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
,
87 class InfoExtractor(object):
88 """Information Extractor class.
90 Information extractors are the classes that, given a URL, extract
91 information about the video (or videos) the URL refers to. This
92 information includes the real video URL, the video title, author and
93 others. The information is stored in a dictionary which is then
94 passed to the YoutubeDL. The YoutubeDL processes this
95 information possibly downloading the video to the file system, among
96 other possible outcomes.
98 The type field determines the type of the result.
99 By far the most common value (and the default if _type is missing) is
100 "video", which indicates a single video.
102 For a video, the dictionaries must include the following fields:
104 id: Video identifier.
105 title: Video title, unescaped.
107 Additionally, it must contain either a formats entry or a url one:
109 formats: A list of dictionaries for each format available, ordered
110 from worst to best quality.
113 * url The mandatory URL representing the media:
114 for plain file media - HTTP URL of this file,
116 for HLS - URL of the M3U8 media playlist,
117 for HDS - URL of the F4M manifest,
119 - HTTP URL to plain file media (in case of
121 - URL of the MPD manifest or base URL
122 representing the media if MPD manifest
123 is parsed from a string (in case of
125 for MSS - URL of the ISM manifest.
127 The URL of the manifest file in case of
129 for HLS - URL of the M3U8 master playlist,
130 for HDS - URL of the F4M manifest,
131 for DASH - URL of the MPD manifest,
132 for MSS - URL of the ISM manifest.
133 * ext Will be calculated from URL if missing
134 * format A human-readable description of the format
135 ("mp4 container with h264/opus").
136 Calculated from the format_id, width, height.
137 and format_note fields if missing.
138 * format_id A short description of the format
139 ("mp4_h264_opus" or "19").
140 Technically optional, but strongly recommended.
141 * format_note Additional info about the format
142 ("3D" or "DASH video")
143 * width Width of the video, if known
144 * height Height of the video, if known
145 * resolution Textual description of width and height
146 * tbr Average bitrate of audio and video in KBit/s
147 * abr Average audio bitrate in KBit/s
148 * acodec Name of the audio codec in use
149 * asr Audio sampling rate in Hertz
150 * vbr Average video bitrate in KBit/s
152 * vcodec Name of the video codec in use
153 * container Name of the container format
154 * filesize The number of bytes, if known in advance
155 * filesize_approx An estimate for the number of bytes
156 * player_url SWF Player URL (used for rtmpdump).
157 * protocol The protocol that will be used for the actual
158 download, lower-case.
159 "http", "https", "rtsp", "rtmp", "rtmp_ffmpeg", "rtmpe",
160 "m3u8", "m3u8_native" or "http_dash_segments".
162 Base URL for fragments. Each fragment's path
163 value (if present) will be relative to
165 * fragments A list of fragments of a fragmented media.
166 Each fragment entry must contain either an url
167 or a path. If an url is present it should be
168 considered by a client. Otherwise both path and
169 fragment_base_url must be present. Here is
170 the list of all potential fields:
171 * "url" - fragment's URL
172 * "path" - fragment's path relative to
174 * "duration" (optional, int or float)
175 * "filesize" (optional, int)
176 * preference Order number of this format. If this field is
177 present and not None, the formats get sorted
178 by this field, regardless of all other values.
179 -1 for default (order by other properties),
180 -2 or smaller for less than default.
181 < -1000 to hide the format (if there is
182 another one which is strictly better)
183 * language Language code, e.g. "de" or "en-US".
184 * language_preference Is this in the language mentioned in
186 10 if it's what the URL is about,
187 -1 for default (don't know),
188 -10 otherwise, other values reserved for now.
189 * quality Order number of the video quality of this
190 format, irrespective of the file format.
191 -1 for default (order by other properties),
192 -2 or smaller for less than default.
193 * source_preference Order number for this video source
194 (quality takes higher priority)
195 -1 for default (order by other properties),
196 -2 or smaller for less than default.
197 * http_headers A dictionary of additional HTTP headers
198 to add to the request.
199 * stretched_ratio If given and not 1, indicates that the
200 video's pixels are not square.
201 width : height ratio as float.
202 * no_resume The server does not support resuming the
203 (HTTP or RTMP) download. Boolean.
204 * downloader_options A dictionary of downloader options as
205 described in FileDownloader
207 url: Final video URL.
208 ext: Video filename extension.
209 format: The video format, defaults to ext (used for --get-format)
210 player_url: SWF Player URL (used for rtmpdump).
212 The following fields are optional:
214 alt_title: A secondary title of the video.
215 display_id An alternative identifier for the video, not necessarily
216 unique, but available before title. Typically, id is
217 something like "4234987", title "Dancing naked mole rats",
218 and display_id "dancing-naked-mole-rats"
219 thumbnails: A list of dictionaries, with the following entries:
220 * "id" (optional, string) - Thumbnail format ID
222 * "preference" (optional, int) - quality of the image
223 * "width" (optional, int)
224 * "height" (optional, int)
225 * "resolution" (optional, string "{width}x{height}",
227 * "filesize" (optional, int)
228 thumbnail: Full URL to a video thumbnail image.
229 description: Full video description.
230 uploader: Full name of the video uploader.
231 license: License name the video is licensed under.
232 creator: The creator of the video.
233 release_timestamp: UNIX timestamp of the moment the video was released.
234 release_date: The date (YYYYMMDD) when the video was released.
235 timestamp: UNIX timestamp of the moment the video was uploaded
236 upload_date: Video upload date (YYYYMMDD).
237 If not explicitly set, calculated from timestamp.
238 uploader_id: Nickname or id of the video uploader.
239 uploader_url: Full URL to a personal webpage of the video uploader.
240 channel: Full name of the channel the video is uploaded on.
241 Note that channel fields may or may not repeat uploader
242 fields. This depends on a particular extractor.
243 channel_id: Id of the channel.
244 channel_url: Full URL to a channel webpage.
245 location: Physical location where the video was filmed.
246 subtitles: The available subtitles as a dictionary in the format
247 {tag: subformats}. "tag" is usually a language code, and
248 "subformats" is a list sorted from lower to higher
249 preference, each element is a dictionary with the "ext"
251 * "data": The subtitles file contents
252 * "url": A URL pointing to the subtitles file
253 "ext" will be calculated from URL if missing
254 automatic_captions: Like 'subtitles'; contains automatically generated
255 captions instead of normal subtitles
256 duration: Length of the video in seconds, as an integer or float.
257 view_count: How many users have watched the video on the platform.
258 like_count: Number of positive ratings of the video
259 dislike_count: Number of negative ratings of the video
260 repost_count: Number of reposts of the video
261 average_rating: Average rating give by users, the scale used depends on the webpage
262 comment_count: Number of comments on the video
263 comments: A list of comments, each with one or more of the following
264 properties (all but one of text or html optional):
265 * "author" - human-readable name of the comment author
266 * "author_id" - user ID of the comment author
267 * "author_thumbnail" - The thumbnail of the comment author
269 * "html" - Comment as HTML
270 * "text" - Plain text of the comment
271 * "timestamp" - UNIX timestamp of comment
272 * "parent" - ID of the comment this one is replying to.
273 Set to "root" to indicate that this is a
274 comment to the original video.
275 * "like_count" - Number of positive ratings of the comment
276 * "dislike_count" - Number of negative ratings of the comment
277 * "is_favorited" - Whether the comment is marked as
278 favorite by the video uploader
279 * "author_is_uploader" - Whether the comment is made by
281 age_limit: Age restriction for the video, as an integer (years)
282 webpage_url: The URL to the video webpage, if given to yt-dlp it
283 should allow to get the same result again. (It will be set
284 by YoutubeDL if it's missing)
285 categories: A list of categories that the video falls in, for example
287 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
288 is_live: True, False, or None (=unknown). Whether this video is a
289 live stream that goes on instead of a fixed-length video.
290 was_live: True, False, or None (=unknown). Whether this video was
291 originally a live stream.
292 start_time: Time in seconds where the reproduction should start, as
293 specified in the URL.
294 end_time: Time in seconds where the reproduction should end, as
295 specified in the URL.
296 chapters: A list of dictionaries, with the following entries:
297 * "start_time" - The start time of the chapter in seconds
298 * "end_time" - The end time of the chapter in seconds
299 * "title" (optional, string)
300 playable_in_embed: Whether this video is allowed to play in embedded
301 players on other sites. Can be True (=always allowed),
302 False (=never allowed), None (=unknown), or a string
303 specifying the criteria for embedability (Eg: 'whitelist')
304 availability: Under what condition the video is available. One of
305 'private', 'premium_only', 'subscriber_only', 'needs_auth',
306 'unlisted' or 'public'. Use 'InfoExtractor._availability'
308 __post_extractor: A function to be called just before the metadata is
309 written to either disk, logger or console. The function
310 must return a dict which will be added to the info_dict.
311 This is usefull for additional information that is
312 time-consuming to extract. Note that the fields thus
313 extracted will not be available to output template and
314 match_filter. So, only "comments" and "comment_count" are
315 currently allowed to be extracted via this method.
317 The following fields should only be used when the video belongs to some logical
320 chapter: Name or title of the chapter the video belongs to.
321 chapter_number: Number of the chapter the video belongs to, as an integer.
322 chapter_id: Id of the chapter the video belongs to, as a unicode string.
324 The following fields should only be used when the video is an episode of some
325 series, programme or podcast:
327 series: Title of the series or programme the video episode belongs to.
328 season: Title of the season the video episode belongs to.
329 season_number: Number of the season the video episode belongs to, as an integer.
330 season_id: Id of the season the video episode belongs to, as a unicode string.
331 episode: Title of the video episode. Unlike mandatory video title field,
332 this field should denote the exact title of the video episode
333 without any kind of decoration.
334 episode_number: Number of the video episode within a season, as an integer.
335 episode_id: Id of the video episode, as a unicode string.
337 The following fields should only be used when the media is a track or a part of
340 track: Title of the track.
341 track_number: Number of the track within an album or a disc, as an integer.
342 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
344 artist: Artist(s) of the track.
345 genre: Genre(s) of the track.
346 album: Title of the album the track belongs to.
347 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
348 album_artist: List of all artists appeared on the album (e.g.
349 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
351 disc_number: Number of the disc or other physical medium the track belongs to,
353 release_year: Year (YYYY) when the album was released.
355 Unless mentioned otherwise, the fields should be Unicode strings.
357 Unless mentioned otherwise, None is equivalent to absence of information.
360 _type "playlist" indicates multiple videos.
361 There must be a key "entries", which is a list, an iterable, or a PagedList
362 object, each element of which is a valid dictionary by this specification.
364 Additionally, playlists can have "id", "title", and any other relevent
365 attributes with the same semantics as videos (see above).
368 _type "multi_video" indicates that there are multiple videos that
369 form a single show, for examples multiple acts of an opera or TV episode.
370 It must have an entries key like a playlist and contain all the keys
371 required for a video at the same time.
374 _type "url" indicates that the video must be extracted from another
375 location, possibly by a different extractor. Its only required key is:
376 "url" - the next URL to extract.
377 The key "ie_key" can be set to the class name (minus the trailing "IE",
378 e.g. "Youtube") if the extractor class is known in advance.
379 Additionally, the dictionary may have any properties of the resolved entity
380 known in advance, for example "title" if the title of the referred video is
384 _type "url_transparent" entities have the same specification as "url", but
385 indicate that the given additional information is more precise than the one
386 associated with the resolved URL.
387 This is useful when a site employs a video service that hosts the video and
388 its technical metadata, but that video service does not embed a useful
389 title, description etc.
392 Subclasses of this one should re-define the _real_initialize() and
393 _real_extract() methods and define a _VALID_URL regexp.
394 Probably, they should also be added to the list of extractors.
396 _GEO_BYPASS attribute may be set to False in order to disable
397 geo restriction bypass mechanisms for a particular extractor.
398 Though it won't disable explicit geo restriction bypass based on
399 country code provided with geo_bypass_country.
401 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
402 countries for this extractor. One of these countries will be used by
403 geo restriction bypass mechanism right away in order to bypass
404 geo restriction, of course, if the mechanism is not disabled.
406 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
407 IP blocks in CIDR notation for this extractor. One of these IP blocks
408 will be used by geo restriction bypass mechanism similarly
411 Finally, the _WORKING attribute should be set to False for broken IEs
412 in order to warn the users and skip the tests.
417 _x_forwarded_for_ip
= None
419 _GEO_COUNTRIES
= None
420 _GEO_IP_BLOCKS
= None
423 def __init__(self
, downloader
=None):
424 """Constructor. Receives an optional downloader."""
426 self
._x
_forwarded
_for
_ip
= None
427 self
.set_downloader(downloader
)
430 def suitable(cls
, url
):
431 """Receives a URL and returns True if suitable for this IE."""
433 # This does not use has/getattr intentionally - we want to know whether
434 # we have cached the regexp for *this* class, whereas getattr would also
435 # match the superclass
436 if '_VALID_URL_RE' not in cls
.__dict
__:
437 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
438 return cls
._VALID
_URL
_RE
.match(url
) is not None
441 def _match_id(cls
, url
):
442 if '_VALID_URL_RE' not in cls
.__dict
__:
443 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
444 m
= cls
._VALID
_URL
_RE
.match(url
)
446 return compat_str(m
.group('id'))
450 """Getter method for _WORKING."""
453 def initialize(self
):
454 """Initializes an instance (authentication, etc)."""
