id, ism_id='mss', fatal=False))
2423 elif src_url.startswith('http') and self._is_valid_url(src, video_id):
2427 'ext': ext or src_ext or 'flv',
2428 'format_id': 'http-%d' % (bitrate or http_count),
2430 'filesize': filesize,
2437 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
2440 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
2441 src = textstream.get('src')
2442 if not src or src in urls:
2445 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
2446 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
2447 subtitles.setdefault(lang, []).append({
2453 def _extract_xspf_playlist(self, xspf_url, playlist_id, fatal=True):
2454 xspf = self._download_xml(
2455 xspf_url, playlist_id, 'Downloading xpsf playlist',
2456 'Unable to download xspf manifest', fatal=fatal)
2459 return self._parse_xspf(
2460 xspf, playlist_id, xspf_url=xspf_url,
2461 xspf_base_url=base_url(xspf_url))
2463 def _parse_xspf(self, xspf_doc, playlist_id, xspf_url=None, xspf_base_url=None):
2465 'xspf': 'http://xspf.org/ns/0/',
2466 's1': 'http://static.streamone.nl/player/ns/0',
2470 for track in xspf_doc.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
2472 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
2473 description = xpath_text(
2474 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
2475 thumbnail = xpath_text(
2476 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
2477 duration = float_or_none(
2478 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
2481 for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP)):
2482 format_url = urljoin(xspf_base_url, location.text)
2487 'manifest_url': xspf_url,
2488 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
2489 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
2490 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
2492 self._sort_formats(formats)
2497 'description': description,
2498 'thumbnail': thumbnail,
2499 'duration': duration,
2504 def _extract_mpd_formats(self, *args, **kwargs):
2505 fmts, subs = self._extract_mpd_formats_and_subtitles(*args, **kwargs)
2507 self.report_warning(bug_reports_message(
2508 "Ignoring subtitle tracks found in the DASH manifest; "
2509 "if any subtitle tracks are missing,"
2513 def _extract_mpd_formats_and_subtitles(
2514 self, mpd_url, video_id, mpd_id=None, note=None, errnote=None,
2515 fatal=True, data=None, headers={}, query={}):
2516 res = self._download_xml_handle(
2518 note='Downloading MPD manifest' if note is None else note,
2519 errnote='Failed to download MPD manifest' if errnote is None else errnote,
2520 fatal=fatal, data=data, headers=headers, query=query)
2526 mpd_base_url = base_url(urlh.geturl())
2528 return self._parse_mpd_formats_and_subtitles(
2529 mpd_doc, mpd_id, mpd_base_url, mpd_url)
2531 def _parse_mpd_formats(self, *args, **kwargs):
2532 fmts, subs = self._parse_mpd_formats_and_subtitles(*args, **kwargs)
2534 self.report_warning(bug_reports_message(
2535 "Ignoring subtitle tracks found in the DASH manifest; "
2536 "if any subtitle tracks are missing,"
2540 def _parse_mpd_formats_and_subtitles(
2541 self, mpd_doc, mpd_id=None, mpd_base_url='', mpd_url=None):
2543 Parse formats
from MPD manifest
.
2545 1. MPEG
-DASH Standard
, ISO
/IEC
23009-1:2014(E
),
2546 http
://standards
.iso
.org
/ittf
/PubliclyAvailableStandards
/c065274_ISO_IEC_23009
-1_2014.zip
2547 2. https
://en
.wikipedia
.org
/wiki
/Dynamic_Adaptive_Streaming_over_HTTP
2549 if not self.get_param('dynamic_mpd', True):
2550 if mpd_doc.get('type') == 'dynamic':
2553 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
2556 return self._xpath_ns(path, namespace)
2558 def is_drm_protected(element):
2559 return element.find(_add_ns('ContentProtection')) is not None
2561 def extract_multisegment_info(element, ms_parent_info):
2562 ms_info = ms_parent_info.copy()
2564 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2565 # common attributes and elements. We will only extract relevant
2567 def extract_common(source):
2568 segment_timeline = source.find(_add_ns('SegmentTimeline'))
2569 if segment_timeline is not None:
2570 s_e = segment_timeline.findall(_add_ns('S'))
2572 ms_info['total_number'] = 0
2575 r = int(s.get('r', 0))
2576 ms_info['total_number'] += 1 + r
2577 ms_info['s'].append({
2578 't': int(s.get('t', 0)),
2579 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2580 'd': int(s.attrib['d']),
2583 start_number = source.get('startNumber')
2585 ms_info['start_number'] = int(start_number)
2586 timescale = source.get('timescale')
2588 ms_info['timescale'] = int(timescale)
2589 segment_duration = source.get('duration')
2590 if segment_duration:
2591 ms_info['segment_duration'] = float(segment_duration)
2593 def extract_Initialization(source):
2594 initialization = source.find(_add_ns('Initialization'))
2595 if initialization is not None:
2596 ms_info['initialization_url'] = initialization.attrib['sourceURL']
2598 segment_list = element.find(_add_ns('SegmentList'))
