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
,
66 parse_m3u8_attributes
,
90 class InfoExtractor(object):
91 """Information Extractor class.
93 Information extractors are the classes that, given a URL, extract
94 information about the video (or videos) the URL refers to. This
95 information includes the real video URL, the video title, author and
96 others. The information is stored in a dictionary which is then
97 passed to the YoutubeDL. The YoutubeDL processes this
98 information possibly downloading the video to the file system, among
99 other possible outcomes.
101 The type field determines the type of the result.
102 By far the most common value (and the default if _type is missing) is
103 "video", which indicates a single video.
105 For a video, the dictionaries must include the following fields:
107 id: Video identifier.
108 title: Video title, unescaped.
110 Additionally, it must contain either a formats entry or a url one:
112 formats: A list of dictionaries for each format available, ordered
113 from worst to best quality.
116 * url The mandatory URL representing the media:
117 for plain file media - HTTP URL of this file,
119 for HLS - URL of the M3U8 media playlist,
120 for HDS - URL of the F4M manifest,
122 - HTTP URL to plain file media (in case of
124 - URL of the MPD manifest or base URL
125 representing the media if MPD manifest
126 is parsed from a string (in case of
128 for MSS - URL of the ISM manifest.
130 The URL of the manifest file in case of
132 for HLS - URL of the M3U8 master playlist,
133 for HDS - URL of the F4M manifest,
134 for DASH - URL of the MPD manifest,
135 for MSS - URL of the ISM manifest.
136 * ext Will be calculated from URL if missing
137 * format A human-readable description of the format
138 ("mp4 container with h264/opus").
139 Calculated from the format_id, width, height.
140 and format_note fields if missing.
141 * format_id A short description of the format
142 ("mp4_h264_opus" or "19").
143 Technically optional, but strongly recommended.
144 * format_note Additional info about the format
145 ("3D" or "DASH video")
146 * width Width of the video, if known
147 * height Height of the video, if known
148 * resolution Textual description of width and height
149 * tbr Average bitrate of audio and video in KBit/s
150 * abr Average audio bitrate in KBit/s
151 * acodec Name of the audio codec in use
152 * asr Audio sampling rate in Hertz
153 * vbr Average video bitrate in KBit/s
155 * vcodec Name of the video codec in use
156 * container Name of the container format
157 * filesize The number of bytes, if known in advance
158 * filesize_approx An estimate for the number of bytes
159 * player_url SWF Player URL (used for rtmpdump).
160 * protocol The protocol that will be used for the actual
161 download, lower-case.
162 "http", "https", "rtsp", "rtmp", "rtmp_ffmpeg", "rtmpe",
163 "m3u8", "m3u8_native" or "http_dash_segments".
165 Base URL for fragments. Each fragment's path
166 value (if present) will be relative to
168 * fragments A list of fragments of a fragmented media.
169 Each fragment entry must contain either an url
170 or a path. If an url is present it should be
171 considered by a client. Otherwise both path and
172 fragment_base_url must be present. Here is
173 the list of all potential fields:
174 * "url" - fragment's URL
175 * "path" - fragment's path relative to
177 * "duration" (optional, int or float)
178 * "filesize" (optional, int)
179 * preference Order number of this format. If this field is
180 present and not None, the formats get sorted
181 by this field, regardless of all other values.
182 -1 for default (order by other properties),
183 -2 or smaller for less than default.
184 < -1000 to hide the format (if there is
185 another one which is strictly better)
186 * language Language code, e.g. "de" or "en-US".
187 * language_preference Is this in the language mentioned in
189 10 if it's what the URL is about,
190 -1 for default (don't know),
191 -10 otherwise, other values reserved for now.
192 * quality Order number of the video quality of this
193 format, irrespective of the file format.
194 -1 for default (order by other properties),
195 -2 or smaller for less than default.
196 * source_preference Order number for this video source
197 (quality takes higher priority)
198 -1 for default (order by other properties),
199 -2 or smaller for less than default.
200 * http_headers A dictionary of additional HTTP headers
201 to add to the request.
202 * stretched_ratio If given and not 1, indicates that the
203 video's pixels are not square.
204 width : height ratio as float.
205 * no_resume The server does not support resuming the
206 (HTTP or RTMP) download. Boolean.
207 * has_drm The format has DRM and cannot be downloaded. Boolean
208 * downloader_options A dictionary of downloader options as
209 described in FileDownloader
210 RTMP formats can also have the additional fields: page_url,
211 app, play_path, tc_url, flash_version, rtmp_live, rtmp_conn,
212 rtmp_protocol, rtmp_real_time
214 url: Final video URL.
215 ext: Video filename extension.
216 format: The video format, defaults to ext (used for --get-format)
217 player_url: SWF Player URL (used for rtmpdump).
219 The following fields are optional:
221 alt_title: A secondary title of the video.
222 display_id An alternative identifier for the video, not necessarily
223 unique, but available before title. Typically, id is
224 something like "4234987", title "Dancing naked mole rats",
225 and display_id "dancing-naked-mole-rats"
226 thumbnails: A list of dictionaries, with the following entries:
227 * "id" (optional, string) - Thumbnail format ID
229 * "preference" (optional, int) - quality of the image
230 * "width" (optional, int)
231 * "height" (optional, int)
232 * "resolution" (optional, string "{width}x{height}",
234 * "filesize" (optional, int)
235 * "_test_url" (optional, bool) - If true, test the URL
236 thumbnail: Full URL to a video thumbnail image.
237 description: Full video description.
238 uploader: Full name of the video uploader.
239 license: License name the video is licensed under.
240 creator: The creator of the video.
241 release_timestamp: UNIX timestamp of the moment the video was released.
242 release_date: The date (YYYYMMDD) when the video was released.
243 timestamp: UNIX timestamp of the moment the video was uploaded
244 upload_date: Video upload date (YYYYMMDD).
245 If not explicitly set, calculated from timestamp.
246 uploader_id: Nickname or id of the video uploader.
247 uploader_url: Full URL to a personal webpage of the video uploader.
248 channel: Full name of the channel the video is uploaded on.
249 Note that channel fields may or may not repeat uploader
250 fields. This depends on a particular extractor.
251 channel_id: Id of the channel.
252 channel_url: Full URL to a channel webpage.
253 location: Physical location where the video was filmed.
254 subtitles: The available subtitles as a dictionary in the format
255 {tag: subformats}. "tag" is usually a language code, and
256 "subformats" is a list sorted from lower to higher
257 preference, each element is a dictionary with the "ext"
259 * "data": The subtitles file contents
260 * "url": A URL pointing to the subtitles file
261 It can optionally also have:
262 * "name": Name or description of the subtitles
263 "ext" will be calculated from URL if missing
264 automatic_captions: Like 'subtitles'; contains automatically generated
265 captions instead of normal subtitles
266 duration: Length of the video in seconds, as an integer or float.
267 view_count: How many users have watched the video on the platform.
268 like_count: Number of positive ratings of the video
269 dislike_count: Number of negative ratings of the video
270 repost_count: Number of reposts of the video
271 average_rating: Average rating give by users, the scale used depends on the webpage
272 comment_count: Number of comments on the video
273 comments: A list of comments, each with one or more of the following
274 properties (all but one of text or html optional):
275 * "author" - human-readable name of the comment author
276 * "author_id" - user ID of the comment author
277 * "author_thumbnail" - The thumbnail of the comment author
279 * "html" - Comment as HTML
280 * "text" - Plain text of the comment
281 * "timestamp" - UNIX timestamp of comment
282 * "parent" - ID of the comment this one is replying to.
283 Set to "root" to indicate that this is a
284 comment to the original video.
285 * "like_count" - Number of positive ratings of the comment
286 * "dislike_count" - Number of negative ratings of the comment
287 * "is_favorited" - Whether the comment is marked as
288 favorite by the video uploader
289 * "author_is_uploader" - Whether the comment is made by
291 age_limit: Age restriction for the video, as an integer (years)
292 webpage_url: The URL to the video webpage, if given to yt-dlp it
293 should allow to get the same result again. (It will be set
294 by YoutubeDL if it's missing)
295 categories: A list of categories that the video falls in, for example
297 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
298 cast: A list of the video cast
299 is_live: True, False, or None (=unknown). Whether this video is a
300 live stream that goes on instead of a fixed-length video.
301 was_live: True, False, or None (=unknown). Whether this video was
302 originally a live stream.
303 live_status: 'is_live', 'is_upcoming', 'was_live', 'not_live' or None (=unknown)
304 If absent, automatically set from is_live, was_live
305 start_time: Time in seconds where the reproduction should start, as
306 specified in the URL.
307 end_time: Time in seconds where the reproduction should end, as
308 specified in the URL.
309 chapters: A list of dictionaries, with the following entries:
310 * "start_time" - The start time of the chapter in seconds
311 * "end_time" - The end time of the chapter in seconds
312 * "title" (optional, string)
313 playable_in_embed: Whether this video is allowed to play in embedded
314 players on other sites. Can be True (=always allowed),
315 False (=never allowed), None (=unknown), or a string
316 specifying the criteria for embedability (Eg: 'whitelist')
317 availability: Under what condition the video is available. One of
318 'private', 'premium_only', 'subscriber_only', 'needs_auth',
319 'unlisted' or 'public'. Use 'InfoExtractor._availability'
321 __post_extractor: A function to be called just before the metadata is
322 written to either disk, logger or console. The function
323 must return a dict which will be added to the info_dict.
324 This is usefull for additional information that is
325 time-consuming to extract. Note that the fields thus
326 extracted will not be available to output template and
327 match_filter. So, only "comments" and "comment_count" are
328 currently allowed to be extracted via this method.
330 The following fields should only be used when the video belongs to some logical
333 chapter: Name or title of the chapter the video belongs to.
334 chapter_number: Number of the chapter the video belongs to, as an integer.
335 chapter_id: Id of the chapter the video belongs to, as a unicode string.
337 The following fields should only be used when the video is an episode of some
338 series, programme or podcast:
340 series: Title of the series or programme the video episode belongs to.
341 season: Title of the season the video episode belongs to.
342 season_number: Number of the season the video episode belongs to, as an integer.
343 season_id: Id of the season the video episode belongs to, as a unicode string.
344 episode: Title of the video episode. Unlike mandatory video title field,
345 this field should denote the exact title of the video episode
346 without any kind of decoration.
347 episode_number: Number of the video episode within a season, as an integer.
348 episode_id: Id of the video episode, as a unicode string.
350 The following fields should only be used when the media is a track or a part of
353 track: Title of the track.
354 track_number: Number of the track within an album or a disc, as an integer.
355 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
357 artist: Artist(s) of the track.
358 genre: Genre(s) of the track.
359 album: Title of the album the track belongs to.
360 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
361 album_artist: List of all artists appeared on the album (e.g.
362 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
364 disc_number: Number of the disc or other physical medium the track belongs to,
366 release_year: Year (YYYY) when the album was released.
368 Unless mentioned otherwise, the fields should be Unicode strings.
370 Unless mentioned otherwise, None is equivalent to absence of information.
373 _type "playlist" indicates multiple videos.
374 There must be a key "entries", which is a list, an iterable, or a PagedList
375 object, each element of which is a valid dictionary by this specification.
377 Additionally, playlists can have "id", "title", and any other relevent
378 attributes with the same semantics as videos (see above).
381 _type "multi_video" indicates that there are multiple videos that
382 form a single show, for examples multiple acts of an opera or TV episode.
383 It must have an entries key like a playlist and contain all the keys
384 required for a video at the same time.
387 _type "url" indicates that the video must be extracted from another
388 location, possibly by a different extractor. Its only required key is:
389 "url" - the next URL to extract.
390 The key "ie_key" can be set to the class name (minus the trailing "IE",
391 e.g. "Youtube") if the extractor class is known in advance.
392 Additionally, the dictionary may have any properties of the resolved entity
393 known in advance, for example "title" if the title of the referred video is
397 _type "url_transparent" entities have the same specification as "url", but
398 indicate that the given additional information is more precise than the one
399 associated with the resolved URL.
400 This is useful when a site employs a video service that hosts the video and
401 its technical metadata, but that video service does not embed a useful
402 title, description etc.
405 Subclasses of this one should re-define the _real_initialize() and
406 _real_extract() methods and define a _VALID_URL regexp.
407 Probably, they should also be added to the list of extractors.
409 _GEO_BYPASS attribute may be set to False in order to disable
410 geo restriction bypass mechanisms for a particular extractor.
411 Though it won't disable explicit geo restriction bypass based on
412 country code provided with geo_bypass_country.
414 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
415 countries for this extractor. One of these countries will be used by
416 geo restriction bypass mechanism right away in order to bypass
417 geo restriction, of course, if the mechanism is not disabled.
419 _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
420 IP blocks in CIDR notation for this extractor. One of these IP blocks
421 will be used by geo restriction bypass mechanism similarly
424 Finally, the _WORKING attribute should be set to False for broken IEs
425 in order to warn the users and skip the tests.
430 _x_forwarded_for_ip
= None
432 _GEO_COUNTRIES
= None
433 _GEO_IP_BLOCKS
= None
437 'any': 'Use --cookies, --username and --password or --netrc to provide account credentials',
439 'Use --cookies for the authentication. '
440 'See https://github.com/ytdl-org/youtube-dl#how-do-i-pass-cookies-to-youtube-dl for how to pass cookies'),
441 'password': 'Use --username and --password or --netrc to provide account credentials',
444 def __init__(self
, downloader
=None):
445 """Constructor. Receives an optional downloader."""
447 self
._x
_forwarded
_for
_ip
= None
448 self
._printed
_messages
= set()
449 self
.set_downloader(downloader
)
452 def _match_valid_url(cls
, url
):
453 # This does not use has/getattr intentionally - we want to know whether
454 # we have cached the regexp for *this* class, whereas getattr would also
455 # match the superclass
456 if '_VALID_URL_RE' not in cls
.__dict
__:
457 cls
._VALID
_URL
_RE
= re
.compile(cls
._VALID
_URL
)
458 return cls
._VALID
_URL
_RE
.match(url
)
461 def suitable(cls
, url
):
462 """Receives a URL and returns True if suitable for this IE."""
463 # This function must import everything it needs (except other extractors),
464 # so that lazy_extractors works correctly
465 return cls
._match
_valid
_url
(url
) is not None
468 def _match_id(cls
, url
):
469 return cls
._match
_valid
_url
(url
).group('id')
472 def get_temp_id(cls
, url
):
474 return cls
._match
_id
(url
)
475 except (IndexError, AttributeError):
480 """Getter method for _WORKING."""
483 def initialize(self
):
484 """Initializes an instance (authentication, etc)."""