455 self
._initialize
_geo
_bypass
({
456 'countries': self
._GEO
_COUNTRIES
,
457 'ip_blocks': self
._GEO
_IP
_BLOCKS
,
460 self
._real
_initialize
()
463 def _initialize_geo_bypass(self
, geo_bypass_context
):
465 Initialize geo restriction bypass mechanism.
467 This method is used to initialize geo bypass mechanism based on faking
468 X-Forwarded-For HTTP header. A random country from provided country list
469 is selected and a random IP belonging to this country is generated. This
470 IP will be passed as X-Forwarded-For HTTP header in all subsequent
473 This method will be used for initial geo bypass mechanism initialization
474 during the instance initialization with _GEO_COUNTRIES and
477 You may also manually call it from extractor's code if geo bypass
478 information is not available beforehand (e.g. obtained during
479 extraction) or due to some other reason. In this case you should pass
480 this information in geo bypass context passed as first argument. It may
481 contain following fields:
483 countries: List of geo unrestricted countries (similar
485 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
486 (similar to _GEO_IP_BLOCKS)
489 if not self
._x
_forwarded
_for
_ip
:
491 # Geo bypass mechanism is explicitly disabled by user
492 if not self
._downloader
.params
.get('geo_bypass', True):
495 if not geo_bypass_context
:
496 geo_bypass_context
= {}
498 # Backward compatibility: previously _initialize_geo_bypass
499 # expected a list of countries, some 3rd party code may still use
501 if isinstance(geo_bypass_context
, (list, tuple)):
502 geo_bypass_context
= {
503 'countries': geo_bypass_context
,
506 # The whole point of geo bypass mechanism is to fake IP
507 # as X-Forwarded-For HTTP header based on some IP block or
510 # Path 1: bypassing based on IP block in CIDR notation
512 # Explicit IP block specified by user, use it right away
513 # regardless of whether extractor is geo bypassable or not
514 ip_block
= self
._downloader
.params
.get('geo_bypass_ip_block', None)
516 # Otherwise use random IP block from geo bypass context but only
517 # if extractor is known as geo bypassable
519 ip_blocks
= geo_bypass_context
.get('ip_blocks')
520 if self
._GEO
_BYPASS
and ip_blocks
:
521 ip_block
= random
.choice(ip_blocks
)
524 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(ip_block
)
525 self
._downloader
.write_debug(
526 '[debug] Using fake IP %s as X-Forwarded-For' % self
._x
_forwarded
_for
_ip
)
529 # Path 2: bypassing based on country code
531 # Explicit country code specified by user, use it right away
532 # regardless of whether extractor is geo bypassable or not
533 country
= self
._downloader
.params
.get('geo_bypass_country', None)
535 # Otherwise use random country code from geo bypass context but
536 # only if extractor is known as geo bypassable
538 countries
= geo_bypass_context
.get('countries')
539 if self
._GEO
_BYPASS
and countries
:
540 country
= random
.choice(countries
)
543 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country
)
544 self
._downloader
.write_debug(
545 'Using fake IP %s (%s) as X-Forwarded-For' % (self
._x
_forwarded
_for
_ip
, country
.upper()))
547 def extract(self
, url
):
548 """Extracts URL information and returns it in list of dicts."""
553 ie_result
= self
._real
_extract
(url
)
554 if self
._x
_forwarded
_for
_ip
:
555 ie_result
['__x_forwarded_for_ip'] = self
._x
_forwarded
_for
_ip
556 subtitles
= ie_result
.get('subtitles')
557 if (subtitles
and 'live_chat' in subtitles
558 and 'no-live-chat' in self
._downloader
.params
.get('compat_opts', [])):
559 del subtitles
['live_chat']
561 except GeoRestrictedError
as e
:
562 if self
.__maybe
_fake
_ip
_and
_retry
(e
.countries
):
565 except ExtractorError
:
567 except compat_http_client
.IncompleteRead
as e
:
568 raise ExtractorError('A network error has occurred.', cause
=e
, expected
=True)
569 except (KeyError, StopIteration) as e
:
570 raise ExtractorError('An extractor error has occurred.', cause
=e
)
572 def __maybe_fake_ip_and_retry(self
, countries
):
573 if (not self
._downloader
.params
.get('geo_bypass_country', None)
575 and self
._downloader
.params
.get('geo_bypass', True)
576 and not self
._x
_forwarded
_for
_ip
578 country_code
= random
.choice(countries
)
579 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country_code
)
580 if self
._x
_forwarded
_for
_ip
:
582 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
583 % (self
._x
_forwarded
_for
_ip
, country_code
.upper()))
587 def set_downloader(self
, downloader
):
588 """Sets the downloader for this IE."""
589 self
._downloader
= downloader
591 def _real_initialize(self
):
592 """Real initialization process. Redefine in subclasses."""
595 def _real_extract(self
, url
):
596 """Real extraction process. Redefine in subclasses."""
601 """A string for getting the InfoExtractor with get_info_extractor"""
602 return compat_str(cls
.__name
__[:-2])
606 return compat_str(type(self
).__name
__[:-2])
609 def __can_accept_status_code(err
, expected_status
):
610 assert isinstance(err
, compat_urllib_error
.HTTPError
)
611 if expected_status
is None:
613 if isinstance(expected_status
, compat_integer_types
):
614 return err
.code
== expected_status
615 elif isinstance(expected_status
, (list, tuple)):
616 return err
.code
in expected_status
617 elif callable(expected_status
):
618 return expected_status(err
.code
) is True
622 def _request_webpage(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
, expected_status
=None):
624 Return the response handle.
626 See _download_webpage docstring for arguments specification.
628 if not self
._downloader
._first
_webpage
_request
:
629 sleep_interval
= float_or_none(self
._downloader
.params
.get('sleep_interval_requests')) or 0
630 if sleep_interval
> 0:
631 self
.to_screen('Sleeping %s seconds ...' % sleep_interval
)
632 time
.sleep(sleep_interval
)
634 self
._downloader
._first
_webpage
_request
= False
637 self
.report_download_webpage(video_id
)
638 elif note
is not False:
640 self
.to_screen('%s' % (note
,))
642 self
.to_screen('%s: %s' % (video_id
, note
))
644 # Some sites check X-Forwarded-For HTTP header in order to figure out
645 # the origin of the client behind proxy. This allows bypassing geo
646 # restriction by faking this header's value to IP that belongs to some
647 # geo unrestricted country. We will do so once we encounter any
648 # geo restriction error.
649 if self
._x
_forwarded
_for
_ip
:
650 if 'X-Forwarded-For' not in headers
:
651 headers
['X-Forwarded-For'] = self
._x
_forwarded
_for
_ip
653 if isinstance(url_or_request
, compat_urllib_request
.Request
):
654 url_or_request
= update_Request(
655 url_or_request
, data
=data
, headers
=headers
, query
=query
)
658 url_or_request
= update_url_query(url_or_request
, query
)
659 if data
is not None or headers
:
660 url_or_request
= sanitized_Request(url_or_request
, data
, headers
)
662 return self
._downloader
.urlopen(url_or_request
)
663 except network_exceptions
as err
:
664 if isinstance(err
, compat_urllib_error
.HTTPError
):
665 if self
.__can
_accept
_status
_code
(err
, expected_status
):
666 # Retain reference to error to prevent file object from
667 # being closed before it can be read. Works around the
668 # effects of <https://bugs.python.org/issue15002>
669 # introduced in Python 3.4.1.
676 errnote
= 'Unable to download webpage'
678 errmsg
= '%s: %s' % (errnote
, error_to_compat_str(err
))
680 raise ExtractorError(errmsg
, sys
.exc_info()[2], cause
=err
)
682 self
.report_warning(errmsg
)
685 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):
687 Return a tuple (page content as string, URL handle).
689 See _download_webpage docstring for arguments specification.
691 # Strip hashes from the URL (#1038)
692 if isinstance(url_or_request
, (compat_str
, str)):
693 url_or_request
= url_or_request
.partition('#')[0]
695 urlh
= self
._request
_webpage
(url_or_request
, video_id
, note
, errnote
, fatal
, data
=data
, headers
=headers
, query
=query
, expected_status
=expected_status
)
699 content
= self
._webpage
_read
_content
(urlh
, url_or_request
, video_id
, note
, errnote
, fatal
, encoding
=encoding
)
700 return (content
, urlh
)
703 def _guess_encoding_from_content(content_type
, webpage_bytes
):
704 m
= re
.match(r
'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type
)
706 encoding
= m
.group(1)
708 m
= re
.search(br
'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
709 webpage_bytes[:1024])
711 encoding = m.group(1).decode('ascii')
712 elif webpage_bytes.startswith(b'\xff\xfe'):
719 def __check_blocked(self, content):
720 first_block = content[:512]
721 if ('<title>Access to this site is blocked</title>' in content
722 and 'Websense' in first_block):
723 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
724 blocked_iframe = self._html_search_regex(
725 r'<iframe src="([^
"]+)"', content,
726 'Websense information URL
', default=None)
728 msg += ' Visit
%s for more details
' % blocked_iframe
729 raise ExtractorError(msg, expected=True)
730 if '<title
>The URL you requested has been blocked
</title
>' in first_block:
732 'Access to this webpage has been blocked by Indian censorship
. '
733 'Use a VPN
or proxy
server (with --proxy
) to route around it
.')
734 block_msg = self._html_search_regex(
735 r'</h1
><p
>(.*?
)</p
>',
736 content, 'block message
', default=None)
738 msg += ' (Message
: "%s")' % block_msg.replace('\n', ' ')
739 raise ExtractorError(msg, expected=True)
740 if ('<title
>TTK
:: Доступ к ресурсу ограничен
</title
>' in content
741 and 'blocklist
.rkn
.gov
.ru
' in content):
742 raise ExtractorError(
743 'Access to this webpage has been blocked by decision of the Russian government
. '
744 'Visit http
://blocklist
.rkn
.gov
.ru
/ for a block reason
.',
747 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
748 content_type = urlh.headers.get('Content
-Type
', '')
749 webpage_bytes = urlh.read()
750 if prefix is not None:
751 webpage_bytes = prefix + webpage_bytes
753 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
754 if self._downloader.params.get('dump_intermediate_pages
', False):
755 self.to_screen('Dumping request to
' + urlh.geturl())
756 dump = base64.b64encode(webpage_bytes).decode('ascii
')
757 self._downloader.to_screen(dump)
758 if self._downloader.params.get('write_pages
', False):
759 basen = '%s_%s' % (video_id, urlh.geturl())
761 h = '___
' + hashlib.md5(basen.encode('utf
-8')).hexdigest()
762 basen = basen[:240 - len(h)] + h
763 raw_filename = basen + '.dump
'
764 filename = sanitize_filename(raw_filename, restricted=True)
765 self.to_screen('Saving request to
' + filename)
766 # Working around MAX_PATH limitation on Windows (see
767 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
768 if compat_os_name == 'nt
':
769 absfilepath = os.path.abspath(filename)
770 if len(absfilepath) > 259:
771 filename = '\\\\?
\\' + absfilepath
772 with open(filename, 'wb
') as outf:
773 outf.write(webpage_bytes)
776 content = webpage_bytes.decode(encoding, 'replace
')
778 content = webpage_bytes.decode('utf
-8', 'replace
')
780 self.__check_blocked(content)
784 def _download_webpage(
785 self, url_or_request, video_id, note=None, errnote=None,
786 fatal=True, tries=1, timeout=5, encoding=None, data=None,
787 headers={}, query={}, expected_status=None):
789 Return the data of the page as a string.
792 url_or_request -- plain text URL as a string or
793 a compat_urllib_request.Requestobject
794 video_id -- Video/playlist/item identifier (string)
797 note -- note printed before downloading (string)
798 errnote -- note printed in case of an error (string)
799 fatal -- flag denoting whether error should be considered fatal,
800 i.e. whether it should cause ExtractionError to be raised,
801 otherwise a warning will be reported and extraction continued
802 tries -- number of tries
803 timeout -- sleep interval between tries
804 encoding -- encoding for a page content decoding, guessed automatically
805 when not explicitly specified
806 data -- POST data (bytes)
807 headers -- HTTP headers (dict)
808 query -- URL query (dict)
809 expected_status -- allows to accept failed HTTP requests (non 2xx
810 status code) by explicitly specifying a set of accepted status
811 codes. Can be any of the following entities:
812 - an integer type specifying an exact failed status code to
814 - a list or a tuple of integer types specifying a list of
815 failed status codes to accept
816 - a callable accepting an actual failed status code and
817 returning True if it should be accepted
818 Note that this argument does not affect success status codes (2xx)
819 which are always accepted.
824 while success is False:
826 res = self._download_webpage_handle(
827 url_or_request, video_id, note, errnote, fatal,
828 encoding=encoding, data=data, headers=headers, query=query,
829 expected_status=expected_status)
831 except compat_http_client.IncompleteRead as e:
833 if try_count >= tries:
835 self._sleep(timeout, video_id)
842 def _download_xml_handle(
843 self, url_or_request, video_id, note='Downloading XML
',
844 errnote='Unable to download XML
', transform_source=None,
845 fatal=True, encoding=None, data=None, headers={}, query={},
846 expected_status=None):
848 Return a tuple (xml as an compat_etree_Element, URL handle).
850 See _download_webpage docstring for arguments specification.
852 res = self._download_webpage_handle(
853 url_or_request, video_id, note, errnote, fatal=fatal,
854 encoding=encoding, data=data, headers=headers, query=query,
855 expected_status=expected_status)
858 xml_string, urlh = res
859 return self._parse_xml(
860 xml_string, video_id, transform_source=transform_source,
864 self, url_or_request, video_id,
865 note='Downloading XML
', errnote='Unable to download XML
',
866 transform_source=None, fatal=True, encoding=None,
867 data=None, headers={}, query={}, expected_status=None):
869 Return the xml as an compat_etree_Element.
871 See _download_webpage docstring for arguments specification.
873 res = self._download_xml_handle(
874 url_or_request, video_id, note=note, errnote=errnote,
875 transform_source=transform_source, fatal=fatal, encoding=encoding,
876 data=data, headers=headers, query=query,
877 expected_status=expected_status)
878 return res if res is False else res[0]
880 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
882 xml_string = transform_source(xml_string)
884 return compat_etree_fromstring(xml_string.encode('utf
-8'))
885 except compat_xml_parse_error as ve:
886 errmsg = '%s: Failed to parse XML
' % video_id
888 raise ExtractorError(errmsg, cause=ve)
890 self.report_warning(errmsg + str(ve))
892 def _download_json_handle(
893 self, url_or_request, video_id, note='Downloading JSON metadata
',
894 errnote='Unable to download JSON metadata
', transform_source=None,
895 fatal=True, encoding=None, data=None, headers={}, query={},
896 expected_status=None):
898 Return a tuple (JSON object, URL handle).
900 See _download_webpage docstring for arguments specification.
902 res = self._download_webpage_handle(
903 url_or_request, video_id, note, errnote, fatal=fatal,
904 encoding=encoding, data=data, headers=headers, query=query,
905 expected_status=expected_status)
908 json_string, urlh = res
909 return self._parse_json(
910 json_string, video_id, transform_source=transform_source,
914 self, url_or_request, video_id, note='Downloading JSON metadata
',
915 errnote='Unable to download JSON metadata
', transform_source=None,
916 fatal=True, encoding=None, data=None, headers={}, query={},
917 expected_status=None):
919 Return the JSON object as a dict.
921 See _download_webpage docstring for arguments specification.
923 res = self._download_json_handle(
924 url_or_request, video_id, note=note, errnote=errnote,
925 transform_source=transform_source, fatal=fatal, encoding=encoding,
926 data=data, headers=headers, query=query,
927 expected_status=expected_status)
928 return res if res is False else res[0]
930 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
932 json_string = transform_source(json_string)
934 return json.loads(json_string)
935 except ValueError as ve:
936 errmsg = '%s: Failed to parse JSON
' % video_id
938 raise ExtractorError(errmsg, cause=ve)
940 self.report_warning(errmsg + str(ve))
942 def report_warning(self, msg, video_id=None):
943 idstr = '' if video_id is None else '%s: ' % video_id
944 self._downloader.report_warning(
945 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
947 def to_screen(self, msg):
948 """Print msg to screen, prefixing it with '[ie_name
]'"""
949 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
951 def report_extraction(self, id_or_name):
952 """Report information extraction."""