2599 if segment_list is not None:
2600 extract_common(segment_list)
2601 extract_Initialization(segment_list)
2602 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
2604 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
2606 segment_template = element.find(_add_ns('SegmentTemplate'))
2607 if segment_template is not None:
2608 extract_common(segment_template)
2609 media = segment_template.get('media')
2611 ms_info['media'] = media
2612 initialization = segment_template.get('initialization')
2614 ms_info['initialization'] = initialization
2616 extract_Initialization(segment_template)
2619 skip_unplayable = not self.get_param('allow_unplayable_formats')
2621 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
2624 for period in mpd_doc.findall(_add_ns('Period')):
2625 period_duration = parse_duration(period.get('duration')) or mpd_duration
2626 period_ms_info = extract_multisegment_info(period, {
2630 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
2631 if skip_unplayable and is_drm_protected(adaptation_set):
2633 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
2634 for representation in adaptation_set.findall(_add_ns('Representation')):
2635 if skip_unplayable and is_drm_protected(representation):
2637 representation_attrib = adaptation_set.attrib.copy()
2638 representation_attrib.update(representation.attrib)
2639 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2640 mime_type = representation_attrib['mimeType']
2641 content_type = representation_attrib.get('contentType', mime_type.split('/')[0])
2643 if content_type in ('video', 'audio', 'text') or mime_type == 'image/jpeg':
2645 for element in (representation, adaptation_set, period, mpd_doc):
2646 base_url_e = element.find(_add_ns('BaseURL'))
2647 if base_url_e is not None:
2648 base_url = base_url_e.text + base_url
2649 if re.match(r'^https?://', base_url):
2651 if mpd_base_url and not re.match(r'^https?://', base_url):
2652 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
2654 base_url = mpd_base_url + base_url
2655 representation_id = representation_attrib.get('id')
2656 lang = representation_attrib.get('lang')
2657 url_el = representation.find(_add_ns('BaseURL'))
2658 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
2659 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
2660 if representation_id is not None:
2661 format_id = representation_id
2663 format_id = content_type
2665 format_id = mpd_id + '-' + format_id
2666 if content_type in ('video', 'audio'):
2668 'format_id': format_id,
2669 'manifest_url': mpd_url,
2670 'ext': mimetype2ext(mime_type),
2671 'width': int_or_none(representation_attrib.get('width')),
2672 'height': int_or_none(representation_attrib.get('height')),
2673 'tbr': float_or_none(bandwidth, 1000),
2674 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
2675 'fps': int_or_none(representation_attrib.get('frameRate')),
2676 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
2677 'format_note': 'DASH %s' % content_type,
2678 'filesize': filesize,
2679 'container': mimetype2ext(mime_type) + '_dash',
2681 f.update(parse_codecs(representation_attrib.get('codecs')))
2682 elif content_type == 'text':
2684 'ext': mimetype2ext(mime_type),
2685 'manifest_url': mpd_url,
2686 'filesize': filesize,
2688 elif mime_type == 'image/jpeg':
2689 # See test case in VikiIE
2690 # https://www.viki.com/videos/1175236v-choosing-spouse-by-lottery-episode-1
2692 'format_id': format_id,
2694 'manifest_url': mpd_url,
2695 'format_note': 'DASH storyboards (jpeg)',
2699 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
2701 def prepare_template(template_name, identifiers):
2702 tmpl = representation_ms_info[template_name]
2703 # First of, % characters outside $...$ templates
2704 # must be escaped by doubling for proper processing
2705 # by % operator string formatting used further (see
2706 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2712 in_template = not in_template
2713 elif c == '%' and not in_template:
2715 # Next, $...$ templates are translated to their
2716 # %(...) counterparts to be used with % operator
2717 if representation_id is not None:
2718 t = t.replace('$RepresentationID$', representation_id)
2719 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
2720 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
2721 t.replace('$$', '$')
2724 # @initialization is a regular template like @media one
2725 # so it should be handled just the same way (see
2726 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2727 if 'initialization' in representation_ms_info:
2728 initialization_template = prepare_template(
2730 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2731 # $Time$ shall not be included for @initialization thus
2732 # only $Bandwidth$ remains
2734 representation_ms_info['initialization_url'] = initialization_template % {
2735 'Bandwidth': bandwidth,
2738 def location_key(location):
2739 return 'url' if re.match(r'^https?://', location) else 'path'