485 self
._printed
_messages
= set()
486 self
._initialize
_geo
_bypass
({
487 'countries': self
._GEO
_COUNTRIES
,
488 'ip_blocks': self
._GEO
_IP
_BLOCKS
,
491 self
._real
_initialize
()
494 def _initialize_geo_bypass(self
, geo_bypass_context
):
496 Initialize geo restriction bypass mechanism.
498 This method is used to initialize geo bypass mechanism based on faking
499 X-Forwarded-For HTTP header. A random country from provided country list
500 is selected and a random IP belonging to this country is generated. This
501 IP will be passed as X-Forwarded-For HTTP header in all subsequent
504 This method will be used for initial geo bypass mechanism initialization
505 during the instance initialization with _GEO_COUNTRIES and
508 You may also manually call it from extractor's code if geo bypass
509 information is not available beforehand (e.g. obtained during
510 extraction) or due to some other reason. In this case you should pass
511 this information in geo bypass context passed as first argument. It may
512 contain following fields:
514 countries: List of geo unrestricted countries (similar
516 ip_blocks: List of geo unrestricted IP blocks in CIDR notation
517 (similar to _GEO_IP_BLOCKS)
520 if not self
._x
_forwarded
_for
_ip
:
522 # Geo bypass mechanism is explicitly disabled by user
523 if not self
.get_param('geo_bypass', True):
526 if not geo_bypass_context
:
527 geo_bypass_context
= {}
529 # Backward compatibility: previously _initialize_geo_bypass
530 # expected a list of countries, some 3rd party code may still use
532 if isinstance(geo_bypass_context
, (list, tuple)):
533 geo_bypass_context
= {
534 'countries': geo_bypass_context
,
537 # The whole point of geo bypass mechanism is to fake IP
538 # as X-Forwarded-For HTTP header based on some IP block or
541 # Path 1: bypassing based on IP block in CIDR notation
543 # Explicit IP block specified by user, use it right away
544 # regardless of whether extractor is geo bypassable or not
545 ip_block
= self
.get_param('geo_bypass_ip_block', None)
547 # Otherwise use random IP block from geo bypass context but only
548 # if extractor is known as geo bypassable
550 ip_blocks
= geo_bypass_context
.get('ip_blocks')
551 if self
._GEO
_BYPASS
and ip_blocks
:
552 ip_block
= random
.choice(ip_blocks
)
555 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(ip_block
)
556 self
._downloader
.write_debug(
557 '[debug] Using fake IP %s as X-Forwarded-For' % self
._x
_forwarded
_for
_ip
)
560 # Path 2: bypassing based on country code
562 # Explicit country code specified by user, use it right away
563 # regardless of whether extractor is geo bypassable or not
564 country
= self
.get_param('geo_bypass_country', None)
566 # Otherwise use random country code from geo bypass context but
567 # only if extractor is known as geo bypassable
569 countries
= geo_bypass_context
.get('countries')
570 if self
._GEO
_BYPASS
and countries
:
571 country
= random
.choice(countries
)
574 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country
)
575 self
._downloader
.write_debug(
576 'Using fake IP %s (%s) as X-Forwarded-For' % (self
._x
_forwarded
_for
_ip
, country
.upper()))
578 def extract(self
, url
):
579 """Extracts URL information and returns it in list of dicts."""
584 self
.write_debug('Extracting URL: %s' % url
)
585 ie_result
= self
._real
_extract
(url
)
586 if ie_result
is None:
588 if self
._x
_forwarded
_for
_ip
:
589 ie_result
['__x_forwarded_for_ip'] = self
._x
_forwarded
_for
_ip
590 subtitles
= ie_result
.get('subtitles')
591 if (subtitles
and 'live_chat' in subtitles
592 and 'no-live-chat' in self
.get_param('compat_opts', [])):
593 del subtitles
['live_chat']
595 except GeoRestrictedError
as e
:
596 if self
.__maybe
_fake
_ip
_and
_retry
(e
.countries
):
599 except ExtractorError
as e
:
600 video_id
= e
.video_id
or self
.get_temp_id(url
)
601 raise ExtractorError(
602 e
.msg
, video_id
=video_id
, ie
=self
.IE_NAME
, tb
=e
.traceback
, expected
=e
.expected
, cause
=e
.cause
)
603 except compat_http_client
.IncompleteRead
as e
:
604 raise ExtractorError('A network error has occurred.', cause
=e
, expected
=True, video_id
=self
.get_temp_id(url
))
605 except (KeyError, StopIteration) as e
:
606 raise ExtractorError('An extractor error has occurred.', cause
=e
, video_id
=self
.get_temp_id(url
))
608 def __maybe_fake_ip_and_retry(self
, countries
):
609 if (not self
.get_param('geo_bypass_country', None)
611 and self
.get_param('geo_bypass', True)
612 and not self
._x
_forwarded
_for
_ip
614 country_code
= random
.choice(countries
)
615 self
._x
_forwarded
_for
_ip
= GeoUtils
.random_ipv4(country_code
)
616 if self
._x
_forwarded
_for
_ip
:
618 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
619 % (self
._x
_forwarded
_for
_ip
, country_code
.upper()))
623 def set_downloader(self
, downloader
):
624 """Sets the downloader for this IE."""
625 self
._downloader
= downloader
627 def _real_initialize(self
):
628 """Real initialization process. Redefine in subclasses."""
631 def _real_extract(self
, url
):
632 """Real extraction process. Redefine in subclasses."""
637 """A string for getting the InfoExtractor with get_info_extractor"""
638 return cls
.__name
__[:-2]
642 return compat_str(type(self
).__name
__[:-2])
645 def __can_accept_status_code(err
, expected_status
):
646 assert isinstance(err
, compat_urllib_error
.HTTPError
)
647 if expected_status
is None:
649 elif callable(expected_status
):
650 return expected_status(err
.code
) is True
652 return err
.code
in variadic(expected_status
)
654 def _request_webpage(self
, url_or_request
, video_id
, note
=None, errnote
=None, fatal
=True, data
=None, headers
={}, query={}
, expected_status
=None):
656 Return the response handle.
658 See _download_webpage docstring for arguments specification.
660 if not self
._downloader
._first
_webpage
_request
:
661 sleep_interval
= float_or_none(self
.get_param('sleep_interval_requests')) or 0
662 if sleep_interval
> 0:
663 self
.to_screen('Sleeping %s seconds ...' % sleep_interval
)
664 time
.sleep(sleep_interval
)
666 self
._downloader
._first
_webpage
_request
= False
669 self
.report_download_webpage(video_id
)
670 elif note
is not False:
672 self
.to_screen('%s' % (note
,))
674 self
.to_screen('%s: %s' % (video_id
, note
))
676 # Some sites check X-Forwarded-For HTTP header in order to figure out
677 # the origin of the client behind proxy. This allows bypassing geo
678 # restriction by faking this header's value to IP that belongs to some
679 # geo unrestricted country. We will do so once we encounter any
680 # geo restriction error.
681 if self
._x
_forwarded
_for
_ip
:
682 if 'X-Forwarded-For' not in headers
:
683 headers
['X-Forwarded-For'] = self
._x
_forwarded
_for
_ip
685 if isinstance(url_or_request
, compat_urllib_request
.Request
):
686 url_or_request
= update_Request(
687 url_or_request
, data
=data
, headers
=headers
, query
=query
)
690 url_or_request
= update_url_query(url_or_request
, query
)
691 if data
is not None or headers
:
692 url_or_request
= sanitized_Request(url_or_request
, data
, headers
)
694 return self
._downloader
.urlopen(url_or_request
)
695 except network_exceptions
as err
:
696 if isinstance(err
, compat_urllib_error
.HTTPError
):
697 if self
.__can
_accept
_status
_code
(err
, expected_status
):
698 # Retain reference to error to prevent file object from
699 # being closed before it can be read. Works around the
700 # effects of <https://bugs.python.org/issue15002>
701 # introduced in Python 3.4.1.
708 errnote
= 'Unable to download webpage'
710 errmsg
= '%s: %s' % (errnote
, error_to_compat_str(err
))
712 raise ExtractorError(errmsg
, sys
.exc_info()[2], cause
=err
)
714 self
.report_warning(errmsg
)
717 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):
719 Return a tuple (page content as string, URL handle).
721 See _download_webpage docstring for arguments specification.
723 # Strip hashes from the URL (#1038)
724 if isinstance(url_or_request
, (compat_str
, str)):
725 url_or_request
= url_or_request
.partition('#')[0]
727 urlh
= self
._request
_webpage
(url_or_request
, video_id
, note
, errnote
, fatal
, data
=data
, headers
=headers
, query
=query
, expected_status
=expected_status
)
731 content
= self
._webpage
_read
_content
(urlh
, url_or_request
, video_id
, note
, errnote
, fatal
, encoding
=encoding
)
732 return (content
, urlh
)
735 def _guess_encoding_from_content(content_type
, webpage_bytes
):
736 m
= re
.match(r
'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type
)
738 encoding
= m
.group(1)
740 m
= re
.search(br
'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
741 webpage_bytes[:1024])
743 encoding = m.group(1).decode('ascii')
744 elif webpage_bytes.startswith(b'\xff\xfe'):
751 def __check_blocked(self, content):
752 first_block = content[:512]
753 if ('<title>Access to this site is blocked</title>' in content
754 and 'Websense' in first_block):
755 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
756 blocked_iframe = self._html_search_regex(
757 r'<iframe src="([^
"]+)"', content,
758 'Websense information URL
', default=None)
760 msg += ' Visit
%s for more details
' % blocked_iframe
761 raise ExtractorError(msg, expected=True)
762 if '<title
>The URL you requested has been blocked
</title
>' in first_block:
764 'Access to this webpage has been blocked by Indian censorship
. '
765 'Use a VPN
or proxy
server (with --proxy
) to route around it
.')
766 block_msg = self._html_search_regex(
767 r'</h1
><p
>(.*?
)</p
>',
768 content, 'block message
', default=None)
770 msg += ' (Message
: "%s")' % block_msg.replace('\n', ' ')
771 raise ExtractorError(msg, expected=True)
772 if ('<title
>TTK
:: Доступ к ресурсу ограничен
</title
>' in content
773 and 'blocklist
.rkn
.gov
.ru
' in content):
774 raise ExtractorError(
775 'Access to this webpage has been blocked by decision of the Russian government
. '
776 'Visit http
://blocklist
.rkn
.gov
.ru
/ for a block reason
.',
779 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
780 content_type = urlh.headers.get('Content
-Type
', '')
781 webpage_bytes = urlh.read()
782 if prefix is not None:
783 webpage_bytes = prefix + webpage_bytes
785 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
786 if self.get_param('dump_intermediate_pages
', False):
787 self.to_screen('Dumping request to
' + urlh.geturl())
788 dump = base64.b64encode(webpage_bytes).decode('ascii
')
789 self._downloader.to_screen(dump)
790 if self.get_param('write_pages
', False):
791 basen = '%s_%s' % (video_id, urlh.geturl())
793 h = '___
' + hashlib.md5(basen.encode('utf
-8')).hexdigest()
794 basen = basen[:240 - len(h)] + h
795 raw_filename = basen + '.dump
'
796 filename = sanitize_filename(raw_filename, restricted=True)
797 self.to_screen('Saving request to
' + filename)
798 # Working around MAX_PATH limitation on Windows (see
799 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
800 if compat_os_name == 'nt
':
801 absfilepath = os.path.abspath(filename)
802 if len(absfilepath) > 259:
803 filename = '\\\\?
\\' + absfilepath
804 with open(filename, 'wb
') as outf:
805 outf.write(webpage_bytes)
808 content = webpage_bytes.decode(encoding, 'replace
')
810 content = webpage_bytes.decode('utf
-8', 'replace
')
812 self.__check_blocked(content)
816 def _download_webpage(
817 self, url_or_request, video_id, note=None, errnote=None,
818 fatal=True, tries=1, timeout=5, encoding=None, data=None,
819 headers={}, query={}, expected_status=None):
821 Return the data of the page as a string.
824 url_or_request -- plain text URL as a string or
825 a compat_urllib_request.Requestobject
826 video_id -- Video/playlist/item identifier (string)
829 note -- note printed before downloading (string)
830 errnote -- note printed in case of an error (string)
831 fatal -- flag denoting whether error should be considered fatal,
832 i.e. whether it should cause ExtractionError to be raised,
833 otherwise a warning will be reported and extraction continued
834 tries -- number of tries
835 timeout -- sleep interval between tries
836 encoding -- encoding for a page content decoding, guessed automatically
837 when not explicitly specified
838 data -- POST data (bytes)
839 headers -- HTTP headers (dict)
840 query -- URL query (dict)
841 expected_status -- allows to accept failed HTTP requests (non 2xx
842 status code) by explicitly specifying a set of accepted status
843 codes. Can be any of the following entities:
844 - an integer type specifying an exact failed status code to
846 - a list or a tuple of integer types specifying a list of
847 failed status codes to accept
848 - a callable accepting an actual failed status code and
849 returning True if it should be accepted
850 Note that this argument does not affect success status codes (2xx)
851 which are always accepted.