953 self.to_screen('%s: Extracting information
' % id_or_name)
955 def report_download_webpage(self, video_id):
956 """Report webpage download."""
957 self.to_screen('%s: Downloading webpage
' % video_id)
959 def report_age_confirmation(self):
960 """Report attempt to confirm age."""
961 self.to_screen('Confirming age
')
963 def report_login(self):
964 """Report attempt to log in."""
965 self.to_screen('Logging
in')
967 def raise_login_required(
968 self, msg='This video
is only available
for registered users
', metadata_available=False):
969 if metadata_available and self._downloader.params.get('ignore_no_formats_error
'):
970 self.report_warning(msg)
971 raise ExtractorError(
972 '%s. Use
--cookies
, --username
and --password
or --netrc to provide account credentials
' % msg,
975 def raise_geo_restricted(
976 self, msg='This video
is not available
from your location due to geo restriction
',
977 countries=None, metadata_available=False):
978 if metadata_available and self._downloader.params.get('ignore_no_formats_error
'):
979 self.report_warning(msg)
981 raise GeoRestrictedError(msg, countries=countries)
983 def raise_no_formats(self, msg, expected=False, video_id=None):
984 if expected and self._downloader.params.get('ignore_no_formats_error
'):
985 self.report_warning(msg, video_id)
987 raise ExtractorError(msg, expected=expected, video_id=video_id)
989 # Methods for following #608
991 def url_result(url, ie=None, video_id=None, video_title=None):
992 """Returns a URL that points to a page that should be processed"""
993 # TODO: ie should be the class used for getting the info
994 video_info = {'_type
': 'url
',
997 if video_id is not None:
998 video_info['id'] = video_id
999 if video_title is not None:
1000 video_info['title
'] = video_title
1003 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None):
1005 self.url_result(self._proto_relative_url(getter(m) if getter else m), ie)
1007 return self.playlist_result(
1008 urls, playlist_id=playlist_id, playlist_title=playlist_title)
1011 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None, **kwargs):
1012 """Returns a playlist"""
1013 video_info = {'_type
': 'playlist
',
1015 video_info.update(kwargs)
1017 video_info['id'] = playlist_id
1019 video_info['title
'] = playlist_title
1020 if playlist_description is not None:
1021 video_info['description
'] = playlist_description
1024 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1026 Perform a regex search on the given string, using a single or a list of
1027 patterns returning the first matching group.
1028 In case of failure return a default value or raise a WARNING or a
1029 RegexNotFoundError, depending on fatal, specifying the field name.
1031 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
1032 mobj = re.search(pattern, string, flags)
1035 mobj = re.search(p, string, flags)
1039 if not self._downloader.params.get('no_color
') and compat_os_name != 'nt
' and sys.stderr.isatty():
1040 _name = '\033[0;34m
%s\033[0m
' % name
1046 # return the first matching group
1047 return next(g for g in mobj.groups() if g is not None)
1049 return mobj.group(group)
1050 elif default is not NO_DEFAULT:
1053 raise RegexNotFoundError('Unable to extract
%s' % _name)
1055 self.report_warning('unable to extract
%s' % _name + bug_reports_message())
1058 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1060 Like _search_regex, but strips HTML tags and unescapes entities.
1062 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1064 return clean_html(res).strip()
1068 def _get_netrc_login_info(self, netrc_machine=None):
1071 netrc_machine = netrc_machine or self._NETRC_MACHINE
1073 if self._downloader.params.get('usenetrc
', False):
1075 info = netrc.netrc().authenticators(netrc_machine)
1076 if info is not None:
1080 raise netrc.NetrcParseError(
1081 'No authenticators
for %s' % netrc_machine)
1082 except (IOError, netrc.NetrcParseError) as err:
1083 self.report_warning(
1084 'parsing
.netrc
: %s' % error_to_compat_str(err))
1086 return username, password
1088 def _get_login_info(self, username_option='username
', password_option='password
', netrc_machine=None):
1090 Get the login info as (username, password)
1091 First look for the manually specified credentials using username_option
1092 and password_option as keys in params dictionary. If no such credentials
1093 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1095 If there's no info available
, return (None, None)
1097 if self._downloader is None:
1100 downloader_params = self._downloader.params
1102 # Attempt to use provided username and password or .netrc data
1103 if downloader_params.get(username_option) is not None:
1104 username = downloader_params[username_option]
1105 password = downloader_params[password_option]
1107 username, password = self._get_netrc_login_info(netrc_machine)
1109 return username, password
1111 def _get_tfa_info(self, note='two-factor verification code'):
1113 Get the two
-factor authentication info
1114 TODO
- asking the user will be required
for sms
/phone verify
1115 currently just uses the command line option
1116 If there
's no info available, return None
1118 if self._downloader is None:
1120 downloader_params = self._downloader.params
1122 if downloader_params.get('twofactor
') is not None:
1123 return downloader_params['twofactor
']
1125 return compat_getpass('Type
%s and press
[Return
]: ' % note)
1127 # Helper functions for extracting OpenGraph info
1129 def _og_regexes(prop):
1130 content_re = r'content
=(?
:"([^"]+?
)"|\'([^\']+?)\'|\s*([^\s"\'=<>`
]+?
))'
1131 property_re = (r'(?
:name|
property)=(?
:\'og
[:-]%(prop)s\'|
"og[:-]%(prop)s"|\s
*og
[:-]%(prop)s\b)'
1132 % {'prop': re.escape(prop)})
1133 template = r'<meta
[^
>]+?
%s[^
>]+?
%s'
1135 template % (property_re, content_re),
1136 template % (content_re, property_re),
1140 def _meta_regex(prop):
1141 return r'''(?isx)<meta
1142 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1143 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1145 def _og_search_property(self, prop, html, name=None, **kargs):
1146 if not isinstance(prop, (list, tuple)):
1149 name = 'OpenGraph
%s' % prop[0]
1152 og_regexes.extend(self._og_regexes(p))
1153 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1156 return unescapeHTML(escaped)
1158 def _og_search_thumbnail(self, html, **kargs):
1159 return self._og_search_property('image
', html, 'thumbnail URL
', fatal=False, **kargs)
1161 def _og_search_description(self, html, **kargs):
1162 return self._og_search_property('description
', html, fatal=False, **kargs)
1164 def _og_search_title(self, html, **kargs):
1165 return self._og_search_property('title
', html, **kargs)
1167 def _og_search_video_url(self, html, name='video url
', secure=True, **kargs):
1168 regexes = self._og_regexes('video
') + self._og_regexes('video
:url
')
1170 regexes = self._og_regexes('video
:secure_url
') + regexes
1171 return self._html_search_regex(regexes, html, name, **kargs)
1173 def _og_search_url(self, html, **kargs):
1174 return self._og_search_property('url
', html, **kargs)
1176 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1177 if not isinstance(name, (list, tuple)):
1179 if display_name is None:
1180 display_name = name[0]
1181 return self._html_search_regex(
1182 [self._meta_regex(n) for n in name],
1183 html, display_name, fatal=fatal, group='content
', **kwargs)
1185 def _dc_search_uploader(self, html):
1186 return self._html_search_meta('dc
.creator
', html, 'uploader
')
1188 def _rta_search(self, html):
1189 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1190 if re.search(r'(?ix
)<meta\s
+name
="rating"\s
+'
1191 r' content
="RTA-5042-1996-1400-1577-RTA"',
1196 def _media_rating_search(self, html):
1197 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1198 rating = self._html_search_meta('rating
', html)
1210 return RATING_TABLE.get(rating.lower())
1212 def _family_friendly_search(self, html):
1213 # See http://schema.org/VideoObject
1214 family_friendly = self._html_search_meta(
1215 'isFamilyFriendly
', html, default=None)
1217 if not family_friendly:
1226 return RATING_TABLE.get(family_friendly.lower())
1228 def _twitter_search_player(self, html):
1229 return self._html_search_meta('twitter
:player
', html,
1230 'twitter card player
')
1232 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
1233 json_ld_list = list(re.finditer(JSON_LD_RE, html))
1234 default = kwargs.get('default
', NO_DEFAULT)
1235 # JSON-LD may be malformed and thus `fatal` should be respected.
1236 # At the same time `default` may be passed that assumes `fatal=False`
1237 # for _search_regex. Let's simulate the same behavior here
as well
.
1238 fatal
= kwargs
.get('fatal', True) if default
== NO_DEFAULT
else False
1240 for mobj
in json_ld_list
:
1241 json_ld_item
= self
._parse
_json
(
1242 mobj
.group('json_ld'), video_id
, fatal
=fatal
)
1243 if not json_ld_item
:
1245 if isinstance(json_ld_item
, dict):
1246 json_ld
.append(json_ld_item
)
1247 elif isinstance(json_ld_item
, (list, tuple)):
1248 json_ld
.extend(json_ld_item
)
1250 json_ld
= self
._json
_ld
(json_ld
, video_id
, fatal
=fatal
, expected_type
=expected_type
)
1253 if default
is not NO_DEFAULT
:
1256 raise RegexNotFoundError('Unable to extract JSON-LD')
1258 self
.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1261 def _json_ld(self
, json_ld
, video_id
, fatal
=True, expected_type
=None):
1262 if isinstance(json_ld
, compat_str
):
1263 json_ld
= self
._parse
_json
(json_ld
, video_id
, fatal
=fatal
)
1267 if not isinstance(json_ld
, (list, tuple, dict)):
1269 if isinstance(json_ld
, dict):
1272 INTERACTION_TYPE_MAP
= {
1273 'CommentAction': 'comment',
1274 'AgreeAction': 'like',
1275 'DisagreeAction': 'dislike',
1276 'LikeAction': 'like',
1277 'DislikeAction': 'dislike',
1278 'ListenAction': 'view',
1279 'WatchAction': 'view',
1280 'ViewAction': 'view',
1283 def extract_interaction_type(e
):
1284 interaction_type
= e
.get('interactionType')
1285 if isinstance(interaction_type
, dict):
1286 interaction_type
= interaction_type
.get('@type')
1287 return str_or_none(interaction_type
)
1289 def extract_interaction_statistic(e
):
1290 interaction_statistic
= e
.get('interactionStatistic')
1291 if isinstance(interaction_statistic
, dict):
1292 interaction_statistic
= [interaction_statistic
]
1293 if not isinstance(interaction_statistic
, list):
1295 for is_e
in interaction_statistic
:
1296 if not isinstance(is_e
, dict):
1298 if is_e
.get('@type') != 'InteractionCounter':
1300 interaction_type
= extract_interaction_type(is_e
)
1301 if not interaction_type
:
1303 # For interaction count some sites provide string instead of
1304 # an integer (as per spec) with non digit characters (e.g. ",")
1305 # so extracting count with more relaxed str_to_int
1306 interaction_count
= str_to_int(is_e
.get('userInteractionCount'))
1307 if interaction_count
is None:
1309 count_kind
= INTERACTION_TYPE_MAP
.get(interaction_type
.split('/')[-1])
1312 count_key
= '%s_count' % count_kind
1313 if info
.get(count_key
) is not None:
1315 info
[count_key
] = interaction_count
1317 def extract_video_object(e
):
1318 assert e
['@type'] == 'VideoObject'
1319 author
= e
.get('author')
1321 'url': url_or_none(e
.get('contentUrl')),
1322 'title': unescapeHTML(e
.get('name')),
1323 'description': unescapeHTML(e
.get('description')),
1324 'thumbnail': url_or_none(e
.get('thumbnailUrl') or e
.get('thumbnailURL')),
1325 'duration': parse_duration(e
.get('duration')),
1326 'timestamp': unified_timestamp(e
.get('uploadDate')),
1327 # author can be an instance of 'Organization' or 'Person' types.
1328 # both types can have 'name' property(inherited from 'Thing' type). [1]
1329 # however some websites are using 'Text' type instead.