2741 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
2743 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2744 media_location_key = location_key(media_template)
2746 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2747 # can't be used at the same time
2748 if '%(Number' in media_template and 's' not in representation_ms_info:
2749 segment_duration = None
2750 if 'total_number' not in representation_ms_info and 'segment_duration' in representation_ms_info:
2751 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
2752 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
2753 representation_ms_info['fragments'] = [{
2754 media_location_key: media_template % {
2755 'Number': segment_number,
2756 'Bandwidth': bandwidth,
2758 'duration': segment_duration,
2759 } for segment_number in range(
2760 representation_ms_info['start_number'],
2761 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
2763 # $Number*$ or $Time$ in media template with S list available
2764 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2765 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2766 representation_ms_info['fragments'] = []
2769 segment_number = representation_ms_info['start_number']
2771 def add_segment_url():
2772 segment_url = media_template % {
2773 'Time': segment_time,
2774 'Bandwidth': bandwidth,
2775 'Number': segment_number,
2777 representation_ms_info['fragments'].append({
2778 media_location_key: segment_url,
2779 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
2782 for num, s in enumerate(representation_ms_info['s']):
2783 segment_time = s.get('t') or segment_time
2787 for r in range(s.get('r', 0)):
2788 segment_time += segment_d
2791 segment_time += segment_d
2792 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
2794 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2795 # or any YouTube dashsegments video
2798 timescale = representation_ms_info['timescale']
2799 for s in representation_ms_info['s']:
2800 duration = float_or_none(s['d'], timescale)
2801 for r in range(s.get('r', 0) + 1):
2802 segment_uri = representation_ms_info['segment_urls'][segment_index]
2804 location_key(segment_uri): segment_uri,
2805 'duration': duration,
2808 representation_ms_info['fragments'] = fragments
2809 elif 'segment_urls' in representation_ms_info:
2810 # Segment URLs with no SegmentTimeline
2811 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2812 # https://github.com/ytdl-org/youtube-dl/pull/14844
2814 segment_duration = float_or_none(
2815 representation_ms_info['segment_duration'],
2816 representation_ms_info['timescale']) if 'segment_duration' in representation_ms_info else None
2817 for segment_url in representation_ms_info['segment_urls']:
2819 location_key(segment_url): segment_url,
2821 if segment_duration:
2822 fragment['duration'] = segment_duration
2823 fragments.append(fragment)
2824 representation_ms_info['fragments'] = fragments
2825 # If there is a fragments key available then we correctly recognized fragmented media.
2826 # Otherwise we will assume unfragmented media with direct access. Technically, such
2827 # assumption is not necessarily correct since we may simply have no support for
2828 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2829 if 'fragments' in representation_ms_info:
2831 # NB: mpd_url may be empty when MPD manifest is parsed from a string
2832 'url': mpd_url or base_url,
2833 'fragment_base_url': base_url,
2835 'protocol': 'http_dash_segments' if mime_type != 'image/jpeg' else 'mhtml',
2837 if 'initialization_url' in representation_ms_info:
2838 initialization_url = representation_ms_info['initialization_url']
2839 if not f.get('url'):
2840 f['url'] = initialization_url
2841 f['fragments'].append({location_key(initialization_url): initialization_url})
2842 f['fragments'].extend(representation_ms_info['fragments'])
2844 # Assuming direct URL to unfragmented media.
2846 if content_type in ('video', 'audio') or mime_type == 'image/jpeg':
2848 elif content_type == 'text':
2849 subtitles.setdefault(lang or 'und', []).append(f)
2851 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
2852 return formats, subtitles
2854 def _extract_ism_formats(self, *args, **kwargs):
2855 fmts, subs = self._extract_ism_formats_and_subtitles(*args, **kwargs)
2857 self.report_warning(bug_reports_message(
2858 "Ignoring subtitle tracks found in the ISM manifest; "
2859 "if any subtitle tracks are missing,"
2863 def _extract_ism_formats_and_subtitles(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
2864 res = self._download_xml_handle(
2866 note='Downloading ISM manifest' if note is None else note,
2867 errnote='Failed to download ISM manifest' if errnote is None else errnote,
2868 fatal=fatal, data=data, headers=headers, query=query)
2875 return self._parse_ism_formats_and_subtitles(ism_doc, urlh.geturl(), ism_id)
2877 def _parse_ism_formats_and_subtitles(self, ism_doc, ism_url, ism_id=None):
2879 Parse formats
from ISM manifest
.