856 while success is False:
858 res = self._download_webpage_handle(
859 url_or_request, video_id, note, errnote, fatal,
860 encoding=encoding, data=data, headers=headers, query=query,
861 expected_status=expected_status)
863 except compat_http_client.IncompleteRead as e:
865 if try_count >= tries:
867 self._sleep(timeout, video_id)
874 def _download_xml_handle(
875 self, url_or_request, video_id, note='Downloading XML
',
876 errnote='Unable to download XML
', transform_source=None,
877 fatal=True, encoding=None, data=None, headers={}, query={},
878 expected_status=None):
880 Return a tuple (xml as an compat_etree_Element, URL handle).
882 See _download_webpage docstring for arguments specification.
884 res = self._download_webpage_handle(
885 url_or_request, video_id, note, errnote, fatal=fatal,
886 encoding=encoding, data=data, headers=headers, query=query,
887 expected_status=expected_status)
890 xml_string, urlh = res
891 return self._parse_xml(
892 xml_string, video_id, transform_source=transform_source,
896 self, url_or_request, video_id,
897 note='Downloading XML
', errnote='Unable to download XML
',
898 transform_source=None, fatal=True, encoding=None,
899 data=None, headers={}, query={}, expected_status=None):
901 Return the xml as an compat_etree_Element.
903 See _download_webpage docstring for arguments specification.
905 res = self._download_xml_handle(
906 url_or_request, video_id, note=note, errnote=errnote,
907 transform_source=transform_source, fatal=fatal, encoding=encoding,
908 data=data, headers=headers, query=query,
909 expected_status=expected_status)
910 return res if res is False else res[0]
912 def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
914 xml_string = transform_source(xml_string)
916 return compat_etree_fromstring(xml_string.encode('utf
-8'))
917 except compat_xml_parse_error as ve:
918 errmsg = '%s: Failed to parse XML
' % video_id
920 raise ExtractorError(errmsg, cause=ve)
922 self.report_warning(errmsg + str(ve))
924 def _download_json_handle(
925 self, url_or_request, video_id, note='Downloading JSON metadata
',
926 errnote='Unable to download JSON metadata
', transform_source=None,
927 fatal=True, encoding=None, data=None, headers={}, query={},
928 expected_status=None):
930 Return a tuple (JSON object, URL handle).
932 See _download_webpage docstring for arguments specification.
934 res = self._download_webpage_handle(
935 url_or_request, video_id, note, errnote, fatal=fatal,
936 encoding=encoding, data=data, headers=headers, query=query,
937 expected_status=expected_status)
940 json_string, urlh = res
941 return self._parse_json(
942 json_string, video_id, transform_source=transform_source,
946 self, url_or_request, video_id, note='Downloading JSON metadata
',
947 errnote='Unable to download JSON metadata
', transform_source=None,
948 fatal=True, encoding=None, data=None, headers={}, query={},
949 expected_status=None):
951 Return the JSON object as a dict.
953 See _download_webpage docstring for arguments specification.
955 res = self._download_json_handle(
956 url_or_request, video_id, note=note, errnote=errnote,
957 transform_source=transform_source, fatal=fatal, encoding=encoding,
958 data=data, headers=headers, query=query,
959 expected_status=expected_status)
960 return res if res is False else res[0]
962 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
964 json_string = transform_source(json_string)
966 return json.loads(json_string)
967 except ValueError as ve:
968 errmsg = '%s: Failed to parse JSON
' % video_id
970 raise ExtractorError(errmsg, cause=ve)
972 self.report_warning(errmsg + str(ve))
974 def _parse_socket_response_as_json(self, data, video_id, transform_source=None, fatal=True):
975 return self._parse_json(
976 data[data.find('{'):data.rfind('}
') + 1],
977 video_id, transform_source, fatal)
979 def _download_socket_json_handle(
980 self, url_or_request, video_id, note='Polling socket
',
981 errnote='Unable to poll socket
', transform_source=None,
982 fatal=True, encoding=None, data=None, headers={}, query={},
983 expected_status=None):
985 Return a tuple (JSON object, URL handle).
987 See _download_webpage docstring for arguments specification.
989 res = self._download_webpage_handle(
990 url_or_request, video_id, note, errnote, fatal=fatal,
991 encoding=encoding, data=data, headers=headers, query=query,
992 expected_status=expected_status)
996 return self._parse_socket_response_as_json(
997 webpage, video_id, transform_source=transform_source,
1000 def _download_socket_json(
1001 self, url_or_request, video_id, note='Polling socket
',
1002 errnote='Unable to poll socket
', transform_source=None,
1003 fatal=True, encoding=None, data=None, headers={}, query={},
1004 expected_status=None):
1006 Return the JSON object as a dict.
1008 See _download_webpage docstring for arguments specification.
1010 res = self._download_socket_json_handle(
1011 url_or_request, video_id, note=note, errnote=errnote,
1012 transform_source=transform_source, fatal=fatal, encoding=encoding,
1013 data=data, headers=headers, query=query,
1014 expected_status=expected_status)
1015 return res if res is False else res[0]
1017 def report_warning(self, msg, video_id=None, *args, only_once=False, **kwargs):
1018 idstr = format_field(video_id, template='%s: ')
1019 msg = f'[{self.IE_NAME}
] {idstr}{msg}
'
1021 if f'WARNING
: {msg}
' in self._printed_messages:
1023 self._printed_messages.add(f'WARNING
: {msg}
')
1024 self._downloader.report_warning(msg, *args, **kwargs)
1026 def to_screen(self, msg, *args, **kwargs):
1027 """Print msg to screen, prefixing it with '[ie_name
]'"""
1028 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg), *args, **kwargs)
1030 def write_debug(self, msg, *args, **kwargs):
1031 self._downloader.write_debug('[%s] %s' % (self.IE_NAME, msg), *args, **kwargs)
1033 def get_param(self, name, default=None, *args, **kwargs):
1034 if self._downloader:
1035 return self._downloader.params.get(name, default, *args, **kwargs)
1038 def report_drm(self, video_id, partial=False):
1039 self.raise_no_formats('This video
is DRM protected
', expected=True, video_id=video_id)
1041 def report_extraction(self, id_or_name):
1042 """Report information extraction."""
1043 self.to_screen('%s: Extracting information
' % id_or_name)
1045 def report_download_webpage(self, video_id):
1046 """Report webpage download."""
1047 self.to_screen('%s: Downloading webpage
' % video_id)
1049 def report_age_confirmation(self):
1050 """Report attempt to confirm age."""
1051 self.to_screen('Confirming age
')
1053 def report_login(self):
1054 """Report attempt to log in."""
1055 self.to_screen('Logging
in')
1057 def raise_login_required(
1058 self, msg='This video
is only available
for registered users
',
1059 metadata_available=False, method='any
'):
1060 if metadata_available and self.get_param('ignore_no_formats_error
'):
1061 self.report_warning(msg)
1062 if method is not None:
1063 msg = '%s. %s' % (msg, self._LOGIN_HINTS[method])
1064 raise ExtractorError(msg, expected=True)
1066 def raise_geo_restricted(
1067 self, msg='This video
is not available
from your location due to geo restriction
',
1068 countries=None, metadata_available=False):
1069 if metadata_available and self.get_param('ignore_no_formats_error
'):
1070 self.report_warning(msg)
1072 raise GeoRestrictedError(msg, countries=countries)
1074 def raise_no_formats(self, msg, expected=False, video_id=None):
1075 if expected and self.get_param('ignore_no_formats_error
'):
1076 self.report_warning(msg, video_id)
1077 elif isinstance(msg, ExtractorError):
1080 raise ExtractorError(msg, expected=expected, video_id=video_id)
1082 # Methods for following #608
1084 def url_result(url, ie=None, video_id=None, video_title=None):
1085 """Returns a URL that points to a page that should be processed"""
1086 # TODO: ie should be the class used for getting the info
1087 video_info = {'_type
': 'url
',
1090 if video_id is not None:
1091 video_info['id'] = video_id
1092 if video_title is not None:
1093 video_info['title
'] = video_title
1096 def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None):
1098 self.url_result(self._proto_relative_url(getter(m) if getter else m), ie)
1100 return self.playlist_result(
1101 urls, playlist_id=playlist_id, playlist_title=playlist_title)
1104 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None, **kwargs):
1105 """Returns a playlist"""
1106 video_info = {'_type
': 'playlist
',
1108 video_info.update(kwargs)
1110 video_info['id'] = playlist_id
1112 video_info['title
'] = playlist_title
1113 if playlist_description is not None:
1114 video_info['description
'] = playlist_description
1117 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1119 Perform a regex search on the given string, using a single or a list of
1120 patterns returning the first matching group.
1121 In case of failure return a default value or raise a WARNING or a
1122 RegexNotFoundError, depending on fatal, specifying the field name.
1124 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
1125 mobj = re.search(pattern, string, flags)
1128 mobj = re.search(p, string, flags)
1132 if not self.get_param('no_color
') and compat_os_name != 'nt
' and sys.stderr.isatty():
1133 _name = '\033[0;34m
%s\033[0m
' % name
1139 # return the first matching group
1140 return next(g for g in mobj.groups() if g is not None)
1141 elif isinstance(group, (list, tuple)):
1142 return tuple(mobj.group(g) for g in group)
1144 return mobj.group(group)
1145 elif default is not NO_DEFAULT:
1148 raise RegexNotFoundError('Unable to extract
%s' % _name)
1150 self.report_warning('unable to extract
%s' % _name + bug_reports_message())
1153 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
1155 Like _search_regex, but strips HTML tags and unescapes entities.
1157 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
1159 return clean_html(res).strip()
1163 def _get_netrc_login_info(self, netrc_machine=None):
1166 netrc_machine = netrc_machine or self._NETRC_MACHINE
1168 if self.get_param('usenetrc
', False):
1170 netrc_file = compat_expanduser(self.get_param('netrc_location
') or '~
')
1171 if os.path.isdir(netrc_file):
1172 netrc_file = os.path.join(netrc_file, '.netrc
')
1173 info = netrc.netrc(file=netrc_file).authenticators(netrc_machine)
1174 if info is not None:
1178 raise netrc.NetrcParseError(
1179 'No authenticators
for %s' % netrc_machine)
1180 except (IOError, netrc.NetrcParseError) as err:
1181 self.report_warning(
1182 'parsing
.netrc
: %s' % error_to_compat_str(err))
1184 return username, password
1186 def _get_login_info(self, username_option='username
', password_option='password
', netrc_machine=None):
1188 Get the login info as (username, password)
1189 First look for the manually specified credentials using username_option
1190 and password_option as keys in params dictionary. If no such credentials
1191 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
1193 If there's no info available
, return (None, None)
1196 # Attempt to use provided username and password or .netrc data
1197 username = self.get_param(username_option)
1198 if username is not None:
1199 password = self.get_param(password_option)
1201 username, password = self._get_netrc_login_info(netrc_machine)
1203 return username, password
1205 def _get_tfa_info(self, note='two-factor verification code'):
1207 Get the two
-factor authentication info
1208 TODO
- asking the user will be required
for sms
/phone verify
1209 currently just uses the command line option
1210 If there
's no info available, return None
1213 tfa = self.get_param('twofactor
')
1217 return compat_getpass('Type
%s and press
[Return
]: ' % note)
1219 # Helper functions for extracting OpenGraph info
1221 def _og_regexes(prop):
1222 content_re = r'content
=(?
:"([^"]+?
)"|\'([^\']+?)\'|\s*([^\s"\'=<>`
]+?
))'
1223 property_re = (r'(?
:name|
property)=(?
:\'og
[:-]%(prop)s\'|
"og[:-]%(prop)s"|\s
*og
[:-]%(prop)s\b)'
1224 % {'prop': re.escape(prop)})
1225 template = r'<meta
[^
>]+?
%s[^
>]+?
%s'
1227 template % (property_re, content_re),
1228 template % (content_re, property_re),
1232 def _meta_regex(prop):
1233 return r'''(?isx)<meta
1234 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
1235 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
1237 def _og_search_property(self, prop, html, name=None, **kargs):
1238 prop = variadic(prop)
1240 name = 'OpenGraph
%s' % prop[0]
1243 og_regexes.extend(self._og_regexes(p))
1244 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
1247 return unescapeHTML(escaped)
1249 def _og_search_thumbnail(self, html, **kargs):
1250 return self._og_search_property('image
', html, 'thumbnail URL
', fatal=False, **kargs)
1252 def _og_search_description(self, html, **kargs):
1253 return self._og_search_property('description
', html, fatal=False, **kargs)
1255 def _og_search_title(self, html, **kargs):
1256 return self._og_search_property('title
', html, **kargs)
1258 def _og_search_video_url(self, html, name='video url
', secure=True, **kargs):
1259 regexes = self._og_regexes('video
') + self._og_regexes('video
:url
')
1261 regexes = self._og_regexes('video
:secure_url
') + regexes
1262 return self._html_search_regex(regexes, html, name, **kargs)
1264 def _og_search_url(self, html, **kargs):
1265 return self._og_search_property('url
', html, **kargs)
1267 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
1268 name = variadic(name)
1269 if display_name is None:
1270 display_name = name[0]
1271 return self._html_search_regex(
1272 [self._meta_regex(n) for n in name],
1273 html, display_name, fatal=fatal, group='content
', **kwargs)
1275 def _dc_search_uploader(self, html):
1276 return self._html_search_meta('dc
.creator
', html, 'uploader
')
1278 def _rta_search(self, html):
1279 # See http://www.rtalabel.org/index.php?content=howtofaq#single
1280 if re.search(r'(?ix
)<meta\s
+name
="rating"\s
+'
1281 r' content
="RTA-5042-1996-1400-1577-RTA"',
1286 def _media_rating_search(self, html):
1287 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
1288 rating = self._html_search_meta('rating
', html)
1300 return RATING_TABLE.get(rating.lower())
1302 def _family_friendly_search(self, html):
1303 # See http://schema.org/VideoObject
1304 family_friendly = self._html_search_meta(
1305 'isFamilyFriendly
', html, default=None)
1307 if not family_friendly:
1316 return RATING_TABLE.get(family_friendly.lower())
1318 def _twitter_search_player(self, html):
1319 return self._html_search_meta('twitter
:player
', html,
1320 'twitter card player
')
1322 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
1323 json_ld_list = list(re.finditer(JSON_LD_RE, html))
1324 default = kwargs.get('default
', NO_DEFAULT)
1325 # JSON-LD may be malformed and thus `fatal` should be respected.
1326 # At the same time `default` may be passed that assumes `fatal=False`
1327 # for _search_regex. Let's simulate the same behavior here
as well
.
1328 fatal
= kwargs
.get('fatal', True) if default
is NO_DEFAULT
else False
1330 for mobj
in json_ld_list
:
1331 json_ld_item
= self
._parse
_json
(
1332 mobj
.group('json_ld'), video_id
, fatal
=fatal
)
1333 if not json_ld_item
:
1335 if isinstance(json_ld_item
, dict):
1336 json_ld
.append(json_ld_item
)
1337 elif isinstance(json_ld_item
, (list, tuple)):
1338 json_ld
.extend(json_ld_item
)
1340 json_ld
= self
._json
_ld
(json_ld
, video_id
, fatal
=fatal
, expected_type
=expected_type
)
1343 if default
is not NO_DEFAULT
:
1346 raise RegexNotFoundError('Unable to extract JSON-LD')
1348 self
.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
1351 def _json_ld(self
, json_ld
, video_id
, fatal
=True, expected_type
=None):
1352 if isinstance(json_ld
, compat_str
):
1353 json_ld
= self
._parse
_json
(json_ld
, video_id
, fatal
=fatal
)
1357 if not isinstance(json_ld
, (list, tuple, dict)):
1359 if isinstance(json_ld
, dict):
1362 INTERACTION_TYPE_MAP
= {
1363 'CommentAction': 'comment',
1364 'AgreeAction': 'like',
1365 'DisagreeAction': 'dislike',
1366 'LikeAction': 'like',
1367 'DislikeAction': 'dislike',
1368 'ListenAction': 'view',
1369 'WatchAction': 'view',
1370 'ViewAction': 'view',
1373 def extract_interaction_type(e
):
1374 interaction_type
= e
.get('interactionType')
1375 if isinstance(interaction_type
, dict):
1376 interaction_type
= interaction_type
.get('@type')
1377 return str_or_none(interaction_type
)
1379 def extract_interaction_statistic(e
):
1380 interaction_statistic
= e
.get('interactionStatistic')
1381 if isinstance(interaction_statistic
, dict):
1382 interaction_statistic
= [interaction_statistic
]
1383 if not isinstance(interaction_statistic
, list):
1385 for is_e
in interaction_statistic
:
1386 if not isinstance(is_e
, dict):
1388 if is_e
.get('@type') != 'InteractionCounter':
1390 interaction_type
= extract_interaction_type(is_e
)
1391 if not interaction_type
:
1393 # For interaction count some sites provide string instead of
1394 # an integer (as per spec) with non digit characters (e.g. ",")
1395 # so extracting count with more relaxed str_to_int
1396 interaction_count
= str_to_int(is_e
.get('userInteractionCount'))
1397 if interaction_count
is None:
1399 count_kind
= INTERACTION_TYPE_MAP
.get(interaction_type
.split('/')[-1])
1402 count_key
= '%s_count' % count_kind
1403 if info
.get(count_key
) is not None:
1405 info
[count_key
] = interaction_count
1407 def extract_video_object(e
):
1408 assert e
['@type'] == 'VideoObject'
1409 author
= e
.get('author')
1411 'url': url_or_none(e
.get('contentUrl')),
1412 'title': unescapeHTML(e
.get('name')),
1413 'description': unescapeHTML(e
.get('description')),
1414 'thumbnail': url_or_none(e
.get('thumbnailUrl') or e
.get('thumbnailURL')),
1415 'duration': parse_duration(e
.get('duration')),
1416 'timestamp': unified_timestamp(e
.get('uploadDate')),
1417 # author can be an instance of 'Organization' or 'Person' types.