1330 # 1. https://schema.org/VideoObject
1331 'uploader': author
.get('name') if isinstance(author
, dict) else author
if isinstance(author
, compat_str
) else None,
1332 'filesize': float_or_none(e
.get('contentSize')),
1333 'tbr': int_or_none(e
.get('bitrate')),
1334 'width': int_or_none(e
.get('width')),
1335 'height': int_or_none(e
.get('height')),
1336 'view_count': int_or_none(e
.get('interactionCount')),
1338 extract_interaction_statistic(e
)
1342 item_type
= e
.get('@type')
1343 if expected_type
is not None and expected_type
!= item_type
:
1345 if item_type
in ('TVEpisode', 'Episode'):
1346 episode_name
= unescapeHTML(e
.get('name'))
1348 'episode': episode_name
,
1349 'episode_number': int_or_none(e
.get('episodeNumber')),
1350 'description': unescapeHTML(e
.get('description')),
1352 if not info
.get('title') and episode_name
:
1353 info
['title'] = episode_name
1354 part_of_season
= e
.get('partOfSeason')
1355 if isinstance(part_of_season
, dict) and part_of_season
.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1357 'season': unescapeHTML(part_of_season
.get('name')),
1358 'season_number': int_or_none(part_of_season
.get('seasonNumber')),
1360 part_of_series
= e
.get('partOfSeries') or e
.get('partOfTVSeries')
1361 if isinstance(part_of_series
, dict) and part_of_series
.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1362 info
['series'] = unescapeHTML(part_of_series
.get('name'))
1363 elif item_type
== 'Movie':
1365 'title': unescapeHTML(e
.get('name')),
1366 'description': unescapeHTML(e
.get('description')),
1367 'duration': parse_duration(e
.get('duration')),
1368 'timestamp': unified_timestamp(e
.get('dateCreated')),
1370 elif item_type
in ('Article', 'NewsArticle'):
1372 'timestamp': parse_iso8601(e
.get('datePublished')),
1373 'title': unescapeHTML(e
.get('headline')),
1374 'description': unescapeHTML(e
.get('articleBody')),
1376 elif item_type
== 'VideoObject':
1377 extract_video_object(e
)
1378 if expected_type
is None:
1382 video
= e
.get('video')
1383 if isinstance(video
, dict) and video
.get('@type') == 'VideoObject':
1384 extract_video_object(video
)
1385 if expected_type
is None:
1389 return dict((k
, v
) for k
, v
in info
.items() if v
is not None)
1392 def _hidden_inputs(html
):
1393 html
= re
.sub(r
'<!--(?:(?!<!--).)*-->', '', html
)
1395 for input in re
.findall(r
'(?i)(<input[^>]+>)', html
):
1396 attrs
= extract_attributes(input)
1399 if attrs
.get('type') not in ('hidden', 'submit'):
1401 name
= attrs
.get('name') or attrs
.get('id')
1402 value
= attrs
.get('value')
1403 if name
and value
is not None:
1404 hidden_inputs
[name
] = value
1405 return hidden_inputs
1407 def _form_hidden_inputs(self
, form_id
, html
):
1408 form
= self
._search
_regex
(
1409 r
'(?is)<form[^>]+?id=(["\'])%s\
1[^
>]*>(?P
<form
>.+?
)</form
>' % form_id,
1410 html, '%s form
' % form_id, group='form
')
1411 return self._hidden_inputs(form)
1414 regex = r' *((?P
<reverse
>\
+)?
(?P
<field
>[a
-zA
-Z0
-9_]+)((?P
<separator
>[~
:])(?P
<limit
>.*?
))?
)?
*$
'
1416 default = ('hidden
', 'hasvid
', 'ie_pref
', 'lang
', 'quality
',
1417 'res
', 'fps
', 'codec
:vp9
.2
', 'size
', 'br
', 'asr
',
1418 'proto
', 'ext
', 'hasaud
', 'source
', 'format_id
') # These must not be aliases
1419 ytdl_default = ('hasaud
', 'quality
', 'tbr
', 'filesize
', 'vbr
',
1420 'height
', 'width
', 'proto
', 'vext
', 'abr
', 'aext
',
1421 'fps
', 'fs_approx
', 'source
', 'format_id
')
1424 'vcodec
': {'type': 'ordered
', 'regex
': True,
1425 'order
': ['av0?
1', 'vp0?
9.2', 'vp0?
9', '[hx
]265|he?vc?
', '[hx
]264|avc
', 'vp0?
8', 'mp4v|h263
', 'theora
', '', None, 'none
']},
1426 'acodec
': {'type': 'ordered
', 'regex
': True,
1427 'order
': ['opus
', 'vorbis
', 'aac
', 'mp?
4a?
', 'mp3
', 'e?a?c
-?
3', 'dts
', '', None, 'none
']},
1428 'proto
': {'type': 'ordered
', 'regex
': True, 'field
': 'protocol
',
1429 'order
': ['(ht|f
)tps
', '(ht|f
)tp$
', 'm3u8
.+', 'm3u8
', '.*dash
', '', 'mms|rtsp
', 'none
', 'f4
']},
1430 'vext
': {'type': 'ordered
', 'field
': 'video_ext
',
1431 'order
': ('mp4
', 'webm
', 'flv
', '', 'none
'),
1432 'order_free
': ('webm
', 'mp4
', 'flv
', '', 'none
')},
1433 'aext
': {'type': 'ordered
', 'field
': 'audio_ext
',
1434 'order
': ('m4a
', 'aac
', 'mp3
', 'ogg
', 'opus
', 'webm
', '', 'none
'),
1435 'order_free
': ('opus
', 'ogg
', 'webm
', 'm4a
', 'mp3
', 'aac
', '', 'none
')},
1436 'hidden
': {'visible': False, 'forced': True, 'type': 'extractor', 'max': -1000},
1437 'ie_pref
': {'priority': True, 'type': 'extractor'},
1438 'hasvid
': {'priority': True, 'field': 'vcodec', 'type': 'boolean', 'not_in_list': ('none',)},
1439 'hasaud
': {'field': 'acodec', 'type': 'boolean', 'not_in_list': ('none',)},
1440 'lang
': {'priority': True, 'convert': 'ignore', 'field': 'language_preference'},
1441 'quality
': {'convert': 'float_none', 'default': -1},
1442 'filesize
': {'convert': 'bytes'},
1443 'fs_approx
': {'convert': 'bytes', 'field': 'filesize_approx'},
1444 'id': {'convert': 'string', 'field': 'format_id'},
1445 'height
': {'convert': 'float_none'},
1446 'width
': {'convert': 'float_none'},
1447 'fps
': {'convert': 'float_none'},
1448 'tbr
': {'convert': 'float_none'},
1449 'vbr
': {'convert': 'float_none'},
1450 'abr
': {'convert': 'float_none'},
1451 'asr
': {'convert': 'float_none'},
1452 'source
': {'convert': 'ignore', 'field': 'source_preference'},
1454 'codec
': {'type': 'combined', 'field': ('vcodec', 'acodec')},
1455 'br
': {'type': 'combined', 'field': ('tbr', 'vbr', 'abr'), 'same_limit': True},
1456 'size
': {'type': 'combined', 'same_limit': True, 'field': ('filesize', 'fs_approx')},
1457 'ext
': {'type': 'combined', 'field': ('vext', 'aext')},
1458 'res
': {'type': 'multiple', 'field': ('height', 'width'), 'function': min},
1460 # Most of these exist only for compatibility reasons
1461 'dimension
': {'type': 'alias', 'field': 'res'},
1462 'resolution
': {'type': 'alias', 'field': 'res'},
1463 'extension
': {'type': 'alias', 'field': 'ext'},
1464 'bitrate
': {'type': 'alias', 'field': 'br'},
1465 'total_bitrate
': {'type': 'alias', 'field': 'tbr'},
1466 'video_bitrate
': {'type': 'alias', 'field': 'vbr'},
1467 'audio_bitrate
': {'type': 'alias', 'field': 'abr'},
1468 'framerate
': {'type': 'alias', 'field': 'fps'},
1469 'language_preference
': {'type': 'alias', 'field': 'lang'}, # not named as 'language
' because such a field exists
1470 'protocol
': {'type': 'alias', 'field': 'proto'},
1471 'source_preference
': {'type': 'alias', 'field': 'source'},
1472 'filesize_approx
': {'type': 'alias', 'field': 'fs_approx'},
1473 'filesize_estimate
': {'type': 'alias', 'field': 'size'},
1474 'samplerate
': {'type': 'alias', 'field': 'asr'},
1475 'video_ext
': {'type': 'alias', 'field': 'vext'},
1476 'audio_ext
': {'type': 'alias', 'field': 'aext'},
1477 'video_codec
': {'type': 'alias', 'field': 'vcodec'},
1478 'audio_codec
': {'type': 'alias', 'field': 'acodec'},
1479 'video
': {'type': 'alias', 'field': 'hasvid'},
1480 'has_video
': {'type': 'alias', 'field': 'hasvid'},
1481 'audio
': {'type': 'alias', 'field': 'hasaud'},
1482 'has_audio
': {'type': 'alias', 'field': 'hasaud'},
1483 'extractor
': {'type': 'alias', 'field': 'ie_pref'},
1484 'preference
': {'type': 'alias', 'field': 'ie_pref'},
1485 'extractor_preference
': {'type': 'alias', 'field': 'ie_pref'},
1486 'format_id
': {'type': 'alias', 'field': 'id'},
1491 def _get_field_setting(self, field, key):
1492 if field not in self.settings:
1493 self.settings[field] = {}
1494 propObj = self.settings[field]
1495 if key not in propObj:
1496 type = propObj.get('type')
1498 default = 'preference
' if type == 'extractor
' else (field,) if type in ('combined
', 'multiple
') else field
1499 elif key == 'convert
':
1500 default = 'order
' if type == 'ordered
' else 'float_string
' if field else 'ignore
'
1502 default = {'type': 'field', 'visible': True, 'order': [], 'not_in_list': (None,), 'function': max}.get(key, None)
1503 propObj[key] = default
1506 def _resolve_field_value(self, field, value, convertNone=False):
1511 value = value.lower()
1512 conversion = self._get_field_setting(field, 'convert
')
1513 if conversion == 'ignore
':
1515 if conversion == 'string
':
1517 elif conversion == 'float_none
':
1518 return float_or_none(value)
1519 elif conversion == 'bytes':
1520 return FileDownloader.parse_bytes(value)
1521 elif conversion == 'order
':
1522 order_list = (self._use_free_order and self._get_field_setting(field, 'order_free
')) or self._get_field_setting(field, 'order
')
1523 use_regex = self._get_field_setting(field, 'regex
')
1524 list_length = len(order_list)
1525 empty_pos = order_list.index('') if '' in order_list else list_length + 1
1526 if use_regex and value is not None:
1527 for i, regex in enumerate(order_list):
1528 if regex and re.match(regex, value):
1529 return list_length - i
1530 return list_length - empty_pos # not in list
1531 else: # not regex or value = None
1532 return list_length - (order_list.index(value) if value in order_list else empty_pos)
1534 if value.isnumeric():
1537 self.settings[field]['convert
'] = 'string
'
1540 def evaluate_params(self, params, sort_extractor):
1541 self._use_free_order = params.get('prefer_free_formats
', False)
1542 self._sort_user = params.get('format_sort
', [])
1543 self._sort_extractor = sort_extractor
1545 def add_item(field, reverse, closest, limit_text):
1546 field = field.lower()
1547 if field in self._order:
1549 self._order.append(field)
1550 limit = self._resolve_field_value(field, limit_text)
1553 'closest
': False if limit is None else closest,
1554 'limit_text
': limit_text,
1556 if field in self.settings:
1557 self.settings[field].update(data)
1559 self.settings[field] = data
1562 tuple(field for field in self.default if self._get_field_setting(field, 'forced
'))
1563 + (tuple() if params.get('format_sort_force
', False)
1564 else tuple(field for field in self.default if self._get_field_setting(field, 'priority
')))
1565 + tuple(self._sort_user) + tuple(sort_extractor) + self.default)
1567 for item in sort_list:
1568 match = re.match(self.regex, item)
1570 raise ExtractorError('Invalid format sort string
"%s" given by extractor
' % item)
1571 field = match.group('field
')
1574 if self._get_field_setting(field, 'type') == 'alias
':
1575 field = self._get_field_setting(field, 'field
')
1576 reverse = match.group('reverse
') is not None
1577 closest = match.group('separator
') == '~
'
1578 limit_text = match.group('limit
')
1580 has_limit = limit_text is not None
1581 has_multiple_fields = self._get_field_setting(field, 'type') == 'combined
'
1582 has_multiple_limits = has_limit and has_multiple_fields and not self._get_field_setting(field, 'same_limit
')
1584 fields = self._get_field_setting(field, 'field
') if has_multiple_fields else (field,)
1585 limits = limit_text.split(":") if has_multiple_limits else (limit_text,) if has_limit else tuple()
1586 limit_count = len(limits)
1587 for (i, f) in enumerate(fields):
1588 add_item(f, reverse, closest,
1589 limits[i] if i < limit_count
1590 else limits[0] if has_limit and not has_multiple_limits
1593 def print_verbose_info(self, write_debug):
1595 write_debug('Sort order given by user
: %s' % ', '.join(self._sort_user))
1596 if self._sort_extractor:
1597 write_debug('Sort order given by extractor
: %s' % ', '.join(self._sort_extractor))
1598 write_debug('Formats
sorted by
: %s' % ', '.join(['%s%s%s' % (
1599 '+' if self._get_field_setting(field, 'reverse
') else '', field,
1600 '%s%s(%s)' % ('~
' if self._get_field_setting(field, 'closest
') else ':',
1601 self._get_field_setting(field, 'limit_text
'),
1602 self._get_field_setting(field, 'limit
'))
1603 if self._get_field_setting(field, 'limit_text
') is not None else '')
1604 for field in self._order if self._get_field_setting(field, 'visible
')]))
1606 def _calculate_field_preference_from_value(self, format, field, type, value):
1607 reverse = self._get_field_setting(field, 'reverse
')
1608 closest = self._get_field_setting(field, 'closest
')
1609 limit = self._get_field_setting(field, 'limit
')
1611 if type == 'extractor
':
1612 maximum = self._get_field_setting(field, 'max')
1613 if value is None or (maximum is not None and value >= maximum):
1615 elif type == 'boolean
':
1616 in_list = self._get_field_setting(field, 'in_list
')
1617 not_in_list = self._get_field_setting(field, 'not_in_list
')
1618 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
1619 elif type == 'ordered
':
1620 value = self._resolve_field_value(field, value, True)
1622 # try to convert to number
1623 val_num = float_or_none(value, default=self._get_field_setting(field, 'default
'))
1624 is_num = self._get_field_setting(field, 'convert
') != 'string
' and val_num is not None
1628 return ((-10, 0) if value is None
1629 else (1, value, 0) if not is_num # if a field has mixed strings and numbers, strings are sorted higher
1630 else (0, -abs(value - limit), value - limit if reverse else limit - value) if closest
1631 else (0, value, 0) if not reverse and (limit is None or value <= limit)
1632 else (0, -value, 0) if limit is None or (reverse and value == limit) or value > limit
1633 else (-1, value, 0))
1635 def _calculate_field_preference(self, format, field):
1636 type = self._get_field_setting(field, 'type') # extractor, boolean, ordered, field, multiple
1637 get_value = lambda f: format.get(self._get_field_setting(f, 'field
'))
1638 if type == 'multiple
':
1639 type = 'field
' # Only 'field
' is allowed in multiple for now
1640 actual_fields = self._get_field_setting(field, 'field
')
1642 def wrapped_function(values):
1643 values = tuple(filter(lambda x: x is not None, values))
1644 return (self._get_field_setting(field, 'function
')(*values) if len(values) > 1
1645 else values[0] if values
1648 value = wrapped_function((get_value(f) for f in actual_fields))
1650 value = get_value(field)
1651 return self._calculate_field_preference_from_value(format, field, type, value)
1653 def calculate_preference(self, format):
1654 # Determine missing protocol
1655 if not format.get('protocol
'):
1656 format['protocol
'] = determine_protocol(format)
1658 # Determine missing ext
1659 if not format.get('ext
') and 'url
' in format:
1660 format['ext
'] = determine_ext(format['url
'])
1661 if format.get('vcodec
') == 'none
':
1662 format['audio_ext
'] = format['ext
']
1663 format['video_ext
'] = 'none
'
1665 format['video_ext
'] = format['ext
']
1666 format['audio_ext
'] = 'none
'
1667 # if format.get('preference
') is None and format.get('ext
') in ('f4f
', 'f4m
'): # Not supported?