2881 1. [MS
-SSTR
]: Smooth Streaming Protocol
,
2882 https
://msdn
.microsoft
.com
/en
-us
/library
/ff469518
.aspx
2884 if ism_doc.get('IsLive') == 'TRUE':
2886 if (not self.get_param('allow_unplayable_formats')
2887 and ism_doc.find('Protection') is not None):
2890 duration = int(ism_doc.attrib['Duration'])
2891 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
2895 for stream in ism_doc.findall('StreamIndex'):
2896 stream_type = stream.get('Type')
2897 if stream_type not in ('video', 'audio', 'text'):
2899 url_pattern = stream.attrib['Url']
2900 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
2901 stream_name = stream.get('Name')
2902 stream_language = stream.get('Language', 'und')
2903 for track in stream.findall('QualityLevel'):
2904 fourcc = track.get('FourCC') or ('AACL' if track.get('AudioTag') == '255' else None)
2905 # TODO: add support for WVC1 and WMAP
2906 if fourcc not in ('H264', 'AVC1', 'AACL', 'TTML'):
2907 self.report_warning('%s is not a supported codec' % fourcc)
2909 tbr = int(track.attrib['Bitrate']) // 1000
2910 # [1] does not mention Width and Height attributes. However,
2911 # they're often present while MaxWidth and MaxHeight are
2912 # missing, so should be used as fallbacks
2913 width = int_or_none(track.get('MaxWidth') or track.get('Width'))
2914 height = int_or_none(track.get('MaxHeight') or track.get('Height'))
2915 sampling_rate = int_or_none(track.get('SamplingRate'))
2917 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
2918 track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
2924 stream_fragments = stream.findall('c')
2925 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
2926 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
2927 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
2928 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
2929 if not fragment_ctx['duration']:
2931 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
2933 next_fragment_time = duration
2934 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
2935 for _ in range(fragment_repeat):
2937 'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
2938 'duration': fragment_ctx['duration'] / stream_timescale,
2940 fragment_ctx['time'] += fragment_ctx['duration']
2944 format_id.append(ism_id)
2946 format_id.append(stream_name)
2947 format_id.append(compat_str(tbr))
2949 if stream_type == 'text':
2950 subtitles.setdefault(stream_language, []).append({
2954 'manifest_url': ism_url,
2955 'fragments': fragments,
2956 '_download_params': {
2957 'stream_type': stream_type,
2958 'duration': duration,
2959 'timescale': stream_timescale,
2961 'language': stream_language,
2962 'codec_private_data': track.get('CodecPrivateData'),
2965 elif stream_type in ('video', 'audio'):
2967 'format_id': '-'.join(format_id),
2969 'manifest_url': ism_url,
2970 'ext': 'ismv' if stream_type == 'video' else 'isma',
2974 'asr': sampling_rate,
2975 'vcodec': 'none' if stream_type == 'audio' else fourcc,
2976 'acodec': 'none' if stream_type == 'video' else fourcc,
2978 'fragments': fragments,
2979 '_download_params': {
2980 'stream_type': stream_type,
2981 'duration': duration,
2982 'timescale': stream_timescale,
2983 'width': width or 0,
2984 'height': height or 0,
2986 'language': stream_language,
2987 'codec_private_data': track.get('CodecPrivateData'),
2988 'sampling_rate': sampling_rate,
2989 'channels': int_or_none(track.get('Channels', 2)),
2990 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
2991 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
2994 return formats, subtitles
2996 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):
2997 def absolute_url(item_url):
2998 return urljoin(base_url, item_url)
3000 def parse_content_type(content_type):
3001 if not content_type:
3003 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
3005 mimetype, codecs = ctr.groups()
3006 f = parse_codecs(codecs)
3007 f['ext'] = mimetype2ext(mimetype)
3011 def _media_formats(src, cur_media_type, type_info={}):
3012 full_url = absolute_url(src)
3013 ext = type_info.get('ext') or determine_ext(full_url)
3015 is_plain_url = False
3016 formats = self._extract_m3u8_formats(
3017 full_url, video_id, ext='mp4',
3018 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
3019 preference=preference, quality=quality, fatal=False)
3021 is_plain_url = False
3022 formats = self._extract_mpd_formats(
3023 full_url, video_id, mpd_id=mpd_id, fatal=False)
3028 'vcodec': 'none' if cur_media_type == 'audio' else None,
3030 return is_plain_url, formats
3033 # amp-video and amp-audio are very similar to their HTML5 counterparts
3034 # so we wll include them right here (see
3035 # https://www.ampproject.org/docs/reference/components/amp-video)
3036 # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/
3037 _MEDIA_TAG_NAME_RE = r'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)'
3038 media_tags = [(media_tag, media_tag_name, media_type, '')
3039 for media_tag, media_tag_name, media_type
3040 in re.findall(r'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE, webpage)]
3041 media_tags.extend(re.findall(
3042 # We only allow video|audio followed by a whitespace or '>'.
3043 # Allowing more characters may end up in significant slow down (see
3044 # https://github.com/ytdl-org/youtube-dl/issues/11979, example URL:
3045 # http://www.porntrex.com/maps/videositemap.xml).