1418 # both types can have 'name' property(inherited from 'Thing' type). [1]
1419 # however some websites are using 'Text' type instead.
1420 # 1. https://schema.org/VideoObject
1421 'uploader': author
.get('name') if isinstance(author
, dict) else author
if isinstance(author
, compat_str
) else None,
1422 'filesize': float_or_none(e
.get('contentSize')),
1423 'tbr': int_or_none(e
.get('bitrate')),
1424 'width': int_or_none(e
.get('width')),
1425 'height': int_or_none(e
.get('height')),
1426 'view_count': int_or_none(e
.get('interactionCount')),
1428 extract_interaction_statistic(e
)
1432 item_type
= e
.get('@type')
1433 if expected_type
is not None and expected_type
!= item_type
:
1435 if item_type
in ('TVEpisode', 'Episode'):
1436 episode_name
= unescapeHTML(e
.get('name'))
1438 'episode': episode_name
,
1439 'episode_number': int_or_none(e
.get('episodeNumber')),
1440 'description': unescapeHTML(e
.get('description')),
1442 if not info
.get('title') and episode_name
:
1443 info
['title'] = episode_name
1444 part_of_season
= e
.get('partOfSeason')
1445 if isinstance(part_of_season
, dict) and part_of_season
.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
1447 'season': unescapeHTML(part_of_season
.get('name')),
1448 'season_number': int_or_none(part_of_season
.get('seasonNumber')),
1450 part_of_series
= e
.get('partOfSeries') or e
.get('partOfTVSeries')
1451 if isinstance(part_of_series
, dict) and part_of_series
.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
1452 info
['series'] = unescapeHTML(part_of_series
.get('name'))
1453 elif item_type
== 'Movie':
1455 'title': unescapeHTML(e
.get('name')),
1456 'description': unescapeHTML(e
.get('description')),
1457 'duration': parse_duration(e
.get('duration')),
1458 'timestamp': unified_timestamp(e
.get('dateCreated')),
1460 elif item_type
in ('Article', 'NewsArticle'):
1462 'timestamp': parse_iso8601(e
.get('datePublished')),
1463 'title': unescapeHTML(e
.get('headline')),
1464 'description': unescapeHTML(e
.get('articleBody')),
1466 elif item_type
== 'VideoObject':
1467 extract_video_object(e
)
1468 if expected_type
is None:
1472 video
= e
.get('video')
1473 if isinstance(video
, dict) and video
.get('@type') == 'VideoObject':
1474 extract_video_object(video
)
1475 if expected_type
is None:
1479 return dict((k
, v
) for k
, v
in info
.items() if v
is not None)
1482 def _hidden_inputs(html
):
1483 html
= re
.sub(r
'<!--(?:(?!<!--).)*-->', '', html
)
1485 for input in re
.findall(r
'(?i)(<input[^>]+>)', html
):
1486 attrs
= extract_attributes(input)
1489 if attrs
.get('type') not in ('hidden', 'submit'):
1491 name
= attrs
.get('name') or attrs
.get('id')
1492 value
= attrs
.get('value')
1493 if name
and value
is not None:
1494 hidden_inputs
[name
] = value
1495 return hidden_inputs
1497 def _form_hidden_inputs(self
, form_id
, html
):
1498 form
= self
._search
_regex
(
1499 r
'(?is)<form[^>]+?id=(["\'])%s\
1[^
>]*>(?P
<form
>.+?
)</form
>' % form_id,
1500 html, '%s form
' % form_id, group='form
')
1501 return self._hidden_inputs(form)
1504 regex = r' *((?P
<reverse
>\
+)?
(?P
<field
>[a
-zA
-Z0
-9_]+)((?P
<separator
>[~
:])(?P
<limit
>.*?
))?
)?
*$
'
1506 default = ('hidden
', 'aud_or_vid
', 'hasvid
', 'ie_pref
', 'lang
', 'quality
',
1507 'res
', 'fps
', 'codec
:vp9
.2
', 'size
', 'br
', 'asr
',
1508 'proto
', 'ext
', 'hasaud
', 'source
', 'format_id
') # These must not be aliases
1509 ytdl_default = ('hasaud
', 'lang
', 'quality
', 'tbr
', 'filesize
', 'vbr
',
1510 'height
', 'width
', 'proto
', 'vext
', 'abr
', 'aext
',
1511 'fps
', 'fs_approx
', 'source
', 'format_id
')
1514 'vcodec
': {'type': 'ordered
', 'regex
': True,
1515 'order
': ['av0?
1', 'vp0?
9.2', 'vp0?
9', '[hx
]265|he?vc?
', '[hx
]264|avc
', 'vp0?
8', 'mp4v|h263
', 'theora
', '', None, 'none
']},
1516 'acodec
': {'type': 'ordered
', 'regex
': True,
1517 'order
': ['opus
', 'vorbis
', 'aac
', 'mp?
4a?
', 'mp3
', 'e?a?c
-?
3', 'dts
', '', None, 'none
']},
1518 'proto
': {'type': 'ordered
', 'regex
': True, 'field
': 'protocol
',
1519 'order
': ['(ht|f
)tps
', '(ht|f
)tp$
', 'm3u8
.+', '.*dash
', 'ws|websocket
', '', 'mms|rtsp
', 'none
', 'f4
']},
1520 'vext
': {'type': 'ordered
', 'field
': 'video_ext
',
1521 'order
': ('mp4
', 'webm
', 'flv
', '', 'none
'),
1522 'order_free
': ('webm
', 'mp4
', 'flv
', '', 'none
')},
1523 'aext
': {'type': 'ordered
', 'field
': 'audio_ext
',
1524 'order
': ('m4a
', 'aac
', 'mp3
', 'ogg
', 'opus
', 'webm
', '', 'none
'),
1525 'order_free
': ('opus
', 'ogg
', 'webm
', 'm4a
', 'mp3
', 'aac
', '', 'none
')},
1526 'hidden
': {'visible': False, 'forced': True, 'type': 'extractor', 'max': -1000},
1527 'aud_or_vid
': {'visible
': False, 'forced
': True, 'type': 'multiple
',
1528 'field
': ('vcodec
', 'acodec
'),
1529 'function
': lambda it: int(any(v != 'none
' for v in it))},
1530 'ie_pref
': {'priority': True, 'type': 'extractor'},
1531 'hasvid
': {'priority': True, 'field': 'vcodec', 'type': 'boolean', 'not_in_list': ('none',)},
1532 'hasaud
': {'field': 'acodec', 'type': 'boolean', 'not_in_list': ('none',)},
1533 'lang
': {'convert': 'ignore', 'field': 'language_preference'},
1534 'quality
': {'convert': 'float_none', 'default': -1},
1535 'filesize
': {'convert': 'bytes'},
1536 'fs_approx
': {'convert': 'bytes', 'field': 'filesize_approx'},
1537 'id': {'convert': 'string', 'field': 'format_id'},
1538 'height
': {'convert': 'float_none'},
1539 'width
': {'convert': 'float_none'},
1540 'fps
': {'convert': 'float_none'},
1541 'tbr
': {'convert': 'float_none'},
1542 'vbr
': {'convert': 'float_none'},
1543 'abr
': {'convert': 'float_none'},
1544 'asr
': {'convert': 'float_none'},
1545 'source
': {'convert': 'ignore', 'field': 'source_preference'},
1547 'codec
': {'type': 'combined', 'field': ('vcodec', 'acodec')},
1548 'br
': {'type': 'combined', 'field': ('tbr', 'vbr', 'abr'), 'same_limit': True},
1549 'size
': {'type': 'combined', 'same_limit': True, 'field': ('filesize', 'fs_approx')},
1550 'ext
': {'type': 'combined', 'field': ('vext', 'aext')},
1551 'res
': {'type': 'multiple
', 'field
': ('height
', 'width
'),
1552 'function
': lambda it: (lambda l: min(l) if l else 0)(tuple(filter(None, it)))},
1554 # Most of these exist only for compatibility reasons
1555 'dimension
': {'type': 'alias', 'field': 'res'},
1556 'resolution
': {'type': 'alias', 'field': 'res'},
1557 'extension
': {'type': 'alias', 'field': 'ext'},
1558 'bitrate
': {'type': 'alias', 'field': 'br'},
1559 'total_bitrate
': {'type': 'alias', 'field': 'tbr'},
1560 'video_bitrate
': {'type': 'alias', 'field': 'vbr'},
1561 'audio_bitrate
': {'type': 'alias', 'field': 'abr'},
1562 'framerate
': {'type': 'alias', 'field': 'fps'},
1563 'language_preference
': {'type': 'alias', 'field': 'lang'}, # not named as 'language
' because such a field exists
1564 'protocol
': {'type': 'alias', 'field': 'proto'},
1565 'source_preference
': {'type': 'alias', 'field': 'source'},
1566 'filesize_approx
': {'type': 'alias', 'field': 'fs_approx'},
1567 'filesize_estimate
': {'type': 'alias', 'field': 'size'},
1568 'samplerate
': {'type': 'alias', 'field': 'asr'},
1569 'video_ext
': {'type': 'alias', 'field': 'vext'},
1570 'audio_ext
': {'type': 'alias', 'field': 'aext'},
1571 'video_codec
': {'type': 'alias', 'field': 'vcodec'},
1572 'audio_codec
': {'type': 'alias', 'field': 'acodec'},
1573 'video
': {'type': 'alias', 'field': 'hasvid'},
1574 'has_video
': {'type': 'alias', 'field': 'hasvid'},
1575 'audio
': {'type': 'alias', 'field': 'hasaud'},
1576 'has_audio
': {'type': 'alias', 'field': 'hasaud'},
1577 'extractor
': {'type': 'alias', 'field': 'ie_pref'},
1578 'preference
': {'type': 'alias', 'field': 'ie_pref'},
1579 'extractor_preference
': {'type': 'alias', 'field': 'ie_pref'},
1580 'format_id
': {'type': 'alias', 'field': 'id'},
1585 def _get_field_setting(self, field, key):
1586 if field not in self.settings:
1587 self.settings[field] = {}
1588 propObj = self.settings[field]
1589 if key not in propObj:
1590 type = propObj.get('type')
1592 default = 'preference
' if type == 'extractor
' else (field,) if type in ('combined
', 'multiple
') else field
1593 elif key == 'convert
':
1594 default = 'order
' if type == 'ordered
' else 'float_string
' if field else 'ignore
'
1596 default = {'type': 'field', 'visible': True, 'order': [], 'not_in_list': (None,)}.get(key, None)
1597 propObj[key] = default
1600 def _resolve_field_value(self, field, value, convertNone=False):
1605 value = value.lower()
1606 conversion = self._get_field_setting(field, 'convert
')
1607 if conversion == 'ignore
':
1609 if conversion == 'string
':
1611 elif conversion == 'float_none
':
1612 return float_or_none(value)
1613 elif conversion == 'bytes':
1614 return FileDownloader.parse_bytes(value)
1615 elif conversion == 'order
':
1616 order_list = (self._use_free_order and self._get_field_setting(field, 'order_free
')) or self._get_field_setting(field, 'order
')
1617 use_regex = self._get_field_setting(field, 'regex
')
1618 list_length = len(order_list)
1619 empty_pos = order_list.index('') if '' in order_list else list_length + 1
1620 if use_regex and value is not None:
1621 for i, regex in enumerate(order_list):
1622 if regex and re.match(regex, value):
1623 return list_length - i
1624 return list_length - empty_pos # not in list
1625 else: # not regex or value = None
1626 return list_length - (order_list.index(value) if value in order_list else empty_pos)
1628 if value.isnumeric():
1631 self.settings[field]['convert
'] = 'string
'
1634 def evaluate_params(self, params, sort_extractor):
1635 self._use_free_order = params.get('prefer_free_formats
', False)
1636 self._sort_user = params.get('format_sort
', [])
1637 self._sort_extractor = sort_extractor
1639 def add_item(field, reverse, closest, limit_text):
1640 field = field.lower()
1641 if field in self._order:
1643 self._order.append(field)
1644 limit = self._resolve_field_value(field, limit_text)
1647 'closest
': False if limit is None else closest,
1648 'limit_text
': limit_text,
1650 if field in self.settings:
1651 self.settings[field].update(data)
1653 self.settings[field] = data
1656 tuple(field for field in self.default if self._get_field_setting(field, 'forced
'))
1657 + (tuple() if params.get('format_sort_force
', False)
1658 else tuple(field for field in self.default if self._get_field_setting(field, 'priority
')))
1659 + tuple(self._sort_user) + tuple(sort_extractor) + self.default)
1661 for item in sort_list:
1662 match = re.match(self.regex, item)
1664 raise ExtractorError('Invalid format sort string
"%s" given by extractor
' % item)
1665 field = match.group('field
')
1668 if self._get_field_setting(field, 'type') == 'alias
':
1669 field = self._get_field_setting(field, 'field
')
1670 reverse = match.group('reverse
') is not None
1671 closest = match.group('separator
') == '~
'
1672 limit_text = match.group('limit
')
1674 has_limit = limit_text is not None
1675 has_multiple_fields = self._get_field_setting(field, 'type') == 'combined
'
1676 has_multiple_limits = has_limit and has_multiple_fields and not self._get_field_setting(field, 'same_limit
')
1678 fields = self._get_field_setting(field, 'field
') if has_multiple_fields else (field,)
1679 limits = limit_text.split(":") if has_multiple_limits else (limit_text,) if has_limit else tuple()
1680 limit_count = len(limits)
1681 for (i, f) in enumerate(fields):
1682 add_item(f, reverse, closest,
1683 limits[i] if i < limit_count
1684 else limits[0] if has_limit and not has_multiple_limits
1687 def print_verbose_info(self, write_debug):
1689 write_debug('Sort order given by user
: %s' % ', '.join(self._sort_user))
1690 if self._sort_extractor:
1691 write_debug('Sort order given by extractor
: %s' % ', '.join(self._sort_extractor))
1692 write_debug('Formats
sorted by
: %s' % ', '.join(['%s%s%s' % (
1693 '+' if self._get_field_setting(field, 'reverse
') else '', field,
1694 '%s%s(%s)' % ('~
' if self._get_field_setting(field, 'closest
') else ':',
1695 self._get_field_setting(field, 'limit_text
'),
1696 self._get_field_setting(field, 'limit
'))
1697 if self._get_field_setting(field, 'limit_text
') is not None else '')
1698 for field in self._order if self._get_field_setting(field, 'visible
')]))
1700 def _calculate_field_preference_from_value(self, format, field, type, value):
1701 reverse = self._get_field_setting(field, 'reverse
')
1702 closest = self._get_field_setting(field, 'closest
')
1703 limit = self._get_field_setting(field, 'limit
')
1705 if type == 'extractor
':
1706 maximum = self._get_field_setting(field, 'max')
1707 if value is None or (maximum is not None and value >= maximum):
1709 elif type == 'boolean
':
1710 in_list = self._get_field_setting(field, 'in_list
')
1711 not_in_list = self._get_field_setting(field, 'not_in_list
')
1712 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
1713 elif type == 'ordered
':
1714 value = self._resolve_field_value(field, value, True)
1716 # try to convert to number
1717 val_num = float_or_none(value, default=self._get_field_setting(field, 'default
'))
1718 is_num = self._get_field_setting(field, 'convert
') != 'string
' and val_num is not None
1722 return ((-10, 0) if value is None
1723 else (1, value, 0) if not is_num # if a field has mixed strings and numbers, strings are sorted higher
1724 else (0, -abs(value - limit), value - limit if reverse else limit - value) if closest
1725 else (0, value, 0) if not reverse and (limit is None or value <= limit)
1726 else (0, -value, 0) if limit is None or (reverse and value == limit) or value > limit
1727 else (-1, value, 0))
1729 def _calculate_field_preference(self, format, field):
1730 type = self._get_field_setting(field, 'type') # extractor, boolean, ordered, field, multiple
1731 get_value = lambda f: format.get(self._get_field_setting(f, 'field
'))
1732 if type == 'multiple
':
1733 type = 'field
' # Only 'field
' is allowed in multiple for now
1734 actual_fields = self._get_field_setting(field, 'field
')
1736 value = self._get_field_setting(field, 'function
')(get_value(f) for f in actual_fields)
1738 value = get_value(field)
1739 return self._calculate_field_preference_from_value(format, field, type, value)
1741 def calculate_preference(self, format):
1742 # Determine missing protocol
1743 if not format.get('protocol
'):
1744 format['protocol
'] = determine_protocol(format)
1746 # Determine missing ext
1747 if not format.get('ext
') and 'url
' in format:
1748 format['ext
'] = determine_ext(format['url
'])
1749 if format.get('vcodec
') == 'none
':
1750 format['audio_ext
'] = format['ext
'] if format.get('acodec
') != 'none
' else 'none
'
1751 format['video_ext
'] = 'none
'
1753 format['video_ext
'] = format['ext
']
1754 format['audio_ext
'] = 'none
'
1755 # if format.get('preference
') is None and format.get('ext
') in ('f4f
', 'f4m
'): # Not supported?