1668 # format['preference
'] = -1000
1670 # Determine missing bitrates
1671 if format.get('tbr
') is None:
1672 if format.get('vbr
') is not None and format.get('abr
') is not None:
1673 format['tbr
'] = format.get('vbr
', 0) + format.get('abr
', 0)
1675 if format.get('vcodec
') != "none" and format.get('vbr
') is None:
1676 format['vbr
'] = format.get('tbr
') - format.get('abr
', 0)
1677 if format.get('acodec
') != "none" and format.get('abr
') is None:
1678 format['abr
'] = format.get('tbr
') - format.get('vbr
', 0)
1680 return tuple(self._calculate_field_preference(format, field) for field in self._order)
1682 def _sort_formats(self, formats, field_preference=[]):
1684 if self._downloader.params.get('ignore_no_formats_error
'):
1686 raise ExtractorError('No video formats found
')
1687 format_sort = self.FormatSort() # params and to_screen are taken from the downloader
1688 format_sort.evaluate_params(self._downloader.params, field_preference)
1689 if self._downloader.params.get('verbose
', False):
1690 format_sort.print_verbose_info(self._downloader.write_debug)
1691 formats.sort(key=lambda f: format_sort.calculate_preference(f))
1693 def _check_formats(self, formats, video_id):
1695 formats[:] = filter(
1696 lambda f: self._is_valid_url(
1698 item='%s video format
' % f.get('format_id
') if f.get('format_id
') else 'video
'),
1702 def _remove_duplicate_formats(formats):
1706 if f['url
'] not in format_urls:
1707 format_urls.add(f['url
'])
1708 unique_formats.append(f)
1709 formats[:] = unique_formats
1711 def _is_valid_url(self, url, video_id, item='video
', headers={}):
1712 url = self._proto_relative_url(url, scheme='http
:')
1713 # For now assume non HTTP(S) URLs always valid
1714 if not (url.startswith('http
://') or url.startswith('https
://')):
1717 self._request_webpage(url, video_id, 'Checking
%s URL
' % item, headers=headers)
1719 except ExtractorError as e:
1721 '%s: %s URL
is invalid
, skipping
: %s'
1722 % (video_id, item, error_to_compat_str(e.cause)))
1725 def http_scheme(self):
1726 """ Either "http:" or "https:", depending on the user's preferences
"""
1729 if self._downloader.params.get('prefer_insecure', False)
1732 def _proto_relative_url(self, url, scheme=None):
1735 if url.startswith('//'):
1737 scheme = self.http_scheme()
1742 def _sleep(self, timeout, video_id, msg_template=None):
1743 if msg_template is None:
1744 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1745 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1749 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1750 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1751 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
1752 manifest = self._download_xml(
1753 manifest_url, video_id, 'Downloading f4m manifest',
1754 'Unable to download f4m manifest',
1755 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1756 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
1757 transform_source=transform_source,
1758 fatal=fatal, data=data, headers=headers, query=query)
1760 if manifest is False:
1763 return self._parse_f4m_formats(
1764 manifest, manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1765 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1767 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1768 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1769 fatal=True, m3u8_id=None):
1770 if not isinstance(manifest, compat_etree_Element) and not fatal:
1773 # currently yt-dlp cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1774 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1775 if akamai_pv is not None and ';' in akamai_pv.text:
1776 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1777 if playerVerificationChallenge.strip() != '':
1781 manifest_version = '1.0'
1782 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1784 manifest_version = '2.0'
1785 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1786 # Remove unsupported DRM protected media from final formats
1787 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
1788 media_nodes = remove_encrypted_media(media_nodes)
1792 manifest_base_url = get_base_url(manifest)
1794 bootstrap_info = xpath_element(
1795 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1796 'bootstrap info', default=None)
1799 mime_type = xpath_text(
1800 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1801 'base URL', default=None)
1802 if mime_type and mime_type.startswith('audio/'):
1805 for i, media_el in enumerate(media_nodes):
1806 tbr = int_or_none(media_el.attrib.get('bitrate'))
1807 width = int_or_none(media_el.attrib.get('width'))
1808 height = int_or_none(media_el.attrib.get('height'))
1809 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1810 # If <bootstrapInfo> is present, the specified f4m is a
1811 # stream-level manifest, and only set-level manifests may refer to
1812 # external resources. See section 11.4 and section 4 of F4M spec
1813 if bootstrap_info is None:
1815 # @href is introduced in 2.0, see section 11.6 of F4M spec
1816 if manifest_version == '2.0':
1817 media_url = media_el.attrib.get('href')
1818 if media_url is None:
1819 media_url = media_el.attrib.get('url')
1823 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1824 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1825 # If media_url is itself a f4m manifest do the recursive extraction
1826 # since bitrates in parent manifest (this one) and media_url manifest
1827 # may differ leading to inability to resolve the format by requested
1828 # bitrate in f4m downloader
1829 ext = determine_ext(manifest_url)
1831 f4m_formats = self._extract_f4m_formats(
1832 manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1833 transform_source=transform_source, fatal=fatal)
1834 # Sometimes stream-level manifest contains single media entry that
1835 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1836 # At the same time parent's media entry in set-level manifest may
1837 # contain it. We will copy it from parent in such cases.
1838 if len(f4m_formats) == 1:
1841 'tbr': f.get('tbr') or tbr,
1842 'width': f.get('width') or width,
1843 'height': f.get('height') or height,
1844 'format_id': f.get('format_id') if not tbr else format_id,
1847 formats.extend(f4m_formats)
1850 formats.extend(self._extract_m3u8_formats(
1851 manifest_url, video_id, 'mp4', preference=preference,
1852 quality=quality, m3u8_id=m3u8_id, fatal=fatal))
1855 'format_id': format_id,
1856 'url': manifest_url,
1857 'manifest_url': manifest_url,
1858 'ext': 'flv' if bootstrap_info is not None else None,
1864 'preference': preference,
1869 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, quality=None, m3u8_id=None):
1871 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1875 'preference': preference - 100 if preference else -100,
1877 'resolution': 'multiple',
1878 'format_note': 'Quality selection URL',
1881 def _extract_m3u8_formats(self, *args, **kwargs):
1882 fmts, subs = self._extract_m3u8_formats_and_subtitles(*args, **kwargs)
1884 self.report_warning(bug_reports_message(
1885 "Ignoring subtitle tracks found in the HLS manifest; "
1886 "if any subtitle tracks are missing,"
1890 def _extract_m3u8_formats_and_subtitles(
1891 self, m3u8_url, video_id, ext=None, entry_protocol='m3u8',
1892 preference=None, quality=None, m3u8_id=None, note=None,
1893 errnote=None, fatal=True, live=False, data=None, headers={},
1896 res = self._download_webpage_handle(
1898 note=note or 'Downloading m3u8 information',
1899 errnote=errnote or 'Failed to download m3u8 information',
1900 fatal=fatal, data=data, headers=headers, query=query)
1905 m3u8_doc, urlh = res
1906 m3u8_url = urlh.geturl()
1908 return self._parse_m3u8_formats_and_subtitles(
1909 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
1910 preference=preference, quality=quality, m3u8_id=m3u8_id,
1911 note=note, errnote=errnote, fatal=fatal, live=live, data=data,
1912 headers=headers, query=query, video_id=video_id)
1914 def _parse_m3u8_formats_and_subtitles(
1915 self, m3u8_doc, m3u8_url, ext=None, entry_protocol='m3u8',
1916 preference=None, quality=None, m3u8_id=None, live=False, note=None,
1917 errnote=None, fatal=True, data=None, headers={}, query={},
1920 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
1923 if (not self._downloader.params.get('allow_unplayable_formats')
1924 and re.search(r'#EXT-X-SESSION-KEY:.*?URI="skd://', m3u8_doc)): # Apple FairPlay
1931 format_url = lambda u: (
1933 if re.match(r'^https?://', u)
1934 else compat_urlparse.urljoin(m3u8_url, u))
1936 split_discontinuity = self._downloader.params.get('hls_split_discontinuity', False)
1939 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
1940 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
1941 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
1943 # We should try extracting formats only from master playlists [1, 4.3.4],
1944 # i.e. playlists that describe available qualities. On the other hand
1945 # media playlists [1, 4.3.3] should be returned as is since they contain
1946 # just the media without qualities renditions.
1947 # Fortunately, master playlist can be easily distinguished from media
1948 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
1949 # master playlist tags MUST NOT appear in a media playlist and vice versa.
1950 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
1951 # media playlist and MUST NOT appear in master playlist thus we can
1952 # clearly detect media playlist with this criterion.
1954 def _extract_m3u8_playlist_formats(format_url=None, m3u8_doc=None, video_id=None,
1955 fatal=True, data=None, headers={}):
1959 res = self._download_webpage_handle(
1960 format_url, video_id,
1962 errnote='Failed to download m3u8 playlist information',
1963 fatal=fatal, data=data, headers=headers)
1968 m3u8_doc, urlh = res
1969 format_url = urlh.geturl()
1971 playlist_formats = []
1974 if split_discontinuity
1981 for line in m3u8_doc.splitlines():
1982 if not line.startswith('#'):
1983 format_info['files'].append(line)
1984 elif split_discontinuity and line.startswith('#EXT-X-DISCONTINUITY'):
1986 playlist_formats.append(format_info)
1992 playlist_formats.append(format_info)
1993 return playlist_formats
1995 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
1997 playlist_formats = _extract_m3u8_playlist_formats(m3u8_doc=m3u8_doc)
1999 for format in playlist_formats:
2002 format_id.append(m3u8_id)
2003 format_index = format.get('index')
2005 format_id.append(str(format_index))
2007 'format_id': '-'.join(format_id),
2008 'format_index': format_index,
2011 'protocol': entry_protocol,
2012 'preference': preference,
2017 return formats, subtitles
2020 last_stream_inf = {}
2022 def extract_media(x_media_line):
2023 media = parse_m3u8_attributes(x_media_line)
2024 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
2025 media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME')
2026 if not (media_type and group_id and name):
2028 groups.setdefault(group_id, []).append(media)
2029 # <https://tools.ietf.org/html/rfc8216#section-4.3.4.1>
2030 if media_type == 'SUBTITLES':
2031 lang = media['LANGUAGE'] # XXX: normalise?
2032 url = format_url(media['URI'])
2035 'ext': determine_ext(url),
2037 if sub_info['ext'] == 'm3u8':
2038 # Per RFC 8216 §3.1, the only possible subtitle format m3u8
2039 # files may contain is WebVTT:
2040 # <https://tools.ietf.org/html/rfc8216#section-3.1>
2041 sub_info['ext'] = 'vtt'
2042 sub_info['protocol'] = 'm3u8_native'
2043 subtitles.setdefault(lang, []).append(sub_info)
2044 if media_type not in ('VIDEO', 'AUDIO'):
2046 media_url = media.get('URI')
2048 manifest_url = format_url(media_url)
2050 playlist_formats = _extract_m3u8_playlist_formats(manifest_url, video_id=video_id,
2051 fatal=fatal, data=data, headers=headers)
2053 for format in playlist_formats:
2054 format_index = format.get('index')
2055 for v in (m3u8_id, group_id, name):
2059 format_id.append(str(format_index))
2061 'format_id': '-'.join(format_id),
2062 'format_index': format_index,
2063 'url': manifest_url,
2064 'manifest_url': m3u8_url,
2065 'language': media.get('LANGUAGE'),
2067 'protocol': entry_protocol,
2068 'preference': preference,
2071 if media_type == 'AUDIO':
2072 f['vcodec'] = 'none'
2075 def build_stream_name():
2076 # Despite specification does not mention NAME attribute for
2077 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
2078 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
2079 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
2080 stream_name = last_stream_inf.get('NAME')
2083 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
2084 # from corresponding rendition group
2085 stream_group_id = last_stream_inf.get('VIDEO')
2086 if not stream_group_id:
2088 stream_group = groups.get(stream_group_id)
2089 if not stream_group:
2090 return stream_group_id
2091 rendition = stream_group[0]
2092 return rendition.get('NAME') or stream_group_id
2094 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
2095 # chance to detect video only formats when EXT-X-STREAM-INF tags
2096 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
2097 for line in m3u8_doc.splitlines():
2098 if line.startswith('#EXT-X-MEDIA:'):
2101 for line in m3u8_doc.splitlines():
2102 if line.startswith('#EXT-X-STREAM-INF:'):
2103 last_stream_inf = parse_m3u8_attributes(line)
2104 elif line.startswith('#') or not line.strip():
2107 tbr = float_or_none(
2108 last_stream_inf.get('AVERAGE-BANDWIDTH')
2109 or last_stream_inf.get('BANDWIDTH'), scale=1000)
2110 manifest_url = format_url(line.strip())
2112 playlist_formats = _extract_m3u8_playlist_formats(manifest_url, video_id=video_id,
2113 fatal=fatal, data=data, headers=headers)
2115 for frmt in playlist_formats:
2118 format_id.append(m3u8_id)
2119 format_index = frmt.get('index')
2120 stream_name = build_stream_name()
2121 # Bandwidth of live streams may differ over time thus making
2122 # format_id unpredictable. So it's better to keep provided
2125 format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
2127 format_id.append(str(format_index))
2129 'format_id': '-'.join(format_id),
2130 'format_index': format_index,
2131 'url': manifest_url,
2132 'manifest_url': m3u8_url,
2135 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')),
2136 'protocol': entry_protocol,
2137 'preference': preference,
2140 resolution = last_stream_inf.get('RESOLUTION')
2142 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
2144 f['width'] = int(mobj.group('width'))
2145 f['height'] = int(mobj.group('height'))
2146 # Unified Streaming Platform
2148 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
2150 abr, vbr = mobj.groups()
2151 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
2156 codecs = parse_codecs(last_stream_inf.get('CODECS'))
2158 audio_group_id = last_stream_inf.get('AUDIO')
2159 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
2160 # references a rendition group MUST have a CODECS attribute.