3046 r'(?s)(<(?P<tag>%s)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>' % _MEDIA_TAG_NAME_RE, webpage))
3047 for media_tag, _, media_type, media_content in media_tags:
3052 media_attributes = extract_attributes(media_tag)
3053 src = strip_or_none(media_attributes.get('src'))
3055 _, formats = _media_formats(src, media_type)
3056 media_info['formats'].extend(formats)
3057 media_info['thumbnail'] = absolute_url(media_attributes.get('poster'))
3059 for source_tag in re.findall(r'<source[^>]+>', media_content):
3060 s_attr = extract_attributes(source_tag)
3061 # data-video-src and data-src are non standard but seen
3062 # several times in the wild
3063 src = strip_or_none(dict_get(s_attr, ('src', 'data-video-src', 'data-src')))
3066 f = parse_content_type(s_attr.get('type'))
3067 is_plain_url, formats = _media_formats(src, media_type, f)
3069 # width, height, res, label and title attributes are
3070 # all not standard but seen several times in the wild
3073 for lbl in ('label', 'title')
3074 if str_or_none(s_attr.get(lbl))
3076 width = int_or_none(s_attr.get('width'))
3077 height = (int_or_none(s_attr.get('height'))
3078 or int_or_none(s_attr.get('res')))
3079 if not width or not height:
3081 resolution = parse_resolution(lbl)
3084 width = width or resolution.get('width')
3085 height = height or resolution.get('height')
3087 tbr = parse_bitrate(lbl)
3096 'format_id': s_attr.get('label') or s_attr.get('title'),
3098 f.update(formats[0])
3099 media_info['formats'].append(f)
3101 media_info['formats'].extend(formats)
3102 for track_tag in re.findall(r'<track[^>]+>', media_content):
3103 track_attributes = extract_attributes(track_tag)
3104 kind = track_attributes.get('kind')
3105 if not kind or kind in ('subtitles', 'captions'):
3106 src = strip_or_none(track_attributes.get('src'))
3109 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
3110 media_info['subtitles'].setdefault(lang, []).append({
3111 'url': absolute_url(src),
3113 for f in media_info['formats']:
3114 f.setdefault('http_headers', {})['Referer'] = base_url
3115 if media_info['formats'] or media_info['subtitles']:
3116 entries.append(media_info)
3119 def _extract_akamai_formats(self, *args, **kwargs):
3120 fmts, subs = self._extract_akamai_formats_and_subtitles(*args, **kwargs)
3122 self.report_warning(bug_reports_message(
3123 "Ignoring subtitle tracks found in the manifests; "
3124 "if any subtitle tracks are missing,"
3128 def _extract_akamai_formats_and_subtitles(self, manifest_url, video_id, hosts={}):
3129 signed = 'hdnea=' in manifest_url
3131 # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
3132 manifest_url = re.sub(
3133 r'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
3134 '', manifest_url).strip('?')
3139 hdcore_sign = 'hdcore=3.7.0'
3140 f4m_url = re.sub(r'(https?://[^/]+)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
3141 hds_host = hosts.get('hds')
3143 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
3144 if 'hdcore=' not in f4m_url:
3145 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
3146 f4m_formats = self._extract_f4m_formats(
3147 f4m_url, video_id, f4m_id='hds', fatal=False)
3148 for entry in f4m_formats:
3149 entry.update({'extra_param_to_segment_url': hdcore_sign})
3150 formats.extend(f4m_formats)
3152 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
3153 hls_host = hosts.get('hls')
3155 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
3156 m3u8_formats, m3u8_subtitles = self._extract_m3u8_formats_and_subtitles(
3157 m3u8_url, video_id, 'mp4', 'm3u8_native',
3158 m3u8_id='hls', fatal=False)
3159 formats.extend(m3u8_formats)
3160 subtitles = self._merge_subtitles(subtitles, m3u8_subtitles)
3162 http_host = hosts.get('http')
3163 if http_host and m3u8_formats and not signed:
3164 REPL_REGEX = r'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
3165 qualities = re.match(REPL_REGEX, m3u8_url).group(2).split(',')
3166 qualities_length = len(qualities)
3167 if len(m3u8_formats) in (qualities_length, qualities_length + 1):
3169 for f in m3u8_formats:
3170 if f['vcodec'] != 'none':
3171 for protocol in ('http', 'https'):
3173 del http_f['manifest_url']
3175 REPL_REGEX, protocol + r'://%s/\g<1>%s\3' % (http_host, qualities[i]), f['url'])
3177 'format_id': http_f['format_id'].replace('hls-', protocol + '-'),
3179 'protocol': protocol,
3181 formats.append(http_f)
3184 return formats, subtitles
3186 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
3187 query = compat_urlparse.urlparse(url).query
3188 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
3190 r'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url)
3191 url_base = mobj.group('url')
3192 http_base_url = '%s%s:%s' % ('http', mobj.group('s') or '', url_base)
3195 def manifest_url(manifest):
3196 m_url = '%s/%s' % (http_base_url, manifest)
3198 m_url += '?%s' % query
3201 if 'm3u8' not in skip_protocols:
3202 formats.extend(self._extract_m3u8_formats(
3203 manifest_url('playlist.m3u8'), video_id, 'mp4',
3204 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
3205 if 'f4m' not in skip_protocols:
3206 formats.extend(self._extract_f4m_formats(
3207 manifest_url('manifest.f4m'),