1756 # format['preference
'] = -1000
1758 # Determine missing bitrates
1759 if format.get('tbr
') is None:
1760 if format.get('vbr
') is not None and format.get('abr
') is not None:
1761 format['tbr
'] = format.get('vbr
', 0) + format.get('abr
', 0)
1763 if format.get('vcodec
') != "none" and format.get('vbr
') is None:
1764 format['vbr
'] = format.get('tbr
') - format.get('abr
', 0)
1765 if format.get('acodec
') != "none" and format.get('abr
') is None:
1766 format['abr
'] = format.get('tbr
') - format.get('vbr
', 0)
1768 return tuple(self._calculate_field_preference(format, field) for field in self._order)
1770 def _sort_formats(self, formats, field_preference=[]):
1773 format_sort = self.FormatSort() # params and to_screen are taken from the downloader
1774 format_sort.evaluate_params(self._downloader.params, field_preference)
1775 if self.get_param('verbose
', False):
1776 format_sort.print_verbose_info(self._downloader.write_debug)
1777 formats.sort(key=lambda f: format_sort.calculate_preference(f))
1779 def _check_formats(self, formats, video_id):
1781 formats[:] = filter(
1782 lambda f: self._is_valid_url(
1784 item='%s video format
' % f.get('format_id
') if f.get('format_id
') else 'video
'),
1788 def _remove_duplicate_formats(formats):
1792 if f['url
'] not in format_urls:
1793 format_urls.add(f['url
'])
1794 unique_formats.append(f)
1795 formats[:] = unique_formats
1797 def _is_valid_url(self, url, video_id, item='video
', headers={}):
1798 url = self._proto_relative_url(url, scheme='http
:')
1799 # For now assume non HTTP(S) URLs always valid
1800 if not (url.startswith('http
://') or url.startswith('https
://')):
1803 self._request_webpage(url, video_id, 'Checking
%s URL
' % item, headers=headers)
1805 except ExtractorError as e:
1807 '%s: %s URL
is invalid
, skipping
: %s'
1808 % (video_id, item, error_to_compat_str(e.cause)))
1811 def http_scheme(self):
1812 """ Either "http:" or "https:", depending on the user's preferences
"""
1815 if self.get_param('prefer_insecure', False)
1818 def _proto_relative_url(self, url, scheme=None):
1821 if url.startswith('//'):
1823 scheme = self.http_scheme()
1828 def _sleep(self, timeout, video_id, msg_template=None):
1829 if msg_template is None:
1830 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1831 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1835 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1836 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1837 fatal=True, m3u8_id=None, data=None, headers={}, query={}):
1838 manifest = self._download_xml(
1839 manifest_url, video_id, 'Downloading f4m manifest',
1840 'Unable to download f4m manifest',
1841 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1842 # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
1843 transform_source=transform_source,
1844 fatal=fatal, data=data, headers=headers, query=query)
1846 if manifest is False:
1849 return self._parse_f4m_formats(
1850 manifest, manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1851 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1853 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, quality=None, f4m_id=None,
1854 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1855 fatal=True, m3u8_id=None):
1856 if not isinstance(manifest, compat_etree_Element) and not fatal:
1859 # currently yt-dlp cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1860 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1861 if akamai_pv is not None and ';' in akamai_pv.text:
1862 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1863 if playerVerificationChallenge.strip() != '':
1867 manifest_version = '1.0'
1868 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1870 manifest_version = '2.0'
1871 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1872 # Remove unsupported DRM protected media from final formats
1873 # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
1874 media_nodes = remove_encrypted_media(media_nodes)
1878 manifest_base_url = get_base_url(manifest)
1880 bootstrap_info = xpath_element(
1881 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1882 'bootstrap info', default=None)
1885 mime_type = xpath_text(
1886 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1887 'base URL', default=None)
1888 if mime_type and mime_type.startswith('audio/'):
1891 for i, media_el in enumerate(media_nodes):
1892 tbr = int_or_none(media_el.attrib.get('bitrate'))
1893 width = int_or_none(media_el.attrib.get('width'))
1894 height = int_or_none(media_el.attrib.get('height'))
1895 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1896 # If <bootstrapInfo> is present, the specified f4m is a
1897 # stream-level manifest, and only set-level manifests may refer to
1898 # external resources. See section 11.4 and section 4 of F4M spec
1899 if bootstrap_info is None:
1901 # @href is introduced in 2.0, see section 11.6 of F4M spec
1902 if manifest_version == '2.0':
1903 media_url = media_el.attrib.get('href')
1904 if media_url is None:
1905 media_url = media_el.attrib.get('url')
1909 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1910 else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1911 # If media_url is itself a f4m manifest do the recursive extraction
1912 # since bitrates in parent manifest (this one) and media_url manifest
1913 # may differ leading to inability to resolve the format by requested
1914 # bitrate in f4m downloader
1915 ext = determine_ext(manifest_url)
1917 f4m_formats = self._extract_f4m_formats(
1918 manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id,
1919 transform_source=transform_source, fatal=fatal)
1920 # Sometimes stream-level manifest contains single media entry that
1921 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1922 # At the same time parent's media entry in set-level manifest may
1923 # contain it. We will copy it from parent in such cases.
1924 if len(f4m_formats) == 1:
1927 'tbr': f.get('tbr') or tbr,
1928 'width': f.get('width') or width,
1929 'height': f.get('height') or height,
1930 'format_id': f.get('format_id') if not tbr else format_id,
1933 formats.extend(f4m_formats)
1936 formats.extend(self._extract_m3u8_formats(
1937 manifest_url, video_id, 'mp4', preference=preference,
1938 quality=quality, m3u8_id=m3u8_id, fatal=fatal))
1941 'format_id': format_id,
1942 'url': manifest_url,
1943 'manifest_url': manifest_url,
1944 'ext': 'flv' if bootstrap_info is not None else None,
1950 'preference': preference,
1955 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, quality=None, m3u8_id=None):
1957 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1961 'preference': preference - 100 if preference else -100,
1963 'resolution': 'multiple',
1964 'format_note': 'Quality selection URL',
1967 def _extract_m3u8_formats(self, *args, **kwargs):
1968 fmts, subs = self._extract_m3u8_formats_and_subtitles(*args, **kwargs)
1970 self.report_warning(bug_reports_message(
1971 "Ignoring subtitle tracks found in the HLS manifest; "
1972 "if any subtitle tracks are missing,"
1976 def _extract_m3u8_formats_and_subtitles(
1977 self, m3u8_url, video_id, ext=None, entry_protocol='m3u8_native',
1978 preference=None, quality=None, m3u8_id=None, note=None,
1979 errnote=None, fatal=True, live=False, data=None, headers={},
1982 res = self._download_webpage_handle(
1984 note='Downloading m3u8 information' if note is None else note,
1985 errnote='Failed to download m3u8 information' if errnote is None else errnote,
1986 fatal=fatal, data=data, headers=headers, query=query)
1991 m3u8_doc, urlh = res
1992 m3u8_url = urlh.geturl()
1994 return self._parse_m3u8_formats_and_subtitles(
1995 m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
1996 preference=preference, quality=quality, m3u8_id=m3u8_id,
1997 note=note, errnote=errnote, fatal=fatal, live=live, data=data,
1998 headers=headers, query=query, video_id=video_id)
2000 def _parse_m3u8_formats_and_subtitles(
2001 self, m3u8_doc, m3u8_url, ext=None, entry_protocol='m3u8_native',
2002 preference=None, quality=None, m3u8_id=None, live=False, note=None,
2003 errnote=None, fatal=True, data=None, headers={}, query={},
2005 formats, subtitles = [], {}
2007 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
2008 return formats, subtitles
2010 has_drm = re.search(r'#EXT-X-SESSION-KEY:.*?URI="skd://', m3u8_doc)
2012 def format_url(url):
2013 return url if re.match(r'^https?://', url) else compat_urlparse.urljoin(m3u8_url, url)
2015 if self.get_param('hls_split_discontinuity', False):
2016 def _extract_m3u8_playlist_indices(manifest_url=None, m3u8_doc=None):
2018 if not manifest_url:
2020 m3u8_doc = self._download_webpage(
2021 manifest_url, video_id, fatal=fatal, data=data, headers=headers,
2022 note=False, errnote='Failed to download m3u8 playlist information')
2023 if m3u8_doc is False:
2025 return range(1 + sum(line.startswith('#EXT-X-DISCONTINUITY') for line in m3u8_doc.splitlines()))
2028 def _extract_m3u8_playlist_indices(*args, **kwargs):
2032 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
2033 # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
2034 # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
2036 # We should try extracting formats only from master playlists [1, 4.3.4],
2037 # i.e. playlists that describe available qualities. On the other hand
2038 # media playlists [1, 4.3.3] should be returned as is since they contain
2039 # just the media without qualities renditions.
2040 # Fortunately, master playlist can be easily distinguished from media
2041 # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
2042 # master playlist tags MUST NOT appear in a media playlist and vice versa.
2043 # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
2044 # media playlist and MUST NOT appear in master playlist thus we can
2045 # clearly detect media playlist with this criterion.
2047 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
2049 'format_id': '-'.join(map(str, filter(None, [m3u8_id, idx]))),
2050 'format_index': idx,
2053 'protocol': entry_protocol,
2054 'preference': preference,
2057 } for idx in _extract_m3u8_playlist_indices(m3u8_doc=m3u8_doc)]
2059 return formats, subtitles
2062 last_stream_inf = {}
2064 def extract_media(x_media_line):
2065 media = parse_m3u8_attributes(x_media_line)
2066 # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
2067 media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME')
2068 if not (media_type and group_id and name):
2070 groups.setdefault(group_id, []).append(media)
2071 # <https://tools.ietf.org/html/rfc8216#section-4.3.4.1>
2072 if media_type == 'SUBTITLES':
2073 # According to RFC 8216 §4.3.4.2.1, URI is REQUIRED in the
2074 # EXT-X-MEDIA tag if the media type is SUBTITLES.
2075 # However, lack of URI has been spotted in the wild.
2076 # e.g. NebulaIE; see https://github.com/yt-dlp/yt-dlp/issues/339
2077 if not media.get('URI'):
2079 url = format_url(media['URI'])
2082 'ext': determine_ext(url),
2084 if sub_info['ext'] == 'm3u8':
2085 # Per RFC 8216 §3.1, the only possible subtitle format m3u8
2086 # files may contain is WebVTT:
2087 # <https://tools.ietf.org/html/rfc8216#section-3.1>
2088 sub_info['ext'] = 'vtt'
2089 sub_info['protocol'] = 'm3u8_native'
2090 lang = media.get('LANGUAGE') or 'und'
2091 subtitles.setdefault(lang, []).append(sub_info)
2092 if media_type not in ('VIDEO', 'AUDIO'):
2094 media_url = media.get('URI')
2096 manifest_url = format_url(media_url)
2098 'format_id': '-'.join(map(str, filter(None, (m3u8_id, group_id, name, idx)))),
2099 'format_note': name,
2100 'format_index': idx,
2101 'url': manifest_url,
2102 'manifest_url': m3u8_url,
2103 'language': media.get('LANGUAGE'),
2105 'protocol': entry_protocol,
2106 'preference': preference,
2108 'vcodec': 'none' if media_type == 'AUDIO' else None,
2109 } for idx in _extract_m3u8_playlist_indices(manifest_url))
2111 def build_stream_name():
2112 # Despite specification does not mention NAME attribute for
2113 # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
2114 # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
2115 # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
2116 stream_name = last_stream_inf.get('NAME')
2119 # If there is no NAME in EXT-X-STREAM-INF it will be obtained
2120 # from corresponding rendition group
2121 stream_group_id = last_stream_inf.get('VIDEO')
2122 if not stream_group_id:
2124 stream_group = groups.get(stream_group_id)
2125 if not stream_group:
2126 return stream_group_id
2127 rendition = stream_group[0]
2128 return rendition.get('NAME') or stream_group_id
2130 # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
2131 # chance to detect video only formats when EXT-X-STREAM-INF tags
2132 # precede EXT-X-MEDIA tags in HLS manifest such as [3].