2161 # However, this is not always respected, for example, [2]
2162 # contains EXT-X-STREAM-INF tag which references AUDIO
2163 # rendition group but does not have CODECS and despite
2164 # referencing an audio group it represents a complete
2165 # (with audio and video) format. So, for such cases we will
2166 # ignore references to rendition groups and treat them
2167 # as complete formats.
2168 if audio_group_id and codecs and f.get('vcodec') != 'none':
2169 audio_group = groups.get(audio_group_id)
2170 if audio_group and audio_group[0].get('URI'):
2171 # TODO: update acodec for audio only formats with
2173 f['acodec'] = 'none'
2174 if not f.get('ext'):
2175 f['ext'] = 'm4a' if f.get('vcodec') == 'none' else 'mp4'
2179 progressive_uri = last_stream_inf.get('PROGRESSIVE-URI')
2182 del http_f['manifest_url']
2184 'format_id': f['format_id'].replace('hls-', 'http-'),
2186 'url': progressive_uri,
2188 formats.append(http_f)
2190 last_stream_inf = {}
2191 return formats, subtitles
2194 def _xpath_ns(path, namespace=None):
2198 for c in path.split('/'):
2199 if not c or c == '.':
2202 out.append('{%s}%s' % (namespace, c))
2203 return '/'.join(out)
2205 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
2206 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
2212 namespace = self._parse_smil_namespace(smil)
2214 return self._parse_smil_formats(
2215 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
2217 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
2218 smil = self._download_smil(smil_url, video_id, fatal=fatal)
2221 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
2223 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
2224 return self._download_xml(
2225 smil_url, video_id, 'Downloading SMIL file',
2226 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
2228 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
2229 namespace = self._parse_smil_namespace(smil)
2231 formats = self._parse_smil_formats(
2232 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
2233 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
2235 video_id = os.path.splitext(url_basename(smil_url))[0]
2239 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
2240 name = meta.attrib.get('name')
2241 content = meta.attrib.get('content')
2242 if not name or not content:
2244 if not title and name == 'title':
2246 elif not description and name in ('description', 'abstract'):
2247 description = content
2248 elif not upload_date and name == 'date':
2249 upload_date = unified_strdate(content)
2252 'id': image.get('type'),
2253 'url': image.get('src'),
2254 'width': int_or_none(image.get('width')),
2255 'height': int_or_none(image.get('height')),
2256 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
2260 'title': title or video_id,
2261 'description': description,
2262 'upload_date': upload_date,
2263 'thumbnails': thumbnails,
2265 'subtitles': subtitles,
2268 def _parse_smil_namespace(self, smil):
2269 return self._search_regex(
2270 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
2272 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
2274 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
2275 b = meta.get('base') or meta.get('httpBase')
2286 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
2287 for medium in media:
2288 src = medium.get('src')
2289 if not src or src in srcs:
2293 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
2294 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
2295 width = int_or_none(medium.get('width'))
2296 height = int_or_none(medium.get('height'))
2297 proto = medium.get('proto')
2298 ext = medium.get('ext')
2299 src_ext = determine_ext(src)
2300 streamer = medium.get('streamer') or base
2302 if proto == 'rtmp' or streamer.startswith('rtmp'):
2308 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
2310 'filesize': filesize,
2314 if transform_rtmp_url:
2315 streamer, src = transform_rtmp_url(streamer, src)
2316 formats[-1].update({
2322 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
2323 src_url = src_url.strip()
2325 if proto == 'm3u8' or src_ext == 'm3u8':
2326 m3u8_formats = self._extract_m3u8_formats(
2327 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
2328 if len(m3u8_formats) == 1:
2330 m3u8_formats[0].update({
2331 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
2336 formats.extend(m3u8_formats)
2337 elif src_ext == 'f4m':
2342 'plugin': 'flowplayer-3.2.0.1',
2344 f4m_url += '&' if '?' in f4m_url else '?'
2345 f4m_url += compat_urllib_parse_urlencode(f4m_params)
2346 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
2347 elif src_ext == 'mpd':
2348 formats.extend(self._extract_mpd_formats(
2349 src_url, video_id, mpd_id='dash', fatal=False))
2350 elif re.search(r'\.ism/[Mm]anifest', src_url):
2351 formats.extend(self._extract_ism_formats(
2352 src_url, video_id, ism_id='mss', fatal=False))
2353 elif src_url.startswith('http') and self._is_valid_url(src, video_id):
2357 'ext': ext or src_ext or 'flv',
2358 'format_id': 'http-%d' % (bitrate or http_count),
2360 'filesize': filesize,
2367 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
2370 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
2371 src = textstream.get('src')
2372 if not src or src in urls:
2375 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
2376 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
2377 subtitles.setdefault(lang, []).append({
2383 def _extract_xspf_playlist(self, xspf_url, playlist_id, fatal=True):
2384 xspf = self._download_xml(
2385 xspf_url, playlist_id, 'Downloading xpsf playlist',
2386 'Unable to download xspf manifest', fatal=fatal)
2389 return self._parse_xspf(
2390 xspf, playlist_id, xspf_url=xspf_url,
2391 xspf_base_url=base_url(xspf_url))
2393 def _parse_xspf(self, xspf_doc, playlist_id, xspf_url=None, xspf_base_url=None):
2395 'xspf': 'http://xspf.org/ns/0/',
2396 's1': 'http://static.streamone.nl/player/ns/0',
2400 for track in xspf_doc.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
2402 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
2403 description = xpath_text(
2404 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
2405 thumbnail = xpath_text(
2406 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
2407 duration = float_or_none(
2408 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
2411 for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP)):
2412 format_url = urljoin(xspf_base_url, location.text)
2417 'manifest_url': xspf_url,
2418 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
2419 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
2420 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
2422 self._sort_formats(formats)
2427 'description': description,
2428 'thumbnail': thumbnail,
2429 'duration': duration,
2434 def _extract_mpd_formats(self, *args, **kwargs):
2435 fmts, subs = self._extract_mpd_formats_and_subtitles(*args, **kwargs)
2437 self.report_warning(bug_reports_message(
2438 "Ignoring subtitle tracks found in the DASH manifest; "
2439 "if any subtitle tracks are missing,"
2443 def _extract_mpd_formats_and_subtitles(
2444 self, mpd_url, video_id, mpd_id=None, note=None, errnote=None,
2445 fatal=True, data=None, headers={}, query={}):
2446 res = self._download_xml_handle(
2448 note=note or 'Downloading MPD manifest',
2449 errnote=errnote or 'Failed to download MPD manifest',
2450 fatal=fatal, data=data, headers=headers, query=query)
2456 mpd_base_url = base_url(urlh.geturl())
2458 return self._parse_mpd_formats_and_subtitles(
2459 mpd_doc, mpd_id, mpd_base_url, mpd_url)
2461 def _parse_mpd_formats(self, *args, **kwargs):
2462 fmts, subs = self._parse_mpd_formats_and_subtitles(*args, **kwargs)
2464 self.report_warning(bug_reports_message(
2465 "Ignoring subtitle tracks found in the DASH manifest; "
2466 "if any subtitle tracks are missing,"
2470 def _parse_mpd_formats_and_subtitles(
2471 self, mpd_doc, mpd_id=None, mpd_base_url='', mpd_url=None):
2473 Parse formats
from MPD manifest
.
2475 1. MPEG
-DASH Standard
, ISO
/IEC
23009-1:2014(E
),
2476 http
://standards
.iso
.org
/ittf
/PubliclyAvailableStandards
/c065274_ISO_IEC_23009
-1_2014.zip
2477 2. https
://en
.wikipedia
.org
/wiki
/Dynamic_Adaptive_Streaming_over_HTTP
2479 if not self._downloader.params.get('dynamic_mpd', True):
2480 if mpd_doc.get('type') == 'dynamic':
2483 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
2486 return self._xpath_ns(path, namespace)
2488 def is_drm_protected(element):
2489 return element.find(_add_ns('ContentProtection')) is not None
2491 def extract_multisegment_info(element, ms_parent_info):
2492 ms_info = ms_parent_info.copy()
2494 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2495 # common attributes and elements. We will only extract relevant
2497 def extract_common(source):
2498 segment_timeline = source.find(_add_ns('SegmentTimeline'))
2499 if segment_timeline is not None:
2500 s_e = segment_timeline.findall(_add_ns('S'))
2502 ms_info['total_number'] = 0
2505 r = int(s.get('r', 0))
2506 ms_info['total_number'] += 1 + r
2507 ms_info['s'].append({
2508 't': int(s.get('t', 0)),
2509 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2510 'd': int(s.attrib['d']),
2513 start_number = source.get('startNumber')
2515 ms_info['start_number'] = int(start_number)
2516 timescale = source.get('timescale')
2518 ms_info['timescale'] = int(timescale)
2519 segment_duration = source.get('duration')
2520 if segment_duration:
2521 ms_info['segment_duration'] = float(segment_duration)
2523 def extract_Initialization(source):
2524 initialization = source.find(_add_ns('Initialization'))
2525 if initialization is not None:
2526 ms_info['initialization_url'] = initialization.attrib['sourceURL']
2528 segment_list = element.find(_add_ns('SegmentList'))
2529 if segment_list is not None:
2530 extract_common(segment_list)
2531 extract_Initialization(segment_list)
2532 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
2534 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
2536 segment_template = element.find(_add_ns('SegmentTemplate'))
2537 if segment_template is not None:
2538 extract_common(segment_template)
2539 media = segment_template.get('media')
2541 ms_info['media'] = media
2542 initialization = segment_template.get('initialization')
2544 ms_info['initialization'] = initialization
2546 extract_Initialization(segment_template)
2549 skip_unplayable = not self._downloader.params.get('allow_unplayable_formats')
2551 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
2554 for period in mpd_doc.findall(_add_ns('Period')):
2555 period_duration = parse_duration(period.get('duration')) or mpd_duration
2556 period_ms_info = extract_multisegment_info(period, {
2560 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
2561 if skip_unplayable and is_drm_protected(adaptation_set):
2563 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
2564 for representation in adaptation_set.findall(_add_ns('Representation')):
2565 if skip_unplayable and is_drm_protected(representation):
2567 representation_attrib = adaptation_set.attrib.copy()
2568 representation_attrib.update(representation.attrib)
2569 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2570 mime_type = representation_attrib['mimeType']
2571 content_type = representation_attrib.get('contentType', mime_type.split('/')[0])
2573 if content_type in ('video', 'audio', 'text'):
2575 for element in (representation, adaptation_set, period, mpd_doc):
2576 base_url_e = element.find(_add_ns('BaseURL'))
2577 if base_url_e is not None:
2578 base_url = base_url_e.text + base_url
2579 if re.match(r'^https?://', base_url):
2581 if mpd_base_url and not re.match(r'^https?://', base_url):
2582 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
2584 base_url = mpd_base_url + base_url
2585 representation_id = representation_attrib.get('id')
2586 lang = representation_attrib.get('lang')
2587 url_el = representation.find(_add_ns('BaseURL'))
2588 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
2589 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
2590 if content_type in ('video', 'audio'):
2592 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
2593 'manifest_url': mpd_url,
2594 'ext': mimetype2ext(mime_type),
2595 'width': int_or_none(representation_attrib.get('width')),
2596 'height': int_or_none(representation_attrib.get('height')),
2597 'tbr': float_or_none(bandwidth, 1000),
2598 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
2599 'fps': int_or_none(representation_attrib.get('frameRate')),
2600 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
2601 'format_note': 'DASH %s' % content_type,
2602 'filesize': filesize,
2603 'container': mimetype2ext(mime_type) + '_dash',
2605 f.update(parse_codecs(representation_attrib.get('codecs')))
2606 elif content_type == 'text':
2608 'ext': mimetype2ext(mime_type),
2609 'manifest_url': mpd_url,
2610 'filesize': filesize,
2612 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
2614 def prepare_template(template_name, identifiers):
2615 tmpl = representation_ms_info[template_name]
2616 # First of, % characters outside $...$ templates
2617 # must be escaped by doubling for proper processing
2618 # by % operator string formatting used further (see
2619 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2625 in_template = not in_template
2626 elif c == '%' and not in_template:
2628 # Next, $...$ templates are translated to their
2629 # %(...) counterparts to be used with % operator
2630 t = t.replace('$RepresentationID$', representation_id)
2631 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
2632 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
2633 t.replace('$$', '$')
2636 # @initialization is a regular template like @media one
2637 # so it should be handled just the same way (see
2638 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2639 if 'initialization' in representation_ms_info:
2640 initialization_template = prepare_template(
2642 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2643 # $Time$ shall not be included for @initialization thus
2644 # only $Bandwidth$ remains
2646 representation_ms_info['initialization_url'] = initialization_template % {
2647 'Bandwidth': bandwidth,
2650 def location_key(location):
2651 return 'url' if re.match(r'^https?://', location) else 'path'
2653 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
2655 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2656 media_location_key = location_key(media_template)
2658 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2659 # can't be used at the same time
2660 if '%(Number' in media_template and 's' not in representation_ms_info:
2661 segment_duration = None
2662 if 'total_number' not in representation_ms_info and 'segment_duration' in representation_ms_info:
2663 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
2664 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
2665 representation_ms_info['fragments'] = [{
2666 media_location_key: media_template % {
2667 'Number': segment_number,
2668 'Bandwidth': bandwidth,
2670 'duration': segment_duration,
2671 } for segment_number in range(
2672 representation_ms_info['start_number'],
2673 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
2675 # $Number*$ or $Time$ in media template with S list available
2676 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2677 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2678 representation_ms_info['fragments'] = []
2681 segment_number = representation_ms_info['start_number']
2683 def add_segment_url():
2684 segment_url = media_template % {
2685 'Time': segment_time,
2686 'Bandwidth': bandwidth,
2687 'Number': segment_number,
2689 representation_ms_info['fragments'].append({
2690 media_location_key: segment_url,
2691 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
2694 for num, s in enumerate(representation_ms_info['s']):
2695 segment_time = s.get('t') or segment_time
2699 for r in range(s.get('r', 0)):
2700 segment_time += segment_d
2703 segment_time += segment_d
2704 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
2706 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2707 # or any YouTube dashsegments video
2710 timescale = representation_ms_info['timescale']
2711 for s in representation_ms_info['s']:
2712 duration = float_or_none(s['d'], timescale)
2713 for r in range(s.get('r', 0) + 1):
2714 segment_uri = representation_ms_info['segment_urls'][segment_index]
2716 location_key(segment_uri): segment_uri,
2717 'duration': duration,
2720 representation_ms_info['fragments'] = fragments
2721 elif 'segment_urls' in representation_ms_info:
2722 # Segment URLs with no SegmentTimeline
2723 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2724 # https://github.com/ytdl-org/youtube-dl/pull/14844
2726 segment_duration = float_or_none(
2727 representation_ms_info['segment_duration'],
2728 representation_ms_info['timescale']) if 'segment_duration' in representation_ms_info else None
2729 for segment_url in representation_ms_info['segment_urls']:
2731 location_key(segment_url): segment_url,
2733 if segment_duration:
2734 fragment['duration'] = segment_duration
2735 fragments.append(fragment)
2736 representation_ms_info['fragments'] = fragments
2737 # If there is a fragments key available then we correctly recognized fragmented media.