3208 video_id, f4m_id='hds', fatal=False))
3209 if 'dash' not in skip_protocols:
3210 formats.extend(self._extract_mpd_formats(
3211 manifest_url('manifest.mpd'),
3212 video_id, mpd_id='dash', fatal=False))
3213 if re.search(r'(?:/smil:|\.smil)', url_base):
3214 if 'smil' not in skip_protocols:
3215 rtmp_formats = self._extract_smil_formats(
3216 manifest_url('jwplayer.smil'),
3217 video_id, fatal=False)
3218 for rtmp_format in rtmp_formats:
3219 rtsp_format = rtmp_format.copy()
3220 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
3221 del rtsp_format['play_path']
3222 del rtsp_format['ext']
3223 rtsp_format.update({
3224 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
3225 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
3228 formats.extend([rtmp_format, rtsp_format])
3230 for protocol in ('rtmp', 'rtsp'):
3231 if protocol not in skip_protocols:
3233 'url': '%s:%s' % (protocol, url_base),
3234 'format_id': protocol,
3235 'protocol': protocol,
3239 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
3241 r'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)(?!</script>).*?\.setup\s*\((?P<options>[^)]+)\)',
3245 jwplayer_data = self._parse_json(mobj.group('options'),
3247 transform_source=transform_source)
3248 except ExtractorError:
3251 if isinstance(jwplayer_data, dict):
3252 return jwplayer_data
3254 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
3255 jwplayer_data = self._find_jwplayer_data(
3256 webpage, video_id, transform_source=js_to_json)
3257 return self._parse_jwplayer_data(
3258 jwplayer_data, video_id, *args, **kwargs)
3260 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
3261 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3262 # JWPlayer backward compatibility: flattened playlists
3263 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
3264 if 'playlist' not in jwplayer_data:
3265 jwplayer_data = {'playlist': [jwplayer_data]}
3269 # JWPlayer backward compatibility: single playlist item
3270 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
3271 if not isinstance(jwplayer_data['playlist'], list):
3272 jwplayer_data['playlist'] = [jwplayer_data['playlist']]
3274 for video_data in jwplayer_data['playlist']:
3275 # JWPlayer backward compatibility: flattened sources
3276 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
3277 if 'sources' not in video_data:
3278 video_data['sources'] = [video_data]
3280 this_video_id = video_id or video_data['mediaid']
3282 formats = self._parse_jwplayer_formats(
3283 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
3284 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
3287 tracks = video_data.get('tracks')
3288 if tracks and isinstance(tracks, list):
3289 for track in tracks:
3290 if not isinstance(track, dict):
3292 track_kind = track.get('kind')
3293 if not track_kind or not isinstance(track_kind, compat_str):
3295 if track_kind.lower() not in ('captions', 'subtitles'):
3297 track_url = urljoin(base_url, track.get('file'))
3300 subtitles.setdefault(track.get('label') or 'en', []).append({
3301 'url': self._proto_relative_url(track_url)
3305 'id': this_video_id,
3306 'title': unescapeHTML(video_data['title'] if require_title else video_data.get('title')),
3307 'description': clean_html(video_data.get('description')),
3308 'thumbnail': urljoin(base_url, self._proto_relative_url(video_data.get('image'))),
3309 'timestamp': int_or_none(video_data.get('pubdate')),
3310 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
3311 'subtitles': subtitles,
3313 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
3314 if len(formats) == 1 and re.search(r'^(?:http|//).*(?:youtube\.com|youtu\.be)/.+', formats[0]['url']):
3316 '_type': 'url_transparent',
3317 'url': formats[0]['url'],
3320 self._sort_formats(formats)
3321 entry['formats'] = formats
3322 entries.append(entry)
3323 if len(entries) == 1:
3326 return self.playlist_result(entries)
3328 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
3329 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3332 for source in jwplayer_sources_data:
3333 if not isinstance(source, dict):
3335 source_url = urljoin(
3336 base_url, self._proto_relative_url(source.get('file')))
3337 if not source_url or source_url in urls:
3339 urls.append(source_url)
3340 source_type = source.get('type') or ''
3341 ext = mimetype2ext(source_type) or determine_ext(source_url)
3342 if source_type == 'hls' or ext == 'm3u8':
3343 formats.extend(self._extract_m3u8_formats(
3344 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
3345 m3u8_id=m3u8_id, fatal=False))
3346 elif source_type == 'dash' or ext == 'mpd':
3347 formats.extend(self._extract_mpd_formats(
3348 source_url, video_id, mpd_id=mpd_id, fatal=False))
3350 formats.extend(self._extract_smil_formats(
3351 source_url, video_id, fatal=False))
3352 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3353 elif source_type.startswith('audio') or ext in (
3354 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
3361 height = int_or_none(source.get('height'))
3363 # Often no height is provided but there is a label in
3364 # format like "1080p", "720p SD", or 1080.