2133 for line in m3u8_doc.splitlines():
2134 if line.startswith('#EXT-X-MEDIA:'):
2137 for line in m3u8_doc.splitlines():
2138 if line.startswith('#EXT-X-STREAM-INF:'):
2139 last_stream_inf = parse_m3u8_attributes(line)
2140 elif line.startswith('#') or not line.strip():
2143 tbr = float_or_none(
2144 last_stream_inf.get('AVERAGE-BANDWIDTH')
2145 or last_stream_inf.get('BANDWIDTH'), scale=1000)
2146 manifest_url = format_url(line.strip())
2148 for idx in _extract_m3u8_playlist_indices(manifest_url):
2149 format_id = [m3u8_id, None, idx]
2150 # Bandwidth of live streams may differ over time thus making
2151 # format_id unpredictable. So it's better to keep provided
2154 stream_name = build_stream_name()
2155 format_id[1] = stream_name if stream_name else '%d' % (tbr if tbr else len(formats))
2157 'format_id': '-'.join(map(str, filter(None, format_id))),
2158 'format_index': idx,
2159 'url': manifest_url,
2160 'manifest_url': m3u8_url,
2163 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')),
2164 'protocol': entry_protocol,
2165 'preference': preference,
2168 resolution = last_stream_inf.get('RESOLUTION')
2170 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
2172 f['width'] = int(mobj.group('width'))
2173 f['height'] = int(mobj.group('height'))
2174 # Unified Streaming Platform
2176 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
2178 abr, vbr = mobj.groups()
2179 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
2184 codecs = parse_codecs(last_stream_inf.get('CODECS'))
2186 audio_group_id = last_stream_inf.get('AUDIO')
2187 # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
2188 # references a rendition group MUST have a CODECS attribute.
2189 # However, this is not always respected, for example, [2]
2190 # contains EXT-X-STREAM-INF tag which references AUDIO
2191 # rendition group but does not have CODECS and despite
2192 # referencing an audio group it represents a complete
2193 # (with audio and video) format. So, for such cases we will
2194 # ignore references to rendition groups and treat them
2195 # as complete formats.
2196 if audio_group_id and codecs and f.get('vcodec') != 'none':
2197 audio_group = groups.get(audio_group_id)
2198 if audio_group and audio_group[0].get('URI'):
2199 # TODO: update acodec for audio only formats with
2201 f['acodec'] = 'none'
2202 if not f.get('ext'):
2203 f['ext'] = 'm4a' if f.get('vcodec') == 'none' else 'mp4'
2207 progressive_uri = last_stream_inf.get('PROGRESSIVE-URI')
2210 del http_f['manifest_url']
2212 'format_id': f['format_id'].replace('hls-', 'http-'),
2214 'url': progressive_uri,
2216 formats.append(http_f)
2218 last_stream_inf = {}
2219 return formats, subtitles
2222 def _xpath_ns(path, namespace=None):
2226 for c in path.split('/'):
2227 if not c or c == '.':
2230 out.append('{%s}%s' % (namespace, c))
2231 return '/'.join(out)
2233 def _extract_smil_formats_and_subtitles(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
2234 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
2240 namespace = self._parse_smil_namespace(smil)
2242 fmts = self._parse_smil_formats(
2243 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
2244 subs = self._parse_smil_subtitles(
2245 smil, namespace=namespace)
2249 def _extract_smil_formats(self, *args, **kwargs):
2250 fmts, subs = self._extract_smil_formats_and_subtitles(*args, **kwargs)
2252 self.report_warning(bug_reports_message(
2253 "Ignoring subtitle tracks found in the SMIL manifest; "
2254 "if any subtitle tracks are missing,"
2258 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
2259 smil = self._download_smil(smil_url, video_id, fatal=fatal)
2262 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
2264 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
2265 return self._download_xml(
2266 smil_url, video_id, 'Downloading SMIL file',
2267 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
2269 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
2270 namespace = self._parse_smil_namespace(smil)
2272 formats = self._parse_smil_formats(
2273 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
2274 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
2276 video_id = os.path.splitext(url_basename(smil_url))[0]
2280 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
2281 name = meta.attrib.get('name')
2282 content = meta.attrib.get('content')
2283 if not name or not content:
2285 if not title and name == 'title':
2287 elif not description and name in ('description', 'abstract'):
2288 description = content
2289 elif not upload_date and name == 'date':
2290 upload_date = unified_strdate(content)
2293 'id': image.get('type'),
2294 'url': image.get('src'),
2295 'width': int_or_none(image.get('width')),
2296 'height': int_or_none(image.get('height')),
2297 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
2301 'title': title or video_id,
2302 'description': description,
2303 'upload_date': upload_date,
2304 'thumbnails': thumbnails,
2306 'subtitles': subtitles,
2309 def _parse_smil_namespace(self, smil):
2310 return self._search_regex(
2311 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
2313 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
2315 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
2316 b = meta.get('base') or meta.get('httpBase')
2327 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
2328 for medium in media:
2329 src = medium.get('src')
2330 if not src or src in srcs:
2334 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
2335 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
2336 width = int_or_none(medium.get('width'))
2337 height = int_or_none(medium.get('height'))
2338 proto = medium.get('proto')
2339 ext = medium.get('ext')
2340 src_ext = determine_ext(src)
2341 streamer = medium.get('streamer') or base
2343 if proto == 'rtmp' or streamer.startswith('rtmp'):
2349 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
2351 'filesize': filesize,
2355 if transform_rtmp_url:
2356 streamer, src = transform_rtmp_url(streamer, src)
2357 formats[-1].update({
2363 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
2364 src_url = src_url.strip()
2366 if proto == 'm3u8' or src_ext == 'm3u8':
2367 m3u8_formats = self._extract_m3u8_formats(
2368 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
2369 if len(m3u8_formats) == 1:
2371 m3u8_formats[0].update({
2372 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
2377 formats.extend(m3u8_formats)
2378 elif src_ext == 'f4m':
2383 'plugin': 'flowplayer-3.2.0.1',
2385 f4m_url += '&' if '?' in f4m_url else '?'
2386 f4m_url += compat_urllib_parse_urlencode(f4m_params)
2387 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
2388 elif src_ext == 'mpd':
2389 formats.extend(self._extract_mpd_formats(
2390 src_url, video_id, mpd_id='dash', fatal=False))
2391 elif re.search(r'\.ism/[Mm]anifest', src_url):
2392 formats.extend(self._extract_ism_formats(
2393 src_url, video_id, ism_id='mss', fatal=False))
2394 elif src_url.startswith('http') and self._is_valid_url(src, video_id):
2398 'ext': ext or src_ext or 'flv',
2399 'format_id': 'http-%d' % (bitrate or http_count),
2401 'filesize': filesize,
2408 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
2411 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
2412 src = textstream.get('src')
2413 if not src or src in urls:
2416 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
2417 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
2418 subtitles.setdefault(lang, []).append({
2424 def _extract_xspf_playlist(self, xspf_url, playlist_id, fatal=True):
2425 xspf = self._download_xml(
2426 xspf_url, playlist_id, 'Downloading xpsf playlist',
2427 'Unable to download xspf manifest', fatal=fatal)
2430 return self._parse_xspf(
2431 xspf, playlist_id, xspf_url=xspf_url,
2432 xspf_base_url=base_url(xspf_url))
2434 def _parse_xspf(self, xspf_doc, playlist_id, xspf_url=None, xspf_base_url=None):
2436 'xspf': 'http://xspf.org/ns/0/',
2437 's1': 'http://static.streamone.nl/player/ns/0',
2441 for track in xspf_doc.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
2443 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
2444 description = xpath_text(
2445 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
2446 thumbnail = xpath_text(
2447 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
2448 duration = float_or_none(
2449 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
2452 for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP)):
2453 format_url = urljoin(xspf_base_url, location.text)
2458 'manifest_url': xspf_url,
2459 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
2460 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
2461 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
2463 self._sort_formats(formats)
2468 'description': description,
2469 'thumbnail': thumbnail,
2470 'duration': duration,
2475 def _extract_mpd_formats(self, *args, **kwargs):
2476 fmts, subs = self._extract_mpd_formats_and_subtitles(*args, **kwargs)
2478 self.report_warning(bug_reports_message(
2479 "Ignoring subtitle tracks found in the DASH manifest; "
2480 "if any subtitle tracks are missing,"
2484 def _extract_mpd_formats_and_subtitles(
2485 self, mpd_url, video_id, mpd_id=None, note=None, errnote=None,
2486 fatal=True, data=None, headers={}, query={}):
2487 res = self._download_xml_handle(
2489 note='Downloading MPD manifest' if note is None else note,
2490 errnote='Failed to download MPD manifest' if errnote is None else errnote,
2491 fatal=fatal, data=data, headers=headers, query=query)
2497 mpd_base_url = base_url(urlh.geturl())
2499 return self._parse_mpd_formats_and_subtitles(
2500 mpd_doc, mpd_id, mpd_base_url, mpd_url)
2502 def _parse_mpd_formats(self, *args, **kwargs):
2503 fmts, subs = self._parse_mpd_formats_and_subtitles(*args, **kwargs)
2505 self.report_warning(bug_reports_message(
2506 "Ignoring subtitle tracks found in the DASH manifest; "
2507 "if any subtitle tracks are missing,"
2511 def _parse_mpd_formats_and_subtitles(
2512 self, mpd_doc, mpd_id=None, mpd_base_url='', mpd_url=None):
2514 Parse formats
from MPD manifest
.
2516 1. MPEG
-DASH Standard
, ISO
/IEC
23009-1:2014(E
),
2517 http
://standards
.iso
.org
/ittf
/PubliclyAvailableStandards
/c065274_ISO_IEC_23009
-1_2014.zip
2518 2. https
://en
.wikipedia
.org
/wiki
/Dynamic_Adaptive_Streaming_over_HTTP
2520 if not self.get_param('dynamic_mpd', True):
2521 if mpd_doc.get('type') == 'dynamic':
2524 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
2527 return self._xpath_ns(path, namespace)
2529 def is_drm_protected(element):
2530 return element.find(_add_ns('ContentProtection')) is not None
2532 def extract_multisegment_info(element, ms_parent_info):
2533 ms_info = ms_parent_info.copy()
2535 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
2536 # common attributes and elements. We will only extract relevant
2538 def extract_common(source):
2539 segment_timeline = source.find(_add_ns('SegmentTimeline'))
2540 if segment_timeline is not None:
2541 s_e = segment_timeline.findall(_add_ns('S'))
2543 ms_info['total_number'] = 0
2546 r = int(s.get('r', 0))
2547 ms_info['total_number'] += 1 + r
2548 ms_info['s'].append({
2549 't': int(s.get('t', 0)),
2550 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
2551 'd': int(s.attrib['d']),
2554 start_number = source.get('startNumber')
2556 ms_info['start_number'] = int(start_number)
2557 timescale = source.get('timescale')
2559 ms_info['timescale'] = int(timescale)
2560 segment_duration = source.get('duration')
2561 if segment_duration:
2562 ms_info['segment_duration'] = float(segment_duration)
2564 def extract_Initialization(source):
2565 initialization = source.find(_add_ns('Initialization'))
2566 if initialization is not None:
2567 ms_info['initialization_url'] = initialization.attrib['sourceURL']
2569 segment_list = element.find(_add_ns('SegmentList'))
2570 if segment_list is not None:
2571 extract_common(segment_list)
2572 extract_Initialization(segment_list)
2573 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
2575 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
2577 segment_template = element.find(_add_ns('SegmentTemplate'))
2578 if segment_template is not None:
2579 extract_common(segment_template)
2580 media = segment_template.get('media')
2582 ms_info['media'] = media
2583 initialization = segment_template.get('initialization')
2585 ms_info['initialization'] = initialization
2587 extract_Initialization(segment_template)
2590 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
2591 formats, subtitles = [], {}
2592 stream_numbers = {'audio': 0, 'video': 0}
2593 for period in mpd_doc.findall(_add_ns('Period')):
2594 period_duration = parse_duration(period.get('duration')) or mpd_duration
2595 period_ms_info = extract_multisegment_info(period, {
2599 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
2600 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
2601 for representation in adaptation_set.findall(_add_ns('Representation')):
2602 representation_attrib = adaptation_set.attrib.copy()
2603 representation_attrib.update(representation.attrib)
2604 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
2605 mime_type = representation_attrib['mimeType']
2606 content_type = representation_attrib.get('contentType', mime_type.split('/')[0])
2608 codecs = representation_attrib.get('codecs', '')
2609 if content_type not in ('video', 'audio', 'text'):
2610 if mime_type == 'image/jpeg':
2611 content_type = mime_type
2612 elif codecs.split('.')[0] == 'stpp':
2613 content_type = 'text'
2615 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
2619 for element in (representation, adaptation_set, period, mpd_doc):
2620 base_url_e = element.find(_add_ns('BaseURL'))
2621 if base_url_e is not None:
2622 base_url = base_url_e.text + base_url
2623 if re.match(r'^https?://', base_url):
2625 if mpd_base_url and base_url.startswith('/'):
2626 base_url = compat_urlparse.urljoin(mpd_base_url, base_url)
2627 elif mpd_base_url and not re.match(r'^https?://', base_url):
2628 if not mpd_base_url.endswith('/'):
2630 base_url = mpd_base_url + base_url
2631 representation_id = representation_attrib.get('id')
2632 lang = representation_attrib.get('lang')
2633 url_el = representation.find(_add_ns('BaseURL'))
2634 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
2635 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
2636 if representation_id is not None:
2637 format_id = representation_id
2639 format_id = content_type
2641 format_id = mpd_id + '-' + format_id
2642 if content_type in ('video', 'audio'):
2644 'format_id': format_id,
2645 'manifest_url': mpd_url,
2646 'ext': mimetype2ext(mime_type),
2647 'width': int_or_none(representation_attrib.get('width')),
2648 'height': int_or_none(representation_attrib.get('height')),
2649 'tbr': float_or_none(bandwidth, 1000),
2650 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
2651 'fps': int_or_none(representation_attrib.get('frameRate')),
2652 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
2653 'format_note': 'DASH %s' % content_type,
2654 'filesize': filesize,
2655 'container': mimetype2ext(mime_type) + '_dash',
2656 'manifest_stream_number': stream_numbers[content_type]
2658 f.update(parse_codecs(codecs))
2659 stream_numbers[content_type] += 1
2660 elif content_type == 'text':
2662 'ext': mimetype2ext(mime_type),
2663 'manifest_url': mpd_url,
2664 'filesize': filesize,
2666 elif content_type == 'image/jpeg':
2667 # See test case in VikiIE
2668 # https://www.viki.com/videos/1175236v-choosing-spouse-by-lottery-episode-1
2670 'format_id': format_id,
2672 'manifest_url': mpd_url,
2673 'format_note': 'DASH storyboards (jpeg)',
2677 if is_drm_protected(adaptation_set) or is_drm_protected(representation):
2679 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
2681 def prepare_template(template_name, identifiers):
2682 tmpl = representation_ms_info[template_name]
2683 # First of, % characters outside $...$ templates
2684 # must be escaped by doubling for proper processing
2685 # by % operator string formatting used further (see
2686 # https://github.com/ytdl-org/youtube-dl/issues/16867).