2738 # Otherwise we will assume unfragmented media with direct access. Technically, such
2739 # assumption is not necessarily correct since we may simply have no support for
2740 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2741 if 'fragments' in representation_ms_info:
2743 # NB: mpd_url may be empty when MPD manifest is parsed from a string
2744 'url': mpd_url or base_url,
2745 'fragment_base_url': base_url,
2747 'protocol': 'http_dash_segments',
2749 if 'initialization_url' in representation_ms_info:
2750 initialization_url = representation_ms_info['initialization_url']
2751 if not f.get('url'):
2752 f['url'] = initialization_url
2753 f['fragments'].append({location_key(initialization_url): initialization_url})
2754 f['fragments'].extend(representation_ms_info['fragments'])
2756 # Assuming direct URL to unfragmented media.
2758 if content_type in ('video', 'audio'):
2760 elif content_type == 'text':
2761 subtitles.setdefault(lang or 'und', []).append(f)
2763 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
2764 return formats, subtitles
2766 def _extract_ism_formats(self, *args, **kwargs):
2767 fmts, subs = self._extract_ism_formats_and_subtitles(*args, **kwargs)
2769 self.report_warning(bug_reports_message(
2770 "Ignoring subtitle tracks found in the ISM manifest; "
2771 "if any subtitle tracks are missing,"
2775 def _extract_ism_formats_and_subtitles(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
2776 res = self._download_xml_handle(
2778 note=note or 'Downloading ISM manifest',
2779 errnote=errnote or 'Failed to download ISM manifest',
2780 fatal=fatal, data=data, headers=headers, query=query)
2787 return self._parse_ism_formats_and_subtitles(ism_doc, urlh.geturl(), ism_id)
2789 def _parse_ism_formats_and_subtitles(self, ism_doc, ism_url, ism_id=None):
2791 Parse formats
from ISM manifest
.
2793 1. [MS
-SSTR
]: Smooth Streaming Protocol
,
2794 https
://msdn
.microsoft
.com
/en
-us
/library
/ff469518
.aspx
2796 if ism_doc.get('IsLive') == 'TRUE':
2798 if (not self._downloader.params.get('allow_unplayable_formats')
2799 and ism_doc.find('Protection') is not None):
2802 duration = int(ism_doc.attrib['Duration'])
2803 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
2807 for stream in ism_doc.findall('StreamIndex'):
2808 stream_type = stream.get('Type')
2809 if stream_type not in ('video', 'audio', 'text'):
2811 url_pattern = stream.attrib['Url']
2812 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
2813 stream_name = stream.get('Name')
2814 stream_language = stream.get('Language', 'und')
2815 for track in stream.findall('QualityLevel'):
2816 fourcc = track.get('FourCC', 'AACL' if track.get('AudioTag') == '255' else None)
2817 # TODO: add support for WVC1 and WMAP
2818 if fourcc not in ('H264', 'AVC1', 'AACL', 'TTML'):
2819 self.report_warning('%s is not a supported codec' % fourcc)
2821 tbr = int(track.attrib['Bitrate']) // 1000
2822 # [1] does not mention Width and Height attributes. However,
2823 # they're often present while MaxWidth and MaxHeight are
2824 # missing, so should be used as fallbacks
2825 width = int_or_none(track.get('MaxWidth') or track.get('Width'))
2826 height = int_or_none(track.get('MaxHeight') or track.get('Height'))
2827 sampling_rate = int_or_none(track.get('SamplingRate'))
2829 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
2830 track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
2836 stream_fragments = stream.findall('c')
2837 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
2838 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
2839 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
2840 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
2841 if not fragment_ctx['duration']:
2843 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
2845 next_fragment_time = duration
2846 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
2847 for _ in range(fragment_repeat):
2849 'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
2850 'duration': fragment_ctx['duration'] / stream_timescale,
2852 fragment_ctx['time'] += fragment_ctx['duration']
2856 format_id.append(ism_id)
2858 format_id.append(stream_name)
2859 format_id.append(compat_str(tbr))
2861 if stream_type == 'text':
2862 subtitles.setdefault(stream_language, []).append({
2866 'manifest_url': ism_url,
2867 'fragments': fragments,
2868 '_download_params': {
2869 'stream_type': stream_type,
2870 'duration': duration,
2871 'timescale': stream_timescale,
2873 'language': stream_language,
2874 'codec_private_data': track.get('CodecPrivateData'),
2877 elif stream_type in ('video', 'audio'):
2879 'format_id': '-'.join(format_id),
2881 'manifest_url': ism_url,
2882 'ext': 'ismv' if stream_type == 'video' else 'isma',
2886 'asr': sampling_rate,
2887 'vcodec': 'none' if stream_type == 'audio' else fourcc,
2888 'acodec': 'none' if stream_type == 'video' else fourcc,
2890 'fragments': fragments,
2891 '_download_params': {
2892 'stream_type': stream_type,
2893 'duration': duration,
2894 'timescale': stream_timescale,
2895 'width': width or 0,
2896 'height': height or 0,
2898 'language': stream_language,
2899 'codec_private_data': track.get('CodecPrivateData'),
2900 'sampling_rate': sampling_rate,
2901 'channels': int_or_none(track.get('Channels', 2)),
2902 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
2903 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
2906 return formats, subtitles
2908 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):
2909 def absolute_url(item_url):
2910 return urljoin(base_url, item_url)
2912 def parse_content_type(content_type):
2913 if not content_type:
2915 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
2917 mimetype, codecs = ctr.groups()
2918 f = parse_codecs(codecs)
2919 f['ext'] = mimetype2ext(mimetype)
2923 def _media_formats(src, cur_media_type, type_info={}):
2924 full_url = absolute_url(src)
2925 ext = type_info.get('ext') or determine_ext(full_url)
2927 is_plain_url = False
2928 formats = self._extract_m3u8_formats(
2929 full_url, video_id, ext='mp4',
2930 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
2931 preference=preference, quality=quality, fatal=False)
2933 is_plain_url = False
2934 formats = self._extract_mpd_formats(
2935 full_url, video_id, mpd_id=mpd_id, fatal=False)
2940 'vcodec': 'none' if cur_media_type == 'audio' else None,
2942 return is_plain_url, formats
2945 # amp-video and amp-audio are very similar to their HTML5 counterparts
2946 # so we wll include them right here (see
2947 # https://www.ampproject.org/docs/reference/components/amp-video)
2948 # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/
2949 _MEDIA_TAG_NAME_RE = r'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)'
2950 media_tags = [(media_tag, media_tag_name, media_type, '')
2951 for media_tag, media_tag_name, media_type
2952 in re.findall(r'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE, webpage)]
2953 media_tags.extend(re.findall(
2954 # We only allow video|audio followed by a whitespace or '>'.
2955 # Allowing more characters may end up in significant slow down (see
2956 # https://github.com/ytdl-org/youtube-dl/issues/11979, example URL:
2957 # http://www.porntrex.com/maps/videositemap.xml).
2958 r'(?s)(<(?P<tag>%s)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>' % _MEDIA_TAG_NAME_RE, webpage))
2959 for media_tag, _, media_type, media_content in media_tags:
2964 media_attributes = extract_attributes(media_tag)
2965 src = strip_or_none(media_attributes.get('src'))
2967 _, formats = _media_formats(src, media_type)
2968 media_info['formats'].extend(formats)
2969 media_info['thumbnail'] = absolute_url(media_attributes.get('poster'))
2971 for source_tag in re.findall(r'<source[^>]+>', media_content):
2972 s_attr = extract_attributes(source_tag)
2973 # data-video-src and data-src are non standard but seen
2974 # several times in the wild
2975 src = strip_or_none(dict_get(s_attr, ('src', 'data-video-src', 'data-src')))
2978 f = parse_content_type(s_attr.get('type'))
2979 is_plain_url, formats = _media_formats(src, media_type, f)
2981 # width, height, res, label and title attributes are
2982 # all not standard but seen several times in the wild
2985 for lbl in ('label', 'title')
2986 if str_or_none(s_attr.get(lbl))
2988 width = int_or_none(s_attr.get('width'))
2989 height = (int_or_none(s_attr.get('height'))
2990 or int_or_none(s_attr.get('res')))
2991 if not width or not height:
2993 resolution = parse_resolution(lbl)
2996 width = width or resolution.get('width')
2997 height = height or resolution.get('height')
2999 tbr = parse_bitrate(lbl)
3008 'format_id': s_attr.get('label') or s_attr.get('title'),
3010 f.update(formats[0])
3011 media_info['formats'].append(f)
3013 media_info['formats'].extend(formats)
3014 for track_tag in re.findall(r'<track[^>]+>', media_content):
3015 track_attributes = extract_attributes(track_tag)
3016 kind = track_attributes.get('kind')
3017 if not kind or kind in ('subtitles', 'captions'):
3018 src = strip_or_none(track_attributes.get('src'))
3021 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
3022 media_info['subtitles'].setdefault(lang, []).append({
3023 'url': absolute_url(src),
3025 for f in media_info['formats']:
3026 f.setdefault('http_headers', {})['Referer'] = base_url
3027 if media_info['formats'] or media_info['subtitles']:
3028 entries.append(media_info)
3031 def _extract_akamai_formats(self, *args, **kwargs):
3032 fmts, subs = self._extract_akamai_formats_and_subtitles(*args, **kwargs)
3034 self.report_warning(bug_reports_message(
3035 "Ignoring subtitle tracks found in the manifests; "
3036 "if any subtitle tracks are missing,"
3040 def _extract_akamai_formats_and_subtitles(self, manifest_url, video_id, hosts={}):
3041 signed = 'hdnea=' in manifest_url
3043 # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
3044 manifest_url = re.sub(
3045 r'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
3046 '', manifest_url).strip('?')