3365 height = int_or_none(self._search_regex(
3366 r'^(\d{3,4})[pP]?(?:\b|$)', compat_str(source.get('label') or ''),
3367 'height', default=None))
3370 'width': int_or_none(source.get('width')),
3372 'tbr': int_or_none(source.get('bitrate')),
3375 if source_url.startswith('rtmp'):
3376 a_format['ext'] = 'flv'
3377 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3378 # of jwplayer.flash.swf
3379 rtmp_url_parts = re.split(
3380 r'((?:mp4|mp3|flv):)', source_url, 1)
3381 if len(rtmp_url_parts) == 3:
3382 rtmp_url, prefix, play_path = rtmp_url_parts
3385 'play_path': prefix + play_path,
3388 a_format.update(rtmp_params)
3389 formats.append(a_format)
3392 def _live_title(self, name):
3393 """ Generate the title
for a live video
"""
3394 now = datetime.datetime.now()
3395 now_str = now.strftime('%Y-%m-%d %H:%M')
3396 return name + ' ' + now_str
3398 def _int(self, v, name, fatal=False, **kwargs):
3399 res = int_or_none(v, **kwargs)
3400 if 'get_attr' in kwargs:
3401 print(getattr(v, kwargs['get_attr']))
3403 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
3405 raise ExtractorError(msg)
3407 self.report_warning(msg)
3410 def _float(self, v, name, fatal=False, **kwargs):
3411 res = float_or_none(v, **kwargs)
3413 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
3415 raise ExtractorError(msg)
3417 self.report_warning(msg)
3420 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
3421 path='/', secure=False, discard=False, rest={}, **kwargs):
3422 cookie = compat_cookiejar_Cookie(
3423 0, name, value, port, port is not None, domain, True,
3424 domain.startswith('.'), path, True, secure, expire_time,
3425 discard, None, None, rest)
3426 self._downloader.cookiejar.set_cookie(cookie)
3428 def _get_cookies(self, url):
3429 """ Return a compat_cookies_SimpleCookie
with the cookies
for the url
"""
3430 req = sanitized_Request(url)
3431 self._downloader.cookiejar.add_cookie_header(req)
3432 return compat_cookies_SimpleCookie(req.get_header('Cookie'))
3434 def _apply_first_set_cookie_header(self, url_handle, cookie):
3436 Apply first Set
-Cookie header instead of the last
. Experimental
.
3438 Some
sites (e
.g
. [1-3]) may serve two cookies under the same name
3439 in Set
-Cookie header
and expect the
first (old
) one to be
set rather
3440 than
second (new
). However
, as of RFC6265 the newer one cookie
3441 should be
set into cookie store what actually happens
.
3442 We will workaround this issue by resetting the cookie to
3443 the first one manually
.