2692 in_template = not in_template
2693 elif c == '%' and not in_template:
2695 # Next, $...$ templates are translated to their
2696 # %(...) counterparts to be used with % operator
2697 if representation_id is not None:
2698 t = t.replace('$RepresentationID$', representation_id)
2699 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
2700 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
2701 t.replace('$$', '$')
2704 # @initialization is a regular template like @media one
2705 # so it should be handled just the same way (see
2706 # https://github.com/ytdl-org/youtube-dl/issues/11605)
2707 if 'initialization' in representation_ms_info:
2708 initialization_template = prepare_template(
2710 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
2711 # $Time$ shall not be included for @initialization thus
2712 # only $Bandwidth$ remains
2714 representation_ms_info['initialization_url'] = initialization_template % {
2715 'Bandwidth': bandwidth,
2718 def location_key(location):
2719 return 'url' if re.match(r'^https?://', location) else 'path'
2721 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
2723 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
2724 media_location_key = location_key(media_template)
2726 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
2727 # can't be used at the same time
2728 if '%(Number' in media_template and 's' not in representation_ms_info:
2729 segment_duration = None
2730 if 'total_number' not in representation_ms_info and 'segment_duration' in representation_ms_info:
2731 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
2732 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
2733 representation_ms_info['fragments'] = [{
2734 media_location_key: media_template % {
2735 'Number': segment_number,
2736 'Bandwidth': bandwidth,
2738 'duration': segment_duration,
2739 } for segment_number in range(
2740 representation_ms_info['start_number'],
2741 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
2743 # $Number*$ or $Time$ in media template with S list available
2744 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
2745 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
2746 representation_ms_info['fragments'] = []
2749 segment_number = representation_ms_info['start_number']
2751 def add_segment_url():
2752 segment_url = media_template % {
2753 'Time': segment_time,
2754 'Bandwidth': bandwidth,
2755 'Number': segment_number,
2757 representation_ms_info['fragments'].append({
2758 media_location_key: segment_url,
2759 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
2762 for num, s in enumerate(representation_ms_info['s']):
2763 segment_time = s.get('t') or segment_time
2767 for r in range(s.get('r', 0)):
2768 segment_time += segment_d
2771 segment_time += segment_d
2772 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
2774 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
2775 # or any YouTube dashsegments video
2778 timescale = representation_ms_info['timescale']
2779 for s in representation_ms_info['s']:
2780 duration = float_or_none(s['d'], timescale)
2781 for r in range(s.get('r', 0) + 1):
2782 segment_uri = representation_ms_info['segment_urls'][segment_index]
2784 location_key(segment_uri): segment_uri,
2785 'duration': duration,
2788 representation_ms_info['fragments'] = fragments
2789 elif 'segment_urls' in representation_ms_info:
2790 # Segment URLs with no SegmentTimeline
2791 # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
2792 # https://github.com/ytdl-org/youtube-dl/pull/14844
2794 segment_duration = float_or_none(
2795 representation_ms_info['segment_duration'],
2796 representation_ms_info['timescale']) if 'segment_duration' in representation_ms_info else None
2797 for segment_url in representation_ms_info['segment_urls']:
2799 location_key(segment_url): segment_url,
2801 if segment_duration:
2802 fragment['duration'] = segment_duration
2803 fragments.append(fragment)
2804 representation_ms_info['fragments'] = fragments
2805 # If there is a fragments key available then we correctly recognized fragmented media.
2806 # Otherwise we will assume unfragmented media with direct access. Technically, such
2807 # assumption is not necessarily correct since we may simply have no support for
2808 # some forms of fragmented media renditions yet, but for now we'll use this fallback.
2809 if 'fragments' in representation_ms_info:
2811 # NB: mpd_url may be empty when MPD manifest is parsed from a string
2812 'url': mpd_url or base_url,
2813 'fragment_base_url': base_url,
2815 'protocol': 'http_dash_segments' if mime_type != 'image/jpeg' else 'mhtml',
2817 if 'initialization_url' in representation_ms_info:
2818 initialization_url = representation_ms_info['initialization_url']
2819 if not f.get('url'):
2820 f['url'] = initialization_url
2821 f['fragments'].append({location_key(initialization_url): initialization_url})
2822 f['fragments'].extend(representation_ms_info['fragments'])
2824 # Assuming direct URL to unfragmented media.
2826 if content_type in ('video', 'audio') or mime_type == 'image/jpeg':
2828 elif content_type == 'text':
2829 subtitles.setdefault(lang or 'und', []).append(f)
2831 return formats, subtitles
2833 def _extract_ism_formats(self, *args, **kwargs):
2834 fmts, subs = self._extract_ism_formats_and_subtitles(*args, **kwargs)
2836 self.report_warning(bug_reports_message(
2837 "Ignoring subtitle tracks found in the ISM manifest; "
2838 "if any subtitle tracks are missing,"
2842 def _extract_ism_formats_and_subtitles(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
2843 res = self._download_xml_handle(
2845 note='Downloading ISM manifest' if note is None else note,
2846 errnote='Failed to download ISM manifest' if errnote is None else errnote,
2847 fatal=fatal, data=data, headers=headers, query=query)
2854 return self._parse_ism_formats_and_subtitles(ism_doc, urlh.geturl(), ism_id)
2856 def _parse_ism_formats_and_subtitles(self, ism_doc, ism_url, ism_id=None):
2858 Parse formats
from ISM manifest
.
2860 1. [MS
-SSTR
]: Smooth Streaming Protocol
,
2861 https
://msdn
.microsoft
.com
/en
-us
/library
/ff469518
.aspx
2863 if ism_doc.get('IsLive') == 'TRUE':
2866 duration = int(ism_doc.attrib['Duration'])
2867 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
2871 for stream in ism_doc.findall('StreamIndex'):
2872 stream_type = stream.get('Type')
2873 if stream_type not in ('video', 'audio', 'text'):
2875 url_pattern = stream.attrib['Url']
2876 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
2877 stream_name = stream.get('Name')
2878 stream_language = stream.get('Language', 'und')
2879 for track in stream.findall('QualityLevel'):
2880 fourcc = track.get('FourCC') or ('AACL' if track.get('AudioTag') == '255' else None)
2881 # TODO: add support for WVC1 and WMAP
2882 if fourcc not in ('H264', 'AVC1', 'AACL', 'TTML'):
2883 self.report_warning('%s is not a supported codec' % fourcc)
2885 tbr = int(track.attrib['Bitrate']) // 1000
2886 # [1] does not mention Width and Height attributes. However,
2887 # they're often present while MaxWidth and MaxHeight are
2888 # missing, so should be used as fallbacks
2889 width = int_or_none(track.get('MaxWidth') or track.get('Width'))
2890 height = int_or_none(track.get('MaxHeight') or track.get('Height'))
2891 sampling_rate = int_or_none(track.get('SamplingRate'))
2893 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
2894 track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
2900 stream_fragments = stream.findall('c')
2901 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
2902 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
2903 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
2904 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
2905 if not fragment_ctx['duration']:
2907 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
2909 next_fragment_time = duration
2910 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
2911 for _ in range(fragment_repeat):
2913 'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
2914 'duration': fragment_ctx['duration'] / stream_timescale,
2916 fragment_ctx['time'] += fragment_ctx['duration']
2920 format_id.append(ism_id)
2922 format_id.append(stream_name)
2923 format_id.append(compat_str(tbr))
2925 if stream_type == 'text':
2926 subtitles.setdefault(stream_language, []).append({
2930 'manifest_url': ism_url,
2931 'fragments': fragments,
2932 '_download_params': {
2933 'stream_type': stream_type,
2934 'duration': duration,
2935 'timescale': stream_timescale,
2937 'language': stream_language,
2938 'codec_private_data': track.get('CodecPrivateData'),
2941 elif stream_type in ('video', 'audio'):
2943 'format_id': '-'.join(format_id),
2945 'manifest_url': ism_url,
2946 'ext': 'ismv' if stream_type == 'video' else 'isma',
2950 'asr': sampling_rate,
2951 'vcodec': 'none' if stream_type == 'audio' else fourcc,
2952 'acodec': 'none' if stream_type == 'video' else fourcc,
2954 'fragments': fragments,
2955 'has_drm': ism_doc.find('Protection') is not None,
2956 '_download_params': {
2957 'stream_type': stream_type,
2958 'duration': duration,
2959 'timescale': stream_timescale,
2960 'width': width or 0,
2961 'height': height or 0,
2963 'language': stream_language,
2964 'codec_private_data': track.get('CodecPrivateData'),
2965 'sampling_rate': sampling_rate,
2966 'channels': int_or_none(track.get('Channels', 2)),
2967 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
2968 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
2971 return formats, subtitles
2973 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):
2974 def absolute_url(item_url):
2975 return urljoin(base_url, item_url)
2977 def parse_content_type(content_type):
2978 if not content_type:
2980 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
2982 mimetype, codecs = ctr.groups()
2983 f = parse_codecs(codecs)
2984 f['ext'] = mimetype2ext(mimetype)
2988 def _media_formats(src, cur_media_type, type_info={}):
2989 full_url = absolute_url(src)
2990 ext = type_info.get('ext') or determine_ext(full_url)
2992 is_plain_url = False
2993 formats = self._extract_m3u8_formats(
2994 full_url, video_id, ext='mp4',
2995 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
2996 preference=preference, quality=quality, fatal=False)
2998 is_plain_url = False
2999 formats = self._extract_mpd_formats(
3000 full_url, video_id, mpd_id=mpd_id, fatal=False)
3005 'vcodec': 'none' if cur_media_type == 'audio' else None,
3007 return is_plain_url, formats
3010 # amp-video and amp-audio are very similar to their HTML5 counterparts
3011 # so we wll include them right here (see
3012 # https://www.ampproject.org/docs/reference/components/amp-video)
3013 # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/
3014 _MEDIA_TAG_NAME_RE = r'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)'
3015 media_tags = [(media_tag, media_tag_name, media_type, '')
3016 for media_tag, media_tag_name, media_type
3017 in re.findall(r'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE, webpage)]
3018 media_tags.extend(re.findall(
3019 # We only allow video|audio followed by a whitespace or '>'.
3020 # Allowing more characters may end up in significant slow down (see
3021 # https://github.com/ytdl-org/youtube-dl/issues/11979, example URL:
3022 # http://www.porntrex.com/maps/videositemap.xml).
3023 r'(?s)(<(?P<tag>%s)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>' % _MEDIA_TAG_NAME_RE, webpage))
3024 for media_tag, _, media_type, media_content in media_tags:
3029 media_attributes = extract_attributes(media_tag)
3030 src = strip_or_none(media_attributes.get('src'))
3032 _, formats = _media_formats(src, media_type)
3033 media_info['formats'].extend(formats)
3034 media_info['thumbnail'] = absolute_url(media_attributes.get('poster'))
3036 for source_tag in re.findall(r'<source[^>]+>', media_content):
3037 s_attr = extract_attributes(source_tag)
3038 # data-video-src and data-src are non standard but seen
3039 # several times in the wild
3040 src = strip_or_none(dict_get(s_attr, ('src', 'data-video-src', 'data-src')))
3043 f = parse_content_type(s_attr.get('type'))
3044 is_plain_url, formats = _media_formats(src, media_type, f)
3046 # width, height, res, label and title attributes are
3047 # all not standard but seen several times in the wild
3050 for lbl in ('label', 'title')
3051 if str_or_none(s_attr.get(lbl))
3053 width = int_or_none(s_attr.get('width'))
3054 height = (int_or_none(s_attr.get('height'))
3055 or int_or_none(s_attr.get('res')))
3056 if not width or not height:
3058 resolution = parse_resolution(lbl)
3061 width = width or resolution.get('width')
3062 height = height or resolution.get('height')
3064 tbr = parse_bitrate(lbl)
3073 'format_id': s_attr.get('label') or s_attr.get('title'),
3075 f.update(formats[0])
3076 media_info['formats'].append(f)
3078 media_info['formats'].extend(formats)
3079 for track_tag in re.findall(r'<track[^>]+>', media_content):
3080 track_attributes = extract_attributes(track_tag)
3081 kind = track_attributes.get('kind')
3082 if not kind or kind in ('subtitles', 'captions'):
3083 src = strip_or_none(track_attributes.get('src'))
3086 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
3087 media_info['subtitles'].setdefault(lang, []).append({
3088 'url': absolute_url(src),
3090 for f in media_info['formats']:
3091 f.setdefault('http_headers', {})['Referer'] = base_url
3092 if media_info['formats'] or media_info['subtitles']:
3093 entries.append(media_info)
3096 def _extract_akamai_formats(self, *args, **kwargs):
3097 fmts, subs = self._extract_akamai_formats_and_subtitles(*args, **kwargs)
3099 self.report_warning(bug_reports_message(
3100 "Ignoring subtitle tracks found in the manifests; "
3101 "if any subtitle tracks are missing,"
3105 def _extract_akamai_formats_and_subtitles(self, manifest_url, video_id, hosts={}):
3106 signed = 'hdnea=' in manifest_url
3108 # https://learn.akamai.com/en-us/webhelp/media-services-on-demand/stream-packaging-user-guide/GUID-BE6C0F73-1E06-483B-B0EA-57984B91B7F9.html
3109 manifest_url = re.sub(
3110 r'(?:b=[\d,-]+|(?:__a__|attributes)=off|__b__=\d+)&?',
3111 '', manifest_url).strip('?')