3051 hdcore_sign = 'hdcore=3.7.0'
3052 f4m_url = re.sub(r'(https?://[^/]+)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
3053 hds_host = hosts.get('hds')
3055 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
3056 if 'hdcore=' not in f4m_url:
3057 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
3058 f4m_formats = self._extract_f4m_formats(
3059 f4m_url, video_id, f4m_id='hds', fatal=False)
3060 for entry in f4m_formats:
3061 entry.update({'extra_param_to_segment_url': hdcore_sign})
3062 formats.extend(f4m_formats)
3064 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
3065 hls_host = hosts.get('hls')
3067 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
3068 m3u8_formats, m3u8_subtitles = self._extract_m3u8_formats_and_subtitles(
3069 m3u8_url, video_id, 'mp4', 'm3u8_native',
3070 m3u8_id='hls', fatal=False)
3071 formats.extend(m3u8_formats)
3072 subtitles = self._merge_subtitles(subtitles, m3u8_subtitles)
3074 http_host = hosts.get('http')
3075 if http_host and m3u8_formats and not signed:
3076 REPL_REGEX = r'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
3077 qualities = re.match(REPL_REGEX, m3u8_url).group(2).split(',')
3078 qualities_length = len(qualities)
3079 if len(m3u8_formats) in (qualities_length, qualities_length + 1):
3081 for f in m3u8_formats:
3082 if f['vcodec'] != 'none':
3083 for protocol in ('http', 'https'):
3085 del http_f['manifest_url']
3087 REPL_REGEX, protocol + r'://%s/\g<1>%s\3' % (http_host, qualities[i]), f['url'])
3089 'format_id': http_f['format_id'].replace('hls-', protocol + '-'),
3091 'protocol': protocol,
3093 formats.append(http_f)
3096 return formats, subtitles
3098 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
3099 query = compat_urlparse.urlparse(url).query
3100 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
3102 r'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url)
3103 url_base = mobj.group('url')
3104 http_base_url = '%s%s:%s' % ('http', mobj.group('s') or '', url_base)
3107 def manifest_url(manifest):
3108 m_url = '%s/%s' % (http_base_url, manifest)
3110 m_url += '?%s' % query
3113 if 'm3u8' not in skip_protocols:
3114 formats.extend(self._extract_m3u8_formats(
3115 manifest_url('playlist.m3u8'), video_id, 'mp4',
3116 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
3117 if 'f4m' not in skip_protocols:
3118 formats.extend(self._extract_f4m_formats(
3119 manifest_url('manifest.f4m'),
3120 video_id, f4m_id='hds', fatal=False))
3121 if 'dash' not in skip_protocols:
3122 formats.extend(self._extract_mpd_formats(
3123 manifest_url('manifest.mpd'),
3124 video_id, mpd_id='dash', fatal=False))
3125 if re.search(r'(?:/smil:|\.smil)', url_base):
3126 if 'smil' not in skip_protocols:
3127 rtmp_formats = self._extract_smil_formats(
3128 manifest_url('jwplayer.smil'),
3129 video_id, fatal=False)
3130 for rtmp_format in rtmp_formats:
3131 rtsp_format = rtmp_format.copy()
3132 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
3133 del rtsp_format['play_path']
3134 del rtsp_format['ext']
3135 rtsp_format.update({
3136 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
3137 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
3140 formats.extend([rtmp_format, rtsp_format])
3142 for protocol in ('rtmp', 'rtsp'):
3143 if protocol not in skip_protocols:
3145 'url': '%s:%s' % (protocol, url_base),
3146 'format_id': protocol,
3147 'protocol': protocol,
3151 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
3153 r'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)(?!</script>).*?\.setup\s*\((?P<options>[^)]+)\)',
3157 jwplayer_data = self._parse_json(mobj.group('options'),
3159 transform_source=transform_source)
3160 except ExtractorError:
3163 if isinstance(jwplayer_data, dict):
3164 return jwplayer_data
3166 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
3167 jwplayer_data = self._find_jwplayer_data(
3168 webpage, video_id, transform_source=js_to_json)
3169 return self._parse_jwplayer_data(
3170 jwplayer_data, video_id, *args, **kwargs)
3172 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
3173 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3174 # JWPlayer backward compatibility: flattened playlists
3175 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
3176 if 'playlist' not in jwplayer_data:
3177 jwplayer_data = {'playlist': [jwplayer_data]}
3181 # JWPlayer backward compatibility: single playlist item
3182 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
3183 if not isinstance(jwplayer_data['playlist'], list):
3184 jwplayer_data['playlist'] = [jwplayer_data['playlist']]
3186 for video_data in jwplayer_data['playlist']:
3187 # JWPlayer backward compatibility: flattened sources
3188 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
3189 if 'sources' not in video_data:
3190 video_data['sources'] = [video_data]
3192 this_video_id = video_id or video_data['mediaid']
3194 formats = self._parse_jwplayer_formats(
3195 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
3196 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
3199 tracks = video_data.get('tracks')
3200 if tracks and isinstance(tracks, list):
3201 for track in tracks:
3202 if not isinstance(track, dict):
3204 track_kind = track.get('kind')
3205 if not track_kind or not isinstance(track_kind, compat_str):
3207 if track_kind.lower() not in ('captions', 'subtitles'):
3209 track_url = urljoin(base_url, track.get('file'))
3212 subtitles.setdefault(track.get('label') or 'en', []).append({
3213 'url': self._proto_relative_url(track_url)
3217 'id': this_video_id,
3218 'title': unescapeHTML(video_data['title'] if require_title else video_data.get('title')),
3219 'description': clean_html(video_data.get('description')),
3220 'thumbnail': urljoin(base_url, self._proto_relative_url(video_data.get('image'))),
3221 'timestamp': int_or_none(video_data.get('pubdate')),
3222 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
3223 'subtitles': subtitles,
3225 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
3226 if len(formats) == 1 and re.search(r'^(?:http|//).*(?:youtube\.com|youtu\.be)/.+', formats[0]['url']):
3228 '_type': 'url_transparent',
3229 'url': formats[0]['url'],
3232 self._sort_formats(formats)
3233 entry['formats'] = formats
3234 entries.append(entry)
3235 if len(entries) == 1:
3238 return self.playlist_result(entries)
3240 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
3241 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3244 for source in jwplayer_sources_data:
3245 if not isinstance(source, dict):
3247 source_url = urljoin(
3248 base_url, self._proto_relative_url(source.get('file')))
3249 if not source_url or source_url in urls:
3251 urls.append(source_url)
3252 source_type = source.get('type') or ''
3253 ext = mimetype2ext(source_type) or determine_ext(source_url)
3254 if source_type == 'hls' or ext == 'm3u8':
3255 formats.extend(self._extract_m3u8_formats(
3256 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
3257 m3u8_id=m3u8_id, fatal=False))
3258 elif source_type == 'dash' or ext == 'mpd':
3259 formats.extend(self._extract_mpd_formats(
3260 source_url, video_id, mpd_id=mpd_id, fatal=False))
3262 formats.extend(self._extract_smil_formats(
3263 source_url, video_id, fatal=False))
3264 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3265 elif source_type.startswith('audio') or ext in (
3266 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
3273 height = int_or_none(source.get('height'))
3275 # Often no height is provided but there is a label in
3276 # format like "1080p", "720p SD", or 1080.
3277 height = int_or_none(self._search_regex(
3278 r'^(\d{3,4})[pP]?(?:\b|$)', compat_str(source.get('label') or ''),
3279 'height', default=None))
3282 'width': int_or_none(source.get('width')),
3284 'tbr': int_or_none(source.get('bitrate')),
3287 if source_url.startswith('rtmp'):
3288 a_format['ext'] = 'flv'
3289 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3290 # of jwplayer.flash.swf
3291 rtmp_url_parts = re.split(
3292 r'((?:mp4|mp3|flv):)', source_url, 1)
3293 if len(rtmp_url_parts) == 3:
3294 rtmp_url, prefix, play_path = rtmp_url_parts
3297 'play_path': prefix + play_path,
3300 a_format.update(rtmp_params)
3301 formats.append(a_format)
3304 def _live_title(self, name):
3305 """ Generate the title
for a live video
"""
3306 now = datetime.datetime.now()
3307 now_str = now.strftime('%Y-%m-%d %H:%M')
3308 return name + ' ' + now_str
3310 def _int(self, v, name, fatal=False, **kwargs):
3311 res = int_or_none(v, **kwargs)
3312 if 'get_attr' in kwargs:
3313 print(getattr(v, kwargs['get_attr']))
3315 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
3317 raise ExtractorError(msg)
3319 self.report_warning(msg)
3322 def _float(self, v, name, fatal=False, **kwargs):
3323 res = float_or_none(v, **kwargs)
3325 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
3327 raise ExtractorError(msg)
3329 self.report_warning(msg)
3332 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
3333 path='/', secure=False, discard=False, rest={}, **kwargs):
3334 cookie = compat_cookiejar_Cookie(
3335 0, name, value, port, port is not None, domain, True,
3336 domain.startswith('.'), path, True, secure, expire_time,
3337 discard, None, None, rest)
3338 self._downloader.cookiejar.set_cookie(cookie)
3340 def _get_cookies(self, url):
3341 """ Return a compat_cookies_SimpleCookie
with the cookies
for the url
"""
3342 req = sanitized_Request(url)
3343 self._downloader.cookiejar.add_cookie_header(req)
3344 return compat_cookies_SimpleCookie(req.get_header('Cookie'))
3346 def _apply_first_set_cookie_header(self, url_handle, cookie):
3348 Apply first Set
-Cookie header instead of the last
. Experimental
.
3350 Some
sites (e
.g
. [1-3]) may serve two cookies under the same name
3351 in Set
-Cookie header
and expect the
first (old
) one to be
set rather
3352 than
second (new
). However
, as of RFC6265 the newer one cookie
3353 should be
set into cookie store what actually happens
.
3354 We will workaround this issue by resetting the cookie to
3355 the first one manually
.
3356 1. https
://new
.vk
.com
/
3357 2. https
://github
.com
/ytdl
-org
/youtube
-dl
/issues
/9841#issuecomment-227871201
3358 3. https
://learning
.oreilly
.com
/
3360 for header, cookies in url_handle.headers.items():
3361 if header.lower() != 'set-cookie':
3363 if sys.version_info[0] >= 3:
3364 cookies = cookies.encode('iso-8859-1')
3365 cookies = cookies.decode('utf-8')
3366 cookie_value = re.search(
3367 r'%s=(.+?);.*?\b[Dd]omain=(.+?)(?:[,;]|$)' % cookie, cookies)
3369 value, domain = cookie_value.groups()
3370 self._set_cookie(domain, cookie, value)
3373 def get_testcases(self, include_onlymatching=False):
3374 t = getattr(self, '_TEST', None)
3376 assert not hasattr(self, '_TESTS'), \
3377 '%s has _TEST and _TESTS' % type(self).__name__
3380 tests = getattr(self, '_TESTS', [])
3382 if not include_onlymatching and t.get('only_matching', False):
3384 t['name'] = type(self).__name__[:-len('IE')]
3387 def is_suitable(self, age_limit):
3388 """ Test whether the extractor
is generally suitable
for the given
3389 age
limit (i
.e
. pornographic sites are
not, all others usually are
) """
3391 any_restricted = False
3392 for tc in self.get_testcases(include_onlymatching=False):
3393 if tc.get('playlist', []):
3394 tc = tc['playlist'][0]
3395 is_restricted = age_restricted(
3396 tc.get('info_dict', {}).get('age_limit'), age_limit)
3397 if not is_restricted:
3399 any_restricted = any_restricted or is_restricted
3400 return not any_restricted
3402 def extract_subtitles(self, *args, **kwargs):
3403 if (self._downloader.params.get('writesubtitles', False)
3404 or self._downloader.params.get('listsubtitles')):
3405 return self._get_subtitles(*args, **kwargs)
3408 def _get_subtitles(self, *args, **kwargs):
3409 raise NotImplementedError('This method must be implemented by subclasses')
3412 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
3413 """ Merge subtitle items
for one language
. Items
with duplicated URLs
3414 will be dropped
. """
3415 list1_urls = set([item['url'] for item in subtitle_list1])
3416 ret = list(subtitle_list1)
3417 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
3421 def _merge_subtitles(cls, *dicts, **kwargs):
3422 """ Merge subtitle dictionaries
, language by language
. """
3424 target = (lambda target=None: target)(**kwargs)
3425 # The above lambda extracts the keyword argument 'target' from kwargs
3426 # while ensuring there are no stray ones. When Python 2 support
3427 # is dropped, remove it and change the function signature to:
3429 # def _merge_subtitles(cls, *dicts, target=None):
3434 for lang, subs in d.items():
3435 target[lang] = cls._merge_subtitle_items(target.get(lang, []), subs)
3438 def extract_automatic_captions(self, *args, **kwargs):
3439 if (self._downloader.params.get('writeautomaticsub', False)
3440 or self._downloader.params.get('listsubtitles')):
3441 return self._get_automatic_captions(*args, **kwargs)
3444 def _get_automatic_captions(self, *args, **kwargs):
3445 raise NotImplementedError('This method must be implemented by subclasses')
3447 def mark_watched(self, *args, **kwargs):
3448 if (self._downloader.params.get('mark_watched', False)
3449 and (self._get_login_info()[0] is not None
3450 or self._downloader.params.get('cookiefile') is not None)):
3451 self._mark_watched(*args, **kwargs)
3453 def _mark_watched(self, *args, **kwargs):
3454 raise NotImplementedError('This method must be implemented by subclasses')
3456 def geo_verification_headers(self):
3458 geo_verification_proxy = self._downloader.params.get('geo_verification_proxy')
3459 if geo_verification_proxy:
3460 headers['Ytdl-request-proxy'] = geo_verification_proxy
3463 def _generic_id(self, url):
3464 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
3466 def _generic_title(self, url):
3467 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
3470 def _availability(is_private, needs_premium, needs_subscription, needs_auth, is_unlisted):
3471 all_known = all(map(
3472 lambda x: x is not None,
3473 (is_private, needs_premium, needs_subscription, needs_auth, is_unlisted)))
3475 'private' if is_private
3476 else 'premium_only' if needs_premium
3477 else 'subscriber_only' if needs_subscription
3478 else 'needs_auth' if needs_auth
3479 else 'unlisted' if is_unlisted
3480 else 'public' if all_known
3484 class SearchInfoExtractor(InfoExtractor):
3486 Base
class for paged search queries extractors
.
3487 They accept URLs
in the format
_SEARCH_KEY(|all|
[0-9]):{query}
3488 Instances should define _SEARCH_KEY
and _MAX_RESULTS
.
3492 def _make_valid_url(cls):
3493 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
3496 def suitable(cls, url):
3497 return re.match(cls._make_valid_url(), url) is not None
3499 def _real_extract(self, query):
3500 mobj = re.match(self._make_valid_url(), query)
3502 raise ExtractorError('Invalid search query "%s"' % query)
3504 prefix = mobj.group('prefix')
3505 query = mobj.group('query')
3507 return self._get_n_results(query, 1)
3508 elif prefix == 'all':
3509 return self._get_n_results(query, self._MAX_RESULTS)
3513 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
3514 elif n > self._MAX_RESULTS:
3515 self.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3516 n = self._MAX_RESULTS
3517 return self._get_n_results(query, n)
3519 def _get_n_results(self, query, n):
3520 """Get a specified number of results
for a query
"""
3521 raise NotImplementedError('This method must be implemented by subclasses')
3524 def SEARCH_KEY(self):
3525 return self._SEARCH_KEY