3444 1. https
://new
.vk
.com
/
3445 2. https
://github
.com
/ytdl
-org
/youtube
-dl
/issues
/9841#issuecomment-227871201
3446 3. https
://learning
.oreilly
.com
/
3448 for header, cookies in url_handle.headers.items():
3449 if header.lower() != 'set-cookie':
3451 if sys.version_info[0] >= 3:
3452 cookies = cookies.encode('iso-8859-1')
3453 cookies = cookies.decode('utf-8')
3454 cookie_value = re.search(
3455 r'%s=(.+?);.*?\b[Dd]omain=(.+?)(?:[,;]|$)' % cookie, cookies)
3457 value, domain = cookie_value.groups()
3458 self._set_cookie(domain, cookie, value)
3461 def get_testcases(self, include_onlymatching=False):
3462 t = getattr(self, '_TEST', None)
3464 assert not hasattr(self, '_TESTS'), \
3465 '%s has _TEST and _TESTS' % type(self).__name__
3468 tests = getattr(self, '_TESTS', [])
3470 if not include_onlymatching and t.get('only_matching', False):
3472 t['name'] = type(self).__name__[:-len('IE')]
3475 def is_suitable(self, age_limit):
3476 """ Test whether the extractor
is generally suitable
for the given
3477 age
limit (i
.e
. pornographic sites are
not, all others usually are
) """
3479 any_restricted = False
3480 for tc in self.get_testcases(include_onlymatching=False):
3481 if tc.get('playlist', []):
3482 tc = tc['playlist'][0]
3483 is_restricted = age_restricted(
3484 tc.get('info_dict', {}).get('age_limit'), age_limit)
3485 if not is_restricted:
3487 any_restricted = any_restricted or is_restricted
3488 return not any_restricted
3490 def extract_subtitles(self, *args, **kwargs):
3491 if (self.get_param('writesubtitles', False)
3492 or self.get_param('listsubtitles')):
3493 return self._get_subtitles(*args, **kwargs)
3496 def _get_subtitles(self, *args, **kwargs):
3497 raise NotImplementedError('This method must be implemented by subclasses')
3500 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
3501 """ Merge subtitle items
for one language
. Items
with duplicated URLs
3502 will be dropped
. """
3503 list1_urls = set([item['url'] for item in subtitle_list1])
3504 ret = list(subtitle_list1)
3505 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
3509 def _merge_subtitles(cls, *dicts, **kwargs):
3510 """ Merge subtitle dictionaries
, language by language
. """
3512 target = (lambda target=None: target)(**kwargs)
3513 # The above lambda extracts the keyword argument 'target' from kwargs
3514 # while ensuring there are no stray ones. When Python 2 support
3515 # is dropped, remove it and change the function signature to:
3517 # def _merge_subtitles(cls, *dicts, target=None):
3522 for lang, subs in d.items():
3523 target[lang] = cls._merge_subtitle_items(target.get(lang, []), subs)
3526 def extract_automatic_captions(self, *args, **kwargs):
3527 if (self.get_param('writeautomaticsub', False)
3528 or self.get_param('listsubtitles')):
3529 return self._get_automatic_captions(*args, **kwargs)
3532 def _get_automatic_captions(self, *args, **kwargs):
3533 raise NotImplementedError('This method must be implemented by subclasses')
3535 def mark_watched(self, *args, **kwargs):
3536 if (self.get_param('mark_watched', False)
3537 and (self._get_login_info()[0] is not None
3538 or self.get_param('cookiefile') is not None)):
3539 self._mark_watched(*args, **kwargs)
3541 def _mark_watched(self, *args, **kwargs):
3542 raise NotImplementedError('This method must be implemented by subclasses')
3544 def geo_verification_headers(self):
3546 geo_verification_proxy = self.get_param('geo_verification_proxy')
3547 if geo_verification_proxy:
3548 headers['Ytdl-request-proxy'] = geo_verification_proxy
3551 def _generic_id(self, url):
3552 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
3554 def _generic_title(self, url):
3555 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
3558 def _availability(is_private=None, needs_premium=None, needs_subscription=None, needs_auth=None, is_unlisted=None):
3559 all_known = all(map(
3560 lambda x: x is not None,
3561 (is_private, needs_premium, needs_subscription, needs_auth, is_unlisted)))
3563 'private' if is_private
3564 else 'premium_only' if needs_premium
3565 else 'subscriber_only' if needs_subscription
3566 else 'needs_auth' if needs_auth
3567 else 'unlisted' if is_unlisted
3568 else 'public' if all_known
3571 def _configuration_arg(self, key):
3572 return traverse_obj(
3573 self._downloader.params, ('extractor_args', self.ie_key().lower(), key))
3576 class SearchInfoExtractor(InfoExtractor):
3578 Base
class for paged search queries extractors
.
3579 They accept URLs
in the format
_SEARCH_KEY(|all|
[0-9]):{query}
3580 Instances should define _SEARCH_KEY
and _MAX_RESULTS
.
3584 def _make_valid_url(cls):
3585 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
3588 def suitable(cls, url):
3589 return re.match(cls._make_valid_url(), url) is not None
3591 def _real_extract(self, query):
3592 mobj = re.match(self._make_valid_url(), query)
3594 raise ExtractorError('Invalid search query "%s"' % query)
3596 prefix = mobj.group('prefix')
3597 query = mobj.group('query')
3599 return self._get_n_results(query, 1)
3600 elif prefix == 'all':
3601 return self._get_n_results(query, self._MAX_RESULTS)
3605 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
3606 elif n > self._MAX_RESULTS:
3607 self.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3608 n = self._MAX_RESULTS
3609 return self._get_n_results(query, n)
3611 def _get_n_results(self, query, n):
3612 """Get a specified number of results
for a query
"""
3613 raise NotImplementedError('This method must be implemented by subclasses')
3616 def SEARCH_KEY(self):
3617 return self._SEARCH_KEY