3116 hdcore_sign = 'hdcore=3.7.0'
3117 f4m_url = re.sub(r'(https?://[^/]+)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
3118 hds_host = hosts.get('hds')
3120 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
3121 if 'hdcore=' not in f4m_url:
3122 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
3123 f4m_formats = self._extract_f4m_formats(
3124 f4m_url, video_id, f4m_id='hds', fatal=False)
3125 for entry in f4m_formats:
3126 entry.update({'extra_param_to_segment_url': hdcore_sign})
3127 formats.extend(f4m_formats)
3129 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
3130 hls_host = hosts.get('hls')
3132 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
3133 m3u8_formats, m3u8_subtitles = self._extract_m3u8_formats_and_subtitles(
3134 m3u8_url, video_id, 'mp4', 'm3u8_native',
3135 m3u8_id='hls', fatal=False)
3136 formats.extend(m3u8_formats)
3137 subtitles = self._merge_subtitles(subtitles, m3u8_subtitles)
3139 http_host = hosts.get('http')
3140 if http_host and m3u8_formats and not signed:
3141 REPL_REGEX = r'https?://[^/]+/i/([^,]+),([^/]+),([^/]+)\.csmil/.+'
3142 qualities = re.match(REPL_REGEX, m3u8_url).group(2).split(',')
3143 qualities_length = len(qualities)
3144 if len(m3u8_formats) in (qualities_length, qualities_length + 1):
3146 for f in m3u8_formats:
3147 if f['vcodec'] != 'none':
3148 for protocol in ('http', 'https'):
3150 del http_f['manifest_url']
3152 REPL_REGEX, protocol + r'://%s/\g<1>%s\3' % (http_host, qualities[i]), f['url'])
3154 'format_id': http_f['format_id'].replace('hls-', protocol + '-'),
3156 'protocol': protocol,
3158 formats.append(http_f)
3161 return formats, subtitles
3163 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
3164 query = compat_urlparse.urlparse(url).query
3165 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
3167 r'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url)
3168 url_base = mobj.group('url')
3169 http_base_url = '%s%s:%s' % ('http', mobj.group('s') or '', url_base)
3172 def manifest_url(manifest):
3173 m_url = '%s/%s' % (http_base_url, manifest)
3175 m_url += '?%s' % query
3178 if 'm3u8' not in skip_protocols:
3179 formats.extend(self._extract_m3u8_formats(
3180 manifest_url('playlist.m3u8'), video_id, 'mp4',
3181 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
3182 if 'f4m' not in skip_protocols:
3183 formats.extend(self._extract_f4m_formats(
3184 manifest_url('manifest.f4m'),
3185 video_id, f4m_id='hds', fatal=False))
3186 if 'dash' not in skip_protocols:
3187 formats.extend(self._extract_mpd_formats(
3188 manifest_url('manifest.mpd'),
3189 video_id, mpd_id='dash', fatal=False))
3190 if re.search(r'(?:/smil:|\.smil)', url_base):
3191 if 'smil' not in skip_protocols:
3192 rtmp_formats = self._extract_smil_formats(
3193 manifest_url('jwplayer.smil'),
3194 video_id, fatal=False)
3195 for rtmp_format in rtmp_formats:
3196 rtsp_format = rtmp_format.copy()
3197 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
3198 del rtsp_format['play_path']
3199 del rtsp_format['ext']
3200 rtsp_format.update({
3201 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
3202 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
3205 formats.extend([rtmp_format, rtsp_format])
3207 for protocol in ('rtmp', 'rtsp'):
3208 if protocol not in skip_protocols:
3210 'url': '%s:%s' % (protocol, url_base),
3211 'format_id': protocol,
3212 'protocol': protocol,
3216 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
3218 r'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)(?!</script>).*?\.setup\s*\((?P<options>[^)]+)\)',
3222 jwplayer_data = self._parse_json(mobj.group('options'),
3224 transform_source=transform_source)
3225 except ExtractorError:
3228 if isinstance(jwplayer_data, dict):
3229 return jwplayer_data
3231 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
3232 jwplayer_data = self._find_jwplayer_data(
3233 webpage, video_id, transform_source=js_to_json)
3234 return self._parse_jwplayer_data(
3235 jwplayer_data, video_id, *args, **kwargs)
3237 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
3238 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3239 # JWPlayer backward compatibility: flattened playlists
3240 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
3241 if 'playlist' not in jwplayer_data:
3242 jwplayer_data = {'playlist': [jwplayer_data]}
3246 # JWPlayer backward compatibility: single playlist item
3247 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
3248 if not isinstance(jwplayer_data['playlist'], list):
3249 jwplayer_data['playlist'] = [jwplayer_data['playlist']]
3251 for video_data in jwplayer_data['playlist']:
3252 # JWPlayer backward compatibility: flattened sources
3253 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
3254 if 'sources' not in video_data:
3255 video_data['sources'] = [video_data]
3257 this_video_id = video_id or video_data['mediaid']
3259 formats = self._parse_jwplayer_formats(
3260 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
3261 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
3264 tracks = video_data.get('tracks')
3265 if tracks and isinstance(tracks, list):
3266 for track in tracks:
3267 if not isinstance(track, dict):
3269 track_kind = track.get('kind')
3270 if not track_kind or not isinstance(track_kind, compat_str):
3272 if track_kind.lower() not in ('captions', 'subtitles'):
3274 track_url = urljoin(base_url, track.get('file'))
3277 subtitles.setdefault(track.get('label') or 'en', []).append({
3278 'url': self._proto_relative_url(track_url)
3282 'id': this_video_id,
3283 'title': unescapeHTML(video_data['title'] if require_title else video_data.get('title')),
3284 'description': clean_html(video_data.get('description')),
3285 'thumbnail': urljoin(base_url, self._proto_relative_url(video_data.get('image'))),
3286 'timestamp': int_or_none(video_data.get('pubdate')),
3287 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
3288 'subtitles': subtitles,
3290 # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
3291 if len(formats) == 1 and re.search(r'^(?:http|//).*(?:youtube\.com|youtu\.be)/.+', formats[0]['url']):
3293 '_type': 'url_transparent',
3294 'url': formats[0]['url'],
3297 self._sort_formats(formats)
3298 entry['formats'] = formats
3299 entries.append(entry)
3300 if len(entries) == 1:
3303 return self.playlist_result(entries)
3305 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
3306 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
3309 for source in jwplayer_sources_data:
3310 if not isinstance(source, dict):
3312 source_url = urljoin(
3313 base_url, self._proto_relative_url(source.get('file')))
3314 if not source_url or source_url in urls:
3316 urls.append(source_url)
3317 source_type = source.get('type') or ''
3318 ext = mimetype2ext(source_type) or determine_ext(source_url)
3319 if source_type == 'hls' or ext == 'm3u8':
3320 formats.extend(self._extract_m3u8_formats(
3321 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
3322 m3u8_id=m3u8_id, fatal=False))
3323 elif source_type == 'dash' or ext == 'mpd':
3324 formats.extend(self._extract_mpd_formats(
3325 source_url, video_id, mpd_id=mpd_id, fatal=False))
3327 formats.extend(self._extract_smil_formats(
3328 source_url, video_id, fatal=False))
3329 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
3330 elif source_type.startswith('audio') or ext in (
3331 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
3338 height = int_or_none(source.get('height'))
3340 # Often no height is provided but there is a label in
3341 # format like "1080p", "720p SD", or 1080.
3342 height = int_or_none(self._search_regex(
3343 r'^(\d{3,4})[pP]?(?:\b|$)', compat_str(source.get('label') or ''),
3344 'height', default=None))
3347 'width': int_or_none(source.get('width')),
3349 'tbr': int_or_none(source.get('bitrate')),
3352 if source_url.startswith('rtmp'):
3353 a_format['ext'] = 'flv'
3354 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
3355 # of jwplayer.flash.swf
3356 rtmp_url_parts = re.split(
3357 r'((?:mp4|mp3|flv):)', source_url, 1)
3358 if len(rtmp_url_parts) == 3:
3359 rtmp_url, prefix, play_path = rtmp_url_parts
3362 'play_path': prefix + play_path,
3365 a_format.update(rtmp_params)
3366 formats.append(a_format)
3369 def _live_title(self, name):
3370 """ Generate the title
for a live video
"""
3371 now = datetime.datetime.now()
3372 now_str = now.strftime('%Y-%m-%d %H:%M')
3373 return name + ' ' + now_str
3375 def _int(self, v, name, fatal=False, **kwargs):
3376 res = int_or_none(v, **kwargs)
3377 if 'get_attr' in kwargs:
3378 print(getattr(v, kwargs['get_attr']))
3380 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
3382 raise ExtractorError(msg)
3384 self.report_warning(msg)
3387 def _float(self, v, name, fatal=False, **kwargs):
3388 res = float_or_none(v, **kwargs)
3390 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
3392 raise ExtractorError(msg)
3394 self.report_warning(msg)
3397 def _set_cookie(self, domain, name, value, expire_time=None, port=None,
3398 path='/', secure=False, discard=False, rest={}, **kwargs):
3399 cookie = compat_cookiejar_Cookie(
3400 0, name, value, port, port is not None, domain, True,
3401 domain.startswith('.'), path, True, secure, expire_time,
3402 discard, None, None, rest)
3403 self._downloader.cookiejar.set_cookie(cookie)
3405 def _get_cookies(self, url):
3406 """ Return a compat_cookies_SimpleCookie
with the cookies
for the url
"""
3407 req = sanitized_Request(url)
3408 self._downloader.cookiejar.add_cookie_header(req)
3409 return compat_cookies_SimpleCookie(req.get_header('Cookie'))
3411 def _apply_first_set_cookie_header(self, url_handle, cookie):
3413 Apply first Set
-Cookie header instead of the last
. Experimental
.
3415 Some
sites (e
.g
. [1-3]) may serve two cookies under the same name
3416 in Set
-Cookie header
and expect the
first (old
) one to be
set rather
3417 than
second (new
). However
, as of RFC6265 the newer one cookie
3418 should be
set into cookie store what actually happens
.
3419 We will workaround this issue by resetting the cookie to
3420 the first one manually
.
3421 1. https
://new
.vk
.com
/
3422 2. https
://github
.com
/ytdl
-org
/youtube
-dl
/issues
/9841#issuecomment-227871201
3423 3. https
://learning
.oreilly
.com
/
3425 for header, cookies in url_handle.headers.items():
3426 if header.lower() != 'set-cookie':
3428 if sys.version_info[0] >= 3:
3429 cookies = cookies.encode('iso-8859-1')
3430 cookies = cookies.decode('utf-8')
3431 cookie_value = re.search(
3432 r'%s=(.+?);.*?\b[Dd]omain=(.+?)(?:[,;]|$)' % cookie, cookies)
3434 value, domain = cookie_value.groups()
3435 self._set_cookie(domain, cookie, value)
3438 def get_testcases(self, include_onlymatching=False):
3439 t = getattr(self, '_TEST', None)
3441 assert not hasattr(self, '_TESTS'), \
3442 '%s has _TEST and _TESTS' % type(self).__name__
3445 tests = getattr(self, '_TESTS', [])
3447 if not include_onlymatching and t.get('only_matching', False):
3449 t['name'] = type(self).__name__[:-len('IE')]
3452 def is_suitable(self, age_limit):
3453 """ Test whether the extractor
is generally suitable
for the given
3454 age
limit (i
.e
. pornographic sites are
not, all others usually are
) """
3456 any_restricted = False
3457 for tc in self.get_testcases(include_onlymatching=False):
3458 if tc.get('playlist', []):
3459 tc = tc['playlist'][0]
3460 is_restricted = age_restricted(
3461 tc.get('info_dict', {}).get('age_limit'), age_limit)
3462 if not is_restricted:
3464 any_restricted = any_restricted or is_restricted
3465 return not any_restricted
3467 def extract_subtitles(self, *args, **kwargs):
3468 if (self.get_param('writesubtitles', False)
3469 or self.get_param('listsubtitles')):
3470 return self._get_subtitles(*args, **kwargs)
3473 def _get_subtitles(self, *args, **kwargs):
3474 raise NotImplementedError('This method must be implemented by subclasses')
3477 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
3478 """ Merge subtitle items
for one language
. Items
with duplicated URLs
3479 will be dropped
. """
3480 list1_urls = set([item['url'] for item in subtitle_list1])
3481 ret = list(subtitle_list1)
3482 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
3486 def _merge_subtitles(cls, *dicts, target=None):
3487 """ Merge subtitle dictionaries
, language by language
. """
3491 for lang, subs in d.items():
3492 target[lang] = cls._merge_subtitle_items(target.get(lang, []), subs)
3495 def extract_automatic_captions(self, *args, **kwargs):
3496 if (self.get_param('writeautomaticsub', False)
3497 or self.get_param('listsubtitles')):
3498 return self._get_automatic_captions(*args, **kwargs)
3501 def _get_automatic_captions(self, *args, **kwargs):
3502 raise NotImplementedError('This method must be implemented by subclasses')
3504 def mark_watched(self, *args, **kwargs):
3505 if (self.get_param('mark_watched', False)
3506 and (self._get_login_info()[0] is not None
3507 or self.get_param('cookiefile') is not None)):
3508 self._mark_watched(*args, **kwargs)
3510 def _mark_watched(self, *args, **kwargs):
3511 raise NotImplementedError('This method must be implemented by subclasses')
3513 def geo_verification_headers(self):
3515 geo_verification_proxy = self.get_param('geo_verification_proxy')
3516 if geo_verification_proxy:
3517 headers['Ytdl-request-proxy'] = geo_verification_proxy
3520 def _generic_id(self, url):
3521 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
3523 def _generic_title(self, url):
3524 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
3527 def _availability(is_private=None, needs_premium=None, needs_subscription=None, needs_auth=None, is_unlisted=None):
3528 all_known = all(map(
3529 lambda x: x is not None,
3530 (is_private, needs_premium, needs_subscription, needs_auth, is_unlisted)))
3532 'private' if is_private
3533 else 'premium_only' if needs_premium
3534 else 'subscriber_only' if needs_subscription
3535 else 'needs_auth' if needs_auth
3536 else 'unlisted' if is_unlisted
3537 else 'public' if all_known
3540 def _configuration_arg(self, key, default=NO_DEFAULT, casesense=False):
3542 @returns A list of values for the extractor argument given by "key"
3543 or "default" if no such key is present
3544 @param default The default value to return when the key is not present (default: [])
3545 @param casesense When false, the values are converted to lower case
3548 self._downloader.params, ('extractor_args', self.ie_key().lower(), key))
3550 return [] if default is NO_DEFAULT else default
3551 return list(val) if casesense else [x.lower() for x in val]
3554 class SearchInfoExtractor(InfoExtractor):
3556 Base
class for paged search queries extractors
.
3557 They accept URLs
in the format
_SEARCH_KEY(|all|
[0-9]):{query}
3558 Instances should define _SEARCH_KEY
and _MAX_RESULTS
.
3562 def _make_valid_url(cls):
3563 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
3566 def suitable(cls, url):
3567 return re.match(cls._make_valid_url(), url) is not None
3569 def _real_extract(self, query):
3570 mobj = re.match(self._make_valid_url(), query)
3572 raise ExtractorError('Invalid search query "%s"' % query)
3574 prefix = mobj.group('prefix')
3575 query = mobj.group('query')
3577 return self._get_n_results(query, 1)
3578 elif prefix == 'all':
3579 return self._get_n_results(query, self._MAX_RESULTS)
3583 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
3584 elif n > self._MAX_RESULTS:
3585 self.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
3586 n = self._MAX_RESULTS
3587 return self._get_n_results(query, n)
3589 def _get_n_results(self, query, n):
3590 """Get a specified number of results
for a query
"""
3591 raise NotImplementedError('This method must be implemented by subclasses')
3594 def SEARCH_KEY(self):
3595 return self._SEARCH_KEY